Diabetes UK Professional Conference 2018

March 14-17, 2018; London, UK; Full Report – Draft

Executive Highlights

  • Diabetes UK featured abundant commentary on the latest diabetes clinical trials: Professor Simon Heller reviewed DEVOTE, arguing that there is a link between severe hypoglycemia and all-cause mortality (as seen in the DEVOTE 3 post-hoc), but acknowledging that this association will be hard to confirm via clinical trials. Dr. Rury Holman reviewed EXSCEL, emphasizing that pragmatic study design likely led to the neutral CV result. Dr. David Matthews reviewed CANVAS, and showed that lower limb amputations are likely a risk inherent to the canagliflozin molecule (though we continue to believe this risk should be manageable in the real world with proper patient selection and diligent monitoring of the feet). We also heard multiple discussions of Professor Roy Taylor’s DiRECT trial, investigating the efficacy of intense caloric restriction in type 2 diabetes remission – our sense is that the diabetes community in the UK is excited about DiRECT, although it remains to be seen how the positive results can be replicated in real-world settings.

  • Diabetes technology discussion was light at Diabetes UK, but we did appreciate two notable exhibit hall updates: (i) Diabetes UK’s impressive advocacy to get FreeStyle Libre covered nationally (done!) and now regionally (in progress – check out the “Flash Map” and letter writing campaign); and (ii) news from Dexcom’s booth that ~40% of their UK patients are now obtaining reimbursement. In sessions, we also heard nice overviews of CGM and FreeStyle Libre from Dr. Raimund Weitgasser; Dr. Pratik Choudhary on Senseonics’ Eversense (positive), and several closed loop overviews (Drs. Roman Hovorka, Helen Murphy, and Hood Thabit).

The 2018 Diabetes UK Professional Conference took place in a classically overcast and rainy London, from March 14-17. This meeting draws some of the very best minds in diabetes from across the UK and even beyond; below, you’ll find our full report from the meeting.


Table of Contents 

Detailed Discussion and Commentary

Diabetes Therapy

DEVOTE: Long Acting Insulin Therapies and Cardiovascular Outcomes

Simon Heller, PhD (University of Sheffield, UK)

The inimitable Dr. Simon Heller gave a precise tour of the DEVOTE CVOT of Novo Nordisk‘s Tresiba (insulin degludec), presented last year at ADA 2017, stating a personal belief that hypoglycemia is linked to mortality but noting that proving this link scientifically is challenging. In DEVOTE, Tresiba was shown to be non-inferior to Sanofi’s Lantus (insulin glargine) on a primary endpoint of three-point MACE (HR=0.91, 95% CI: 0.78-1.06, consistent across all components) but superior for severe hypoglycemia and nocturnal severe hypoglycemia (40% and 53% risk reductions, respectively). A follow-up pooled analysis of both arms, presented at EASD 2017, showed a strong temporal link between severe hypoglycemia and mortality. But while this seemed like a smoking gun to the audience, Dr. Heller concluded that “avoiding hypoglycemia may reduce the risk of adverse outcomes and overall death but designing studies to establish this with certainty will be challenging,” for ethical and other reasons. Indeed, the counterargument is that hypoglycemia occurs in people who are high-risk (“brittle” or “frail”) patients, more likely to experience CV events or death. That said, during Q&A Dr. Heller did admit that he personally believed in the association, with the caveat that the effect is likely multifactorial and there are probably confounding factors involved; he also advocated for the use of agents without hypoglycemia risk in older, frail individuals, rather than just raising A1c targets to lowering mortality risk. A fantastic debate at IDF 2017 drew similar conclusions on hypoglycemia and death.

  • DEVOTE 3 pooled data across both DEVOTE arms to demonstrate a significant association between CV death at any time and a prior severe hypoglycemia event (HR=2.51, 95% CI: 1.79-3.50). We note that the hazard ratios were higher 15 days following an event (HR=4.20), 30 days post-event (HR=3.66), and even at one year (HR=2.78), though the confidence intervals are consistently wide (but significant). The risk of death after severe hypoglycemia declines over time, leading Dr. Thomas Pieber, who initially presented these results, to call the temporal link suggestive rather than conclusive. But with a more than 300% increased risk of dying in the 15 days after severe hypoglycemia, at the very least these results underscore the importance of avoiding severe hypoglycemia. Moreover, we think these data should drive payers to better cover diabetes therapies that don’t cause (or cause less) hypoglycemia, as well as technology like CGM that can help avoid hypoglycemia. Despite the quality of this evidence, however, causality hasn’t been scientifically established and DEVOTE 3 is therefore hypothesis generating.

  • As a reminder, DEVOTE was presented at ADA 2017, with a follow-up at EASD 2017. After the usual phase 3 trial of Tresiba, Novo Nordisk presented a data set that was neutral for cardiovascular events (MACE). The trial was conducted after FDA requested a post hoc analysis of phase 3 data for insulin degludec, which strangely showed a higher number of events in the Tresiba arm, despite overall neutral results on MACE. DEVOTE randomized over 7,600 participants with high CV risk to once daily degludec or insulin glargine (Sanofi’s Lantus), in a treat to target (oral therapy was used), double blind, event driven trial. The primary endpoint was three-point MACE (CV death, non-fatal MI and non-fatal stroke). At baseline, mean age was 65 years, BMI 34 kg/m2, duration of diabetes 16 years, and A1c 8.4%. The trial duration lasted two years.

EXSCEL: GLP1 Analogs and Cardiovascular Outcomes

Rury Holman, MD (University of Oxford, UK)

Dr. Rury Holman gave his usual polished and professional read out of study results – in this case of the EXSCEL CVOT for once-weekly GLP-1 agonist Bydureon – explaining that adherence in the trial wasn’t ideal but also maintaining the value of trials with wide enrollment criteria. While EXSCEL proved the safety of AZ’s Bydureon (once-weekly exenatide), it narrowly missed superiority (HR=0.91, 95% CI: 0.83-1.00). In providing some of the first commentary on EXSCEL at EASD 2017, Dr. Holman echoed the opinion of many others that EXSCEL’s narrow miss on superiority was largely due to trial design: For example, the study enrolled a 27% primary prevention cohort, higher than LEADER’s 19% and SUSTAIN 6’s 17%, and the study was designed to mimic real-world circumstances more closely than other CVOTs. He also added, during Q&A, that the clumsiness of the device used in EXSCEL (complex reconstitution kits, rather than the pen, were used) were a big problem; once recognized, investigators implemented training videos – we can’t help but wonder what results would have looked like if the new Bydureon BCise autoinjector was used in EXCEL. Moreover, we wonder if the low convenience was associated with the high (43-45%) rate of premature discontinuation cited by Dr. Holman, which ultimately lessened drug exposure time to 2.4 years. Further, we were particularly intrigued by Dr. Holman’s comment during Q&A: “The ideal [study of CV outcomes] would be to take all-comers – a complete spectrum, to examine the effect of duration of diabetes and the various therapies. The problem is that it takes too long. I would like to see a sequential trial design. The companies set up a long-term trial which produces efficacy in year 1, key CV outcomes in the medium term, and in the longer term we get safety. But we couldn’t do it unless we can extend the marketing license for those companies that participate.” This sums up, so well, how we feel about CVOTs: They’ve offered so much valuable insight on the benefits of GLP-1 agonists and SGLT-2 inhibitors in lowering CV risk, but there are significant financial incentives for getting them done as quickly as possible, and it’s easier to show benefit quickly when you enroll a high-risk population. Dr. Lawrence Leiter just commented at ACC that these “crash tests,” while offering exciting results, may not be showing the long-term benefit that diabetes drugs can have when started earlier and used longer in lower-risk patients. Moreover, at poster there showed that SGLT-2 inhibitor CVOT enrollment criteria exclude ~50%-80% of people with diabetes. While we admittedly can’t wait for SGLT-2s and GLP-1s to go generic, we do see how limitations on patent protection might actually be working against patients and we’re interested to hear more on this front.

  • Dr. Holman also presented a meta-analysis of GLP-1 agonist CVOTs, supporting a class effect in reducing cardiovascular risk and all-cause mortality – all very safely, while lowering glucose and weight. Calling GLP-1s a ‘major therapeutic advance,’ Dr. Holman cited a 10% reduction in three-point MACE and a 12% reductions in all-cause mortality for the class, based on ELIXA (n=6,068), LEADER (n=9,340), SUSTAIN-6 (n=3,297), EXSCEL (n=14,752). Only LEADER (liraglutide) and SUSTAIN-6 (semaglutide) demonstrated individual superiority, but most results were in the right direction for each trial. Of late, we have been hearing more commentary on the cardioprotective class effect front for both GLP-1 agonists and SGLT-2 inhibitors, and thought leaders as prominent as Dr. Mikhail Kosiborod have come out in favor of both – though we do sense more controversy on GLP-1s than SGLT-2s.

LEADER: GLP-1 Analogs and Renal Outcomes

Steve Bain, MD (Swansea University, Wales, UK)

Dr. Steve Bain emphasized the role of liraglutide in preventing the onset of kidney disease, referencing results from the LEADER trial of Novo Nordisk’s Victoza (liraglutide). LEADER found a 22% risk reduction (95% CI: 0.67-0.92) on the composite four-point renal outcome of (i) new onset of persistent macroalbuminuria (this drove the effect), (ii) persistent doubling of serum creatinine, (iii) need for continuous renal replacement therapy, and (iv) death due to renal disease. At present, the data do not indicate a mechanism of action for liraglutide offering renal protection, be it glucose or weight lowering or another independent mechanism. Dr. Bain ended his talk by briefly considering empagliflozin and the EMPA-REG trial – indeed, empagliflozin was associated with a similar improvement in microvascular outcomes, driven by an endpoint of new or worsening nephropathy; Dr. Bain explained that empagliflozin was able to demonstrate an improvement in eGFR over time and a risk reduction across a broader range of ‘hard’ endpoints. For that reason, he believes that the SGLT-2 class will be ‘more exciting’ from a renal perspective going forward, clarifying that the 22% risk reduction seen in LEADER was not driven by a delay in “hard” endpoints but rather a delay in the onset of macroalbuminuria. Moreover, in LEADER, the risk of the composite outcome was only significantly lowered in patients with eGFR above 60 ml/min/1.73 m2 – implying that liraglutide offers prevention for those with more preserved renal function but little benefit for those with stage 3 or 4 CKD. However, the data do support the use of liraglutide in patients with impaired kidney function – a small victory in its own right, given the various challenges associated with using most drugs in patients with impaired renal function.

  • LEADER was a trial of liraglutide (Novo Nordisk’s Victoza) vs. placebo in 9,340 patients, which reported a 13% reduced risk for three-point MACE, a 15% risk reduction for all-cause mortality, and a 16% reduction on all microvascular outcomes. The lower risk of microvascular outcomes was driven by a 22% risk reduction on renal outcomes. At baseline, LEADER participants had an A1c of 8.7%, diabetes duration of 13 years, BMI of 32 kg/m2, systolic blood pressure of 136 mmHg, and an eGFR of 80 ml/min/1.73 m2. However, there was a range of renal function, with about a quarter of participants having low eGFR (<60 ml/min/1.73 m2), a quarter with proteinuria, and 12% in the high risk group (having both).

Questions and Answers

Q: We use Victoza in our dialysis population, mainly to lose weight so they can be transplanted. What do you think of this?

A: The FDA discourages liraglutide for end stage renal disease, but there is no reason not to consider it. I personally would continue.

Q: How should we be treating patients with microalbuminuria?

A: I have always been a microalbuminuria skeptic so you should be treating blood pressure and using any appropriate blood glucose lowering agent.

CANVAS: SGLT-2 Inhibitors and Cardiovascular Outcomes

David Matthews, MD (Oxford University, UK)

Dr. David Matthews elegantly discussed CANVAS results: He very much feels that the amputation signal for canagliflozin is true, going so far as to say that, “If it’s specific to the drug, then I suspect that will be the end of that particular drug.” These are strong words but somewhat challenging to refute given the seriousness of amputations in the eyes of any given patient. Still, we’re not sure there isn’t a role for canagliflozin in treatment paradigms: Anecdotal evidence supports that canagliflozin carries the best glucose-lowering efficacy of any SGLT-2 inhibitor, and some experimental evidence supports this hypothesis as well. Indeed, the majority of commentary we’ve heard on the CANVAS amputation signal has been that the risk should be manageable with careful patient selection (no prior amputations or PAD) and monitoring (careful and vigilant foot exams, both at home and in office). Dr. Matthews actually seemed to agree with this, explaining that the overall risk of amputations remains low and emphasizing that these precautions further minimize risk. Moreover, he says, “I suppose the trade-off is that you might have that risk of amputation, but you avoid a fatal heart attack,” reminding us all of the significant CV benefits offered by canagliflozin. Certainly, we eagerly await more data from the ongoing trials of SGLT-2 inhibitors (DECLARE, VERTIS CV, Dapa-CKD, Dapa-HF, EMPEROR, CREDENCE). Dr. Matthews is not willing to say, yet, whether increased risk for amputations is a class effect or specific to canagliflozin, particularly because ertugliflozin has shown a non-definitive signal for amputations. He did emphasize that, when given to patients without prior amputations, the risk/benefit is still strongly in favor of using an SGLT-2 inhibitor – but there should be a strong focus on ongoing routine foot care. Ultimately, despite increasing amputation risk nearly two-fold across both CANVAS program trials, canagliflozin also demonstrated an exciting 14% risk reduction on three-point MACE.

DiRECT Clinical Trial Outcomes at One Year

Mike Lean, MD (Glasgow University, Scotland, UK)

Dr. Mike Lean discussed the results of the groundbreaking DiRECT trial – the buzz of Diabetes UK this year, first presented at IDF 2017 in December. The slides are available from http://www.directclinicaltrial.org.uk. The idea behind DiRECT is that we should address type 2 diabetes by addressing obesity – the root cause. Not only does the DiRECT trial demonstrate that weight loss-mediated remission is possible in many patients, it also demonstrates that weight loss can be delivered and maintained out to 12 months in a real world clinical setting, using local healthcare providers. Dr. Lean commented, “We should understand type 2 diabetes as a disease process of excess body fat accumulation” – i.e., a complication of obesity. In this paradigm, weight loss is the primary goal of diabetes care, raising interesting questions of whether providers can ethically use drugs without an enthusiastic attempt at weight reduction? To be sure, the implications of the DiRECT trial for clinical care and treatment of diabetes do seem truly significant, potentially shifting focus more toward weight management than is already the case and even compelling providers to treat a diagnosis of type 2 diabetes as a medical emergency, since there is only a limited time before it’s impossible to obtain remission. Presenters noted that, as difficult as it is to lose weight, the healthcare system has an ethical responsibility to provide weight loss programs and support, even before prescribing anti-diabetic drugs, since excess obesity is the underlying cause of type 2 diabetes and the vast majority of patients have tried many times to lose weight by themselves and failed. For good reason, the big question at hand now is whether remission can be sustained. Of course, even a few years of remission has huge health and cost benefits. Moreover, Dr. Taylor was clear that if patients regain the weight, they will regain the diabetes. The team at Newcastle has focused on behavioral strategies for weight maintenance in the DiRECT trial, which will now have a three year follow up. But it does appear that people with diabetes in remission (‘post-diabetics’) will need lifetime support from the healthcare system – though we suspect this cost is nominal compared to the cost of diabetes.

  • As a reminder, the DiRECT trial (n=298) reported that 24% of participants receiving an intensive weight loss intervention (36 out of 149) achieved at least 15 kg of weight loss and 46% (68 individuals) achieved diabetes remission. Only 4% (6 of 149) of the control group achieved remission, and no one in the group achieved 15 kg of weight loss. Average weight loss for the intervention group was 14.5 kg (~30 lbs), and average weight regain at 12 months was ~2 kg (~4.4 lbs). Remission was obtained in 34% of those who lost 5-10kg, 57% of those who lost 10-15kg and 86% of those who lost over 15kg (which is equivalent to the results of bariatric surgery). This trial is already viewed as groundbreaking, as evidenced by the way it is being discussed in other presentations. Dr. Lean noted that because we can’t seem to manage type 2 diabetes, and “we are missing the main target by focusing on blood glucose.” Similarly, bariatric surgery has been shown to result in diabetes remission. Importantly, DiRECT was administered in a real world clinical setting across 49 centers, and average baseline age was 54 years, BMI was 35 kg/m2, and duration of diabetes was 3.0 years.

  • It’s notable that the intervention was carried out in a real life primary care setting with no additional staff, and patients were highly enthusiastic and motivated. The intervention, Counterweight Plus, aims for >15kg weight loss in three phases – weight loss (830kcal/day using total diet replacement shakes), stepped food reintroduction after four weeks, then weight maintenance up to 12 months. All diabetes and blood pressure medicines are discontinued on day #1. The shakes are 61% carb (although low absolute carbs) and weight maintenance is 50% carb. If patients gained more than 2kg, then there is a relapse management program. Despite the challenges of the intervention, only 20% withdrew. The population also skewed to more deprived groups, which are often thought of as more resistant to intervention.

  • DiRECT has received funding for a two and three year follow up. In future, work needs to be done to recruit high risk minorities.

Questions and Answers

Q: Do we take people off the diabetes register if they are in remission?

A: They will still be on the register and be recalled for annual checks even in remission, because the code for diabetes in remission allows for this.

Q: Why is the dropout rate low?

A: Behind the scenes we have worked on behavioral theories for change, to enhance the long-term maintenance. The trial performed better than many expected because of the pilot work.

Personal Insight: What Was It Like to Take Part in DiRECT?

David Paul, DiRECT Trial Participant

Mr. Paul, an engaging, middle-aged man offered his perspective as a DiRECT trial participant. He was able to achieve diabetes remission through the trial – but it was clearly very challenging. He spoke at length about trying to find behavioral strategies to cope with the very low energy diet. Below are some key quotes from Mr. Paul:

  • “I am so blooming grateful to Roy Taylor. I am not a diabetic anymore. I feel like a lottery winner. I gained weight as a young person. Four years ago, at the age of 52 I was diagnosed with type 2 diabetes. At that time, I was nudging 280 pounds.”

  • “The trial sounded awful, but it was a chance to get rid of diabetes. It is hard living on four sachets of dust a day. I thought the flavors were bearable at best. I found the soups horrible. I soon missed chewing on solid food. The hunger pangs were hard, but I didn’t crack. I kept telling myself that it was only 12 weeks.”

  • “I lost 10 lbs after one week, my blood pressure was down, and my blood glucose dropped from 241 mg/dl to 88 mg/dl. I lost 7 pounds in week two, five pounds in week three. Losing weight was mentally a massive boost to continue.”

  • “I couldn’t watch others eat, so I stopped having meals with my family. I realized that images of food are around us everywhere. There was no escape. I needed to keep my mind off things and pass the time. I went out in the evenings to stay away from the kitchen fridge. But I couldn’t escape food wherever I went. I realized it was down to me. My clothes stopped fitting, I began to feel cold all the time. My arm and leg strength diminished. I got ever more grumpy. My wife needs an award for staying with me. I couldn’t have a social life any more – no dinners out, no nights at the pub. Attending the weigh-ins became the highlight of the week. But I wasn’t hungry for most parts of the day. You lose perspective and an extreme diet recalibrates your attitude towards food.”

  • “After the trial I had lost 56 pounds. After a year, my weight has crept up – I’ve gained 21 pounds. But I still don’t have diabetes, and I am going to the gym and planning to cycle the length of Britain.”

  • “This may have saved my life. I would say to any doctors here today ‘Take a look at Professor Taylor’s research. Give the patients you encounter a chance.’”

DiRECT in Primary Care

Alison Barnes, RD (Newcastle University, Newcastle upon Tyne, UK)

Ms. Alison Barnes, a leading dietitian who helped develop the DiRECT program and train GP practices to deliver the intervention, explained how practitioners (mainly nurses and dieticians) were trained to perform the required medical supervision for the intervention. Four specialist dietitians from DiRECT provided support to the practice (nurse/dietitian) lead. The specialists gave input on diet, protocol, behavior change skills, support for decreasing medications, provided cover and continuity (because of staff changes during the trial) and practitioner mentoring and motivation. Only eight hours of baseline training were provided, which covered type 2 diabetes remission (many weren’t aware that this is possible), total diet replacement, and food reintroduction. After training, the intervention could be immediately delivered. After four months, extra support was given for weight loss maintenance and rescue plans. Ms. Barnes mentioned that there has been a ripple effect in these practices, beyond the study, as it has changed the skills and outlook of practice staff regarding weight loss and the management of people with diabetes.

Against All Odds: How and Why Do Patients Succeed in Dietary Diabetes Remission

Falko F. Sniehotta, PhD (Newcastle University, Newcastle upon Tyne, UK)
  • Dr. Falko Sniehotta, a professor of behavioral medicine who helped develop the DiRECT intervention, emphasized that the team chose very low energy diets because they produce greater weight loss and are not harder to follow than other diets. He reviewed how the team ran pilots and focused on learning about weight loss and weight maintenance coping strategies prior to initiating the full intervention. Dr. Sniehotta says they initially chose a very low energy diet (VLED) because it has a strong evidence base for the most weight loss. There is a linear relationship between daily calorie deficit and weight loss, but VLED doesn’t seem to be more difficult to follow than other diets, and it has a larger calorie deficit. Interestingly, the first pilot study showed that VLEDs are easier than expected because rapid weight loss is a great motivator. That said, participants also have a need for social support, and they need to develop a range of behavior regulation strategies to overcome barriers to adherence. They typically remove food from the house, avoid TV or shopping malls, plan ahead, carry healthy snacks, develop strategies for dealing with hunger (hobbies, self-talk), tell other people about their weight loss plan, find a weight loss ‘buddy’. Eventually, the team developed the Counterweight-Plus protocol (details at www.directclinicaltrial.org.uk), and Dr. Sniehotta believes that there is good reason to think that the protocol can be implemented on a nationwide scale as a game changer for the treatment of diabetes. Most people regain weight after a diet, so weight maintenance is critical to success. Weight maintenance is also less investigated than weight loss. But even after significant weight loss, most participants wanted to keep losing weight. The key to weight maintenance, Dr. Sniehotta says, is controlling behavior. The most important strategies were monitoring (food diary, weighing), planning (menus, shopping lists), finding alternative foods, and compensation (exercising or fasting before/after eating too much).

Understanding the How and Why of Remission

Roy Taylor, MD (Newcastle University, Newcastle upon Tyne, UK)

Through his work, Dr. Roy Taylor is both demonstrating that GPs can help patients lose significant weight to obtain diabetes remission and helping to elucidate the underlying pathophysiology of the disease. As we noted at last year’s Diabetes UK, Dr. Taylor has used in vivo imaging studies to show that diabetes results from excess fat accumulation in the liver and pancreas, starting a vicious cycle of metabolic stress and leading to beta cell dysfunction. The “twin-cycle hypothesis” for type 2 diabetes shows that excess obesity triggers two codependent vicious cycles of fat increase in the liver and pancreas that results in the symptoms of diabetes. In the liver, excess fat leads to hepatic insulin resistance, increased basal insulin secretion, creating more liver fat and increased fat in the blood (VLDL/triglycerides). This in turn leads to fat in the islet cells, lowering the acute insulin response, causing increased plasma glucose. The two vicious cycles feed off each other, leading to metabolic stress that impacts the function of the beta cell. The cycle can be broken if one lowers body fat to below an individual threshold. One benefit of this model is that it predicts that in diabetes we will see liver fat and blood fat, which was measured during the trial. Dr. Taylor explained that beta cells will stop working under conditions of increased metabolic stress, but for a short while the process is reversible. Unfortunately, there comes a point at which beta cells, although living, will never produce insulin again and diabetes remission is impossible. In his words, a diabetes diagnosis should be considered a medical emergency.

  • In the DiRECT trial, some participants achieved diabetes remission and some did not – even though weight loss was similar. Why did some respond to weight loss and others didn’t? Dividing the intervention group into responders (A1c <6.5% post initial weight loss) and non-responders, the two groups were not defined by weight loss. While the responders had slightly more weight loss (16.2 kg vs. 13.4 kg) and slightly better weight retention to 12 months, Dr. Taylor doesn’t think this explains the difference.

    • Responders were able to sustainably lower liver fat in the trial. Both groups started the DiRECT trial with very high liver fat of around 15%, which fell to around 4% after the initial weight loss phase. However, in the non-responders it rose again out to 12 months, and in the responders it remained flat. As Dr. Taylor explained, this implies that responders had dipped below their personal fat threshold and no longer had ‘spillover’ fat going to the liver.

    • Responders were also able to control VLDL1-triglycerides out to 12 months. But the non-responders saw their levels rise after initial weight loss to become equivalent to control.

    • There was similar levels of pancreatic fat in both groups after initial weight loss, but only the responders were able to restore first phase insulin response. This endured out to twelve months.

    • The two groups had overlapping duration of diabetes, but a significant difference of 2.6 versus 3.7 years. Otherwise, the two groups were similar at baseline for age, weight, and sex. A1c and fasting glucose were a little better in the responder group. Dr. Taylor was also investigating to see if any participants had delayed receiving their diagnosis for any reason.

Questions and Answers

Q: You need to wait longer to see if this approach is successful. In Torbay, after fifteen years, all subjects put all the weight back on.

Dr. Lean: We recognize that the duration of remission is very important. And then how long of a duration of remission is worth pursuing for clinical significance? We have spent a lot of time with psychologists on the maintenance of weight loss. There is little work in the literature on maintenance – all the work is on weight loss. We could justify continuing the intervention and monitoring if diabetes is absent.

Dr. Taylor: It’s important to separate the human factors from the biological factors. Diabetes doesn’t come back if you keep the weight off.

Q: What is the fall off of remission?

A: Diabetes will come back for sure if fat comes back.

Dr. Sattar: Even five years free of diabetes is worth it.

Q: Can we determine who the responders and non-responders will be at baseline?

A: There is a big range. But there is not a predictive index yet. We are looking at this carefully.

Q: Why not keep the patients on metformin post-diabetes?

Dr. Taylor: People who are post diabetes have much lower cardiovascular risk than even prediabetes people, so am I worried about a slightly high glucose? I say no. I would rather want to get the weight off.

Q: During the weight loss phase, what was your main motivation?

Mr. Paul: My motivation was the ticking time bomb of diabetes. It wasn’t to look good – I don’t think that’s ever going to happen!

Hot Topics: Translating DiRECT to Practice: Is the Evidence Sufficient?

Stephen Bain, MD (Swansea University, Swansea, Wales)

Dr. Stephen Bain, faced with the lonely job of trying to be a skeptic in the face of great enthusiasm about the DiRECT trial, listed as many points of concern as he could. Although most of his punches failed to land in our view, he introduced fascinating and important food for thought as we consider real-world implementation of a DiRECT-like approach for type 2 diabetes treatment and remission. Moreover, the talk set up a nice set of comments from Dr. Mike Lean, the principal investigator, who addressed the issues of recruiting and cost in the trial. Dr. Bain aptly captured the key concern with DiRECT – the unresolved question of the duration of remission. The investigators are concerned themselves, and Diabetes UK has funded a two- and three-year follow-up. Earlier at the conference, there was talk suggesting that a “post-diabetes” patient will need weight maintenance follow-up for life (what are the implications here for cost-effectiveness?). The idea that PCPs are now in the weight loss and weight maintenance business might seem strange, but Dr. Naveed Sattar (chairing the session) commented that when it comes to weight loss, “we need to do better than we have done in the past.”

  • The twin cycle theory behind the DiRECT trial introduces the notion of a personal fat threshold, determined by genetics. This means that we can’t play the “blame game” with diabetes, and in this way, DiRECT could be an important factor in reducing stigma. People at all BMI levels can be susceptible to diabetes. The twin cycle theory led to the realization that excess obesity is the root cause of diabetes. In 2011, the remission of diabetes with rapid weight loss was demonstrated, which ultimately led to the DiRECT trial.

  • While the DiRECT trial results were promising, the intervention involved very strict caloric reduction. Dr. Bain pointed out that real life eating is out of control, with massive portion sizes, energy-dense foods, and added sugar – begging the question of whether we can implement DiRECT more broadly. In the DiRECT trial, 24% of participants achieved a >15 kg weight loss, and remission of diabetes was achieved in nearly half of the intervention group. The (pre-selected) participants volunteered after being invited to take part by mail. Are they representative of the mainstream?

  • Dr. Bain asserted that recruitment was slower than other trials – implying a lack of enthusiasm – even to have diabetes cured (see Dr. Lean’s response below). From 1,510 trial invitations by mail, 423 (28%) individuals agreed to participate – the others declined or didn’t answer.

  • Participants had to remove themselves from a real-world environment in order to participate in four shakes/day – how reasonable is this to scale up? Dr. Bain was taken aback at the social isolation required.

  • The trial cost £18,000 per person in the intervention group (~$25,500). Only 2.7% of the diabetes population today is getting structured education in Wales. In this sense, the NHS simply doesn’t have the resources to allocate to this work. Each of the 23 intervention centers in the study needed eight hours of training which required four trainers – itself a scarce resource. The total cost of the trial was £2.8 million – approximately £18,000 per person in the intervention group. Patients, HCPs, and payers will need more evidence that this is financially feasible in real-world settings.

  • All similar trials, such as Look AHEAD, have shown a rebound in weight and A1c after the trial. Only long-term data can tell us if this approach will be of value.

Response from Dr. Mike Lean, Principal Investigator

  • Recruitment: A 28% response to a mailed clinical trial invitation is actually stellar! Even 10% would have been good. People were in fact queuing up to take part. Others wanted to start immediately and not even wait for the trial.

  • Cost: Only 10% of the trial cost was intervention; the rest was the research aspect, as there was lots of MRI scanning in Newcastle and so forth. The four expert trainers were used to keep the trial running, and wouldn’t fully translate into ongoing NHS costs. Diabetes UK has also funded a detailed economic analysis (we are so eager to see this). Dr. Lean hypothesized, at this stage, that the intervention might come out to be about £800 per patient (~$1,100).

  • Duration of effect: It’s true that there has been weight regain in every trial so far. So, Dr. Lean asserted, the DiRECT team has worked hard to create a better longer-term program for weight maintenance. He emphasized that it’s definitely a lot better than Look AHEAD, with an aim to have >20% in remission after two and three years.

Questions and Answers

Q: We have got really nihilistic about weight loss. We take the view that unless it’s bariatric surgery, people can’t lose weight. This trial has shown us that people can. And if diabetes goes away it’s really powerful. So actually, people can lose weight if they are motivated and supported to do it.

Dr. Sattar: This trial reminds us that we need to do better than we have done in the past.

HFpEF in Diabetes: What Is It and How to Recognize It?

Martin Cowie, MD (Imperial College, London, England)

London’s Dr. Martin Cowie gave an overview of heart failure with preserved ejection fraction (HFpEF) as a complication of diabetes. As he described, diabetes patients who present with the symptoms of heart failure (e.g. tiredness, swollen ankles) could have HFpEF, which is closely linked to the metabolic syndrome. In HFpEF, ejection fraction is normal and there are no structural abnormalities of the heart. It can be diagnosed using a BNP test and an echo. HFpEF has been gaining more attention in the diabetes field of late, because of the connection to insulin resistance and obesity; Dr. Mikhail Kosiborod spoke to this at CODHy 2018, highlighting HFpEF as an unmet need since the condition is less responsive to traditional myocardium-targeting treatments (which follows, since there are no structural abnormalities of the heart). There is excitement brewing that SGLT-2 inhibitors may be an effective treatment for HFpEF, and this is currently being investigated in the EMPEROR HF-Preserved trial for Lilly/BI’s Jardiance (empagliflozin), expected to complete June 2020.

  • Most practitioners are familiar with HFrEF (heart failure with reduced ejection fraction), Dr. Cowie explained. This CV complication presents with “typical” heart failure symptoms like tiredness or swollen ankles. But if the ejection fraction is normal or near normal, then we may have HFpEF. To diagnose HFpEF we need to exclude a plethora of structural heart problems (by echo), which is another feasibility challenge in the real world. If a patient complains of heart failure symptoms, then an exam, a BNP test, and an echo should help diagnose most patients. The BNP (serum natriuretic peptide) test is the single most useful blood test.

  • According to Dr. Cowie, HFpEF accounts for about half of all heart failure, is slightly more common in women, and has a slightly better prognosis. The common comorbidities include lung disease and diabetes.

  • The only treatment for HFpEF at present is a diuretic. This is in contrast to HFrEF where there is a long list of options. But typically, statins and ACE inhibitors/beta blockers are prescribed. Again, we’re excited about SGLT-2 inhibitors for this potential application, although we are still at least two years away from data that could demonstrate an SGLT-2’s efficacy in heart failure with preserved ejection fraction.

Questions and Answers

Q: Didn’t we miss something? Shouldn’t weight loss be the treatment for HFpEF?

A: Weight loss is fantastic. We need solutions to help achieve it.

Epidemiology: The Burden of Cardiovascular Disease in People with Diabetes

Martin Rutter, MD (University of Manchester, UK)

Dr. Martin Rutter gave an elegant presentation that set out the case for simple, comprehensive primary prevention of CV disease for people with type 2 diabetes. This includes smoking cessation, lifestyle modification, cardioprotective glucose-lowering therapy, blood pressure management, and aggressive lipid-lowering. As he described it, “holistic personalized care for people with diabetes that doesn’t just focus on glucose.” Dr. Rutter also noted an important role for digital health – decision support systems for PCPs – and a place for community diabetes specialist nurses. We appreciated the push past a glucose-centric view of diabetes management, especially now that we have anti-hyperglycemic agents indicated for CV risk reduction (Novo Nordisk’s GLP-1 Victoza and Lilly/BI’s SGLT-2 Jardiance), so we can really start to implement outcomes-based medicine in the real world.

  • CV disease is the leading cause (37%) of death in type 2 diabetes, occurring five-six years younger than in the general population. In the United Kingdom, there are 3.8 million people with diabetes, and 90% have type 2 diabetes. Ethnic minorities with diabetes have an even more reduced life expectancy, for unknown reasons.

  • Primary CV disease prevention in type 2 diabetes is of key importance, and Dr. Rutter advocated for a combination of smoking cessation, lifestyle intervention, and pharmacotherapy. Stopping smoking carries a slight weight increase and glucose elevation, but there is a more meaningful reduction in CV risk. The Look AHEAD trial was underpowered and had other weaknesses, but an overall 10% weight reduction led to a 21% reduction in the primary composite CV outcome – so lifestyle does work! Dr. Rutter outlined drugs that impact CV risk, including TZD pioglitazone, SGLT-2 inhibitors, and GLP-1 agonists. Although an aggressive glucose-lowering strategy in ACCORD was associated with greater mortality, hypoglycemia, and weight gain, it seems that it was how the drugs were used that was the source of the problem, not the aim of lower blood glucose targets itself.

  • Dr. Rutter underscored that management of blood pressure in type 2 diabetes is also paramount. In ACCORD, the blood pressure arm had a negative result, but it was underpowered and there was a significant reduction in stroke. It seems that the best approach is to select a personalized target that is more or less stringent depending on the patient situation – Dr. Rutter presented a clinical framework. He also suggested self-testing blood pressure in a room without a healthcare practitioner. Notably, AHA/ACC released updated blood pressure guidelines late last year, strengthening the recommendation for ≤130/80 mmHg in patients with diabetes. ADA has yet to follow suit, and still recommends ≤140 mmHg systolic in its 2018 Standards of Care.

  • Aggressive management of lipids also seems to be warranted, and it’s safe. Dr. Rutter assured that the benefits of statins are well-established. Trials such as PROVE-IT TIMI 22 and Steno-2 highlight the benefit of high-dose statin therapy in diabetes across most CV endpoints. In Steno-2, the number needed to treat to prevent one event or one death was only five patients over eight years – a 20% reduction in risk. With a multifactorial intervention (glucose-lowering, blood pressure-lowering, and lipid-lowering simultaneously), the median survival was eight years longer!

Questions and Answers

Q: What is the best mechanism to make further gains in reducing CV disease?

Dr. Rutter: The answer has to be new therapies, new technology, bringing everyone up to the same standard. Community nurses are going to be the catalyst for change in primary care.

New Clinical Trials: The Promise of Added Benefit

David Preiss, PhD (University of Oxford, UK)

Cardiologist Dr. David Preiss presented a pragmatist’s view on CV disease prevention in type 2 diabetes. Like many cardiologists, he doesn’t strongly believe in glucose-lowering as a priority, focusing instead on managing blood pressure, lipids, and anti-thrombotics. In this lecture, he reviewed new approaches and CV risk mitigation strategies, suggesting that lack of treatment options is not the problem today. Rather, it’s an issue of clinical inertia, less than ideal adherence, benefit/harm tradeoffs, and cost-effectiveness. Dr. Preiss argued that we need to deliver these interventions at mass scale, and that we need to improve real-world medication adherence as well, noting that “it only works if you take it – and 50% of drugs don’t get taken.”

  • Dr. Preiss reviewed four main areas of pharmacotherapy for prevention of CV disease in type 2 diabetes: (i) glucose-lowering, (ii) anti-thrombotic, (iii) blood pressure-lowering, and (iv) LDL-lowering. Evidence abounds to show that lowering LDL and blood pressure prevents CV morbidity/mortality. For glucose-lowering, however, the effect may depend on the particular drug, but in the other areas, it’s likely that the mechanism of the drug doesn’t matter. We’ve certainly heard similar commentary from diabetes thought leaders – that HCPs need to pay attention to how an agent is lowering blood glucose, not just the magnitude of A1c reduction. GLP-1 agonists like Victoza (Novo Nordisk’s liraglutide) and SGLT-2 inhibitors like Jardiance (Lilly/BI’s empagliflozin) and Invokana (J&J’s canagliflozin) seem to exert their CV benefit independent of their glucose-lowering effects, which also invites cardiologists officially into the realm of diabetes care.

    • On GLP-1 agonists, Dr. Preiss suggested that there is a cardioprotective class effect. He implied that there is no clear difference between the drugs – even with semaglutide (Novo Nordisk’s Ozempic). Instead, he attributed positive vs. neutral trial results to variations in trial design, an opinion we’ve heard several times in response to AZ’s EXSCEL for Bydureon (exenatide) – many thought leaders attribute the neutral finding here to a “pragmatic study design” (no run-in period to exclude patients with low medication adherence, wide range of concomitant drugs allowed, single-dose reconstitution kits for Bydureon instead of the pen or new autoinjector, etc.). Indeed, Drs. Rury Holman and Angelyn Bethel have each presented meta-analyses highlighting CV benefit as a class effect across exenatide, liraglutide, and semaglutide. Dr. Preiss further suggested that longer trials are needed.

    • On SGLT-2 inhibitors, Dr. Preiss also emphasized the similarities between the two agents that have reported CVOT results to-date (J&J’s canagliflozin, Lilly/BI’s empagliflozin). The DECLARE CVOT for dapagliflozin (AZ’s Farxiga) is expected to complete in July 2018 and to report results in 2H18. DECLARE will further illuminate whether or not cardioprotection – and heart failure benefit – may be class effects of SGLT-2s. Merck/Pfizer also have a CVOT ongoing for their recently approved/launched ertugliflozin (Steglatro), and that study (VERTIS CV) is expected to complete in October 2019.

  • The COMPASS trial (n=10,341 people with diabetes) showcased the anti-thrombotic rivaroxaban (a factor Xa inhibitor), plus aspirin. The combination arm saw CV risk reduction of 24% vs. aspirin alone. For this reason, the dual approach may become standard of care, according to Dr. Preiss. The ASCEND trial is studying aspirin vs. placebo for n=15,000 participants entirely remotely, and will report at ADA 2018.

  • Cholesterol management has been sensationally influenced by PCSK9 inhibitors. Evolocumab (Amgen’s Repatha) lowers LDL by ~60% and MACE by ~18%, although it is expensive, Dr. Preiss reviewed. Alirocumab (Sanofi/Regeneron’s Praluent) just demonstrated CV risk reduction of 15% in ODYSSEY Outcomes (reported at ACC in Orlando), and is also a very attractive (but expensive) drug. Inclisiran is a next-generation PCSK9 inhibitor that is a once per six-month injection (still in clinical development). Thought leaders have called for more PCSK9 utilization in diabetes care, though across the board, experts acknowledge the high cost/poor reimbursement surrounding this therapy class. We’re hopeful that with two positive CVOTs now on the books (FOURIER for Repatha + ODYSSEY for Praluent), payers will see the potential cost-savings of these very effective, efficient lipid-lowering drugs. Amgen has suggested that prior authorization requirements have already started to relax somewhat since Repatha received a CV indication from FDA based on FOURIER results.

Questions and Answers

Q: What are the CV benefits of SGLT-2s?

Dr. Preiss: The combination of blood pressure-lowering and effects in the kidney, which have knock on effects, seem to be the underlying cause.

Comment: We are getting too carried away with all the new drugs that are really expensive and not cost-effective.

Dr. Naveed Sattar: It’s true that doing the simple things really well is going to have the biggest impact. [Editor’s note: Dr. Sattar argued at ADA 2016 that PCSK9 inhibitors are not a practical solution for population-level lipid control, due to their extremely high cost. We wonder how/when this will start to change, and if it already has, with more outcomes data hinting at long-term cost-savings.]

Comment: We must use the latest pharmacotherapy because it can pay for itself down the road.

Dr. Preiss: For people with type 2 diabetes, we are looking more and more at how we can use technology to influence self-actors. Lifestyle change has to be such a large part of what we do.

Making Cardiovascular Disease Risk Reduction Happen in Primary Care

Tony Willis, MD (Clinical Director for Diabetes, North West London)

Dr. Tony Willis gave one of the most inspirational “clinical perspective” talks we’ve heard in a long time. He’s responsible for diabetes in North West London, and over the last five years, has initiated a program to “flatten the cost curve” by providing better care and reducing complications. He is clearly an exceptional change agent, and has managed to prove that a number of techniques can be used effectively to improve results. These tools include standardized care, integrated care, dashboards, and the use of technology at the point of care.

  • The statistics for primary care of people with diabetes in North West London (NWL) are quite shocking. There are ~143,000 people with diabetes in NWL who account for 41% of hospital admissions and 63% (!) of bed days. These patients account for 22% of the entire NWL health budget (diabetes cost is £598 million). Future projections indicate that a further 377 beds will be required by 2028 if nothing is done. On the same basis, the UK as a whole will need another 22 hospitals in 10 years – just to handle diabetes inpatients. The current trajectory of rising prevalence of diabetes (and its complications, leading to more hospitalizations) is clearly unsustainable.

  • Since total diabetes cost was projected to increase by £100 million over the next 10 years, Dr. Willis’ team decided to “flatten the cost curve” by introducing a disciplined transformation plan to improve standards of care and reduce complications. The team introduced new standards – nine key care processes, three treatment targets, care planning, and structured education. They also offered psychological support, integrated care, and digital interventions for people with diabetes. The measures also included early intensification of therapy and broader adoption of the National Diabetes Prevention Program (NDPP) – we love this focus on intervening earlier, whether in prediabetes or soon after diabetes diagnosis, with effective lifestyle strategies or pharmacotherapy. What was perhaps most striking and impressive to us was the highly structured approach, and the reliance on dashboards to monitor performance across the region.

  • The results after only 12-24 months have been encouraging. For example, in one test area, there was a 10% improvement in treatment targets and a 3% increase in patients at target A1c. Average A1c decreases were modest, but significant at a public health level. Dr. Willis reported on a range of outcomes, including 24,000 more people receiving the nine key care processes, 4,300 reaching A1c target, and 11,000 taking part in the NDPP. Remember that Dr. Willis’ team is still ramping up, and the diabetes population overall is 143,000.

  • Dr. Willis outlined his key ingredients for success: patient empowerment, clinical education, networks, dashboards, contracts, clinical systems optimization, and clinical guidelines. This is a significant shopping list, and to execute it properly required tireless leadership, tenacity, and change management. Dr. Willis noted that it is hard to get the decentralized funding bodies (CCGs) to invest for longer-term benefits.

  • Digital initiatives were critical to improvement. Technology enabled best practices, clinical guidance at the point of care, pre-written letters, care plans, an integrated care record across healthcare, mental health support, and community/social care.

  • A pilot study of self-care apps was successful. The supported apps were Oviva, OurPath, and Changing Health. The team studied n=430 people with type 2 diabetes, and the three apps provided coaching, motivation, nutritional support, and activity tracking. Feedback from patients was very positive and the most deprived deciles were over-represented – contrary to popular thinking. The trial (results aggregated across all three apps) delivered a 2.5 kg (~5.5 lbs) weight reduction along with modest but significant A1c and blood pressure reduction.

  • Moving forward, Dr. Willis wants to reach more of his population by introducing a single point of referral call center that triages patients and initiates lifelong support via technology. His team intends to layer on a digital personalized self-care support pathway involving email, web, and video.

Questions and Answers

Q: What has been the biggest ingredient in improving primary care?

Dr. Willis: Mainly the digital support that we give clinicians has been the biggest factor. This helps in reducing the variability.

Q: How do we improve diet and reduce obesity?

Dr. Willis: Public Health England is getting better. We need to change the food environment in our country.

Debate: In Light of Real-World Evidence Should SGLT-2 Inhibitors Now Be Initiated as First Choice Oral Therapy?

Kamlesh Khunti, MD (University of Leicester, England); Philip Newland-Jones (University Hospitals Southampton NHS Foundation Trust, Southampton, UK)

Dr. Kamlesh Khunti and Mr. Philip Newland-Jones debated the prospect of SGLT-2 inhibitors as first-line therapy in type 2 diabetes, taking the pro and con stances, respectively. Dr. Khunti’s primary argument appeared to be that patients want to live longer without complications, and SGLT-2s will do a better job of this than metformin. We noticed that he didn’t talk about weight loss in any detail or the legacy effects of early intervention. Mr. Newland-Jones’ argument appeared to be that for first line therapy, metformin is extremely cost-effective, safe, and easy.

  • SGLT-2s have an A1c impact of 0.6%-1.0%, and they offer 4-7 lbs weight loss. Beyond this, they also improve insulin resistance and beta cell function. Dr. Khunti emphasized that we need to correct the underlying pathophysiology of hyperglycemia, not just treat the glucose – and for this, it may be better to start patients on SGLT-2 treatment as soon as possible.

  • Since CV disease is the biggest killer of people with diabetes, we need a drug that is cardioprotective, and SGLT-2 inhibitors have shown remarkable CV benefit. If you are 60 years-old and have both diabetes and CV disease, you lose 12 years of life expectancy. Enter SGLT-2s, which provide cardio and renal protection, alongside clinically-meaningful blood pressure-lowering.

  • Dr. Khunti turned to empirical evidence: UKPDS showed a 39% CV risk reduction with metformin after 15 years of follow-up, but UKPDS was really a primary prevention trial, while SGLT-2s have demonstrated CV efficacy in secondary prevention. There is a huge difference between modern CVOT trial design and UKPDS, Dr. Khunti explained – patients are older, sicker, already on a range of anti-hyperglycemic therapies. (The objection that these effects are on top of metformin don’t bear weight according to the subgroup analyses). EMPA-REG and CANVAS showed ~14% risk reduction for the primary outcome of three-point MACE (non-fatal MI, non-fatal stroke, CV death) with empagliflozin (Jardiance) and canagliflozin (Invokana) vs. placebo. These CVOTs also found significant risk reduction for dialysis and many other favorable renal outcomes. Real-world analyses like CVD-REAL and EASEL have also reported lower rates of heart failure hospitalization and all-cause death with SGLT-2s vs. other glucose-lowering drugs, which is a good sign for the translation of CVOT results into real clinical settings.

  • Overall, Dr. Khunti established SGLT-2s as safe. Genital infections – well, they only occur once, and patients can get medicine direct from the pharmacy. CANVAS did show a signal for lower limb amputations, but the same signals haven’t been seen with other members of the class. Moreover, we continue to believe that this risk should be manageable in the real world with careful patient selection (e.g. don’t start Invokana in someone with peripheral artery disease), strong education, and diligent monitoring of the feet (i.e. at home and at every in-person visit between patient/care team). DKA is a severe side-effect requiring hospitalization, but it is rare (~70 per 100,000 patient-years).

  • NICE has assessed SGLT-2 inhibitors as £20,000 per QALY which is below the threshold for use in the NHS. Treatment costs for diabetes account for only 20% of the total – it’s the complications that cost money and we need to reduce them. That’s why NICE approved SGLT-2s. This was a very compelling argument by Dr. Khunti, in our view.

  • Mr. Newland-Jones clarified that the question is not whether SGLT2s are good medicine, but should they be used as first-line therapy. He maintained that we should use metformin first on the basis of cost and benefit. There are a range of options for second-line therapy. DPP-4s are creeping up as second-line therapy because they are very safe, he explained, while SGLT-2 does have the specter of DKA.

  • He continued, metformin is a potent medicine, lowering A1c by 1.3% when used as monotherapy, which is equivalent to many alternatives, including SGLT-2s. While we don’t see many metformin monotherapy or head-to-head trials these days, there was little difference between exenatide, pioglitazone, and metformin in DURATION-4. Mr. Newland-Jones added that pioglitazone (a TZD) is somewhat cardioprotective and its side-effect profile is only problematic in patients with pre-existing heart failure.

  • As expected, Mr. Newland-Jones focused on cost. There’s no way around it – SGLT-2s are expensive drugs. He cited the monthly cost of an SGLT-2 inhibitor at £36.59; for metformin, this is only £5.54. That means we could treat six people on metformin for only one on SGLT-2. Alternatively, a patient taking metformin instead of an SGLT-2 could get a new bike each year, plus extra test strips for structured testing. SGLT-2s are a good medicine, Mr. Newland-Jones acknowledged, but the financial constraints of the NHS at the moment mean that we couldn’t actually afford it.

Questions and Answers

Q: Do we have to watch out for euglycemic DKA with SGLT-2s?

A: The randomized controlled trials didn’t show differences in DKA, but we do see it in the real world, particularly for patients taking insulin or other therapies. We need to warn them about inter-current illness and have sick day rules.

Q: Do we give a drug with a rare and serious side-effect or a drug with a very common and inconvenient side-effect (diarrhea). These GI side-effects can even emerge after many years.

A: We should take metformin with meals and titrate slowly. Be up front with the patient about the side-effects, and have patients drop the dose down if they experience any diarrhea.

A: We know that lots of patients get common side-effects. But in SGLT-2 trials, less than 2% actually discontinued because of side-effects.

Early Intervention Saves Lives in Diabetes/Kidney Disease: Insights from the Steno-2 Trial

Maarten Taal, MD (University of Nottingham, England)

Dr. Maarten Tall from Nottingham outlined a compelling case for early intervention in type 2 diabetes, drawing on a study organized by the Steno Diabetes Center in Denmark. The Steno-2 trial is a classic long-term outcomes trial, similar to DCCT and UKPDS in the sense that we see a legacy effect of intensive therapy even after many years since the intervention. Steno-2 trial focused on early multi-factorial intervention to optimize multiple risk factors – glucose, blood pressure, and lipids. Early, aggressive, multifactorial treatment essentially halved the risk of macro and microvascular complications, increasing survival. This was despite the failure of the intensive arm to reach treatment goals during the eight-year intervention period. The benefits persisted out to 21 years of follow-up, at which point the intensive group had gained a median extra 7.9 years of life. The lessons should be obvious, said Dr. Taal – use combination therapy early and intensively.

  • Diabetes is the most common cause of end-stage renal disease in the UK. Almost one-third of patients on dialysis have diabetes. Also, people with diabetes do worse on dialysis than other patients – they have double the risk of death. We don’t have any good treatment options for end-stage renal disease, apart from a transplant, so it would be much better if we could intervene earlier. This was the idea behind the Steno-2 trial, which Dr. Taal reviewed in his talk. We were also reminded of remarks by Dr. Mark Molitch at ENDO 2018 – he stressed that the main goal of renal care in diabetes should be to slow (or stop) progression to end-stage renal disease.

  • As background, Steno-2 began in 1993 to assess the impact of early intensive treatment of multiple CV risk factors. The trial randomized n=160 participants to standard or intensive treatment with a follow-up of 7.8 years. The “intensive” group aimed for blood pressure <140/85 mmHg, A1c ≤6.5%, and total cholesterol ≤190 mg/dl. Arguably, by today’s standards, this intervention wasn’t particularly intensive. There was a lifestyle component to the intensive treatment, aimed at lowering dietary fat, increasing exercise, and supporting smoking cessation. At baseline, participants were 55 years-old, had a mean blood pressure of 149/86 mmHg, mean A1c of 8.8%, and mean cholesterol at 223 mg/dl. Multifaceted management worked well, decreasing risk for microvascular outcomes by 60%, despite minimal treatment differences on A1c (compared to the control group, the intensive group lowered A1c by <1%, ending up at 7.9%). Blood pressure ended up at 135/75 mmHg, cholesterol at 160 mg/dl on average.

  • After 5.5 years of additional follow-up (13 years from randomization), the risk of death in the intensive group was still around half – even though the two groups had more or less converged on A1c, blood pressure, and LDL. A1c after 5.5 more years was around 8%, blood pressure was about 145/75 mmHg, and cholesterol was around 150 mg/dl. Dramatically, after 21.2 years, eight years of intensive therapy was associated with a median increase in survival of 7.9 years; this data was presented by Dr. Jens Øligaard at ENDO 2017. Intensive, multifactorial treatment reduced risk for all-cause mortality by 45% vs. conventional treatment, while the relative risk reduction for CV death was 62%. The median time without a CV event increased by 8.1 years. All the microvascular value was preserved. In other words, the benefits seen at the end of randomization were maintained – a legacy effect.

  • eGFR declined 22% more slowly over the 21-year period in the intensive control arm, leading to a 40% relative risk reduction for the composite renal endpoint. That said, Dr. Taal underscored that it took some time to see this benefit on renal outcomes.

Harnessing the Power of the Kidney with SGLT-2 Inhibitors to Improve Outcomes in Chronic Kidney Disease and Diabetes

Patrick Mark, PhD (University of Glasgow, Scotland, UK)

As Dr. Patrick Mark laid out, we don’t have too many options for treating kidney disease in people with diabetes, and diabetes-related nephropathy remains an area of high unmet need. Several candidates are in development for a nephropathy indication, and SGLT-2 inhibitors in particular are generating distinct excitement in the field. Dr. Mark pointed out that SGLT-2s were associated with ~40% lower risk of renal events in major diabetes CVOTs, including EMPA-REG OUTCOME for Lilly/BI’s Jardiance (empagliflozin) and CANVAS for J&J’s Invokana (canagliflozin). In his view (and ours), this data is highly promising; the renal benefits in these studies were a very happy surprise, the icing on the cake of CV efficacy. Dr. Mark further explained that these outcomes have been positive for all different stages of kidney disease, although to-date, SGLT-2s have only been evaluated in people at high risk for CV disease. The next step is to assess their renal protective properties in people who don’t have diabetes, and trials are already underway. AZ’s Dapa-CKD study will enroll an estimated ~4,000 patients with chronic kidney disease (CKD), with or without diabetes, randomizing them to SGLT-2 Farxiga (dapagliflozin) or placebo for ~four years. The study was initiated in February 2017, and it’s expected to complete in November 2020. Lilly/BI have announced plans for a dedicated renal outcomes study of Jardiance in people with CKD with or without diabetes, and Lilly management recently confirmed that this is on track to begin sometime this year.

  • According to Dr. Mark, renal outcomes for people with diabetes have been improving over the last few decades, with the exception of end-stage renal disease. At ENDO last week, Dr. Mark Molitch described a favorable drop in the number of diabetes patients progressing to end-stage renal disease starting in the mid-1990s, which is now starting to plateau. Despite a plethora of drug candidates (see our competitive landscape for the full list, and write us if you notice anything missing!), there hasn’t been anything new for the treatment of diabetes-related kidney complications for a while (since 2000). However, Dr. Mark again emphasized great excitement in the nephrology community about SGLT-2 inhibitors. This therapy class is gaining favor among cardiologists as well, since empagliflozin and canagliflozin both reduced risk for three-point MACE in their respective CVOTs and also demonstrated a profound heart failure benefit.

TZDs: Where Are We Now?

Clifford Bailey, PhD (Aston University, Birmingham, UK)

Dr. Clifford Bailey provided a comprehensive review on the TZD therapy class, which he described as quite effective but plagued with a checkered and controversial history of safety concerns. He emphasized that TZDs improve insulin sensitivity (very few drugs do), have an overall positive CV impact (as seen in the IRIS trial), but also come with side-effects of edema (which affects heart failure) and some weight gain. Dr. Bailey suggested that questions still remain over bladder cancer and bone fractures, though the dominant commentary we’ve heard from thought leaders recently is that these particular safety concerns should be put to bed. Pioglitazone, now generic, is remarkably inexpensive and is currently recommended by NICE as second-line therapy after metformin. In our view, we’d much rather see patients go on TZD pioglitazone vs. a sulfonylurea as the generic drug of choice (SUs are associated with weight gain, significant hypoglycemia risk, long-term beta cell burnout, and possible CV harm). Dr. Bailey showed that compared to a sulfonylurea, pioglitazone might lower A1c more slowly, but the A1c-lowering effect is more durable over time. Notably, Dr. Ralph DeFronzo has chosen pioglitazone as part of his “Fab Four” diabetes drugs, alongside a GLP-1 agonist, an SGLT-2 inhibitor, and metformin. The “Fab Four” is designed to address the pathophysiology of diabetes, not just the symptom (glucose). Moreover, in his talk on the “Fab Four” at CODHy, Dr. DeFronzo denounced sulfonylureas and also criticized an over-reliance on metformin.

Questions and Answers

Q: This is a class of drugs that have been vilified. All the guidelines for NAFLD have recommended pioglitazone. Our problem is that we don’t know how to identify the 50% that respond.

A: Yes, I agree. [Editor’s note: Drs. Jay Skyler and Robert Eckel also expressed excitement about the potential for pioglitazone in NASH at CMHC last year.]

NASH: When Do I Refer? What Should I Measure?

Quentin Anstee, PhD (Newcastle University, Newcastle on Tyne, England)

Dr. Quentin Anstee discussed NAFLD (non-alcoholic fatty liver disease) and NASH, which are highly prevalent conditions strongly linked to metabolic syndrome. There is immense overlap between type 2 diabetes and NASH, and yet NASH is hard to diagnose and even harder to treat, since no therapies have been FDA-approved to-date. Dr. Anstee recommended that all patients with type 2 diabetes be screened for NASH, regardless of liver enzyme levels. Right now, liver disease in type 2 diabetes typically gets picked up as an incidental finding, after the condition has already progressed quite far. The FIB-4 calculation, based on simple biomarkers, followed by a non-invasive scan can rule out patients that don’t have significant disease. The standard of care for precise diagnosis of NASH is still an expensive, invasive liver biopsy, but we appreciated Dr. Anstee’s practical strategy to get the ball rolling on improved diagnosis.

  • According to Dr. Anstee, of all the diseases of organs or systems (including the heart, brain, circulatory, endocrine) in the UK, only liver disease has increasing mortality rates. The hypothesis is that the root cause is obesity.

  • NAFLD is extremely common and is often comorbid with diabetes. It occurs in 25% of normal weight patients, two-thirds of people with overweight, and 94% of people with obesity. Dr. Anstee also described how 20%-30% of the general population has NAFLD, and how 3%-16% of people with NAFLD have NASH. Progression rates to more advanced stages are high, and the leading risk factor is diabetes. Over the longer term, it is the amount of fibrosis in the liver that is the strongest predictor of mortality. People with NAFLD die with increased rates of CV disease and cancer, not just from liver disease.

  • The ratio of AST to ALT liver enzymes can illuminate disease progression, and Dr. Anstee recommended the FIB-4 calculation to assess fibrosis. Although routine clinical biochemistry doesn’t really help diagnosis, we need to find better ways than a liver biopsy. Liver enzymes are normal in up to 80% of NAFLD patients. AST/ALT >1 with the presence of low platelets suggests NASH or cirrhosis. There are no lab tests for steatohepatitis. There are more sophisticated markers for fibrosis (such as the ELF test), but the FIB-4 test is a good way to exclude advanced fibrosis using simple lab tests, before a patient is referred to secondary care. At that stage, transient elastography (by ultrasound) is a non-invasive technique that can triage patients before a liver biopsy is needed.

Questions and Answers

Q: 80%-90% of our patients have NAFLD. Isn’t the ALT/AST ratio good enough?

A: The FIB-4 score is really easy to calculate; there is a website. That’s as easy as the ratio and a little bit better.

Banting Memorial Lecture: Precise Diagnosis of the Subtype of Diabetes is Key for Appropriate Treatment of Patients

Andrew Hattersley, PhD (University of Exeter, England)

Dr. Andrew Hattersley created a group at Exeter that is a world leader in the genetics of diabetes. Delivering this year’s prestigious Banting Memorial Lecture, Dr. Hattersley focused on subtypes of diabetes. Initially, subtypes were discovered as genetic variants, but more recently, patient outcome data has created “utilitarian” subtypes that help predict how a particular patient will respond to certain types of therapy. In fact, differentiating appropriate therapy and outcomes is precisely the value of having a subtype. In type 1 diabetes, there are ~60 genetic subtypes known at present. In type 2, donated patient outcome data has led to an understanding that some drugs work better in some subtypes than others. For example, TZDs work better than sulfonylureas in female patients with obesity, and SGLT-2s work better than DPP-4s in patients with higher starting A1c, more obesity, and better kidney function. This has strong implications for personalized medicine, leading Dr. Hattersley to state, “in type 2 diabetes, you ain’t seen nothing yet.” See below for additional takeaways from this outstanding lecture:

  • MODY was clinically defined as a genetic type of diabetes in the 1980s. It is non-insulin dependent, it’s not type 1, and it has a dominant inheritance. It represents only 3% of diabetes diagnosed under 30 years. It was linked to genetic changes in the glucokinase and HNF1A, HNF4A, HNF1B genes. The different subtypes have different clinical features, which helps provide the right treatment to the right patients. Diagnosing MODY is difficult, but all the clinical signs were combined in a probabilistic model – the MODY probability calculator which is available today.

  • Neonatal diabetes (diabetes which occurs before six months of age) is not type 1 diabetes either. It is monogenic, caused by mutations in the potassium channel. It responds well to high-dose sulfonylureas and 90% of patients can stop insulin. Samples come from all over the world for genetic testing at Exeter. There are 24 subtypes of neonatal diabetes, of which 50% can be treated with a sulfonylurea.

  • Theresa May (UK Prime Minister) was incorrectly diagnosed as having type 2 diabetes and was given the wrong treatment, which didn’t work. Only 58% of patients with type 1 diagnosed over 30 were given insulin at diagnosis. It’s hard for the healthcare system because <4% of diabetes is type 1 for the over 30s. So, we should measure C-peptide as a matter of routine.

  • In type 1 diabetes, there are two different immune signatures, which present at different ages. Subtype A are diagnosed under seven years and those diagnosed over 12 years have subtype B. Subtype B still produce a little bit of insulin, while subtype A have very low C-peptide. This matters when testing immune therapies – and there may be a differentiation of treatment one day.

  • In type 2 diabetes, the age of diagnosis is useful for subtyping because it predicts glycemic deterioration. Those diagnosed over age 65 have a much slower deterioration in beta cell function than those diagnosed under 50. In a few years, this implies a completely different drug regimen.

Diabetes Technology

What’s New? CGM and Flash Glucose Monitoring

Raimund Weitgasser, MD (Paracelsus Private Medical University Salzburg, Austria)

CGM and Flash systems continue to improve – the best sensors are now at ~10% MARD, with reduced/eliminated need for calibration and approval for making insulin dosing decisions. They benefit a broad spectrum of patients, including those using MDI, those at higher risk for hypoglycemia, in pregnancy, and for children. Additionally, CGM is having a positive impact on the trials for new therapeutic agents.

  • Dr. Weitgasser presented a rapid-fire list of clinical trial results related to CGM. These were:

    • HypoDE (ATTD 2018): in MDI type 1 diabetes patients (n=149) with hypoglycemia unawareness, the use of a Dexcom G5 for 28 days reduced hypoglycemia incidence (£54 mg/dl for at least 20 minutes) by 72%. Severe hypoglycemia was nearly halved with CGM, a major win for cost-effectiveness. Importantly, this is the very first randomized controlled trial to examine the efficacy of real-time CGM in this specific patient population, making a strong case for the benefits of CGM in those treated with MDI and with hypoglycemia unawareness. It also begs the question: what would a head-to-head study show comparing CGM+MDI+decision support vs. automated insulin delivery (pump+CGM)?

    • CONCEPTT (EASD 2017): The JDRF-funded randomized controlled trial testing CGM in pregnant women with type 1 diabetes (n=215), showed positive neonatal outcomes with Medtronic’s older Guardian CGM. CGM drove a significant reduction in the incidence of large for gestational age (OR =0.51), fewer NICU admissions lasting 24+ hours (OR=0.48), fewer incidence of neonatal hypoglycemia (OR=0.45), and one-day shorter length of hospital stay. While there was only small -0.2% A1c reduction and non-significant difference in hypoglycemia (3% vs 4%; ~-14 minutes/day), mothers on CGM spent a significant 100 more minutes/day in range (68% vs 61%) and 72 fewer minutes/day in hyperglycemia (27% vs 32%).

    • The IN-CONTROL study (EASD 2016): type 1 diabetes patients (n=52) with hypoglycemia unawareness for 16 weeks showed a significant 53% fewer severe hypoglycemia events with CGM vs SMBG (14 events vs 34 events). Results were achieved in a population with baseline A1c 7.5%, mean age 49 years, and 56% MDI users.

    • GOLD (JAMA 2017): A multi-center, crossover, randomized trial with Dexcom CGM (n=161) in type 1 MDI users an A1c reduction of 0.43% (baseline: 8.6%).

    • DIaMonD (JAMA 2017, ADA 2016): In the type 1 phase (n=158 ), MDI users not at goal (baseline A1c: 8.6%) achieved a mean A1c difference of 0.6% favoring CGM over SMBG (-1.0% vs. -0.4). CGM users spent a median of 43 minutes per day in hypoglycemia (a 34% improvement from baseline) vs. 80 minutes per day in the control group on SMBG (11% MORE hypoglycemia). All quality of life metrics improved. Similar results were observed for type 2 diabetes, albeit with a smaller A1c reduction (-0.3%) and no difference in hypoglycemia (Annals of Internal Medicine 2017).

    • REPLACE-BG: T1D Exchange’s 26-week randomized controlled trial study comparing Dexcom’s G4 CGM with and without confirmatory SMBG (n=77 for CGM+BGM; n=149 for CGM-only) in people with type 1 diabetes unsurprisingly showed no difference in time in range with adjunctive vs. non-adjunctive use. There were also no differences in hypoglycemia, hyperglycemia, A1c, mean glucose or coefficient of variation. Importantly, outcomes did not significantly differ in subgroups based on age, diabetes duration, and education level – a promising indication that non-adjunctive CGM use is safe for many patient groups.

  •  Dr. Weitgasser also covered recent data for Abbott’s FreeStyle Libre:

    • IMPACT and REPLACE: IMPACT compared FreeStyle Libre to SMBG in type 1 patients with a low baseline A1c (6.7%) and significant hypoglycemia at baseline (3+ hours/day), with FreeStyle Libre driving ~74 fewer minutes/day <70 mg/dl. There was not a significant difference in A1c between groups. The REPLACE study comparing FreeStyle Libre to SMBG in type 2s with a high baseline A1c (8.8%) missed its primary endpoint, showing similar 0.3% A1c reductions in both groups, but demonstrated compelling hypoglycemia improvements, with FreeStyle Libre users spending ~30 minutes fewer per day <70 mg/dl relative to the control group.

    • IMPACT MDI analysis: in a pre-specified MDI subgroup analysis there was a 46% reduction in hypoglycemia for n=167 patients. The Flash group scanned 16 times/day compared to 5.6 tests/day in the SMBG group, demonstrating the value added by FreeStyle Libre in those on injections – in line with DIaMond.

    • SELFY-Study: This 10-week study of n=76 children 4-17 years old showed that FreeStyle Libre improved time in range by 1.0 hours/day compared to SMBG. A1c also dropped significantly by 0.4% (baseline 7.9%) and time >180 mg/dl fell by 1.2 hours/day. There were no significant changes in hypoglycemia, despite the very positive results achieved in the IMPACT study. FreeStyle Libre is approved for use in children as young as four years old in the EU, but has not yet achieved a pediatric indication in the US.

    • Cost calculation for Flash: The annual UK cost of Flash glucose monitoring is £970 (~$1,371). This slightly exceeds the cost of routine SMBG with fingersticks (5.6 strips/day) – estimated at £675 (~$954) – but is notably less than the recommended 10 fingersticks/day £1,205 (~$1703). If the differential costs of severe hypoglycemia are calculated and included, then SMBG costs £1,103 at 5.6 tests/day (~$1559) vs. £1,191 (~$1683) for FreeStyle Libre.

What’s New? Implantable Sensors

Pratik Choudhary, MD (Kings College Hospital, London, England)

Dr. Choudhary covered implantable continuous glucose sensors. They offer accuracy and quality of life benefits, although they require implantation/explantation. He covered Eversense XL, Senseonics’ 180-day wear implantable CGM available in the UK and Sweden. Trial data showed a 0.5% A1c reduction, MARD of 8.9% and a promising 84% of patients in the trials who would choose to have the sensor inserted again. Endocrinologists have found the device “very easy to insert and remove” and said that the outpatient insertion takes “less than two minutes.” As of Senseonics 4Q17, Eversense is used by 2,000 patients across 14 countries and is actively preparing for its FDA advisory Committee meeting to take place this week.

  • NICE guidelines for CGM in type 1 patients have not been updated since 2015 and advise use in adults who are willing to wear the sensor at least 70% of the time and have any of the following despite optimized use of insulin and conventional blood glucose monitoring: (i) >1 episode/year of severe hypoglycemia with no obvious cause; (ii) hypoglycemia unawareness; (iii) hypoglycemia at least twice a week; (iv) extreme fear of hypoglycemia; or (v) A1c above 9% despite at least 10 fingersticks per day. It is not advised to offer CGM routinely to adults with type 1 diabetes, despite clear data showing CGM to be superior in achieving targets relative to MDI. NICE guidelines for type 2 diabetes management recommend not routinely offering SMBG for adults with type 2, unless: (i) the person is on insulin; (ii) there is evidence of hypoglycemia; (iii) the person is on a medication that may increase risk of hypoglycemia; (iv) the person is pregnant or planning to become pregnant. By contrast, FreeStyle Libre has national reimbursement in the UK, giving it a significant leg-up on traditional CGM.

  • Senseonics’s implantable 180-day Eversense XL CGM has been launched in the UK and Sweden. It consists of a small pill-sized sensor implanted in the arm in only a few minutes, and a transmitter which powers the sensors and connects the sensor to a mobile app. The sensor is changed every six months (versus roughly once per week with non-implanted sensors). During Senseonics’ 4Q17 earnings call, we learned that the company has plans to expand to other countries (Italy was mentioned as a “very attractive market”) and a US clinical trial for the Eversense XL sensor with “reduced calibration” is expected to begin enrolling patients this summer. Senseonics anticipates that the entire 90-day wear installed base of ~2,000 users will have moved to the Eversense XL sensor by 3Q18.

  • Two PRECISE studies were conducted for the Eversense. The EU pivotal study PRECISE (n=71) was single arm, non-blinded with predominantly type 1 patients, and PRECISE II (n=90) had two-thirds type 1 and one-third type 2 participants. PRECISE II had a blinded control group. Senseonics modified the algorithm and introduced a new sensor configuration between the two trials. PRECISE showed a 0.5% A1c reduction through 90 days. Mean wear time was 23.5 hours per day. (The only time without the sensor is when charging the transmitter.) MARD in PRECISE was 11.6%, which improved to 8.9% in PRECISE II. Only one procedure-related adverse event was reported in ~500 insertions. Promisingly, 84% of participants would choose to have the sensor inserted again.

  • Other CGM technology under development includes fluorescence-based sensors and smart tattoos. Implanted fluorescent hydrogel fiber can measure glucose using an optical sensor on the skin. Smart tattoos incorporate nano-sensors implanted in the skin which fluoresce in a glucose dependent manner. Dr. Choudhary described these technologies as fairly crude at present. See below for brief descriptions of the groups Dr. Choudhary mentioned:

    • Glysens - developing an implantable sensor, with a predicted one-year life and infrequent calibration. Last we heard, the one-year implant is waiting a pivotal trial.

    • Clinitech – their LabPatch is a sensor on a band-aid, but no public information is available yet.

    • Glysure – has an intravascular CGM for use in the ICU (attached to a central line). However, Dr. Stanley Nasraway told us in February that Glysure is now targeting the outpatient market. There have been some questions raised as to the practicality of entering this space, with Emory’s Dr. Guillermo Umpierrez commenting on a “limited future” for CGM in the ICU at the recent Endo Fellows meeting. Still, we’d point out that reducing the need for hourly fingersticks would significantly reduce nurse burden not to mention patient pain. Currently, the OptiScanner is the only inpatient CGM cleared by the FDA for surgical intensive care unit (SICU) use.

    • IVBG – an IV blood glucose monitoring system for the hospital. They are conducting a n=100 trial of an intravascular catheter.

Questions and Answers

Q: What is the failure rate for Eversense?

A: 86% got to three months, and a similar amount to six months. There have been only 10-15 insertions so far in UK so its early days. But it looks like the manufacturer expects a small amount of failures.

Closed Loop Insulin Delivery and Diaport

Roman Hovorka, PhD (University of Cambridge, England)

In an action-packed talk, the renowned Dr. Hovorka toured the manufacturers and academic groups who are developing closed loop systems, although he didn’t say too much about their anticipated timescales or chances of success – see our updated competitive landscape here. He covered a new meta-analysis of closed loop trials, which showed over two hours more per day in the target range and a reduction of hypoglycemia. He also touched on four trials occurring at Cambridge and a randomized controlled trial of the 670G. If that wasn’t enough, he presented initial data on closed loop with Diaport. He concluded by underscoring how implementing closed loop represents huge challenges for existing clinical staff – that’s because, he explained, it’s less about biology and more about technology.

  • The Medtronic MiniMed 670G is the first commercially available hybrid closed loop system and has been available for a year in the US. In ‘Auto Mode’ the system doses every 5 minutes with a PID controller. It’s a first-generation product and not particularly aggressive – we’ve heard from Dr. Irl Hirsch and others that some users are even entering “fake carbs” to make the system more aggressive. Still, 670G users are seeing strong overnight outcomes, despite some frustrations with exiting auto-mode during the day. A single arm pivotal trial demonstrated that it is safe (JAMA 2016, DT&T 2017). Real world data from ~14,000 users show that time in range increases from 63% in manual mode to 71% in Auto mode – see our coverage from ATTD 2018.

  • Insulet is developing the Horizon Automated Glucose Control system which houses the control algorithm on the pod. The hybrid closed loop system is in its third IDE trial (five-day hotel study), and depending on the data, the company will either move to a pre-pivotal study later this year or conduct another study before moving ahead. Still, a 2018 pivotal trial seems unlikely, meaning a pivotal could come in 2019, followed by a launch with Dexcom G6 in ~2020 (if we had to guess). Insulet has focused on testing Horizon in very young populations, with this ongoing trial in pediatrics down to age two years.

  • Bigfoot Biomedical is developing a system which uses the former Asante pump platform and a next generation FreeStyle Libre sensor. At DTM, we learned that the Bigfoot algorithm starts with a single basal rate, carb ratio, and insulin sensitivity factor, and then adjusts these variables plus glucose targets on an hourly basis. Topline feasibility results from a single-arm safety trial (n=20) presented at DTM showed 65% of time between 70-180 mg/dl, 0.9% time <70 mg/dl, and 34% of time <180 mg/dl - consistent with other AID studies, but not a super challenging test of the system. Bigfoot expects to begin its pivotal trial this year, with a possible US launch in 2020, pending FDA approval.

  • Inreda Diabetic is a EU company developing a dual hormone fully closed loop system. The company has not updated its website in some time.

  • Tandem revealed positive results from the pivotal trial (PROLOG) of its predictive low glucose suspend (PLGS) device with Dexcom’s G5 at ATTD.  The system drove a solid 31% reduction in mean time <70 mg/dl, meeting its primary endpoint. Suspensions occur without alarms – a huge patient win – and the product is currently under FDA review, with a launch expected this Summer. The second-gen TypeZero Control-IQ hybrid closed loop is expected to enter its pivotal study in 2Q18.

  • Boston University is developing a fully integrated dual hormone dual chamber pump system, although they plan to commercialize and insulin only system first. As of ATTD 2018, an insulin-only pivotal trial is expected to begin in April 2019, followed by a bihormonal pivotal trial beginning in June 2019.

  • University of Virginia/Type Zero Group are working with partners to develop their connectivity-based approach using commercial products (Dexcom/Roche/Tandem). At ATTD, Dr. Boris Kovatchev highlighted a ski camp study showing a remarkable six-hour/day improvement in time-in-range for the Dexcom G6/Tandem t:slim X2/TypeZero system in 13-18 year olds (n=25).

  • University of Cambridge is also taking a connectivity-based approach using the Abbott Navigator and Florence pump. The team’s commercialization plans are unknown, but there are four longer term studies happening: (i) APCam11 (with JDRF) in children 6-18 years-old over three months; (ii) Dan05 (with NIH) in children 6-18 years-old over six months; (iii) KidsAP (­with EU) in children 1-7 years-old studying diluted (U20) vs undiluted (U100) insulin aspart in closed loop; and (iv) CLOuD (with Helmsley Charitable Trust and NHS) in newly diagnosed 10-17 year-olds over 24 months. The aim of CLOuD is to find participants within six weeks of diagnosis and assign them to 24 hour closed loop for two years. The trial endpoint is residual C-peptide secretion.

  • A 2017 meta-analysis of closed loop outpatient studies (Weisman et al., Lancet D&E 2017) shows an average increase of 12.6 percent points in time in target range, which is over two hours per day. The analysis demonstrated a high variability between the various results, showing that optimization of the technology has some way to go.

  • Dr. Hovorka also mentioned other ongoing trials:

    • Australian JDRF/NHMRC outpatient trial in youth, expected to complete by the end of 2018. This is the first randomized controlled trial of the Medtronic 670G versus standard care in children and young adults, with a parallel adult study.

    • Type Zero: IDCL Trial is an international multi-center $13M trial assessing closed loop versus open loop pump and CGM in two systems: (i) Tandem t:slim X2 with Dexcom G6 and TypeZero inControl; and (ii) Roche Accu-Chek Insight pump with Senseonics Eversense XL (180-day wear) and Type Zero inControl.

  • The DiaPort (Gen 2) is an intraperitoneal insulin delivery device, meant to be a ‘last resort’ when subcutaneous delivery fails. A catheter is implanted close to the liver, connected to a port on the surface of the skin. The first UK child on DiaPort is Taylor Banks who has a severe skin-deep insulin allergy. He has improved tremendously. A closed loop study (n=12) of DiaPort 2 showed improvement in A1c, time in range and glycemic variability versus closed loop on a regular pump, although 20% had a failure or major issue. Only one type of insulin is approved for use with Diaport, a potential commercial risk for the system if Sanofi does not keep making U400. At the JDRF Helmsley Charitable Trust Closed Loop Intra-Peritoneal Infusion Workshop last April, Harvard’s Dr. Eyal Dassau showed the DiaPort achieved 66% time-in-range in full closed loop, while the subcutaneous delivery group spent 44% time-in-range. At the time, an automated insulin delivery trial with intraperitoneal delivery was expected in a canine model in collaboration with PhysioLogic Device (pump-sensor system), Vanderbilt’s Dr. Justin Gregory (who will conduct the trial), and UCSB/Verily’s Dr. Howard Zisser.

Questions and Answers

Q: Which patients should be first in the queue for closed loop?

Dr. Roman Hovorka: I think that technology support is key. First, we need providers to understand the technology. Some HCPs don’t have a clue about technology. It’s really data rich. As for patients, it would most benefit those with a high A1c who can’t be bothered. But even those with lower A1c can reduce their hypos.

Dr. Pratik Choudhary: We need to set patient expectations very well and train carefully. We don’t want them to fail and set the technology back.

Comment [Clinician from USA]: We have 57 patients on the 670G in our practice. We have a collaborative approach with Medtronic. All healthcare providers are getting trained by Medtronic. But when people are eligible for a new pump, those who were on Medtronic or Animas will get the 670G. It can be used as a manual pump, but then most patients switch to auto mode. Note that it shouldn’t be in auto mode for exercise.

Closed Loop Insulin Delivery Systems: Who First? Patients with Hypoglycemia

Michael Rickels, MD (University of Pennsylvania, Philadelphia, PA)

Although this talk was intended to make the case for closed loop in patients with problematic hypoglycemia, Dr. Rickels also spent a lot of time comparing pumps/CGM with islet transplantation. As he argued, islet transplantation is superior, but comes with its own set of issues.

  • Hypoglycemia is common and it’s a vicious cycle – hypoglycemia begets hypoglycemia. About 20% of patients reported an episode of severe hypoglycemia in the prior year and there is increased risk with hypoglycemia unawareness. The prevalence of hypoglycemia is significant at both low (<7%) and high (>9%) A1cs, suggesting how A1c is a limited metric when it comes to providing a comprehensive overview of an individual patient’s experience with diabetes. Typically patients with diabetes rely on the sympathoadrenal defense to hypoglycemia, but they can develop HAAF (hypoglycemia-associated autonomic failure) which leaves them unaware until it’s too late. For this reason, there are many ‘frequent fliers’ – patients with recurring hypoglycemia It might require six months of near-complete avoidance of hypoglycemia to reverse HAAF. Closed loop systems with micro-dosed glucagon might hold promise in achieving this goal.

  • The current recommendations for treating hypoglycemia are: (i) first line – education; (ii) second line - pump therapy or MDI/CGM; (iii) third line - pump/CGM with/without low glucose suspend; and (iv) fourth line - islet or pancreas transplant.

  • A comparison of islet call transplantation versus real time CGM demonstrated the superiority of islet transplantation on hypoglycemia severity and frequency, impaired awareness of hypoglycemia, and glucose variability. Islet transplantation also restores glucagon and epinephrine response after 18 months. Of course, it comes with its own unique set of issues, requiring surgical intervention and therefore raising the chance of infection and related complications.

Automated insulin delivery (AID) performs better than real time CGM (open loop) for time spent above 70 mg/dl, but neither are a match for islet transplantation. According to Dr. Rickels, time spent without hypoglycemia is likely what’s most important in the restoration of symptom awareness. He presented data showing hypoglycemia declines by 92% with islet transplantation, ~40% with CGM, and ~50% with AID. Equally, islet transportation is superior to AP in A1c reduction. Of course, scalability and cost must be factored in, where technology has the advantage.Questions and Answers

Q: Can we use dual hormone closed loop in hypoglycemia unawareness?

A: Dual hormone systems have great potential for those at increased risk for hypoglycemia. There is still a delay in glucose predictive suspend systems while glucose can still fall. So, having a backup of glucagon would be advantageous.

Q: For the dual hormone closed loop – are there any long-term implications of using glucagon?

A: Currently available glucagon is only for immediate use because it forms fibrils in aqueous solution, but there are two products in development – one non-aqueous (Xeris) and the other structurally modified to avoid fibril formation (Zealand). I think we will have options that will work longer term and be delivered in micro-doses.

Closed Loop Insulin Delivery Systems: Who First? Diabetes in Pregnancy

Helen R. Murphy, MD (University of Cambridge
, England)

Dr. Murphy made a strong case for CGM and closed loop in pregnancy. Small changes in time-in- range appear to have major implications for pregnancy outcomes. Dr. Murphy conducted two excellent closed loop trials in pregnancy, demonstrating that closed loop is safe and effective for continuous wear through all stages of pregnancy.

  • The CLIP_03 (n=16) and CLIP_04 (n=16) studies demonstrated the value of closed loop in pregnancy. They found that overnight closed loop reduces hyperglycemia without increases in hypoglycemia or insulin dose (CLIP_03), and that day-and-night closed loop reduces hypoglycemia and is suitable for all phases of pregnancy (CLIP_04). In CLIP_03, a randomized, crossover study comparing overnight closed-loop therapy vs sensor-augmented pump therapy, time in target increased from 60% to 75% (+216 minutes/day) and mean glucose dropped from 133 mg/dl to 119 mg/dl at four-weeks. Participants used Sooil’s DANA Diabecare R Insulin Pump and Abbott’s FreeStyle Navigator II. In CLIP_04, also a randomized crossover trial, the average number of hypoglycemia events were reduced from 12.5 to 8 over 28 days and time <63 mg/dl decreased from 2.7% to 1.6% (-16 minutes/day). There were no significant differences between time-in-range, hyperglycemia, and time >140 mg/dl. It was found that day and night closed loop is feasible during admissions, steroids, labor and delivery (including c-section) and rapid reductions of insulin post-delivery. The closed loop system was developed by Dr. Roman Hovorka and his team. Read our coverage of the NEJM publication here.

  • The CONCEPTT randomized controlled trial of real time CGM in pregnancy demonstrated only a small improvement in time in target, but a significant reduction in NICU admission, justifying funding CGM. At 34 weeks, time in target was 68% for the CGM group (using Medtronic’s older Guardian CGM) and 61% for control (a significant 100 minutes per day difference), while the A1c advantage for CGM was a small -0.2%. Mothers on CGM also spent 72 fewer minutes/day in hyperglycemia (27% vs 32%) and a non-significant ~14 fewer minutes/day in hypoglycemia (3% vs 4%). Importantly, CGM drove a reduction in babies born large for their gestational age (LGA; OR=0.51), a reduction in babies being treated for hypoglycemia at birth (OR=0.45), and a reduction in NICU admissions lasting 24+ hours from 43% to 27%. Results were published in The Lancet – a major win for visibility. One day in the NICU exceeds the cost of CGM for the duration of pregnancy! Given use of a very old CGM, we imagine a modern system paired with remote monitoring/decision support would do miles better. Read our in-depth report of the full CONCEPTT results here (EASD 2017).

  • The majority of women with type 1 diabetes don’t achieve optimal control in pregnancy. In the National Pregnancy in Diabetes (NPID) register, only 14% of mothers with pre-existing type 1 diabetes achieved the 6.5% A1c first trimester target in pregnancy. This has stayed the same for 15 years. There is a large clinic to clinic variation (zero to 44% of patients achieving target) likely due to large discrepancies in quality of care, healthcare access, and available resources. In the 2016 NPID Audit, one in two babies have complications, 43% babies are delivered pre-term, and 40% spend time in the neonatal intensive care unit (NICU).

  • Phase III clinical studies (AiDAPT) will examine the long-term impact of closed loop on maternal fetal outcomes.

Questions and Answers

Q: Should CGM be the standard of care in pregnancy?

A: Yes, for sure. We have shown the benefits in terms of less days of intensive care. We hope NICE will evaluate it. We will need a formal healthcare economics calculation, but my ‘back of the postage stamp’ calculations suggest that it will work out.

Automating Glucose Control In-Hospital: Bringing Closed Loop to the Wards

Hood Thabit, MD (University of Manchester, England)

Dr. Thabit made a good case for using closed loop in the hospital where glycemic control can be especially challenging. The advantages of a closed loop system are that it is automated, continuous, has a better algorithm, and reduces the burden on healthcare providers and patients. Early studies have shown a strong increase in time-in-range for a diverse patient population with no increase in hypoglycemia.

  • Better glucose control in the hospital can lead to a 50% reduction in infections and improved outcomes, but attempts to achieve tighter control have led to an increase in hypoglycemia, which becomes a limiting factor for insulin administration. Hospital targets are 100-180 mg/dl but one in 10 inpatients experience blood glucose of less than 50 mg/dl during their stay. Administering insulin takes up a lot of staff time, glucose can be unpredictable, and CGM is not currently recommended for inpatient use.A study of subcutaneous closed loop on a general ward for 72 hours increased time in target from 38% to 60% (+~5 hours!) without any hypoglycemia. It especially reduced glucose variability overnight. The trial excluded patients on dialysis, steroids, or parenteral nutrition, but nonetheless administered insulin in a granular way that couldn’t be replicated by healthcare staff. A second study (n=43) of up to 15 days had essentially identical time in target improvement and also reduced time in hypoglycemia by 28%. This trial included any kind of patient, requiring adjustment to large variation in insulin needs. See our coverage of the Cambridge team’s impressive work in the hospital from ADA 2016 and ADA 2017 – there is so much potential here!

  • Intravenous closed loop in the ICU is being prototyped. This device (we believe Admetsys) uses insulin and dextrose pumps (through an IV) and a standard CGM device. In principle it should be possible to achieve very tight control without hypoglycemia, although patients in the ICU are much more sick and volatile.

  • Currently 17% of inpatients in the NHS have diabetes. In Manchester, one in three have diabetes, and one in three of these will require insulin therapy. Additionally, the number of hospital beds occupied by people with diabetes is increasing in the UK due to increasing age and prevalence of obesity and type 2 diabetes.

Questions and Answers

Q: Can closed loop reduce the length of stay?

A: That’s quite challenging to address. We know that glucose control reduces infection rates. There are many other factors that affect length of stay – the hospital, discharge policies etc. I would look at outcomes first and the impact on HCP time.

Q: What is the training for inpatient closed loop?

A: Patients get zero training – that is the goal and in any case, they are unwell. The staff were only taught to calibrate the sensor and the research staff changed the infusion set.

What’s New? CSII and Bolus Calculators

Lalantha Leelarathna, PhD (University of Manchester, England)

Given a very short time, Dr. Leelarathna gave a shopping list of notable innovations. These were:

  • GoCARB: is a computer vision smartphone app that estimates carbs based on volumetric assessment of the food components from two photos of a plated meal. Food has to be organized in separate areas of the plate. In a weeklong, prospective randomized controlled cross-over study (n=20), GoCARB significantly reduced hyperglycemia and glycemic variability in type 1 adults. Participants decreased time in hyperglycemia (>216 mg/dl) from 18% to 15% (-42 minutes/day). Glucose standard deviation decreased by ~5 mg/dl (baseline ~58 mg/dl). GoCarb is part of Roche’s integrated diabetes ecosystem and was featured at the Roche booth at EASD. The app is expected to launch in 2018 following optimization of its machine learning algorithm, and we’re very excited to see how it works in real-world use – obviously accuracy and usability will be key, especially if two photos of the meal are required. That said, given how hard it is to accurately estimate carbs, even moderate carb-counting accuracy could be highly beneficial.

  • VoiceDiab: is a bolus calculator app with speech recognition. The user speaks their meal into the app to calculate insulin bolus from a carb estimate. In a four day trial (n=44), the proportion of patients who achieved two hour post-prandial glucose in-range (70-180 mg/dl) differed by a meaningful 12 percentage points (59% vs 47%). Like GoCARB above, we’ll be following the accuracy and real-worldy usability closely on this one – obviously the speech recognition part is doable with current technology, but accurately estimating “Spaghetti” or “rice” or “pizza” or “fruit” is still a guessing game.

  • ABC4D: leverages case based reasoning (an AI technique) incorporated into an insulin advisory app. A six-week observational pilot (n=10), resulted in a “trend towards lower episodes of hypoglycemia,” with the number of postprandial hypoglycemia episodes within six hours of a meal dropping from 4.5 in week one to 2.0 in week six.

  • Fiasp (fast acting insulin aspart) in pumps:  The double-blinded Onset 5 trial compared Fiasp vs Novolog (insulin aspart) in pump users (n=472) for 16 weeks. There was no change in A1c (non-inferiority) but one hour post-prandial glucose was a superior 16 mg/dl lower (95% CI: -25.73 mg/dl to -7.06 mg/dl). At ATTD, Dr. David Klonoff concluded from Onset 5 results that Fiasp is safe and effective for pump use, an important finding – Fiasp is currently approved for use in pumps in Europe, but not in the US or Canada. Fiasp was launched in US pharmacies in February.

  • Medtronic 640G PLGM in children: In a 14-day trial (n=96) of predictive low glucose management (PLGM) in children already on pumps, switching PLGM ‘on’ reduced the number of hypoglycemia events by 41% (events <65 mg/dl decreased from 4.4 to 7.4), at the small expense of 9% longer time spent in hyperglycemia (>140 mg/dl; +76 minutes). In Australia, a similar trial in n=154 adolescents reduced time spent in hypoglycemia (<65 mg/dl) by roughly 40% (-16 minutes) with no change in A1c. This system has been out in Europe for three years, following the major coming-out party at ATTD 2015.

  • Adding CGM to MDI or Pump: In the second phase of the DIaMonD study in type 1 diabetes, when patients on the G4 CGM were further randomized to continue on MDI (n=38) or switch to an Omnipod (n=37), the OmniPod group increased time-in-range (7-180 mg/dl) by 78 minutes/day. Time >180 mg/dl was also strongly in favor of pump therapy, with the OmniPod group demonstrating an improvement of -47 minutes/day vs +59 minutes/day compared to MDI. However, time <70 mg/dl actually favored the MDI group, who spent 9 fewer minutes/day compared to 15 minutes/day in the Omnipod group. This data was published in Lancet D&E last year, adding to the type 2 data (Annals of Internal Medicine) and type 1 MDI data (JAMA).

  • All pumps can achieve the same results? A study of pumps from four manufacturers (Medtronic, Animas, Roche, Omnipod) showed that they performed similarly on A1c lowering. We did not catch the citation, but assume automation was not included and the study was not randomized. It might have been from Dr. Leelarathna’s practice, as he presented in Insulet’s symposium at EASD. He emphasized that patients should select their preferred pump based on lifestyle, design, and on-body form factor. Moving forward, we expect this could change as automation differs and user experiences differentiate between pumps and phones.

Big Picture

Plenary Session with Mr. Simon Stevens, NHS England

Simon Stevens (Chief Executive, National Health Service, England)

It was refreshing to see how knowledgeable Mr. Simon Stevens, the Chief Executive for NHS England, appeared to be when it came to diabetes, given that the National Health Service is one of the planet’s largest employers (1.5m people) and NHS England has an annual budget of £100 billion. Of course, the fact that diabetes is already 10% of the NHS budget and is growing faster than Mr. Stevens’ funding has probably focused his mind a lot. In his talk and Q&A he went on the record as saying that weight loss should be the “primary goal in the treatment of diabetes”, and that there will be a “revolution” in digital health. NHS England launched a National Diabetes Prevention Program in 2016, which is still scaling up (but is now bigger than any similar European or US Program). NHS DPP is a joint commitment from NHS England, Public Health England, and Diabetes UK. In November, the same three organizations teamed up to launch a pilot aimed at evaluating the efficacy of digital diabetes prevention programs. Up to 5,000 people who are overweight or affected by obesity will be recruited over a six-month period, receiving access to digital intervention for up to 12 months.

  • The NHS is investing in diabetes prevention– in fact, Mr. Stevens explained that diabetes is an index of how the NHS is future proofing its resources. The NHS has launched the biggest DPP of any country in Europe. Currently 66,000 adults are enrolled -  a quarter are ethnic minorities and about half are men. Weight loss is ahead of what’s been modeled.

  • According to Mr. Stevens, the root of the problem is the obesity epidemic. To this end, the NHS is “kicking out junk food and pop [soda] from hospital premises”. In the last 18 months, one NHS onsite retailer sold 1.1m fewer chocolate bars and 175,000 more pieces of fruit instead. Healthy sandwiches went up from 22% to 52%. The NHS has also developed the Childhood Obesity Strategy, an initiative to cut added sugar for children by 2020. But more forceful action is needed, Mr. Stevens argued. He believes that the environment is right to have a more assertive UK government stance on the fight against obesity. 80% of people think that the government should be taking action to force manufacturers to make lower calorie foods and label food more clearly. A post Brexit regulatory arrangement could allow tighter rules. Mr. Stevens emphasized that the DiRECT trial revealed “weight loss should be the primary goal in the treatment of diabetes”. On current trends 70% of millennials will be overweight or obese by the time they are 35-45 years old. We were shocked to hear that one in five children are obese when they leave primary school. There is a strong connection between obesity and deprivation. Mr. Steven argued that we live in an obesogenic environment, showing data suggesting that fast food outlet density and obesity are correlated.

  • NHS England is trying to move away from a system in which care is delivered in separate settings and towards an integrated care model. Mr. Stevens noted that when NHS was founded in 1948, the system was in the ‘brief encounter’ mode of delivering acute care. Today, he explained, we are now in the ‘continuing relationship’ era. Still, there are continuing gaps in care – lack of coordination or quality of care, including worrying errors in inpatient insulin management – requiring enhanced integration of care for people with diabetes. Mr. Stevens finds early results to show significant improvements.

The problems faced today in England are significant. The UK is already the most overweight nation in western Europe, and in the last 12 months, 100,000 more people have been diagnosed with diabetes. The UK has more than 3.5 million people with diabetes, up from 2 million over the last 20 years, plus 12 million people at risk of type 2 diabetes. One sobering statistic shows that 22,000 people in the UK will die early this year, in part because of their diabetes.Questions and Answers

Q: The National Diabetes Audit showed problems for the most deprived groups. How do we improve care of this neglected group of people?

Dr. Partha Kar: Integration of care will make a big difference.

Q: Diabetes and obesity are worsening. What are we doing about community pharmacists? They see the patient more often.

Mr. Simon Stevens: We are investing in clinical pharmacists working alongside GPs in primary care and in direct patient-facing clinics. We don’t have shortages of clinical pharmacists like we do for GPs.

Dr. Partha Kar: They have a big role to play from the start to the finish.

Q: We are moving from testing therapies to testing improvements in clinical practice. Why isn’t there funding for research in these new areas?

A: The NHS is trying to look at how to invest in these areas. We are entering a new world of data. We need to figure out how to use it better. We are addressing this with the digital arm of the NDPP.

Q: How can digital access help?

Mr. Stevens: Structured [diabetes] education doesn’t work for people in deprived areas. Support services have been organized around the system rather than the patient. On the other hand, think about the ubiquity of smartphones. We are trying to be evidence based, but I think we could probably see a complete sea change over the next few years, and the touchstone is what happens in diabetes. I’m confident that this will be a revolution.

Q:  What is the role of surgery in weight loss?

A: Bariatric surgery has an important role to play. The results are impressive. But now we have evidence that we can get similar degrees of weight loss with lifestyle intervention. The emphasis has to be upstream on prevention – we need a spread of interventions, and of course that does include bariatric surgery.

Diabetes UK’s Evidence-Based Nutrition Guidelines for the Prevention and Management of Diabetes

Douglas Twenefour (Deputy Head of Care, Diabetes UK, London, UK); Louise Goff, PhD (Kings College, London, UK); Pamela Dyson, PhD (Oxford University, UK)

Diabetes UK recently embarked on a process to update its nutritional guidelines for people with diabetes based on a thorough review of the scientific evidence. We salute the nutrition science community in the UK in its efforts to actively improve its reputation – in the past, badly designed studies and sensational newspaper headlines were just two of the many problems that prevented clarity and consensus. The detailed report is valuable in itemizing specific foods for particular situations. For weight loss, prevention of type 2 diabetes, reduction of cardiovascular risk and glycemic control, the advice is to reduce red and processed meat, refined carbohydrates and sugar sweetened beverages, and increase vegetables, fruits, wholegrains, fish, nuts and pulses. Eggs and dairy are not a concern, in fact, the report finds fermented milk products (yogurt and cheese) to help prevent diabetes. The complete report can be downloaded from the Diabetes UK website. Below are some of the highlights:

  • The 2018 Diabetes UK nutritional guidelines take an evidence-based approach in covering the prevention and management of type 1 and type 2 diabetes. Specific recommendations are made for reducing the risk of type 2 diabetes, remission of type 2 diabetes, glycemic control in type 1 and type 2 diabetes, as well as special considerations for those with gestational diabetes and cystic fibrosis- related diabetes. We’re glad to see the emphasis placed across the full spectrum of diabetes, as nutrition plays a key role regardless of diabetes type. The guidelines recommend adopting an individualized, holistic approach, covering not only nutrition, but also alcohol consumption and physical exercise. This latest edition of the guidelines incorporated existing evidence plus additional studies published between January 2010 and July 2017, with one exception of a major UK diabetes remission study published in December 2017 (DiRECT; see above). The guidelines were unable to use only randomized controlled trials or higher quality of evidence because of the limited amount of such evidence – this continues to be a shame and such a solvable problem. For this reason, they also had to adjudicate across prospective cohort studies. Although these studies are often problematic and not definitive, there are at least a large number and evidence can be aggregated and adjudicated. The panel reviewed an incredible 560 references of which 40% were systematic reviews or meta-analyses – it’s amazing how often nutrition guidelines are ‘reviewing the reviews.’ They also consulted key experts and other organizations.

  • The new emphasis is on foods rather than macronutrients. For example, instead of talking about saturated fat percentage, the guidelines will suggest reducing red meat. This has the benefit that it simplifies a lot of complex science (e.g., some types of saturated fats appear to be neutral for diabetes risk) and is easier for patients to understand and implement. People with type 2 diabetes should prioritize weight loss, eat a lower carb and a lower glycemic index diet, take 150 mins/week of exercise and adopt a healthy eating pattern. Below are some of the specific recommendations.

    • Type 2 Diabetes Remission: aim for a weight loss of approximately 15 kg as soon as possible after diagnosis.

    • Low Carb Diets: are an option. Carbohydrate is the prime determinant of post-prandial glycemia (weight is probably the determinant of fasting glucose) and so low carb diets show a greater reduction in diabetes medications. But there is currently a lack of evidence that low carb diets work better in the long term. However, given Virta’s recently published one-year data (n=262) demonstrating 60% type 2 diabetes reversal (A1c <6.5% and elimination of all medications except metformin) with its low-carb/high-fat ketogenic diet and tech-enabled remote care, we may begin to see more consensus on going low carb. While one might assume such approaches are challenging to stick with, Virta reported a very impressive 83% retention rate, suggesting that long-term outcomes may be achievable. For those interested in more low-carb nutrition hacks, we’d recommend Adam’s book Bright Spots & Landmines.

    • Foods to eat more of: Vegetables, whole grains, fruit, lean meat or fish (particularly oily fish), unrefined carbohydrates, nuts and legumes. We’d love to see more clarity on “whole grains,” since these carbs can easily stack up and include high-glycemic foods like granola.

    • Foods to eat less of: salt, red and processed meat, potatoes, refined carbs and sugar sweetened beverages. (Replace saturated fats with unsaturated oils and fats, not refined carbohydrates.)

    • Foods not to worry about: Eggs, dairy (fermented milk products seem to lower risk of type 2 diabetes), tea and coffee

  • The guidelines are tailored for ethnic minorities. 15% of the UK population are ethnic minorities, who suffer a disproportionate burden of diabetes compared to the general population. Minority groups also suffer higher levels of deprivation. Diabetes prevalence is particularly high in African/Caribbean and South Asian groups. Typically, diabetes is diagnosed 10-12 years younger in these populations. The average age at diagnosis is 48 years for African/Caribbean people and 46 for South Asians, compared to 58 years in the general population. In South Asians, 70% of people with diabetes are under 60. However, there is strong evidence that the relative risk of type 2 diabetes in high risk groups can be reduced by around 50% through lifestyle interventions, of which the most dominant factor is weight loss. In trials, the strategies typically used are a goal of 5-7% weight loss, moderating fat, reducing saturated fat, increasing fiber, and increasing moderate physical activity to 150 minutes/week.

  • Diabetes UK has found that there are a lot of unresolved important issues: (i) we still don’t yet know the ideal proportion of calories from carbs, fat, and protein for all people with diabetes (the KEY questions!); (ii) we don’t know the perfect diet for long term weight loss; and (iii) we don’t know enough to predict who is more likely to adhere to and benefit from a particular diet. We suspect that finding the “ideal proportion” or “perfect diet” may be a bit of a moonshot, but it’s clear to us that CGM can help significantly in this domain – where are the large, randomized trials testing different diets with CGM?

Questions and Answers

Q: We have to tell people to eat less in order to lose weight.

Dr. Pamela Dyson: I think we would agree with that. We have to reduce weight by whatever means. Portion control is a key factor for weight loss.

Q: What is your view on artificial sweeteners?

Mr. Douglas Twenefour: People should move to lessen their urge for sweetness by not using sweeteners at all. But artificial sweeteners don’t impact glycemic control, so they are to be preferred over nutritive sweeteners.

Q: What is the role of gut microbes?

A: Evidence for food patterns that impact gut microbes requires more research

Q: We have lumped together all saturated fats because of ignorance or to try and keep it simple. But, fats with 12,14,16 chain carbons will increase LDL but the saturated fats in dairy won’t increase it.

A: That’s why we should say ‘less processed meat and more fermented dairy’ and not talk about types of fat.

Q: Do you recommend very low carb diets?

A: Very low carb diets (<50 grams daily) are fine in the short term, and we support people who are willing to take them with appropriate medical supervision. But there is no long-term evidence for them and people tend to revert back. Most of the current evidence is on restricted carb (< 130 grams daily) diets, which seem beneficial. (Editors’ Note: Given the outcomes many people with diabetes (especially CGM users) report on low-carb diets, we think the field needs to move beyond, “Not enough long-term evidence exists. See this review in Nutrition for a strong case.)

Q: Shouldn’t we encourage low carb diets because they reduce drug costs and make patients feel better?

A: Yes. Also, low carb diets tend to improve food choices. UK fruit consumption is only 1.5 portions per day. We want to encourage more fruit intake and eliminate cakes, sugar sweetened beverages etc. But with low carb diets, people need more support for maintaining the changes. (Editor’s: We hope to see advice get more specific than “increase fruit consumption,” since some fruits, like bananas, are very high in sugar.)

Exhibit Hall

The Diabetes UK exhibition was reasonably compact, with most exhibitors occupying smaller spaces than at ADA or EASD, for example. New product and drug introductions were thin on the ground.

  • The Lilly/BI and Lilly booths offered the broadest product line of the show, to our eye, and there was excitement around Trulicity (dulaglutide), which Lilly promoted as the “most initiated GLP-1” in the UK (based on IQVIA data). Abasaglar (biosimilar insulin glargine) was also prominently advertised, with list price in the UK at 85% of Sanofi’s Lantus.

  • Abbott has received unprecedented support from Diabetes UK for the FreeStyle Libre, which received national reimbursement under the NHS last year. Prior to reimbursement, the Libre was immensely popular on a self-pay basis. Diabetes UK’s major petition effort was massively influential in accelerating the national coverage. Still, in the UK, there are over 200 local Clinical Commissioning Groups (CCGs) which have to agree to put a new product on formulary, so Diabetes UK has started an influencing campaign (normally the job of the drug or device company) and has published a Flash Map so people with diabetes can see where they can obtain the Libre. For those who do not have access to the Libre in their area, Diabetes UK makes it quick and easy to send an email to local NHS healthcare decision makers – nice! Diabetes UK also provides a “How to Obtain Flash” fact sheet, based on its guidance recommending the product be used more widely. A quick look at the Flash map shows that the campaign is doing well.

  • Dexcom also seems to be growing – they were advertising job vacancies at the booth! The company reported exponential growth over the last few years and about 40% of their UK patients are now getting reimbursement. NICE guidelines do support real time CGM (with alarms) for specific circumstances in adults and children, but unlike the Libre, broad national reimbursement hasn’t yet been obtained. As a reminder, the no-cal G6 is currently under FDA review with a global launch expected this year, possibly by mid-year in the US, per Dexcom’s 4Q17 call.


-- by John Close, Ann Carracher, Payal Marathe, Maeve Serino, Adam Brown, and Kelly Close