International Diabetes Federation - World Congress 2011

December 4-8, 2011; Dubai, United Arab Emirates; Full Report – Published May 17, 2012 – Draft

Executive Highlights

In this final report, we provide our complete coverage of the International Diabetes Federation’s 21st World Congress, held in Dubai, United Arab Emirates from December 4-8, 2011. A record 15,100 attendees from 172 countries packed the Dubai International Exhibition and Convention Centre for the five-day conference, up 21% from 12,500 attendees in Montreal, Canada in 2009. The 2011 World Congress marks the first time the conference was held in the IDF Middle East and North Africa Region, which has six of the world’s top 10 countries for diabetes prevalence – these are nations, unbelievably, where one out of five adults has diabetes. The Middle East and North Africa region of the world will also be one of the hardest hit by diabetes in coming years, with the number of cases expected to rise from 32.8 million to 60 million (83% growth) between 2011 and 2030 according to the IDF’s recently releasedFifth Edition of the Diabetes Atlas (see our November 15, 2011 Closer Look at

IDF 2011 showcased six tracks, 1,500+ speakers, and was primarily focused on drugs and type 2 diabetes – there were virtually no device presentations at this meeting, except for a JDRF-sponsored set of presentations on the last day. Although few new data were shown, we learned a great deal from the thoughtful, diverse perspectives of the speakers and attendees. To help you sort through all of the nuanced learnings, we’ve organized our commentary into nine sections: (1) Incretins; (2) Treatment Algorithms and Combination Therapies; (3) Insulin and Non-Incretin Therapies; (4) Diabetes Technology; (5) Obesity and Bariatric Surgery; (6) Emerging Markets; (7) Diabetes Prevention and Complications; (8) Government and Policy; (9) Exhibit Hall Report. Coverage of select corporate symposia held at IDF 2011 is included within each section. Below, we outline our take on the meeting’s major themes, followed by a comprehensive table of contents for ease of reference.

  • Amidst enthusiasm for the GLP-1 analogs, speakers voiced strong interest in the combination of GLP-1 and basal insulin. While presenters appreciated the theoretically complementary mechanisms of the therapies (targeted postprandial and basal control while minimizing hypoglycemia and weight gain), clinical evidence was cited as well – Dr. J. Hans DeVries (Academic Medical Center, Amsterdam, Netherlands) reprised results of a 52-week study investigating the effects of add-on insulin detemir to patients on background liraglutide, and Dr. John Buse (University of North Carolina, Chapel Hill, NC) highlighted results from the 30-week GWCO study, which showed a strong reduction in A1c (8.3% to 6.7% vs. 8.5% to 7.4% with placebo – impressive, especially given the average diabetes duration of 12 years) when adding exenatide BID to background insulin glargine. Notably, given cost, effects on postprandial glucose, and potential synergy, Dr. Buse suggested the combination could also offer a niche for short-acting versus long-acting GLP-1 agonists. Supporting this theory, Dr. Juris Meier (St. Josef- Hospital, Bochum, Germany), noted that early infusion studies suggest a greater effect on postprandial glucose levels with short-acting agonists, likely due to differences in gastric emptying., We await later-stage clinical evidence with the combination to discern any meaningful difference – an angle we suspect Sanofi may take in the future in hopes of distinguishing lixisenatide from the long-acting agonists. Cost-effectiveness will surely come up as a question aswe believe ultimately some patients will want to take both, especially to stave off rapid acting insulin.
  • We were surprised by the strong support for pharmacologic interventions (especially incretins) for prediabetes and expect to see much more of this going forward, especially as regulatory agencies are pressured to come up with real pathway guidelines. In a headliner debate with Dr. Jaako Tuomilehto (University of Helsinki, Helsinki, Finland), Dr. Ralph DeFronzo (University of Health Science Center, San Antonio, TX) summed up his perspective with an emphatic title slide: “Answer is simple: non-pharmacologic therapy (diet plus exercise) DOES NOT WORK on a long-term basis.” He argued in favor of prediabetes treatment with combination low-dose pioglitazone/metformin, GLP-1 agonists, or DPP-4 inhibitors. Regarding the latter, we took note of his thoughts during Q&A: “I’ve suggested to every pharmaceutical company with a DPP-4 that this is where the future is going.” The very smart Dr. Steven Kahn (University of Washington, Seattle, WA – we love listening to him anytime he speaks) shared this enthusiasm for DPP-4 use in prediabetes, although he more strongly emphasized the need for further study. Regarding GLP-1s for diabetes prevention, Dr. John Buse (University of North Carolina, Chapel Hill, NC) felt liraglutide obesity studies were “quite positive” and noted some excitement for the prediabetes arm in the SCALE trial (estimated primary completion in March 2013; identifier: NCT01272219). We suspect the results from SCALE will fall in line with previous positive results for liraglutide in prediabetes (see page 22 of our TOS 2010 Full Report at Finally, Dr. Luc Van Gaal (Antwerp University Hospital, Antwerp, Belgium) said that while he would wait to make final conclusions, he would “most likely” want to prescribe GLP-1 agonists in prediabetes. Aside from a scarcity of data on the topic, we wonder how payers will think about reimbursement of pharmacotherapy for prediabetes – while population-wide interventions are unlikely to be cost-effective, we suspect forward-looking payers will support early pharmacotherapy in particular high-risk subgroups, especially as more data emerges. And wow if there is ever a time patients would be adherent, we would think this would be it. Combination low-dose pioglitazone and metformin may prove particularly attractive for pre-diabetes as generic pioglitazone is expected to reach the US market by August 2012 although controversy surrounding safety may limit interest to some degree.
  • Treatment algorithms and combination therapies were one of the most frequently debated topics at IDF, with DPP-4s in particular receiving widespread support from speakers. Although we heard a number of arguments supporting second-line therapy (after metformin) using sulfonylureas (cost, rapid efficacy, long track record), TZDs (long-term data, upcoming generic status, addressing insulin resistance), GLP-1s (durability, efficacy, weight loss, low hypoglycemia risk), or insulin (glucose-lowering efficacy), a plurality of speakers seemed to concur that DPP-4 inhibitors were the most ideal second-line therapy. Dr. Bernard Zinman (University of Toronto, Ontario, Canada) went a step further to suggest DPP-4/metformin combination therapy right from diagnosis, citing the robust and durable reductions in A1c, no association with hypoglycemia or weight gain, lower GI side effects, and an enhanced GLP-1 response. Even with this enthusiasm, we heard many lament the lack of data on where certain therapies should be used in the course of type 2 diabetes. With this in mind, we look forward to Dr. David Nathan’s (Massachusetts General, Boston, MA) upcoming four-year GRADE study, which will compare five different diabetes medications combined with metformin in recent onset type 2s. This is taking some time to get going, but should be most interesting; as we understand it, there has been some controversy on combinations. ORIGIN data to be presented at ADA 2012 should shed more light on how early insulin treatment should fit in the treatment paradigm – we look forward to this data and wonder what it will show as most long-term outcomes trials havehad to be extended since number of events has been so low. Finally, given the cardiovascular burden of diabetes, many implied the benefits of shifting away from glucose-centric diabetes treatment towards addressing blood pressure, cholesterol, and smoking. For example, Dr. Wiley Chan discussed Kaiser Permanente’s preference for treating diabetes patients with aspirin, ACE inhibitors, and statins over intensive glycemic control. A1c targets are only chased after patients are on these CV drugs. We are glad to see payers taking a longer-term approach and we hope that this translates into earlier and earlier reimbursement of screening and preventative interventionson the other hand, we certainly would not want to see reduction of microvascular riskminimized.
  • On the insulin front, much of the buzz surrounded Novo Nordisk’s phase 3 candidate insulin degludec, although experts also pushed for broader use of currently available insulin analogs worldwide. In his very well-attended talk on the application of insulin therapy in type 2 diabetes, Dr. Irl Hirsch (University of Washington, Seattle, WA) set the stage, noting that while early initiation of insulin therapy provides substantial glycemic improvements, insulin use remains limited worldwide due to cost and availability issues. Novo Nordisk’s A1CHIEVE study, the largest observational study on insulin analog therapy including about 66,000 people from Asia, Africa, Latin America, and Europe, bolstered the argument for broader insulin analog use – results indicated that insulin analogs could deliver the gains seen in clinical trials. Presenter Dr. Philip Home (Newcastle University, Newcastle, UK) argued that insulin analog use could also “be an opportunity for enhanced lifestyle behaviors… and improved self-management” – though in the Q&A session he did allow that the improvements seen in the trial may have been the result of the patients being given better care due to trial enrollment rather than the specific use of insulin analogs.
  • Diabetes technology occupied a small fraction of the IDF program, though a few speakers reported new updates. Speaking at a Bayer corporate symposium, Dr. Peter Schwarz (University of Dresden, Dresden, Germany) talked about a study comparing patient outcomes on the Contour USB (plug-and-play data downloading) to the Contour (data kept in a handwritten logbook), with results expected in 11-13 months – we hope that the quality-of-life benefits of simpler downloading translate into improved clinical outcomes. We note that the Contour USB has been particularly highly rated among patients with type 2 diabetes; contact for more information on this front. Dr. Schwarz’s co-presenter Dr. Tim Bailey (University of California, San Diego, San Diego, CA) looked forward to the second- generation Contour XT blood glucose meter, which will include built-in data management software, another innovation that is becoming adopted across the Big Four blood glucose meter companies (e.g., the OneTouch Verio Pro). As for continuous glucose monitoring, we especially enjoyed the insights of star and beloved endo Dr. Anne Peters (Keck School of Medicine at the University of Southern California, Los Angeles, CA), who described her three-month, four-visit method of starting patients on CGM. We also attended a symposium on the bleeding edge of artificial pancreas research, with Dr. Aaron Kowalski (JDRF, New York, NY) leading Q&A with some of the leading minds in closed-loop research. Dr. Eric Renard (University of Montpellier, Montpellier, France) emphasized that intraperitoneal delivery and implantable pumps remain active research areas, Dr. Claudio Cobelli (University of Padova, Padova, Italy) discussed ongoing studies on how real-world events affect blood sugar, and Dr. Boris Kovatchev (University of Virginia, Charlottesville, VA) explained the cost savings of mobile phone-based platforms. Also on the topic of mobile devices, Dr. Robert Istepanian (Kingston University, London, United Kingdom) called for IDF to establish a mobile diabetes task force that could set the path of technological development. We expect mobile health to develop especially quickly in the nextseveral years due to the novelty of the field, and we share Dr. Istepanian’s hope that IDF takes a lead in this key area of 21st-century care.
  • The treatment of obesity through pharmacotherapy and bariatric surgery were major discussion topics in Dubai. The great Dr. Nancy Bohannon (St. Luke’s Hospital, San Francisco, CA) presented new data on Vivus’ Qnexa from a CONQUER subanalysis of patients with type 2 diabetes. Participants in the full-dose treatment arm experienced least-squares mean weight loss of 12.1%, compared to 6.6% in the half-dose arm, and 2.8% in the placebo arm. Notably, the 12-month excess weight loss observed with full-dose Qnexa (32.6%) was similar to that observed in two laparoscopic gastric banding studies. While this data was only exploratory, the great efficacy associated with Qnexa was something that came up time and time again at Vivus’ advisory committee on February 22, 2012 – we believe the overwhelmingly positive recommendation was largely associated with the composite profile of such strong efficacy and tolerability. The comparison between bariatric surgery and pharmacotherapy was also addressed in a talk from Dr. Luc Van Gaal (Antwerp University Hospital, Antwerp, Belgium) – he argued that pharmacotherapy will play an important role in curbing the obesity epidemic given the limited population that bariatric surgery can address. Dr. Francesco Rubino (Weill Cornell Medical College, New York, NY) took a different tack, calling bariatric surgery “the most high impact discovery in diabetes since the discovery of insulin” and noting, “If a pill or shot could control glycemia, body weight, improve lipids, blood pressure, and increase survival, would it be acceptable that over 99% of patients do not have access to that treatment?” This fervor was not shared by a number of speakers that emphasized the significant quality of life issues (e.g., lack of freedom to eat, vomiting, vitamin deficiency, osteomalacia, and occasional neurologic syndromes) associated with bariatric surgery. Finally, we noticed that the traditional first-line obesity treatment, lifestyle intervention, was not discussed very broadly at IDF – an exception was a compelling overview of childhood obesity from Dr. Fran Kaufman (Chief Medical Officer, Medtronic Diabetes, Los Angeles, CA) and data from the UK-wide MEND Programme in 10,173 children enrolled in a 24-month lifestyle intervention program.
  • We heard many insights and new data about the particular problems facing diabetes patients in emerging markets – a bigger topic at IDF than at the Western-focused meetings we typically attend and a growing subject of interest in company updates that we hear. One major theme was the developing world’s high expenditure ratios for diabetesi.e., getting diabetes increases someone’s healthcare costs by a much higher multiple than in the developed world. As explained by Dr. Jonathan Brown (Chair, IDF Task Force on Diabetes Health Economics), higher expenditure ratios can reflect lack of early diagnosis, which explains why Africa (where diabetes is typically diagnosed only at advanced stages) has much higher expenditure ratios than China. Availability of diabetes supplies is also a critical issue. Dr. Anant Nigam (CEO, Diabetes India) highlighted two fundamental problems in this regard: the “insulin dilemma” (that insulin often costs more in poor countries than in rich countries due to access issues) and the “insulin paradox” (that more-expensive forms of insulin have replaced less- expensive ones, despite what some characterize as marginal differences in safety and efficacy – we do not see it as such). Governments, companies, and non-profit organizations (such as Dr. Sharad Pensey’s DREAM Trust in Nagpur, India) have made great strides in distributing insulin and other therapies, but the room for improvement is immense.
  • By far the most common oral drugs in emerging markets are metformin and sulfonylureas, while demand for newer therapies is limited largely by cost. Dr. Anoop Misra (Fortis Hospitals, New Delhi, India) was notably positive on DPP-4 inhibitors because of their weight neutrality and benign side effect profile – especially key advantages in India, wheremetabolic syndrome occurs at relatively low BMI and where few physicians are on hand to address adverse events. We heard several Westerners at the meeting express skepticism at the motivations of drug companies, sometimes in inexplicable ways (e.g., some people worried that profit-motivated companies were trying to sell pills for prediabetes, even though a) companies are essentially ignoring this area and b) the public health gains could be massive). By contrast, IDF attendees from emerging markets seemed quite positive on pharmaceutical therapies – as noted above, one of the biggest problems they cited was too few therapies in their home countries. That said, many obstacles remain besides the cost of any particular medication. In a large survey of the Chinese populace, Dr. Jianzhong Xiao (China-Japan Friendship Hospital, Beijing, China) found that fewer than half of people with diabetes were using oral drugs.
  • IDF 2011 gave us a sense of how difficult political change will be to make. While IDF leaders felt that the 2011 UN Summit on Non-Communicable Diseases (NCDs) was an important turning point (speakers compared it to the pivotal UN AIDS summit in 2001), many felt that little actually came of it and intense follow-up will be required. IDF President Dr. Jean Claude Mbanya said that he was “so disappointed [at the summit] when issues of human interest and life and death were considered behind politics.” IDF CEO Dr. Ann Keeling aimed her criticism at the World Health Organization, which did not shift resources after the summit and still has only one person who works specifically on diabetes in its Geneva headquarters of 3,000 people. Drs. Jean Claude Mbanya, Ann Keeling, and Sir George Alleyne (Johns Hopkins University, Baltimore, MD) all emphasized the need to frame the diabetes epidemic in terms of social injustice, likening it to the feminist movement, the civil rights movement, and the fight against AIDS. Of course, these movements were not an easy government fix, just as tackling diabetes will not be – Dr. Paul Zimmet (Baker IDI Heart and Diabetes Institute, Melbourne, Australia) had a notable take: “We have some very simplistic people who think if you ban television advertising and Coca-Cola and some of these other things, you’re going to solve obesity.” Looking to the future, epidemiologist Dr. Nick Wareham (Institute of Metabolic Science, Cambridge, UK) gave a valuable overview of IDF’s Global Diabetes Plan 2011-2021, sharing insights on how we can better integrate diabetes care and prevention. Dr. Mbanya was “very proud” of the plan, which gives a general outline of the key attributes of diabetes care systems. We like that the plan is designed to be simple and enable implementation by community health workers and non-diabetologists, groups that are becoming increasingly important given the wide scope of the diabetes epidemic and the lack of diabetes specialists in many countries. The full document can be downloaded at
Table of Contents 

1. Incretins

Oral Presentations: Late-Breaking Abstracts Session

Efficacy and Safety of Lixisenatide QD Morning and Evening Injections vs Placebo in T2DM Inadequately Controlled on Metformin (GETGOAL-M)

Ronnie Aronson, MD (LMC Endocrinology Centers, Toronto, Canada)

Dr. Aronson presented 24-week data from GetGoal-M, a Sanofi-sponsored study comparing morning and evening doses of lixisenatide to placebo in adult type 2 diabetes patients failing metformin. Participants in the study (n=680) had mean A1c 8.06%, BMI 32.9 kg/m2, age 54.7 years, and diabetes duration 6.1 years. Both lixisenatide arms achieved statistically significant benefits relative to placebo in mean A1c change (-0.9% and -0.8% with lixisenatide vs. -0.4% with placebo), percent of patients requiring rescue therapy, and mean fasting plasma glucose change (with greater FPG benefits seen in AM than PM dosing or placebo: 21.4, 14.4, and 3.0 mg/dl, respectively). Weight loss was slightly, statistically non-significantly greater with Sanofi’s GLP-1 receptor agonist than with placebo. Lixisenatide caused greater rates of nausea, vomiting, and symptomatic hypoglycemia; no severe hypoglycemia was reported. Overall these results are consistent with others from lixisenatide clinical trials, suggesting an efficacy and safety profile that seems at best, comparable with those of exenatide and liraglutide.

  • GetGoal-M was designed to compare morning and evening once-daily doses of lixisenatide to placebo in type 2 diabetes patients failing metformin. Patients in the four-arm, double blind study were randomized in a 3:3:1:1 ratio to receive morning (AM) lixisenatide, evening (PM) lixisenatide, AM placebo, or PM placebo. Lixisenatide dosing was uptitrated in a weekly stepwise pattern (10 ug for one week, 15 ug for the next week, and 20 ug thereafter). Primary efficacy was assessed at 24 weeks, but GetGoal-M also included an additional 52-week follow-on that will be presented at a later date.
  • The study enrolled 680 adults with type 2 diabetes whose A1c was between 7% and 10%, inclusive. Mean baseline characteristics were: A1c 8.06%, BMI 32.9 kg/m2, age 54.7 years, and diabetes duration 6.1 years. Mean weight was roughly 90 kg, and mean metformin dose was nearly at the recommended maximum of 2,000 mg daily. Over 95% of patients in each arm completed the 24-week protocol, and roughly 92% of patients in each lixisenatide arm tolerated the full 20 ug dose. For the presentation of data, the two placebo groups were pooled.
  • Both lixisenatide arms achieved statistically significant benefits relative to placebo in mean A1c change, mean fasting plasma glucose change, and percent of patients requiring rescue therapy. Notably, the change in fasting glucose was greater with AM than PM lixisenatide dosing; Dr. Aronson was unsure of the statistical significance of this difference, and he did not speculate as to its cause. (In general, we have heard that stability in the morning makes glucose control easier throughout the day.) In terms of categorical results, a greater percentage of lixisenatide patients ended the 24-week period with A1c below 7.0%. Mean body weight change was slightly (less than one pound) greater with lixisenatide than placebo.

Lixi AM (n=255)

Lixi PM (n=255)

Placebo (n=170)

LS Mean Change in A1c, %




% Patients Achieving A1c < 7.0%




LS Mean Change in FPG, mg/dl




% Patients Requiring Rescue Therapy




LS Mean Body Weight Change, kg (lbs)

-2.0 (-4.4)

-2.0 (-4.4)

-1.6 (-3.6)

*p<0.05 vs. placebo

  • In a post-meal glucose test conducted only in the AM lixisenatide and AM placebo groups, AM lixisenatide caused significantly lower two-hour postprandial glucose and glucose excursion. Compared to AM placebo, AM lixisenatide gave two-hour post-meal glucose that was lower by 4.51 mmol/l (81.2 mg/dl) and glucose excursion that was lower by 3.88 mmol/l (69.8 mg/dl) (both figures are least-squares mean differences).
  • Rates of overall adverse events and severe adverse events were comparable among all groups, though lixisenatide caused greater GI side effects and symptomatic hypoglycemia (no severe hypoglycemia was observed in any group). Rates of withdrawal due to adverse events were more than twofold greater in the lixisenatide groups relative to placebo.

Adverse Event Prevalence, %

Lixi AM (n=255)

Lixi PM (n=255)

Placebo (n=170)

All Adverse Events




All Severe Adverse Events












Symptomatic Hypoglycemia




Severe Hypoglycemia




Questions and Answers

Q: How would you explain the greater fasting plasma glucose effect in AM vs. PM lixisenatide dosing?

Dr. Aronson: I agree there appears to be a greater effect of AM dosing. We didn’t plan to compare AM to PM dosing in this regard, so I don’t know if it’s statistically significant. It looks as though it may be.

Q: Do you have any information on glucose throughout the day?

Dr. Aronson: Some seven-point glycemic profiles were collected but I don’t have this data to share.


Oral Presentations: Ins and Outs of Incretin Therapies

Beneficial Effects of Insulin and Exendin-4/GLP-1 on the Viability/Apoptosis and Function of GLP-1-Secreting Cells

Camilla Kappe (Karolinska Institute, Solna, Sweden)

Ms. Kappe presented the results of an in vitro study investigating the effects of insulin, exendin-4, and metformin administration on isolated GLP-1-secreting cells. Results suggested both insulin and exendin- 4 were able to stimulate GLP-1 release in the presence of glucose, with increased release following prolonged exposure; exendin-4 was also shown to increase GLP-1 receptor expression. When incubated with palmitate to simulate diabetic hyperlipidemia, both compounds, as well as metformin, protected cells from lipoapoptosis. Given type 2 diabetes patients show decreased GLP-1 levels, Ms. Kappe indicated these results suggest potential additional benefits with the examined compounds through increased L-cell mass and GLP-1 release – though it is still unclear if these results would translate to the in vivo setting.

Questions and Answers

Q: Do you think there is a difference between the GLUTag cells you used and the cells in the human body?

Ms. Kappe: I think GLUTag cells are generally good model. I don’t know if you should say these results are contradictory to previous results. If you have these elevated levels of incretin analogs in the body, you may get some receptor desensitization in the body that may lead to the reversal of this effect.

Dr. Michael Nauck (Diabeteszentrum Bad Lauterberg, Harz, Germany): Usually L-cells have a polarity. I doubt that you can maintain this polarity in cell cultures, though it is important. The response in the body depends on what side you offer the GLP-1. Can you get that type of response in cell culture?

Ms. Kappe: Probably not. Of course there are differences, and these data do need to be confirmed in vivo.

Dr. Nauck: Given GLP-1 is an incretin, if you now say insulin stimulates GLP-1 secretion, that would be a feed forward mechanism. What would the purpose of that be in the body?

Ms. Kappe: The feed forward mechanism is not common, but it does exist elsewhere in the body.


Mechanism of Exenatide-Induced Hypoglycemic Effect in Type 1 Diabetes

Ernest Adeghate, MD, MFM, PhD (United Arab Emirates University, Al Ain, United Arab Emirates)

Dr. Adeghate discussed the results of an in vivo study examining the effects of exenatide on pancreatic islet morphology in an animal model of type 1 diabetes. In the study, exenatide was injected intraperitoneally into normal and diabetic (induced by streptozotocin injection) rats five days a week for 10 weeks. Exenatide treatment increased the distribution of insulin-positive cells in both control and diabetic animals, with a mild increase in the number of cells observed in diabetic rats. Treatment also caused a significant increase in the number of catalase- and glutathione-positive cells (measures of antioxidant activity), suggesting exenatide may protect islet cells by increasing expression of endogenous antioxidants – while we suspect any comparable benefit in patients with type 1 diabetes likely would require administration early in the disease course, we certainly have heard a great deal about benefit from patients with type 1 who have very established type 1 diabetes.

Questions and Answers

Q: Did you measure glucagon at all?

Dr. Adeghate: Not at all. We plan to do it.

Dr. Michael Nauck (Diabetezenstrum Bad Lauterberg, Harz, Germany): One concern one might have is that the beta cells you saw are the result of the anti-apoptotic effect of exenatide treatment.

Dr. Adeghate: If it has anti-apoptotic effects, it may come out that way. We haven’t looked at that.

Dr. Nauck: If you give an insulin secretagogue to cells already stressed by hyperglycemia, you could make cells more susceptible to an immune attack. Probably you did this study in order to suggest some beneficial effect in treating early type 1 diabetes with exenatide.

Dr. Adeghate: That’s a good comment. It might be much more appropriate to look at early type 2 diabetes where you still have many functional beta cells. We’re looking at different animal models soon.


Sitagliptin Exerts an Anti-Inflammatory Effect

Paresh Dandona, MD, PhD (State University of New York at Buffalo, Buffalo, NY)

Dr. Dandona withdrew his presentation from the session. This was very disappointing.


Significantly Better Glycemic Control Achieved with Liraglutide in Patients with the Greatest Estimated Beta-Cell Mass at Baseline

Juris Meier, MD (St. Josef-Hospital, Bochum, Germany)

Dr. Meier presented the results of a pooled analysis examining whether the effects of liraglutide are altered by baseline beta cell mass. In the analysis, subjects from seven of the phase 3 LEAD trials (n=3,995) were stratified into quartiles based on fasting C-peptide/glucose ratios (a surrogate for beta cell mass). As expected given the superiority shown in the LEAD trials, across all quartiles 1.8 mg liraglutide was significantly more effective at lowering A1c from baseline versus comparators (thiazolidinedione, placebo, exenatide, sulfonylurea, sitagliptin). However, the A1c reduction was significantly greater in patients in the highest quartile of beta cell mass compared to patients in the lowest quartile – Dr. Meier suggested these results supported initiation of treatment earlier, as patients of shorter diabetes duration would have a greater chance of achieving glycemic control.

  • Given the progressive loss of beta cells in the development of diabetes, this pooled analysis examined whether the effects of liraglutide were altered by baseline beta cell mass.Twenty-six week data from seven of the phase 3 LEAD trials (n=3,995) were included, with individual subjects stratified into quartiles of baseline beta cell mass – previous studies conducted by Novo Nordisk correlating various measures of beta cell function with beta cell mass observed during pancreatic surgery suggested that the fasting C-peptide/glucose ratio most strongly associated with beta cell area, and thus the metric was used in this study.
  • Across all quartiles, 1.8 mg liraglutide was significantly more effective at lowering A1c from baseline versus comparators (thiazolidinedione, placebo, exenatide, sulfonylurea, sitagliptin). Similarly, a significantly greater percentage of patients achieved target with liraglutide versus comparators across the continuum of beta cell mass – this is more or less expected given the results of the LEAD trials.
  • The A1c reduction with 1.8 mg liraglutide was significantly greater in patients in the highest quartile of beta cell mass compared to patients in the lowest quartile. Likewise, patients in the highest quartile were significantly more likely to achieve target compared to patients in the lowest quartile. This pattern was observed in thiazolidinedione and placebo-treated patients as well, though exenatide, sulfonylurea, and sitagliptin-treated patients showed similar responses regardless of beta cell mass. Dr. Meier suggested these results supported initiation of treatment earlier for liraglutide and thiazolidinediones, with patients of shorter diabetes duration having a greater chance of achieving glycemic control.

Questions and Answers

Q: You showed that even in the lowest quartile, liraglutide had efficacy though sitagliptin lost efficacy quite a bit. Can you explain why?

Dr. Meier: It is difficult to explain. Maybe that reflects the increased GLP-1 concentration achieved with liraglutide, but mechanistically I cannot give you a straight answer.

Q: The patients in LEAD were on different background agents. Did you do a similar experiment before the study drugs were started to examine the effect of C-peptide?

Dr. Meier: That has not been performed, but its something we might look into.

Q: What were those patients where you had an indication to remove the pancreas representative of?

Dr. Meier: That is an excellent question and an important caveat. Those with a healthy pancreas don’t undergo surgery. It was a mixture of patients with benign pancreatic tumors. It’s really a mixed group though, so you’re right. These studies cannot be translated directly to type 2 diabetes, but I think it is something someone might generalize. It was out of necessity.


'Real-World' Clinical Outcomes of Exenatide Bid Compared to Insulin Glargine in Patients with Type 2 Diabetes

Byron Hoogwerf, MD (Eli Lilly, Indianapolis, IN)

Dr. Hoogwerf discussed the results of retrospective analysis aiming to compare clinical outcomes in patients treated with exenatide BID versus insulin glargine in an ambulatory care setting. Patients initiating exenatide (n=2,683) or glargine (n=2,683) from the General Electric electronic medical records database were initially matched to ensure similar baseline characteristics. Results suggested patients initiating exenatide showed a stronger decline in A1c versus insulin glargine (-0.7% vs. -0.4%; baseline 8.1%). Weight was also more greatly reduced, by 2.6 kg (5.7 lbs) versus a 0.2 kg (0.4 lbs) gain with glargine. These trends are consistent with those observed in clinical trials with exenatide, suggesting some translation to the ambulatory setting – while difficult using medical records, we would hope for more data on ease of use and co-administration in this setting as well.

  • This study aimed to compare clinical outcomes in patients treated with exenatide BID versus insulin glargine in an ambulatory care setting. In the study, a retrospective analysis was performed of patients initiating exenatide (n=2,683) or glargine (n=2,683) from the General Electric electronic medical records database; patients were matched to ensure nostatistical differences in baseline mean age, gender, race, rates of diabetes-related complications, and oral anti-diabetic drug use. Patients were excluded if follow-up data was not available from 90-365 days following initiation.
  • Patients initiating exenatide showed a stronger decline in A1c versus insulin glargine (-0.7% vs. -0.4%; baseline 8.1%). This translated to 25% achieving A1c targets compared to 17% with glargine. Weight was also more greatly reduced, by 2.6 kg (5.7 lbs) versus a0.2 kg (0.4 lbs) gain with glargine. Blood pressure showed a slightly greater decline as well, by 2 mmHg compared to 0.1 mmHg with glargine. These trends are consistent with those observed in clinical trials with exenatide, suggesting some translation to the ambulatory setting.
  • Dr. Hoogwerf noted some limitations intrinsic to the study design. These included missing values in the database, the lack of randomization, and under-reporting of hospitalizations and complications (given data was limited to clinical practice) – thus, Dr. Hoogwerf suggested no causal inferences were possible.

Questions and Answers

Q: Was it possible to adjust for the duration of diabetes?

Dr. Hoogwerf: Unfortunately in databases like these we don’t always capture that. We do believe the matching would adjust for that, but we don’t have that specifically.

Comment: I think that the benefits of these compounds for cardiovascular risk are incredibly important to the weight loss. I think it’s fantastic that the actual weight goes down and blood pressure goes down so nicely.

Dr. Hoogwerf: I should comment that there is an analysis that the blood pressure effects are independent of weight reduction.

Q: Any information on hypoglycemia?

Dr. Hoogwerf: One of the difficulties in capturing data in electronic medical records is we only get results from serious hypoglycemia usually, but we didn’t capture that in this particular study.

Q: Do you have any co-administration?

Dr. Hoogwerf: We did not in this study, but we do have studies that show beneficial metabolic effects to the combination. We do not see an increased risk of hypoglycemia versus insulin alone.


Differences in Response Between Exenatide and Liraglutide in the Association of British Clinical Diabetologists (ABCD) Nationwide Audits

Ken Thong, MD (City Hospital, Birmingham, United Kinegdom)

Dr. Thong reported the results of a study comparing data from two nationwide audits conducted by the Association of British Clinical Diabetologists in the UK on exenatide and liraglutide use. In the audits, diabetes centers were invited to provide retrospective data on patients started on exenatide (n=4,589; conducted 2007-2009) or liraglutide (n=1,740; 2009-2011). At three- and six-month follow-up, A1c decline was greater with liraglutide (1.12% and 0.98%; baseline ~9.4%) versus exenatide (0.74% vs. 0.75%; baseline ~9.5%); however, a greater weight reduction was seen with exenatide (4.4 kg [9.7 lbs] and 6.5 kg [14.3 lbs]; baseline ~114 kg [251 lbs]) versus liraglutide (3.0 kg [6.6 lbs] and 3.5 kg [7.7 lbs]; baseline ~111 kg [245 lbs]). Dr. Thong suggested this pattern was likely due to differences in medication use, as many in the exenatide audit discontinued insulin use when exenatide was initiated compared to those in the liraglutide audit who favored reducing the dose – making any comparisons difficult. Information on use proved interesting though, with rates of co-insulin use high at 31% in the exenatide group and 41% in the liraglutide group, which Dr. Thong attributed in the Q&A to increasing confidence with the GLP-1s overtime.

  • This study compared data from two nationwide audits conducted by the Association of British Clinical Diabetologists in the UK on exenatide and liraglutide use. In the audits, diabetes centers were invited to provide retrospective data on patients started on exenatide (n=4,589) or liraglutide (n=1,740); importantly, the audits occurred subsequently (2007-2009 with exenatide; 2009-2011 with liraglutide). Patients were excluded if they had no A1c and weight results at three- and six-month follow-up and from the liraglutide audit if they were on background exenatide or 1.8 mg liraglutide (due to agency guidelines in the UK, which negate the higher dose). While the groups showed statistical differences in age and BMI at baseline, Dr. Thong indicated the differences were likely not clinically significant.
  • At three- and six-month follow-up, A1c decline was greater with liraglutide (1.12% and 0.98%; baseline ~9.4%) versus exenatide (0.74% vs. 0.75%; baseline ~9.5%). In contrast, a greater weight reduction was seen with exenatide (4.4 kg [9.7 lbs] and 6.5 kg [14.3 lbs]; baseline ~114 kg [251 lbs]) versus liraglutide (3.0 kg [6.6 lbs] and 3.5 kg [7.7 lbs]; baseline ~111 kg [245 lbs]). Dr. Thong indicated this pattern was likely due to differences in medication use, however – in the exenatide audit, many discontinued use of thiazolidinediones and insulin when exenatide was initiated, whereas in the liraglutide audit fewer patients stopped insulin treatment instead reducing the dose (overall, rates of co-insulin use were high at 31% in the exenatide group versus 41% in the liraglutide group). In the Q&A, Dr. Thong attributed this to increasing confidence in the GLP-1s, given the differences in the timing of the audits.
  • All reported GI side effects were slightly higher with exenatide (23.7%) versus liraglutide (21.8%). Discontinuation before six months was slightly higher as well, at 14.7% versus 13.0% with liraglutide – discontinuation was primarily due to GI side effects (7.2% with exenatide and 5.1% with liraglutide).

Questions and Answers

Q: Did you see any evidence of GI disease or thyroid disorder?

Dr. Thong: We do plan to provide a presentation on this in a symposium. In terms of thyroid, given there’s a lot of awareness we do get some report of thyroid disorder, but when we clarified it patients had a previous disorder so I don’t think there’s a signal in our database.

Q: Can you speculate on why you have less diabetes treatment reductions with liraglutide versus exenatide?

Dr. Thong: I think we’ve learned in the UK the dangers of stopping insulin when starting exenatide, so that may have trickled down to the liraglutide audit. From personal experience, we’ve seen that physicians felt they were too cautious when they started exenatide as well and they were also rigidly trying to stay in the guidelines. I think people got braver.


Pharmacokinetics, Pharmacodynamics, and Safety of Exenatide Once Weekly Formulation in Chinese Subjects with Type 2 Diabetes

Michael Trautmann, MD (Eli Lilly, Indianapolis, IN)

Dr. Trautmann presented the results of clinical study examining the delivery of exenatide once-weekly (EQW) in native Chinese subjects with type 2 diabetes treated with metformin. Patients (n=26) received weekly injections of 2 mg EQW for 10 weeks, followed by a 10-week washout period. Results suggested a steady state of 300-400 pg/l EQW was achieved by eight weeks of dosing, which was maintained for three weeks following treatment cessation; likewise, mean fasting plasma glucose lowered to a steady state (9.1 mmol/l [163.8 mg/dl] to 6.3 mmol/l [113.4 mg/dl]) at six to eight weeks of dosing. Body weight began to decline after two weeks of dosing, leading to a 4.5 kg (9.9 lbs) loss at 10 weeks of treatment with slight regain to 4.0 kg loss (8.8 lbs) after washout (baseline 69.8 kg [153.8 lbs]). As expected, GI side effects were most frequent, at somewhat higher rates than seen in clinical trials (40% diarrhea, 40% vomiting, 16% nausea) – in the Q&A, Dr. Trautmann suggested this was likely due to the controlled nature of the trial. Overall, he concluded that results were similar to those seen in Western patients, supporting further clinical research in Chinese patients with EQW.

Questions and Answers

Dr. Juris Meier (St. Josef-Hospital, Bochum, Germany: I was surprised by the relatively high nausea rates and diarrhea seen. On the other hand you emphasized the body weight differences. Wouldn’t all this in argue for dose adjustment meaning that in such populations we need a lower dose and maybe escalate the dose in Caucasian patients?

Dr. Trautmann: I don’t think the dose matters for the effect, but the plasma concentrations. I think all the evidence we’ve seen thus far suggests the concentration we achieved gives reasonable side effects and efficacy. I would refrain from judging adverse events in a short study with patients in an institution versus ambulatory, so I would wait for clinical trials. The question remaining is if subjects in these regions are more sensitive to the same GLP-1 concentrations, and that remains to be answered. We have demonstrated that bodyweight only has a minimal contribution to plasma levels – that has allowed for a simple dosing scheme.

Q: To get the full response, some doctors titrate twice a week in early weeks. Do you have some thoughts on the safety of this?

Dr. Trautmann: It is reasonable to expect that if you go more aggressively, you would get faster rise to steady state plasma concentrations. As I showed you, plasma state concentrations begin to decrease in the first week. I really doubt that in a chronic situation like diabetes there is an emergency situation where you would need to reduce blood sugar levels so quickly. The result with slow titration once a week can make the transition easier in terms of side effects. It’s a balance of speed for glucose control and best tolerability.

Comment: I’m a bit puzzled by this study. You recruited a mean baseline at around 7.4% and a BMI of 25 kg/m2. I’m not quite sure what the indication was for exenatide once- weekly in the first place.

Dr. Trautmann: This was a pharmacokinetic study, not a therapeutic study. There were no limitations for patient enrollment except they had to have type 2 diabetes.


Long-Term, Injection-free Treatment with ITCA 650 Leads to Sustained Glycemic Control and Weight Loss in Metformin-Treated Type 2 Diabetes

Robert Henry, MD (University of California at San Diego, La Jolla, CA)

Dr. Henry reprised Dr. Julio Rosenstock’s oral presentation from ADA 2011 (see the June 16, 2011 Closer Look) on phase 2 dose-ranging study results with Intarcia’s ITCA650. As a reminder, ITCA650 is an implantable DUROS device that allows continuous subcutaneous administration of exenatide for six to 12 months. While the results of the presentation were repeated, the submitted abstract elaborated that the 60 mcg/day dose (selected for phase 3 studies) produced changes of -3.1% in total cholesterol, -9.9% in triglycerides, -5.2% in LDL, +0.5% in HDL, -1.71 mmHg in systolic blood pressure, and -7.8 mmHg in diastolic blood pressure at 48 weeks of treatment. Similar to Dr. Rosenstock, in the Q&A Dr. Henry suggested that implantation was easy to perform in-office using an included kit, with few adverse events – we look forward to learning more about the insertion/removal process from both the patient and provider perspective to better assess the clinical and commercial potential of the device. Offhand, this certainly sounds very positive from an adherence perspective. We also believe endos would be positive on the chance to perform procedures, even small ones. Notably, in Q&A, we heard positive things on the stability front.

Questions and Answers

Q: Can you tell us a little bit about the device? Is it the same for three, six, and 12 months of delivery? Do you reload or change it?

Dr. Henry: Depending on the size of the device, it can infuse continuously for a period from three to 12 months. In this case, the device was placed for three months, and then another was inserted for the next three months.

Q: Is there any problem with fibrosis?

Dr. Henry: There was no significant problem with site events. I’ve personally put some in, and they’re very easy to put in and remove. They’re placed in abdominal region with a small kit that is provided with a little anesthetic, a scalpel for a small incision, and an introducer to put it beneath the skin. A Steri-strip is then placed.

Q: When you showed the final A1c, there seemed to be quite some variability in the A1c achieved. Does that correlate with the plasma levels you achieved? Is there a wide spread in concentrations achieved?

Dr. Henry: The exenatide levels during the infusion were very steady, with much less variation than compared to injections. Between subjects, there is data that suggests it’s steady, but I don’t have it here.


Symposium: Incretin Based Therapy: Now and Down the Road

DPP-4 Inhibitors and Glucose Control

Juan Gagliardino, MD (CENEXA, La Plata, Argentina)

Dr. Gagliardino opened the symposium with a discussion of the mechanism of action behind the DPP- inhibitors. He detailed that DPP-4 endogenously breaks down GLP-1 into the 9-36 amide and a His-Ala fragment, and thus that inhibition increases GLP-1 levels. He also noted an animal study in which rats fed a high fructose diet were shown to have increased levels of reactive oxygen species and glycerol, altering glucokinase activity – DPP-4 inhibitor and GLP-1 analog administration were able to counter these effects.

Questions and Answers

Q: Could you comment on the relationship of insulin resistance and incretin dysfunction?

Dr. Gagliardino: The target is dyslipidemia. We know that you can decrease the insulin sensitivity, and as you can see we can significantly modify the triglyceride levels.


GLP-1 Receptor Agonists - Clinical Observations

John Buse, MD, PhD (University of North Carolina, Chapel Hill, NC)

Dr. Buse gave a thorough presentation on the relevant considerations for when using the GLP-1 analogs in clinical practice. For when choosing a treatment, he opened with a review of topline results from head-to-head clinical trials of the short (exenatide twice daily) and long (liraglutide) acting agonists, noting superior effects on A1c, fasting plasma glucose, and GI side effects with liraglutide though less effect on postprandial glucose levels. He also examined the results between liraglutide and longer- acting options (exenatide once weekly and albiglutide) and pointed out that liraglutide was associated with greater reduction in A1C and weight but greater GI adverse events. In the Q&A, while he saw strong potential for once-weekly and once-monthly dosing, he suggested the dose for these agents may not have been optimized in development. Compared to basal insulin, Dr. Buse summarized evidence indicating that the GLP-1 agonists were associated with similar or greater A1C reduction, weight loss instead of weight gain, though similar rates of hypoglycemia and increased rates of GI side effects. However, he voiced a stronger interest in the combination of GLP-1 and basal insulin, highlighting the results of the 30-week GWCO study, which showed a strong reduction in A1c (8.3% to 6.7% vs. 8.5% to 7.4% with placebo – impressive given the average diabetes duration of 12 years and the “failure” of prior insulin therapy) when adding exenatide BID to background insulin glargine which was subsequently optimized. Notably, given cost, effects on postprandial glucose and potential synergy, he suggested the combination could offer a niche for short-acting GLP-1 agonists versus long-acting GLP-1 agonists (“I suspect it might be unique”), though further research is necessary.

  • Dr. Buse opened with a review of topline results from head-to-head clinical trials of the short- and long-acting GLP-1 agonists. Initial comparisons of trials evaluating “short- vs. long-acting” agonists (DURATION-1 and -5 [exenatide vs. exenatide once-weekly]; LEAD-6 [exenatide vs. liraglutide]) suggested superior effects on A1c, fasting plasma glucose, and GI side effects with longer-acting options, though potentially less effect on postprandial glucose levels (likely due to limited effects on gastric emptying with long-acting agents). In trials comparing “long- vs. longer-acting” agonists (DURATION-6 [liraglutide vs. exenatide once-weekly]; HARMONY-7 [liraglutide vs. albiglutide]), the long-acting agents proved superior on A1c and weight but with more GI side effects than longer-acting agents. Dr. Buse indicated these differences could impact treatment selection in clinical practice, though in the Q&A he suggested that dosing for the once-weekly agonists might not have been fully optimized in development, causing the differences in efficacy.
  • Dr. Buse then moved to a review of trials comparing GLP-1 analogs to basal insulin. He summarized evidence indicating greater weight loss and A1c declines with the GLP-1 analogs, with potentially similar rates of hypoglycemia and increased rates of GI side effects (he did note, however, that head-to-head trials of GLP-1 analogs and basal insulin have been criticized for inadequate dosing – fearing hypoglycemia, he suggested investigators may shy from aggressively uptitrating insulin). One of the more notable effects of the GLP-1 analogs versus basal insulin, he suggested, was the potential for the preservation of beta cell function – he highlighted a recent study that indicated a slight improvement in the disposition index of patients treated with GLP-1 therapy versus insulin glargine, though the effect only became apparent after three years of treatment.
  • Dr. Buse voiced strong interest in the combination of GLP-1 analogs and basal insulin. He highlighted the results of the 30-week GWCO study, which examined the effects of exenatide BID versus placebo in patients failing on background insulin glargine – findings suggested a strong decline in A1c (8.3% to 6.7% vs. 8.5% to 7.4% with placebo – impressive given the average diabetes duration of 12 years) and weight (-1.78 kg [3.92 lbs] vs. +0.96 kg [+2.12 lbs]), with similar rates of minor hypoglycemia (25% vs. 29%) but increased GI side effects (41% vs. 8% nausea, 18% vs. 8% diarrhea, 18% vs. 4% vomiting). He also was encouraged by the progress being made in this area, noting ongoing head-to-head trials comparing MDI to GLP-1/basal insulin and the development of a GLP-1/insulin single injection device. He hoped future research would investigate whether there was a niche for short-acting GLP-1 agonists in combination with basal insulin versus long-acting GLP-1 agonists (“I suspect it might be unique”), given the targeted effects on postprandial glucose.
  • Dr. Buse concluded with a summary of the ongoing safety concerns with the GLP-1 analogs. Overall, Dr. Buse suggested the most clinically relevant safety concern currently was renal impairment – as noted in the label, prolonged nausea/vomiting and the subsequent dehydration can lead to acute renal failure, and there is limited experience with the drug in patients with more advanced chronic kidney disease. On pancreatitis, Dr. Buse suggested the concern was mostly “among the media and in settings like this where we lecture on drugs,” with evidence largely based on post-marketing reports. The best data, he suggested, was from large- scale pharmacoepidemiologic studies – while in three such publications there was no significant association, he noted that confidence intervals were wide, indicating increased risk was possible. On thyroid and pancreatic cancer, he indicated that further research was necessary but that to date there was virtually no clinical evidence.

Questions and Answers

Q: Dr. Michael Nauck (Diabeteszentrum Bad Lauterberg, Kirchberg, Germany): With liraglutide and once-weekly agents, I don’t understand the differences in A1c results given both regimens provide relatively stable levels of the GLP-1 agonist. There is relatively little fluctuation. Isn’t it dangerous to compare the treatments based off a single dose given we don’t know that the dose-response relationship was maximized in phase 2 properly?

Dr. Buse: I think the difference is because the dose in longer agents hasn’t been optimized. Unfortunately the way in drug development now is that pharmaceutical companies have to pick a single dose. I think there may be opportunities to develop higher doses for better responses. As someone who is a terrible drug taker for his hypercholesterolemia, I do think there is tremendous opportunity for once-weekly or once-monthly agents.

Q: Do you think the glucose lowering after a meal is due to the beta cell or off beta cell effects?

Dr. Buse: I don’t have data to directly address that question. There was no mechanistic explanation in LEAD-6. There was work done previously by Dr. Orville Kolterman in which a five-day crossover period with patients given NPH insulin at bedtime and exenatide versus placebo showed a rather complete exploration of the various effects of the GLP-1 receptor agonists. The best evidence is that the dominant effect on postprandial glucose in that setting was based on glucagon and gastric emptying versus the insulin releasing effect. That’s a limited population, but my sense is that those are the dominant effects.

Q: Can you comment on use in obese non-diabetics?

Dr. Buse: I think that’s an area of opportunity for exploration. There have been publications regarding the use of liraglutide for the treatment of obesity, which I think is quite positive in regard to the weight loss achieved. The major problem is the regulatory environment – for obesity getting drug approval is a path that’s been less well-paved, and it remains to be seen what hurdles would be in the way of a GLP-1 based therapy for obesity. But I think the evidence suggests they’re as good or better than the other agents approved for weight loss worldwide.

Q: How about lixisenatide?

Dr. Buse: Lixisenatide is basically an analog of exenatide. It’s relatively on the shorter-acting side of the spectrum – it’s longer than exenatide but less then liraglutide. It is administered once a day, but its half- life may suggest that to achieve 24-hour coverage it might ideally be administered twice a day. There are ways to take advantage of a peak and a trough though – there might be less tachyphylaxis of gastric emptying. The phase 3 study with combination with insulin that I think is very interesting will be available very soon.

Q. How about the combination of GLP-1 agonists with DPP-4 inhibitors?

From a theoretical perspective, with a long-acting GLP-1 agonist combined with basal insulin you probably won’t have a further increase in GLP-1 levels, but you would get an effect on GIP, so there is a theoretical hypothesis that you might see a benefit. But with two expensive agents, that combination is not likely to be cost effective.

Q: Can you comment on the use of long-acting versus short-acting GLP-1 analogs?

Dr. Buse: I think they’re both good. For patients in worse control, I think the long-acting analog might be better – that was best shown in the DURATION-1 trial. In very poorly controlled patients, these long- acting agents are extremely powerful glucose lowering agents. For people earlier in their natural history, exenatide BID may be adequate. In particular, the combination of long-acting insulin and rapid-acting GLP-1 might have a unique niche.


Do Current GLP-1 Receptor Agonists Exploit the Full Potential of this Mechanism of Action?

Michael A. Nauck, MD (Diabeteszentrum Bad Lauterberg, Kirchberg, Germany)

Reviewing early studies with GLP-1, Dr. Nauck questioned whether further research could maximize the therapeutic potential of the GLP-1 analogs. In early intravenous infusion studies with GLP-1, Dr. Nauck noted a full normalization of glucose levels in type 2 diabetes patients in the absence of nausea and vomiting. However, with subcutaneous administration, nausea appeared before a full normalization of glucose levels could be achieved – Dr. Nauck compiled results from clinical trials to confirm this relationship between the intensity of decline in fasting glucose levels and nausea rates. With a further understanding of this relationship, he proposed that maximal glucose control could be achieved with the GLP-1 analogs in the absence of nausea.

  • Early studies suggested intravenous administration of GLP-1 could normalize glucose levels without nausea or vomiting. In early infusion studies with GLP-1, Dr. Nauck noted a full normalization of glucose levels in type 2 diabetes patients in the absence of nausea and vomiting, even at lower doses. Dr. Nauck suggested this was matched by slowing in gastric motility, indicating the role of gastric emptying in promoting strong postprandial control.
  • With subcutaneous administration, however, nausea appears before a full normalization of glucose levels can be achieved. When GLP-1 was infused subcutaneously, glucose lowering could be achieved, though not to normal levels and not without increasing nausea. This was confirmed in clinical trials – compiling data from clinical trials with the GLP-1 analogs available and in development, Dr. Nauck noted a significant correlation between the decline in fasting glucose achieved and rates of nausea.
  • Given intravenous administration demonstrated maximal glucose control is theoretically possible without nausea, Dr. Nauck suggested a better understanding of the relationship could maximize the therapeutic potential of the GLP-1 analogs. He posited numerous hypotheses for the differences between intravenous and subcutaneous administration, including modification of GLP-1 when traversing the subcutaneous compartment, activity at potential GLP-1 receptors in the autonomic nervous system, or the triggering of nausea by very high local concentrations in the subcutaneous depot.

Questions and Answers

Q: I know it’s difficult to say the mechanism behind the nausea, but do you know anybody reporting the GI motility effects when GI is infused intravenously?

Dr. Nauck: So the motility effects are there when GLP-1 was used intravenously, but with no nausea reported. I think that is representative of all the trials we’ve looked at.

Q: Do you expect any body weight reduction when GLP-1 is administered intravenously?

Dr. Nauck: That has not been tried because no one would infuse a peptide over a long period of time. My prediction would be yes, based on a visual analog scale of hunger and satiety that was tested with intravenous GLP-1.

Dr. Michael Trautmann (Eli Lilly, Indianapolis, IN): One comment – with a weekly load of exenatide in the subcutaneous tissue, it would not be available to stimulate locally given it is trapped in microspheres. The other thing was that I think in your comparison of fasting glucose achieved versus the rate of nausea one factor was missing: tolerance development for nausea develops so I think we need to factor that in if there’s a way to achieve better tolerability using slow dose escalations to finally get to the most effective plasma concentrations in the end.

Dr. Nauck: I think you’re right; the differences in nausea rates are not fully appreciated in these analyses. I think what you described is probably why the correlation is not perfect.

Dr. John Buse (University of North Carolina, Chapel Hill, NC): Some people seem to be exclusively sensitive to subcutaneous GLP-1 receptor agonists. With the liraglutide pen they’ll have nausea even with the 0.6 mg dose whereas some have no nausea at all. Do you now if any large datasets have been examined to search for characteristics of patients experiencing nausea?

Dr. Nauck: I think it has been tried but unsuccessfully. There is not a definite characteristic, though it could be due to some autonomic neuropathy.

Dr. Buse: True – the easier thing would be to compare those prone to nausea versus those resistant and look for differences.


DPP-4 Inhibitor Therapy of T2DM: Present Status and Future Perspectives

Carolyn F. Deacon, PhD, DMSc (University of Copenhagen, Copenhagen, Denmark)

Dr. Deacon discussed the current market landscape with the DPP-4 inhibitors, offering potential future directions for the class. Reviewing the currently available options, Dr. Deacon noted similar levels of efficacy due to maximal inhibition of DPP-4 activity – differences, she suggested, were mostly in metabolism, elimination, and selectivity, though these showed little impact clinically beyond in dosing. Notably, Dr. Deacon highlighted that linagliptin is the only option that can be used in all forms of renal insufficiency and hepatic insufficiency without dose adjustments. Moving forward, with the opportunities in efficacy and adverse event profile limited, she focused on the effects of DPP-4 inhibitors in indications outside of type 2 diabetes, including type 1 diabetes and wound healing. The wound healing part was interesting – definitely new news to us.

  • Dr. Deacon opened with a review of the current market landscape with the DPP-4 inhibitors. Though numerous DPP-4 inhibitors have been approved, she highlighted the differences between the available options, noting variations in metabolism (Merck’s sitagliptin, Takeda’s alogliptin, and BI’s linagliptin are not appreciably metabolized, whereas Novartis’ vildagliptin and BMS/AZ’s saxagliptin are hepatically hydrolyzed), elimination (linagliptin is excreted in the bile versus the kidneys with all other options), and selectivity (vildagliptin, saxagliptin, and linagliptin show some activity at other members of the DPP family). While this translates to some differences in dosing (she noted linagliptin is the only option that can be used in all forms of renal insufficiency and hepatic insufficiency without dose adjustments), Dr. Deacon noted that all offered similar levels of inhibition at the DPP-4 enzyme and thus efficacy.
  • Dr. Deacon questioned whether improvements could be made to the drug class, though overall opportunities appeared limited. With maximal inhibition of the DPP-4 enzyme achieved, she indicated efficacy could not be improved. Similarly, she suggested the absorption, distribution, metabolism, and elimination of current options offered simple dosing with minimal side effects. With only short- and medium-term data available, however, she indicated other signals (or differentiating factors) could emerge with time.
  • Dr. Deacon concluded with an exploration of DPP-4 inhibitor use for indications outside type 2 diabetes. While data is scarce and mostly limited to preclinical studies, she noted that the DPP-4 inhibitors have shown some impact on cardiovascular function, neuroprotection, and wound healing. She also showed results indicating a reduction in postprandial glucose excursions in type 1 diabetes patients, likely mediated by suppression of glucagon secretion.
  • The Q&A was really interesting; we wish that more discussion of various combos would have come up.

Questions and Answers

Q: You showed us that linagliptin can be used in renal failure and hepatic insufficiency. Why is that?

Dr. Deacon: For linagliptin, it can be used in renal failure and hepatic failure patients without dose reduction. That’s because it’s not metabolized in the liver. So when you look at patients with linagliptin with hepatic insufficiency or renal insufficiency the exposure is not increased.

Comment: I’ve been using sitagliptin, and I observed a very severe unexpected hypoglycemia.

Dr. Deacon: It’s very unusual to see severe hypoglycemia unless it’s used together with sulfonylureas. I don’t know the answer.

Q: DPP- 4 is released from adipocytes, so it might be important. If we inhibit it what might happen?

Dr. Deacon: I think we’re seeing the effect of the inhibitors DPP-4 wherever it’s from. So I think it will be inhibited to the same effect and will be accounted for in trials.

Q: What can you say about the combination of DPP-4 inhibitors with insulin?

Dr. Deacon: They’ve been giving very good results – the combination seems beneficial in terms of the actions of GLP-1 and insulin, giving additional glycemic control without additional risk of hypoglycemia.

Q: What are the side effects if you give DPP-4 inhibitors to a patient with renal insufficiency?

Dr. Deacon: The adjustments in dosing are simply to bring the plasma concentration to the same level in normal renal function patients. As far as I’m aware there’s no safety issue; you can inhibit the enzyme more than 100%.


Symposium: Basic Science of Incretins for the Clinician

Incretin Control of SS-Cell Mass and Function

Bernard Thorens, PhD (University of Lausanne, Lausanne, Switzerland)

Dr. Thorens discussed the mechanism behind the effects of GLP-1 on beta cell mass. Potentially countering the decline in beta cell function underlying type 2 diabetes, GLP-1 and GIP have been shown to have trophic effects on the beta cells, causing differentiation and proliferation as well as protecting against apoptosis in rat models. Dr. Thorens reviewed animal studies describing the molecular pathway underlying these effects, in which GLP-1 administration was shown to induce IGF-1R expression (protecting against apoptosis through stimulation of activity of the IGF-2/IGF-1R autocrine loop – this pathway was also stimulated by glucose and glutamine); GLP-1 also increased expression of the genes involved in its own inhibition (RGS2, CREM, ICER, DUSP14), suggestive of a feedback mechanism. Dr. Thorens suggested further research would be necessary to confirm these pathways in humans and examine the potential for manipulation.

Questions and Answers

Dr. Michael Trautmann (Eli Lilly, Indianapolis, IN): Do you think this effect is involved in other cell types beyond beta cells? Can you speculate on how much this is a rodent phenomenon or for humans as well?

Dr. Thorens: It has been seen in other cell types such as myoblasts, so it’s not specific to the beta cell. It’s also seen in the thyroid cells. What we know is that blocking IGF-2 in human beta cells shows a similar effect, though data suggests no proliferation through GLP-1 on human cells. But further careful experiments are suggesting that might not be so.

Q: You showed nicely that GLP-1 blocks its own feedback pathway. Is that possibly manipulated by time, given the different duration therapies we have?

Dr. Thorens: I’ve not shown the data, but really the induction of this response occurs within 30 minutes. So you have the balance between this rate and stimulation.

Dr. Juris Meier (St. Josef-Hospital, Bochum, Germany): Some of the results don’t translate to humans, which has been suggested might be due to the age of the animals. Might that be related?

Dr. Thorens: Really we have not done the experiments yet to be sure.


Control of Vascular Function by GLP-1

Hirotaka Watada, MD, PhD (Juntendo University Graduate School of Medicine, Tokyo, Japan)

Dr. Watada described the effects of GLP-1 on vascular function. According to Dr. Watada, GLP-1 has been shown to have inhibitory effects on atherosclerosis, potentially altering cardiovascular outcomes. Research suggests these effects are two-fold – firstly through the ability of GLP-1 to stabilize glucose without hypoglycemia (rat studies suggest inhibiting the postprandial glucose excursion reduces monocyte adhesion [a proxy for the progression of atherosclerosis], though the opposite occurred when glucose fluctuation and hypoglycemia were instead introduced – interesting given the conflicting results from large-scale glycemic control trials), and secondly through direct effects (GLP-1 reduces expression of factors involved in monocyte adhesion and induces vasorelaxation, with a reduction of MCP-1 and TNF-alpha activity in macrophages). Though of course the results of long-term trials will be necessary to confirm any impact on outcomes, we do appreciate additional evidence suggestive of a potential cardiovascular benefit with the GLP-1 receptor agonists.

Questions and Answers

Dr. Juris Meier (St. Josef-Hospital, Bochum, Germany): Is this specific to the vascular endothelium or are there also systemic effects on macrophage function?

Dr. Watada: Actually I know data to answer your question, but we think that this effect also occurs systemically. Therefore we should think that GLP-1 should reduce the inflammatory action.

Q: When you administered insulin and saw the increase in adhesion, you pinned that to the induced hypoglycemia. What about insulin itself?

Dr. Watada: We did experiments with insulin adding glucose to prevent the hypoglycemia, and the effect was removed – so it was the hypoglycemia, not the insulin.


Does One Incretin Receptor Explain All the Effects?

Laurie L. Baggio, PhD (University of Toronto, Toronto, Canada)

Dr. Baggio presented results suggesting the presence of a yet unidentified GLP-1 receptor involved in the hormone’s effects on cardioprotection. Given the rapid breakdown to the 9-36 metabolite, Dr. Baggio posited that many of the effects of endogenous GLP-1 activity could be due to the activity of the metabolite rather than the active hormone. Animal studies have shown that 9-36 reduces postprandial glucose and improves left ventricular performance, absent of effects on insulin, glucagon, or gastric emptying – interestingly, improvements in functional recovery were demonstrated in GLP-1R knockout mice as well, suggestive of activity at an unidentified receptor (GLP-1’s ability to reproduce these effects was blocked with DPP-4 administration, suggesting metabolism to 9-36 was necessary for activity). While it is unclear if these results translate into humans or outcomes, we find the implications relative to the available GLP-1 agonists on the market of interest and look forward to further research – while liraglutide is known to breakdown into a 9-36-like metabolite, both exenatide and lixisenatide are not (likewise, DPP-4 inhibitors are known to reduce 9-36 levels).

Questions and Answers

Q: We have to remember that this is GLP-1, a member of a very large family of related peptides. They all are promiscuous at each other’s receptors, showing cross reactivity. I suppose the question therefore is how physiologically relevant this is. Is there evidence to show 9-36 is physiologically relevant?

Dr. Baggio: These all are pharmacologic studies, so they are all hypothesis driven. We need to do the physiologic studies to confirm the effects.

Dr. Michael Trautmann (Eli Lilly, Indianapolis, IN): I’m still totally confused. I see the some of the actions of 9-36 are replicated by exenatide, but some are not. Do you see this as evidence of at least two different receptors?

Dr. Baggio: You’re correct. That’s exactly what we see – in some studies, exendin-4 acts like 9-36 but in others not. This is what gives us the idea that there might be a second receptor.

Dr. Juris Meier (St. Josef-Hospital, Bochum, Germany): I’ve heard some say the effects of GLP-1 on the heart might be mediated by the glucagon receptor. Is there perhaps a GLP- 1/glucagon receptor knockout mouse to test the hypothesis?

Dr. Baggio: There is some data on that mouse in progress, but I can’t really talk about it until it’s confirmed. I think glucagon has an important role in the heart, but it’s difficult to say if GLP-1 mediates these effects. I don’t think GLP-1 has a very high affinity for the receptor.


The Gut to Brain Endocrine Signals

Stephen Bloom, MA, MD, DSc, FRCPath, FRCP (Imperial College London, London, United Kingdom)

Dr. Bloom discussed current research on the control of appetite through hormonal signals to the brain. Following a review of the orexigenic and anorexigenic signaling pathways in the brain, Dr. Bloom identified a number of the gut hormones known to affect appetite, including GLP-1, PYY, ghrelin, and oxyntomodulin. Molecular and neuroimaging studies suggest GLP-1 may mediate appetite through activity at the vagus nerve, hypothalamus, and brainstem – given knockout of receptors in any of these sites negates the effect, Dr. Bloom posited all three may be required to work in tandem to function properly. He concluded by noting evidence suggesting glucagon/GLP-1 dual agonists produce stronger weight loss versus pure GLP-1 analogs (we are familiar with Transition Therapeutics’ TT-401 and 402 in this class, which were set to begin phase 1 trials in 4Q11 – the company has suggested the combination produces similar reductions in blood glucose levels as current the GLP-1 agonists in addition to greater weight loss and improvements in the lipid profile in preclinical studies).

Questions and Answers

Dr. Juris Meier (St. Josef-Hospital, Bochum, Germany): Oxyntomodulin always confuses me. It targets GLP-1 and the glucagon receptor. Wouldn’t the glucagon receptor usually increase the glucose levels and increase the heart rate?

Dr. Bloom: When aiming to treat a diabetic, you might want to produce weight loss with glucagon, but you don’t want to make the diabetes worse. There are a number of glucagon antagonists that have been tested. I can’t answer it.

Dr. Meier: I’m also puzzled with the PYY story. We’ve been seeing this data, but there’s nothing on the market based on PYY and nothing in late stages of development. What are the limitations to using PYY?

Dr. Bloom: I think it’s pharmaceutical company intelligence. They did the dumb experiment of administering PYY nasally. This produces a massive peak in the blood level and 80% projectile vomiting, given appetite and vomiting are on a continuum.

Q: What about compounds altering MSH – what is the mechanism?

Dr. Bloom: That’s certainly mimicking the normal appetite inhibiting circuit. It has the same problem that rimonabant had – while it will inhibit appetite, it does a large number of other things, such as bladder contraction and activity in the limbic system. With the melanocyte receptor you get a wide range of effects so it’s non-targeted. The advantage of a peripheral gut hormone is it only affects the appetite-related functions.


Corporate Symposium: Recent Developments in GLP-1 Based Therapy for People with Type 2 Diabetes Mellitus (Sponsored by Sanofi)

GLP-1 Agonists in the Diabetes Landscape

Juris Meier, MD (St. Josef-Hospital, Bochum, Germany)

Dr. Meier opened the Sanofi-sponsored symposium with a broad introduction to the GLP-1 agonists. Following a review of GLP-1 wide range of physiological effects, he focused primarily on the differences between the short-acting (Sanofi’s lixisenatide, Amylin’s exenatide) and long-acting (Novo Nordisk’s liraglutide, Amylin’s exenatide once-weekly) agonists, noting a greater effect on postprandial glucose levels with short-acting agonists. He pinned this to differences in gastric emptying, proposing that GLP- 1’s effects on gastric emptying decline with continuous exposure.While in the Q&A Dr. Meier indicated the clinical impact of this on complications was still unclear, he suggested short-acting agonists might better serve patients with large postprandial excursions. This certainly seems logical.

  • Dr. Meier opened with a review of GLP-1’s wide range of physiological effects. In addition to the hormone’s effects on insulin release and glucose levels, he described the reduction in glucagon levels, increase in beta cell mass, slowing of gastric emptying, and decrease in appetite observed after GLP-1 administration. In particular, he highlighted animal studies indicating the potential for cardioprotection (lixisenatide reduced myocardial infarct area induced by transient artery occlusion in rats by 40% vs. 37% with liraglutide) and neuroprotection (liraglutide prevented memory impairment and reduced histological hallmarks of disease in a mouse model of Alzheimer’s), though this has yet to be confirmed clinically. Cardioprotection, in particular, has been discussed for years (nearly a decade) – we’re looking forward, of course, to what will emerge in the long-term CVD trials currently ongoing.
  • Dr. Meier suggested the short-acting GLP-1 agonists produce a stronger reduction in postprandial glucose levels as compared to the long-acting agonists. Highlighting differences in half-life (1.5-4.5 hr for lixisenatide and 2.4 hr for exenatide vs. 12-13 hr for liraglutide and 24 hr for exenatide once-weekly), Dr. Meier presented clinical data indicating a greater postprandial reduction with exenatide versus liraglutide (though liraglutide showed stronger effects on fasting glucose levels). This data has been seen many times before so was not a surprise; one of the questions that is emerging, is whether some patients could even take both a short-acting and a long-acting GLP-1 or whether some combination could be developed.
  • Interestingly, Dr. Meier indicated that the difference in postprandial effect was likely due to varying effects on gastric emptying. Research suggests inhibiting GLP-1’s effects on gastric emptying reduced postprandial control. As such, Dr. Meier proposed that continuous exposure to GLP-1 weakened effects on gastric emptying, leading to the differences observed between the short-acting and long-acting agonists. This was confirmed in a study in which participants received a continuous infusion of GLP-1 for five hours – test meals given at the beginning and end of infusion suggested that gastric emptying was reduced after prolonged exposure.


Emerging Data from the GETGOAL Program

Denis Raccah, MD, PhD (University Hospital Sainte Marguerite, Marseille, France)

Dr. Raccah provided an overview of the phase 3 GETGOAL program with Sanofi’s lixisenatide. In addition to previously reported results (GETGOAL-F1 [one-step vs. two-step titration] and -S [+sulfonylurea] reported at EASD 2011 [see the October 1st, 2011 Closer Look]; GETGOAL-X [vs. exenatide] at ADA 2011 [see the July 16th, 2011 Closer Look]), he noted topline results from the GETGOAL-M study, which examined the effects of lixisenatide administered before either the morning or evening meal versus placebo. Results suggested similar declines in A1c (-0.92% in the morning vs. - 0.83% in the evening vs. -0.42% with placebo; baseline ~8.0%) and fasting plasma glucose, as well as similar rates of hypoglycemia, indicating flexibility in dose timing (though always before a meal as compared to liraglutide, which can be administered irrespective of meals).

  • The phase 3 GETGOAL program includes nine trials of similar design. The program includes over 3,000 patients on lixisenatide with a baseline A1c of 7-10%; with some variation, the trials uptitrate exposure from 10, to 15, to 20 g lixisenatide over the course of 24 weeks. For an overview of all nine trials, see the October 4, 2010 Closer Look.
  • Compiling the results of the GETGOAL-M (dose flexibility), -F1 (one-step vs. two- step titration), and -S (+sulfonylurea) trials, Dr. Raccah suggested similar efficacy and safety findings. As a reminder, GETGOAL-S and -F1 were previously reported at EASD 2011 (see the October 1st, 2011 Closer Look), with additional analyses scheduled for Day #2 and Day #5 of this World Congress, respectively. He previewed results from the GETGOAL-M trial (scheduled for Day #5 of this meeting) as well, which examined the effects of lixisenatide administered before either the morning or evening meal versus placebo. Topline results suggested similar declines in A1c (-0.92% in the morning vs. -0.83% in the evening vs. -0.42% with placebo; baseline ~8.0%) and fasting plasma glucose (-1.19 mmol/l [-21.4 mg/dl] and -0.81 mmol/l [-14.6 mg/dl] vs. -0.25 mmol/l [-4.5 mg/dl] with placebo; baseline 9.4 mmol/l [169.2 mg/dl]). No difference in hypoglycemia was observed between lixisenatide-treated and placebo patients.
  • Dr. Raccah included a comparison of lixisenatide versus Novo Nordisk’s liraglutide. Noting an open label trial comparing 20 g lixisenatide and 1.8 mg liraglutide (two-week initiation/two-week maintenance), he suggested a greater reduction in postprandial glucose levels with lixisenatide (129% vs. 41% AUC0:30-4:30 from baseline to day 28), as well as a greater percentage of patients achieving postprandial glucose levels <140 mg/dl (69.3% vs. 29.4%). He also indicated a lower percentage of patients with GI side effects (58% vs. 73%), though the breakdown and definition of adverse events was not described.


Complementing Basal Insulin: Combining Existing and New Therapeutic Pathways

Luc van Gaal, MD, PhD (Antwerp University Hospital, Antwerp, Belgium)

Dr. van Gaal reviewed the supporting evidence for combination basal insulin/GLP-1, with a focus on data from the GETGOAL program. Following a review of the potentially complementary effects of basal insulin and GLP-1, Dr. van Gaal discussed previously reported results from the GETGOAL-L (by press release; see the June 2, 2011 Closer Look) and GETGOAL-L-Asia (at ADA 2011; see the July 16, 2011 Closer Look) trials. He concluded with a preview of GETGOAL-Duo1 results (12-week insulin glargine treatment followed by 24-week add-on lixisenatide or placebo), which he suggested would be given by press release in the next week. Patients achieved an average A1c of 7.6% (baseline 8.6%) during the initial 12-week run-in, which declined to 6.96% after 24-week treatment (placebo not given); 2 hr PPG levels declined 3.16 mmol/l (56.8 mg/dl) relative to placebo as well, which seemed quite a good result. Lixisenatide-treated patients showed 0.89 kg (1.96 lbs) weight loss relative to placebo, with 22.4% experiencing symptomatic hypoglycemic events versus 13.5% with placebo – this was unsurprising given the Lantus taken with GLP-1. Given the structured nature of this trial, we were encouraged to see the strong improvement in control confirmed with add-on GLP-1 – we look forward to learning more about the ease of add-on over the short treatment period. With a combined lixisenatide/glargine pen in development, an easy transition in combination with a facilitated delivery method could greatly help lixisenatide overcome some of the barriers of late market entry. A successful launch could be quite negative for Novolog and Humalog – Sanofi, of course, has not put big commercial effort into its rapid acting insulin analog, Apidra.

  • Combination basal insulin/GLP-1 offers potentially complementary effects. While GLP-1 receptor agonists primarily target postprandial glucose levels, basal insulins address nocturnal and fasting levels; Dr. van Gaal also highlighted evidence indicating GLP-1 use may ease the weight gain observed with insulin without increasing hypoglycemia. Noting that postprandial glucose contributes more to overall A1c as A1c levels decline (the often-cited Monnier, Diabetes Care 2003), he suggested the addition of GLP-1 could help patients achieve target as well. We certainly believe this is true and expect to see negative implications for rapid acting insulin over time (though there are other positive expectations for rapid acting insulin that will counter this, such as more people being around to take rapid acting insulin once their beta cells are burned out).
  • Dr. van Gaal reviewed topline results of GETGOAL-L (lixisenatide added to basal insulin ± metformin). As a reminder, results were first given by press release earlier last year (see the June 2, 2011 Closer Look). Dr. van Gaal summarized that the addition of lixisenatide improved glycemic control with significant reductions in postprandial glucose, A1c, and body weight and no increase in hypoglycemia – he suggested full results would be submitted to “one of the major scientific meetings” in 2012, presumably ADA.
  • Dr. van Gaal summarized GETGOAL-L-Asia (lixisenatide added to basal insulin ± sulfonylurea) as well. As a reminder, results were first published at ADA 2011 (see the July 16, 2011Closer Look at After 24 weeks, lixisenatide-treated patients (n=146) showed a 0.77% decline in A1c versus a 0.11% gain in placebo (n=154) patients (baseline 8.5%). Dr. van Gaal also highlighted the strong reduction in postprandial glucose (-7.96 mmol/l [143.3 mg/dl] in 2 hr PPG vs. -0.14 [-2.5 mg/dl]), which he said was of particular interest given Asian patients have been observed to have greater postprandial spikes in glucose. While hypoglycemia differed between the groups (42.9% vs. 23.6%), sub-analyses suggested the increase in incidence occurred primarily in patients on sulfonylureas at screening.
  • Dr. van Gaal concluded with a preview of GETGOAL-Duo1 results, which he implied would be released by press release in the next week. As a reminder, patients in GETGOAL-Duo1 were initially force-titrated on insulin glargine for 12 weeks; patients then were randomized to add-on lixisenatide or placebo for an additional 24-week double-blind period. Patients achieved an average A1c of 7.60% (baseline 8.6%) during the initial 12-week run-in, which declined to 6.96% after 24-week treatment (placebo not given); 56.3% of patients were able to meet a target of <7.0% with lixisenatide versus 38.5% with placebo. Two-hour PPG levels declined 3.16 mmol/l (56.8 mg/dl) relative to placebo as well. Lixisenatide-treated patients showed 0.89 kg (1.96 lbs) weight loss relative to placebo, with 22.4% experiencing symptomatic hypoglycemic events versus 13.5% with placebo.


Panel Discussion

Moderator: Jack Leahy (University of Vermont, Burlington, VT)

Juris Meier, MD (St. Josef-Hospital, Bochum, Germany); Denis Raccah, MD, PhD (University Hospital Sainte Marguerite, Marseille, France); Luc van Gaal, MD, PhD (Antwerp University Hospital, Antwerp, Belgium)

Questions and Answers

Q: Can you talk about the mechanism behind the effect of GLP-1 on neuroprotection?

Dr. Meier: We know GLP-1 receptors are expressed in cardiomyocytes, in the vascular system, and also in neuronal cells. There’s most likely a direct effect; we believe it’s an anti-apoptotic receptor. The cardioprotective effect is probably multifaceted, with a reduction in risk factors as well as direct mechanisms involved.

Q: You focused on the differences between short- and long-acting GLP-1. What about glucagon?

Dr. Meier: Glucagon is an extremely important point. I think the glucagon part plays a role in both the short- and long-acting agonists. This hasn’t been examined in the postprandial state, though there is a reduction in glucagon postprandially for both long and short.

Q: What about potential differences on weight?

Dr. Meier: Results are exactly identical, with the same reduction with exenatide, once-weekly, and liraglutide. I would have expected a stronger reduction with short-acting agonists because of the effect on the stomach, but as you can see I am often wrong. I think the CNS effects are mediating the body weight loss, versus reductions in gastric emptying.

Q: What’s the best method for measuring gastric emptying?

Dr. Meier: There are many measures. We have used a carbohydrate-labeled breath test that’s non-invasive and cheap and reliable. This is what we use in clinical practice. Gastric emptying is often disturbed in patients with diabetes. There are other methods, such as ultrasound and MRI-based methods, but that’s not for clinical practice I don’t think.

Q: How can you explain the low rates of nausea in GETGOAL?

Dr. Raccah: When we compare exenatide with lixisenatide you have a compound with little longer half-life with lixisenatide. Exenatide is a shorter. With only one shot of short-acting GLP-1 agonist with lixisenatide it could explain the lower incidence of nausea compared to exenatide. The lower incidence of hypoglycemia is likely for the same reason.

Q: Because lixisenatide is a short-acting compound, what about the potential for antibody production?

Dr. Raccah: There is antibody production for all GLP-1 analogs and receptor agonists. So the titers are elevated, but it is the same feature for lixisenatide and liraglutide. From a clinical point of view there is no relationship between antibody titers and pharmacokinetic factors and efficacy. So probably it has no clinical relevance.

Q: The half-life of lixisenatide is similar to exenatide. Why do you use it once versus twice?

Dr. Raccah: It is a short-acting GLP-1 receptor agonist, but it is little longer acting than exenatide. We have seen with once versus twice daily there are comparable efficacies.

Q: You talked about flexible dosing with lixisenatide. It is recommended with exenatide it goes before a meal. Does flexible dosing mean independent of meal or just at different times?

Dr. Raccah: From the GETGOAL-M study we showed it can be used either in the morning or evening, but both before meals. So when I say flexibility I mean morning or evening, but always before a meal.

Q: Is it possible to combine lixisenatide and glargine in one shot?

Dr. van Gaal: This may become possible in the future. I know Sanofi is working on that potential system.

Q: For which type of patient would you consider a combination?

Dr. van Gaal: As you may have seen from the preliminary data in the Duo1 study, I think optimization is the first goal. The nice additional effects you get from adding lixisenatide is the first consideration. I think secondarily the argument is for patients with large postprandial excursions, if you can see that with CGM.

Q: With a hemoglobin A1c of more than 9.0%, should we control fasting or the postprandial to target? Or both?

Dr. van Gaal: Irrespective of the Monnier hypothesis, I think both are as important. For that reason the combination of basal and lixisenatide has to be considered.

Q: How long do you think it will take before GLP-1 receptor agonists are taken up in international treatment guidelines?

Dr. van Gaal: I think it depends from country to country and agency to agency. Some agencies have already positioned them as second line after metformin. The outcome trial might have a huge impact in the future, though I personally would not wait that long.

Dr. Raccah: Already in some European countries there is the positioning of incretins and DPP-4 inhibitors as very early. In France currently the recommendations are written to put the position of incretin-based therapies.

Q: What clinical benefit would you expect from lowering in postprandial glucose? Do you believe there is a specific clinical benefit?

Dr. Meier: I think this an extremely difficult question, and I think it’s unclear. Some studies suggest postprandial is more detrimental in terms of CV risk, but others challenged that. My personal view is the field is still out and both are important. If you have managed hyperglycemia fasting you need to address postprandial. If you have a problem with fasting but not postprandial, you need to address fasting.

Dr. Raccah: In terms of postprandial, there are two features. The first is the contribution in overall glycemia. Usually you cannot fix an A1c without a postprandial drug. The second aspect is the potential specific effect on endothelial cells and CV outcomes.

Q: What do you think about GLP-1 and the risk of pancreas and thyroid cancer?

Dr. Meier: I would say this has been even over-discussed in the past year. The reality is some animal data showing increased risk of pancreatitis. I have not seen an animal model with pancreatic cancer. Liraglutide showed thyroid cancer in animal models, but none of that has shown any correlation in clinical data. All retrospective analyses have not found pancreatitis and pancreatic cancer. The paper suggesting an association has been criticized heavily for methodology, even by EASD. We cannot rule out side effects but I have not seen any strong evidence supporting the link.

Dr. Raccah: I would like to add that concerning the GETGOAL program, I was part of the data committee, and we checked for these specific things. And to answer your question specifically, no signal was determined in all the GETGOAL programs with lixisenatide.

Dr. van Gaal: On the thyroid cancer with liraglutide, I was involved in the obesity study where they were given up to 3.0 micrograms. Those patients have been followed for more than two years and no changes in calcitonin levels have been observed thus far. There’s the argument that rodents and humans are different.

Q: Where will this combination fall in the future of diabetes management?

Dr. Meier: I think we need more data, but I think we will see more patients on this combination. Off-label many of us have tried it, and I think it will play an increasing role.

Dr. Raccah: In terms of action, I think it makes sense to use basal insulin to address fasting and the short- acting GLP-1 for marked effects on postprandial glucose. I see a strategy for on top of basal insulin when it’s no longer working you add short-acting GLP-1. I see that over rapid-acting insulin in the beginning.

Dr. van Gaal: I am very interested. This is the future with one condition, which is if it is economically affordable for diabetics around the world.


Corporate Symposium: The Effects of GLP-1 on Satiety and Metabolism (Sponsored by Novo Nordisk)

Welcome and Introduction

Ken Fujioka, MD (Scripps Clinical, La Jolla, CA)

Dr. Fujioka began the session with a review of obesity’s prevalence and comorbidities, as well as the benefits (and methods of achieving) weight loss. He said that the past decade has seen great advances in understanding the physiology of obesity, noting that the “biggest player” in these signaling pathways seems to be GLP-1. The only weight-loss drugs available in the US currently are orlistat and phentermine, while the FDA has removed rimonabant (because of side effects related to the central nervous system) and sibutramine (because of side effects related to norepinephrine signaling and the cardiovascular system). In light of these risks, Dr. Fujioka told the audience that peripheral-acting weight-loss mechanisms are now of much greater interest. He briefly mentioned several peripheral- acting therapies that have been observed to cause weight loss, including metreleptin (in development), exenatide and liraglutide (approved for type 2 diabetes), pramlintide (approved for type 1 diabetes), and roux-en-Y gastric bypass surgery (approved for morbid obesity).


Does GLP-1 Play a Role in Satiety in Normal Physiology?

Niels Vrang, MD, PhD (University of Copenhagen, Copenhagen, Denmark)

Dr. Vrang focused his talk on animal studies about GLP-1’s effects on appetite signaling. He reviewed that GLP-1 is produced by intestinal L-cells, while its receptors are found not only in the gut but also communicating to the vagus nerve (which is known to carry signals from the gut to the central nervous system), the brain stem, and the hypothalamus (especially in appetite-regulating pathways that are highly conserved between rodents and humans). Dr. Vrang noted that exogenous GLP-1 receptor agonist administration has been found to increase fullness and decrease hunger in both rodents and humans. Animal studies suggest that the vagal nerve is not important for GLP-1’s effects on appetite (Kanoski et al., Endocrinology 2011), though it may play a role in glucose-regulating effects. The brain stem’s area postrema seems fairly unimportant (Vrang et al., Diabetes 2011), but other parts of the brain stem may play a role. Dr. Vrang concluded that whatever the mechanism, circulating GLP-1 seems somehow to send signals to the appetite-regulating neurons in the hypothalamus, with increases in the mRNA of the fullness neurotransmitter CART and suppression of mRNA of the hunger neurotransmitters NPY and AgRP (Vrang et al., Diabetes 2011).


Gastric Bypass and GLP-1

Ken Fujioka, MD (Scripps Clinic, La Jolla, CA)

Dr. Fujioka provided an overview of the role of GLP-1 in the body and the effect of gastric bypass on gut hormones and diabetes. He began by noting the different diabetes remission rates associated with adjustable gastric banding (48%) and roux-en-y gastric bypass (84%). Dr. Fujioka proposed that the difference may be due to GLP-1 – while roux-en-y surgery drastically increases plasma GLP-1 levels, gastric banding does not. He transitioned to a discussion of the role of GLP-1 in obesity, calling it a “master hormone” and reminding the audience that the level of GLP-1 secretion is reduced with increasing BMI (i.e., morbidly obese individuals make very little GLP-1). After a quick review of the LEAD-3 trial (liraglutide as a monotherapy), he concluded with an interesting case study of a bariatric surgery patient with type 2 diabetes and a surgical complication (Dirksen et al., Diabetes Care 2010). Due to a leak, the patient was tube fed through the normal-sized stomach, which permitted comparison of pre- and post-gastric bypass feeding in the same patient. Feeding the ‘new way’ (orally, which traveled through the small stomach pouch following gastric bypass surgery) resulted in significantly higher GLP-1 and insulin secretion and lower blood glucose compared to feeding the ‘old way’ (through the tube in the full stomach). In his view, it’s the exaggerated GLP-1 secretion following bariatric surgery that helps explain the procedure’s beneficial effects on diabetes (Editors note – this presentation was quite similar to Dr. Holst’s identically titled talk at EASD. For more information, see page 62 of our EASD full report at


GLP-1 and Weight Loss: Emerging Science

Arne Astrup, MD (University of Copenhagen, Copenhagen, Denmark)

The very highly regarded Dr. Astrup reviewed evidence that GLP-1 mediates protein’s effects on satiety and generally enhances feelings of fullness. Also, clinical research indicates that GLP-1 therapy causes weight loss and can even reverse prediabetes (with particular efficacy seen at 2.4 mg and 3.2 mg doses, likely reflecting metabolic effects beyond weight loss alone). Finally, although liraglutide-induced weight loss reaches a stable level over time, Dr. Astrup argued that the drug is still working during this period. Indeed, when patients stop using liraglutide, they tend to regain lost weight – suggesting that ongoing GLP-1 receptor agonist therapy is key to maintaining the new body weight. (Editor’s note – this presentation resembled Dr. Astrup’s identically titled talk at EASD. For further details, see page 61 of our EASD full report at


Panel Discussion

Moderator: Ken Fujioka, MD (Scripps Clinic, La Jolla, CA)

Nick Finer, MD (University College Hospital, London, UK); Arne Astrup, MD (University of Copenhagen, Copenhagen, Denmark); Niels Vrang, MD, PhD (University of Copenhagen, Copenhagen, Denmark)

Questions and Answers

Q: Rather than just a lifestyle intervention, should prediabetics be treated with medications? If so, which one?

Dr. Finer: This is an important question, but I’m not sure we have enough data. The most commonly used drug in prediabetes is metformin. Some data show metabolic improvements, such as in insulin resistance and, for women with prediabetes seeking to become pregnant, fertility rates. So perhaps you see improved “hard outcomes.” However, a note of caution: In DPP, the lifestyle group did much better than the metformin group. So it’s important that we don’t just rely on a prescription for a drug at the expense of full lifestyle intervention.

Q: Why is it that incretin mimetics like liraglutide regulate body weight, but DPP-4 inhibitors don’t?

Dr. Vrang: The levels of GLP-1 generated by DPP-4 inhibitors won’t get to the levels of GLP-1 receptor agonism that have clear effects on food intake. You need higher levels that can only be achieved by adding a long acting GLP-1. It’s a dosing effect.

Q: How does GLP-1 secretion work, in the sense that it’s down-regulated by DPP-4?

Dr. Fujioka: GLP-1 is around for only roughly 90 seconds before it gets degraded by dipeptidyl peptidase 4. The reason it works in normal physiological doses is that it’s right next to the pancreas. But the analogs are around for hours, which explains the different effects seen with drugs rather than physiological GLP-1.

Q: What is the mechanism of the “metabolic memory” preventing type 2 diabetes despite weight regain?

Dr. Finer: This is a very important question. Weight is only one measure of behavior change. Perhaps even when patients re-gain weight, they are more active. Perhaps their nutritional content has improved, or perhaps stress management – an element of some cognitive therapies – has been maintained. Whether the reduced incidence of diabetes is true metabolic memory, or whether it reflects maintained elements of lifestyle change, is unproven. My own view is that it’s probably the latter.

Q: Are higher doses of liraglutide, such as 3.0 mg, useful for obese people with diabetes?

Dr. Astrup: I would expect that higher doses of liraglutide will achieve greater weight loss. There is no doubt about that. Liraglutide is not approved for higher doses than 1.8 mg. We need to see the phase 3 program results before we can be sure.

Q: It’s interesting to investigate satiety with GLP-1 in non-diabetes. How about for eating disorders, like anorexia?

Dr. Vrang: I would expect levels of GLP-1 are low in anorexics, since its secretion is driven by intake of nutrients.

Dr. Astrup: There is no evidence that the anorexic effect is driven by satiety hormones. I think the problem is not related to hunger and satiety.

Q: How many grams of protein were used in the high-protein diet you described?

Dr. Astrup: We looked at the percentage of calories that were protein. We went from 15-25% to 50%. If you have a meal with chicken, cottage cheese, bread, and vegetables, you can easily come up to 50% protein per meal. The goal is to decrease refined carbohydrates and increase protein, whether that’s from meat, fish, lentils, beans, or other sources.

Dr. Fujioka: I learned form my nutritionist that with some new yogurts, you can go from a yogurt that is 7 g protein to 15 g protein in a single serving. (Editor’s note: We believe this is a reference to “Greek” yogurt.)

Q: I’m wondering about the decreased level of GLP-1 associated with visceral and central fat. Does bringing down central obesity help the GLP-1 problem?

Dr. Finer: The holy grail of obesity treatment is to find something to selectively reduce ectopic fat stores. As far as I’m aware, nothing has this selective effect. You tend to lose fat from your visceral and ectopic stores first. Thus, any weight loss will appear to be selective, but there is no good evidence of an intervention that works specifically.

Dr. Astrup: That’s true – losing 10% of one’s body weight typically means losing 20% of one’s body fat and 30% of one’s visceral fat. As a result, any weight loss will preferentially reduce visceral fat.

Dr. Finer: I think this reinforces the benefits of weight loss, especially amounts of weight loss that patients might think is modest.

Q: Does exercise increase GLP-1 levels, either in the brain or blood?

Dr. Vrang: I’m not aware of any effect. I wouldn’t anticipate it. I would anticipate you have to send nutrients down the gut to get GLP-1 signaling, and this wouldn’t normally be combined with exercise.

Q: Certain antipsychotic drugs closely related to increase in appetite and weight. Might these act through an effect on GLP-1?

Dr. Vrang: I’m not sure. In one study, we used olanzapine to induce obesity in rats. We found that GLP-1 mitigated this weight gain. So GLP-1 receptor has potential to reduce this weight gain, but not sure related. Antipsychotics are probably related to more complicated pathways, such as those involving dopamine.

Q: Instead of high doses of liraglutide, perhaps you could use lower doses of liraglutide and add a DPP-4 inhibitor? Would this have the same effect on weight?

Dr. Astrup: It’s an interesting hypothesis, but it needs to be tested in trials.

Q: Obesity and malignancy are correlated. Is there any evidence for a decrease in malignancy after receiving GLP-1 therapy?

Dr. Finer: No, the data are not long enough. But data from the SOS and some US research show that weight loss from bariatric surgery induces up to 50% decreases in cancer mortality at 7-10 years. I think this is an area of obesity management that has been largely ignored so far.

Dr. Astrup: I heard data presented in Copenhagen last week that breast cancer mortality in women went down by 50% in bariatric surgery patients. Thus weight loss mediated by GLP-1 may reduce cancer mortality.

Q: Why is GLP-1 secretion impaired in type 2 diabetes?

Dr. Astrup: I have no idea. Perhaps a predisposition to develop type 2 diabetes is due to impaired GLP-1 secretion. But we don’t have that evidence. There are a few studies looking at the GLP-1 response before and after weight loss. It looks like it is improving after weight loss but not entirely normalized. Impaired GLP-1 secretion could be a primary defect, but more research is needed.

Q: Dr. Finer, you talked about inflammation and various inflammatory factors coming out of fat cells pushing prediabetes into diabetes. How about if you attack inflammation itself?

Dr. Finer: The problem is that most of our anti-inflammatory drugs (e.g., prostaglandins, prostacyclins) act on short-term inflammation. Drugs that affect longer-term markers, such as TNF-alpha, are not only expensive but also involve risks of suppressing immunity. I suspect work is going on in this area, but I haven’t yet seen anything in clinical practice.

Dr. Astrup: CRP levels can fall 40% with 10 kg weight-loss. A whole grain, low-GI diet has been shown to give a 30-40% reduction in inflammatory markers; exercise on top of this provides additional benefit.

Dr. Fujioka: It’s kind of like the healthy obese, who are obese but seem not to have high-risk levels of biomarkers. Exercise seems prevalent in this group.


Corporate Symposium: Incretins in Diabetes – From Pathophysiology to Therapeutic Use (Sponsored by Novo Nordisk)

New Insights into the Pathophysiology of Incretin Regulation

Filip Knop, MD, PhD (Gentofte Hospital, Copenhagen, Denmark)

Dr. Knop gave a comprehensive overview of the history of incretin research and the role the hormones play in the body. He began by summarizing some of the foundational and more recent studies on incretins, arguing that the reduced incretin effect in people with type 2 diabetes is a consequence of the diabetic state (rather than a cause of type 2 diabetes). In his view, studies show that this reduced incretin effect in type 2 diabetes is attributable to the reduced insulinotropic effect of incretin hormones. Dr. Knop also provided evidence that the reduced incretin effect is associated with obesity and increased insulin resistance. However, studies show that improving glycemic control can ameliorate the defects in the incretin effect in people with type 2 diabetes. He concluded by showing data received in the last week on the importance of the vagal nerve, which he believes may play an important role in the insulinotropic effect of GLP-1. (Editor’s note – this presentation was similar to Dr. Knop’s talk at EASD. For more information, see page 64 of our EASD full report at

  • A study in press (Knop et al., Diabetes Obes Metab) shows that a reduced incretin effect is present in obese individuals with normal glucose tolerance and lean individuals with type 2 diabetes. In one study, four groups of patients were studied: healthy individuals with normal glucose tolerance, obese individuals with normal glucose tolerance, lean individuals with type 2 diabetes, and obese individuals with type 2 diabetes. The incretin effect was 35% in the healthy group, 20% in the obese group with normal glucose tolerance, 10% in the lean individuals with type 2 diabetes, and 2% in the obese individuals with type 2 diabetes When the incretin effect was plotted against insulin sensitivity (as measured by HOMA-IR), there was a “clear correlation.”
  • Recent studies suggest that defects in the incretin effect are reversible. Improvements in beta cell responsiveness to GLP-1 and GIP have been observed in people with type 2 diabetes after near-normalization of blood glucose following insulin therapy. Additionally, patients with gestational diabetes exhibit a reduced incretin effect during the third trimester of pregnancy (30%), but following delivery and reestablishment of normal glucose tolerance, the incretin effect returns to normal (54%).
  • Emerging data suggests that the vagal nerve may play a role in the insulinotropic effect of GLP-1. A recent study compared individuals receiving a truncal vagotomy (resection of the vagus nerve) to a control group. Both groups were given a 50-gram oral glucose tolerance test and an intravenous glucose infusion. The incretin hormone responses were compared between the two groups. Relative to the control group, patients who had received the truncal vagotomy produced a little more GIP and a 6-7-fold increase in GLP-1. However, the group that received the truncal vagotomy exhibited a significantly decreased incretin effect compared to the healthy control group.


What Next, When Metformin Fails?

Richard Pratley, MD (Sanford-Burnham Medical Research Institute, Orlando, FL)

Dr. Pratley discussed the benefits of GLP-1 receptor agonists (in particular liraglutide) relative to other second-line therapeutic options after failure of metformin monotherapy. He noted that compared to DPP-4 inhibitors, TZDs, and sulfonylureas, GLP-1 receptor agonists have demonstrated better or comparable effects on A1c, weight loss, and hypoglycemia. In particular, compiled phase 3 data suggest that liraglutide causes a greater percentage of patients to reach a composite endpoint of A1c < 7.0%, no weight gain, and no hypoglycemia, as compared to various other second-line options (including exenatide, insulin glargine, sitagliptin, sulfonylureas, and TZDs). Despite requiring injections and being associated with GI side effects, GLP-1 receptor agonists are associated with similar or better treatment satisfaction relative to sulfonylureas and DPP-4 inhibitors. Dr. Pratley also noted that GLP-1 receptor agonists may improve adherence and that they have been found cost-effective relative to sitagliptin and sulfonylureas (with liraglutide still more cost-effective than exenatide). (Editor’s note – Dr. Pratley’s presentation was similar to an identically titled one that he delivered at EASD. For details, see pg 65 of our EASD 2011 full report at

  • Patient satisfaction seems greater with liraglutide than sitagliptin, as measured with the diabetes treatment satisfaction questionnaire (DTSQ) administered in a randomized, head-to-head trial. From baseline to week 52, treatment satisfaction increased to a greater degree with liraglutide than sitagliptin (statistically significant for liraglutide 1.8 mg vs. sitagliptin). At week 52, sitagliptin patients were switched to liraglutide, a change that was associated with increases in treatment satisfaction at week 78 (statistically significant for 1.2 mg liraglutide and non-significant for 1.8 mg liraglutide). Also, Dr. Pratley noted that the sitagliptin- to-liraglutide switch was not associated with any change in perceived convenience of treatment even though a once-daily injection might intuitively be thought less convenient than a once-daily pill (Montanya et al., Diabetes 2011). Dr. Pratley suggested that GLP-1 receptor agonists could also be associated with greater compliance since, relative to other commonly used second-line therapies, they have low risk of hypoglycemia, benefits on body weight, and (in the case of Victoza or upcoming GLP-1 receptor agonists) they require dosing daily or less frequently.
  • Dr. Melanie Davies and colleagues found that over the long term, liraglutide is likely to be cost-effective relative to sulfonylureas (Diabet Med 2011). The researchers used a model in which patients were given either liraglutide or sulfonylureas for five years and then switched to basal insulin. The probability that liraglutide would be cost-effective varied depending on the monetary value assigned to each quality-adjusted life year (QALY). For example, liraglutide 1.2 mg achieved an 80% probability of being cost-effective relative to sulfonylureas when each QALY was worth roughly 17,000 pounds; the liraglutide 1.8 mg dose reached 80% likelihood of cost effectiveness only when the QALY value was raised to roughly 27,000 pounds. Currently in the United Kingdom, cost-controlling agencies are generally willing to spend 20,000 pounds per QALY. At this level, 1.2 mg liraglutide has a 90% chance of cost-effectiveness relative to sulfonylureas. Evidently under this model, liraglutide’s greater per-dose expense is outweighed by its greater benefits on weight loss and lower risk of hypoglycemia.


Panel Discussion

Moderator: Melanie Davies, MD (University of Leicester, UK)

Filip Knop, MD, PhD (Gentofte Hospital, Copenhagen, Denmark); Stephen Gough, MD (University of Oxford, Oxford, UK); Richard Pratley, MD (Sanford-Burnham Medical Research Institute, Orlando, FL)

Questions and Answers

Q: Are there data on intra-subject variability and the incretin effect?

Dr. Knop: That’s a good question. The incretin effect is a plastic character. It depends on how large the glucose stimulus is. But what also matters is how you measure the pancreatic beta cell response: insulin, C-peptide, or insulin secretory rates. When you measure the incretin effect, you must take all these variables into consideration. I’m not aware of studies that have evaluated intra-subject variability by looking at the same person several times. Across the same population, however, there is great variation in the incretin effect. It mostly depends on the way you measure it.

Q: On the experiment in vagotomized patients that Dr. Knop presented, can you speculate on what causes the increase in GLP-1 and how your data is different from some of the previous data we’ve seen?

Dr. Knop: These are very new data, and we are not sure how to interpret them in a perfectly clear way. Firstly, I didn’t show these data, but the vagotomized patients experienced decreased gastric emptying time tendency. The oral glucose load was propelled more rapidly downward. But against that, we saw only a small rise in GIP, the other incretin hormone. I think maybe there is a compensatory mechanism, given that the vagotomized patients lack a compelling way to convey an important effect of GLP-1. Perhaps they are producing a greater level of GLP-1 to compensate, though the potential mechanism for this is not clear.

Q: Why was the effect of sitagliptin diminished compared to liraglutide or exenatide at lower A1c levels?

Dr. Pratley: There is not a good explanation. At lower A1c levels, you remember that the majority of hyperglycemia is from the postprandial response. Exenatide would be expected to have a better effect there, as I presented data from the DeFronzo study. Liraglutide may have more effect on the fasting glucose and could knock the whole curve down. We don’t have the studies to say for sure.

Q: Do you think that the UK NICE guidelines, which recommend use of 1.2 mg liraglutide, should these be revised in light of evidence presented today?

Dr. Gough: NICE recommends that 1.2 mg is preferred for most patients. I think this was based on earlier data on glycemic control. Originally we felt the major benefit of lira 1.8 mg was on weight reduction. A number of physicians in UK are now using the 1.8 dose. It is difficult to tease out the relative benefit of the higher dose or to deprive patients of this benefit when they are achieving weight loss. I think there are some differences in glycemic control, but the main ones are in weight. And most of our patients are overweight.

Q: You showed data on treatment satisfaction and quality of life with liraglutide. Is this explained by the weight loss?

Dr. Pratley: That’s an excellent question. In the comparison versus sitagliptin, it was principally explained by body weight and the improved effect on glycemic control. Hypoglycemia and convenience were about the same. The combination of improved efficacy and weight loss really drove those differences.

Q: Can you explain the rationale for using a GLP-1 and a DPP-4 inhibitor?

Dr. Knop: Exogenous GLP-1 analogs also activate nerve endings in the intestine. The question is, ‘Does this elicit the same effect as endogenous GLP-1?’ Maybe DPP-4 can potentiate some of the neural defects of endogenous GLP-1. If true, it does make sense to use a DPP-4 inhibitor and a GLP-1.

Q: Is there any hope for an effective oral GLP-1 agonist?

Dr. Knop: Certainly it would be very nice, but it would be more nice to have a substance that could increase endogenous GLP-1 secretion from L cells without needing nutrients deposited somewhere.

Q: In the study of insulin and GLP-1 combination therapy, what dose of insulin was used, and how was it titrated?

Dr. Gough: Insulin detemir was titrated to a fasting plasma glucose of 4.o-6.0 mmol/l. The mean dose of insulin detemir was 39.2 units per day at six months and 42 units per day at the end of 12 months. So there was a slight difference between six and 12 months, but it was generally around 40 units per day.

Q: Can you discuss the risk of pancreatitis and pancreatic cancer with incretin-based therapy? What about use in patients with non-alcoholic fatty liver disease?

Dr. Pratley: Clearly, there has been a concern about pancreatitis with GLP-1s and case reports with DPP-4 inhibitors. This has been addressed with exenatide in at least five studies. These were claims in large managed care databases. There is no sense of an increase in pancreatitis with either GLP-1s or DPP-4s. You obviously must pay careful attention to this in people with diabetes. The question about pancreatic cancer stems from a study in Gastroenterology. The overwhelming data does not support pancreatic cancer risk. Studies have addressed this question as well.

Dr. Gough: On the topic of non-alcoholic fatty liver disease, we have just conducted a meta-analysis of the LEAD program. We’ve seen quite clearly a benefit with liraglutide. It’s hard to tease out whether this improvement is independent of A1c and weight improvement. But there is no deterioration in the disease and we can be reassured by that.

Dr. Pratley: A recent study in Diabetes Care in the last few months looked at exenatide in combination with pioglitazone in patients with fatty liver disease. It found a regression in fatty liver disease, suggesting a potential benefit.


Corporate Symposium: Cut the Burden – Break through Diabetes (Sponsored by Bristol-Myers Squibb / AstraZeneca)

Welcome and Introduction

Ghaida Kaddaha, MBBS, MRCP, FRCP (Rashid Hospital, Dubai, United Arab Emirates)

Although not evident from its title, this symposium was about incretins, especially DPP-4 inhibitors, and especially saxagliptin. It’s possible with a clearer indication of subject matter, the turnout would have been better; as it was, the hall was only about 15% full. Dr. Kaddaha, from Dubai, chaired the session and noted in her introduction that the UAE was an appropriate location for the meeting, given that the Gulf States have such a high prevalence of diabetes (Editor’s note: Bahrain, Kuwait, Qatar, and the UAE all rank in the top 10 for diabetes prevalence by percent; IDF Diabetes Atlas 5th ed., 2011).

Facing the Hurdles in Type 2 Diabetes Management

Ian Blumer, MD, FRCPC (University of Toronto, Toronto, Canada)

The general theme of the session was that with the majority of people with diabetes still not at target, we are in the market for a “wonder drug,” in Dr. Blumer’s words. Was it implied that saxagliptin and its colleagues might shape up to be it? Of note, he described the features that a perfect drug might have, all factors that tend to play a role in a doctor’s decision to prescribe: it needs to have efficacy (in terms of A1c and pre and postprandial glucose control); it would have durability of glycemic control (i.e., beta cell preservation); it would have additional benefits (especially on cardiovascular risk factors, such as blood pressure, lipids, and weight); it should have little to no adverse events (such as hypoglycemic episodes, weight gain, bone effects, or cardiovascular events); it would have long-term safety data to support it, have a labeled indication (vs. off-label use), and be incorporated in guidelines and supported by expert opinion; it should not have prohibitive cost or coverage issues; and it should not be a hassle (i.e., easy to explain, easy to take). What we need in diabetes, he said, is a “statin-like” therapy – easy, effective, other benefits. He and the other speakers laid out how DPP-4 inhibitors fill or seem poised to fill many of these criteria.


DPP4-Inhibition: Taking Diabetes Management to a New Level

Ghaida Kaddaha, MBBS, MRCP, FRCP (Rashid Hospital, Dubai, United Arab Emirates)

Dr. Kaddaha spoke about the promise of incretins to “take diabetes care to a new level”: as effective as sulfonylureas without weight gain and hypoglycemia, but then also with animal studies showing induction of beta cell proliferation and differentiation, and even with evidence, human or animal, of direct beneficial effects on blood pressure, lipids, cardiac function, neurons, and fatty liver. Several cardiovascular outcomes trials for DPP-4 inhibitors will soon more definitively determine the extent of cardiovascular effect: SAVOR (BMS/AZ), TECOS (Merck), EXAMINE (Takeda), and CAROLINA (Boehringer-Ingelheim).


Saxagliptin: A DPP4-Inhibitor in Action

Robert Henry, MD (University of California, San Diego, San Diego, CA)

Dr. Henry, in addition to a rundown of basic clinical trials on saxagliptin, presented a study of saxagliptin efficacy in varying degrees of renal failure: there is significant A1c reduction compared to placebo in patients with moderate and severe renal failure, but not in those with end-stage renal disease, after both 12 weeks and 52 weeks. Saxagliptin was very safe across all levels of renal impairment, he said, with few notable events.


Panel Discussion

Moderator: Ghaida Kaddaha, MBBS, MRCP, FRCP (Rashid Hospital, Dubai, United Arab Emirates)

Ian Blumer, MD, FRCPC (University of Toronto, CA); Robert Henry, MD (University of California, San Diego, San Diego, CA)

Questions and Answers

Q: Why wasn’t saxagliptin effective in end-stage renal disease?

Dr. Henry: People who have end-stage renal disease have had diabetes for a long time and they would have very low c-peptide levels and poor response to most oral anti-diabetics. Most end up on insulin and in the study I presented, 75-80% were on insulin. I wouldn’t have expected that patients with end-stage renal disease would be responsive to DPP-4 inhibitors.

Q: How long do you keep patients on saxagliptin?

Dr. Henry: You keep giving it while it still has efficacy. My experience tells me all agents fail over time. The majority of individuals do fail, and I think that includes the GLP-1 class. I would predict that DPP-4s have a more durable effect than sulfonylureas, more like metformin and glitazones.

Dr. Blumer: I agree.

Q: Is there anything we have to worry about when prescribing DPP-4 inhibitors, such as allergies? Some have reported some vasculitic changes…

Dr. Blumer: I find the class to be very well tolerated.

Dr. Henry: With most drugs, you can get an allergic reaction. I have seen reports of it in DPP-4 inhibitors. Is it common? No. The reason they’ve done so well is because they are reasonably efficacious and easy to use. They are most effective in combination.

Dr. Blumer: Yes I seldom would use one in monotherapy; mainly in combination.

Q: Metformin also increases GLP-1 levels. How does it do that?

Dr. Henry: Though metformin has been on the market for a while, I still can’t tell you how it works anywhere in the body… so I couldn’t tell you a mechanism in the GI tract. But I think it’s clear that there is an effect; the studies have been done and repeated. It is clear that metformin gives you increased GLP-1 over and above DPP-4 inhibition alone. The reason we haven’t picked it up before is because it is difficult to detect a GLP-1 increase when you don’t have DPP-4 inhibition.

Q: Would you combine GLP-1 agonists and DPP-4 inhibitors?

Dr. Blumer: I’ve been tempted. Sometimes I think GLP-1 agonists seem like a sledgehammer compared to the small hammer of DPP-4 inhibitors but there is reason to think that it would help.

Dr. Kaddaha: Studies show that they are more effective together, because there are other hormones involved such as GIP.

Q: Can the benefits you mentioned for saxagliptin be generalized across the board to incretins?

Dr. Kaddaha: Myocardial protection has been shown with DPP-4 inhibitors and GLP-1 agonists. In studies on neuroprotection and fatty liver, most of them are with GLP-1 agonists or exendin-4. Except for the study on myocardial protection, it’s hard to extrapolate to the whole class – I’d wait for data.

Q: What do you think about comparisons between different DPP-4 inhibitors?

Dr. Henry: Each of them has very small advantages over each other. It’s a very good class and they are probably very similar in terms of their A1c reductions. So it’s just small advantages and disadvantages. Safety across the board is very good.

Dr. Blumer: I think they are much more similar than they are different.

Q: Can we use DPP-4 inhibitors safely in people with liver disease?

Dr. Henry: I’m careful with any medications in people with severe liver disease or cirrhosis, because of clearance.

Dr. Blumer: I also avoid oral agents in liver failure and use insulin in those cases.

Dr. Kaddaha: I thought I heard saxagliptin can be used in liver disease since it is partly excreted by the kidney.


Corporate Symposium: The Evolution of DPP-4 Inhibition in Type 2 Diabetes (sponsored by Merck Sharp & Dohme)

Can Intervention Change the Course of Beta Cell Function?

Steven Kahn, MB, ChB (University of Washington, Seattle, WA)

Dr. Steven Kahn led off the symposium with a focus on the progression from prediabetes to type 2 diabetes. Using the oral disposition index (product of insulin sensitivity and beta cell response), Dr. Kahn discussed the DPP and ADOPT study results from a physiological perspective. He noted the

beneficial effects of lifestyle intervention, metformin, and rosiglitazone (better oral disposition index) and the negative impact of glyburide (worse oral disposition index due to progressive beta cell failure over time). Dr. Kahn transitioned to genetics, stressing the importance of the TCF7L2 gene – interestingly, individuals in the DPP who inherited both alleles for the gene had an 80% increased risk of developing diabetes. This group also had a reduced incretin response, which suggests the potential for incretin therapy to slow the progression from prediabetes to diabetes. To close, Dr. Kahn called for more studies on GLP-1 and DPP-4 use in prediabetes, the need for more data on the mechanisms of bariatric surgery, and a greater understanding of genomics, proteomics, and metabolomics.

Questions and Answers

Q: Can you speculate more on the impact of sulfonylureas and metformin on the beta cell? What exactly happens?

A: I don’t think we have the answer. The problem is that the mechanistic studies are done in rodents, but the proteins expressed in human islets are different. In my group, we’re very interested in amyloid deposition. You see this in humans and not in rodents. Excessive secretion of this one particular peptide could result in beta cell dysfunction. Sulfonylureas are intense secretagogues. With rosiglitazone and metformin, we can improve things. We’re doing work at the University of Washington with DPP-4 inhibitors to measure beta cell function over a year.

Q: Is the abnormal incretin response the cause of the beta cell dysfunction?

A: My personal opinion is that the data suggests both sides. There is data to suggest deficiency and data to suggest no deficiency. With the TCF7L2 gene, you have higher GLP-1 and GIP levels, but a reduced incretin effect. This is not reduced secretion of GLP-1, but a deficient effect of GLP-1 at the level of the beta cell. I don’t think GLP-1 deficiency on its own is a major explanation for the beta cell dysfunction in type 2 diabetes.

Q: Can you use sitagliptin in patients with impaired glucose tolerance?

A: My guess is yes, it can work – there is not doubt in my mind. But we need the trials to support this.

Q: What’s the best combination — a TZD plus metformin?

A: From a glucose lowering perspective, the combination of a DPP-4 and a TZD is better than what was seen in the DPP with metformin. When you use combination therapies in individuals without diabetes, you must weigh the positive and negative impact. With TZDs, you have the CV risk, weight gain, edema, bone loss, and bone fractures. If I had to make a choice for my mother, it would be metformin plus DPP-4 or DPP-4 alone.

Q: I noticed in one of the studies that sitagliptin lost its efficacy after a few weeks?

A: Those curves go up again with both metformin and sitagliptin. We do not yet have a miracle drug that will lower glucose and keep it there. You still have individuals with rosiglitazone that progressed and failed. I don’t think the curves are that different with metformin and DPP-4 inhibitors. We need to understand the mechanisms, and I think pharmacogenomics could give us a clue. Some individuals with certain genes could have a dramatic effect.

Q: In the future, will we be able to treat type 1 or type 2 by genetics?

A: There is data emerging. For instance, one study looked at sulfonylureas and metformin and the TCF7L2 gene. You see a differential response to the sulfonylurea depending on the genotype. We need more studies and more time.

Q: What about the effect of sulfonylureas on the beta cell. Are all sulfonylureas created equal?

A: In vitro data suggests no marked differences. For clinical studies in general, I’d like to make an observation. It’s all very well to compare drug x to drug y. But it must be a head to head study matched at baseline. Comparisons between drug studies not done head to head are not appropriate.

Q: What about DPP-4 safety during pregnancy? Is it logical to use in gestational diabetes?

A: There is no data of value that I know of. I would strongly discourage its use until we know more.


The Expanding Role of Sitagliptin in Diabetes Management: New Clinical Data

Michael Nauck, MD, PhD (Diabeteszentrum Bad Lauterberg, Harz, Germany)

Dr. Nauck gave a comprehensive overview of an array of sitagliptin clinical studies, including use as a monotherapy, as an add-on to metformin, in combination with other therapies, and in patients with impaired renal function. He began by citing a study (Aoki Endocr J 2010) in non-diabetic men that found GLP-1 levels were most increased when a DPP-4 inhibitor was combined with miglitol. In his view, this suggests alpha-glucosidase inhibitors would be “very good additives” to DPP-4s. He next highlighted a number of other sitagliptin trials, noting its noninferiority to metformin and sulfonylureas, its superiority to voglibose, and its safety/effectiveness in renally impaired patients (at half the dose). He concluded with an emphasis on the potential to combine DPP-4s with other drugs that increase the release of GLP-1 – this would lead to more substrate and likely greater effect than just using the DPP-4 alone.

Questions and Answers

Q: Is there a difference between sitagliptin and other DPP-4s?

A: All the DPP-4s approved for type 2 diabetes are different chemical entities. That is a difference. They may differ with respect to their off-target activities, such as other protesases like DPP-8 and DPP-9. There are some thoughts that off-target activity can be a problem leading to safety problems. I would say all the approved drugs have not created obvious problems. The other difference is how long we have had these on the market and how much patients have been exposed to them. Sitagliptin is the number one – it was the first one to market and has a very long record. Until now, we can conclude what we did at the end of phase 3: there are not substantial toxicity or safety concerns. As long as this is the case, sitagliptin has a strong standing in the class.

Q: Are there different characteristics of patients that respond to or fail sitagliptin therapy. Does duration of diabetes play a role?

A: I cannot give an easy, clear answer. In the beginning, you need a substrate to act on. GLP-1 needs beta cells. You would think that patients with different functional status of beta cells would see different efficacy, but this is not the case. You can treat patients with short and long diabetes duration and different BMIs. Any simple clinic characteristic does not help us explain the difference. Regarding incretins, there are studies that suggest some changes in the GLP-1 receptor gene, which seem to determine how beta cells respond to exogenous GLP-1. Incretin-based medications may have a pharmacogenomic basis for being very active or having reduced activity. The fact that simple clinical characteristics don’t help us is both a bad and good message. The good is that you can treat anyone with a DPP-4 without fearing reduced effectiveness.

Q: What about use of sitagliptin in patients with renal problems?

A: Yes, as I presented in my talk, sitagliptin appears to be very safe for patients on dialysis. It must be used at a reduced dose because it is eliminated renally.

Q: What about the use of DPP-4 in type 1 diabetes?

A: This is an interesting topic. The rationale is that even in type 1 diabetes, several of the mechanisms of action of GLP-1, which are augmented with DPP-4, should work. One example is suppression of glucagon. There would be no stimulation of insulin secretion due to little residual capacity. It is of interest whether you can reduce the insulin dose. You might even see lower hypoglycemia risk in patients with type 1 diabetes. There are initial studies that suggest some benefit to adding sitagliptin for type 1s. These are small and short studies, so it’s much too early for getting approval for this treatment. It’s an area of further research and I’ll be interested in the results.

Q: What about the risk of pancreatitis?

A: There have been observations of acute pancreatitis with the treatment. This is not a real problem. There are now epidemiological studies looking at how many patients develop pancreatitis. The number has never been found elevated except in FDA reports of adverse events by Peter Butler. This is prone to reporting bias and it occurred around the time this was a public issue. My take is we would like to see more data, larger databases analyzed, and maybe randomized trials showing that. I’m pretty convinced that pancreatitis will not be an issue. Caution is advised to use sitagliptin in patients with previous pancreatitis.

Q: If cost and side effects are not an issue, would you add a GLP-1 analog or a DPP-4 inhibitor in a patient failing metformin?

A: It depends on the patient. Rarely do I come into a conflict. Patient preferences will determine how you proceed. Those who fear injections will prefer DPP-4. Generally, I find that the subgroup that wants to be treated with a GLP-1 are those who want to lose weight and haven’t succeeded in the past. They see this as a last chance to lose weight. So this question will depend on the patient’s view of their body weight, patient preferences, and the distance from glycemic target.

Q: Is there data on the use of sitagliptin in intensive insulin therapy in type 1 diabetes?

A: There is one published study. We need more information.

Q: Will DPP-4s replace sulfonylureas? If so, when?

A: If you look at the numbers of medications used in the US and Europe, use of sulfonylureas is slowly but steadily declining, and use of DPP-4s is steadily decreasing. I would only need a ruler to determine when DPP-4s will surpass sulfonylureas. If you read the European guidelines on developing new drugs, sulfonylureas would never get approval under today’s circumstances. The cost of drugs will be important, but the trend is in favor of DPP-4s.


2. Treatment Algorithms and Combination Therapies

Oral Presentations: New News on Combination Therapies of Type 2 Diabetes

Adding Insulin Detemire to Liraglutide and Metformin Improves Glycemic Control with Sustained Weight Reduction: 52 Week Results

J. Hans DeVries, MD (Internal Medicine, Academic Medical Center, Amsterdam, Netherlands)

Dr. DeVries reprised Dr. Stephen Bain’s presentation from EASD 2011 (see the October 1, 2011 Closer Look) on a 52-week study investigating the effects of add-on insulin detemir to patients on background liraglutide and metformin. While in the Q&A Dr. DeVries refrained from comparing results to those with exenatide BID/insulin glargine, as evidence for combination GLP-1/basal insulin builds we will be interested to see how the GLP-1 agonists compare given their differing effects on postprandial glucose levels.

Questions and Answers

Q: If you had hypoglycemia did you back off on insulin or the liraglutide dose?

Dr. DeVries: The insulin dose.

Dr. John Buse (University of North Carolina, Chapel Hill, NC): We additionally published the GWCO study, and the A1c achieved with glargine and exenatide was somewhat lower. Any thoughts on the comparison?

Dr. DeVries: That’s apples to oranges. But the positive conclusion from your trial is that I think your patients were further down the road in diabetes, and still they could be brought to target.


Initial Combination of the DPP-4 Inhibitor Linagliptin with Metformin Improves Poor Glycemic Status in Patients with Type 2 Diabetes

Thomas Haak, MD (Diabetes Center Mergentheim, Bad Mergentheim, Germany)

Dr. Haak reiterated results from a 24-week phase 3 trial examining the safety and efficacy of initial combination therapy with linagliptin and metformin, which he first presented at ADA 2011 (see the July 16, 2011 Closer Look). He again concluded with support for initial combination linagliptin/metformin, speculating that the cost of treatment could be justified by reductions in complications down the road.

Questions and Answers

Dr. John Buse (University of North Carolina, Chapel Hill, NC): You do lose the advantage of one dose once daily when you go to the combination right away, as well as occasional GI side effects. The idea of starting two drugs at once doesn’t provide a great deal of advantage and has the chance of causing confusion. In my practice, I would start with one first and then add on the other. Any comments?

Dr. Haak: If you look at German guidelines, we usually start with metformin twice daily, so patients are used to taking drugs twice a day. If you consider the cost, which is €2 in Germany for DPP-4, it might be easier to start once a day. I’m not sure guidelines will be changed.

Dr. Philip Home (Newcastle University, Newcastle, UK): We wouldn’t change the guidelines based on this study. In clinical practice, when you add metformin at the beginning you don’t get to target, so then you would add linagliptin. I don’t think payers would support the initial combination. This was also a proof of concept study. Are you planning on doing the correct studies with increased trial length?

Dr. Haak: I’m not so sure whether Boehringer Ingelheim will plan such a study. If they do, I’d expect the results to be the same as the sitagliptin studies. We are very experienced with the DPP-4 concept. I think the novel aspect of linagliptin is it is better to handle and has less risk with patients with renal and hepatic issues, though I’d expect clinical results to be similar.

Dr. Robert Ratner (MedStar, Washington, DC): Data exists that the addition of metformin increases total GLP-1 levels. Do you have any information suggesting the combination increases active GLP-1 levels?

Dr. Haak: I don’t have the data, but I would expect this would be.


Anti-Inflammatory Effects of Add-on Therapy with Pioglitazone, Metformin or the Combination of Both in Type 2 Diabetes

Andreas Pfützner, MD, PhD (Institute for Clinical Research and Development, Mainz, Germany)

Dr. Pfützner reiterated results from the PIOcomb study, which examined the effects of pioglitazone on various measures of cardiovascular risk. Supplementing results first presented at ADA 2011 (see the July 16, 2011 Closer Look), Dr. Pfützner noted significantly improved total adiponectin (4.29 to 13.20 mg/l with pioglitazone and 4.83 to 13.42 mg/l with pioglitazone/metformin vs. 4.43 to 4.33 mg/l with metformin alone), hs-CRP (3.30 to 2.57 mg/l and 2.62 to 1.78 mg/l vs. 3.22 to 2.99 mg/l), and HOMA-S scores (4.60 to 2.39 and 3.40 to 1.80 vs. 3.87 to 4.14) in patients receiving add-on pioglitazone. He suggested these results, in addition to improvements in the lipid profile, may have contributed to the reduced risk of all-cause mortality/non-fatal MI/stroke observed in the PROactive trial.

Questions and Answers

Q: It’s becoming more confusing to us as practitioners – many studies show pioglitazone causes weight gain. In your study, it did not cause weight gain. What determines that in your opinion?

Dr. Pfützner: As I mentioned, people gain weight when they eat more than what they burn. I believe it’s a matter of educating them about the uptake of food, which we did in our trial. It takes a lot of time.

Q: How does it compare to a statin in terms of cholesterol?

Dr. Pfützner: We’ve published a paper looking at the impact of pioglitazone with and without a statin, and results suggest additive effects on the lipid profile given the different mechanisms of action. We can try to select those patients that are in the process of developing that glycemia representative of diabetes. I treat patients at risk with the combination of pioglitazone and a statin.

Q: In the group on pioglitazone plus metformin and insulin, it had the greatest effect. They also had the greatest change in A1c. How much is that due to glucose lowering?

Dr. Pfützner: I do not consider pioglitazone to be the most potent glucose-lowering agent we have. I think we should look at it like aleglitazar – a drug for cardiovascular benefit with the side effect of lowering glucose. Do we want to focus on glucose or anti-inflammatory drugs?

Q: But how much is due to glucose lowering?

Dr. Pfützner: If you look, the anti-inflammatory effect in the combination was not as good as the monotherapy with pioglitazone alone.

Dr. John Buse (University of North Carolina, Chapel Hill, NC): Studies have implied that cardiovascular benefit may be due to particle number and not size. Any comments?

Dr. Pfützner: When you look at the numbers, you run into laboratory issues I’ll confess. I believe it’s a combination of both, though it seems an unclear situation.

Q: Can pioglitazone be used post-myocardial infarction?

Dr. Pfützner: We do not have any evidence suggesting that there is an improvement acutely. Personally, from the anti-inflammatory activity I think we can expect something. Obviously we do have a lot of data on secondary prevention and definitely then I would suggest to use it.


Oral Presentations: Late-Breaking Abstracts Session

Efficacy and Safety of Initial Combination Treatment with Sitagliptin and Pioglitazone - a Factorial Study

Robert Henry (University of California San Diego, La Jolla, CA)

Dr. Henry presented 24-week results from a factorial study of sitagliptin (100 mg) and pioglitazone (15, 30, or 45 mg), alone or in combination for people with type 2 diabetes. The seven groups were well matched at baseline, with mean A1c 8.7-8.9% and mean age 50.1-53.5 years. Each of the sitagliptin+pioglitazone groups was statistically significantly superior to the respective pioglitazone monotherapy group, with respect to change in A1c, fasting plasma glucose, and two-hour postprandial glucose. Co-administration of sitagliptin seemed to blunt pioglitazone-associated edema but to increase pioglitazone-associated weight gain. Rates of hypoglycemia were similar across groups but relatively high given the safety profile of the two drugs (roughly 10% for any reported hypoglycemia, including some occurrences of severe hypoglycemia requiring assistance), based on 54-week results. As noted during the question-and-answer period, Merck dropped development of its fixed-dose combination of sitagliptin and pioglitazone in November 2011 for unspecified reasons. But given the similar effects seen with all DPP-4 inhibitors, these results have relevance to Takeda’s aloglitpin+pioglitazone fixed-dose combination, which was resubmitted to the FDA in July 25 and which now has a PDUFA date of April 25, 2012.

  • Type 2 diabetes patients with baseline A1c from 7.5% and 11.0% were randomized to sitagliptin (100 mg), pioglitazone (15, 30, or 45 mg), or a combination of sitagliptin and one of the three pioglitazone doses (n=183-198 per group). Enrollees had to be drug-naïve or (if A1c was ≤ 9.0%) on sulfonylurea or metformin monotherapy. The groups were similar in mean age (50.1-53.5 years), baseline A1c (8.7-8.9%), fasting plasma glucose (177-189 mg/dl), and two-hour postprandial glucose (248-267 mg/dl). Changes in these endpoints were assessed after 24 weeks of treatment.
  • Each of the sitagliptin+pioglitazone groups was statistically significantly superior to the respective pioglitazone group, with respect to change in A1c, fasting plasma glucose, and two-hour postprandial glucose. These benefits were generally sustained over the entire 54-week study. For example, in the sitagliptin + pioglitazone 45 mg group, the percentage of patients with A1c <7.0% was 56.9% at week 24 and 56.4% at week 54.


Pio 15

Pio 30

Pio 45




LS Mean A1c ∆, %








LS Mean FPG ∆, mg/dl








  • Co-administration of sitagliptin seemed to blunt pioglitazone-associated edema but to increase pioglitazone-associated weight gain; rates of hypoglycemia were similar across groups (but were relatively high given the safety profile of the two drugs). Over the full year of the study, one or more adverse event was reported in 57.4%-64.2% of patients, one or more treatment-related adverse event was reported in 9.3%-14.6% of patients, one or more serious adverse event was reported in 2.5-7.0% of patients, and one or more serious, treatment-related adverse event was reported in 0.0%-0.5% of patients. Rates of unconfirmed hypoglycemia were roughly 10%, rates of symptomatic hypoglycemia were roughly 5%, and rates of severe hypoglycemia requiring medical assistance were as follows: 0.5% with sitagliptin alone, 0.5% with sitagliptin+pioglitazone 45 mg, and 0.0% in all other groups.


Pio 15

Pio 30

Pio 45




Mean Body Wt ∆, lbs








Edema, %








Questions and Answers

Q: Could you explain the relatively high rate of hypoglycemia? 10% seems high for two drugs that are not really expected to give hypoglycemia at all.

Dr. Henry: I agree. Generally one thinks of those agents as being very low in hypoglycemia. For symptomatic hypoglycemia, patients could report however they felt. Severe hypoglycemia requiring assistance of another is tough to explain. I didn’t think it was likely either.

Q: Do you know why Merck has stopped developing a fixed-dose combination of these drugs?

Dr. Henry: I think you’d have to ask Merck, not me. The drugs will continue to be indicated for use together. Why Merck decided last month not to make this combination is beyond me.


Symposium: What Did We Learn from the Recent Megatrials?

Initial Combination Therapy for Type 2 Diabetes: Is It Ready for Prime Time?

Bernard Zinman, MDCM, FRCPC, FACP (University of Toronto, Ontario, Canada)

Dr. Zinman gave a very persuasive and comprehensive presentation on initial combination therapy for type 2 diabetes. He built a strong case for the merits of combination therapy “right from the get-go,” and supported use of metformin/DPP-4s and metformin/TZDs (at a submaximal dose) – he was also very negative on use of sulfonylureas, emphasizing their lack of long-term durability. Dr. Zinman closed by mentioning the GRADE study, which will compare initial combination therapy with metformin plus either sulfonylureas, TZDs, DPP-4s, GLP-1s, or insulin. The study sounds well designed to answer not only ‘What therapy is best,’ but whether initial combination treatment (vs. sequential treatment) is actually better. We hope it begins soon, although the four-year follow-up period means we won’t see the data for a while.

  • In general, cellular pathways operate more like webs than superhighways – there are multiple redundancies and alternate routes that may be activated in response to the inhibition of a pathway (Woodcock et al., NEJM 2011). For this reason, combination therapies are oftenneeded to effectively treat conditions. While early use of combination therapy is well established and used from the get-go in cancer, hypertension, asthma, and tuberculosis, Dr. Zinman believes experts have not readily adopted combination therapy for type 2 diabetes.
  • Combination therapy for diabetes offers important advantages in efficacy, clinical practice, physiology, and side effects. Combination therapies provide more robust lowering of A1c, help avoid clinical inertia, and early combination therapy may improve beta cell function. Dr. Zinman also noted that using therapies with complementary mechanisms of action is particularly rational for combination therapy. Finally, he emphasized that combination therapy allows clinicians to use less than the maximal dose of individual agents, which helps reduce side effects (Zinman, American Journal of Medicine 2011).
  • “We end up with combination therapy anyways.” Dr. Zinman showed data from UKPDS on combination therapy demonstrating that three years after diagnosis, 50% of patients in the trial required combination therapy. At nine years after diagnosis, the number grew to 75%. Additionally, data on metformin monotherapy suggests that most patients fail to achieve glycemic targets. For those with an A1c less than 7% at diagnosis, 12% fail metformin each year; in patients with an A1c over 7%, the failure rate is 20% per year. At the 58-month mark following diagnosis, over 80% of individuals have failed metformin (Brown et al., Diabetes Care 2010).
  • “Clinical inertia is alive and well.” Dr. Zinman reviewed data from the US suggesting therapy intensification in patients with an A1c >8% takes 14 months with background metformin and 20 months with background sulfonylurea therapy. Additionally, Dr. Zinman mentioned a study from Shah et al. (Diabetes Care 2005) suggesting only 45% of specialists’ patients and 37% of PCPs’ patients had their therapy intensified when A1c was over 8%, and intensification to insulin therapy occurred in just 8.5% of patients seen by specialists and in just 1.7% of patients seen by PCPs.
  • If diabetes was a six month disease, sulfonylureas would be the therapy of choice.” Dr. Zinman was very negative on metformin/sulfonylurea combination therapy, citing the strong short-term efficacy but lack of durability of sulfonylureas in the long run, as well as more frequent hypoglycemia and weight gain. Dr. Zinman also mentioned data suggesting the adverse impact of sulfonylureas on beta cell function.
  • Initial combination therapy using metformin and a TZD (with the TZD at a submaximal dose) should be considered. Dr. Zinman argued that the small glycemic benefits of going to a full dose of a TZD are outweighed by the large increase in side effects. Using data he published in the Lancet in 2010, Dr. Zinman described how a submaximal dose of rosiglitazone combined with metformin was very effective in preventing diabetes, and was even comparable in efficacy to the results of the ACTNow study. However, the weight gain seen in ACTNow was not observed in his study because rosiglitazone was used at the submaximal dose.
  • Initial combination therapy with metformin/DPP-4 provides robust and durable reductions in A1c, no association with hypoglycemia or weight gain, lower GI side effects, and an enhanced GLP-1 response. Dr. Zinman showed data on sitagliptin, vildagliptin, saxagliptin, and linagliptin documenting the benefits of initial combination therapy with metformin. Interestingly, one study found lower rates of GI side effects with the combination compared to metformin monotherapy. He also noted that the complementary mechanisms of action associated with metformin/DPP-4 therapy leads to enhanced GLP-1 secretion compared to DPP-4 therapy alone.
  • The Glycemia Reduction Approaches in DiabetEs (GRADE) is a comparative effectiveness study examining five medications combined with metformin to achieve and maintain an A1c <7%. At this point, the study is “past the early planning stage” and will be chaired by Dr. David Nathan (Massachusetts General Hospital, Boston MA) – we last heard it mentioned in passing by Dr. Rury Holman at EASD 2011 (see page 197 of the full report at but Dr. Nathan also mentioned it at ADA 2011 – see page 86 of our full report at This seems like it is taking some time to get going – but of course careful planning and design is of the utmost importance. As a reminder, GRADE will be a randomized controlled trial in patients with recent-onset (<three years duration) type 2 diabetes – 7,500 drug-naïve or metformin monotherapy patients will be included in the study, which will compare the following medications combined with metformin: sulfonylureas, TZDs, DPP-4s, GLP-1s, and insulin. Two treatment strategies will also be compared: sequential therapy (start a therapy, wait until failure, add another) vs. early combination therapy from the get-go (Dr. Zinman’s preference). The expected follow-up period is four years. We’ve heard a lot of debate at this meeting about how to intensify therapy following metformin, but the default answer seems to be, ‘We need more data.’ We are very glad there is a trial in the works to answer this question, and we’re glad to see all of the major drug classes represented. We also like the trial’s comparison of sequential vs. early combination therapy – if the results pan out, it could have important implications for treatment algorithms in the future. That said, we suspect harder-to-take drugs will do far better in this trial than they do in “real life” – we look forward to hearing more about design and support and we hope the support is minimal since it usually is in real life. Although the four-year follow-up implies it will be a long time before the results are presented, we suspect it will be worth the wait. We hope if it takes much longer that other new drug classes (e.g., SGLT-2s) will be added as they are approved.

Questions and Answers

Q: All the trials conducted on type 2 diabetes – UKPDS, VADT, ACCORD – all of them are neutral towards macrovascular complications. Why so much stress on A1c?

Dr. Zinman: Those trials have done a terrible disservice to managing diabetes. They have downplayed the impact of improved A1c on microvascular disease. Microvascular complications can be devastating. We have made a huge impact in reducing kidney failure, blindness, and amputation. For me, that’s enough of a reason to aim for an A1c under 7%. We have good drugs for cardiovascular disease. We also have very large ongoing CV outcomes studies with DPP-4s and GLP-1s. But we should still target an A1c <7%. (Editor’s note – long term follow up data for UKPDS was not neutral toward CVD complications, it showed a benefit from intensification.)

Q: Is GRADE going to assess microvascular and macrovascular outcomes?

Dr. Zinman: The principal aim is to look at A1c, as well as biomarkers and surrogate outcomes – not micro- or macrovascular outcomes. The trial is not powered to do that and we would need a much larger study. Nonetheless, microvascular and macrovascular events will be documented.

Q: I would disagree about microvascular endpoints in the major trials. A meta-analysis looked at microvascular complications and found they weren’t affected in the pooled analysis. It was not significant to the point that 599 people out of 600 would not benefit from lowering A1c from 7.9% to 7%.

Dr. Zinman: Right – that was going from 7.9% to 7%. But most people are not at 7.9%. One needs a target, and if you have a target of <7%, many will still be above 7%.


Symposium: What Should Follow Metformin?


David Matthews, MA, BM BCh (University of Oxford, Oxford, United Kingdom)

Dr. Matthews spoke to a relatively full hall in the first of four segments assessing options for second-line therapy. His area was sulfonylureas (SFUs), about which he was very positive, a different tune from what we have been hearing from most in the last couple of years (even here). He went through each of the major concerns with regard to SFUs and addressed them. Besides being effective early in the disease course and having a long track record, their low cost is a particularly important consideration, given that the diabetes epidemic disproportionately affects low- and middle-income countries. While he said he was not against trying new therapies, he used a quote that was to the effect of: in excitement to try new therapies, let’s not in haste leave the old aside. (To which many would ask, but what if they cause weight gain, hypoglycemia, and beta cell burnout? We understand that often far more basic needs are more critical from an economic perspective – but we also point out that for many, the longer-term costs of using drugs like SFUs are only seen in hindsight.)

  • Dr. Matthews said that hypoglycemia is actually not common with sulfonylureas – more of an overdose effect than a side effect. “Although there is a big story about hypoglycemia with sulfonylureas, and you will get some when you move glycemia into the normal range, overall rates are nevertheless small.” Hypoglycemic events are rare in the UKPDS, he said, even using glibenclamide (which causes more hypoglycemia than other SFUs like gliclazide). In ACCORD, minor hypoglycemic events occurred 120 times in 100 patient-years with intensive therapy, and severe hypoglycemia only occurred 0.7 times in 100 patient-years with intensive therapy. We wondered about the study completed by the UK Department of Transportation that we learned about at EASD in 2010. A major takeaway from that study had been that hypoglycemia is widely underreported. There, we learned from Professor Anthony Barnett (the University of Birmingham, Birmingham, UK), that the study showed severe hypoglycemia was just as common in type 2 patients as type 1 patients (7%) and nearly as many people on SFUs (sulfonylureas) suffer from hypoglycemia as do those on insulin (20% vs. 22%). In that talk, Professor Barnett emphasized that there is massive underreporting of hypoglycemia in the UK; he implied this was because patients don’t want to share information with their doctors that could put their jobs at risk. A large (n=654 including 250 HCPs and the rest patients) follow up survey done by Merck found that 37% of patients experienced hypoglycemia while working. While 82% (likely over-reporting) of doctors claimed they always asked patients about hypoglycemia, nearly as many (81%) said they suspected patients did not tell them the truth. In 2009, Barnett said, there were over 5,000 patients with cases of severe hypoglycemia admitted to the emergency room in the UK; Professor Michael Nauck estimated at the same session there were 80 deaths a year in the US for SFU-induced hypoglycemia, based on four million patients still on SFUs.
  • Beta cell failure is not a convincing argument, according to Dr. Matthews. It is not a contraindication to early use and it is “not a huge problem in itself.” A comment in Q&A brought this point to contention, but Dr. Matthews strongly and convincingly stood his ground, pointing out that many people have misconceptions about the actual effect of SFUs on beta cells. In ADOPT, it looks like you’re getting quicker beta cell loss, because you get a great effect with glibenclamide early on but then it declines with time to become worse than metformin and rosiglitazone in terms of A1c at five years. The way it’s often shown plotted on a graph, with anabridged y-axis, it looks like this is a large difference in A1c. However, Dr. Matthews noted that if you zoom out to see the whole y-axis, you see how small the difference actually is – less than 0.3% A1c at five years (not to mention that glibenclamide is known to fail earlier than other SFUs). In fact, A1c averaged over this time period is lower in patients on glibenclamide than on metformin or rosiglitazone, because of that very strong initial decrease.
  • Weight gain happens in people who are normal weight, not overweight. Weight gain is another popularly cited problem with SFUs. Yet, in the intensive group in ACCORD, when patients were stratified according to BMI, weight gain occurred only in people with BMI under 25 and not in those with BMIs 25-30 or 30+. People with BMIs over 30 actually tended to lose weight. Weight gain too then, according to Dr. Matthews, is not a convincing attack on SFUs.
  • The effect of drug cost should not be underestimated, Dr. Matthews said, especially in this day and age. Exenatide is about $1,200 per year and sitagliptin about $600 per year, while in comparison glibenclamide and gliclazide are close to zero. One month of treatment with exenatide would pay for 16.7 years of glibenclamide or 8.3 years with gliclazide. In low-resource countries, this is very attractive, and as we heard it, the implication was that this was even the more responsible choice.

Questions and Answers

Q: By the time a patient gets to us, 75% of beta cells are dead. Then by the time metformin fails, 10-15% of beta cells might be left. Sulfonylureas are cytotoxic – putting them on it kills the rest of the beta cells. As a physician, my first mission is to do no harm. How can we do this to patients? It is unethical.

Dr. Matthews: What you know is wrong. Beta cells are not killed. These studies look at beta-cell function, not death. As we know, if you do bariatric surgery you can resolve diabetes rather quickly, which tells you absolutely that beta cells aren’t dead – these people can have restoration of function in one or two days before they even leave hospital. You need to follow the best evidence base – and that is UKPDS for early intervention. It is terribly important not to distress your patients by saying beta cells are dead. We can wake beta cells up with so many things: sulfonylureas, bariatric surgery, GLP-1 receptor agonists, DPP-4 inhibitors. The research agenda should really be, “what puts them to sleep?”

Q: Will the minor hypoglycemic events like in ACCORD contribute to cognitive decline and weight gain?

Dr. Matthews: The problem with hypoglycemia research is that you ask patients at infrequent intervals if they felt funny, etc. People often feel funny for various reasons; I sometimes feel funny. They tend to go off and eat and then it’s too late for you to confirm with a blood glucose measurement. So, minor hypoglycemic events are very difficult to capture. I’m sure cognitive impairment in major hypoglycemia is a serious concern, but you don’t often get major hypoglycemia. Nevertheless, you’re right to raise that as an issue.

Q: What are your thoughts on the fact that sulfonylurea receptors are present in the GI tract, blood vessels, and the brain?

Dr. Matthews: The pleiotropic effects of sulfonylureas will differ among different drugs. You’re right, they don’t bind only to the beta cell. But we would need more research to know how they affect other tissues. But the interesting thing about understanding the sulfonylurea receptor is that it opens another area of research into drugs that are more selective, can give you less hypoglycemia, and so on.



Bernard Charbonnel, MD (University of Nantes, Nantes, France)

Dr. Charbonnel did not purport to say pioglitazone (Takeda’s Actos) should be second-line treatment for all or even most patients, but his goal was to convince us that there were some patients in which pioglitazone might be the best option. There have been more than 15,000,000 patient years of glitazone use, and it is reassuring that we have long term real-life follow up. Pioglitazone will also go generic later this year. Dr. Charbonnel passionately debunked the concerns people tend to have about TZDs and convincingly suggested that in particularly insulin insensitive patients, especially those with cardiovascular risk factors (but not heart failure) or chronic kidney disease, pioglitazone is likely to be very useful. Despite his assigned drug class, he mentioned that most patients, barring cost considerations, would end up on DPP-4 inhibitors as second line.

  • If cost weren’t an issue, DPP-4 inhibitors would be second-line drug for most patients, Dr. Charbonnel said. He said he believed SFUs are incorporated into guidelines as second-line treatment only because of cost considerations. Pioglitazone would be an option in patients who do not respond to DPP-4 inhibitors. It might also be preferable to incretin-based therapy under certain criteria, namely if a person is known to have marked insulin insensitivity. In patients with at least two out of the following markers of insulin resistance, pioglitazone might be predicted to show good response: abdominal obesity, increased liver enzymes, and/or low HDL. He said that in these select patients, one may see remarkable and durable drops in A1c. The case for pioglitazone is especially strong for patients at high cardiovascular risk, he said, such as those who are post-MI (with no congestive heart failure) or post-stroke. It is also strong in patients with chronic kidney disease.
  • Heart failure is nonfatal, easily treatable, and should not occur if one prescribes smartly. It arises from fluid retention, a well-known side effect of TZDs, and even though PROactive saw more serious heart failures in pioglitazone than placebo, they were without prognostic consequence in terms of mortality. This has been corroborated in other studies. The heart failure in TZDs, said Dr. Charbonnel, represents decompensation of pre-existing dysfunction that is easy to manage (stop pioglitazone, add diuretics), and it can be avoided simply by avoiding pioglitazone in patients with heart failure. Even in patients with heart failure, there was no increase in mortality in those on pioglitazone, according to a 2007 meta-analysis in the British Medical Journal, he pointed out.
  • In fact, evidence suggests that pioglitazone is cardioprotective. The 10% risk reduction in the composite endpoint of PROactive was not significant because it was too short of a study, according to Dr. Charbonnel, and because the composite endpoint included considerations of the diabetic foot which many diabetologists know, he said, is not due to the same mechanism as other vascular complications. The more important endpoint of “death, MI, or stroke” was significant at 16% relative risk reduction. The 47% relative risk reduction in strokes in patients with a previous stroke was particularly dramatic, and pioglitazone should especially be considered in post-stroke patients.
  • Pioglitazone-associated weight gain is without any metabolic consequences. It is well known that treatments causes fat redistribution from visceral to subcutaneous fat. This is especially helpful for non-alcoholic fatty liver, where fat in the liver decreases by about 50% with pioglitazone treatment and insulin resistance in the liver is massively reduced.
  • Increased bladder cancer risk is very low, even if it is there. Analysis of the Kaiser Permanente Diabetes Registry showed that the increase in the risk of bladder cancer wasstatistically significant, though small, in patients who had been taking pioglitazone for over two years. The absolute risk is very small, amounting to only three extra events in 10,000 patient- years, compared, presumably, to many more patients who see benefit.
  • There is a possible disease-modifying effect of glitazones. In other words, there may be some kind of beta cell preservation. In ADOPT, TZDs were the most durable therapy, and in ACT- NOW, pioglitazone reduced diabetes incidence by 80% at 2.5 years in pre-diabetic subjects. In ACT-NOW, the pioglitazone and placebo curves actually diverge rather than shift, suggesting a disease-modifying rather than disease-masking effect.

Questions and Answers

Q: You said to choose patients that were insulin resistant. Is this based on any evidence that these patients do better?

Dr. Charbonnel: Yes. We did an analysis of insulin-treated patients in PROactive, of which there were about 2,000. When you add pio on top, as expected you get a decrease in A1c with fewer doses of insulin. But we split these patients into quartiles at baseline. On one end you have people who have good A1c on low doses of insulin – we called them insulin-sensitive patients. If you add pio on top to these patients it was without clear effect. On the other end are patients who at baseline were not well controlled but on high doses of insulin – in other words, insulin-resistant patients. Adding pio on top of insulin to these patients resulted in a dramatic drop in A1c, by almost 2%, and this was without changing insulin regimens (in the placebo arm, to get any decrease, you had to increase insulin doses). So I think we can conclude from that and from clinical experience that it works well in insulin-resistant individuals but not in the insulin-sensitive.



Richard Pratley, MD (Florida Hospital Diabetes Institute, Orlando, FL)

Dr. Pratley did not actually argue that DPP-4 inhibitors and GLP-1 analogs should be added as second- line after metformin, though he probably could have without much trouble given all the positivity surrounding incretins in the diabetes pharmacopoeia. Incretins have many benefits, and many of the same benefits as metformin itself – they are somewhat durable, weight-neutral or -beneficial, and have low risk of hypoglycemia. Dr. Pratley feels that incretins as add-ons to metformin are almost there, but not quite. When starting incretins, he said in Q&A, he would start with a long-acting GLP-1 agonist.

  • Two weeks ago, an article in NEJM by Budnitz et al showed that among the top drugs that lead to hospitalizations, insulin is number two just behind warfarin, and oral hypoglycemic agents are in the top five as well. Hypoglycemia is a major problem in diabetes, and incretin-based therapies of course avoid this.
  • Metformin and DPP-4 inhibitors seem to have complimentary mechanisms. This might suggest that DPP-4 inhibitors are the natural next step after metformin. In terms of increasing GLP-1, they are more potent synergistically than either alone. All the available DPP-4 inhibitors are similar, with not many differences besides clearance – linagliptin (BI/Lilly’s Tradjenta) does not need to be adjusted in chronic kidney disease.
  • Thyroid C-cell tumors in liraglutide and pancreatitis in incretins are not major concerns. C-cell tumors were seen in mice and rats, not primates, and there have been no cases of C-cell tumors in humans. Regarding pancreatitis, large observational studies do not seem to show increased risk of pancreatitis.
  • Longer-acting GLP-1 analogs are better, said Dr. Pratley. Exenatide LAR and liraglutide have lower rates of nausea, which seem to be highest in exenatide dosed twice daily (as nausea is likely related to GLP-1 peaks), and the longer-acting drugs also produce better A1c reductions.
  • In terms of adding incretins on to metformin, “I would say we are almost there.” GLP-1 mimetics have some ways to go with tolerability, he said, referring to nausea. And we still don’t know about longer-term cardiovascular safety or durability, he said, as trials only go out to two to three years. Liraglutide is associated with an increase in heart rate, and longer trials are underway to examine longer-term cardiovascular safety for liraglutide and other incretins. Regarding durability, “As you might expect, it turns out DPP-4 inhibitors are not quite as durable as TZDs,” Dr. Pratley said. And regarding cost: “They clearly fall down in the cheap category – they’re not.”

Questions and Answers

Q: You mention pancreatitis. What would you advise about following up on this once you start DPP-4 inhibitors? And what about with pioglitazone and bladder cancer?

Dr. Pratley: Well, I tell patients about the pancreatitis because I know they’ll read it on the web. I say that in general, people with diabetes have higher risk, and if you have any symptoms then report to the clinic immediately. In terms of bladder cancer, I wouldn’t start pioglitazone if the patient had a history of bladder cancer. Otherwise though, the small increased risk has not changed my practice.

Q: If you were given a choice of using one incretin-based therapy in an individual for whom cost is not an issue, what would be your first, second, and third choices?

Dr. Pratley: I’d favor a long-acting GLP-1 agonist – for convenience, improved efficacy, and the weight loss effect, which most of my patients would benefit from. If we can get them over the nausea, which is not unlike the challenge with metformin, people do fairly well.

Q: Are there any data on risk to the kidney, such as incidence of microvascular events or progression of chronic kidney disease?

Dr. Pratley: Proteinuria is usually looked at in clinical trials and mostly there is no effect of GLP-1 analogs or DPP-4 inhibitors on progression of proteinuria – but you wouldn’t expect any effects in these short trials. In exenatide, there have been reports of acute renal failure, but it seems to be more prerenal [Editor’s note: in other words, occurring from low plasma volume, such as in dehydration] rather than from a direct toxic effect of the drug.



Leon Litwak, MD (Argentina Diabetes Federation, Buenos Aires, Argentina)

Dr. Litwak argued that insulin must have an important role early in the treatment of type 2. His mantra seemed at first to be that we have to maintain glucose “as low as possible, as early as possible, as intensive control can reverse diabetes in some patients and prevent more beta cell damage from glucotoxicity. But then he strongly added “as safely as possible,” saying that in most people insulin should not be second-line therapy. He said he uses the AACE/ACE treatment algorithm (for under A1c of 9.0%) rather than the ADA/EASD algorithm, for this very reason – because in the former, insulin does not appear as a second-line treatment. He flashed a slide of degludec for about a second, saying that newer insulins may produce “better control without so much hypoglycemia.” His bottom line was that though we should be cautious with insulin, it should nevertheless not be considered the last resort in our arsenal, as it is sometimes treated in practice. In Q&A, Dr. Litwak revealed he too might choose a DPP-4 inhibitor as his second-line treatment.

Questions and Answers

Q: What do you think about intermittent insulin therapy, abandoning it once you’ve reached substantial control?

Dr. Litwak: There is this new option, which entails starting very early with insulin treatment and then you take out the insulin and continue with orals, such as metformin plus a DPP-4 inhibitor or something like that. It could work as a temporary treatment at the beginning of disease.


Corporate Symposium: GLP-1 Therapy in Type 2 Diabetes: Recent Advances and Controversy (sponsored by Lilly Diabetes)

Opening Remarks

Bernard Zinman, MDCM, FRCPC, FACP (University of Toronto, Ontario, Canada)

Dr. Zinman opened the symposium with a brief outline of the agenda. We were interested to hear him discuss Lilly’s newest innovation in corporate symposium technology: placing QR codes in the handout book to allow audience members to obtain key slides from speakers’ presentations (we would have tested this out, but our US cell phone technology was not advanced enough to work in Dubai!). In addition to this novel approach, Lilly is also using another smart technology and marketing strategy at IDF 2011: to use the conference’s wireless, attendees must visit the Lilly booth to get the network password.


Which Should Be Used to Intensify After Oral Antihyperglycemic Therapy - Basal Insulin

Guntram Schernthaner, MD (University of Vienna, Vienna, Austria)

Dr. Guntram Schernthaner opened the debate by arguing in favor of basal insulin after oral therapy failure, stating, “Michael Nauck is more than biased. He is fighting for his own baby.” In a straw poll of the audience, roughly two-thirds of respondents said they would use a GLP-1 receptor agonist (GLP-1 RA) following oral antihyperglycemic therapy, while the remainder would go with insulin. Using four head-to-head GLP-1/glargine studies (including one from debate opponent Dr. Michael Nauck!), Dr. Schernthaner systematically addressed just about every therapy consideration. He started with efficacy data and built a case for comparable A1c benefits of GLP-1 therapy and basal insulin. He next cited adherence and side effect data in favor of basal insulin, and he spent significant time arguing that basal insulin had the same rate of hypoglycemia as GLP-1 RAs – while the four studies he cited backed him up for the most part, this seemed like a potential case of cherry picking favorable data in our view. Most memorable was his discussion of weight effects, where Dr. Schernthaner argued that weight loss actually has a detrimental impact on cardiovascular outcomes. That’s right – he argued basal insulin deserves the edge in the weight department precisely because it causes weight gain. Despite this unconventional tack, Dr. Schernthaner showed interesting data from a meta-analysis of five major trials (UKPDS, PROactive, ADVANCE, VADT, and ACCORD), specific data from PROactive, and the TRIAD study suggesting that those with higher BMIs had better CV and mortality outcomes – we presume this was a chance association and not a causal link as Dr. Schernthaner seemed to imply. He closed with the commonly cited cost argument by noting the significantly higher cost of GLP-1 RAs in Austria relative to basal insulin.

  • Dr. Schernthaner used four head-to-head GLP-1/glargine studies to build a case for the use of basal insulin over GLP-1 following intensification after oral therapy: Heine et al.,Ann Intern Med 2005; Nauck et al., Diabetologia 2007; Russell-Jones et al., Diabetologia 2009; and Diamant et al., Lancet 2010.


Which Should Be Used to Intensify After Oral Antihyperglycemic Therapy - GLP-1 Receptor Agonist

Michael Nauck, MD (Diabeteszentrum Bad Lauterberg, Harz, Germany)

Dr. Nauck countered Dr. Schernthaner by supporting intensification with GLP-1 receptor agonists (GLP-1 RAs) after oral therapy failure. In a similar fashion, he addressed a wide range of factors that affect therapy choice. On efficacy, Dr. Nauck asserted that GLP-1 RAs have equal or even slightly better A1c improvement versus basal insulin. He next cited the very low rates of hypoglycemia with Bydureon, the glycemic dependent action of GLP-1, and the well-documented weight benefits of the GLP-1 RA class. Dr. Nauck also passionately argued for the convenience of GLP-1 RAs, emphasizing the simple and relatively infrequent dosing (compared to basal insulin), the challenges of titrating insulin (“It’s an art to treat patients with the right amount at the right time”), and the little necessity for glucose monitoring with GLP-1. Dr. Nauck lastly addressed costs, arguing that liraglutide 1.2 mg is the same price as 50 units of basal insulin and supplies (“You can’t use costs as a consideration”). To compare the whole package, Dr. Nauck developed a scoring system (ranging from -2 [worst possible] to 2 points [best possible]) to look at the bundled criteria: A1c, complications, mechanism of action, adverse events, cardiovascular effects, weight, costs, hypoglycemia, blood glucose testing, durability, etc. After scoring both GLP-1s and basal insulin, Dr. Nauck’s system resulted in a score of one for insulin and six for GLP-1 RAs – while we thought this was quite creative, the subjectivity of this last analysis left much to be desired. Nevertheless, we thought Dr. Nauck more persuasively argued his side of the debate, and given the audience poll after his talk, two-thirds seemed to agree with him that GLP-1 is the way to go.


Panel Discussion

Moderator: Bernard Zinman, MDCM, FRCPC, FACP (University of Toronto, Ontario, Canada)

Guntram Schernthaner, MD (University of Vienna, Vienna, Austria), Michael Nauck, MD (Diabeteszentrum Bad Lauterberg, Harz, Germany);

Dr. Schernthaner: Great job, Michael. I would say that some of the study protocols unfairly favored GLP-1 because they used low doses of basal insulin. And what is your opinion on the increase in heart rate seen in some studies?

Dr. Nauck: I’ve been a part of these studies. The protocols were very fair. They gave guidance on titration. If you look at the fasting glucose achieved, it’s not so much off the target. I believe this reflects the clinical practice of using insulin. If anything, it proves it’s not easy to use insulin to reach the glycemic targets. It’s another argument to use a drug where dosing is not so important.

Dr. Schernthaner: But why was the dosing 50% less?

Dr. Nauck: I do not know why. I don’t know whether the investigators had a bias. The protocols were easy, repeated instructions with all the consequences, I think there are other studies, where you find similar insulin doses. It’s not the best use of insulin. Maybe that just proves how difficult it is to use insulin. We want treatments with more ease and without complications. On heart rate, this is a concern. This is based on studies in certain cohorts of people, and it’s based on their cardiorespiratory fitness. It wasn’t a pharmacologically changed heart rate. Heart rate is representative of fitness. I would not be so concerned.

Dr. Zinman: What about the atherogenic effect of insulin? And is there increased cancer with glargine?

Dr. Schernthaner: For glargine, there is this huge debate. I believe high insulin dosages are problematic in obese patients and this is a clear advantage of GLP-1. I was wondering why Michael didn’t take this into his argument. This is the long-term question that we cannot answer at this moment. We don’t have any head-to-head outcome studies. This is a topic that’s especially important when thinking about the safety of insulin in patients with a history of cardiovascular disease. I believe that many patients cannot be treated with GLP-1 – at the end, most patients will need insulin. Michael and I will have the same discussion ten years later: they will need insulin later on. GLP-1s are effective at the beginning, but after some period they are less effective.

Dr. Nauck: We do a lot of insulin treatment at our institution: 80% of patients, maybe more. The majority of patients come to the hospital, get good education, are instructed to intensify insulin regimens, are titrated to target, and go home. When they come back, they have an A1c of 9% and are not responding to glucose readings. This is not a small population of patients. Insulin treatment is not so easy – we have to accept it as a fact. Against this background, it’s easier to get decent diabetes control with GLP-1 in some patients.

Dr. Schernthaner: If you look at large groups, you see the same effect of GLP-1 and insulin. In a large database, I think it’s more related to the patients and less to the drug.

Q: Can you explain the combination use of insulin and a GLP-1 agonist?

Dr. Nauck: There is experience and studies to suggest that if you titrate well, you will be successful. This would involve using a long acting GLP-1 to get better fasting control, and use exenatide for postprandial control. John Buse and colleagues brought the average A1c way below 7%. Insulin should make weight go up and with exenatide weight goes down, but net effect of combination therapy is weight loss. If costs don’t play an undue role, a combination of a long acting insulin and a short acting GLP-1 will be the most potent regimen to treat difficult patients with type 2 diabetes in the future.

Dr. Schernthaner: I presented this over the summer, and I believe that this will be the future for the complicated patient, particularly if we have one syringe with the same treatment.

Q: Why initiate a GLP-1 if the patient is going reach insulin anyways?

Dr. Nauck: We don’t know if insulin truly is the last resort if we see lots of patients with bad control on suboptimal insulin regimens. Maybe they could benefit from GLP-1.

Q: Can you speak about the magnitude of weight gain with insulin or the weight loss with GLP-1?

Dr. Nauck: It’s very bearable. Not all patients gain a lot on insulin, and not all lose a lot on GLP-1.

Q: A note on diabetes duration. I can share my experience in Italy, where all patients are registered in a nationwide registry. Only specialists can prescribe incretin therapies. We did a two-year follow-up of exenatide in 21,000 patients. After one year of follow-up, those with diabetes for longer than 10 years were performing worse than people with a shorter duration of diabetes. After two years of follow-up, the two groups were matching together.

It was not that younger people became worse, but the longer-duration group was gaining. After two years, it looks like the groups were performing equally well.

Dr. Nauck: This is backed by proof-of-principle studies.


Corporate Symposium: Incretins in Diabetes – From Pathophysiology to Therapeutic Use (Sponsored by Novo Nordisk)

Further Intensification: Present and Future

Stephen Gough, MD (University of Oxford, Oxford, UK)

Dr. Gough provided an overview of therapy intensification following metformin, highlighting the differences between basal insulin, GLP-1 analogs, and the potential combination of the two. He started by discussing the choice between therapy intensification with GLP-1 or basal insulin. Using the results of the LEAD-5 study, he noted that liraglutide demonstrated greater A1c decline and a better effect on weight compared to insulin glargine. Dr. Gough also took the opportunity to address Bydureon, reminding the crowd of the drug’s failure to demonstrate non-inferiority to liraglutide in DURATION-6 (see pages 32-34 of our EASD full report at He next turned to the “intriguing question” of combining a GLP-1 with basal insulin, noting that recently published trials have demonstrated the benefits of doing so: Buse et al., Ann Intern Med 2010 and GetGoal-L presented at ADA 2011 (see page 30 of our ADA full report at To further bolster the case, he gave a detailed overview of a 26-week extension trial comparing liraglutide alone to insulin detemir plus liraglutide (see page 27 of ADA 2011 Full Report for our detailed coverage at He closed by summarizing his beliefs that 1) GLP-1s should be considered early on after metformin and prior to insulin, and 2) treatment can be successfully intensified by adding a basal insulin to a GLP-1, and doing so can provide sustained improvements in glycemic control, weight benefits, and a lower risk of hypoglycemia. (Editor’s note – this presentation closely resembled one we saw at EASD; see page 65 of our EASD 2011 full report at

  • Dr. Gough gave a detailed overview of a 26-week extension trial comparing liraglutide alone to insulin detemir plus liraglutide (see page 27 of ADA 2011 Full Report for our detailed coverage at As a reminder, after a 12-week run-in period with 1.8 mg liraglutide, individuals who were unable to achieve an A1c of < 7.0% (average A1c was 7.6%) were randomized to intensify treatment with the addition of insulin detemir or to continue with liraglutide plus metformin. At 26 weeks, the detemir arm achieved greater additional reductions in A1c (-0.51% vs. + 0.02%; baseline post run in was 7.6%), similar weight loss (-4.0 kg [-8.8 lbs] vs. -4.7 kg [-10.4 lbs]), and a similar rate of hypoglycemia (≈0.2 events/patient-year) as compared with the non-detemir arm. Dr. Gough added that A1c levels and weight loss remained stable throughout the 26-week extension phase.


Insight into Guidelines: An Interactive Discussion

Bernard Zinman, MD (University of Toronto, Canada) and Melanie Davies, MD (University of Leicester, Leicester, United Kingdom)

Dr. Zinman closed the symposium by discussing the ADA/EASD consensus statement (Nathan et al., Diabetes Care 2008), which took a unique “prescriptive approach” to managing type 2 diabetes. He began by noting the one part of the algorithm that all participants unanimously agreed on: starting with both lifestyle intervention AND metformin at diagnosis (“No other algorithm has that”). Dr. Zinman next proposed a number of personal modifications to the algorithm, which included 1) renaming ‘tier 1’ and ‘tier 2 ’therapies ‘option 1’ and ‘option 2’ (“They’re both good options”). Dr. Zinman was greeted with hearty applause from the audience after stating his second modification: sulfonylureas should be cut out of the algorithm. In his view, the only benefit of the class is that it’s “cheap, cheap, cheap,” but the hypoglycemia, weight gain, and lack of durability all support discontinuing its use. Dr. Zinman concluded by arguing for use of pioglitazone only when used at less than the maximal dose and use of GLP-1 and DPP-4s prior to insulin. In the discussion that followed, Dr. Davies pushed him to clarify his criticism of sulfonylureas (he believes glipizide and glimepiride are “clearly superior” to first-generation secretagogues).

Questions and Answers

Dr. Davies: I want to challenge you on sulfonylureas. We must deal with this issue of sulfonylureas: doctors are familiar with them, they are inexpensive, and there is good access to them around the world. How does this compare to incretin therapies?

Dr. Zinman: For incretins, there are ongoing, very large clinical trials that demonstrate their effectiveness, and perhaps a reduction in cardiovascular outcomes. They are very well tolerated, although the question is whether they have long-term safety. I think once we demonstrate that, we’ll see incretins replace sulfonylureas. And let’s be clear – not all sulfonylureas are created equal. Glipizide and glimepiride are clearly superior to first-generation secretagogues. When I have a patient on metformin that needs an added intervention, I consider many things. What is the patient’s weight? How important is weight reduction? How much of a barrier is an injection? I think this has reduced substantially. I give patients a choice of an oral incretin therapy or an injectable therapy. I provide the differences with respect to efficacy and weight loss. For some people, that’s an easy question.

Dr. Davies: Going down the treatment algorithm, I’m wondering about GLP-1 and insulin. What’s your favorite combination and what do you go with first?

Dr. Zinman: I go with GLP-1 before insulin. The weight loss and dosing is so much easier. These are therapies that can be used by family physicians, as they’re much easier to implement and dose. They also have the added value of no risk of hypoglycemia and weight gain. Adding insulin to a GLP-1 or vice versa is where I think we’re going in the future.


Corporate Symposium: The Evolution of DPP-4 Inhibition in Type 2 Diabetes (sponsored by Merck Sharp & Dohme)

Combination Therapy: Aligning Treatment Goals with Therapeutic Options

John Prins, MD (Mater Medical Research Institute, Brisbane, Australia)

Dr. Prins began his discussion of combination therapies by noting the mechanisms of hyperglycemia: insulin resistance, insulin deficiency, impaired beta cell function, unsuppressed hepatic glucose output, and an impaired incretin response. He then highlighted the therapies that best address each of these areas: metformin and TZDs (insulin resistance; unsuppressed hepatic glucose output), sulfonylureas and insulin (insulin deficiency), DPP-4s, TZDs, incretins (beta cell dysfunction; impaired incretin response). Dr. Prins concluded with a discussion of pleiotropic and adverse events, which in his view support use of metformin/DPP-4 combination therapy and argue against use of either sulfonylureas or TZDs. Despite his opinions against use of TZDs during the presentation, we were interested to hear Dr. Prins recommend their use in a number of cases during Q&A.

Questions and Answers

Q: What about use of sitagliptin for gastroparesis?

Dr. Prins: The impact of DPP-4 on gastric emptying is minimal. I don’t think there is any consideration, and this should not be a significant determinant of what drug combination to use.

Q: Besides the beta cell, there are other cells that have GLP-1 receptors like the liver. What is the effect on these cells?

Dr. Prins: GLP-1 receptors are ubiquitous. There are off-target effects on the liver and cardiovascular effects that are also under investigation. Most of the data is in animals. The effect in the liver can be explained as a result of glucagon changes.

Q: What about use of sitagliptin for extreme insulin resistance?

Dr. Prins: For insulin resistance, TZDs are the most efficacious, followed by metformin.

Q: For a newly diagnosed individual with type 2 diabetes, what combination therapy would you use to delay progressive beta cell decline?

Dr. Prins: The best data is going from the impaired glucose tolerance state and progressing to type 2 diabetes –TZDs and metformin. The question for established diabetes is slightly different. There is not much information. Logically, insulin sensitizers would be the way forward.

Q: For predicting the response to sitagliptin, you mentioned that it’s more effective for long duration of type 2 diabetes. Why is that?

Dr. Prins: The mechanism is not known, but has been reproduced in more than one study. The glucagon effect is much greater than we anticipated. The drugs that manipulate the incretin system have reinforced the importance of glucagon in our patients. It’s counter intuitive. We would anticipate drugs would be more effective earlier, but that’s not the case here.

Q: Can you comment on DPP-4s after bariatric surgery?

Dr. Prins: I have not seen published work on this. Anecdotally, there is lots of bariatric surgery in my clinic. The drugs are used, although there are ongoing concerns for insulin users about hypoglycemia.

Q: How about a DPP-4 and statin combination?

Dr. Prins: The fixed dose combination is on the market in the US. The combination therapy provides exactly the predicted response in terms of A1c and lipid response with no increase in adverse events.

Q: In type 2 diabetes, do you start with lifestyle, metformin, or sitagliptin? Should orals be started right away?

Dr. Prins: I would ask the patient in that circumstance. Sulfonylurea might accelerate beta cell failure. After effective dual combination therapy, I would move to insulin. A patient needs to be happy with what it decided.

Q: Can you comment on the beta cell response after bariatric surgery? Do beta cells lie dormant with morbid obesity and diabetes, and come back after surgery?

Dr. Prins: This has been highlighted as one of the reasons for improvement in beta cell function. The type of surgery can influence that. Lap banding is less effective than other procedures. There is growing evidence of an incretin contribution – changing gut motility, particularly in the distal gut, can account for that. Overall, yes, there appears to be dormant beta cell function and capacity for function later. This has been partly shown by the lifestyle intervention studies.


3. Insulin and Non-Incretin Therapies

Oral Presentations: Insulin Therapy – From Monitoring to Control

Flexibile Once-Daily Dosing of Insulin Degludec is as Efficacious and Safe as Insulin Glargine Given the Same Time Each Day in Type 2 Diabetes

Stephen Bain, MD (Swansea University, Swansea, United Kingdom)

Dr. Bain reviewed the results of a flexible dosing trial from insulin degludec’s phase 3 program, which we previously covered in a poster at ADA 2011 (see page 163 of our full report at and an oral presentation at EASD 2011 (see page 119 of our full report at We were interested to hear many questioners attack the study design during Q&A. Additionally, it remains to be seen whether flexibility and a reported reduction in nocturnal hypoglycemia will be enough to differentiate degludec from Sanofi’s glargine. We’ll certainly be interested to see how Novo Nordisk markets and position the insulin once it is launched.

Questions and Answers

Dr. Allan Vaag (Steno Diabetes Center, Copenhagen, Denmark): What is the pharmacodynamic profile of insulin degludec in type 2 diabetes patients? How does it compare to glargine?

Dr. Bain: We’ve been debating for years whether we have a true once-daily basal insulin. With degludec, I think we can truly say that we have once-daily basal insulin. The question is, what advantage will that offer to patients? Certainly, one is flexibility with respect to lifestyle.

Dr. Vaag: With glargine, you assumed that it had to be taken at the same time every day. Couldn’t you have achieved flexible dosing with glargine as well?

Dr. Bain: We don’t have that experimental data. The label for glargine says that this is what should be done. For degludec, the rationale for flexible dosing was the long profile.

Q: What about hypoglycemia during the daytime? Was it increased during the day with degludec?

Dr. Bain: Overall, the rate of hypoglycemia was the same – there was no significant difference between the two insulins. The hypoglycemia difference tends to emerge over the nocturnal period. I note that nocturnal hypoglycemia is a tricky and difficult issue because you only get a documented hypoglycemia with confirmation of a blood glucose or third-party assistance. It’s hard to know how many patients are having hypos that aren’t documented. This area does require additional research.

Q: Glargine has traditionally been used at night because of the dawn phenomenon in patients with type 1 diabetes. However, I always use glargine in the morning to avoid the peak that would occur in the middle of the night with evening dosing. It would be useful to know what percentage of patients was taking glargine at night vs. in the morning.

Dr. Bain: I fully agree. It would be useful data to know when the glargine was administered. I didn’t mention it because I don’t know the answer to that question. In this comparison, degludec was given at night with the final meal of the evening, allowing a comparison of degludec to glargine. In my own practice, I have many patients giving glargine in the morning as well. In the UK, evening administration of glargine tends to be the norm.

Q: In a study of insulin degludec in type 1 diabetes, you showed less nocturnal hypoglycemia with a fixed dose time. Should fixed or flexible dosing be used?

Dr. Bain: The idea was to show whether flexible dosing is possible. If that’s something patients want to use, they can do it.

Q: But isn’t the problem of flexible dosing that its gets too chaotic?

Dr. Bain: Yes – I don’t think flexible dosing would be a treatment recommendation, but rather, an option for patients if they need it.

Q: Is there any experience with insulin degludec in children with type 1 diabetes?

Dr. Bain: There will be trials in younger patients, but the focus has been on adults with type 1 diabetes.


Oral Presentations: Late-Breaking Abstracts Session

Active-Controlled Dose-Finding Study of Efficacy, Safety, and Tolerability of Multiple Doses of Ipragliflozin in Type 2 Diabetes Patients

Vivian Fonseca, MD, FRCP (Tulane University Medical Center, New Orleans, LA)

Dr. Fonseca presented 12-week results from a dose-finding study of ipragliflozin, Astellas Pharmaceuticals’ SGLT-2 inhibitor. The study enrolled patients with type 2 diabetes (n=412), whether drug-naïve or treated with oral antihyperglycemic agents (the former went through a six-week washout period). Patients were randomized to placebo, metformin, or one of four ipragliflozin doses (12.5, 50, 150, or 300 mg once daily). Ipragliflozin caused dose-dependent glycemic benefits that were statistically significantly better than those of placebo and, for doses ≥ 50 mg, comparable to those of metformin (i.e., roughly 0.5% from a baseline of roughly 8.0%. Ipragliflozin also caused a trend toward dose-dependent weight loss, with a notable magnitude given the study’s short duration (5.1 lbs for the 300 mg dose vs. 1.5 lbs for placebo). All groups were comparable in rates of treatment-emergent adverse events (TEAEs) as well as urinary tract infections (UTIs), but genital tract infections (GTIs) were more common with ipragliflozin than placebo (4.1% vs. 1.4%). We think that safety and tolerability will be critical for the success of any SLGT-2 inhibitor given the relatively modest glycemic benefits in the class; thus far it is not clear how particular candidates compare to each other in this regard.

  • Patients with type 2 diabetes (n=412) were randomized to receive one of four doses of ipragliflozin (12.5, 50, 150, or 300 mg once daily), placebo, or metformin (begun at 1,000 mg and titrated to 1,500 mg). The study included both drug-naïve patients and those previously on oral antidiabetic medicines (the latter had a six-week washout period prior to the start of the trial). Before the treatment period began, all patients were on a two-week placebo run-in.
  • Over 12 weeks, ipragliflozin caused dose-dependent glycemic benefits that were statistically significantly better than those of placebo and, for doses ≥ 50 mg, comparable to those of metformin. A trend toward dose-dependent body-weight reduction was seen with ipragliflozin; we look forward to longer trials that can shed more light on weight- loss effects.

Pbo n=69

Met n=69

12.5 mg n=70

50 mg n=67

150 mg n=68

300 mg n=68

Mean Baseline A1c, %







LS Mean Change in A1c, %*







LS Mean Change in FPG, mg/dl*







% Patients Achieving A1c < 7.0%*







Mean Change in Body Weight, lbs







*p<0.5 between ipragliflozin and placebo

  • Rates of overall treatment-emergent adverse events (TEAEs) were similar among groups, with higher prevalence of genital tract infection in the ipragliflozin groups relative to placebo and metformin. Hypoglycemia was rare in all groups (0-1.5%).

Adverse Event Prevalence, %

Placebo n=69

Metformin n=69

Ipragliflozin n=273





Urinary Tract Infection




Genital Tract Infection




Questions and Answers

Q: You said the effects on body weight were modest. I think they are quite impressive given the short duration of the study.

Dr. Fonseca: I was being modest.

Q: Do you have any idea about long-term treatment and body weight loss? What about the physiology of the weight loss?

Dr. Fonseca: This is a relatively short-term study so I can’t answer based on these data, which I’d like to stick with for now. This cannot be dehydration – one cannot stay dehydrated long term. You see maintenance of weight loss and continued weight loss. It appears to occur in a large percentage of people; this is not just a case of responders vs. non-responders. The precise mechanism is difficult to tease out in this sort of trial.

Q: What are the main pluses of the drug?

Dr. Fonseca: Efficacy in terms of A1c lowering; potential for combination therapy (though it was used as monotherapy in this study); sodium loss; water loss; body-weight effects that appear to be long-standing, possibly related to calorie loss in the urine.


Oral Presentations: New News on Combination Therapies of Type 2 Diabetes

Efficacy and Safety of Different Dosing Regimens of the Glucokinase Activator AZD1656 During 4-Months in T2DM Patients on Metformin

John Wilding, DM (University Hospital Aintree, Liverpool, United Kingdom)

Dr. Wilding presented the results of a phase 2b dose-ranging study of AstraZeneca’s now discontinued glucokinase activator AZD1656. In the study, patients on background metformin (n=458) were treated with fixed daily doses of 20 or 40 mg AZD1656, titrated doses of 10-140 or 20-200 mg AZD1656, glipizide, or placebo for four months. A1c declines (baseline 8.3%) were significantly improved with titrated 10-140 (0.80%) and 20-200 mg (0.80%) AZD1656 versus placebo (~0.2%), though similar to that observed with glipizide (0.85%). Fixed doses of 20 (0.16%) and 40 mg (0.22%) AZD1656 showed similar declines to placebo. Patients treated with titrated 10-140 (12%) and 20-200 mg (13%) AZD1656 showed relatively increased hypoglycemia versus placebo (1.1%), with a roughly 20% increase in triglyceride levels though no effect on HDL, LDL, total cholesterol, body weight, or blood pressure. While these rates of hypoglycemia appear less than that observed with some of the other discontinued glucokinase activators, management has suggested AZD1656 was discontinued due to inability to meet “predefined internal criteria for the product.” Forest/TransTech, which continue to develop a glucokinase activator, have stressed that their program is liver-selective, minimizing the risk for hypoglycemia; however, unless the side effect profile allows for relatively innocuous administration or the drugs shows additional metabolic benefit, we remain uncertain on the future of the class relative to other options.

  • This phase 2b dose-ranging study investigated the efficacy and safety of AZD1656. As a reminder, the activation of glucokinase has been demonstrated to induce glucose-stimulated insulin release from beta cells, promote hepatic glycogen formation, and potentially regulate the secretion of glucagon and incretin hormones. In the study, patients on background metformin (n=458) were treated with fixed daily doses of 20 or 40 mg AZD1656, titrated doses of 10-140 (mean dose achieved 102 mg) or 20-200 mg AZD1656 (mean dose achieved 144 mg), glipizide (titrated doses of 5-20 mg), or placebo for four months; after initial treatment (~80% completion), patients could continue in an elective two-month continuation phase. Mean age across the groups was 57 years with a diabetes duration of 5-7 years.
  • A1c declines at four-month follow-up (baseline 8.3%) were significantly improved with titrated 10-140 (0.80% decline/48.3% of patients meeting <7.0% goal) and 20- 200 mg (0.80%/47.3%) AZD1656 versus placebo (~0.2%/18.6%). A1c decline with glipizide was similar (0.85%/54.3%) to the titrated doses of AZD1656, though fixed doses of 20 (0.16%/27.5%) and 40 mg (0.22%/28.6%) AZD1656 showed similar declines to placebo. In patients choosing to enter the two-month continuation phase (30-40% entered), A1c levels began to slowly climb, with titrated doses showing similar declines to glipizide and fixed doses showing similar declines to placebo at six-month follow-up.
  • Patients treated with titrated 10-140 (12%) and 20-200 mg (13%) AZD1656 showed relatively increased hypoglycemia versus placebo (1.1%), with a roughly 20% increase in triglyceride levels at four-month follow-up. Percentages of patients experiencing hypoglycemia with fixed 20 and 40 mg AZD1656 were 3% and 2%, respectively, and 18% with glipizide. Interestingly, Dr. Wilding suggested the increase in triglycerides trended toward baseline during the two-month extension, potentially correlating with the decline in efficacy. He indicated no relevant effects on LDL, HDL, total cholesterol, body weight, or blood pressure. While these rates of hypoglycemia appear less than that observed with some of the other discontinued glucokinase activators (a large graveyard including Merck’s MK-0599, Roche’s R1440 and R1511, Eli Lilly’s LY2599506, and AstraZeneca’s AZD6714, AZD5658, and AZD1656), management has suggested AZD1656 was discontinued due to inability to meet “predefined internal criteria for the product.”

Questions and Answers

Dr. John Buse (University of North Carolina, Chapel Hill, NC): The hypoglycemia conclusion was that rates were less than that with glipizide. Is that based on the combination of all the doses studied?

Dr. Wilding: The rates were approximately 18% with glipizide but around 12% with the high doses. They were lower, but no statistics of course were done on those adverse events.

Dr. Buse: What was the rationale for the fixed doses?

Dr. Wilding: This was really just to establish what the minimally effective dose was, and the sample size statistically was adequate to give that answer. The minimally effective dose may have been somewhere between 40 and the higher doses achieved.

Q: Were there any studies done at the liver level to see if the drug had any effect on carbohydrate intake or glucose production?

Dr. Wilding: Is it mainly working on insulins secretion on the liver; I think you’re asking theoretically if it acts on both. There is data suggesting some increases in c-peptide with this, but not as much as glipizide so there may be liver effects as well. But formal clamp studies have not been done yet.

Q: Given the rise in triglycerides, it would suggest something is going on at the level of the liver. I’m curious in terms of your conclusions that there’s a trend to correction – was that solely based on the optional two-month extension or did you see that in the original four- month study?

Dr. Wilding: That was based purely on the extension data. It’s a very interesting question to see if that’s related to the fall off in efficacy. I’m sure you’re aware of the publication with the Pfizer drug that suggested it wasn’t related, but I think more sophisticated studies are needed.


Efficacy and Safety of Dapagliflozin as an Add-On to Glimepiride in T2DM Inadequately Controlled with Glimepiride Alone Over 48 Weeks

Krzysztof Strojek, MD, PhD (Silesian Medical University, Zabrze, Poland)

Following up 24-week data initially presented at EASD 2010 (see the January 1, 2011 Closer Look), Dr. Strojek discussed 48-week results of a phase 3 trial evaluating the effects of dapagliflozin as an add-on therapy to glimepiride. In the trial, 597 type 2 diabetes patients on at least a half-maximal stable dose of sulfonylurea monotherapy were randomized to receive dapagliflozin 2.5 mg (n=154), 5 mg (n=142), 10 mg (n=151), or placebo (n=145) in addition to open-label glimepiride 4 mg/day. All doses of dapagliflozin showed greater declines in A1c (baseline ~8.1%; -0.41% with 2.5 mg, -0.56% with 5 mg, and -0.73% with 10 mg) versus placebo (-0.04%). As has been demonstrated previously, treatment also produced improvements in body weight, fasting plasma glucose, and systolic blood pressure. Perhaps related to the background therapy, dapagliflozin-treated patients showed only slightly higher rates of any hypoglycemia (9.7% with 2.5 mg, 10.3% with 5 mg, 11.3% with 10 mg) versus placebo (6.8%). As a reminder, dapagliflozin’s PDUFA date has been postponed until January 28, 2012 due to requests for further submission of phase 3 data. Presumably, concerns predominantly surround cancer risk – in the Q&A Dr. Strojek indicated no cases of bladder or breast cancer were observed in this trial.

  • This phase 3 trial aimed to evaluate the effects of dapagliflozin as an add-on therapy to glimepiride. In the trial, 597 type 2 diabetes patients on at least a half-maximal stable dose of sulfonylurea monotherapy were randomized to receive dapagliflozin 2.5 mg (n=154), 5 mg(n=142), 10 mg (n=151), or placebo (n=145) in addition to open-label glimepiride 4 mg/day for 48 weeks. Baseline A1c across all groups was ~8.1%.
  • All doses of dapagliflozin showed greater declines in A1c (-0.41% with 2.5 mg, - 0.56% with 5 mg, and -0.73% with 10 mg) versus placebo (-0.04%). As has been demonstrated previously, treatment also produced improvements in body weight (-1.36 kg [-3.00 lbs], -1.54 kg [-3.40 lbs], and -2.41 kg [-5.31 lbs] vs. -0.77 kg [-1.70 lbs] with placebo), fasting plasma glucose (-16.7 mg/dl, -16.5 mg/dl, and -28.8 mg/dl vs. +2.6 mg/dl with placebo) and systolic blood pressure (-2.53 mmHg, -3.01 mmHg, and -4.21 mmHg vs. +1.81 mmHg with placebo).
  • Dapagliflozin-treated patients showed only slightly higher rates of any hypoglycemia (9.7% with 2.5 mg, 10.3% with 5 mg, 11.3% with 10 mg) versus placebo (6.8%). A concern more or less dismissed at the FDA’s Advisory Committee meeting on the drug, dapagliflozin-treated patients showed higher rates of events suggestive of genital infections (5.2%, 6.2%, and 8.6% vs. 1.4% with placebo) – though similar rates of urinary tract infections (4.5%, 7.6%, and 7.9% vs. 7.5% with placebo). In the Q&A, Dr. Strojek suggested no cases of bladder or breast cancer were observed in the trial.

Questions and Answers

Q: Was there any dropout between 24 and 48 weeks?

Dr. Strojek: The dropout was similar in all groups. I can’t give exact figures, but the dropouts were similar to placebo.

Q: Was there any recurrence of genital infections and did you identify any patient characteristic making them more prone?

Dr. Strojek: There were higher rates in those with dapagliflozin especially at higher doses. We could not identify such a specific characteristic.

Q: Given the fall in body weight and systolic blood pressure, any data on pulse rate changes or body composition, specifically volume?

Dr. Strojek: We analyzed the data and there was no change in hematocrit as a measurement of volume. We did not analyze the body composition in this study though. No increase in pulse rate was observed.

Dr. John Buse (University of North Carolina, Chapel Hill, NC): What were the exclusion criteria around renal function?

Dr. Strojek: A GFR below 60 was exclusion.

Dr. Buse: You didn’t mention cancer in this study.

Dr. Strojek: It was a point of interest, but there were no cases in this study.

Dr. Buse: Is circumcision a risk factor for genital infections?

Dr. Strojek: We did not analyze this condition.


Symposium: Late-Breaking Clinical Trials

The A1Chieve Study: Evaluation of Insulin Analogue therapy in 66,000 People from Four Continents

Philip Home, BM BCh (Newcastle University, Newcastle, UK)

Novo Nordisk’s A1CHIEVE is the largest observational study of insulin therapy with about 66,000 subjects. Dr. Home presented data from the final time point, 24 weeks, which was published today in Diabetes Research and Clinical Practice. Patients were started on biphasic insulin aspart 30, insulin detemir, detemir plus aspart, or aspart, in a nonrandomized way – it was up to the discretion of participating physicians. Q&A was particularly interesting, as people seemed to think most of the benefit is due to the fact that previously underserved patients are obtaining better diabetes care, possibly independent of insulin analogs or even of insulin at all.

  • The goal of A1CHIEVE was to show that insulin analogs can safely improve A1c to the extent that they have done in randomized controlled trials. Participating doctors decided which patients to put on NovoMix 30, Levemir, Novolog, or Levemir plus Novolog. It was not meant to be interventional but rather an observational study, though the validity of this designation was questioned in Q&A. There were 44,661 patients across Asia, Africa, South America, and Europe that ended up included in the final analysis.
  • A1c fell on average 2.1% at 24 weeks from a baseline of 9.5%, more than would be expected from trials, with no increased adverse effects. If anything, Dr. Home said, you would expect observational studies to show less efficacy than RCTs. This represents a population that is poorly controlled and probably many are in resource-poor areas. As expected, drug-naïve patients had the largest drop in A1c – a 2.8% decline from a baseline of 10%, and those already on human insulin had the lowest drop – a 1.8% decline from a baseline of 9.4%. There was no difference between insulin types, with all groups falling about two A1c points.
  • Major and minor hypoglycemic events declined over the course of the study and weight gain was not clinically significant. There were too few major events to be meaningful. The reduction in hypoglycemic events was most prominent in the arm of patients who were switched from human insulin to an analog. Weight change was essentially neutral: 0.1 kg gain in 24 weeks, which is not clinically significant even if it is statistically. This was completely attributable to the 1.4 kg weight gain in people who started as drug-naïve.
  • Dr. Home’s conclusion was that in routine clinical practice, insulin analogs can deliver the gains seen in clinical trials. He also emphasized that lack of weight gain, blood pressure improvement, and no increase in hypoglycemic events suggest that starting insulin analogs “can be an opportunity for enhanced lifestyle behaviors…and improved self- management.” Others in Q&A felt that simply being enrolled in a study provided this opportunity, independent of insulin analogs.

Questions and Answers

Q: Have you thought about compliance? Part of the results could be explained by compliance being poor before and just an improvement in compliance caused better control. Also, what about regression to the mean when you start at such a high A1c?

Dr. Home: In terms of regression to the mean, these people weren’t selected as having a high blood glucose. Except in the sense that those are the people who tend to transfer to insulin. This does represent an opportunity to enhance self-management and self-management behaviors, so if you want to call that compliance, then yes compliance is improved with these therapies – if that is happening, fantastic.

Q: Were patients randomized to the four therapies?

Dr. Home: No they were put on insulin before they were entered. This was not an interventional study. It was observational.

Q: So, if you were going to do a post marketing survey, wouldn’t you rather look at fewer patients but over more years, such as looking at long-term outcomes etc., rather than things we’ve already seen before?

Dr. Home: I think if I were going to spend $20 million on a study, I wouldn’t do a five-year study, no. Putting together a randomized controlled trial is about 10 times the cost. You have to consider what you want to get. We wanted to get information from parts of the world where RCTs traditionally have not been done.

Q: You call this an “observational study.” Is it really though? Usually in observational studies you look at people who are already on certain therapies. In this study, patients were started on insulin at baseline. Doctors were recruited to actually start patients on insulin. The Hawthorne effect says that if you have a doctor being recruited first and then deciding to start patients on therapy, he’s going to choose a different treatment than he would have otherwise and he will be induced to try to improve everything. And moral hazard – obviously doctors are reimbursed. In resource poor countries, you are paying $125 per patient enrolled let’s say, then no doubt treatment is going to change. You are hanging a carrot in front of doctors to get them to change insulin for their patients. And anyway, there is no evidence from RCTs that analogs provide better glucose control than non-analogs.

Dr. Home: There is always a risk in these kinds of studies that people will be recruited who would have otherwise not made the change. We don’t think there’s much of a risk of that here. The high baseline A1c suggests that people are being brought in appropriately. We do not pay doctors for each patient recruited; they are required to enter just a couple patients per month – we do pay for completion of CRF and the fees vary from country to country to make sure they are not inappropriately high for that country. Look, if you want to do a study, you have to do it in some way.

Dr. Steven Kahn (University of Washington, Seattle, WA): I liked the talk but I think your conclusions are a bit weighted to the analogs and short-change oral agents or other types of management. I think a lot is changing and I think it is just a function of giving them better care. I don’t think we should ascribe this to analogs. Anyone given better care would have a reduction in A1c.

Dr. Home: If you listened to what I said, that was our conclusion. That is exactly what I said.


Glargine and Cancer Risk

Peter Boyle, PhD (International Prevention Research Institute, Lyon, France)

The day before this talk, a small study was presented (Olsson 2011) at the 34th Annual San Antonio Breast Cancer Symposium showing that the use of glargine was highly associated with risk of breast cancer (OR: 2.88, 95% CI: 1.15 - 6.64). In this presentation, Dr. Boyle, an international expert in prevention research, presented what seemed like a very thorough meta-analysis of glargine and cancer risk. He concluded that there is no association any way you analyze it. Even taking the worst case scenario, i.e., the worst odds ratios from every study, or introducing a fake, very large trial where the odds ratio was very bad for glargine – the relative risk is less than one. In other words, if anything, glargine is protective, although not significantly as the relative risk overlaps 1.0. If looking at breast cancer in particular, the relative risk is over 1.0 but still overlaps 1.0 “comfortably” as to not be significant. He added the Olsson paper to the meta-analysis at the last minute and found that it barely changed the numbers, causing him to say: “All the fuss we saw in the press yesterday – if you absorb it into all the data, it adds nothing.” He also said it was “extremely unfortunate” that this issue has been built up into such a worry and implied it has wasted much research time and money. His final report with even more data will be an abstract at ADA 2012.

  • There is no increased risk of cancer with the use of glargine. Dr. Boyle conducted a large meta-analysis using all available data, including observational studies and randomized controlled trials on cancer risk and glargine. Taking all data together ended up including 1,049,927 person- years. The relative risk of all forms of cancer with glargine was 0.87 (CI: 0.78-0.96), significant for a protective effect of glargine (“To be conservative, I will not say glargine is protective, just that there is no indication that it increases cancer risk.”) Even if you take the “worst case scenario” and use the worst odds ratios from every study, the relative risk is still 0.92, this time not significant but still trending toward protective. Including a “phantom study” with very large odds ratios (e.g., up to 10.0, showing hugely increased risk of cancer with glargine), you still can’t get the meta-analysis to move toward significant harm. Finally, all this is true even with the likely impact of publication bias, i.e., that positive results are more likely to be published than negative.
  • Breast cancer risk with glargine was a non-significant 1.08 (CI: 0.92-1.27). When looked at individually, breast and prostate cancer both had relative risks greater than 1.0, but still “comfortably overlapping 1.0.” A study was presented yesterday (Olsson 2011) showing hugely increased risk of breast cancer with glargine (OR: 2.88, 95% CI: 1.15-6.64). However, after incorporating this study into the meta-analysis at the last minute, it changes the relative risk to(0.94-1.30) – still not significant. This led Dr. Boyle to say: “All the fuss we saw in the press yesterday – if you absorb it into all the data, it adds nothing.”
  • “It is extremely unfortunate that this has grown into such a big topic,” Dr. Boyle said in Q&A. Dr. Boyle pointed out that the databases from which cancer risk with glargine use has been taken have been abused - “they have not been designed to be put to the uses everyone is putting them to these days.” They’re designed to look at economic data more than epidemiologic.

Questions and Answers

Q: This is important data. How do we handle the fact that diabetes itself increases the risk of some cancers?

Dr. Boyle: There is always possible bias and that may have an impact. It is extremely unfortunate that this has grown into such a big topic. It’s a business. My background is cancer epidemiology, and I think people have been torturing these epidemiologic databases a little too much. How can you call it a proper study if you don’t know why a patient got a particular treatment, other outcomes, etc.? But no matter how much you torture this data, there is no way you’re going to move this into a positive finding. Though in general if you torture data with so many patient-years and so many different kinds of cancers with so many different statistical tests, I’m sure you can find something.



Insulin Therapy in Type 2 Diabetes: Who, When, and How

Irl Hirsch, MD (University of Washington, Seattle, WA)

In one of the most crowded talks we have ever attended, Dr. Hirsch shared clinical wisdom about insulin therapy in type 2 diabetes. He noted that early initiation of insulin therapy seems to provide substantial glycemic gains (and possibly long-term health benefits, as being studied in the ORIGIN trial), but insulin use in type 2 diabetes remains limited worldwide – mainly due to cost and availability issues. He discussed the rationale for using analog insulin vs. human insulin (less hypoglycemia risk, which is especially a concern for patients who will be near A1c of 7.0% after treatment) and for using premix insulin vs. basal-bolus therapy (premix insulins give less flexibility, which is especially problematic for patients who have had diabetes for several years). Dr. Hirsch also discussed how to surmount several barriers to insulin initiation (e.g., using a clinical “team” to address patient concerns, prescribing metformin and/or a GLP-1 receptor agonist to mitigate insulin-associated weight gain). He closed by noting that although much remains unknown and insulin therapy remains imperfect, access and understanding have come a long way in the past two decades.

  • Research suggests that early, aggressive initiation of insulin therapy can be beneficial to type 2 diabetes patients’ “beta-cell health” – an issue explored by the recently completed ORIGIN study. In recently diagnosed type 2 diabetes patients, insulin therapy has been associated with one-year remission rates of roughly 40-50% – roughly double the rate seen with oral antihyperglycemic agents (Li et al., Diabetes Care 2004). The longer-term benefits of early insulin initiation are not yet known, but results from the seven-year, 12,000- plus-patient ORIGIN study will be presented at ADA 2012. As a reminder, the study compares standard step therapy to an aggressive regimen of insulin glargine metformin. Dr. Hirsch said that although his clinic was one of the study centers, he has no idea of the results and can only hope that basal insulin therapy will be linked to wide-ranging benefits on cardiovascular health and other endpoints.
  • Barriers to access and/or the availability of other effective diabetes therapies will likely continue to delay these patients’ insulin initiation; overcoming these barriers represents a huge opportunity for industry. Today, few type 2 patients worldwide initiate insulin early in the course of disease, either because they have many other treatment options that they exhaust first (a common issue in the developed world) or because they do not have access to affordable insulin (a common issue in the developing world, and the major barrier to insulin use overall). Dr. Hirsch noted that the use of incretin therapies so far seems to have had minimal impact on when patients initiate insulin, though he said that the impact of incretins on insulin decisions is increasing of late – especially in the US.
  • On the fundamentals of basal insulin initiation, Dr. Hirsh reviewed a series of relevant studies and offered three main “teaching points.”
    • Most patients with A1c in the mid-8% range can drop below 7% using basal insulin therapy alone, as consistently seen in treat-to-target studies of glargine such as Treat-to-Target, INSIGHT, and APOLLO.
    • The final A1c achieved with basal insulin is dependent on baseline A1c. In other words, patients with lower baseline A1c are more likely to drop below a particular target value, even though the A1c change is greater in patients with higher baseline A1c. However, baseline A1c does not seem to determine patients’ ultimate hypoglycemia risk,even though one would expect that patients with lower baseline A1c would experience more hypoglycemia after starting on insulin. Dr. Hirsch said he had been surprised by this result, which we speculate may reflect generally higher levels of glycemic variability among patients with higher baseline A1c.
    • Above 1.0 U/kg, increasing doses of glargine prolong therapeutic duration but do not statistically significantly increase the magnitude of effect. For highly insulin-resistant patients who require such high doses, one possibility is to switch to U-500 insulin. However, Dr. Hirsch said that before he even considers U-500, he recommends that patients split each basal dose into two shots of NPH insulin. (Using NPH instead of basal insulin makes the regimen more affordable, and the patients’ extreme insulin resistance means that hypoglycemia risk is low.) In a small (n=31) study of highly insulin-resistant patients presented at EASD 2011, those who split each NPH dose into two injections improved their glycemic control at six months and 12 months (under 9.0% in both periods compared to 10.0% at baseline; p<0.05), while patients randomized to single-site NPH injections saw no improvement (A1c > 9.8% throughout trial).
  • Dr. Hirsch reviewed results from an insulin initiation study in type 2 diabetes patients (n=572) in which basal-bolus therapy was similarly effective whether the regimen involved one or three mealtime shots. (For more data on this 60-week study by Rosenstock et al., see pg 161 of our ADA full report at Each basal-bolus regimen led to bigger glycemic reductions than twice-daily doses of premix aspart 70/30, though the premix regimen was also quite effective (2.2% and 2.3% for basal-bolus vs. 1.8% for premix). During the question and answer period, Dr. Hirsch said that premix insulin could be useful early in diagnosis when patients still have a sizable reserve of beta cells, but he believes that basal-bolus therapy is clearly better for patients who have had type 2 diabetes for several years.
  • To mitigate insulin-associated weight gain, Dr. Hirsch recommended concurrent use of metformin and/or, where available, a GLP-1 receptor agonist. Dr. Hirsch said that aside from hypoglycemia, weight gain is the biggest source of concern about initiating insulin therapy. In general, patients who gain more weight after initiating insulin tend to be those with higher baseline A1c, greater use of prandial insulin, and higher overall insulin doses. Dr. Hirsch said that metformin is safe and well studied but that its weight-maintenance effect is fairly modest, whereas GlP-1 receptor agonists seem to have more dramatic body-weight effects that allow people to “really do something positive.” Indeed, exenatide and insulin combination therapy has been associated with weight loss ranging from 2.7 kg to 6 kg (5.9 lbs to 13.2 lbs). (He noted that the FDA recently approved concurrent use of exenatide and glargine, though he added that GLP-1/basal insulin combination therapy had already been widely used.) Dr. Hirsch also noted that access to GLP-1 receptor agonists is relatively limited, given the drugs’ higher cost and narrower availability.
  • Dr. Hirsch concluded by cataloguing several real-life barriers to insulin initiation and expressing his optimism about the trends in insulin therapy worldwide. Worldwide, insulin initiation may be hindered by each patient’s individual or cultural concerns (which need to be understood by physicians), lack of clinical team (e.g., if a lone doctor tries to start a patient on insulin without help from a nurse or educator), lack of adequate glucose testing material (e.g., if strips are too expensive or hard to supply), fear of hypoglycemia (particularly difficult to overcome for many patients), and – with regard to initiating prandial insulin – a lack of primary care resources to guide patients through the process. Despite these ongoing challenges,Dr. Hirsch said that many parts of the world now do a better job with insulin therapy than in the past, and he noted that the quality of best-practices research has improved dramatically relative to 25-to-30 years ago.

Questions and Answers

Q: You gave a very strong recommendation for analog insulins. But evidence very clearly states that between human and analog insulin, analogs’ benefits are not so great to justify wide use.

Dr. Hirsch: If I wasn’t clear I will say again: there is no evidence of better A1c in type 2 diabetes with analogs vs. human insulin. The question of whether to use analogs has to do with someone’s glucose control – not where you start, but where you end up. If at the end of insulin treatment, you have A1c of 7.0%, your hypoglycemia risk will be higher than when your A1c was 8.5%. I know that there are many, many populations where after insulin therapy, people’s A1cs are in the eights and nines – still better than 12-13%, but some distance from goal. If their post-treatment A1c is that high, I don’t disagree that regular insulins can be quite useful. But as you get to lower A1c, I do feel relatively strongly that you do better with insulin analogs. For someone with an A1c quite high who is very insulin-resistant, it doesn’t matter.

Q: I saw that you came out very strongly in favor of basal followed by one-to-three prandial injections, rather than premix insulin. In my experience in India with lots of carbohydrate in the diet, premix works very well – starting with once daily and then adding multiple injections as needed.

Dr. Hirsch: I’ve never been a big fan of premix insulin, possibly in part because of most of what I do clinically is with type 1 diabetes. Still, I have used it quite a bit over the years. The biggest benefit is in people new to insulin with a good beta-cell reserve. But once someone has been on insulin five or ten years or more, premix insulins don’t work as well because the patients need more flexibility than premix insulins can give. I should note that in many parts of the world they have all sorts of premix combinations; in the US we are very limited in the variety of premix options.

Comment: I agree that eventually basal-bolus is good. But in the initial five-to-seven years, why not use premix?

Q: Do you give insulin in prediabetes, either for impaired fasting glucose or impaired glucose tolerance?

Dr. Hirsch: I’m not currently giving insulin for IGT and IFG. I can’t predict how ORIGIN will end up even though we were one of the sites (Editor’s note: the ORIGIN study population included both type 2 diabetes patients as well as people with prediabetes). My hope is for less CV events and less progression from prediabetes to diabetes, but I have no idea.

Q: There is a lot of literature on the relationship of cancer to diabetes medications such as liraglutide, pioglitazone, and others. Can you comment?

Dr. Hirsch: There are clear signals out there, either positive or negative, with every drug. As a diabetologist, I have been thinking of this too simplistically at times, but the reality is complex. For example, in men with diabetes we see less prostate cancer but more colon cancer. I think the data are strongest for metformin, which seems to reduce cancer risk. I think that the data on other drugs is still at too early a stage to interpret confidently. I continue to use insulin quite a bit; I’m not concerned about cancer risk with glargine, detemir, or any other insulin.

Q: in the hospital setting, what do you do for the glycemic management of patients with enteral feeding, particularly continuous feeding?

Dr. Hirsch: There are several ways to do this, and none is right or wrong. A lot of people use NPH and regular human insulin or NPH and a rapid-acting analog. In our hospital, once patients are on the enteral dose, we put them on intravenous insulin to ballpark their needs, then we start them on NPH thrice daily with a rapid-acting analog. I don’t like using glargine with these patients because its long duration creates all kinds of problems if the tube-feeding is discontinued.

Q: You said that one way to counter insulin-associated weight gain is to use metformin. If dealing with type 2 who is emaciated, when should we add metformin?

Dr. Hirsch: if you have an emaciated patient with low BMI, I’m not sure I would use metformin at all. If they get ketosis while taking metformin, that’s not a good combination.

Q: Are there circumstances when you would not pursue strict glycemic control? Could you comment on insulin failure?

Dr. Hirsch: Some people won’t reach their target. Sometimes, you don’t want to aim for a low target – for example, with someone who has hypoglycemia unawareness or who has had diabetes for 40-to-50 years and is hypoglycemic all the time. For some other patients, you try again and again, but can’t reach target. We often blame ourselves in these cases, but often it’s the patients’ decision not to be invested in better control.

Q: If a patient comes in on a GLP-1 receptor agonist and multiple oral drugs, how would you shift the regimen to prepare for insulin initiation?

Dr. Hirsch: As a rule of thumb I leave sulfonylureas in until I start basal insulin. When they just come in, the first thing I stop is the DPP-4 inhibitor. They are all on it where I’m from, and they are often at A1c of 9% or 10% – too big a distance from goal for a DPP-4 inhibitor to make up. My goal is to get patients to as few drugs as they need, for variety of reasons. If they are on a TZD, I probably would take that away, too… though in Washington state, TZDs haven’t been widely used for years. GLP-1 is a tough one. I generally keep it if they are on it; the reason is the weight issue. It’s a hard one for me to rationalize keeping it if the A1c is that high, though. Once I get them on a pretty good insulin dose, I give them a trial period without GLP-1 and then we decide whether to restart it.

Q: Could you comment on the choice between regular human insulin and rapid-acting analogs?

Dr. Hirsch: I generally like to use rapid-acting analogs; regular insulin hangs around in middle of night and can lead to hypoglycemia. What I said before is true – this risk isn’t as big for someone who will get to an A1c of 9.0% compared to an A1c of 7.0%.

Q: According to the ADA, A1c should be less than 7.0%. The IDF and AACE say less than 6.5%. What do you recommend for Pakistan?

Dr. Hirsch: If you read the fine print of current guidelines, the word that comes up again and again is individualization – based on patients’ health, demographics, psychosocial factors, and so on.


Corporate Symposium: Steps Towards Future Insulin Therapy – How Far Have We Come? (Sponsored by Novo Nordisk)

The Idea - Multi-Hexamer Soluble Depots

Peter Kurtzhals, PhD (Senior Vice President, Diabetes Research, Novo Nordisk, Bagsvaerd, Denmark)

Dr. Kurtzhals reviewed the history of research into ultra-long-acting insulins to mimic the flat basal insulin profile seen in people without diabetes. He noted that although insulin glargine and insulin detemir both offer longer and more consistent action profiles than insulin NPH, neither has a profile that is truly flat or consistently 24 hours long – a goal pursued by the Novo Nordisk researchers who developed insulin degludec. Degludec’s mechanism of action involves the formation of multihexamer depots in the subcutaneous space, thanks to careful selection of the amino acid chain (the same as that of human insulin, but with the deletion of the B-30 residue), the attached fatty acid moiety (chosen to be a 16-carbon chain that includes two carboxylic acid groups, to facilitate multihexamer formation), and the linker group that connects the two. In solution with phenol and zinc, degludec exists as a dihexamer. Phenol diffuses immediately after injection to lead to the formation of the multihexameric depot, and zinc diffuses gradually thereafter to facilitate the breakup of the depot into active insulin monomers.


The Action - Pharmacological Profiles

Thomas Pieber, MD (Medical University of Graz, Graz, Austria)

Dr. Pieber reviewed pharmacokinetic and pharmacodynamic data on insulin degludec, which suggest a longer and more consistent action profile than that of insulin glargine. Indeed, degludec has been shown to have lower day-to-day variability (coefficient of variation 20% vs. 82% in a study of people with type 1 diabetes), lower within-day variability (with roughly 50% of the day’s insulin action consistently occurring in the first 12 hours after injection) and a longer half-life (25.4 hours vs. 12.5 hours in a study of people with type 2 diabetes). Dr. Pieber explained that if a patient achieves lower glycemic variability, he or she can theoretically target a lower mean blood glucose without as much risk of hypoglycemia.


The Start - Clinical Data on Basal Only Therapy

Chantal Mathieu, MD, PhD (Catholic University of Leuven, Leuven, Belgium)

Emphasizing her status as a patient-focused clinician, Dr. Mathieu presented phase 3 data on a flexible degludec dosing regimen (designed to accommodate patients’ varying day-to-day schedules and to reduce both patient and provider barriers to initiating insulin) as compared to a fixed dosing regimen of insulin degludec or a fixed dosing regimen of insulin glargine. In addition to being a poster at IDF, the study was presented as a poster at ADA 2011 (see page 163 of our ADA Full Report) and as an oral at EASD 2011. As a reminder, the flexible degludec schedule required rotating doses by 40 hours and eight hours (i.e., alternating morning and evening dosage). All patients had type 2 diabetes (n=687) and were targeted to the same pre-breakfast fasting goal of 4.0-4.9 mmol/l (72.0-88.2 mg/dl), such that A1c was reduced by equivalent amounts in both groups (though fasting plasma glucose was significantly lower with the flexible degludec arm than the glargine arm [105 vs. 112 mg/dl; p=0.04]). The treatment arms were similar with regard to weight gain and hypoglycemia (though flexible degludec dosing produced non-significant reductions in nocturnal hypoglycemia). During Q&A, Dr. Mathieu explained that although degludec doses can be spread out without compromising safety or efficacy, hypoglycemia rates and A1c do worsen when doses are doubled and given every-other-day – the problems seem to arise in large part due to the higher volumes of insulin injected at the same time.


The Next Step - Clinical Data on Basal-Bolus Therapy

Miles Fisher, MD (University of Glasgow, Glasgow, United Kingdom)

Dr. Fisher presented data from late-stage treat-to-target studies comparing insulin degludec and insulin glargine with regard to glycemic efficacy and hypoglycemia risk. He particularly focused on one phase 3 study in type 1 diabetes (n=629, randomized 3:1 degludec vs. glargine) and another in type 2 diabetes (n=992, randomized 3:1 degludec vs. glargine). In both studies, the relative risk of nocturnal hypoglycemia was statistically significantly lower by 25% with degludec, and in the type 2 diabetes study the overall rate of hypoglycemia was statistically significantly lower by 18% as well. Finally, Dr. Fisher reviewed a meta-analysis of phase 3 data indicating that degludec is associated with statistically significantly lower risk of nocturnal hypoglycemia as well as statistically significantly lower risk of overall hypoglycemia in the “maintenance period” starting at 16 weeks into each trial.

  • Dr. Fisher presented results from a prespecified meta-analysis of seven phase 3 studies comparing once-daily regimens of insulin degludec and insulin glargine, either as part of basal-bolus or basal-only therapy, in patients with type 1 or type 2 diabetes. Overall, degludec was associated with a statistically significant 9% reduction in overall confirmed hypoglycemia (including measured blood glucose < 3.1 mmol/l [<54.6 mg/dl] and/or hypoglycemia requiring assistance) and a statistically significant 26% reduction in nocturnal hypoglycemia (confirmed hypoglycemia occurring between 11:00 pm and 5:59 am). In a prespecified analysis of the “maintenance period” (starting at 16 weeks), degludec was linked with 32% lower nocturnal hypoglycemia risk among all patients, a 49% lower nocturnal hypoglycemia risk among type 2 patients on basal insulin only, a 16% lower risk of all confirmed hypoglycemia among all patients, and a 28% lower hypoglycemia among type 2 patients on basal insulin only.


Panel Discussion

Moderator: Bernard Zinman, MD (Mount Sinai Hospital, University of Toronto, Toronto, Canada)

Peter Kurtzhals, PhD (Novo Nordisk, Bagsvaerd, Denmark), Thomas Pieber, MD (Medical University of Graz, Graz, Austria), Chantal Mathieu, MD, PhD (Catholic University of Leuven, Leuven, Belgium), and Miles Fisher, MD (University of Glasgow, Glasgow, UK)

Questions and Answers

Q: What is the fate of the phenol? Why do phenol and zinc not leave at the same time?

Dr. Kurtzhals: Phenol is eliminated unmetabolized in the kidneys with a half-life of a few hours. Why don’t they bind with the same affinity? Zinc is bound to the central cavity of insulin molecule tightly, but phenol binding is looser and further to the outside of the insulin molecule.

Q: Is there a safety concern with regard to phenol?

Dr. Kurtzhals: No. Phenol is probably the most widely used preservation agent in any injectable. It is in all insulin preparations, and it has been used for 70 years.

Dr. Zinman: We don’t hear about phenol in other formulations because it doesn’t play role in PK/PD, but it is perfectly common.

Q: Could you address whether we should be concerned about hypoglycemia due to the buildup of insulin degludec?

Dr. Pieber: Per the definition of steady state, insulin is not stacking up, and you are not at increased risk of hypoglycemia. In fact, there is a reduced risk of hypoglycemia.

Dr. Zinman: Dose adjustments are much further away. You don’t adjust on a daily basis, you do it once a week based on the previous three glucose measurements. When you adjust appropriately and at appropriate intervals, you get much less hypoglycemia.

Q: Many of us have patients who seem to have the dawn phenomenon. Would it be better to use NPH or detemir, rather than degludec, with these patients?

Dr. Mathieu: Surely not NPH. When you counterbalance the huge variability seen with NPH, I don’t think it’s even in the same ballpark as degludec. On the contrary, I think stable profile at night will be the reason degludec is the next-step basal insulin.

Dr. Fisher: Some would say that the dawn phenomenon simply reflects insulin at night running out.

Dr. Zinman: I think that’s a good point. The studies that Dr. Pieber showed were elegant with regard to the shortcomings of current insulins and their less-than-24-hour action profiles.

Q: In the UK, there is a big drive use more regular human insulin and NPH and less analogs.

Dr. Fisher: I think there is indeed a drive and that it is misguided. If you are trying to intensify insulin therapy, you should use the best therapy possible. If you have an older patient with comorbidities and you are just trying to manage symptoms, maybe NPH is fine. But for other patients we should fight against attempts to make us use second-rate treatments.

Dr. Mathieu: You can’t help but see the variability of NPH action and the glycemic variability it induces at night. Elderly people are also entitled not to have hypoglycemia at night. Even in our bankrupt Belgium and Eurozone, we have to have good insulins to treat our patients.

Q: Dr. Kurtzhals, if the injection of degludec goes intravascular, the insulin is actually a dihexamer. What happens?

Dr. Kurtzhals: It would be similar to what you’d find with accidental injection of human insulin. No formation of multihexamers in plasma.

Q: Is there any downregulation of insulin receptors more so than with other insulins?

Dr. Zinman: Insulin multihexamers are in the subcutaneous space, not the circulation. There is no especially high hyperinsulinemia.

Q: You showed lots of comparisons to glargine, what about to detemir?

Dr. Pieber: we know that detemir covers 24 hours in half-to-2/3 of patients, similar to insulin glargine. A direct head-to-head comparison of detemir’s and degludec’s action profiles has not been conducted to my knowledge.

Dr. Kurtzhals: We did study detemir and degludec with regard to regular clinical endpoints in phase 3. The study showed similar superiority of degludec to detemir on fasting plasma glucose and hypoglycemia that we’ve seen comparing degludec to glargine.

Q: What about mitogenicity?

Dr. Kurtzhals: The mitogenic potential of degludec is low, even lower than with glargine or other insulins. There is no concern in that regard.

Q: What about weight?

Dr. Mathieu: No difference in weight.

Q: So it doesn’t have detemir’s purported advantage of less weight gain?

Dr. Mathieu: We didn’t show data on weight, but the protocol for the studies involved weight- maintenance training, and the weight gain with all insulins was minimal. No difference was observed in the arms.

Dr. Fisher: There was a very slight weight gain that was similar between arms.

Q: Is there a possibility to use degludec every other day?

Dr. Zinman: A small phase 2 proof-of-concept study was conducted in the Lancet. The advantage of lowering hypoglycemia with daily degludec was lost in every-other-day injection, but A1c reductions were similar. But in a larger phase 3 study reported as a press release, these advantages were not sustained. Might this be due to large doses of insulin? Dr. Mathieu, you were involved in these studies.

Dr. Mathieu: The difference between the flexible study I presented on and the phase 2 study was that in the Monday-Wednesday-Friday study, double doses of insulin were injected. The big conclusion for the moment is that you can stretch the dosing as far as 40 and eight hours, but giving double doses of insulin is tough.

Dr. Zinman: How many of you get phone calls from patients in the morning saying that they forgot to take insulin the night before? I usually tell them to take half their dose right away, but it’s tricky to work this out. The elegant study Dr. Mathieu presented shows that you can take insulin in the AM and be perfectly safe. There are studies where people would like to delay insulin for several hours. We don’t encourage people to do it, but it can be done.

Q: Do we expect to see more skin reactions due to the subcutaneous depots of insulin degludec?

Dr. Kurtzhals: Site reactions were low and didn’t differ from comparator insulins.

Q: On the detachment of zinc from injected degludec – does this vary from patient to patient? Is it affected by age or ethnicity?

Dr. Kurtzhals: The binding of zinc to insulin is a physicochemical property that is not dependent on the individual.

Q: Will clinical studies compare insulin degludec to NPH?

Dr. Zinman: My understanding is that the focus has been on comparing degludec to glargine and detemir, as these are considered standard of care for basal insulin therapy. I doubt there will be comparisons to NPH.

Dr. Kurtzhals: Correct, that is not currently the plan.

Q: Are there any differential effects of degludec on hepatic vs. peripheral action, compared to current analogs?

Dr. Kurtzhals: This is something we would like to explore in later clinical pharmacological studies but we haven’t studied yet.

Q: Where can I buy it?

Dr. Zinman: It’s not approved anywhere to my knowledge. Dr. Kurtzhals: That’s correct.

Dr. Zinman: Hopefully Canada will be one of the first.

Q: It’s going to be very cheap, right?

Dr. Kurtzhals: (Smiles)

Dr. Zinman: New things generally aren’t cheap.


4. Diabetes Technology

Oral Presentations: Insulin Therapy – From Monitoring to Control

Results of a Three-Year Experience of Continuous Subcutaneous Insulin Infusion in Type 1 Diabetes Patients

Alexandros Kamaratos, MD, PhD (Tzanio General Hospital, Diabetes Centers, Athens, Greece)

Dr. Kamaratos presented follow-up data on a group of type 1 pumpers (n=66) three years after initiation of pump therapy. Baseline A1c was 9.4%, which declined to 7.28% by the end of year two and rebounded to 8.08% in year three. Dr. Kamaratos could not confidently explain the deterioration in glycemic control between years two and three, but he suggested that an increase in hypoglycemia, an increased proportion of insulin devoted to bolus infusion, or relaxation of patients’ self-observation might have been involved. Cholesterol and triglycerides declined throughout the study, which we speculate was related to the improved glycemic control. Overall, we found the study quite interesting, although we believe its observational nature and inconsistent results prevent any definitive conclusions from being drawn.

  • Sixty-six type 1 patients (10-17 years of age) were started on insulin pumps and followed over a three-year period. All subjects had been on intensive insulin therapy with MDI prior to starting on the pump. However, patients were not achieving satisfactory control on MDI, as mean baseline A1c was 9.4%. Data were collected at years one, two, and three on BMI, blood pressure, A1c, lipid profile, 24-hour urine albumin excretion, aminotransferases, frequency of hypoglycemia, and insulin dosing.
  • A1c declined by 1.3% over the course of the study, although a deterioration in glycemic control was observed between years two and three (p<0.001 for A1c in year three vs. baseline).


Year 1

Year 2

Year 3






  • The mean frequency of hypoglycemic episodes declined between baseline and year two, with an unexplained increase in year three The time measure of this statistic (e.g., weekly, monthly) was not specified. Dr. Kamaratos noted that this year-three finding resembles a previously reported study appearing in Diabetes Care (Bode et al.) (p<0.001 for year one vs. baseline; p=0.04 for year two vs. year three).


Year 1

Year 2

Year 3

Mean number of hypoglycemic episodes





  • Total cholesterol, HDL cholesterol, and triglycerides all declined between baseline and year three. Dr. Kamaratos noted that the drop in HDL in the third year was particularly unexpected (p=0.01 for total cholesterol in year three vs. year one; p<0.001 for HDL in year two vs. year three; p=0.004 for triglycerides in year three vs. year two).


Year 1

Year 2

Year 3

Total Cholesterol (mg/dl)





HDL (mg/dl)





Triglycerides (mg/dl)





  • The mean total daily dose declined by almost 20 units between baseline and year three (p=0.001). As the table illustrates, the drop in dosing occurred between baseline and year one, with that level then sustained up to year three.


Year 1

Year 2

Year 3

Total Daily Dose (units)





  • Bolus insulin infusion declined between baseline and year two, with a notable increase in year three.


Year 1

Year 2

Year 3

% of Daily Dose as Bolus Insulin Infusion





Questions and Answers

Q: On the changes in cholesterol levels – these were young type 1 kids. Can you be sure they did not get statins?

Dr. Kamaratos: No, very few were on statin therapy. We were looking for all the reasons that lower HDL levels and trying to figure it out. We are calling back the patients to follow-up, but we don’t have any explanations yet.

Q: How did you collect hypoglycemia data?

Dr. Kamaratos: It was self-reported by patients.

Q: Were you following these patients to reprogram the pump, increase basal rates, reduce boluses, etc.?

Dr. Kamaratos: Yes, we did. The problem with patients is many postponed their appointments and did not come in. We are trying to get them back and retrain them.

Dr. Roman Hovorka (University of Cambridge, Cambridge, UK): Do you think the reduction in control from year two to three could be explained by progression to less compliance?

Dr. Kamaratos: Yes, I believe the reason is less compliance.

2 Years Metabolic Evolution in Type 2 Diabetes Patients Treated by Subcutaneous Ambulatory Insulin Pump

Jean-Pierre Courrèges, MD (General Hospital, Narbonne, France)

We were extremely disappointed that Dr. Courrèges did not show up to the session, as we were really looking forward to this two-year data on insulin pumps in type 2 diabetes. According to the abstract, 189 patients (mean age: 58 years; mean A1c 9.1%; mean diabetes duration 17.5 years) were originally switched from MDI therapy to pump therapy. At the two-year mark, follow-up data was collected on 54 patients. A1c declined by 1.26% after two years (p <0.0001), with 44% of patients recording an A1c of <7.5%. Weight gain was a miniscule 0.32 kg (0.7 lbs; p=0.85), average insulin dose declined by 0.24 units per kilogram per day (p=0.03), and rate of severe hypoglycemia declined from 0.36 events per patient per year to 0.03 events per patient per year. This is the first time we’ve seen data on pumps in type 2 over such a long period of time, and the results seem strong. Although the A1c drop was not so impressive in the context of such a high baseline A1c, these patients had very long-standing diabetes and were failing MDI – moving the needle by this magnitude is notable in our view. The abstract does not give rates of therapy usage, but we’d be interested to see what proportion of patients were using other drug classes like GLP-1 receptor agonists, DPP-4 inhibitors, and TZDs.


Patient- or Physician-Driven Continuous Glucose Monitoring in Poorly-Controlled T1D Patients: A One-Year Randomised Multicenter Study

Regis Radermecker, MD (University of Liège, Liège, Belgium)

Dr. Radermecker presented the results of a one-year study comparing physician-driven and patient- driven CGM use to self-monitoring of blood glucose. We’ve previously seen this same study presented at both ATTD 2011 (page 21 at and ADA 2011 (page 122 at As a reminder, the study randomized 178 patients with a mean A1c of 9% to one of three groups for one year: standard care (self-monitoring of blood glucose), physician-driven use of CGM (sensor use prescribed and titrated by the physician), or patient-driven use of CGM. At one year, A1c was significantly reduced compared to the SMBG control group in both the patient-driven arm (-0.52%, p=0.0006) and the physician-driven arm (-0.47 %, p=0.0008). The CGM groups were not statistically significantly different from each other in terms of A1c improvement. Interestingly, individuals in the physician-driven group used 34% fewer sensors for the same improvement in A1c – Dr. Radermecker very loosely hinted that this might support the greater cost-effectiveness of the physician-driven approach.

Questions and Answers

Dr. Roman Hovorka (University of Cambridge, Cambridge, UK): What was the retention of the sensor use? Did you observe a decrease over time? We’ve seen in other studies that over longer periods of time, patients tend to use less sensors.

Dr. Radermecker: That’s hard to answer because it’s an incremental use in our physician prescribed group. In the patient group, we saw this trend of decreasing use over time. Additionally, patients with CGM decreased the number of self-monitoring of blood glucose tests, and it was significant.

Dr. Hovorka: You said 60 patients didn’t satisfy the entrance criteria. What was the reason?

Dr. Radermecker: It was a 10-day test period to see if patients were able to use the CGM device. There were significant differences when we compared the 60 dropouts to the rest of the patients participating in study. They were younger, they had experienced more ketoacidosis in their diabetic life, and they had a shorter duration of type 1 diabetes.

Q: Your A1c improved by about 0.5%. This means that A1c was still above target – it would have left them at about 8%. Can you explain the lack of achieving target control with so much intervention?

Dr. Radermecker: When you compare with other CGM trials of six months or 12 months, it’s the same reduction. The message is that it’s an interesting tool, but for a specific patient with a specific education. The question of whether it’s a cost-effective approach is very important. In a specific population, 0.5% could justify its use as an add-on tool.


Self-Monitoring of Blood Glucose (SMBG), Medication Adherence and Change in A1C in Non-Insulin Using Type 2 Diabetes Pateints

Naunihal Virdi, MD, MBA (LifeScan, Milpitas, CA)

Dr. Virdi presented an analysis of claims data associating self-monitoring of blood glucose and medication adherence with change in A1c. This observational study of non-insulin-dependent type 2 patients was previously presented at ADA 2011 (see page 131 of our full report at Overall, the study found that presence of SMBG in type 2s not on insulin was associated with greater medication adherence, and SMBG was also associated with larger improvements in A1c in both medication-adherent and non-adherent patients.

Questions and Answers

Dr. Roman Hovorka (University of Cambridge, UK): This had the limitation of an observational study. Based on this study, would you recommend a randomized controlled trial?

Dr. Virdi: The decision to use self-monitoring of blood glucose is between the patient and physician. I think further studies would be needed to address this question.

Dr. Jayendra Shah (University of Arizona College of Medicine, Tucson, AZ): Nice study – it’s similar to one published a few years ago. That was also an observational study, and the interesting part was why these patients dropped their A1c. We found that our patients are smarter than we think. They made diet modifications and reduced the number of carbohydrates they were eating. Everything else remained the same – insulin, weight, BMI, etc.


Symposium: The Artificial Pancreas

Insulin Delivery Devices and Continuous Glucose Monitoring

Eric Renard, MD, PhD (University of Montpellier, France)

Dr. Renard gave a historical presentation on the artificial pancreas, starting with the Biostator of the 1970s and culminating in UVA’s mobile-phone-based system (for more information, see our DTM 2011 report at He focused his presentation around the problem of delay, noting the time lag inherent in insulin absorption, insulin action, and glucose sensing. To get around this problem, early work on the artificial pancreas focused on IV insulin infusion-IV glucose sensing systems. While these projects had solid results, Dr. Renard emphasized their usability problems: blood clotting and lack of portability. Research transitioned to implantable pumps and subcutaneous (SC) sensing, which again demonstrated efficacy but have limited applicability due to high cost and poor distribution worldwide. Dr. Renard closed with a review of the recent progress on SC insulin-SC sensing systems, emphasizing the importance of model predictive control for countering the time lag associated with these routes. He believes that current subcutaneous systems are feasible for the artificial pancreas, although Dr. Renard asserted that intraperitoneal and implantable pump delivery is not dead – a sentiment that he reiterated during an interesting Q&A.

Questions and Answers

Dr. Aaron Kowalski (JDRF, New York, NY): I’m a big believer in subcutaneous- subcutaneous approaches. When you think about the speed of IP delivery, is it faster absorption of insulin or suppression of glucagon?

Dr. Renard: Faster absorption. Also, when using the IP route, there is less variability of infusion. This is an issue that comes up with subcutaneous delivery for unknown reasons. You have variability in insulin action in the same patient from one day to another. The age of the catheter also plays a role – the absorption is not the same by the third day, which may influence the data. I think the IP and implantable pump route is not dead. It’s hard to distribute these around world and they are expensive. But alternatives like Roche’s port system could be interesting.

Dr. Roman Hovorka (University of Cambridge, UK): You are the IP authority. What are the roadblocks of taking the IP route forward?

Dr. Renard: It is an alternative route. We are going forward with the subcutaneous route with great progress. We will soon start a study from Roche with an implantable port, which is easier to manage than an implantable pump. It could take care of the problem of delay.

Dr. Hovorka: What about implantable pumps? Are they too costly? Is it the insulin?

Dr. Renard: We just finished the study comparing the old IP to a new one from Sanofi. The results are even better with the new one. It is sold on a medical and scientific point of view. Medtronic is now starting new production of an implantable pump. We could have this option, but the price will remain quite high. It will be a sophisticated device. But at least it will be considered in a few years as a possible closed-loop system.

Q: How well does your algorithm perform in response to exercise? What is a smart way to take exercise into consideration?

Dr. Eric Renard: There are few investigations on this point. Exercise is a poorly reproducible model. It depends on the intensity and duration and there are no good models used for the closed loop. There is some data on exercise challenges, but it remains very standardized exercise. It will be difficult to close the loop with full safety without feeding the system some information.

Q: What about adding an accelerometer?

Dr. Renard: Good suggestion. I see this as a supplementary addition to the algorithm model.


Closing the Loop - Technological Advances

Boris P. Kovatchev, PhD (University of Virginia, Charlottesville, VA)

Dr. Kovatchev reviewed the most important milestones in artificial pancreas research, discussed the recent advances in technology, and provided his outlook on the future. His presentation resembled one he gave at the recent Diabetes Technology Meeting (see pages 8-9 in our report at He began by noting the major landmarks in AP research in the last few years: the FDA-accepted in silico simulator, the UCSB/Sansum Artificial Pancreas System, and the design of a modular architecture framework. He also reviewed some results from the jointly developed modular control-to-range system and gave a useful review and pictures of the Diabetes Assistant (DiAs), UVA’s very recently developed portable artificial pancreas system running on a Sony Ericsson cell phone (for a comprehensive overview of the system, including images, see pages 3-5 of our DTM 2011 report at We appreciated Dr. Kovatchev’s clearer explanation of the system’s traffic light system, although no update was given on the outpatient trials aside from what we heard at DTM. We hope to hear more on DiAs at ATTD, as well as see the most updated data from JDRF’s multicenter trial of control-to-range. Looking to the future, Dr. Kovatchev forecasted home studies and clinical acceptance of the artificial pancreas, control algorithms built into pumps, and closed-loop systems capable of control-to-target (what he deemed ‘replacement’ for basal-bolus therapy as opposed to ‘adjunct’ use with control-to range).

  • Dr. Kovatchev showed a meta-analysis of the previous JDRF inpatient studies performed at UVA, Padova, and Montpellier from 2008-2010. In 60 adults, the control- to-range system dropped average glucose from 154 mg/dl to 146 mg/dl, increased time in target (70-180 mg/dl) from 72% to 83%, improved time in tight control (80-140 mg/dl) from 44% to 58%, and reduced hypoglycemia from an average of 1.15 events per person to 0.54 events per person over the 24-hour duration of the study.
  • DiAs’ home screen features hypoglycemia and hyperglycemia “traffic lights” to help users quickly understand the system’s state (various screenshots are available at: As Dr. Patrick Keith Hynes mentioned at DTM, the idea of the lights is that “a user can pull the phone out of their pocket and if the lights are green, things are okay.” Dr. Kovatchev explained that the hypoglycemia and hyperglycemia traffic lights appear green if there is no risk of either excursion, yellow if the system detects risk and mitigates it automatically, and red if the system cannot mitigate the excursion. In this case, it would alarm the user that action is needed.

Questions and Answers

Dr. John Wilding (University of Liverpool, UK): From the glucose sensing technology aspect, how robust and reliable are these systems? Will patients be able to use them day in and day out, month after month?

Dr. Kovatchev: Insulin pumps have been around for a long time. CGM is a newcomer – it’s only been ten years. Contemporary CGMs are good enough for control-to-range; they are not good enough yet for control-to-target. They are sufficiently reliable over several days. We use the Dexcom Seven Plus, which must be changed once every seven days. The accuracy allows for glycemic control that is not perfect yet, so it’s a notch below control-to-target. It’s the middle two blocks of the modular system I described and we can do it now. Further improvements in CGM accuracy and reliability will be needed for control to target. For now, control-to-range has padding built in and it can absorb more errors.

Dr. Aaron Kowalski (JDRF, New York, NY): In the UAE and back in the United States, many can afford pumps and sensors. By utilizing a phone platform, I would suspect that it will bring AP technologies to many more people, particularly with less expensive patch insulin pumps. And you your system will have the capability to talk to multiple devices, right?

Dr. Kovatchev: Yes, the basic idea is to make devices cheaper and more accessible. The diabetes assistant is not solely dedicated to run closed-loop control. It can also assist with other diabetes-related things. It can be used as a bolus calculator and pump interface when not in closed-loop mode. If connected to a sensor alone, it can provide sensor-based information. It also features the remote connection to a central location. Such a remote monitoring system would parents to monitor their children’s CGM values. When connected to a sensor and pump, it has the function of closed-loop control. It’s meant to be accessible. It’s only connected to the Insulet OmniPod and Dexcom Seven Plus, but will be expanded in the near future. (Editors note – we wonder if Medtronic will be added to the system even though the company has a proprietary mobile-phone based artificial pancreas system. This was also showcased at DTM – see pages 5-6 of our report at


Closing the Loop - Models, Algorithms, and Clinical Results

Claudio Cobelli, PhD (University of Padova, Padova, Italy)

Dr. Cobelli discussed the complex process of modeling glucose control and designing closed-loop algorithms. He explained that glucose models are useful both from a regulatory perspective (preclinical research moves faster and costs less with in silico models than animal experiments) and for building real-time control software, such as the model predictive control (MPC) algorithms used by Dr. Cobelli and his colleagues. MPC algorithms, which recommend insulin dosage based on continuously updated forecasts about glucose levels, have been validated in several inpatient (and recently outpatient) studies at different centers. Ongoing efforts and next steps include a randomized inpatient study at the University of Padova, the JDRF Multicenter Trial, the European AP@Home Project, and NIH-funded research on modeling meals and physical activity.

  • Expressing his optimism for the upcoming JDRF Multicenter trial and the ongoing European AP@Home Project, Dr. Cobelli also highlighted two projects through the NIH Artificial Pancreas Initiative. First he described a triple tracer study, so-called because it uses three different types of radiolabeled glucose to assess postprandial metabolism. The study’s protocol involves three days, each with one mixed meal test (breakfast, lunch, or dinner). The non-diabetic arm of the study has already finished (n=20), and 17 patients with type 1 diabetes have also participated in the experiment (toward an enrollment goal of n=20). Dr. Cobelli also looked forward to a new project designed to quantify the glycemic effects of physical activity (as measured by a sensing system developed by Dr. James Levine and colleagues). He implied that the goal of this study is to improve the algorithms of existing artificial pancreas systems, rather than to build a new system that incorporates continuous monitoring of physicalactivity (as we understand it, real-time monitors of non-glucose factors seem likely to introduce too much noise to be useful at the current stage of technology).

Questions and Answers

Dr. Aaron Kowalski (Senior Vice President, Treatment Therapies, JDRF): Could you talk about the mathematics of algorithm getting smarter from an individual patient. Might this be something we’ll see in the future?

Dr. Cobelli: I think this is already happening. This parameter Q is individually personalized based on person’s insulin delivery rate, body weight, and carb ratio. The other way is to have not just a single parameter but an individual model that is much richer. This is much more difficult because you have to have informative experiments on that particular patient. Research is ongoing in this area, and at some point I’m sure we’ll come up with something.

Dr. John Wilding (University of Liverpool, Liverpool, United Kingdom): As a clinician treating patients, I would fear whether the system can distinguish between a failing sensor and real values.

Dr. Cobelli: The bottom layer of the architecture is the safety module, which predicts what will happen to glucose in the future. If glucose starts to fall, you can aggressively decrease insulin (full suspension can be dangerous because it can causes oscillation). Based on the models, you intelligently turn down the pump. Thanks to this safety layer, the number of hypoglycemia decreases dramatically, as you saw.


Closing the Loop - Clinical Results

Roman Hovorka, PhD (University of Cambridge, Cambridge, United Kingdom)

Dr. Hovorka reviewed his group’s clinical experience with closed-loop control, covering much of the same ground as in his presentation at ADA 2011 (see pg 55 at Closed-loop control has been shown superior to open-loop control in overnight studies, though challenges have arisen in more challenging conditions (e.g., hypoglycemia due to exercise, overdosing at meals, or the delayed glucose disposal seen late in the pregnancy of women with type 1 diabetes). His group continues to research food and exercise effects, with the goal of building algorithms that require the user to “announce” as little information as possible. Unfortunately, Dr. Hovorka’s three-week outpatient trials are now expected to start next year, pushed back from the September 2011 target he mentioned at ADA. His group’s research (using Abbott’s FreeStyle Navigator CGM and Aviator pump) has historically been hindered by supply interruptions, and we hope that this situation improves in the near future.

Questions and Answers

Dr. John Wilding (University of Liverpool, Liverpool, United Kingdom): You talked about the adaptivity of the algorithms. You showed that hypoglycemia was in issue in pregnancy. Is a different algorithm used in this situation? If not, should a different one be used?

Dr. Hovorka: Initially it was the same algorithm. We learned that glucose disposal is slower and insulin sensitivity is lower late in pregnancy and were able to adjust the algorithm for this. We can make adaptations for specific groups based on knowing certain characteristics.

Dr. Wilding: In the future, might algorithms be able to adapt to individual patients?

Dr. Hovorka: Yes, possibly. Given the short duration of existing studies, the system could adapt only to the next two-to-three hours. In the future adaptation could occur at a longer time-scale, though.

Dr. Claudio Cobelli (University of Padova, Padova, Italy): With regard to meal absorption studies, we have been solving this problem by having to prepare a special meal. Otherwise the glucose is not consistently mixed throughout the food. I was interested in understanding how you solved the complex meal tracing of glucose, though this may be too complicated to discuss in this setting.

Q: It seems that during exercise, knowing the rate of glucose disposal (Rd) is essential to predict how the glucose profile will respond to a bolus. Does the Mayo database have Rd data following certain exercise activities or intensities? If not, other databases might.

Dr. Hovorka: These algorithms could use information from these databases. I think it’s the way forward to reduce post-exercise hypoglycemia. I think it depends on what type of closed-loop system we want to create. Our aim is to avoid announcing exercise to make the system as simple as possible, but we may have to announce exercise or even type of exercise.

Dr. Ragnar Hanas (Uddevalla Hospital, Uddevalla, Sweden): Would it be possible to analyze meals by analyzing chewing patterns? [Editor’s Note – This was similar to a question Dr. Hanas asked of Dr. Edward Damiano at ADA 2011.]

Dr. Hovorka: I have looked into chewing sensors, and none is currently available. There are some challenges in monitoring meals, and it might be nice to measure this better – perhaps a GLP-1 monitor. A chewing monitor is a possible way forward. I tried to find something and couldn’t.


Symposium: Diabetes and Technology

Benefits of Insulin Pump Therapy in Everyday Life: Now and In the Near Future

Eric Renard, MD, PhD, FRCP (University of Montpellier, Montpelier, France)

Dr. Renard reviewed several studies and meta-analyses on insulin pump therapy, with an eye toward answering: what are the reasons that roughly 600,000 people use pumps today – a number expected to rise to one million by 2025 (Novo Nordisk 2010) – and what benefits do these people receive? Dr. Renard said that patients typically start insulin pump therapy because they seek better glycemic control or greater flexibility, and they can expect clinical benefits that include lower A1c, better quality of life, and a reduction in severe hypoglycemia (which can be relevant for patients whether their baseline A1c is high or low, though Dr. Renard said that the hypoglycemia benefits seem greatest in patients with longer duration of diabetes and impaired glucagon responses). He said that the biggest benefits are traditionally seen in patients who test their blood glucose most frequently, but he added that the burden of fingersticks is being reduced by products that integrate pumps and continuous glucose monitoring (CGM). The long-term goal is to develop closed-loop systems in which the pump delivers insulin automatically based on CGM readings. He concluded by describing a new mobile-phone-based closed- loop system that began being used in outpatient studies in France and Italy this October (for more on this joint initiative between the University of Virginia, University of Montpellier, and University of Padova, see our Diabetes Technology Meeting Days #2-3 coverage at

Questions and Answers

Q: I am one of the people who has the benefit of insulin pump use. Before I was not in good control and now I am. I hear about patients asking their doctors for insulin pumps and being told, “Not yet, because you don’t need it.” Why would I have to wait for bad control? How can we change this idea?

Dr. Renard: When you have evidence that control is better with pump whatever the initial situation, there is no reason not to start with a pump if you feel more comfortable with it. In many European countries, if you are a childe and have type 1 diabetes, you can start with a pump. It’s just an option.

Q: You commented that people with longer duration of diabetes have more benefits with pumps because their glucagon response is impaired. Could you comment the on the evidence base for this issue?

Dr. Renard: If you have longer duration of t1dm, you are more prone to hypo. Some Italian studies have studied glucagon response relative to diabetes duration. It was lower in long history of diabetes, and hypoglycemia unawareness was also higher in these patients. Even if you have hypo unawareness for years, studies suggest you can restore awareness. It doesn’t reflect the death of alpha cells, but rather a condition that can be addressed with restoration of glucose control.


CGMs: Identifying Problems and Adjusting Treatment Programs

Anne Peters, MD, CDE (Keck School of Medicine at the University of Southern California, Los Angeles, CA)

Dr. Peters shared tips on continuous glucose monitoring (CGM) based on treating some of the richest and poorest patients in Los Angeles at her several clinics there. “I love CGM…I can’t imagine why any patient wouldn’t be using it. It just makes sense to me,” Dr. Peters said. “But,” she added, “it doesn’t make sense to all my patients.” She discussed a three-month, four-visit plan designed to ease patients’ transition to CGM, and she also specifically addressed challenges such as discordance with fingersticks (in Dr. Peters’ experience the biggest patient frustration, which she attributed largely to time lag), bolus stacking (bolus-on-board information is useful for patients who have pumps), poor literacy and numeracy (recommending treatment changes based on absolute numbers rather than percentages can be helpful), and skin irritation (treatment with steroid cream up to 12 hours ahead can be helpful, as can IV3000 or Tegaderm bandages with a hole for the sensor to go through).

  • Dr. Peters explained that she tries to hold four clinic visits during the first three months of each patient’s CGM use, and she outlined the goals and emphases at each visit. She stressed the importance of continual encouragement and management of expectations. “As you can see,” smiled the USC-based physician, “I’m a bit of a cheerleader.”
    • Visit One is about training on the system and setting expectations, especially about the need for calibration and the phenomenon of lag time. Dr. Peters initially sets the alarms widely and sometimes does not turn on the high alarm at all (so that neither the patient nor their spouse get annoyed at day-and-night buzzing).
    • During Visit Two, held two-to-four weeks later, Dr. Peters focuses on high- level adaptation. Her foremost priority is adjusting overnight basal rates, given the importance of avoiding nocturnal hypoglycemia. She also reiterates education on lag time and how it can be especially apparent during times of rapid glucose change.
    • Visit Three occurs six-to-eight weeks after the start of CGM and includes a retrospective data analysis to help optimize settings. For example, she mightexplain how to override bolus wizard recommendations based on the trend displayed by the CGM. Dr. Peters emphasized that many patients have trouble with math, so it can be important to describe changes in terms of insulin units rather than a percentage of the doses.
    • When patients reach three months of use and come to their fourth clinic visit, Dr. Peters discusses their first A1c value that reflects CGM-related changes. Although she typically sees A1c reductions over time, these don’t always show up on the first post-CGM A1c value – especially if CGM has helped the patient reduce nocturnal hypoglycemia. Dr. Peters said it is important to explain this to patients with unchanged (or higher) A1c so that they do not get discouraged.

Questions and Answers

Q: What do you recommend for small kids using the sensor – for instance, those less than six years old?

Dr. Peters: I am not a pediatrician; my youngest patient on CGM is 10 years old. In general, especially with adolescents, I try to let pediatric patients make their own decisions rather than letting their parents drive treatment choices.

Q: How often do your patients change sensor sites?

Dr. Peters: Typically patients wear the Dexcom sensor for about 14 days and the Medtronic sensor for about six. This is longer than indicated, but the sensors are expensive, and most patients don’t have 100% coverage. To prolong sensor life, patients should blow-dry the adhesive tape after they shower. For young adults that swim and surf, the sensors are toast after a couple days.

Q: If you don’t have accurate insulin:carb ratios to begin with, how do you use patient results to adjust to more appropriate values?

Dr. Peters: With the MiniMed pump, I tell patients that they can correct their settings if they don’t like them. I try to get their data on what the ratio is programmed to do and to see what they actually do. I use logs – I get these data along with logs on what they are eating and how the calculated the carbs for those meals. The effects of these things on control aren’t random, and I need to know what they are doing when they’re not doing well. My dietitian is wonderful and very helpful in these efforts.

Q: What are your thoughts on Medtronic’s upcoming Enlite sensor?

Dr. Peters: I can’t wait to use the Enlite. I take care of many MiniMed staff studying this, so I get to see them. Anything that makes the accuracy better makes it easier for patients. I think it will reduce stacking and other problems with CGM.


Advances on Mobile Phone Technologies for Diabetes Management

Robert Istepanian, PhD (Kingston University, London, United Kingdom)

Dr. Istepanian gave an overview of mobile health for diabetes management, highlighting evidence of glycemic benefits. He mentioned several major telecommunications companies (e.g., AT&T and Qualcomm) and diabetes-focused companies in the field (e.g., Cellnovo, PositiveID, and Entra – whose Bluetooth-integrated MyGlucoHealth blood glucose meter received FDA clearance back in 2009), though we had hoped his talk would be more comprehensive in this regard (e.g., he did not mention

WellDoc or Telcare). He is advocating the IDF to establish a mobile diabetes task force that can connect various stakeholders and make a roadmap for future technologies. Given the huge potential benefits worldwide (especially in developing countries, where the relative lack of current infrastructure makes it easier to introduce new standards), we hope IDF follows his advice and strengthens the growing field of mobile diabetes care.

Questions and Answers

Q: What is the cost of such a mobile health service?

Dr. Istepanian: It differs by country. There is no universal system, and the challenge is to build one – this is relatively simpler in poorer countries.

Q: What is the cost-effectiveness compared to regular care?

Dr. Istepanian: I am not an economist, but I know it can be cost-effective, and all the major companies are involved because they think they can make money in this area.

Q: Have you made your recommendation to the IDF?

Dr. Istepanian: I clearly addressed this issue at the IDF forum on December 4, and hopefully they will act on these recommendations. The UN’s M-Health Alliance has begun, but I think it is important for IDF to have a special task force on mobile health in diabetes.


Corporate Symposium: Innovations for Diabetes Management: Looking Toward the Future (Sponsored by Bayer Diabetes Care)

Data Management and Patient Empowerment: What's the Evidence?

Peter Schwarz, MD (University of Dresden, Dresden, Germany)

Dr. Schwarz discussed the benefits of Bayer’s Contour USB, a blood glucose meter that can plug into USB ports, and announced the design of a recently initiated clinical trial comparing the Contour USB (with data management software) to the Contour (with a handwritten logbook). The nine-month CONGO study is designed to assess the importance of data management for behavior change and ultimately glycemic control in patients with type 2 diabetes using insulin (n=240). CONGO was initiated after glycemic benefits were seen in over 1,000 patients who participated in the Contour USB’s usability study (mean blood glucose down by 11.2 mg/dl and postprandial blood glucose down by 19.2 mg/dl after four-to-six weeks of Contour USB use). For his part, Dr. Schwarz hypothesized that self- monitoring of blood glucose in conjunction with effective data management “has the ability to give unique additional reductions for patients with type 2 diabetes.” Overall we think that this is a clever and important study topic. If clear benefits are seen from simplified SMBG data management, CONGO could further raise the market’s demand for elegant and user-friendly technologies – a goal of both mobile health startups and, increasingly, major device manufacturers.

  • Dr. Schwarz discussed a prospective randomized controlled trial, called Evaluation of the Impact of the Contour USB Blood Glucose Monitoring System with Integrated Data Management on Glycemic Control in Insulin-treated Diabetic Patients (CONGO), which Bayer recently initiated to study the effects of diabetes data management on behavior change and glycemic control. CONGO will enroll 240 adults with insulin-treated type 2 diabetes across six clinical centers. Patients will be randomized toreceive either the Contour USB (with Bayer’s Glucofacts Deluxe software used for data management) or the older Contour meter (with a handwritten logbook required for data management); full training will be provided for whichever meter the patient uses. The primary endpoint is A1c at nine months (the end of the study); secondary endpoints include frequency of severe hypoglycemia, changes in patient empowerment and therapy adherence, satisfaction with their assigned data management regimen, increased number of preprandial and postprandial fingersticks, glucose variation, change in insulin dose per injection (which the researchers hope will go down if glycemic variability decreases), and change in microalbuminuria. Dr. Schwartz anticipated that CONGO will have data in 11-13 months, though we were not clear whether this referred to the study end date or Bayer’s timeline for announcing the data.


Innovations in SMBG

Tim Bailey, MD (University of California, San Diego, San Diego, CA)

Dr. Bailey reviewed clinical performance data on Bayer’s new Contour XT meter, which has performed with extreme accuracy in several early studies. (See our coverage of Days #2 and #3 of the Diabetes Technology Meeting in the November 1, 2011 Closer Look at As a reminder, the first- generation Contour XT has been launched in Italy and Iberia. Notably, Dr. Bailey said that the second- generation Contour XT will include data management software directly on the meter. We are curious to hear more about this in-meter software and how it compares to the built-in bolus calculators of Abbott’s FreeStyle InsuLinx and Roche’s Accu-Chek Aviva Expert; so far we do not know the approval status or launch timeline of the second-generation Contour XT.


5. Obesity and Bariatric Surgery

Oral Presentations: Marking, Bypassing, and Treating Obesity

Weight Loss and Diabetic Status in Obese (BMI>35) Diabetic Adults After 56 Weeks on Phentermine/Topiramate CR (PHEN/TPM CR)

Nancy Bohannon, MD, FACP, FACE (St. Luke’s Hospital, San Francisco, CA)

Dr. Bohannon reported on results in patients with type 2 diabetes and BMI > 35 kg/m2 who were randomized to placebo or low- or standard-dose phentermine/topiramate controlled release (PHEN/TPM CR, aka Qnexa) as part of Vivus’ 56-week CONQUER study. Patients in the full-dose treatment arm (n=65) experienced least-squares mean weight loss of 12.1%, significantly better than the 2.8% least-squares mean weight loss in the placebo group (n=58). Benefits were also seen in A1c, fasting glucose, and rates of diabetes. We understand that this analysis was performed in light of the IDF’s controversial position statement that encourages the use of bariatric surgery in patients with diabetes and BMI greater than 35 kg/m2 (for more on this position statement, see our coverage of its announcement during the first day of last year’s WCIDT at In a sort of response to this position statement, Dr. Bohannon noted the similarities in excess weight loss seen with full-dose PHEN/TPM vs. two separate, observational studies of laparoscopic adjustable gastric banding surgery in obese people with diabetes. She noted that the comparison is only directional, due to the differences in study designs and patient populations. Nonetheless, we think it is highly encouraging to see new therapies starting to fill the gap between current medications and surgery – especially therapies that are effective in this high-risk group of obese patients.

  • As a reminder, CONQUER was a 56-week study comparing placebo, phentermine/topiramate controlled release 7.5 mg / 46 mg (PHEN/TPM CR 7.5/46), and PHEN/TPM CR 15/92 in adults with BMI between 27 and 45 kg/m2 (inclusive) and at least two obesity-related comorbidities. Among other comorbidities, the enrollment criteria included type 2 diabetes treated with diet and exercise or metformin. Participants were randomized to placebo, low-dose PHEN/TPM CR, and full-dose PHEN/TPM CR in a 2:1:2 ratio. For more details on CONQUER study design as well as the study’s results, see page 16 of our AADE 2010 full report at
  • Dr. Bohannon presented a subanalysis of patients in CONQUER with type 2 diabetes and BMI greater than 35 kg/m2. All three study arms were generally well matched, with mean age slightly over 50 years, mean BMI roughly 40 kg/m2, and mean A1c roughly 6.7%, and mean weight roughly 108 kg (~238 lbs) in the placebo and low-dose groups and roughly 115 kg (~253 lbs) in the high-dose group – the only statistically significant difference in baseline characteristics. Mean diabetes duration was roughly four years, and mean previous metformin use ranged from 40% in the placebo group to 54% in the treatment groups.
  • Both doses of PHEN/TPM CR led to statistically significantly greater weight loss than placebo, and full-dose PHEN/TPM CR also caused statistically significant benefits in mean A1c and fasting glucose. Full-dose treatment also caused significantly greater excess weight loss (weight loss divided by the difference between a patient’s current weight and normal weight). As would be expected given these results, weight loss was highly significantly correlated with improvements in fasting glucose (r=0.475; p<0.0001), fasting insulin (r=0.253; p=0.0024), and A1c (r=0.504, p<0.0001). Also, study group was significantly correlated with resolution of diabetes, as defined by maintenance of fasting blood glucose below5.5 mmol/l (99 mg/dl) and discontinuation of diabetes medications (χ2=7.1043, p=0.0287). Amongthe patients whose diabetes resolved, mean (SD) 56-week A1c was 5.5% (0.3%).




CR 7.5/46



CR 15/92


Least Squares Mean % Weight Loss 2.8 6.6* 12.1**
LS Mean % Excess Weight Loss 7.4 17.8* 32.6**
Basline A1c, mean (SD), % 6.7 (1.0) 6.8 (1.0) 6.6 (0.9)
LS Mean A1c change at 56 wks, % 0.05 -0.19 -0.50^
BL Fasing Glucose, mean (SD), mg/dl 126.4 (29) 135.5 (36) 124.7 (29)
LS Mean FG change at 56 wks, mg/dl 0.72 3.24 -11.7*
Resolution of Diabetes, % 1.7 8.3 15.4

* p<0.05 vs. placebo, ^p=0.0002 vs. placebo, **p<0.0001 vs. placebo. SD=standard deviation

  • The nature and prevalence of adverse events were similar to those seen in previous PHEN/TPM studies. Rates of constipation, upper respiratory tract infarction, paraesthesia, insomnia, dry mouth, headache, and dysgeusia seemed to increase in response to treatment.
Adverse Event (%) Placebo (n=58)

PHEN/TPM CR 7.5/46








Upper Respiratory Tract Infection












Dry mouth
















  • Ending on an exploratory and provocative note, Dr. Bohannon showed that the trial’s excess weight loss (EWL) results were similar to those from two laparoscopic gastric banding studies. Dixon and O’Brien, prospectively examining 50 LAP-BAND patients with diabetes, reported mean excess weight loss of 38% from a mean baseline BMI of 48.2 kg/m2 (Diabetes Care 2002). Segato and colleagues reviewed data from 1993 through 2005 on laparoscopic adjustable gastric banding procedures at their clinic in Italy, finding that in 52 type 2 diabetes patients all receiving diabetes medication, excess weight loss was 32.4% from a baseline BMI of 49.1 kg/m2 (Surg Obes Relat Dis 2010). Dr. Bohannon emphasized that the study populations and designs do not enable true comparison among these three studies. (E.g., as we note below, baseline diabetes medication use was higher in each of the bariatric surgery studies.) Nonetheless, it is encouraging to see that pharmacotherapy can get highly obese diabetes patients into a similar ballpark as laparoscopic banding, potentially giving patients and doctors another strong option for weight loss.

PHEN/TPM CR 15/92 (n=65)

Dixon & O’Brien 2002 (n=50)

Segato et al., 2010 (n=52)

Baseline BMI, kg/m2




% Excess Weight Loss at 12 mos




Baseline % OAD use




Baseline % Insulin use




*Represents “previous metformin use,” not baseline OAD use. OAD = oral antidiabetes drug

Questions and Answers

Dr. Fran Kaufman (Chief Medical Officer, Medtronic Diabetes): What about stopping the drug for a period. There might be weight re-gain. What are the metabolic effects in this case?

Dr. Bohannon: This is being studied. We are looking at the effects on a variety of metabolic parameters and risk factors.


Reducing Diabetes Risk by Managing Childhood Obesity: UK National Effectiveness Data Following the Mend Programme

Paul Sacher, RD, MBDA (UCL Institute of Child Health, MRC Childhood Nutrition Research Centre, London, United Kingdom)

Mr. Sacher presented on the excellent work of the MEND program (Mind, Exercise, Nutrition, Do it!), a nationwide, family-based, childhood obesity intervention being delivered at over 300 locations throughout the United Kingdom. The group’s work has been published in a number of academic journals, and the data in this presentation was recently presented at TOS 2011 (24-OR). In 10,173 children (mean age ~10 years) that went through the 24-month program between 2007 and 2010, mean BMI and waist circumference decreased by 0.8 kg/m2 (27.2 to 26.4 kg/m2) and 2.6 cm (one inch) (p<0.0001). The study also found that parent-reported levels of their children’s physical activity increased 3.6 hours per week and sedentary activities decreased 5.8 hours per week (all p <0.0001). Other improvements were noted in heart rate (decrease of 8.7 beats per minute on a step-test recovery), self-reported body image measures, and even a slight decline (1.2%) in parental BMI. While the time commitment (four hours per week for the ten-week intensive phase) is obviously not insignificant, Mr. Sacher noted that mean attendance of the MEND sessions was quite high (78.5%) and the dropout rate was only 11.4%. Notably, an independent UK cost effectiveness study found that the ICER of MEND is £1,671, well below NICE’s cost-effective threshold of £20,000-30,000; this translates to a 10-13 fold return on investment – this is certainly valuable to bring to policymakers. We’re glad to hear that the group is also rolling out the program in a number of other countries, including the US, Canada, and Australia, and will be adding other age groups as well as primary prevention.

  • The MEND Programme ( is a 10-week, family-based intervention followed by an online-based maintenance and support program until 24 months. The initial phase consists of twice-weekly two-hour group sessions that include behavior modification, nutrition education, and physical activity. Mr. Sacher emphasized that the intervention involves parents, which many childhood obesity programs neglect to do. To bring it to scale, the program is designed for delivery by community health workers and non-obesity specialists. This is essential in our view for both cost-control and wide-scale expansion.
  • Participants in the study had a mean age of 10 years and a BMI >91st percentile. Parental and caregiver attendance was required and children were either referred by healthcare providers or families self referred. The mean parental BMI at baseline was 32 kg/m2 and 80% of parents were overweight or obese. The study population was 24% minorities, 34% single parents, and 47% did not own home.
  • We were impressed to see the low dropout rate considering the program’s schedule and length – it certainly takes a fairly high level of commitment from families. However, since many study participants were self-referred into the study (actual number not given), we presume the general motivation level was quite high – we wonder if and how these results would translate to a less motivated population.

Questions and Answers

Q: Dr. Francine Kaufman (Medtronic Diabetes, Northridge, CA): This is an amazing mass of research and real world effort. Did you consider delivering this as part of a school curriculum?

Mr. Sacher: Our programs involve parents, so they are delivered after school when parents can attend with their children. Some material is suitable for a school curriculum. We have developed school resources that can be delivered by teachers. In an ideal world, we shouldn’t exist, as all of this should be taught in schools. The education of the family unit is very important.

Dr. Kaufman: You could imagine that this could be done in schools with older children. Then, they could bring the learnings home to the family.

Mr. Sacher: Yes. Data suggests that parent-only interventions work better in younger children. In these cases, the parent applies the learnings and makes changes.

Q: We performed a similar study in India by intervening in one school and comparing it to a control group. We published this in the European Journal of Clinical Nutrition. We included more parameters, including A1c and C-reactive protein. We showed improvement in almost every parameter. However, the cost of counseling such students in schools was quite high. Each student was counseled 3-4 hours per week. The time taken was also high. I think this cost effective analysis needs to be done in developing countries as well.

Mr. Sacher: Just one point to add: a systematic review of school-based interventions revealed they have not shown good results. For preventing obesity, I think it’s because they don’t involve parents. Our intervention takes a family-based approach that is quite different from many other studies.


Normalised Metabolic Risk Markers Two Years After Gastric Bypass Surgery on Obese Adolescents in Amos, A Swedish Nationwide Study

Claude Marcus, MD (Karolinska Institute, Stockholm, Sweden.)

This presentation covered the Adolescent Morbid Obesity Surgery (AMOS) study, discussing the two- year follow-up results of 81 adolescents (mean age: 16.5 years) undergoing laparoscopic gastric bypass surgery. Mean weight decreased from a baseline of 133 kg (292 lbs; BMI: 45.5 kg/m2) to 89 kg (195 lbs; BMI: 30 kg/m2) at the two-year mark, with the vast majority of this weight loss occurring during the first year following the operation. Interestingly, 43% of the subjects had a weight gain during the second year, and compared to non-gainers, the only significant difference was higher insulin levels. We were surprised to see the negative side effects of the operation, which included a staggering 67% of individuals recording low vitamin levels despite prescribed supplementation (adherence to supplementation was not specified) – this is a careful reminder to us that the effectiveness of bariatric surgery has to be weighed against the lifelong nutritional deficiencies and other side effects, especially when individuals as young as 13 years old are undergoing the operation. Five reoperations (1%) occurred during the study period. According to the speaker, the treatment was generally well tolerated and quality of life was improved in all parameters of SF-36 (i.e., physical and psychological). We’ll be interested to see long term follow-up of these and other young patients undergoing bariatric surgery to better understand the durability of weight loss and prevalence of side effects and complications.

  • In addition to significant weight loss, other cardiometabolic parameters improved as well. A1c went from 4.37% at baseline to 4.17% at two years, fasting insulin decreased from31.7 mu/l to 7.4 mu/l, and fasting glucose went from 5.1 mmol/l (91.8 mg/dl) to 4.9 (88.2 mg/dl). Blood pressure went from 124/78 to 117/71 mmHg. LDL cholesterol, triglycerides, Apo A and Apo B lipoprotein, CRP, and white blood cell count, also significantly declined, while HDL increased significantly (all changes p<0.001).

Questions and Answers

Dr. Arya Sharma (University of Alberta, Edmonton, Canada): About 80% of your subjects had significant psychosocial issues prior to surgery. You said quality of life improved. Did you get measures of psychosocial function, such as coping in school?

A: We do not have this specific data. One can say that in the group of morbidly obese children, they have psychosocial problems. It was an improvement going to school after surgery, but I can’t answer specifically.

Q: Among those who gained weigh after year one, did you consider that the kids were 13-18 years old and might be growing? Were there any correlations? And what will you do with those who gained weight? Would you perform another procedure like gastric banding?

A: I don’t know about the first question. No, the gainers will not have a second procedure done. We’ll now see what happens after five years.

Q: You chose to do gastric bypass over the banding. Why?

A: There have been some studies with banding in adolescents with good results. Those adolescents had a good psychiatric background and were selected. The profile of severely obese in Sweden is not good and we can’t select in that way. Therefore, gastric bypass was used. Additionally, there was a small study in Sweden that used banding in adolescents – the results were very poor. The study had to be stopped because of complications.


Symposium: Approaches to the Treatment of Obesity: Why and What?

Is There Still a Future for Pharmacotherapy in the Treatment of Obesity?

Luc Van Gaal, MD PhD (Antwerp University Hospital, Antwerp, Belgium)

Dr. Van Gaal argued that pharmacotherapy is going to be important to curb the obesity problem. While bariatric surgery is “one of the best options to deal with obesity,” it’s not going to be a big picture solution given the numbers of obese. Regulatory agencies unfortunately tend to shy away from approving and paying for obesity drugs, despite the fact that obesity has many features that make it an appropriate disease for pharmacological treatment: it is an “epidemic,” it has a large role in predisposing to other disease and risk factors, reversal of these abnormalities is possible, and it is multifactorial in origin and therefore difficult to treat. It has been very difficult for any weight loss drug candidate to achieve more than 10% weight loss and this rarely occurs. Two combinations that were able to achieve this were Qnexa (Qnexa was recently recommended for approval at the FDA) at about 12%, and Amylin’s pramlintide/metreleptin, which “unfortunately” has been discontinued, at about 13%. Dr. Van Gaal mentioned liraglutide during Q&A, explaining that the SCALE trial will explore the drug’s ability to prevent diabetes – Dr. Van Gaal’s inkling is that it will.

  • Drugs almost never achieve more than 10% body weight loss unless used in combination. In drug trials for anti-obesity candidates, weight loss is nearly always 7-10 kg (14-or ~7-10% body weight loss, no matter the drug, the delivery, or the mechanism. Weight loss drugs in monotherapy have rarely if ever surpassed this. Certain combinations reached it, however, and it may only be with combinations that this is possible. Future drug options have this challenge: “to break the 10% weight loss target.” Although rimonabant was physiologically elegant, “we are from reaching 10% with cannabinoid antagonists.” MTP inhibitors have also beenstudied with limited success – so far about 4% weight loss, so “certainly not the immediate future in my interpretation.” Tesofensine, a triple monamine reuptake inhibitor like sibutramine, has achieved weight loss over 10% at higher doses but with potentially higher risk.
  • Lorcaserin, Contrave (bupropion/naltrexone), and Qnexa (topiramate/phentermine) are centrally active oral drugs recently filed with the FDA, with Qnexa showing the largest weight loss. Contrave has had good results, with weight loss of about 8-9%. However, the FDA will require a cardiovascular outcomes trial before approval. Lorcaserin has had an effect of similar magnitude, and was also shown to have durability up to two years (with wash-out subjects at one year going back to placebo weight). Lorcaserin is expected to be resubmitted to the FDA. Controlled release topiramate alone is efficacious in terms of both A1c and weight loss, with the latter around 7%. Qnexa, however, a combination of phentermine and topiramate, achieved about 12% weight loss in the EQUIP trial and other studies. Other parameters like FPG and A1c also appear to be improved, which we assume bodes well. Qnexa’s new application was accepted with new proposed labeling that includes a contraindication for women of childbearing potential.
  • Orlistat limits progression from IGT to diabetes and raises GLP-1 levels. It is the only obesity drug that is still available worldwide, though it is not used much. It seems to raise GLP-1 levels similar to metformin. It is also the only obesity drug with a diabetes outcomes trial, though liraglutide now has the ongoing SCALE trial examining diabetes prevention with liraglutide in non-diabetic obese individuals.
  • Injectable peptides may be promising for obesity treatment, especially liraglutide. Many have advocated for GLP-1 agonists for moderate weight loss. Most of these drugs induce average loss of up to only 3 kg (6 lbs) or so. However, liraglutide is the only one that has been studied extensively for weight loss, and Dr. Van Gaal noted it was “interesting for the future.” In non-diabetic obese individuals, high doses of liraglutide induce about 8-9 kg (16-19 lbs) or 8-9% loss at 24 weeks. The effect is relatively sustained, with 5.8 kg (12 lbs) weight loss over two years.
  • It is “unfortunate” that Amylin discontinued work with pramlintide/metreleptin. Dr. Van Gaal said that pramlintide alone was able to achieve a nice change in daily caloric intake in a six-week observational study and also has an effect on binge eating, which no previous drug has tackled. It achieved solid weight loss sustained over 12 months, but still only at about 7-8%. Combining it with the leptin analog metreleptin, however, boosts weight loss efficacy to 13%, “one of the highest I’ve seen,” in a 20 week observational period.

Questions and Answers

Q: Maybe instead of focusing on body weight per se, which is unlikely to be addressed by any drug in the near future, perhaps drugs should focus on other metabolic effects? Ten percent is an arbitrary cut-off and these percent body weights are difficult to sustain by pharma alone.

Dr. Van Gaal: I agree that comorbid risk factors are more important and it is difficult to reflect efficacy purely in weight loss. But you know how the FDA and other regulatory agencies work.There are millions of people worldwide at BMIs 27 to 35 and if we can do weight maintenance even at 8% loss there is still a future for these therapies. Authorities should be more open to weight loss drugs, to be given not to everyone, but to the right patients.

Q: Would you prescribe orlistat and acarbose in combination?

Dr. Van Gaal: The effect on body weight of acarbose is very limited, from a number of trials. And in terms of tolerability, both of these drugs have unpleasant GI side effects. My clinical feeling is that people would fall into two cohorts with regard to these, where one would benefit from one and the other would benefit from the other. I don’t think that acarbose is a good option for the future.

Q: Would you want to prescribe GLP-1 agonists in prediabetes?

Dr. Van Gaal: Excellent question. So far only liraglutide is involved in trials in non-diabetic obese individuals. The ongoing trial SCALE specifically looks at not only weight reduction but also prevention of type 2 diabetes. The pre-diabetes arm will address this specifically and we will see. For prediabetes it is difficult to convince authorities to approve drugs and reimburse. But scientifically, most likely yes, though I will wait to say final conclusions.


Bariatric Surgery

Josep Vidal, MD (University of Barcelona, Barcelona, Spain)

Dr. Vidal said that “from a surgeon’s perspective, diabesity is a surgical disease.” Dr. Vidal felt that for patients with BMI over 35 kg/m2, bariatric surgery likely impacts the natural history of diabetes and benefits glycemic measures and other risk factors. However, complete remission has turned out to be less frequent than publicized and seems to be less the longer the follow up in studies. The duration of benefits therefore needs to be established. He “disagreed with the IDF” in terms of recommending bariatric surgery for people with diabetes in the 30-35 BMI range, stating he’d rather wait for results from some of the long list of trials looking at this population.

  • Dr. Vidal is not ready to recommend bariatric surgery for the 30-35 kg/m2 BMI range. There is a very long list of ongoing or enrolling trials looking at bariatric surgery in patients with type 2 diabetes in this range. He respectfully “disagreed with the IDF,” saying that further studies are needed to better establish benefits for persons with diabetes and BMI under 35 kg/m2. He would prefer to wait for the results of some of these trials.
  • Long-term partial remission is seen in a significant proportion of patients (two thirds in Dr. Vidal’s study) but complete remission (i.e., off all diabetes medications) is less common than previously reported. Unlike 78% complete remission famously reported in the Buchwald meta-analysis, Dr. Vidal’s group saw a complete remission in only 29%. His results have been accepted for publication in the Annals of Surgery (Jimenez et al.). If you look across bariatric surgery studies – not statistically valid but still of interest – the longer the follow-up the lower the remission rates. Though there is marked improvement in A1c (in many patients ~2%), and often a slight reduction in blood pressure and LDL, the majority of patients are not spared from all medications to achieve optimal metabolic control.
  • GLP-1 is not the sole determinant of improvements after bariatric surgery. It is a significant contributor to postprandial insulin secretion after gastric bypass. However, in a study by Dr. Vidal currently in review, patients who had type 2 diabetes at baseline and underwent surgery were separated into those whose diabetes remitted, those whose diabetes had remitted and then recurred, and those who failed any remission. They had different glucose responses to a meal, i.e., different glucose tolerance, but their GLP-1 response to the meal was exactly the same.
  • Baseline patient characteristics influence outcome, especially residual beta cell function. Interventions earlier in the course of diabetes increase the likelihood of remission. As such, insulin use prior to surgery was associated with more frequent recurrence. In a multivariatemodel that included several glycemic markers, the only characteristic that significantly predicted diabetes remission after bariatric surgery was beta cell function. Dr. Vidal noted that impaired beta cell function is potentially thought to be a greater contributor to hyperglycemia in people with BMI under 35, which may have implications about whether bariatric surgery should be used in this group.

Questions and Answers

Q: Which procedure exactly benefits diabetes patients more?

Dr. Vidal: There is no strong data to support either one. On the basis of my experience and knowledge, I would say gastric bypass but I have to admit that in our cohort, sleeve gastrectomy was as effective as bypass in lowering glucose. I don’t have strict criteria to select on procedure over the other.

Q: What is the complication rate at your center and what are common complications?

Dr. Vidal: In centers of excellence, mortality is approaching zero, similar to procedures like cholecystectomy. In other centers, surgeons may not be as skilled. But surgical complications are not a big issue in big hospitals. It is more long-term nutritional problems that we need to survey over time.

Q: Indian BMI cut-offs are lower. For that ethnic group, would you consider operating at lower BMIs?

Dr. Vidal: Yes, the 35 BMI cut-off is for Caucasians. Fat distribution varies in different ethnic groups. It does make sense to have lower cut-offs for South and East Asians.

Dr. Francesco Rubino, MD (Weill Cornell Medical College, New York, NY): I want to make an IDF statement clarification – they are not recommending surgery for anyone with diabetes under 35 kg/m2. They are just saying that if nothing else is working and a patient has a lot of risk factors, then surgery might be valuable. It also is important for ethnic groups other than white.


Diabesity: A Global Threat to Health and Economies

Paul Zimmet, MD PhD (Baker IDI Heart and Diabetes Institute, Melbourne, Australia)

Dr. Zimmet gave an overview of the staggering extent of the problem of “diabesity,” using this term to simultaneously describe both the obesity and diabetes epidemics and calling it “the biggest epidemic in human history.” We have heard the figures from the recent IDF atlas many times – 366 million with diabetes in the world and over 500 million expected by 2030, with about two thirds in the developing world. “We have some very simplistic people,” he said a bit harshly, “who think if you ban television advertising and Coca-Cola and some of these other things, you’re going to solve obesity.” This epidemic, however, is a “societal challenge” that is on par with climate change in how far-reaching, devastating, and complex it is, influenced and influencing hundreds of factors: social services, education, public safety, agriculture, development, labor, city planning, and maternal and child health to name a few. Similar to global warming, it will require an “integrated approach” with partnerships between governments, businesses, and civil society.

  • Along with environmental factors, epigenetic factors may actually be more important than genetic factors in predisposing to diabesity, said Dr. Zimmet. In other words, programming in the fetal environment can predispose to disease. Mothers exposed to very low calorie diets in the Dutch Winter Famine had babies with higher prevalence of obesity, hypertension, type 2 diabetes, cardiovascular disease, and even schizophrenia. It is possible thatgreater maternal nutrition efforts now, according to Dr. Zimmet, especially in the developing world, may be effective in reducing burden 20 or 30 years from now.

Questions and Answers

Q: Epidemiologic studies sometimes give some hint of an etiology. Is there any pattern that you have seen that might suggest one?

Dr. Zimmet: My view is that we have to spend a lot more time looking at the epigenetic story. This is based only on epidemiologic and public health data. Maternal and fetal issues have not been addressed significantly.



Elizabeth Venditti, PhD (University of Pittsburgh, Pittsburgh, PA)

Dr. Venditti gave a talk asking, ‘Why bother with lifestyle intervention at all?’ It is messy, difficult, and weight loss is disappointing. Ultimately, she believes behavior remains crucial to energy balance and there is no getting around that. Lifestyle interventions can influence broad-spectrum physical outcomes even with lesser weight loss. She reviewed the intervention techniques, most of which are very interesting and very old, such as operant conditioning. Studies show that eating habits are largely a product of stimulus control (you only crave a mega-tub of popcorn when you’re at the movies, i.e., it is learned) and that self-control strategies, though different in different individuals, can nevertheless be taught.


Symposium: Diabetes and Children

Childhood Obesity and Diabetes

Francine Kaufman, MD (Chief Medical Officer, Medtronic Diabetes, Los Angeles, CA)

Dr. Fran Kaufman, a tireless advocate for children with diabetes, discussed the growing and increasingly alarming prevalence of childhood obesity and diabetes. She emphasized the seriousness of type 2 diabetes in youth, some of the studies to date (HEALTHY, SEARCH), the limited approved treatment options (lifestyle intervention, metformin, and insulin), the need for more data, and the massive importance of prevention. We always relish hearing Dr. Kaufman’s wisdom and passion for improving the health of America’s children, and we hope policymakers are taking careful notes as they craft solutions to the biggest public health problem of our time.

  • The HEALTHY study examined the impact of a middle school-based intervention to improve risk factors for diabetes in high-risk children around the US (NEJM 363; 5, July 2010). Sixth grade students (n=6,367) were randomized to intervention or control schools and followed until eighth grade. The intervention schools had their food service and PE programs rearranged by the study investigators, with a social marketing campaign and behavior change curriculum also implemented. The mean age in the study was 11 years, with schools in seven cities and five states participating. Schools in the study were located in disadvantaged areas with high rates of minority students (53% Hispanic, 21% African American, 16.8% Caucasian). In families of study participants, 52% had a high school education or less and 16% had existing diabetes in the family. Forty-nine percent of kids in the study were overweight and 29% were obese.
    • The intervention resulted in significantly greater reductions in various indexes of adiposity compared to the control group. The intervention schools had greater reductions in the secondary outcomes of BMI z score, percentage of students with waist circumference at or above the 90th percentile, fasting insulin levels (p=0.04 for all comparisons), and prevalence of obesity (p=0.05). However, the comprehensive school- based program did not result in greater decreases in the combined prevalence of overweight and obesity compared to the control schools. Dr. Kaufman noted that many secular trends occurred during the study, which may have confounded the results (e.g., banning soda in schools, changes to physical education requirements).
  • Obesity in youth is associated with a number of very serious risk factors, comorbidities, and complications. Dr. Kaufman called the “most significant problem” the psychological issues these children face, explaining that they often feel as fragile and depressed as a child diagnosed with cancer. She also noted a number of other common risk factors in children with obesity: dyslipidemia (12-17% have a lipid abnormality), elevated blood pressure (7-16% in obese children and a 3.26 RR for hypertension for overweight and obese children), metabolic syndrome (4.2% in 12-19 year olds, 50% in those with a BMI >40 kg/m2), acanthosis nigricans (51% of African American youth, 8% in Caucasians), earlier onset of pubarche and thelarche, hyperandrogenism, polycystic ovarian syndrome, glomerulosclerosis, obstructive sleep apnea, non-alcoholic fatty liver disease, gallstones (2%), orthopedic problems, and pseudotumor cerebri.
  • In the last two decades, the incidence of type 2 diabetes in youth has increased around the world, with minority and socioeconomically disadvantaged children particularly affected. Dr. Kaufman mentioned that in her own clinic in Los Angeles, 20% of children with new-onset diabetes have type 2. Among American Indian youth with diabetes, two- thirds have type 2 diabetes. In the US, there are ~3,700 children diagnosed with type 2 diabetes every year, compared to ~16,000 diagnosed with type 1. Of these young type 2 cases, Dr. Kaufman emphasized that the vast majority occur after puberty.
  • Diagnosis of type 2 diabetes in children relies on standard cut points for A1c and glucose tolerance, which may be inappropriate. Unfortunately, we don’t have the data to concretely support one cut point or method over another. Dr. Kaufman also asserted the uselessness of a random C-peptide or insulin value for diagnostic purposes. To differentiate between type 1 and type 2 in children, she recommends testing for antibodies that signal evidence of autoimmunity (e.g., GAD).
  • Dr. Kaufman explained that treatment of type 2 diabetes in children is a confusing area, with approved options limited to metformin and insulin. Lifestyle intervention is always recommended, but doesn’t typically work well as the only treatment. Using approved therapies, metformin would be next, eventually followed by insulin. Unfortunately, there are not many clinical trials examining GLP-1s, sulfonylureas, DPP-4s, or combination therapies for treatment of youth with type 1 diabetes – while these are often used off-label, she reminded the audience that they may not be reimbursed. Overall, Dr. Kaufman called for more data on these topics.
  • Type 2 diabetes in youth is often accompanied by significant complications. Data from Philadelphia suggests 3.7% of children with type 2 diabetes fall into hyperglycemic hyperosmolar comas, with a 14% mortality rate. Among Pima Indians diagnosed under age 20, 22% had microalbuminuria, while 45% of indigenous Canadians had hypertension and 6% required dialysis, culminating in a 9% mortality rate. Finally, data from the SEARCH study suggests that 92% of children with type 2 diabetes have at least two cardiovascular risk factors.
  • Dr. Kaufman concluded on the important topic of prevention, emphasizing the need to change the obesogenic environment in which kids are growing up. She believes that children with obesity and type 2 diabetes need be treated within the context of the family, and that we need to focus our efforts on modifiable factors like home life, community, workplace, school/childcare, government, public health policy, social norms and behavior. We absolutely agree, and we hope that the Partnership for a Healthier America, the Alliance for a Healthier Generation, and other anti-childhood-obesity groups are putting us on the road to these changes.

Questions and Answers

Q: The predominance of obesity among children is evident among African American and Hispanic populations. Is it genetic? Environmental?

Dr. Kaufman: Genes do play a role, but it really is a different environment in the US depending on where you live in the cities. In Los Angeles, many more minority children live in parts of the city where there are no parks that are safe to go to. There are no grocery stores to buy healthy food options – what we call ‘food deserts.’ Also, those with low economic resource buy more calories from sugar and fat and less from fiber and macronutrients. Many ask, ‘Who is responsible?’ Some say it’s a person’s responsibility to prevent obesity. But this is not a level playing field across socioeconomic racial groups. There are many disadvantaged people who are in a much greater obesogenic environment.

Q: Is there any data regarding bariatric surgery in children with type 2 diabetes?

Dr. Kaufman: There is data on bariatric surgery in youth – mainly on the LAP-BAND rather than full gastric bypass. It has been successful in reducing obesity and improving or ameliorating type 2 diabetes. This has mostly been from weight loss due to banding rather than gastric bypass. There are centers in the US that are performing these procedures in adolescents. The criteria are surgery for extremely obese children in the 99th percentile of BMI and comorbidities.

Q: Can you explain what is appropriate to discriminate type 1 diabetes and type 2 diabetes in youth?

Dr. Kaufman: There was data presented yesterday from the SEARCH trial. A risk factor score was used that includes waist circumference, A1c, and triglycerides. Short of that, if you look at pre-pubertal children, they most likely have type 1 diabetes. In post-pubertal and intra-pubertal children, if they’re not overweight or obese, they should be presumed to have type 1 diabetes. If they are overweight or obese, particularly in a high-risk ethnic group, the differentiator we are using is antibodies, GAD or IAD-2. Ketoacidosis would make you think more type 1. Even with antibodies, however, 15% of type 1 patients have no antibodies and 15% of type 2 patients have antibodies. They may have a hybrid form of diabetes.

Q: I would like to thank you for your lecture. In the Gulf countries, we are facing a big disaster in with the increasing prevalence of obesity. In my country of Bahrain, 25% of school-age children are overweight or obese. Do you have a strategic plan for physical activity and nutrition?

Dr. Kaufman: From our HEALTHY study, we learned that a comprehensive plan in schools can have an effect on obesity. Getting children more active needs to start in preschool and childcare centers. We also need to look at the prenatal period. When we looked at the physical education periods in the HEALTHY study, they were 45 minutes long. The average child had eight minutes of physical activity. It would be great to just get that to 22 minutes.


Debate: Is Bariatric Surgery an Appropriate Treatment for the Majority of Type 2 Diabetic Patients?


Francesco Rubino, MD (Weill Cornell Medical College, New York, NY)

Dr. Rubino refrained from saying that bariatric surgery should be done on the majority of patients with type 2 diabetes, although he does believe that surgery would help most patients with diabetes, either directly or by influencing diabetes care in general. He said that it was “the most high impact discovery in diabetes since the discovery of insulin.” The fact that you can get complete remission – “whether 80% do or 30% do” – deserves “maximal attention” by everyone in this field.

  • Less than 2% of eligible patients (based on old NIH criteria) actually undergo bariatric surgery in the US. Dr. Rubino said, “If a pill or shot could control glycemia, body weight, improve lipids, blood pressure, and increase survival, would it be acceptable that over 99% of patients do not have access to that treatment?” He thinks barriers are not only cost and coverage but predominantly cultural – contending with the stigma of obesity and the general feeling that bariatric surgery is “radical, dangerous, or drastic.” It has the same mortality as one of the safest surgeries done today, a laparoscopic cholecystectomy, which is “done all the time and doesn’t have a bad reputation.” Just as smoking cessation cannot cure someone’s lung cancer, Dr. Rubino said, so it is that modifying behavior is not enough to get the disease to go away. Unfortunately, he said, this is how most physicians and laypeople think. Surgery often produces partial or complete remission, with some improvement in cardiac risk factors, reduced medication need, sustainable weight loss, and improved survival (over years – “it takes time to see this because diabetes takes time to produce mortality”). The IDF statement presented at the 2nd World Congress of Diabetes Surgery states that surgery should be considered an alternative treatment option in patients with BMI of 30-35 kg/m2 when diabetes cannot adequately be controlled by an optimal medical regimen and especially in the presence of other risk factors (in Asians, the BMI range is 27.5-32.5 kg/m2).
  • Strict BMI-based criteria for selection of candidates for bariatric surgery is “discriminatory.” It discriminates based on age, gender, and other factors that may or may not impact whether a person could benefit from surgery. Excess weight is a symptom of the disease, not the disease itself, said Dr. Rubino. “There are metabolically healthy obese individuals and there are non-obese individuals with the metabolic syndrome – we should be operating on the latter.” We need to stratify by risk, he said, not BMI.



Jonathan Pinkney, MD (Peninsula College of Medicine and Dentistry, Plymouth, UK)

Dr. Pinkney admitted that bariatric surgery is a good treatment but argued that it is not appropriate for most patients. Beyond the argument that we just can’t roll it out to everyone in the world with diabetes, he also argued that data on its benefits was not yet sufficient. For example, it harms quality of life in a way that should not be overlooked. Dr. Rubino felt in Q&A that most patients are very happy after surgery. It was determined, not surprisingly, that whether to recommend surgery should be decided on a case-by-case basis.

  • It was acknowledged in the IDF statement that we need studies to map out long- term effects, complications, differences between procedures, and evidence that surgery stabilizes or improves microvascular complications of diabetes, said Dr. Pinkney. Unlike medical treatments, there is of course mortality from surgical complications. He says medications are often “thrown to the bin” for mortality rates much lower than those for surgery. There is also data emerging on higher rates of diabetes relapse than previously thought. This would require that patients continue to be monitored for recurrence of their diabetes and also calls into question long-term durability of surgery.
  • “There are quality of life issues” with the side effects of bariatric surgery. This includes “freedom to eat.” If we don’t look specifically at whether we restore normal duration and quality of life, it is “paternalistic” to put patients through this. Eating problems can persist he said, with some patients experiencing continued vomiting. He said that the patient-reported number one reason diabetes impacts quality of life is the “lack of freedom to eat as I wish,” which is not helped with bariatric surgery. He “guaranteed” that a patient who undergoes surgery will become anemic, will probably have osteomalacia, and occasionally have neurologic syndromes, all from vitamin deficiency. You “substitute diabetes care for this” – annual monitoring and micronutrient replacement.

Questions and Answers

Q: If bariatric surgery is so safe, why do I face reluctance from surgeons to operate on adolescents, even with BMIs in the 40s and 50s?

Dr. Rubino: Bariatric surgery will become more popular in the next few years. But right now, most bariatric surgeons are not used to dealing with pediatric patients. The thought of changing things around in a patient who is still growing could be a concern. We are starting to perform adolescent surgery for metabolic disease at our center though.

Q: Is there any study on whether diabetic patients who have had the surgery would recommend it to others?

Dr. Rubino: I think there are but I don’t remember the numbers. I would disagree with Dr. Pinkney’s point that patients do not have good quality of life after surgery. Patients tell you, with a few exceptions, “Doctor, I don’t know what you did but I’ve never been so happy before.” It’s not necessarily true that they’re not eating well. Some have vomiting, but they are not the majority.

Q: Has anyone looked at the impact of surgery on end-organ damage, obstructive sleep apnea, arthritis, etc.?

Dr. Pinkney: There are good data of regression of obstructive sleep apnea and fatty liver and this is one of the many reasons to consider surgery in people who are very obese. But I think this is different from the argument that we should perform it in people who may not be concerned about weight.

Q: Which type of bariatric surgery is more effective to control diabetes?

Dr. Rubino: We don’t have data comparing the surgeries. In general, the more complex, the more effective. Bypass is more effective than banding. Biliopancreatic diversion is even more complex, with more effect on glycemia, but significant side effects with malabsorption. So most people say bypass. Recently sleeve gastrectomy has been shown to be very powerful, but bypass still has an edge.

Q: How many patients need cosmetic surgery for skin folds after all this weight loss?

Dr. Rubino: Some patients who are very obese need cosmetic surgery for skin folds, but that is not usually the case with a BMI that is on the lower side to start with.


Teaching Lecture

An Effective Obesity Program for Diabetes - What Are The Key Components

Arya Sharma, MD, PhD (University of Alberta, Alberta, Canada)

Dr. Arya Sharma – noted obesity-focused physician, blogger, and sometimes-FDA-advisory-committee- testifier – delivered a rich set of insights on clinical obesity management. A firm believer that obesity should be managed like other disease, Dr. Sharma described a staging system for classifying high body weight based on its health effects, and he emphasized that obesity is a lifetime diagnosis (even if a successful treatment can bring it “into remission” for a long time) – in both cases, language reminiscent of cancer and other seriously regarded medical conditions. He emphasized that differential diagnosis for obesity should include a full consideration of factors that can contribute to weight gain (e.g., mental health, sociocultural influences, side effects of medications, lack of education, etc.) so that referrals and treatment plans are more effective (e.g., he noted that a nutritionist is not trained to help patients who are overweight because they are depressed). Dr. Sharma was generally positive on drugs and surgery as options for some patients, and he commended the IDF for recommending that bariatric surgery be considered as a type 2 diabetes treatment (see our coverage of the IDF’s position statement on bariatric surgery at Day #1 of last year’s WCIDT in the March 29, 2011 Closer Look at However, he stressed that no drug or surgery is a miracle cure and that good obesity management involves setting realistic expectations and educating patients about the benefits of modest weight loss (or even a slowed pace of weight gain, if that is all that can be achieved). A rather gritty truth, but one we good for the clinical community (and the world as a whole) to hear.

  • Dr. Sharma described the Edmonton Obesity Staging System (EOSS), a rubric for classifying obesity by its health effects rather than simply the numerical value of BMI. Stage 0 denotes no medical, mental, or functional problems; Stage 1 denotes mild obesity- related issues; Stage 2 denotes moderate complications; Stage 3 denotes severe side effects (e.g., clinically relevant end-organ damage); and Stage 4 refers to obesity-induced damage so severe that weight loss cannot reverse it (e.g., blindness from diabetic retinopathy). Dr. Sharma and colleagues found that compared to BMI, EOSS was much more predictive of mortality in the NHANES III cohort (Padwal et al., CMAJ 2011). He said that this has important implications for how we define obesity. For example, roughly 20% of people with BMI >35 kg/m2 have mild or no comorbidities, while nearly 60% of people considered merely “overweight” (BMI < 25 kg/m2) already have moderate or severe weight-related comorbidities.
  • When thinking about the obesity assessment for any given patient, Dr. Sharma recommended asking three questions: 1) Why is this person obese? 2) How is this obesity affecting the patient? (a question that can in part be answered with the EOSS staging system), and 3) What is the best treatment plan? To help answer these questions, he shared “the four M’s of obesity assessment” (as a mnemonic, he said, “Think of a large person eating two M&Ms”). These are: Mental (“If there is an uncontrolled mental health issue, you can forget about obesity management.”), Mechanical (“If a patient has plantar fasciitis, don’t expect them to walk 10,000 steps a day; if they have stress incontinence, don’t expect them to lift weights.”), Metabolic (e.g., diabetes, dyslipidemia, and other comorbidities that can add make weight-loss more difficult), and Monetary (“Not having enough money can make you obese,believe it or not…though in some countries, having too much money can make you obese.”). Further complicating matters is the fact that many common treatments for mental, mechanical, and metabolic disease can themselves contribute to weight gain.
  • Dr. Sharma said that obesity should be thought of as “obesities,” since energy imbalance can reflect a variety of etiologies and require a variety of treatments. Drawing on his background as a nephrologist, he explained that physicians are trained to know that edema – “positive fluid balance” – can be a symptom of many different underlying disorders. But when it comes to obesity, doctors often simply advise, “eat less and exercise more.” Dr. Sharma said that this would be as reductionist as telling all edema patients to “drink less and pee more” (or telling depression patients that they should be happier). In fact, he noted that energy balance is complex, especially since metabolic rate can drop by 10-20% due to weight loss – a downturn that is extremely difficult to counteract with exercise. (Indeed, Dr. Sharma stressed that exercise makes up a very small fraction of most people’s total energy expenditure.) To further illustrate the complexity, he said that even something like “overeating” is not a diagnosis, but a symptom of some more basic cause – possibly depression or poorly managed hunger (relatively easy to treat), possibly a personality disorder (much harder to treat). Dr. Sharma told the audience that until they have considered the wide range of potential root causes (sociocultural, biomedical, pharmaceutical, etc.), they have not truly performed a differential diagnosis of obesity.
  • Dr. Sharma emphasized that patients who achieve 3-5% sustainable weight loss with lifestyle intervention are doing “pretty well,” with two-to-fourfold greater results to be expected from medical or surgical interventions. (He cautioned against paying much attention to “crazy” outliers who go from morbid obesity to running marathons through diet and exercise: “If you think your guy with a BMI of 35 kg/m2 is going back to 25 kg/m2, you are dreaming – you are not practicing evidence-based medicine.”) The good news about lifestyle intervention is that 5% weight loss is associated with significantly better metabolic health, as seen in the Diabetes Prevention Program and Look AHEAD. Dr. Sharma spent little time discussing current or upcoming pharmaceutical interventions (he mentioned that liraglutide is in late development for a weight-loss indication), but he noted that lifestyle-plus-medication can provide 5-10% weight loss. As for bariatric surgery, Dr. Sharma said to expect only about 20-30% body weight loss – not a miracle cure. He skeptically reviewed the image of a skinny model on a LAP- BAND billboard, displaying one of his patients (who looked obese even after losing 200 pounds through surgery) as a more typical example of bariatric surgery success. He also noted that bariatric surgery is = associated with risks of surgical complications and psychological distress (“If someone is treating their depression with food, what happens when you take away their treatment?” Dr. Sharma asked rhetorically: “They get depressed.”)
  • Dr. Sharma outlined “the five A’s” of obesity treatment in primary care (Sharma, unpublished). Paraphrased slightly, they are as follows:
    • Ask for permission to discuss weight and explore readiness for change.
    • Assess obesity-related risk and potential ‘root causes’ of weight gain
    • Advise on the need for long-term efforts and the benefits of modest weight loss
    • Agree on realistic weight-loss expectations
    • Assist in identifying and addressing treatment barriers and, perhaps most importantly, in arranging follow-up

Questions and Answers

Q: You’ve shown us the graph of typical weight reduction with lifestyle intervention. Does your center get better results than 4-5%?

Dr. Sharma: It depends on the cause. Depression is easy, but addictions are difficult. It depends on the patient. Think of obesity as a chronic progressive condition. If someone has been gaining five pounds per year for past 10 years, but you put on only 2.5 pounds per year after an intervention, that is a successful intervention. “Weight loss” is just a number on the scale.

Q: What do you think about obesity in children?

Dr. Sharma: In many cases it’s a family problem. If you can’t get parents in, you will probably be wasting time on the kids.

Q: What do you think about diabesity?

Dr. Sharma: I wish I came up with it. However, in some ways it reduces obesity to a diabetes problem. Obesity is so much more – sleep apnea, blood pressure, etc.

Q: It is a term that is getting attention worldwide.

Dr. Sharma: Diabetes and obesity are definitely very related. If I could snap my fingers and make obesity disappear, a lot of you guys might be out of business.


Corporate Symposium: The Effects of GLP-1 on Satiety and Metabolism (Sponsored by Novo Nordisk)

Obesity and Prediabetes: A Disease Continuum

Nick Finer, MD (University College Hospital, London, UK)

Dr. Finer gave a valuable overview of prediabetes in the symposium’s opening presentation. He began with definitions of obesity, prediabetes, and diabetes, emphasizing the utility of the new Edmonton Obesity Staging System. The basis of the system is similar to the staging of cancer, and NHANES data suggests it can better explain the differences in mortality compared to just using BMI. Dr. Finer transitioned to a discussion of fat distribution and insulin resistance, noting that where fat is stored is “critically important” – he called subcutaneous fat “relatively healthy,” while visceral depots are deleterious. To define the scope of the prediabetes problem, he showed data that 314 million people worldwide have prediabetes, and those with both prediabetes and obesity have a 17 times greater risk of developing type 2 diabetes. Dr. Finer also reviewed the health risks of prediabetes, which include increased risk of CVD and mortality (DECODE study, EPIC Norfolk study) and can be reduced or eliminated with only modest weight loss. He closed with a review of some of the “highly effective” lifestyle intervention trials (e.g., DPP, DPS, Malmo, Da Qing), arguing that these should be widely applied. Even though weight loss was often not maintained in the long term after these study ended, Dr. Finer argued that the benefits were still substantial and sustained years after the fact. (Editors note – this presentation was quite similar to Dr. Finer’s identically titled talk at EASD. For more information, see page 165 of our EASD full report at

  • The Edmonton obesity staging system classifies obesity along a clinical continuum (Padwal et al., CMAJ 2011). Using NHANES data, different categories of BMI and Edmonton obesity stages were plotted against mortality. While different levels of BMI resulted in littledifferences in mortality rates, using the Edmonton obesity staging system resulted in very distinct mortality rates.
    • Stage 0: No apparent risk factors (e.g., blood pressure, fasting glucose levels), physical symptoms, psychopathology, functional limitations, and/or impairment of well-being related to obesity.
    • Stage 1: Presence of obesity-related subclinical risk factors (e.g., borderline hypertension, impaired fasting glucose), mild physical symptoms, mild psychopathology, mild functional limitations, and/or mild impairment of well-being.
    • Stage 2: Establishment of obesity related chronic disease (hypertension, type 2 diabetes), moderate limitations in activities of daily living and/or well-being.
    • Stage 3: Established end organ damage such as myocardial infarction, heart failure, stroke, significant psychopathology, significant functional limitations, and/or impairment of well-being.
    • Stage 4: Severe (potentially end-stage) disabilities from obesity-related chronic diseases, severe disabling psychopathology, severe functional limitations, and/or severe impairment of well-being (i.e., “equivalent to disseminated cancer malignancy”).


6. Emerging Markets

Symposium: Social and Economic Impact of Diabetes: New Data

Social and Economic Impact of Diabetes in China: Challenge and Opportunity

Jianzhong Xiao, MD (China-Japan Friendship Hospital, Beijing, China)

Dr. Xiao presented data from IDF’s Diabetes Impact study in China. The initiative interviewed 3,035 people from 10 provinces in China and gathered particularly interesting data on costs and medication use. Overall, people with diabetes in China spend two to four times more on medical expenses than their non-diabetic counterparts (age, sex, and location-matched) and use about twice the number of hospital services. We found the data on medication use quite valuable – less than half of patients in the study were on orals, less than one in ten was on insulin, and only one out of fifty was on a statin. Metformin was the most commonly prescribed medication (22% of individuals), followed by sulfonylureas (16%), insulin (9%), and alpha-glucosidase inhibitors (8%). Dr. Xiao closed by highlighting China’s heavily ageing population, which will only exacerbate the burden diabetes will place on society.

  • This study involved in-person interviews with 1,482 individuals with diabetes and 1,553 people with normal glucose tolerance. On average, those with diabetes were 57 years old, 58% female, 75% urban residents, and had an income of 3,550 RMB (the time period associated with this income was not specified). The study used a representative sample in 10 out of 14 provinces in China, and participants were recruited from the 2007-2008 national study (Yang et al., NEJM 2010). People with diabetes were matched to a non-diabetic control group based on similar age, sex, and residence. The original Yang et al., study had a response rate of 87%, compared to 67% in this study.
  • In this Chinese study, less than half of study participants were on orals and less than 10% were on insulin. Rates of therapy use increased along with longer diabetes duration, although only 29% of participants were on insulin after ten or more years of diabetes – this is justanother reminder of the challenges in moving patients onto insulin (logistical, socioeconomic, philosophical, behavioral, and others).


<2 years diabetes duration


































  • Metformin was the most commonly prescribed therapy in this study, followed by sulfonylureas, and insulin. With longer diabetes duration, sulfonylureas and insulin were increasingly used.


<2 years diabetes duration






















Alpha-glucosidase inhibitor












  • Compared to the control group, those with diabetes had 2-4 times higher annual medical care expenditures.
  Diabetes Non-diabetic Adjusted Odds Ratio
Total annual payment per person for inpatient care 2,527 RMB 508 RMB 4.2*
Total annual payment per person for outpatient care 2,352 804 2.44*
Total annual point of service expenditure 5,768 1,454 3.33*

*P <0.05

  • People with diabetes in China used approximately double the hospital services and had double the annual medical events of those without diabetes.



Adjusted Odds Ratio

Any Hospital Encounter

(last 90 days for inpatient, emergency, outpatient)




Inpatient Admissions (days)

0.19 (2.47)

0.08 (1.08)

1.93* (1.98)*

Outpatient Visits




*P <0.05


Social and Economic Impact of Diabetes: Africa

Kaushik Ramaiya, MBBS (Shree Hindu Mandal Hospital, Dar es Salaam, Tanzania)

In a similar fashion to the presentation on China, Dr. Ramaiya discussed the IDF Diabetes Impact Study in Africa. Data was collected from a total of 2,139 people in Kenya, Mali, South Africa, Tanzania, and Cameroon. Overall, people with diabetes had out of pocket expenses that were ten times higher than those without diabetes and had ten times higher point of service expenses (per most recent admission). Additionally, those with diabetes in the study were five times more likely to visit a hospital, four times more likely to report an inpatient hospital stay in the last 90 days, and used 32.5 times more inpatient days than those without diabetes. Self-monitoring of blood glucose was very infrequent, with 3.7% of patients testing at least once per day, 13.7% testing less than once per month, and 6.3% testing less than once per week. Sulfonylureas were the most commonly used therapy (47%), followed by metformin (46%), insulin (30%), and acarbose (0.3%). To close, Dr. Ramaiya emphasized that diagnosed diabetes has a much larger relative impact in developing African countries than it does in the industrialized world. We’d agree and we’re glad to see data collection studies like this, as they will certainly be needed to build a case for more government resources in the future.


Introduction: IDF Diabetes Impact Studies

Jonathan Brown, MPP, PhD (Chair, IDF Task Force on Diabetes Health Economics)

Dr. Jonathan Brown began this sparsely attended afternoon session by describing the methods and rationale of the IDF Diabetes Impact studies in Africa, China, and Kazakhstan – data on each country was subsequently presented during the session. The studies are intended to better characterize diabetes in low and middle-income countries by interviewing people with diabetes about the costs, quality, and nature of their medical care. Dr. Brown noted that for comparison, total medical cost ratios (people with diabetes compared to those without) in industrialized countries are as low as 1.5, with Germany at 2.0, the US average at 2.5, and the IDF Diabetes Atlas assuming 2.5.

  • The IDF Diabetes Impact studies aim to address the lack of data on the economic and social impact of diabetes in low- and middle-income countries. The studies stem from the difficulty of accurately measuring the incremental impact of diabetes and the need for information on more than just medical care expenditure. Additionally, standardized measures areneeded to allow cross-country comparisons. In these studies, this was addressed by using cost ratios.
  • The 25-page interview asked a wide variety of questions to better characterize the social and economic impact of diabetes. These included demographics, work situation, income, wealth, health utility index, use of medical care services over the last three months, out of pocket payments, travel costs, the impact of ill health on the person and family, possession of medicines, prices, amounts purchased, quality of care (process measures), and access to care.
  • The studies used mostly standardized methods of data collection and analysis. Individuals in Africa were identified through hospital registries, while Chinese individuals were identified from a population-based screening study. In China, total expenditures for medical care were based on what patients paid; in Africa, institutions supplied the charges and costs. Total costs combine expenses for medical services, payments for medicines, travel costs, and informal payments. Annual rates were calculated by multiplying three-month rates by four. Medication use and costs were calculated based on self-reported dosing, adherence, out of pocket payments, and days supply. Overall differences in use cost per unit of service were calculated using a two-stage multivariable OLS and logistic models.


Panel Discussion

Jianzhong Xiao, MD (China-Japan Friendship Hospital, Beijing, China); Kaushik Ramaiya, MBBS (Shree Hindu Mandal Hospital, Dar es Salaam, Tanzania); Jonathan Brown, MPP, PhD (Chair, IDF Task Force on Diabetes Health Economics); Katarzyna Kissimova- Skarbek, PhD (Jagiellonian University of Krakow, Poland).

Q: I had a question on the insulin treatment rate of 9.2% in China. This number seems to be lower than what we have believed – we’ve thought it was about 20-25%.

Dr. Xiao: Our data come is based on the national epidemiological studies. The majority of patients showed a duration of diabetes less than five years. Nearly half of patients with diabetes weren’t treated. This may not have been representative data.

Dr. Brown: To summarize the results of the studies quickly – overall expenditure ratios in the developed world are around 1.5-2.5. In China, we were at 3.3 on a population of diagnosed and undiagnosed patients. Many were not treated and not followed up from the original study. This figure in China is higher than in Europe and the US. That’s something right there. But when we look at Africa and Kazakhstan, it’s up in the range of a ten times difference. That’s really quite huge. In Africa, the explanation is that people aren’t diagnosed with diabetes until they are ill. We’re comparing the sickest ‘n’ of the diabetes population. It’s not clear in Kazakhstan that this is the case. To me, this is the weirdest thing.

Q: I work on the IDF Diabetes Atlas. What advice can you give us for the future diabetes expenditure data from the results of your study? How can we update that number?

Dr. Kissimova-Skarbek: I know that for Kazakhstan, it’s not so convenient to use the ratio provided from countries with well-controlled patients.

Dr. Brown: The Atlas shows that $466 million dollars were spent in 2011. The general picture is that 90% of the spending – using the same ratio in every country – is spent in Europe, US, Australia, and Japan. Fifty percent of all money is spent in the US. After these studies, we know the ratios are much higher in the rest of the world. That will change the picture; that means a lot more money is being spent outside of the first world. Instead of 90%, it might be 50%. This is a good thing. We should spend more money in some places around the world and less in others. This will change the big picture quite a bit.


Symposium: Diabetes in China

Prevalence of Diabetes and CVD Risk Factors in China

Wenying Yang, MD (China-Japan Friendship Hospital, Beijing, China)

Dr. Yang is a very big deal – as a reminder, she is the first author on the groundbreaking 2007-08 China National Diabetes and Metabolic Disorders Study (n>46,000), reviewing data on diabetes prevalence (Yang et al., NEJM 2010), CVD and risk factor prevalence (Yang et al., Euro Heart J 2011), and the relationship between glucose tolerance and cardiovascular disease (CVD). Notably, in this talk, she had new data on complications. Specifically, she said that the prevalence of CVD was notably higher among people with diabetes (3.83%) than in those with normoglycemia (1.06%) or impaired glucose regulation (1.79%), with a similar relationship seen for both stroke and coronary heart disease. Given that a larger follow-up study confirmed Dr. Yang’s initially controversial estimate for Chinese diabetes prevalence (9.7%), we expect that her group’s daunting data on glycemia and CVD are also right on the money. Though no questions were asked after this (extremely clear) presentation, it wasn’t for lack of appreciation – as seen when moderator and IDF president-elect Sir Michael Hirst led the crowd in enthusiastic applause for Dr. Yang’s work.

  • With the same cohort used to estimate Chinese diabetes prevalence, Dr. Yang presented unpublished data on the relationship between glycemic status and cardiovascular disease.As might be expected, risk for stroke, heart failure, and composite CVD were highest in people with diabetes, lowest in those with normoglycemia, and intermediate in people with impaired glucose regulation. She noted that the prevalence of stroke was greater at higher age, higher BMI, and higher systolic blood pressure. Dr. Yang also stratified the impaired glucose regulation group according to the specific glucose defect, showing that CVD prevalence was highest among people with isolated impaired glucose tolerance (2.09%) as compared to isolated impaired fasting glucose (0.66%) or both impaired fasting glucose and impaired glucose tolerance (1.59%).
















IGR = impaired glucose regulation, CHD = coronary heart disease, CVD = cardiovascular disease, including stroke and/or CHD.


Symposium: Diabetes and Children

Socioeconomic Problems of Children with Type 1 Diabetes

Sharad Pendsey, MD (DREAM Trust, Nagpur, India)

Dr. Pendsey gave a sobering overview of type 1 diabetes in children in India, noting the incredible economic and social challenges that accompany treatment. He began by explaining some of the realities in India: 40% of the population lives below the poverty line, 26% are illiterate, government healthcare is overburdened (by communicable diseases), and there is no health insurance and widespread gender discrimination. He stated that there are 97,600 cases of children with type 1 diabetes in India, and treatment often involves expenditures that consume 25-30% of family income. As a result, families often look to cut corners by reducing or skipping insulin doses, minimally buying syringes, and rarely buying test strips. The situation inspired Dr. Pendsey to establish the DREAM Trust, which seeks to bring diabetes supplies to type 1s in India. When it started, the organization’s goal was to get A1cs into the single digits (they were as high as 15-18%!). The Trust has distributed 40,000 vial of human insulin, 3,500 pens, educational and travel grants, and few strips. Overall, DREAM’s work in a cohort of 600 children has reduced the school dropout rate, decreased the rate of growth retardation, and dropped the mortality rate from 12% in 1996 to 2% – impressive and inspiring work. We took special note of Dr. Pendsey’s closing statement, “No child with type 1 diabetes should ever die for want of insulin.”


Meet the Expert

The Role of the Newer Drugs for Diabetes in Developing Countries

Anoop Misra, MD (Fortis Hospitals, New Delhi, India)

This informative talk gave us a strong sense of what is required for diabetes pharmacotherapy in the developing world. Dr. Misra spoke specifically about India but felt that most of his points could be generalized to most of the developing world, i.e., low-resource settings. Cost of drugs is a paramount consideration, but ignoring this, DPP-4 inhibitors came away as the real favorite. Only 5% of Dr. Misra’s patients can afford them, but it sounded like as though if cost weren’t an issue, he and others would prescribe them to a large majority of patients (especially in a fixed dose combination with metformin, he mentioned in Q&A). Because resource-poor settings mean patients see the doctor infrequently and briefly (three minutes every two years was his example), have limited access to hospitals, and often have poor health literacy, hypoglycemia is unacceptable, as it is not unlikely to cause death. Similarly, durable and easy treatment is important. Criteria for an ideal drug in the developing world are: cheap, simple dosing, oral, widely available, alters basic pathophysiology of diabetes, and has few adverse effects so you can “administer and forget.”

  • Three quarters of the 200 million people currently living with diabetes are in low- and middle-income countries. This population is also growing rapidly, leading to incredible financial drains on countries like China and India. India has a particularly severe epidemic, which is compounded by physiological factors that cause Indians to get diabetes at a lower BMI and almost a decade earlier than their Western counterparts, not infrequently in their late 20s. (Editor’s note – IDF estimates 366 million. The difference may related to diagnosed vs. undiagnosed.)
  • Hypoglycemia must be avoided in the developing world, because it kills. Many people have very low health literacy and diabetes awareness and are very far from medical facilities.Sulfonylureas are cheap and available, so they are used frequently, but as Dr. Misra said, “We don’t know how many people are dying.” Starting insulin is also a major concern because of hypoglycemia, he noted. There is a massive divide in psychosocioeconomic conditions in the developing world. If a patient is motivated, educated, capable, and has access, then Dr. Misra recommends going with intensive therapy to get that patient to goal. However, there are manypatients who don’t match this description, and in these patients, one goes easy and has loose targets, simply in order to avoid hypoglycemia.
  • Dr. Misra said that DPP-4 inhibitors would come away as far and away the favorite for the developing world, if only they weren’t so expensive. These are the drugs,” Dr. Misra said almost lovingly. They control glucose without the risk of hypoglycemia. They have convenient dosing making them easy for patients to understand and take. They are relatively durable, which is important given follow up is so infrequent. They are weight-neutral, which is of great importance in India, where very small weight changes increase insulin resistance greatly. And they have no evidence yet of major adverse effects, which again is an important consideration where people have such poor access to doctors. Besides cost and availability, DPP-4 inhibitors fill all the criteria that Dr. Misra cited as important for a diabetes drug in the developing world. In particular, he mentioned, linagliptin (BI/Lilly’s Trajenta) might be a good option because it seems to be safe in renal impairment – renal function is of course difficult to monitor in resource-poor settings. He mentioned vildagliptin (Novartis’ Galvus) as well for its possible lipid benefits, which may be especially helpful for Indians, who seem to get dyslipidemia at an earlier age than other ethnicities. Besides DPP-4 inhibitors, Dr. Misra mentions a couple of injectables: exenatide LAR (Amylin/Alkermes’ Bydureon) “may be attractive” given that it is powerful and yet dosing is only once weekly; insulin degludec, which he said has to be delivered only three times per week, “could be a good option in those that have poor compliance” (Editor’s note: after unsuccessful phase 3 results, Novo Nordisk is no longer pursuing an indication for every-other-day dosing of insulin degludec).
  • The ADA guidelines tend not to be followed. Dr. Misra showed a modified ADA algorithm. After lifestyle modification and metformin fail, second line treatment in India actually rarely consists of basal insulin, due to cost, hesitance to use injections, and hypoglycemia. Adding a sulfonylurea to metformin is favored, because, despite the risk of hypoglycemia, it is effective at a very low cost. Adding pioglitazone to metformin is also widely accepted due to low cost but, yet again, in spite of adverse effects. GLP-1 analogs are not favored due to their high cost and low availability as well as hesitance to use an injectable.

Questions and Answers

Comment: In diseases like tuberculosis and malaria, we follow strict guidelines, but in diabetes, individualized treatment is more important.

Dr. Misra: Very well said. ADA guidelines are not being followed in India. There are so many factors at play, unlike in malaria or TB. You individualize treatment greatly.

Q: What do you think about the fixed dose combination of metformin and sitagliptin?

Dr. Misra: Sitagliptin and metformin in a single tablet is a very attractive therapy. Any patient with diabetes has 10-15 medications and this decreases the number of tablets. Also, it has powerful glucose- lowering effect and it may produce less GI irritation than metformin or a gliptin alone. This means that it is more powerful, but also more tolerable and with better compliance. I have personally been using it every day in my patients, and gliptins I think will move up in the algorithm to number one.

Q: In Dubai, we see a lot of Indian and Pakistani patients, but the different kind of fixed- dose combination drugs are not available. In that case, what drug is the most suitable?

Dr. Misra: You’re right. Each country and scenario is different. Sitagliptin (Merck’s Januvia) is not in the national drug formulary, and insurance probably does not list it as a first-line drug. In my clinic, hardly 5% can actually afford sitagliptin. But 90% of your patients will probably be controlled with some combination of the three first-line drugs – TZDs, sulfonylureas, and metformin – and in the others you will just have to use insulin. We have to work within our confines.

Q: We know that diabetes is related to many other conditions, such as dyslipidemia and hypertension. Do you think that such drug combinations are useful?

Dr. Misra: As far as I am concerned, I never believe in combinations except a very few, like gliptin and metformin, and maybe aspirin and statin. I say confine yourself to one drug unless the drugs are clearly proven to be synergistic.

Q: How do you choose a patient for gliptins?

Dr. Misra: The number-one consideration is cost, no doubt. If a poor patient comes in that can be treated with a sulfonylurea and metformin, I would never treat with a gliptin. I would currently put gliptin second line. Some people are using it as first line, but I think we have to wait for that. Second line definitely and combining it with metformin is a very attractive option.


Teaching Lecture

Challenges of Insulin Therapy in Developing Societies

Anant Nigam, MD (CEO, Diabetes India)

Dr. Nigam discussed a plethora of obstacles to diabetes care in developing countries, especially (but not solely) insulin access. He highlighted what he called the “insulin dilemma” – the fact that insulin often costs even more in poor countries than rich countries (due in part to taxes, poor reimbursement, kickbacks for intermediaries, retailer markups, etc.) – and the “insulin paradox” – the fact that market forces and individual beliefs have caused human insulin to replace animal insulin and analog insulin to replace human insulin. The insulin paradox persists despite international consensus statements that for most patients, analogs’ benefits on nocturnal hypoglycemia do not justify their much-greater cost and their unknown long-term cancer risk (e.g., as recently stated at WHO’s 18th Expert Committee on the Selection and Use of Essential Medicines in Accra, Ghana in March 2011). Dr. Nigam also mentioned some bright spots, such as the International Insulin Foundation’s Rapid Assessment Protocol for Insulin Access (RAPIA) that has helped improve availability in several developing countries, “twinning initiatives” whereby a developed nation partners with a developing one for better diabetes care, and ongoing multi-year supply distribution programs from Lilly and Novo Nordisk. Still, for the majority of people living with diabetes worldwide, the conditions remain bleak– everyone involved (especially governments, the private sector, and international health organizations) needs to play a role in the solutions.

  • Dr. Nigam discussed some of the pricing issues facing patients in developing countries, with a focus on his home nation of India. He noted sizable pricing discrepancies between human and analog insulin ($3.2-$4.6 vs. $7.5-$10.6 for a 300-unit cartridge) and between glargine manufactured by an Indian company (we assume this is Biocon) and that manufactured by Sanofi ($32 vs. $53.5 for a 1000-unit supply), as reflected in July 2011 IMS data. Staggering price differentials occur throughout the developing world: according to the Fourth Edition of the IDF Diabetes Atlas (2009), the price of insulin in Zambia was $7.4 from a central medical store but as high as $17 from the private sector.

Questions and Answers

Dr. Amanda Adler (Addenbrookes Hospital, Cambridge, UK): Do you think it’s a problem when industry supports patient groups? Is there a conflict of interest?

Dr. Nigam: I wouldn’t say there’s a conflict of interest. But as I said, there is the insulin dilemma – the price discrepancy between developed and developing countries – and the insulin paradox – that though there is not too much evidence analogs are better than human or animal insulins, the companies push this and we as doctors believe it. Industry has played is part by supporting Insulin for Life (an Australia-based charity that collects donated insulin supplies and distributes them in the developing world – Ed.). But I think they also need to get prices in developing countries very different from their current levels. Hopefully the UN Summit on Non-Communicable Diseases will help governments to realize the need and in turn maybe force companies to take action.

Dr. Marja-Ritta Taskinen (Helsinki University, Helsinki, Finland): I still must say that in some European countries, there was almost no attention paid to the Summit. I think this is part of the problem. We are writing guidelines to improve care, but there is no option to follow these guidelines for 80-90% of the people suffering from diabetes.

Dr. Ragnar Hanas (Uddevalla Hospital, Uddevalla, Sweden): Can you address the issue of insulin actually getting to patients without the cost increasing along the way? Even if insulin is donated, it sometimes still winds up being sold to patients.

Dr. Nigam: There are kickbacks, transportation problems, and the fact that expiry dates often pass by the time insulin gets to patients, meaning that they must buy a replacement that they cannot afford.

Dr. Hanas: From Sweden we sent insulin by individual post to Baltic countries, because otherwise it would disappear along the way.


Corporate Symposium: The Impact of Socio-Economic Determinants: Can We Do Better for People with Diabetes in Emerging Countries? (Sponsored by Sanofi)

Partnerships for Improving Type 1 Diabetes in China: The 3-C Study

Linong Ji, MD (Peking University, Peking, China), Jean Claude Mbanya, MD, PhD (University of Yaoundé, Yaoundé, Cameroon), and Ricardo Perfetti, MD, PhD (Sanofi, Paris, France)

Drs. Ji, Perfetti, and Mbanya jointly described the 3-C study, a collaboration of the IDF, Sanofi, and the Chinese Diabetes Society to gather data on the coverage and cost of type 1 diabetes in China. Early results from patients surveyed in Beijing (more urban) and Shantou (more rural) indicated type 1 patients were frequent and intensive users of medial care, with the majority of patients on premixed insulin – in rural areas patients averaged two injections per day, while in urban areas patients averaged four or more injections. However, a number of findings pointed towards suboptimal care: 30% of patients presented with signs of diabetic ketoacidosis as diagnosis, the majority showed no evidence of A1c, lipid, or blood pressure measurements in the last 24 months, and only 46% of patients received any form of diabetes education in the past 12 months. Notably, the total cost of diabetes care out-of-pocket accounted for 29.1% of the patient’s household annual income – likely a large barrier to access to care.

Questions and Answers

Q: Are you sure the increasing age in children with type 1 in China is not due to contamination of early type 2 diabetic children?

Dr. Mbanya: We recently showed there was not contamination because the tests were taken to diagnose type 1 correctly based on the ISPAD guidelines.


7. Diabetes Prevention and Complications

Oral Presentations: Primary and Secondary Prevention of Diabetes and its Complications

Screening for Impaired Glucose Tolerance and Type 2 Diabetes Using Risk Factors Association

YC Woo, PhD (University of Hong Kong, Hong Kong, China)

Dr. Woo discussed a five-year predictive study, in a Chinese cohort, that looked at the role of obesity- related serum biomarkers in predicting the development of type 2 diabetes. Increases in pro- inflammatory markers leptin, TNF-α, and C-reactive protein are elevated in obese individuals and mark an increased risk for developing type 2 diabetes. Adipocyte fatty acid binding protein (A-FABP) has also shown to be predictive of development of type 2 diabetes in a Chinese cohort. The anti- inflammatory marker adiponectin has been shown to be decreased in obese individuals; it is also thought to be insulin-sensitizing and to protect from type 2 diabetes. The question is whether measuring these markers in patients adds value to predict diabetes risk. Of 1,315 non-diabetic subjects, 75 developed diabetes in a median of 5.4 years, and as expected they had higher conventional risk factors including BMI, waist circumference, blood pressure, lipids, and so on. They also had significantly lower adiponectin and significantly higher TNF-α, leptin, CRP, and A-FABP. After further statistical analysis and adjusting for other factors, only adiponectin was both an independent risk factor and enhanced the predictive value of a basic risk assessment model. It may thus one day be useful to add adiponectin measurements to risk assessment models.

Questions and Answers

Q: How expensive is it to measure an adiponectin level in real life? Age and the other factors are cheap. How does it compare to an oral glucose tolerance test (OGTT)?

Dr. Woo: Checking adiponectin by itself is not as good as OGTT. You would combine it with clinical variables, and if you add it to the clinical risk assessment model, it will enhance it and render it equal to a two hour OGTT. OGTT is good but it is not very convenient or well tolerated by patients. For future development of risk assessment models, if adiponectin assays get cheaper and more available, you could consider including it.

Q: Out of the five markers, only one was left. Do you think there is co-linearity of them or was it a matter of study power?

Dr. Woo: I think both. The number of subjects who progressed to diabetes was only 75 out of over 1,000. But reports do suggest that there is co-linearity between markers.

Q: Is adiponectin just a marker, or could it play a causal role in diabetes?

Dr. Woo: I think it is more than a bystander. Evidence suggests it does play a partial causal role.


Oral Presentations: Getting to the Heart of the Matter in Diabetic Cardiovascular Disease

Applicability of the Advance Cardiovascular Absolute Risk Equation to a Multinational Cohort with Diabetes: An External Validation Study

Andre-Pascal Kengne, MD, PhD (University of Cape Town, Cape Town, South Africa)

Dr. Kengne presented the results of a study investigating the applicability of the ADVANCE cardiovascular absolute risk equation to an external population of multinational patients. Using the patients in the ADVANCE trial, Dr. Kengne previously developed a model for cardiovascular risk showing strong predictability for events at 4.5 years. In this study, he applied the model to eligible patients in the DIABHYCAR trial (n=1,836), a clinical trial investigating the effects of ramipril in type 2 diabetes. Results suggested discrimination (ability to distinguish high-risk subjects from low-risk subjects; area under the curve 0.69) similar to that observed in internal validation, with the model identifying 39% of the DIABHYCAR participants in whom 66% of outcomes occurred. With external validation, Dr. Kengne indicated the risk calculator had been made publicly available for use in clinical practice (

Questions and Answers

Q: Does the ADVANCE risk engine consider therapeutic interventions such as blood pressure and lipid lowering in its value?

Dr. Kengne: It is based on treated hypertension, so we use that in the model.

Q: Does the population in South Africa equal that in Australia in the risk engine?

Dr. Kengne: There is reason to think the coefficient predicts risk in a similar way.

Q: When you use a risk score in a clinical setting, what is your aim? Are you aiming at intensifying risk modification in high-risk patients or relaxing in less significant risk patients?

Dr. Kengne: It has essentially two applications. The fist is to assist physicians in decision-making, and the second is to inform the patient on their risk. But it is going to be really context specific. Treatment may also have side effects, and you don’t want to expose people to excess side effects while maximizing benefit.


Symposium: Cardiovascular Risk and Diabetes Medications – The Ups and Downs

Incretin Therapies

Daniel Drucker, MD (University of Toronto, Toronto, Canada)

Dr. Drucker’s presentation on the cardiovascular risk of the incretin therapies was cancelled due to conflict with Dubai’s stance on Israel. “I believe that it is important to build bridges to new parts of the world, but they must be solidly constructed on a foundation of honesty, sincerity, and transparency,” said Dr. Drucker. We were extremely disappointed not to hear him speak at this meeting.


Clifford Bailey, PhD (Aston University, Birmingham, United Kingdom)

Dr. Bailey discussed the microvascular and macrovascular effects of metformin removed of its metabolic effects. Noting the cardiovascular benefit first observed in the UKPDS trial, Dr. Bailey reviewed a number of underlying actions with metformin that possibly contributed to its cardioprotective effects, including anti-atherogenic and anti-thrombotic properties. In addition to addressing many of the components of metabolic syndrome, he also suggested metformin has a variety of direct effects on the heart, improving function and viability.

  • The UKPDS trial first suggested a cardiovascular benefit with metformin. Dr. Bailey reviewed that metformin-treated patients showed prolonged survival as well as a considerable decrease in myocardial infarctions. However, he cautioned that the first evidence of benefit did not appear until after nine years of follow-up – relevant to the cardiovascular outcomes trials being conducted with novel therapies today.
  • Studies indicate that metformin has anti-atherogenic properties. For instance, Dr. Bailey noted that metformin reduces aortic lipid accumulation in animal models and inhibits monocyte adhesion to the endothelium in in vitro studies. Clinically, metformin has also been shown to increase post-ischemic blood flow in non-diabetic patients (though Dr. Bailey warned that the drug should not be used acutely, given the risk of lactic acidosis).
  • Metformin has been shown to have anti-thrombotic properties as well. Studies suggest metformin reduces platelet aggregation and improves hemodynamics (reduces blood viscosity/improves blood filterability), as well as producing thinner fibers and smaller pores in fibrin clots, easing degradation.
  • In addition to addressing many of the components of metabolic syndrome, metformin may also have direct effects on the heart. Dr. Bailey suggested metformin decreases insulin resistance, hyperinsulinemia, and impaired glucose tolerance while also potentially promoting weight loss, all components of metabolic syndrome. Additionally, the drug has been shown to reduce infarct size and improve the ejection fraction in animal models, as well as improving the viability of cardiomyocytes in in vitro studies.

Questions and Answers

Q: Can you comment on the combination of metformin and sulfonylureas and the CV risk?

Dr. Bailey: I don’t know whether it was a statistical aberration or not in the UKPDS, but when this study was undertaken if you look in that subgroup it was compared to those on sulfonylurea alone. Interestingly, at that stage in the study, the sulfonylurea group had a particularly low rate of events. Therefore, it made metformin with sulfonylurea look worse. But if you look at metformin with sulfonylurea against conventional treatment, you will see not an adverse situation. What happened with the passage of time? We do know that in the follow-up we can see quite clearly that those who had metformin at any time in that study had fewer cardiac events.

Q: Can you comment on those patients who were on different diabetes medications in the post-infarct period?

Dr. Bailey: I’d like to know the answer to that, but we don’t really know. It’s not possible to know the extent of any infarct. But as far as I can see it’s certainly not worse.

Q: Could you elaborate on metformin’s effect on the lipid profile?

Dr. Bailey: Most people are concerned about the diarrhea possibilities. This appears associated with suppression of the rate of some bile salts. There’s likely to be a bile salt effect, which could be of some value in hypercholesterolemic individuals. As far as triglycerides, if they are raised metformin can bring them down.

Q: Nowadays our main focus is the effects on CV safety. Are there any trials looking at this aspect in combination with DPP-4 or GLP-1?

Dr. Bailey: All novel drugs have undertaken meta-analyses of phase 3 trials. Adding on to metformin in all of these trials has not produced a negative signal. That’s not to say it’s a positive one, but it’s not a negative one.

Q: What should be the maximum ideal dose of metformin?

Dr. Bailey: The maximum dose is likely to be 2250 mg in some countries and 3 g in others. The only adverse effect of too much is on the GI tract; I’m not aware of any dose-related other significant adverse events unless of course it used when the renal function is affected.

Q: The anti-thrombotic effects should be acute. Why does it take so long to appear?

Dr. Bailey: It isn’t actually a disconnect here. If you look at metformin compared against null treatment, then you see the separation. The point I was making is the comparison with other treatments – in the current designs, you can’t leave patients inadequately treated. It takes a good number of years to see the arms separate out.



Agostino Consoli, MD (University of Chieti, Chieti, Italy)

Dr. Consoli took on the challenging task of discussing the cardiovascular risk associated with the thiazolidinediones. Reviewing the history of the class, he noted that early results suggested the thiazolidinediones would have cardiovascular benefit, both in preclinical and clinical studies. But then came the Nissen meta-analysis – Dr. Consoli highlighted the “meta-analytic frenzy” of more than 13 meta-analyses with rosiglitazone published in the two years following the negative results, noting that all except one indicated a hazard ratio greater than one, which he took as a strong suggestion of increase risk. In contrast, in the several meta-analyses of pioglitazone, none showed a hazard ratio above one – he attributed this to different effects on lipid metabolism.

  • Early studies suggested the thiazolidinediones would have cardiovascular benefit. Results indicated benefits to proxies of CV risk, such as carotid intima-media thickness and carotid restenosis. Clinical trials also suggested benefit – the PERISCOPE trial showed improvements in atheroma changes in contrast with glimepiride, and the PROactive study showed a 16% risk reduction in a composite of death from any cause, nonfatal myocardial infarction, or stroke.
  • Then came the “cold shower,” as Dr. Consoli called it – the Nissen meta-analysis. Dr. Consoli highlighted the “meta-analytic frenzy” of more than 13 meta-analyses with rosiglitazone published in the two years following these negative results – while studies showed results of varying significance, all except one indicated a hazard ratio greater than one, which he took as a strong suggestion of increase risk. In contrast, in the several meta-analyses of pioglitazone, none showed a hazard ratio above one.
  • Pioglitazone and rosiglitazone have different effects on lipid metabolism, likely causing this disparity. Dr. Consoli noted that the effects of both drugs on VLDL and LDL particle size is different; the lipid profile with pioglitazone shows increases in HDL and decreases in non-HDL cholesterol versus rosiglitazone as well.

Questions and Answers

Q: Probably one of the things that would help in assessing the risk and benefit is deciding what is the clinical profile of subjects that would most benefit. Could you tell us something about that?

Dr. Consoli: I agree with you. As a matter of fact, I would summarize saying a young person with a good heart – they would benefit the most. On the other hand, some studies show those doing the best are those with secondary prevention. An old frail lady with bad bones would be a poor candidate. We don’t have precise guidelines – I believe it goes back to how we clinically judge a patient.

Q: Regarding heart failure, the glitazones show an increase. What are the criteria to diagnose heart failure – is it the increase in fluid retention?

Dr. Consoli: In the PROactive study, they were adjudicating the heart failure as a symptomatic disease of getting to the hospital. This was increased in the PROactive study. If you look at an observational study, pioglitazone seems to do better than rosiglitazone. The thing that should be kept in mind is the mortality rate did not increase in PROactive.

Q: When we’re looking at applying CV risk I try to apply this to a person sitting in front of me. If I have the choice of pioglitazone or GLP-1, what do you choose?

Dr. Consoli: Why not use both?



Baptist Gallwitz, MD (Eberhard Karls University, Tuebingen, Germany)

Dr. Gallwitz reviewed the existing data on the cardiovascular risk associated with the sulfonylureas. While the sulfonylureas have been in use for decades, only two hard cardiovascular endpoint trials have been performed, to mixed results – the UDGP indicated an increase in CV mortality with tolbutamide, and the UKPDS showed no increase in mortality, though combination with metformin did produce a benefit. Dr. Gallwitz presented a number of possible mechanisms for increased mortality, including the enlargement of infarct size due to inhibition of ischemic preconditioning, QTc prolongation during hypoglycemia, increased rate of neoplasia due to hyperinsulinemia, and enhancement of atherosclerosis due to increases in visceral and liver fat. While we have certainly heard an array of negative opinions surrounding this class, we do note that Dr. Rury Holman (University of Oxford, Oxford, United Kingdom) commented in the Q&A that decisive evidence on the cardiovascular risk of the sulfonylureas was still lacking.

Questions and Answers

Q: Should we be generalizing when there is great heterogeneity amongst the sulfonylureas?

Dr. Gallwitz: Well, we certainly have to differentiate between the various sulfonylureas. This is also why I showed you the retrospective analysis where glipizide was the sulfonylurea with best results in terms of CV risk. What they all have in common is the glucose-independent way of action, so there’s potential hypoglycemia risk. This is why we should choose sulfonylureas with short duration of action in clinical practice.

Q: How much more data will be necessary before regulatory authorities ban sulfonylureas forever?

Dr. Gallwitz: We do need hard endpoint studies for novel drugs to have more arguments in order to slowly get away from the older drugs. The sulfonylureas have very few endpoints after 50 years of being around. I think the authorities will come up with a new decision once we have better endpoint studies with novel drugs.

Comment: There have been two endpoint studies, and they show no increase in events.

Dr. Gallwitz: I agree that the data is out. I do think there will be a new discussion with the endpoint studies coming out with novel therapies.

Dr. Rury Holman (University of Oxford, Oxford, United Kingdom): I’d like to agree with the last speaker. It was an unbalanced review. There were only two endpoint trials. So there was no suggestion of increased mortality with either the first or second generation. Any comparison with metformin is going to come out negative given the CV benefit of metformin. The observational studies you showed are also flawed. The studies coming out are also not against a comparative agent, so I don’t think we can hope to learn more about the sulfonylureas from that. But the sulfonylureas are a good drug, and you may have a biased presentation and scared people unnecessarily.

Dr. Gallwitz: You may also note that there were attempts in the past to set up trials with sulfonylureas, and they have not been set up.


Symposium: Late-Breaking Clinical Trials

New Results from Advance

John Chalmers, MD PhD (The George Institute for Global Health, Sydney, Australia)

Dr. Chalmers presented several small sub-analyses of ADVANCE data. He found that there was a point around 6.5-7% A1c where higher A1c values led to worse outcomes but where lower values did not offer any more improvement. He also looked more closely at renal outcomes, and presented a new risk assessment score for predicting five-year risk of diabetic nephropathy.

  • Higher A1cs lead to worse outcomes only above a certain A1c threshold: 6.5-7% for macrovascular events and death and 6-6.5% for microvascular events. These results are in press in Diabetologia (Zoungas et al.). When you plot outcome vs. A1c you can see an inflection point at these values. Below them, there is no statistical association between outcomes and A1c, and no conclusion can be made about the effect of lowering A1c. This is true for both the intensive and standard treatment arms. After the inflection point, averaging standard and intensive therapy, for every increase in A1c by 1%, you have a 38% higher risk of macrovascular events, 38% higher risk of all-cause death, and 40% higher risk of microvascular events. This lower limit of 6.5-7% for macrovascular outcomes and death is consistent with ACCORD, which showed no cardiovascular and mortality benefits of treating to <6.5%.
  • There were no significant differences among people of different regions/ethnicities in ADVANCE. This is in press in Diabetes Care (Woodward et al.). This was despite marked differences in health care systems, clinical practice, and use of drugs. This suggests that treatments used in ADVANCE can be safely recommended for all ethnicities.
  • Soon-to-be published data suggests intensive therapy reduced progression to end- stage renal disease by 65% (ESRD; defined as dialysis or transplant). Renal death tends toward improvement but is not significant due to small numbers. Dr. Chalmers’ group also usedADVANCE data to build a risk assessment score that can predict a patient’s five-year risk of developing new onset microalbuminuria and five-year risk of developing a major renal event. This is meant to be used in clinical practice. The most important components are eGFR (estimated glomerular filtration rate) and ACR (albumin creatinine ratio). A separate study showed that the addition of two biomarkers to the assessment enhances accuracy - TGFβ1 and BMP7.

Questions and Answers

Q: What you are doing with renal data is unique and we have never done anything like that. We never separate end-stage renal disease and death. I think you should stick to the rules…

Dr. Chalmers: My conclusion was that we need larger numbers, but I think it shows that’s there’s a real need to examine it in a larger study.

Q: The risk score is beautiful because it allows you to predict individual risk. Regarding, the inflection points for A1c/outcome data. The steeper rise of the intensive group – does that show there’s a benefit of being in the standard group independent of A1c?

Dr. Chalmers: The intensive group had the steeper slope. The difference is probably just that the intensive group had less variation since everyone was aiming at a target. The larger standard deviation of the standard curve would cause regression dilution that we did not correct for and would make a shallower curve.


New Results from ACCORD - The Mind Study

Lenore Launer, PhD (National Institutes of Health, Bethesda, MD)

Dr. Launer looked at “global cognitive decline,” a pre-dementia loss of cognitive function, in people with type 2 diabetes in the Memory in Diabetes (MIND) sub-study in ACCORD. She found that there was no difference in cognitive function between patients in the intensive and standard arms, but that the intensive treatment arm had less decline in brain volume than the standard. She said that although intensive control might prevent some brain atrophy in type 2 diabetes, it nevertheless does not support the use of an intensive glucose control strategy given the other outcomes of ACCORD. Over time, however, the cognitive differences between treatment groups will emerge, she things. It is known that type 2 diabetes increases the risk for dementia; it also seems that brain volume loss in this population is occurring about 15 years before it does in the general population.

Questions and Answers

Q: Were there enough hypo events to look at the effect of severe hypoglycemia on cognitive function?

Dr. Launer: We are in the process of doing these analyses now.

Q: And what about the effects of hypertension and lipid control?

Dr. Launer: Yes, we’re working on those results right now and they will be available soon.

Q: Some medications may also be beneficial and some harmful regarding cognitive decline; could that be a confounder?

Dr. Launer: Yes. We’re looking at rosiglitazone now, but based on how ACCORD was designed, as you may know, it is hard to pick out one drug vs. another.


Symposium: Neural Complications of Diabetes

Detecting Diabetic Neuropathy: Can We Do Better?

Rayaz Malik, MBChB, PhD (Manchester University, Manchester, United Kingdom)

Diabetic peripheral neuropathy (DPN) is often thought to be near-impossible to treat, but Dr. Malik suggested that we may simply need better diagnostic methods. On that note he described corneal confocal microscopy (CCM), a two-minute eye-imaging test that correlates quite well with skin biopsies (a sensitive but invasive DPN diagnostic). (Researchers from the University of Manchester have commercialized a software package that automatically analyzes CCM results, called CCMetrics; during Q&A Dr. Malik noted that the CCM machine costs roughly 45,000 pounds [~$30,000].) The tests have been shown sensitive enough to detect neuropathy even in impaired glucose tolerance (Petropoulos et al., Neurodiab 2011) and even to detect therapeutic effect – it seems that unmyelinated nerve fibers can regenerate in response to diabetes interventions, even if the myelinated nerve fibers cannot (Tavakoli et al., Diabet Med 2011). We are glad to see progress on diagnosing DPN from the Manchester group and from companies like US-based NeuroMetrix, and we hope that up-front investments in new technologies can start to translate into fewer ulcers and amputations. Given that complications are particularly rampant in the developing world, we also look forward to research on new methods that merge high sensitivity and lower cost.

  • Dr. Malik explained that traditional tests (e.g., 10 g monofilament, vibration perception threshold, electromyography) are not sensitive enough to detect early- stage neuropathy, in part because they measure only myelinated nerve fibers – not the small, unmyelinated fibers that are the first to be damaged. Analyses of ACCORD, ADVANCE, VADT, and Steno 2 suggest that interventions do not address peripheral neuropathy, causing many researchers and companies to lose hope in the field. He quoted an anonymous head of late complications at a major pharmaceutical company: “Diabetic neuropathy is a nightmare. There are no clear endpoints, so we focus on nephropathy.” Dr. Malik said that DPN researchers may have “scored an ‘own-goal’” by conducting high-profile studies using flawed tools; he said the results might have been different with the use of better diagnostics.

Questions and Answers

Q: Is there any false positivity or false negativity?

Dr. Malik: It’s not a perfect tool, but the sensitivity and sensitivity are both in the 90% region, which EMG doesn’t get anywhere near.

Q: Are other causes of neuropathy reflected in the results?

Dr. Malik: Sometimes we wonder why the test more damage than you’d expect from people’s diabetes state, then we go back to the clinic and find that they also have other reasons for neuropathy – a double hit.

Q: Does this technique require specific training for optometrists?

Dr. Malik: Anyone comfortable with examining eyes can do this. Even I can do it, and I am useless.

Q: What is the cost?

Dr. Malik: The instrument itself costs around 45,000 pounds. And people say, “Oh, that’s terrible.” But an EMG machine costs around 20,000 pounds, and a skin biopsy costs $2,000 dollars every time. There is an initial outlay but massive return.

Q: So what is the mechanism for this nerve loss, and why are the small fibers so sensitive? Also, it looks like they can re-grow normally?

Dr. Malik: with skin biopsy, you see earliest hit is to skin biopsy. They seem v sensitive to diabetes and other abnormal processes. Even after myelinated neurons have given up the ghost, the small unmyelinated fibers seem to have an inherent capacity to regenerate.

Q: Does this study apply to autonomic nerve damage as well as sympathetic?

Dr. Malik: The CCN predominantly detects c-sensory fibers. There are some autonomic components detected, but we can’t differentiate between autonomic and sympathetic with this test.

Q: How long does it take to detect clinical improvement using this test?

Dr. Malik: In the pancreas transplant setting with patients who’d had severe damage, we showed improvement at six months. In a multifactorial intervention study where the changes were less aggressive, it took 24 months.


Symposium: Biological Mismatching and Modernization – Unraveling Multi- Causality

Developmental Epigenetics, Obesity and Diabetes

Paul Zimmet, MD, PhD (Baker IDI Heart and Diabetes Institute, Melbourne, Australia)

Dr. Zimmet discussed the role of epigenetic factors in diabetes risk. As a reminder, epigenetics refers to factors that modify the translation of genes and can change with environment and behavior. Dr. Zimmet described the “mismatch pathway” in diabetes, in which influences acting in early life (such as in the maternal environment) set the trajectory for a phenotype better adapted to a low-energy environment – placing the child at risk when they arrive in a high-energy environment. Given the mother’s lifestyle and behavior could potentially predispose children to increased risk, Dr. Zimmet hoped future prevention measures would begin to address the maternal environment.

  • Initial signs that genes could be modified through environment and behavior were discovered in World War II, in which women exposed to famine in the 2nd and 3rd trimesters of pregnancy delivered children with increased risk of obesity and diabetes as adults. Dr. Zimmet labeled this an example of the “mismatch pathway,” in which influences acting in early life set the trajectory for a phenotype better adapted to a low-energy environment – placing the child at risk when they arrive in a high-energy environment. He noted a study in which retinoid X receptor alpha (a transcription factor) levels were increased in the umbilical cord of low birth weight children who later became obese, suggestive of epigenetic involvement.
  • Given the mother’s lifestyle and behavior could potentially predispose children to increased risk, Dr. Zimmet hoped future prevention measures would begin to address the maternal environment. Animal studies have led researchers to question if the epigenetic state is reversible as well – when undernourished and control rat pups were given saline or leptin following birth, leptin was shown to block all of the consequences of dietary change when rats were fed a high-fat diet later.

Questions and Answers

Q: I wonder whether some of the data you presented could explain in part that we are seeing the onset of non-communicable diseases earlier in low-income countries. What would your recommendation be for these low-income countries?

Dr. Zimmet: Our hope is that epigenetic changes might be reversible. There are likely experts who know more than me. While we do focus on non-communicable diseases as the outcomes of these effects, there are certainly strong effects on the brain and also susceptibility to infectious diseases. I think we need to change the paradigm of how we think about these diseases. Had the babies of the mothers during the Dutch famine been studied at the time, we might better understand this whole mechanism.

Q: Do we have any leads on dietary advice to give to women during pregnancy?

Dr. Zimmet: I think sensible nutrition advice and proper education of mothers about alcohol, smoking, and diet would go a long way to solving the problem.

Q: We see trends to be thin as possible during pregnancy. Would that have an effect on diabetes risk?

Dr. Zimmet: I think that’s a good comment. I’ll leave it at that.


Migration and Globalization

Anne Sumner, MD, FAHA (NIDDK, Bethesda, MD)

Using an NIDDK cohort of African immigrants living in the Washington, DC area, Dr. Sumner described the impact of migration and globalization on diabetes risk. Comparing immigrants from the cohort to African Americans, Dr. Sumner noted at a lower BMI immigrants show higher systolic blood pressure, diastolic blood pressure, visceral adipose tissue, and glucose tolerance levels – she posited this was due to the rapid weight gain observed in the population after migration to the United States. While migration to this obesogenic environment clearly impacts risk, Dr. Sumner noted that globalization is exporting risk as well, with increasing sedentary activity and availability of fast food abroad. Interestingly, she concluded with a discussion of how globalization of screening tests developed in the West is also impacting healthcare abroad. As an example, she noted that while Africans show increased blood pressure and visceral adipose tissue compared to African Americans, the prevalence of metabolic syndrome in both groups remains the same – this is due to low fasting triglyceride levels in Africans, which leads to fewer diagnoses based on Western definitions of metabolic syndrome.

Questions and Answers

Q: I was really struck by the Africans moving to the US having gained 3-5 lbs per year. At the NIH we have a lot of foreigners coming in, and everyone will tell you they start gaining weight. Have you been able to measure acculturation and assimilation of those Africans coming to the US and their relationship?

Dr. Sumner: We’re beginning to collect the data to relate how long people have been here to their weight. We’re actually going to start a study at the NIH to compare workers at the NIH to get a handle of some of those issues.

Comment: We have a big migration of people from Ethiopia to Israel. They had no diabetes before; now 10% of them have diabetes. This could be change in diet and also an increase in lifespan.

Dr. Sumner: I had the privilege to be in Tel Aviv to address some of these issues last year. When the Ethiopians came in the 1980s, they were malnourished when they came, and now they have a quick change in diabetes rates.

Q: You mentioned that people who moved to the States gained weight quickly. Is that due to different foods or the amount of food?

Dr. Sumner: We think it has to do with people preparing food themselves or outside. Men have more issues, and they tend to eat out versus women.

Q: What can we do to put pressure on big multinational corporations to make healthier food products?

Dr. Sumner: It’s an economic engine. I know in the United States, McDonald’s has changed greatly in its offerings. If you can create a market in which people change their desires, the restaurants will follow the people. People are becoming more aware of diabetes abroad and its complications, and they are becoming more willing to begin making a change.


Debate: Early Intervention to Prevent Diabetes Must Include Medication


Ralph DeFronzo, MD (University of Health Science Center, San Antonio, TX)

Dr. DeFronzo gave a thorough presentation in support of pharmacologic therapy for diabetes prevention. Against lifestyle intervention, he argued that non-pharmacologic therapy is neither effective nor cost-effective in the long-term. Citing the Diabetes Prevention Program (DPP), he noted that after six years of follow-up, the metformin and lifestyle groups converged at around 2 kg (4 lbs) weight loss. Moreover, 10-year DPP data indicated a $1,500 cost per patient versus placebo with lifestyle intervention compared with a $30 cost-savings with metformin. For pharmacological therapy, he presented a detailed summary slide indicating a relative risk reduction with pharmacotherapy across all prevention trials – in particular, he was excited about the potential for reducing side effects using combination low-dose pioglitazone/metformin, especially as generic pioglitazone becomes available (expected to reach the US market by August 2012). Notably, he concluded with strong support for the incretin therapies in diabetes prevention, citing what he deemed “one of the most dramatic studies ever carried out” (Chang et al., Diabetes2003), in which a single dose of liraglutide was able to normalize insulin sensitivity in type 2 diabetes patients. In the Q&A, he also included the DPP-4 inhibitors (“I’ve suggested to every pharmaceutical company with a DPP-4 that this is where the future is going”) – while benefits to diabetes risk factors have not proven as strong with DPP-4 inhibitors as with GLP-1 agonists, ease of administration is certainly striking with this oral class.

  • Dr. DeFronzo opened with a strong thesis: “The answer is simple – non- pharmacologic therapy does not work on a long-term basis.” Citing the DPP, Dr. DeFronzo noted that while the lifestyle group showed greater weight loss initially, by six years follow-up the metformin and lifestyle groups converged at around 2 kg (4 lbs) weight loss. Additionally, despite successful weight loss, 40-50% of individuals with impaired glucose tolerance progressed to type 2 diabetes in the trial. Dr. DeFronzo pinned difficulty in maintaining weight loss to long-term persistence of hormonal adaptations – in a study of 50 obese patients achieving 13 kg (28 lbs) weight loss during a 10-week intervention, follow-up at 13 months showed physiological changes promoting regain (decreased energy expenditure, leptin, PYY, CCK, and amylin; increased appetite and ghrelin).
  • Effective lifestyle intervention is not cost-effective. Listing the numerous of components of the lifestyle intervention in the DPP, Dr. DeFronzo posited that broader application would not be feasible in the majority of countries. He also showed 10-year follow-up data indicating $1,500 cost per patient versus placebo with lifestyle intervention compared with a $30 cost-savings with metformin. Given that interventions are rarely cost saving, he suggested this strongly supported pharmacological treatment.
  • A wide range of evidence supports the preventative effects of pharmacotherapy, potentially without exorbitant costs and side effects. Presenting a detailed summary slide, Dr. DeFronzo indicated that all prevention trials have shown a relative risk reduction with pharmacotherapy (DPP, Chinese, IDPP for metformin; DPP, TRIPOD, ACT NOW, DREAM for thiazolidinediones). He focused particularly on the ACT NOW trial, highlighting the reduced conversion rate with pioglitazone (2.1% vs. 7.6% with placebo) and improvements in insulin sensitivity and beta cell function. He was also excited about the potential for reducing side effects using combination low-dose pioglitazone/metformin, especially as generic pioglitazone becomes available (expected to reach the US market by August 2012).
  • In the future, incretin therapies should be considered for diabetes prevention. In particular, Dr. DeFronzo extolled the preservative effects of GLP-1 therapy on beta cell function – in “one of the most dramatic studies ever carried out” (Chang et al., Diabetes 2003), he highlighted that a single dose of liraglutide was able to revert insulin sensitivity in type 2 diabetes patients to levels observed in normal glucose tolerance patients. He also cited results from the obesity study of liraglutide (Astrup et al., The Lancet 2009), which indicated an 84-96% reduction in prediabetes and 60% decline in metabolic syndrome, in support of its use in diabetes prevention. Though not mentioned in his presentation, in the Q&A he strongly supported the use of the DPP-4 inhibitors as well, especially given their ease of use versus GLP-1 therapy.



Jaako Tuomilehto, MD, PhD (University of Helsinki, Helsinki, Finland)

In methodical fashion, Dr. Tuomilehto argued against the use of pharmacotherapy for diabetes prevention. While he conceded that pharmacotherapy does reduce the incidence of diabetes, he suggested its effectiveness comes with excessive stipulations – citing the DREAM trial, he noted that while rosiglitazone produced a 60% reduction in diabetes over a five-year period, the observed relative risk (0.39) was similar to that expected (0.44) solely accounting for the drug’s effects on blood glucose. Thus, once treatment is stopped, patients revert in risk, requiring lifelong treatment to maintain preventative effects. With many treatments showing side effects and extensive costs, he suggested this widely limits the applicability of pharmacotherapy.

  • Evidence suggests the impact of pharmacotherapy on diabetes progression was not any greater than expected due to the glucose-lowering effects of the drugs. Citing the DREAM trial, Dr. Tuomilehto noted that while rosiglitazone produced a 60% reduction in diabetes over a five-year period, the observed relative risk (0.39) was similar to that expected (0.44; calculated assuming a 0.6 change in risk for each 1 mmol/l [18 mg/dl] decrease in 2 hr glucose) solely accounting for the drug’s effects on blood glucose.
  • Pharmacotherapy must be given lifelong to maintain preventative effects. According to Dr. Tuomilehto, given the observed benefit of pharmacotherapy is solely due to its glucose lowering effects, once treatment is stopped patients revert in risk – citing the short-livedtroglitazone arm of the DPP, he noted that the incidence of diabetes became approximately the same as that of the placebo group six months after treatment ceased. Likewise, if patients progress to diabetes regardless of therapy, even further pharmacotherapy will be required once treatment fails.
  • Side effects and costs also limit the applicability of pharmacotherapy. Citing the DREAM trial, Dr. Tuomilehto highlighted the increase in congestive heart failure observed with rosiglitazone as well as the €750 price per patient – with no obvious end to therapy. We do note, however, that Dr. DeFronzo argued against the use of rosiglitazone given its adverse event profilegeneric pioglitazone would also reduce costs.


Panel Discussion

Ralph DeFronzo, MD (University of Health Science Center, San Antonio, TX) and Jaako Tuomilehto, MD, PhD (University of Helsinki, Helsinki, Finland)

Questions and Answers

Q: Have you looked at DPP-4 inhibitors?

Dr. DeFronzo: Actually this has not been looked at. I’ll tell you I’ve suggested to every pharmaceutical company with a DPP-4 that this is where the future is going, but the studies have not been carried out. Clearly the GLP-1 analogs produce higher GLP-1 levels, but the ease of administration with DPP-4s is much greater.

Q: Because prediabetic patients can have complications, why not start earlier?

Dr. Tuomilehto: We should start earlier to treat diabetes with pharmacotherapy compared to current guidelines.

Dr. DeFronzo: I would agree with that. Once you’ve been diagnosed with prediabetes, you’ve already got dysfunction – you’re starting too late.

Q: Is there a kind of metabolic memory that sets the body to starvation mode if you start extensive lifestyle intervention mode and when you ease that do you go back to baseline?

Dr. DeFronzo: I tried to show you that that is indeed the case. When you try to lose weight, every single thing works against you. These people are much hungrier than when they started. It’s difficult to lose weight and keep it off. If you can keep it off, that’s a great therapy. But personally I think we need both, because our patients don’t always do as well as they do in Finland.

Comment: I think all the anti-diabetic medications have complications, and we don’t have long-term data on them.

Dr. DeFronzo: I have to disagree. It is not at all clear that pioglitazone causes bladder cancer, and if it does it is infinitesimally small. If you do the numbers, treatment would reduce mortality by 1,600 patients versus the three additional patients with bladder cancer. I think the whole bladder cancer thing has been blown out of proportion. I think with pioglitazone that’s the only drug that’s shown it can affect cardiovascular events. To me it’s a natural intervention. As we go forward, I do think the GLP-1s and the DPP-4s will be considered.

Comment: We need to know if preventing diabetes prevents long-term complications.

Dr. DeFronzo: We don’t have the 10-year study, so I think you have to go on the clinical evidence. If you don’t have the glucose, you don’t get the microvascular complications. I don’t feel comfortable saying that about the macrovascular complications, but I think we have good evidence that a drug like pioglitazone from PROactive can prevent complications. I wish they would start the 10-year study so we could get a definitive answer.

Q: The majority of patients newly diagnosed are Asians. There was a study from Beijing that suggested if you add acarbose to lifestyle you get an additional effect, which was not seen with metformin. Why did you not propose that drug for Asian patients?

Dr. DeFronzo: I did show that acarbose was successful in decreasing the conversion rate. From the pathophysiologic standpoint, however, I have trouble believing acarbose will reverse the progression. I also don’t understand how it could address cardiovascular events. The number of events in that study was so small that it’s hard to accept it as well. I would favor pioglitazone, not rosiglitazone, as it is a bad example with a bad lipid profile. But I think low-dose pioglitazone and low-dose metformin, which both will be generic, make more sense pathophysiologically.

Q: Where do you put bariatric surgery?

Dr. DeFronzo: If your BMI is 40 kg/m2, that’s great. If it’s 30 kg/m2 or 32 kg/m2 I would think it would be a little extreme and not supported.

Q: Why did you not talk about nutritional interventions?

Dr. Tuomilehto: Two years ago the last Cochrane review was done on nutritional dietary therapy, and the authors concluded there were no proper studies. I agree though – the question to ask is why don’t we do these proper studies.


Debate: Primary Versus Secondary Prevention: What Should Be the First Priority?


Simon Griffin, MSc, DM (Medical Research Council Epidemiology Unit, Cambridge, United Kingdom)

Dr. Griffin posited that there is insufficient evidence in support of diabetes screening, arguing instead for primary prevention measures. In contrast to primary prevention, Dr. Griffin indicated that screening has proven neither effective nor cost-effective in clinical trials. Using statistical arguments, he also suggested that population-level measures would be necessary to shift the risk distribution and curb the incidence of diabetes, rather than exclusively focusing on high-risk patients.

  • Screening for diabetes has proven neither effective nor cost-effective. According to Dr. Griffin, modeling data has provided very little support for the cost-effectiveness of diabetes screening. Additionally, in a trial conducted in England, 32 general practices were randomized to provide screening during visits – results indicated no benefit in mortality (adjusted HR 0.99, 95% CI 0.84-1.16).
  • Primary prevention has been shown to be effective in clinical trials. Dr. Griffin cited both the Finnish Diabetes Prevention Study (lifestyle intervention halved the risk of diabetes, with benefit maintained even after the intervention was stopped) and the Da Qing study (long- term follow-up indicated intervention halved the risk of retinopathy; Dr. Griffin suggested the small number of events limited results for CV endpoints) to support this.
  • In light of the striking increase in diabetes prevalence, primary prevention is necessary to shift the distribution of risk. Moderate-risk patients make up the bulk of therisk distribution but also progress to diabetes. Thus, in order to curb the incidence of diabetes, the entire risk distribution must be shifted, rather than focusing exclusively on high-risk patients through screening. This requires population-level primary prevention measures.



KM Venkat Narayan, MD, MSc, MBA (Emory University, Atlanta, GA)

Given the proven efficacy of currently available diabetes treatments, Dr. Narayan countered Dr. Griffin, arguing that screening would help identify patients early on and allow for treatment and the prevention of complications. He opened with a broad review of the efficacy of current diabetes treatments, indicating a decline in mortality rate amongst men with diabetes in addition to decreases in complications. While evidence suggests that lifestyle intervention and some pharmacotherapy are efficacious in people with prediabetes as well, Dr. Narayan noted that effectiveness is lost when interventions are offered to the total population. Thus, he suggested that screening must come first, to identify whom to treat. Policy interventions (such as a soda tax), he concluded, do not yet have the evidential support necessary for broad application.


Panel Discussion

Simon Griffin, MSc, DM (Medical Research Council Epidemiology Unit, Cambridge, United Kingdom) and KM Venkat Narayan, MD, MSc, MBA (Emory University, Atlanta, GA)

Questions and Answers

Q: There is limited data on how the disease develops early on. How can we detect the natural cause of the increase of risk?

Dr. Griffin: The burden of diabetes is largely cardiovascular disease, and glucose is a very weak predictor actually. Of course we can deal with blood pressure and cholesterol and smoking without needing to screen the population for their blood glucose. So we could actually make a big difference using primary prevention strategies without screening for glucose and labeling those people.

Dr. Narayan: Prevention of cardiovascular disease is actually secondary prevention.

Q: Why did both of you avoid talking about vitamin D?

Dr. Narayan: I haven’t seen strong data on the connection between vitamin D and diabetes. Last year there was an Institute of Medicine report in the US raising issues with the benefits of vitamin D.

Dr. Griffin: I have read the observational data on vitamin D, and I find it quite persuasive and am pursuing a trial. I didn’t include it because of the lack of time, and also because there is not any data on the cost-effectiveness of such an intervention.

Q: In one slide I saw the LDL cholesterol kept very low. What level is safe?

Dr. Narayan: The slide showed it was reduced by that amount, not that it was brought to that amount.


Debate: Glucose Variability and Diabetic Complications: Should We Care?


Antonio Ceriello, MD (IDIBAPS, Barcelona, Spain)

Dr. Ceriello argued that glycemic variability is an important clinical challenge to address moving forward. CGM results have shown that even at decent A1c, glucose is highly variable. Fluctuating glucose levels in cell culture experiments induce death more effectively than consistently high glucose levels, probably due to poor adaptation to reactive oxygen species from hyperglycemia. Outcomes studies from the ICU are clear that glucose variability is associated with poor outcomes. Outpatient studies are less common but available. Finally, the fact that glycemic variability would matter just makes more sense. For these reasons, Dr. Ceriello argued that we should start paying attention to this very challenging aspect of blood glucose control.

  • Dr. Ceriello believes evidence exists to suggest that high glycemic variability is harmful. The most convincing data linking glycemic variability with poor outcomes comes from the ICU – in many studies, high variability is associated with negative ICU outcomes. In an outpatient setting, the Verona Diabetes Study showed that long-term variability of FPG is a predictor of cardiovascular mortality. Dr. Ceriello said that the Taichung Diabetes Study, which is currently in press at the American Journal of Medicine, also showed that variation of FPG is a predictor of mortality in type 2 patients. Though in theory HEART2D showed that improving glucose variability did not impact cardiovascular events in type 2 diabetes patients post- myocardial infarction (it actually compared basal insulin vs. prandial insulin treatment), a post- hoc analysis showed that in older patients and in patients with a longer duration of disease, it did (Siegelaar et al., Diabetes Care 2011). Nevertheless, Dr. Ceriello said in the panel discussion that it was a “stupid study.”
  • Glucose variability is dangerous because of the production of reactive oxygen species in hyperglycemia. In cell culture studies, death is induced more frequently in endothelial cells in alternating glucose environments than a consistently high glucose environment. This is because high glucose environments increase intracellular antioxidants, likely as an adaptation to the environment, in a way that fluctuating glucose in the environment cannot. Monnier showed in 2009 that isoprostane excretion, as a marker for oxidative stress, was associated with MAGE (mean amplitude of glycemic excursions, a measure of glycemic variability), in people with type 2 diabetes. This relationship did not hold true in Monnier’s similar study in type 1s.
  • Theoretically, it just makes sense that glycemic variability would be harmful. CGM now allows us to look at what happens in non-diabetic individuals, and we find that glucose remains remarkably stable, generally staying under 140 mg/dl and over 70 mg/dl. It is hard to see how something so complex and yet so strictly regulated would not have great importance for our health, Dr. Ceriello said.

Questions and Answers

Q: What levels of glucose variability will we consider benign and what won’t we? Are there any clinical trials looking at this?

Dr. Ceriello: I hope to get this data in order to win the Nobel Prize. How to measure this is a challenge. Different types of glucose changes might have different impacts – high to less high, normal to hypo, high to hypo, and so on. Unfortunately, we don’t know which are important. I encourage you to look at what happens in normal people – this is how we know what should be happening.

Q: Do you believe there is a difference in the effect glucose variability has on micro- and macrovascular endpoints?

Dr. Ceriello: There is an impact on retinopathy that Dr. Kilpatrick found – so he can’t deny it in his talk. I would say that there is more impact on the microvascular. This is my opinion but the data is still not there.

Q: What is the methodology of measuring glucose variability? How often would you have to measure blood glucose in daily practice, if this is found to be important? And what are normal values for this index? We need prospective trials in order to consider using this index in practice.

Dr. Ceriello: Nobody knows. Monnier suggests MAGE is a better marker than SD, and I believe him, as he’s also a mathematician. You need CGM to see variability, not just a few points during the day. Trials for this would have to be massive.

Dr. Irl Hirsch (University of Washington, Seattle, Washington): You have mostly focused on postprandial hyperglycemia. I’d argue that in talking about glycemic variability we have to talk about both highs and lows. I think that hypoglycemia needs to be included in this definition, because hypoglycemia itself predicted both microvascular and macrovascular complications in ADVANCE. We need to talk about ups and downs, not just ups.

Dr. Ceriello: We still don’t have enough basic research looking at hypoglycemia. I am convinced the mechanism would not be so different, but there is just no evidence available.



Eric Kilpatrick, MD (Hull York Medical School, York, United Kingdom)

Dr. Kilpatrick took us through the story of glycemic variability, focusing on analyses of DCCT, UKPDS, and HEART2D and showing that glycemic variability in these studies did not predict outcome. We worry about these conclusions given these studies weren’t constructed to look at glycemic variability per se. Of course, seven point glucose measurements throughout the day in DCCT is not the same as data obtained from CGM. Additionally, HEART2D also used a roundabout, possibly ineffective way of approximating glycemic variability and is generally considered a flawed study. Overall, Dr. Kilpatrick’s point was that by far and away the most important parameter for predicting complications is mean blood glucose. This is in fact heartening, he said, because you can tell your patients with good A1c not to worry that their fluctuations will lead to complications. To close the debate, we took note of a compromising point: even if it is all about mean blood glucose, it might be that there is no way to reduce mean blood glucose below a certain point without targeting variability.

  • In the DCCT, the only predictor of microvascular outcomes was mean blood glucose. Based on seven-point blood glucose measurements in the DCCT, two patients with the same mean blood glucose over time will have the same A1c, no matter the glucose variability. The fact that intensive treatment appeared to produce better outcomes than conventional treatment even at the same A1c really threw people off and got them thinking about the role of glycemic variability, until it was discovered that the difference was an artifact from a statistical error. Both mean blood glucose and glycemic variability (as measured by standard deviation) seem to predict microvascular outcomes until you do a multivariate analysis, after which it is clear only mean blood glucose is predictive. In other words, at any given mean blood glucose, variability does not add predictive value. The reason glycemic variability dropped out as a predictor was because it was actually correlated with mean blood glucose. Although it was a messy correlation, patients with higher mean blood glucose levels tended to have higher glycemic variability.
  • In UKPDS, there were no differences in diabetes-related endpoints in patients treated with insulin and those treated with SFUs. These two groups must surely have had different patterns of glycemic variability, Dr. Kilpatrick said, with insulin causing more variability than SFUs. He therefore inferred that glycemic variability does not play a strong role in microvascular outcomes – Dr. Irl Hirsch warned against such an assumption in Q&A. (Editor’s note: We note that until recently there was no CGM so we caution against interpreting GV in any “real” way from DCCT and UKPDS. )
  • People with postprandial hyperglycemia, which has been connected to macrovascular disease, probably just have hyperglycemia. As postprandial glucose increases, presumably so does A1c and mean blood glucose. The HEART2D study showed that the same outcomes were seen in patients on basal insulin and prandial insulin, despite prandial presumably doing a better job of controlling glycemic variability. Follow up studies show that glycemic variability was indeed higher in the basal group as expected, yet there was no correlation with cardiovascular rate.

Questions and Answers

Q: Are there data to support that there may be individuals or groups particularly susceptible to glycemic variability independent of mean blood glucose?

Dr. Kilpatrick: At a given mean blood glucose, an individual may be very fluctuating or be very stable. But when we look at everyone taken together, we have found no particular relationship that has emerged.

Q: What’s your view about people who have a normal OGTT but very high level at one hour? Could this be related to glycemic variability?

Dr. Kilpatrick: Right, two hours is not actually the peak; around one hour and 19 minutes is. I’m not sure at the moment that a one-hour measurement gives much more value. People who tend to be higher at one hour tend to be higher at two hours as well. Is it more sensitive? Maybe, and maybe we’re missing a trick there.

Q: Have you looked into other risk factors that are influenced by glycemic variability?

Dr. Kilpatrick: High mean blood glucose correlates with glycemic variability. Is there an additional effect of glycemic variability beyond that? We don’t have that, which is reassuring to me. I think it’s quite heartening to say, “your risk of severe hypoglycemia may be high, but your risk of microvascular and macrovascular complications may not be higher than your average should suggest.” (Editors note – From a patient or HCP perspective, we do not think this is a “heartening statement”.)

Dr. Irl Hirsch: I was surprised about your review of UKPDS data. You were making speculations about glycemic variability in different groups when there were no measurements done. And I’m surprised you’d make comments about it since we don’t know.

Dr. Kilpatrick: That’s a fair comment. In clinical practice, people on insulin have much higher variability than on SFUs. Bur you’re right – it was an act of faith.

Dr. Ceriello: To respond: please forget the HEART2D study. It was a stupid study. In DCCT, we are talking about people who were followed for ten years; we were probably able to catch some variability and probably lost some too. And let me say something naïve as well. This data in non-diabetic people is normal glucose levels continuously. What is the reasonable explanation? I can’t believe that whatever God you believe in would create something so difficult to maintain just by chance.

Dr. Kilpatrick. On this relationship between mean blood glucose and glycemic variability – we can look at it another way: maybe the only way we can decrease mean blood glucose is to decrease glycemic variability.


8. Government and Policy

Symposium: Acting on a Larger Stage 1 – Advocacy and Reform in a Post-UN- Summit World

Working with Governments

Sir George Alleyne, MD (Johns Hopkins University, Baltimore, MD)

Dr. Alleyne, who was appointed in 2003 as the United Nations Secretary-General’s Special Envoy for HIV/AIDS in the Caribbean, was truly moving in his presentation on the role of activism and advocacy in the non-communicable diseases (NCD) movement. He took cues from two great social movements of the last century, the feminist movement and the civil rights movement, and proposed we see and frame the NCD movement in the same way: in terms of social injustice. The UN Summit was invaluable as it represented a global commitment to this cause. However, its success still requires we hold political organizations to the commitments (especially the financial commitments) that they have made.

  • Social injustice: this is how the great social movements of the last century have framed their cause, and this is the way the NCD movement should frame its. A social movement is key, as these can help governments make difficult decisions. Despite organizational strength, good communication networks, and leaders, the NCD movement does not have the momentum we would hope for. We tend to frame the issue in economic or intellectual terms, but we need to “frame it at the level of the heart…social injustice…the rift between the rich and the poor,”Said Dr. Alleyne. “Amputations in my country Barbados are three times greater than they are in the US. That is not right.” “Frame it in terms of amputations, blindness, impotence…these things make the disease real.” “We are kidding ourselves if we believe the private sector does not have a legitimate role in matters of the state.” We need to work within the contemporary pluralist state, which consists of the government, the private sector, and civil society (such as NGOs).
  • We know global political priority is present when: 1) international and national political leaders publicly and privately express sustained concern, 2) the organizations and political systems they lead enact policies to address the problem, and 3) these organizations and political systems devote levels of resources commensurate to the problem. We need more money, Dr. Alleyne said, and it is unrealistic to expect some new global fund will emerge. The developed world made a commitment to support the development agendas of low and middle-income countries and we must hold them to it. In doing so, we must never assume governments have the information they need.
  • Our governments should be involved in four ways: promotion, regulation, legislation, and taxation. This paradigm is useful because each goal requires one or some combination of these, and knowing what that is may help us direct our efforts. For example, on a national level, early diagnosis and treatment goals, consisting of screening and increased access to medicines, requires regulation and legislation; primary and secondary prevention require all four aspects. On a global level, global partnerships require promotion and regulation.

Questions and Answers

Q: NCDs are not very good at appealing to the heart compared with diseases like AIDS. Cancer does generate that terror, but the others don’t seem to…

Dr. Alleyne: I only mention diabetes because we are here. But there are other great problems – allowing people to die without opiates; cancers; strokes – I could make you weep with the stories I have. NCDs are one of our greatest injustices.

Q: Can you elaborate on what you said about legislation required in terms of screening for diabetes.

Dr. Alleyne: I hope I didn’t imply that we should legislate screening for diabetes because I don’t that is going to happen. But regulation: when people come to clinics, they would be screened. Regulation is appropriate but I am not advocating for legislation.

Q: How long do you think it will be from now until there is action from the people we are trying to convince?

Dr. Alleyne: The speed of action depends on the intensity of the input. If we don’t do that much it could take forever. If we are optimally aggressive and urge action, we can see it in a relatively short term. We’ll know it’s happened when we start to see increased resources being applied. We already have indication of some countries increasing allocation to NCDs. In terms of the impact on morbidity and mortality…that is going to take longer; I couldn’t say how long.


The NGO Perspetive

Ann Keeling, MA (CEO of IDF, USA)

Dr. Keeling stepped in for the WHO’s Dr. Ala Alwan, who could not be here. She covered all aspects of the NCD movement, from how to frame the debate, to thoughts about what the Summit did and did not accomplish, timeline, and partnerships. The UN Summit is a very important beginning, but of course, it is only a beginning. Hopefully, as the AIDS Summit in 2001 was a turning point for that epidemic, there can be a similar turning point for diabetes and other NCDs.

  • “The diabetes paradox”: it causes more deaths than AIDS and malaria combined, and yet it is still not a global health priority. It is under-resourced and there is “far too little outrage.” We need to get this on the international agenda. The IDF produced the first ever International Charter of Rights and Responsibilities of People with Diabetes. We need to frame this in a human way as well as in an economic way. “We need to ask the question: why, 90 years after insulin was discovered, do children with type 1 diabetes die in Africa?” Unnervingly, type 1 diabetes is not a huge financial drain on low-income countries, because children die before they can become a drain, she said. “It is not okay for a child to die of diarrhea and it is not okay for a child to die of diabetes. Full stop.” 
  • There are many ways to frame the issue. Besides the human angle, we also need to better make the economic case. We should frame it as an excellent investmentWe can also frame it as a security issue: governments like the US are having increasing difficulty finding healthy young people to recruit for armies and police forces.
  • “NCD is a horrible label.” The public does not know what it means, and it is generally undesirable to define something by what it isn’t. Nevertheless, in 2009, it was clear that there was a political space for an NCD civil society movement. There is and was a risk that diabetes would be marginalized within that movement, for example as simply a risk factor for heart disease. However, she said in Q&A, that when the WHO defined NCDs, they did not break it into very many categories – diabetes, cardiovascular disease, chronic respiratory disease, and cancer – even though it actually is a much larger group of diseases. This was advantageous for diabetes politically.
  • “The summit in September has already changed the global landscape forever, but we never thought we’d get everything we wanted from one event. It was meant to be a landmark to move the whole agenda…it did that…but it was never the end.” It is an official declaration by the world’s 193 governments that “we have a problem.” The only other UN summit on a health-related issue, on AIDS in 2001, is regarded as being the turning point in that epidemic. This is what we hope for diabetes and the other NCDs, said Keeling – “…we are now 10 weeks out.” As we understand it, the timeline going forward includes a UN discussion of global NCD targets in 2012, a progress report on NCDs in 2013, an assessment of NCDs in 2014, and then in 2015, the “next major milestone,” the end of the current Millennium Development Goals.
  • “We thought there would be a shift of resources in the WHO after the Summit.” Diabetes and NCDs are still hugely under-resourced in this organization. In the Geneva headquarters of 3,000 people, there is only one person who works specifically on diabetes. We need a major shift in resources, and the WHO “needs to be extremely strong in a way that it isn’t right now.”
  • The private sector has to be part of the solution. We need to change the confrontational language that is often used, Ms. Keeling said. Rather than talking of “industry” and “conflicts of interest,” Ms. Keeling said we should say the “the private sector,” language that invites effective partnerships. With the private sector’s powerful marketing machinery, she said in Q&A, it can help get the message across, maybe make the cause more “glamorous,” in the words of one attendee.

Questions and Answers

Q: Do you think the NCD label might change?

Ms. Keeling: The NCD label has stuck and I think it is here to stay. One advantage is that when the WHO defined NCDs, they limited its categorization (to diabetes, cardiovascular disease, chronic respiratory disease, and cancer) even though it is actually a much bigger group of diseases. In a sense it worked for us politically. But as much as I hate it, it is here to stay.

Q: Do you think we should be talking about the epidemic of obesity, which leads to whole range of NCDs?

Ms. Keeling: Obesity, definitely. This did not come out strongly enough from the UN Summit. There is not a strong anti-obesity lobby, and that needs to be grown. We should be careful, though, about linking type 2 diabetes and obesity too strongly. Type 2 diabetes in its early origins is also linked to maternal malnutrition, so it is linked to both over- and undernutrition.

Q: Isn’t framing the debate in terms of economic issues the most powerful message for governments, in this economy?

Ms. Keeling: We also have to emphasize that it is not okay to let those type 1 children die, even though they do not make an economic impact. In my opinion, this has to be about human rights as well as economics.

Comment: Dr. Keeling, you have done an incredible job for diabetes. Without you, diabetes would not be on the global agenda as it is today.

Q: As a diabetologist, I’ve noticed that diabetes as a disease and as a specialty, lacks glamour. It is not seen as on the cutting edge of medicine or science, but more like just old age – you live with it. In this day and age, glamour is important…

Ms. Keeling: I would agree with that. Absolutely. Lack of glamour leads to complacency. Cancer has that fear factor, but diabetes doesn’t – and it should. The messaging in diabetes has to have to two aspects – it is a serious disease with serious consequences, but we know how to manage it and allow people to live normal and healthy lives. I don’t know how to make it sexy. Other people have spoken to me about that. But the private sector, with its strong marketing machinery, might help us. I don’t know if it should be glamorous, but we do need help getting the message across.


Beyond Disease

Richard Smith, MB ChB (Ovations Initiative, UK)

Dr. Smith filled in for the World Health Organization’s Michael Marmot, who did not come to IDF. Dr. Smith was the editor of the British Medical Journal, a journal that took up advocacy for NCDs. His talk was called beyond disease, because he spoke about two ways that the idea of “disease” hinders us – it allows us to lose focus on pre-diabetes, and it also obscures the reality that most care in high income countries goes to people with multiple conditions. He then spoke about how to construct healthcare systems that do not mimic the flawed systems of high-income countries but that are specific to low- and middle-income countries.

  • Thinking of diseases as such may be counterproductive – we should be responding to pre-diabetes. “I’d argue we haven’t thought enough about before the disease and putting more resources and energy here.” We know that we can prevent the progression of pre-diabetes to diabetes so this is where efforts should be focused.
  • Low- and middle-income countries should avoid the problems in high-income countries by having fewer hospitals and specialists, Dr. Smith said. Hospitals and specialists capture resources away from the network of primary care providers. Primary care providers, community health workers, and patient self-management are the key ingredients to build up health care systems in these countries.

Questions and Answers

Dr. Alleyne: I think what we need to do is actually to reconceptualize primary care rather than to just emphasize it. This is the wave of the future.

Dr. Smith: We definitely need to rethink the concept of primary care. In Britain, people call it General Practice, but it is not the same as primary care. In the US, gynecologists are considered primary care doctors, and this is not how we would think about it in other countries.

Q: Rolling out something like the DPP would be too expensive, so of course people opt for drugs. So the 79 million people cited to have pre-diabetes in the US is just a carrot for pharmaceutical companies to go after. Pre-diabetes is a risk factor for diabetes, which is itself a risk factor for cardiovascular disease etc. Intensive glucose lowering in diabetes doesn’t significantly impact those risks, so why would we start doing that in pre-diabetes? Nevertheless, targeting pre-diabetes has become the new mantra.

Dr. Smith: I sympathize with a lot of what you say, but on the other hand, why let’s wait until people develop diabetes and then plug them in to this expensive system. It seems much more effective to me to start early on and the less doctor-y, the less hospital-y, the better.

Comment: I support your keep message of prevention, and I think there is room for policies to promote healthy lifestyles. I’m a lawyer and former politician so I don’t know: is there a consensus within the movement about the word “pre-diabetes?” In terms of public debate it is very powerful if we can use that term because it increases the numbers and sends a message that we have a huge problem that is bigger than what you think.

Dr. Smith: I feel the ambivalence myself because I’ve written a lot about medicalization. Suddenly, you turn some 90 million formerly healthy people in the US into now having a condition, a disease, pre- diabetes. And I sympathize with the former questioner about using this to sell more drugs. But if you can encourage all these people to change their lifestyle, why would you sit and wait for them to develop diabetes.

Q: In the UK, how would you deliver prevention? Who pays? Not GPs…

Dr. Smith: Excellent question. How do you set up a system for people who don’t have diabetes? In Britain, it is very unclear to me who would pay. I’m skeptical the GPs would want to pay for this. It means money going out of the health system, into gyms or something. I would suspect this would be an issue in other countries as well – who is willing to pay for this? In the US, when employers provide health insurance, companies would benefit from diabetes prevention, so they might pay. But in other countries, it is a problem.


Symposium: Creating Access to Affordable Medicines and Cost-Effective Technologies

Cost Effective Choices in the NHS

Amanda Adler, MD, PhD (Institute of Metabolic Science, Cambridge, UK)

Dr. Adler gave a fascinating review of UK’s NICE, focusing her discussion on the institute’s approach, analysis, and examples from insulin detemir, glargine, and exenatide. She emphasized opportunity costs throughout her presentation, arguing that limited resources mandate that cost-effectiveness be considered, “We cannot pay for everything. We must make choices.” Most notable in Dr. Adler’s presentation was a fascinating cost-effectiveness analysis of insulin glargine and detemir compared to NPH. NICE appraised the drugs at £320,029/QALY (glargine) and £417,625/QALY (detemir), far exceeding the typical cost-effective threshold of £20,000 - 30,000/QALY. Dr. Adler likened the basal insulins for type 2 diabetes to a Porsche ambulance in Germany – they may have some extra bells and whistles, but are generally comparable to the old standard (i.e., NPH insulin) in accomplishing their main objective. We were also interested to learn more about how NICE determines these figures, which includes use of the UKPDS outcomes model. Interestingly, body weight (and therefore weight loss) is not taken account into the complications model, a surprise to us. Nevertheless, NICE used weight loss considerations in its review of GLP-1 agonists, a potential sign that the institute is moving away from a glucose-centric point of view. We found her presentation quite informative, although we wish she had devoted more time to GLP-1s and how NICE is thinking about their costs, benefits, ideal therapeutic use, and patient populations.

  • “Regulators decide whether Britain can use a drug, NICE tells Britain whether it should use a drug.” Dr. Adler described the rationale for NICE, which was created 11 years agoto advise on the cost-effectiveness of drugs and devices. She made it clear that companies must first get marketing authorization from the EMA or MHRA before NICE appraises the drug. Dr. Adler alternatively described the difference between NICE and regulators as the respective questions they answer – regulators must answer, ‘Does this new drug or device do what the label says’ while NICE answers, ‘Is it better than what we already have, and does it do so at a cost that reflects good value for money?’ She also noted that over the last eight or nine years, the lag time between marketing authorization and NICE recommendation has sped it up, typically lasting a matter of months.
  • Establishing the effectiveness of new drugs for diabetes is a challenge compounded by the need for more than placebo-controlled trials. Dr. Adler noted the different studies of diabetes drugs: placebo-controlled trials with an A1c outcome (least useful for NICE), trials comparing a new treatment to standard therapy with an A1c outcome, and trials comparing a new treatment to standard therapy and measuring the differential incidence of complications (most useful for NICE). Unfortunately, the latter is not typically feasible, so NICE must do the best it can with the other types of trials.
  • In Britain, NPH insulin is the standard third-line therapy (following metformin and sulfonylureas) and the comparator in NICE’s appraisal of basal insulin and exenatide. A third agent is recommended with an A1c >7.5%, although ongoing review of the use of sulfonylureas and hypoglycemia should occur. Premix insulin is recommended if A1c is >9%.
  • NICE uses the UKPDS outcomes model to assess length of life, and interestingly, body weight is not taken account in the calculations. NICE recommendations are based on cost (in British pounds) per quality adjusted life year (QALY), which allows comparison across different treatments and diseases. Dr. Adler reminded the audience that the model assigns a QALY of 0.75 to people with diabetes, meaning someone with diabetes has the quality of life of about three-fourths of a normal person – this figure has always struck us as quite arbitrary, especially when we remember our favorite Dr. Bill Polonsky quote, “Well-controlled diabetes is the leading causes of nothing.” When evaluating new drugs for type 2 diabetes, NICE uses the UKPDS outcomes model to mathematically relate risk factors that drive survival. Factors in the model include A1c, blood pressure, lipids, age, sex, ethnicity, duration of diabetes, and smoking status. She emphasized what is not included in the model, “Body weight, hair color, and astrologic sign.” We were interested to hear that body weight was not factored into the model, an especially important differentiator (or detraction) for many current classes of drugs. We wonder if these complications models need to be updated considering the rapid advances in treatment, detection, and research that have taken place since UKPDS was reported.
  • NICE’s analysis of insulin glargine and detemir deemed the analogs equivalent to NPH (and each other) for glycemic control, and better in terms of hypoglycemia.

Glargine vs. NPH

Detemir vs. NPH

Δ A1c


0.08% (favors NPH)

Δ Weight

-0.28 kg (not significant)

-1.2 kg (significant)

Severe Hypoglycemia

No difference

No difference

Any Hypoglycemia

OR: 0.74 (0.63-0.89)

OR: 0.51 (0.35-0.76)

Quality of Life Not enough information Not enough information
  • A cost-effectiveness analysis revealed insulin glargine and detemir dramatically exceed NICE’s threshold of £20,000 - 30,000/QALY. The analysis was based off a male with a BMI of 30 kg/m2 and complications (more specifics not given). According to Dr. Adler, the costs include insulin, nurse time, adverse events, complications, and annual diabetes care. The analysis was published in Health Technology Assessment 2010; Volume 14: No. 36.








Drugs Costs




Cost per Qaly



  • After conducting an analysis of exenatide, NICE recommended use of the GLP-1 agonist as a third-line therapy when certain obesity-related considerations and complications are taken into account. Unfortunately, Dr. Adler did not provide any cost- effectiveness analyses or give further color on how NICE is thinking about the drugs.

Exenatide as a Third Line Therapy for Type 2 Diabetes

Starting exenatide

BMI >35 kg/m2 and specific psychological or medical problems associated with high body weight


BMI <35 kg/m2 and therapy with insulin would have significant occupational implications or weight loss would benefit other significant obesity- related comorbidities.

Stopping exenatide

Reduction of >1% in A1c and weight loss of at least 3% initial body weight

Questions and Answers

Q: In Israel, we are having the exact same problems that are hard to work against. Put, simply, patients will not accept old drugs if they are not as effective as new drugs. We are working hard against the PR of companies.

Dr. Adler: It is a challenge. One way to help people understand is to have them think about opportunity cost within the diabetes area. In Britain, we have successful diabetes education programs that people love - more spending on drugs means less to spend on these programs. It’s difficult when pharmacies supportthe patient groups. Sometimes the patient groups, rather than be frustrated, need to turn to pharmaceutical companies to bring down prices. It’s very tough for us if a new drug is better than the old drug, but it’s not worth the price the manufacturer chooses to charge. Effectiveness is God-given, while price is set.

Q: I’m from Finland and just wanted to thank you. This is one of the most important symposia at this conference and I was hoping the hall would be full of people. Reimbursement and availability of medicines really influences diabetes care. I think we need more transparency in the market. I’ve followed the NICE process as I personally have type 1 diabetes. I used to use animal insulins and now have an insulin pump. I understand how you evaluate drugs for type 2 diabetes such as analog insulins, but in the future, I would ask you that as we get new medicines, please consider the quality of life of people with diabetes. Especially people with type 1 diabetes. Quality of life is not evaluated in isolation, but how drugs affect everyday life on different days – working days, weekends, and holidays. In short, what is the best way to live a normal life like other people can.

Dr. Adler: Thank you. It’s Jonathan Brown’s credit that we have this session. Quality of life is very important and you can imagine it’s a challenge to assign a number. It’s an area of research that is exceedingly important – it’s easy to put a number of length of life, but putting a number on quality of life is much more difficult.

Q: A question about weight. We know that at any moment in time in the UK, 9% of people in the hospital for diabetes are there for the primary diagnosis of diabetes. When you make these decisions, would you consider broadening the questions?

Dr. Adler: Generic measures of effectiveness are a way to broaden the question. We’re not sure in the NHS whether we want to pay for someone to look better in bathing suit. But there are real issues with weight. Maybe someone who loses weight is less likely to have knee surgery. With exenatide, you need to have an obesity-related complication vs. a diabetes-related complication. Being obese was associated with either frustration or an actual medical problem. You could take an incredibly broad look for every new diabetes intervention. NICE is considering new ways to look at things. We must not only consider that what we spend on diabetes cannot be spent elsewhere, we should take into account what caregivers cost, what stroke from diabetes costs, for example, and how that affects quality of life of people with diabetes and those caring for them. Diabetes complications can prevent people from working, which means they will ask the government to pay for unemployment insurance. NICE does not take that into account at this time, but we may broaden that. One impressive thing about NICE is they continually reevaluate and attempt to improve things.

Q: What happens if you’ve got a product and you don’t have the data to make a decision.

Dr. Adler: Regulators want to see new drugs that work better than pretty much nothing – in others words, new drugs must prove they can lower blood glucose better than placebo. That’s not clinically relevant in Britain. I don’t offer my patients placebo. When the comparative data is not available, some clever people are taking studies of drug A vs. placebo and drug B vs. placebo and estimating how drug A would fair against B in a theoretical head-t0-head comparison. It’s a modeling exercise that may sound a bit fiddly to you. But in the absence of anything else, it’s the best we’ve got.


Cost-Effective Choices in an HMO

Wiley Chan, MD (Kaiser Permanente Northwest, Portland, OR)

Dr. Chan gave an instructive talk on Kaiser Permanente, the largest non-profit healthcare program in the US. The organization takes a very integrated, future-oriented approach towards patient care – a relative rarity among US payors. We took particular note of Kaiser’s use of POEMs: patient oriented evidence that matters. Such thinking takes into account things patients actually care about (e.g., symptoms and quality of life rather than an A1c or LDL value), which we presume results in prioritization and utilization of therapies that better meet patient needs. Dr. Chan also discussed the organization’s preference for a trio of CV therapies (aspirin, ACE inhibitors, and statins) over intensive glycemic control. Kaiser’s population- and evidence-based approach suggested this was the way to go, and it certainly complements the shift we’ve seen away from a glucose-centric view of diabetes treatment.

  • The forward-looking Kaiser Permanente is the largest non-profit healthcare program in the US. Kaiser is a prepaid integrated healthcare delivery system that combines medical groups, hospital groups, and an insurance arm in eight regions spread across nine states. The organization’s 15,000 physicians serve nine million patients. Kaiser’s three entities share accountability for the program’s success, use a common EMR, and are aligned in goal and patient care. Perhaps most notably, Kaiser uses a population-based perspective to facilitate its ultimate goal: improving the health of its patients on a broad scale.
  • Clinical care within Kaiser is informed by the GRADE methodology and POEMs. GRADE is an international collaboration intended to assess the quality of evidence behind an intervention – it includes Grading of Recommendations, Assessment, Development, and Evaluation. Factors that influence the strength of a recommendation include the balance between desirable and undesirable effects, the quality of the evidence, values and preferences, and resource use (costs). A ‘strong’ recommendation means most patients/providers would want to use it/prescribe it, while a ‘weak’ recommendation implies (a) most patients would want it, but many would not; or (b) different provider choices are appropriate for different patients. Kaiser also uses POEMs – Patient Oriented Evidence that Matters – in its evaluations. This mindset addresses questions and outcomes that patients care about: symptoms, morbidity, quality of life, and mortality (rather than A1c and LDL values, for example). Kaiser offers only a single recommendation, clinical effectiveness is addressed before cost.
  • Kaiser treats type 2 patients with an ACE inhibitor, a statin, and aspirin before targeting intensive glycemic control. Ten years ago, the organization carried out a formal cost-effectiveness analysis and decided that this trio of therapies reduced the annual risk of micro- and macrovascular complications more than intensive glycemic control. A 2009 publication analyzing 170,000 patients with diabetes substantiated this strategy (Dudl Am J Manag Care 2009). For diabetes, Kaiser first ‘tests the untested’ – LDL, blood pressure, and A1c. Then it ‘treats the untreated’ by putting patients interventions known to reduce the risk of cardiovascular disease: aspirin, high dose statins, ACE inhibitors, and sometimes beta-blockers. Only after that are the LDL, blood pressure, and A1c targets chased.

Questions and Answers

Q: I’m from an Israeli HMO that is second in the world behind Kaiser. We have been doing the same kind of intervention as Kaiser. We have found that it lowers costs and resulted in 25% less amputations, 50% less blindness, 50% fewer patients in end renal stage, and 20% fewer invasive procedures for cardiac disease. On aspirin, there are so many articles that say you don’t need aspirin in diabetes if a person didn’t have CVD. Additionally, taking one drug away from a patient promotes adherence. What are your thoughts?

Dr. Chan: Those are very impressive results. I didn’t mention the emerging evidence that caused us to withdraw our routine systematic push for aspirin in patients with diabetes without CV disease. The point estimate favors benefit over harm with aspirin, although the confidence interval crosses one. The risk of complications is fairly low. We’re not actively discouraging aspirin in diabetes, but we’re not pushing a hard metric either.

Q: You mentioned you centralized procurement. Primary care physicians are an independent species. What was the reaction from clinicians to centralized procurement? Are there rules as to what’s available and clinicians follow those rules?

Dr. Chan: Very few interventions are actually being restricted in Kaiser Permanente. Clinicians can prescribe what they like. In some cases we go after those physicians, but in most cases not. There are very few cases where we would prevent a clinician from using an intervention.

Q: Do you keep track of how many patients choose to use fancier insulin or smaller gauge needles?

Dr. Chan: Most of the time, it’s the patient who initiates those questions about an intervention. I don’t have those stats.

Q: The UK NHS has a culture of audit and feedback. Does Kaiser Permanente have that as well? How is that transmitted to providers? Is it a name-in-shame sort of thing?

Dr. Chan: We didn’t have internally reported unblinded performance metrics. Some are around utilization resource issues. But physicians are part of the solutions and all share responsibility for the success of our programs. It’s not a hard sell to clinicians when a specific drug does not have enough benefit. We typically don’t have many tough discussions with clinicians. One we did have was for colonoscopy and colon screening. It was a challenge to convince them to use chemical testing.

Q: Do you use pay for performance?

Dr. Chan: We do have pay for performance – most are around quality metrics. I cannot think of a utilization metric, but that’s not to say we couldn’t use one.

Q: Can Kaiser’s principles apply internationally to other organizations?

Dr. Chan: The first thing is that our unique structure allows us to do things that might be hard to replicate. We have shared accountability between physicians, the hospitals, and the insurance group. Prioritization is developed in collaboration and everyone must be on the same page. Getting patients to comply is a tough sell in the US – they always have the option to switch insurance companies. Most patients will understand if you describe the population approach. That can be a more difficult conversation that we have with our clinicians.

Q: Sounds like impressive physician-related metrics. Do patients have access to this?

Dr. Chan: Patients have access to some. Some are publicly reported, but not at the individual physician level. We post many of these metrics at a clinic module level, which includes five to ten physicians in a group. Patients can look up the performance of one location and see how they’re doing vs. other locations. We’re not aware that other HMOs are doing that. In the future, at least in Oregon, we might report at a lower level.

Dr. Jonathan Brown (Chair, IDF Task Force on Diabetes Health Economics): I wanted to add to the question about bringing this model to other countries. The key thing in the US is that insurance companies’ customers are businesses for the most part. As a result, businesses are trying to save money and not spend much and they are looking for plans that do a good job and provide good value for money. There is a national system of measurement of performance in diabetes and cardiovascular disease – it is published. While you might not have an integrated system in every country, the publication of metrics makes a big difference in the United States. This might be exportable to some extent.


Pricing and Availability of Essential Medicines

Margaret Ewen, Dip Pharm (Health Action International, Amsterdam, Netherlands)

Ms. Ewen’s presentation gave a useful overview of the high cost and low access to essential medicines in many countries around the world. She began with a review of data published in the Lancet (2009) by Cameron et al. The study found that public and private sector prices for originator and generic medicines were substantially higher than would be expected if purchasing and distribution were efficient and mark-ups were reasonable. The limited statistics she provided on markups on essential medicines were particularly disturbing – these ranged from 25% in Pakistan to 500% in Khartoum to >6,000% in El Salvador. We were most interested to hear diabetes-specific data on metformin and human insulin, which showed notable variances in prices around the world. Ms. Ewen also discussed a surprising diabetes advertisement from Nicaragua: “I have diabetes. If my medication fails, I could suffer a diabetic coma…I don’t take chances, I only use originals.” The ad appeared in 2008 and 2010 and was supported by Abbott, AstraZeneca, Boehringer Ingelheim, Bristol-Meyers Squibb, GlaxoSmithKline, Lilly, Merck, Pfizer, Roche, Sanofi, and others. Ms. Ewen closed with a discussion of steps to improve the availability and pricing of essential medicines, including improved government policies, lower taxes, incentives for pharmacists to dispense generics, and more efficient supply systems.

  • Ms. Ewen showed data on the affordability of three-times daily metformin (500 mg), suggesting the medication is still very expensive for many consumers around the world. The figures are based on the number of days the lowest paid unskilled worker would have to work to purchase a one-month supply. In the Democratic Republic of the Congo, a worker would need to work 14 days out of the month to purchase the generic brand and 13 day per month to purchase the originator brand; this compares to five days per month in Shaanxi Province, China.For what is likely the cheapest diabetes medication on the market, we found these high costs quite surprising.
  • On May 11, 2010, Health Action International collected the price of a 10 ml vial of human insulin in 60 countries around the world. The price was taken at between one and six private retail pharmacies in each country, so Ms. Ewen conceded the figures are not that reliable – still, we found them quite interesting. Austria was the most expensive in the world ($77 USD), followed by the US (~$53). The lowest prices were in Cairo, Egypt, where Lilly’s Humulin S sells for $2.52, and in Tehran, Iran, where Exir-Iran’s Lansulin R sells for $1.55. The complete list is available at


Questions and Answers

Dr. Jonathan Brown (Chair, IDF Task Force on Diabetes Health Economics): I have about ten questions for you. For World Diabetes Day, we light up buildings all over world. Could Health Action International coordinate doing the one day global survey you talked about?

Ms. Ewen: That would be great. We need more attention, particularly on insulin and some of the newer diabetes oral medications.

Dr. Brown: Right. We also forgot statins and blood pressure medications for patient with diabetes. A question about pre qualification and joint purchasing – in the asthma arena, a union is buying large quantities of medicine so many countries can get a low price. This seems like a good way to go.

Ms. Ewen: I think we should lobby the WHO to extend prequalification to medicines for diabetes. In the long term, it’s a more sustainable option – I agree.

Dr. Brown: There is a debate about generic insulin – IDF has been criticized for not promoting them. But we’ve often heard that when generics are available, they are sold for the same price as originator brands. Also, insulin can be tricky because of quality issues. Is nationally produced insulin the solution to lowering prices?

Ms. Ewen: Prices have been driven down because of competition. For insulin, we have very limited suppliers and that’s a problem. We have to ensure the quality of other supplies. If we had more suppliers, I would think we would see a lowering of prices. What is so difficult with the process that there can’t be more suppliers? If it was easier to produce them, we would see a fall. I think it’s more than financial. It can be very valuable to have local production, but the insulin must be of the quality needed. It’s interesting when the price is higher, because it might reflect that there is not enough demand. There could also be perceptions about quality issues.

Q: This is absolutely fascinating. It’s astonishing what an imperfect market we face. Rather than beat up any particular part, do we not have a responsibility to help the market develop? When there are ridiculous margins, you’ve got a poor market. Competition helps and transparency helps. It seems we have very little transparency in the market. At IDF, we need to help develop this.

Ms. Ewen: That’s a very valid point. There is no reason to get angry about prices – we need solutions.

Q: I’ve worked in many markets and I have never seen margins like that in any sector.

Ms. Ewen: Let me be clear: El Salvador was a one-off. Although some countries in Africa have cumulative markups well over 200%.

Q: But when you take a luxury good product, say gems for example, work at 100%.

Ms. Ewen: Right, and we’re talking essential medicines here.


Presidential Oration: Turning the Tide of Diabetes Together

Jean Claude Mbanya, MD PhD (President of IDF, Cameroon)

Dr. Mbanya’s talk was similar to the morning symposium on a post-UN-Summit world, in that he expressed pride in the strides already taken but still awed by the problem we face – a battle cry of sorts. Though 14,000 attendees at the conference last year seems quite substantial, when you compare it to the 366 million in the world who have diabetes, according to the new Diabetes Atlas 5th edition, it seems like a wooden fence in a flash flood. No country is immune, Dr. Mbanya said, and no country has the epidemic under control. Rising health costs are crippling governments. Diabetes is related to other non- communicable disease, infectious disease, economic and environmental sustainability, and human development – failure to address diabetes means failure to address all of these. And yet: “Where is the outrage?” He asks, “Where is the outrage for lack of insulin?” A handful of Dr. Mbanya’s other quotable quotes and interesting points are outlined below.

  • The number of people with diabetes is roughly equivalent to the entire population of the Middle East and North African region: “the combined population of 20 nations.”
  • “It is impossible to understand why these diseases are not the top priority.” For some reason, deaths from pneumonia or malaria cause outrage but not deaths from diabetes. Fourdiseases that represent over 60% of burden – diabetes, cardiovascular disease, chronic respiratory disease, and cancer – receive only 3% of official development assistance for health. “It just doesn’t make sense…We can’t allow them the excuse of the global financial crisis to compromise the health of millions.” To this, there was applause. “Don’t let power brokers skirt this responsibility with pleas of destitution.”
  • “Our long-held assumptions about what constitutes a good life, development, progress – all these are being questioned in a devastating way.” We felt this was quite thought-provoking: throughout history, leisure and plenty have been seen as goals of human progress – now diet and exercise must become a priority.
  • “We have the solutions. We know what to do to turn the global epidemic around.” Dr. Mbanya called for a solution-oriented approach. The IDF’s Global Diabetes Plan 2011-2021 has these answers, and he is “extremely proud of it.” Its goals are to 1) Improve health outcomes for people with diabetes, 2) Prevent development of type 2 diabetes, and 3) Stop discrimination against people with diabetes.
  • “We can’t fix this just by changing lifestyles – and the thought that doing so would be free – nonsense.”
  • “I’m filled with optimism for the future. I didn’t think we’d come so far in these few short years. But it’s all just words until we see a difference.”


Meet-the-Expert: 21st Century Vision for an Integrated System to Control and Prevent Diabetes

21st Century Vision for an Integrated System to Control and Prevent Diabetes

Nick Wareham, MD (Institute of Metabolic Science, Cambridge, UK)

The renowned epidemiologist Dr. Nick Wareham gave a valuable overview of IDF’s Global Diabetes Plan 2011-2021, sharing insights on how we can better integrate diabetes care and prevention. He focused his presentation on the essential pillars of a diabetes care system, what he calls the seven R’s: recognize/reconsider, registration, risk assess, respond, recall, and review. We found this quite intuitive and simple to understand, although we wish more specifics had been included. Dr. Wareham had some interesting ideas for preventing diabetes, including widespread adoption of an A1c-based definition of diabetes and use of low-cost, multicomponent interventions to stimulate modest weight loss across populations. Interestingly, an analysis found that one kilogram of weight loss would be cost-effective at NICE’s threshold of £20,000/QALY if it cost less than £100 per person. Weight loss of three to four kilograms would make the same intervention cost saving. Finally, a population level intervention that achieved 0.25 kilograms of weight loss would be cost effective at the NICE threshold if the cost per person was £10. We are very encouraged to see that such modest weight loss interventions would be cost-effective. The major question is whether they can be delivered at the cited per-person prices. And of course one of the primary challenges is convincing policymakers that they are worth the money and time. We found the stimulating Q&A following the presentation quite notable, chaired by IDF President Dr. Jean-Claude Mbanya.

  • Dr. Wareham feels we should be striving for integration in all aspects of diabetes care. This includes clinical care, systems of early detection (screening and prevention), between different diabetes risk groups, diabetes treatment, and other disease prevention programs. Healso noted the importance of finding a balance between societal and individualistic approaches to prevention – a particularly challenging and controversial task. Finally, Dr. Wareham asserted the importance of integrated international, national, and local strategy/action.
  • IDF’s Global Diabetes Plan 2011-2021 features an overarching model based on answering the question, ‘What are the key attributes of diabetes care systems?’ According to Dr. Wareham, the framework was designed to avoid the complicated “medical textbook discussions” and drill down to the key pillars of diabetes care systems. Encouragingly, it is designed to be simple and enable implementation by community health workers and non- diabetologists. The framework is described by seven R’s:
    • Recognize/Reconsider: This is the first essential step in a diabetes care system – you must recognize whether someone has diabetes, and if so, what type. Dr. Wareham talked about the difficulty of drawing a line between abnormality and normality. For example, when a group of experts convened at an NIH meeting a few years ago, they were unable to create an operational definition of type 1 and type 2 diabetes. In his view, the purpose of a diagnostic label is to bring about action by the health system or practitioners. Unfortunately, he believes we’ve lost that way of thinking, perhaps because diagnostic labels sometimes get in the way and lead us down the wrong path.
    • Registration: Registration is an essential pillar of a diabetes care system. It enables tracking, recording, and subsequent patient follow-up. Dr. Wareham noted that the registration system can be as simple as a piece of paper and a pen, or as complicated as an electronic medical records system.
    • Risk Assess: Dr. Wareham devoted the majority of his attention to risk assessment. He began by reiterating the central idea in his Diabetologia 2011 – the tools you use depend on what you’re trying to accomplish. Next, Dr. Wareham provided a clinical case study to demonstrate the value of risk assessment. For instance quantifying an individual’s risk (e.g., Framingham Risk Score) would facilitate risk ranking and targeting those with the highest risk – this is especially useful because resources are limited and not everyone . Risk assessment also provides estimation of likely benefit from an intervention. Finally, risk assessment can motivate patients to change their behavior.
  • Presentation of risk (e.g., numerical, graphical) can dramatically affect a patient’s response, and more data is needed to determine best practices. Dr. Wareham explained that numerical presentation of risk, as opposed to categorical presentation, leads to more accurate risk perception by patients. Interestingly, the relative risk reduction format encourages acceptance of treatment the most. In other words, “If I want you to take statins, I will tell you it reduces your relative risk of heart disease by 25%.” As might be expected, the somewhat abstract numbers needed to treat format encourages the lowest acceptance of treatment. Dr. Wareham likes the Framingham Heart Risk score, the National Cholesterol Education Program risk calculator, and an innovative risk calculator developed by Unilever. The latter gives patients both a chronological age (e.g., 60 years old) and a heart age (e.g., 80 years old). Finally, Dr. Wareham considered the personalization of risk, such that a patient would be shown a picture of their arteries or a body fat scan to motivate them to change. Unfortunately, more research is needed to ascertain what method of risk presentation patients most readily respond to.
    • Respond: Of course, it’s not enough to simply carry out risk assessment. The clinician and/or the patient must take action.
    • Recall and Review: Dr. Wareham described the importance of recall and review, summarizing some of the scientific literature supporting their use (Griffin et al., BMJ 1998; Renders et al.,Diabetes Care 2001).
  • Dr. Wareham believes that an A1c-based definition of diabetes would aid and simplify the development of an integrated system of early detection and high risk stratification. For prevention, Dr. Wareham also favored the strategy used in the Finnish Diabetes Prevention study. In this trial, participants were asked to reach five targets: (1) normal BMI; (2) increasing physical activity; (3) increasing fiber; (4) decreasing total dietary fat; (5) decreasing saturated fat. As might be expected, patients in both the intervention and control groups had zero incidence of diabetes if they achieved four or five of these targets. Those who achieved zero targets had a very high risk of developing diabetes.
  • We need a balanced approach to diabetes prevention that trades off individual approaches with population-based approaches. An individual approach to diabetes prevention involves targeting a relatively small group of high-risk individuals; while this would have low numbers needed to treat, the public health impact is also relatively low. On the other hand, population-based approaches have high public health impact, but small individual benefit, high numbers needed to treat, and high cost. Dr. Wareham noted that a multifaceted approach is likely needed, which would involve trying lots of interventions. Provided they are cheap and done across a lot of people, he believes these would be cost effective across the population.

Questions and Answers

Dr. Wareham: What do people think of this?

Dr. Jean-Claude Mbanya (University of Yaoundé, Yaoundé, Cameroon): I think it helps people to think. To do something rather than following guidelines blindly.

Comment: This is human nature – the seven R’s makes it inquisitive and slightly more appealing. I haven’t taken any notes on your presentation other than writing down the seven R’s.

Comment: I think it’s a bit cute. Many of the words are very clear. But some aren’t. What does respond mean for instance?

Dr. Wareham: Responding to risk assessment.

Dr. Mbanya: How many people think this would be useful in clinical practice?

[Majority of the 75 people in the room raise their hands]

Comment: In my practice, the thing that’s made the most difference in quality has been our recall and review system. I don’t know whether you need all the rest. I think you have too many R’s and too many E’s.

Comment: We’re naïve if we have this idea that everyone is living with good care. Everyone is not. We need to get the message out there. We need to get them to think simply and logically about what the key components are.

Dr. Mbanya: Most of you don’t know, but in most middle- and low-income countries, nurses and healthcare providers not trained in diabetology offer the care. This is intended to be something simple for them to follow.

Comment: I think lots of structure is good when dealing with a patient who is educated. But this might be overwhelming.

Dr. Wareham: The seven R’s are for the people designing the system, not the patient.

Comment: I’m from Canada and I’ve been involved in developing our guidelines. We reviewed the literature about what good care looks like. Convincing people to change is a part of the provider’s job. But people need support to change. Family physicians are who largely care of type 2s in Canada. There is resistance against recall. They say, ‘My patients don’t want to change, so why do I have to baby them and call them back in?’ Well, the evidence shows they’ll do better. Also, many don’t have electronic medical records so they don’t know whom to recall. However, I think this is very excellent and very simple and I’m going to steal it.

Comment: All of this is really brilliant and great, but if you don’t motivate the patient it doesn’t matter. We almost need marketing training. Everything we do is sold to us – there is this element of being told what to do. You must negotiate with the patient; what will I gain as the patient? Then, it changes to the doctor helping you achieve the goal. I like how you personalize and help the patient own the report. I think they act on it.

Dr. Wareham: I think intention was not to talk about details. It is an overarching structure for a system.

Comment: I think public perception has a lot to play with it. Some don’t believe they are at high risk and won’t get screened. Reaching those people is a challenge. We need a way to educate people that you don’t have to become a diabetic – being healthy is an option. You are not definitely going to have diabetes, and it’s still okay to be healthy and well. I’m a freelance pharmacist. When I work in supermarkets, I know that the people being screened are elderly upper middle-income people.

Dr. Wareham: Remember that approaches in medicine are very individualistic and often require people to respond. They are almost always socially regressive, targeting the better off. The problem in the UK is health inequalities: diabetes disproportionately affects those of lower socio-economic status. We need to find ways to reach out to those people.

Q: What if everyone got an annual A1c?

Dr. Wareham: It would be a waste of money. We have to target to those who are most at risk. Comment: Governments or individuals are not going to donate 100% of their income to health. Looking after patients who already have diabetes is really fundamental. It gets extremely expensive. Dr. Wareham: But it would be a shame if we only did that.

Q: I think the Diabetes Atlas is fascinating. Have you looked at the characteristics of countries with the lowest incidence of diabetes?

Dr. Wareham: This is semantic, but the Diabetes Atlas is one of prevalence. We would love to have one of incidence. If our target is to reduce diabetes prevalence, it will go up with a good system of detection and care. People will be detected more and they will live longer.

Q: Is there a way of teasing out the difference?

Dr. Wareham: Yes, but it’s very difficult.

Comment: I don’t think diabetes-specific interventions are the way forward. There are risk factors shared across many chronic diseases.

Dr. Wareham: Yes, which is why Jean-Claude stood up with the other NCD groups at the UN Summit.

Comment: I think what’s more attractive to the population is statements about risk. What’s my risk of this, this, and this if I don’t do this?

Dr. Wareham: We published a paper looking at alcohol intake, smoking status, fruit and vegetable intake, and physical activity. If you take those who did all four, they had 14 extra years of life expectancy compared to those who didn’t.

Comment: [Referring to the Global Diabetes Plan diagram] From a pictorial sense, I’m not seeing that your schematic includes population and community-based approaches. You need population and environments in that diagram. I would also say the word ‘screening’ is not ideal. We use the phrase ‘risk assessment and detection’ in Australia. It still says screening in the diagram.

Dr. Wareham: Good points, thanks.

Comment: I come from an area where prevalence of diabetes is 20%. Since then, I suspect prevalence has gone up. Policymakers have not taken enough notice.

Dr. Wareham: For too long, studying prevalence and incidence has been the subject of academic work. It ought to be the subject of routine surveillance.

Q: Given that 10-pound weight-loss interventions are cost-effective, what stops governments from using legislative interventions to cut trans fats, sugar sweetened drinks, etc.

Dr. Mbanya: Lobbying. Lobbying. In New York, [at the UN Summit] they said voluntary reduction: reduction by this percent by this time. It’s heavy lobbying that goes straight to the white house. When there is a strong lobbying community, we don’t have the financial muscle to counteract that. I was so disappointed when issues of human interest and life and death were considered behind politics.

Q: I would advocate something besides a reactive approach. There is this book called the Blue Zones that looks at the longest living people around the world. Why don’t we look at what these people have done? Wouldn’t that be more effective?

Dr. Wareham: Some of those populations have chosen to live a certain way going back a long time. It’s very difficult to go to developing countries and say don’t develop.

Q: But couldn’t the principles be incorporated?

Dr. Wareham: It’s worse than the don’t develop argument. It’s saying by virtue of your ethnicity, you’re at increased risk.

Comment: I come from the country of Georgia and we have many troubles. In any stage we have different barriers. We need to figure out why patients have barriers. Why don’t they trust us? We need to have good information.

Dr. Mbanya: All this has to be updated with what we can put in the system to adapt and change it.

Comment: On the messaging content with people with diabetes, I like using graphic images. There’s lots of work in Australia with tobacco advertising. We need risk messaging testing in diabetes. We also tend to focus on CV risk and death, but to many people, that risk is less important. People know they will die. There is this sense of inevitability about CV death. We’re finding the motivators are in other areas.

Dr. Wareham: I think you’re perfectly right. We assume CV risk is important because of the medical perception. Maybe we ought to give risk scores for loss of income or level of stress or erectile dysfunction.

Comment: I think erectile dysfunction would be a strong motivator for men to change. [Laughter]


9. Exhibit Hall Report

The large IDF 2011 Exhibit Hall featured a wide variety of company exhibits and over 160 IDF member associations from countries around the world. Many exhibitors reprised their booths from EASD 2011, so we found little new news in the exhibit hall overall. One unique exception was a much publicized hall visit from his highness Sheikh Mohammed bin Rashid Al Maktoum, Vice President and Prime Minister of the UAE and Ruler of Dubai. We thought this was a nice gesture and hope it’s a sign of growing political attention to diabetes in both the UAE and the larger Middle East region.

  • Abbott: The majority of Abbott’s blue and white booth was occupied by a quiz show game about lipanthyl, the company’s fenofibrate for the treatment of high cholesterol (We have to admit, these were easily some of the most difficult quiz show questions we’ve ever faced in an exhibit hallwe personally tend to love these quizzes!) Participants received a combination BMI calculator and tape measure for playing the game. The backside of the booth had more information about lipanthyl and one brochure on Abbott’s meters. We were surprised to see no major signage devoted to blood glucose monitoring, especially the Insulinx, which was introduced in five European countries in May and recently approved in Canada and the US (for more information about the InsuLinx, please see our EASD Exhibit Hall report at At the time we visited the booth, no meters were on display and reps on staff were not able to answer our questions about diabetes-specific products.
  • Bayer: Bayer’s booth showcased a worldwide sample of blood glucose monitoring devices, including its flagship product Contour, the developed-world success Contour USB, the Contour TS (which is targeted to developing markets and has its own strips), and – most excitingly – the Contour XT, the first meter to use Bayer’s new highly accurate Contour Next strips. The Contour XT was similar in size and shape to the Contour itself, but we expect that the form factor will change in the second generation (which will reportedly include built-in data management software, as we learned at IDF). So far the Contour XT has launched in Italy and Iberia; we have heard little about the launches but assume progress is good. Indeed, the rep we spoke with said that Bayer plans to switch all of its meter franchises over to the Contour Next strip. As at EASD, Bayer also drew attention to Glucobay (acarbose), its alpha-glucosidase inhibitor (“the simple way to flatten the peaks” – an excellent tagline in our view). To match the mainly white décor, frothy cappuccinos were available for visitors.
  • BMS/AZ: Bristol-Myers Squibb and AstraZeneca had two exhibits in one, nearly completely separated by a wall. On one side was a nice open green and white space with signs promoting Onglyza (saxagliptin). We immediately noticed a large screen with attendees’ names and photos appearing along with a personalized “pledge of support” for the fight against type 2 diabetes. A booth within the exhibit allowed attendees to enter this pledge and have their photo taken – we thought this was a cool interactive and personalized addition to the exhibit. Near this was a smallseating area where attendees could relax and enjoy coffee from an espresso bar. Beyond a wall was the second exhibit, smaller and containing its own espresso bar and tables. We noticed less traffic overall despite the booth’s focus on the novel SGLT-2 inhibitor class. No brand information was listed and signs and an animation talked generally about the “science of glucose regulation.” Boehringer Ingelheim: In a sharply decorated, white and purple colored booth, Boehringer Ingelheim prominently featured its recently approved DPP-4 inhibitor Tradjenta (linagliptin). Throughout the booth, brightly lit display panels and video screens displayed prescribing information for Tradjenta, clinical trial results, and slogans such as “Kidneys Matter in Type 2 Diabetes” and “Control and Care Matter.” The emphasis was on the therapeutic benefits of Tradjenta, including meaningful and durable glycemic control efficacy, a low risk for hypoglycemia, weight neutrality, greater simplicity through a single dose, and a lack of dose adjustment required for individuals with renal or hepatic impairment.
  • J&J/LifeScan: While the OneTouch Verio Pro meter was the clear star of LifeScan’s EASD 2011 booth, the company’s IDF exhibit featured an ensemble cast of meters from around the world. For the first time we got to see the Select Simple, a no-button meter that has been launched in India and Pakistan (see September 12 Closer Look at The meter’s size and shape are similar to that of the Select, but the icon-driven interface and automatic alerts for high and low results are notable changes. (One downside of the simple interface is that retrospective review is possible only by USB cable download.) We noticed that LifeScan made a strong effort to emphasize its place in Arabic culture, highlighting its Lamasat program for caregiver support that launched in Dubai during IDF. LifeScan reps also made a strong effort to share glasses of juice – a change from EASD, where Jay Jay’s Juice Bar required orders to be placed by scanning one’s badge at a kiosk. We were very impressed by LifeScan’s TV ad featuring the Arabic matriarch – as we understand it, the average mother in the Middle East influences eight people with diabetes on average (parents, spouse, children, close friends, etc.).
  • Lilly: Similar to EASD, Lilly’s IDF booth focused on Bydureon – in partially enclosed podiums at the front of the exhibit, representatives confidently demonstrated the injection procedure for Bydureon. While some doctors are uncertain about how patients will fare with once-weekly dosing, the company’s plan of sending reminder text messages to Bydureon patients seems a very smart one. As well, we thought it was clever for reps to ask attendees about how many weekly activities people already complete (e.g., TV shows, garbage collection, religious observance, ‘extracurriculars’ like music lessons, etc.). Overall, we think many patients will value the reduced time commitment with once-weekly dosing compared to once or twice-daily injections. Also featured (albeit less prominently) at Lilly’s booth were displays for Tradjenta, Humalog, Byetta, and Actos – a testament to Lilly’s broad track record of partnerships with Amylin, BI, and Takeda.
  • Medtronic: Medtronic largely reprised its booth from EASD with its new“Connecting the Stars: Integrated Technology for Better Diabetes Care” campaign. A starry blue and white billboard advertised the company’s full suite of products: Veo, Enlite, CareLink, Support & Services, Research, and Education. Half of the booth was occupied by a large, comfortable seating area and coffee table. The other side of Medtronic’s exhibit had a prominent kiosk displaying demo Veo pumps flanked by two plasma touchscreens identical to those at EASD. A display case built-in to the kiosk was devoted to comparing Medtronic’s new Enlite sensor to the previous-gen Sof-Sensor. We thought the needle size comparison was the most compelling improvement highlighted in the case. Both sensors’ inserters were also shown, demonstrating a clear improvement with the Enlite’s simpler 90-degree inserter. The rep informed us that insertion is “painless” because the Enlite inserter needle cannot be seen – we arestill surprised to hear companies making this claim, as we hear many patients still citing CGM insertion as a painful process. A corner of the display case also had a children’s book, “Lenny Explains Diabetes,” in both English and Arabic. As we understand it, Medtronic distributes these colorful books all over the world to emphasize to kids with diabetes that they can and should live full and active lives. The display case was rounded out with a packaged Mio, the company’s popular all-in-one infusion set/insertion device. As we exited the backside of the booth, a massive billboard of a sleeping child displayed, “How can I help parents to overcome their fear of hypoglycemia? Medtronic insulin pump systems feature predictive alerts which allow you to put the brake on glycemic excursions.” The corner of the poster showed a Veo pump and Enlite sensor and small font at the bottom cited the STAR-3 study (see page 172-175 of our ADA 2010 full report at and Dr. Thomas Danne’s (Hannover Medical School, Germany) recently presented and published study of the Veo in children (see page 65 of our ADA 2011 report at As a reminder, the FDA recently approved the pivotal study design of both technologies, which at the earliest, puts them on track for approval in late 2012 (Enlite) and early 2013 (Veo).
  • Merck Sharp & Dohme: MSD had a large booth, enclosed on two sides by colorful signs advertising the “most widely prescribed DPP-4 inhibitor in the world,” Januvia (sitagliptin). Janumet (sitagliptin plus metformin) and “once-daily” Januvia were also prominently advertised. The interior of the exhibit was white and clinical-looking. MSD offered A1c and lipid testing, with a long line of people waiting to be phlebotomized in one of a few available cubicles. A coffee bar enticed those who had completed their tests. Besides lipid testing, there were other hints at the combination of Januvia plus simvastatin (Juvisync) – we noted a sign displaying a track runner hurdling both a “glucose” bar and a “cholesterol” bar, with the caption: “Coming soon: a new approach to treatment.” Another smaller sign also talked about how guidelines recommend statin therapy for most patients with type 2 diabetes.
  • Novartis: Novartis had a large and very bright exhibit with an illuminated archway as its centerpiece. At one corner, signage promoted Rasilenz (aliskiren), described as the “first and only” direct renin inhibitor, specifically appropriate for controlling hypertension in type 2 patients with nephropathy. The other three corners of the booth, however, advertised Galvus (vildagliptin) and GalvusMet (vildagliptin plus metformin). One of the exhibit’s main attractions was a booth that measured attendees’ height and weight and calculated their BMI. The other draw was a video game on a large screen that challenged attendees to keep their avatars within a normal glucose range – we saw this previously at EASD 2011 in September. The video game characters sped along on a screen resembling a CGM trace, with sudden hyperglycemic spikes occurring when snacks or meals dropped from the top. One button bolused insulin while the other dispensed glucagon or glucose tablets. A charismatic British woman emceed. There was no mention of Galvus in renal impairment, an indication approved by the FDA during the conference. We will look for this at ADA!
  • Novo Nordisk: The Danish diabetes giant showcased its wide variety of products in a modern, wood-floored booth largely similar to what we saw at last year’s EASD. The major focus of the booth was Victoza and the ‘Grab Type 2 Diabetes by the Roots’ campaign – representatives gave a narrated presentation on the GLP-1 and booth-goers were encouraged to play a Victoza iPad game. Given the enthusiasm for DPP-4s at IDF, we were not surprised to see comparative Victoza marketing phrases such as, “Six times the beta cell function improvement relative to sitagliptin.” Along the side of the booth and throughout the conference hall, Novo Nordisk also had pictures of patients and signs for its Changing Diabetes campaign with words like “Inspire” “Impact” “Change” and “Innovation.” (As a reminder, the program has various worldwide initiatives toimprove access to diabetes care, raise awareness about diabetes, and study the attitudes, wishes, and needs of people with diabetes.) While walking through Novo’s booth, we noticed a bold sign above the main desk, “Hypoglycemia is the principal limitation in glycemic management.” Less hypoglycemia is of course a major benefit of Victoza relative to other therapies, and as we learned in the medical information area at the back of the booth, an advantage of insulin degludec as well. A representative gave us a PowerPoint presentation on the many benefits of the next-gen basal insulin, including its extensive clinical data, flexible dosing, and flat profile. When we specifically asked about hypoglycemia, the rep described the nocturnal benefits of degludec. Once it is approved, we’ll be interested to see how hypoglycemia fits into degludec’s marketing – we believe that a label containing a superiority claim for nocturnal hypoglycemia would be an important differentiator for many patients currently on Lantus or Levemir, although price and reimbursement will no doubt be major factors (and uncertainties) as well.
  • Roche: This maroon, orange, and white booth highlighted a plethora of new and upcoming Accu- Chek products. We first noticed the second-generation Accu-Chek Mobile, which like the first generation, is a ‘no-strip’ device that uses an internal drum of test material good for 50 fingersticks (for more information, see our EASD 2011 Exhibit Hall Report at As of IDF, Roche had launched the next-generation Accu-Chek Mobile in Denmark, Germany, Italy, UK, Switzerland, and Australia, with wider international launch still on track for next spring (including a Spanish debut to coincide with ATTD in Barcelona in early February). As a reminder, neither version of the Accu-Chek Mobile has ever been cleared in the United States, and IDF booth reps said that US launch would not come before 2013. The Accu-Chek Aviva Expert was also on display, which features a built-in bolus calculator based on the same software as the bolus calculator in the Accu-Chek Spirit pump. This software is somewhat complex to understand (see Dr. Bruce Bode’s comparison of bolus calculator algorithms during DTM 2011 at, and the user interface feels more crowded than that of Abbott’s FreeStyle InsuLinx (which also features a larger screen). However, the Aviva Expert has the advantage of a color display compared to the FreeStyle InsuLinx’s monochrome, and the Aviva Expert does not require a healthcare provider to ‘unlock’ settings like insulin sensitivity and insulin to carb ratio. Also of note, the Aviva Expert does not provide the FreeStyle InsuLinx’s option for storing a typical mealtime dose and adjusting it roughly based on meal size. While we agree with the company’s reasoning that many patients can perform these simple calculations on their own with no problems, we think that this sort of option is valuable for the large number of people with numeracy issues (and is generally convenient for any patient). Given Roche’s understated marketing strategy (the Expert was not at all a focus of the company’s EASD 2011 booth), we speculate that the company will modify the device based on initial feedback and release the second-generation Expert more widely. We expect that US launch would wait until an updated product comes out.
    • Roche has also begun a study comparing the Accu-Chek Aviva Expert to traditional blood glucose measurement in poorly controlled (A1c > 7.5%) patients with type 1 or type 2 diabetes (n=285) on MDI The primary outcome measure is A1c at 26 weeks, potentially making this study key for increasing adoption of meters with built-in bolus calculators. The study includes 33 centers in the UK and Germany, and we understand that postmarketing research is also under way in Denmark. We hope that the trial shows positive glycemic results (primary completion is expected in May 2012, according to and are curious about secondary outcomes as well – we understand that these include glycemic variability as measured by blinded CGM. Finally, reps were showing off the next-generation SmartPix scanner for meter downloads and data management (version 2.0, due out in 2012, features a simpler user interface and a “traffic-light” icon for quick interpretation of glucose values) and the Solo MicroPump (scheduled for soft launch in the Netherlands in mid-2012 and Germany in late 2012).
  • Sanofi: Decorated with the mottos “Going Beyond Together” and “Efficient and Simplified Care for a Life to Live,” Sanofi’s elegantly designed booth showcased the company’s insulins, delivery devices, and blood glucose meters. The iBGStar/BGStar corner of the booth drew attendees eager to see the meters in action, while information about Lantus, Apidra, and Amaryl was also distributed. The SoloSTAR and ClikSTAR pens were a major focus as well. Finally, flyers provided information on, the company’s new resource for teens with type 1 diabetes. Outside the exhibit hall, Sanofi’s picture-frame advertisements were similar to those we saw at EASD in Lisbon (although less prominent).

-- by Daniel Belkin, Adam Brown, Eric Chang, Joseph Shivers, and Kelly Close