March 16-20, 2018; Chicago, IL; Days #4-5 Highlights – Draft

The 100th (!) annual meeting of the Endocrine Society wrapped up earlier this week in Chicago, and we have one more installment of conference highlights for you. This report covers days #4-5 of the meeting, which were a whirlwind of learning on technology and therapy alike, not to mention the several talks we attended emphasizing unmet need in obesity care. For more on ENDO, click for our earlier highlights coverage:

Day #1 featured Dr. Irl Hirsch on the landscape of injectable diabetes drugs (he had lots of praise for next-gen basals Toujeo and Tresiba, and also highlighted fixed-ratio basal insulin/GLP-1s), Dr. Vivian Fonseca on typical and atypical complications of diabetes, and Dr. Michael Schwarz on preclinical studies showing diabetes remission with FGF1.

Days #2-3 featured full phase 2 data on Novo Nordisk’s GLP-1 semaglutide in obesity, news of a metformin CVOT in prediabetes (VA-IMPACT), and several exciting product theaters (just-launched Ozempic, Salix’ Cycloset, and more). NIH’s Dr. Judith Fradkin outlined her research agenda for hypoglycemia, in which CGM could play a major part (hooray!), Dr. Lori Laffel discussed CGM, pumps, and AID. We also heard insightful commentary from Dr. Steven Nissen on ODYSSEY Outcomes, from Dr. Donna Ryan on translating lifestyle interventions from RCTs into the real world, and from Dr. Francine Kaufman on women in endocrinology.

In case you missed it, you can also read our coverage of this year’s Endo Fellows series on type 1 diabetes, which immediately preceded ENDO 2018 in Chicago. Here are highlights from days #1-2 and day #3. 

Diabetes Technology Highlights

1. Dr. Bergenstal: Only Abuse of CGM is Not Using It in People Who Could Benefit; Next Standardization Project – CGM Trend Arrows?; Thinking, Fast and Slow for Optimal CGM Use

In a 45-minute talk on the “uses and abuses” of CGM, IDC’s Dr. Rich Bergenstal devoted 44 minutes to the “uses,” reserving just the last line of the talk to one “abuse”: “not utilizing CGM in those with the potential to benefit.” Hear, hear! Dr. Bergenstal illustrated the stepwise progression we’ve seen in the past few years with a hopscotch board, ranging from CGM (phone, pump, receiver) to optimal glucose management (“good” A1c and time-in-range, little hypoglycemia). He shared lots of insight on where the field stands in CGM, with notable highlights on CGM trend arrow standardization (we need it!), benchmark goals (670G pivotal is a good starting point), estimated A1c (moving to “Glucose Management Indicator (GMI)”), and “thinking, fast and slow” for using CGM – more below!

• Having gained some footing in the standardization of glycemic reporting bins and metrics display, Dr. Bergenstal seems energized and ready to take on the next frontier: CGM trend arrow standardization. “As you can tell, part of my life has been spent on standardizing. Now it’s time to “standardize” CGM trend arrow definitions and display. It would be nice if we could just do this, and it doesn’t seem like a big stretch. Then again, it took seven years to get to the point that we agree on whether hypoglycemia is ‘<70’ or ‘≤70 mg/dl.’ Yeah, that took us seven years.” Indeed, standardization is often a heavier lift than it seems, but the benefits of the field coming together can usher in new paradigms – e.g., widely using CGM in trials, hopefully getting such data in labels, improving HCPs’ understanding of CGM data, etc. Currently, as the Endocrine Society’s table below depicts, the meaning of Dexcom, Abbott, and Medtronic trend arrows are meaningfully different, complicating algorithms designed to help people dose insulin off of them. We hope that manufacturers could actually come to a rapid consensus around what these arrows should like and mean – we are happy to lend our opinion, as we’re sure Dr. Bergenstal and other KOLs are, to help the different players arrive at a “Trend Arrow Accord” in a timely manner. Hard work from countless standardization papers, debates, and meetings does pay off: As Dr. Bergenstal pointed out, there are now five glycemic bins (<54 mg/dl, <70 mg/dl, 70-180 mg/dl, >180 mg/dl, and >250 mg/dl) that are becoming the “international standard”; a coefficient of variation (CV) <36% is considered a stable profile with relatively low risk of hypoglycemia; and the one-page AGP report is now used by Abbott, Dexcom, and Medtronic (plus, we would add Glooko and Roche). Read our coverage of the Diabetes Care publications.

• Dr. Bergenstal’s notion is to consider “good management” a low A1c with very little hypoglycemia, or “better yet,” high time-in-range with little hypoglycemia. As we’ve heard before, he suggested benchmark goals could be similar to the outcomes seen in the Medtronic MiniMed 670G trial – ~72% time-in-range (70-180 mg/dl), 3% of time <70 mg/dl, and 1% of time <54 mg/dl. “At least we have some benchmarks to aim for.” It is nice to have an idea of what can be expected from modern technology, though the field is still lacking a sensitivity analysis on this front – how much do outcomes differ for 60% vs. 70% vs. 80% time-in-range and 3% vs. 5% vs. 10% time-in-range? Some of this data should not be too hard to gather, given some of the early Big Data analyses Abbott, Dexcom, and Medtronic have shared at recent conferences (most recently at ATTD: Clarity, Libre, 670G). Short-term, what time-in-hypoglycemia predicts severe hypoglycemia? What time-in-hyperglycemia predicts healthcare costs and DKA? Longer-term, it would also be good to know if people who spend 1% <70 mg/dl rather than 4% experience significantly less cardiovascular events later in life. And crucially, do people with different pheno- and genotypic make-ups have different thresholds of times in different ranges that they should strive for to maximize healthy years later on? Lastly, what kind of times in ranges are possible with different technologies; surely, a type 1 on MDI and CGM may not be able to achieve the same time between 70-180 mg/dl as a type 1 on closed loop or a type 2 on metformin. Many questions remain unanswered, but arriving at standard bins and other metrics was one very positive step in the right direction. For more on this topic, see diaTribe’s interview with experts on time-in-range goals and Adam’s column on an “achievable goal” here.
• Dr. Bergenstal echoed Dr. Roy Beck’s comments at The diaTribe’s Foundation’s ATTD event that the term “eA1c” may be replaced with “glucose management indicator” (GMI). He alluded to questions about whether eA1c would remain on reports, presumably referring to complaints that CDRH has received that eA1c values (based on CGM data) were not aligning with lab-measured A1c values. While the CDRH staff likes the metric, according to Dr. Beck, they believe it needs a new name to avoid confusion and even distrust in either CGM or lab values. We’re on board and think this metric has value, assuming it is interpreted in the context of time-in-range, below-range, and above-range. Of course, someone with a very low GMI (e.g., 5.7%) could still think they had optimal glucose management, even if they were spending 24% of their day <70 mg/dl. We like the term “glucose management indicator (GMI),” as the abbreviation is catchy and implies the former-“estimated-A1c” metric is an indicator among other valuable indicator metrics like lab A1c, time-in-range, and time-in-hypoglycemia. Obviously no single metric is sufficient!  
• Pulling a page from Nobel Laureate Dr. Daniel Kahneman’s book, Dr. Bergenstal explained how optimal CGM use requires both “fast” (real-time adjustments) and “slow” (retrospective) thinking. By fast thinking, he was referring to in-the-moment therapeutic or behavioral adjustments based on real-time CGM numbers and trends – this entails using arrows pre- and post-meals, avoiding extreme highs and lows, and generally “using the data on the fly.” As for thinking slow, Dr. Bergenstal was referring to using AGP and retrospective data to look at patterns and adjust therapies and behaviors moving forward – this would encompass adjusting background insulin, basal rates, insulin:carb ratios, timing of insulin, amount of food, etc. He provided one example where thinking slow (retrospective review) was extremely valuable and necessary: He had a patient who took 42 units of Lantus at night, and six units of aspart at every meal. The patient’s average fasting glucose was 112 mg/dl, and his A1c was 7.8% (“I’m not happy with that, but perhaps ACP is” – a nice dig at the ACP relaxed A1c guidelines). After using FreeStyle Libre Pro for two weeks, it became obvious that the patient “needed less Lantus, probably more aspart at lunch, and probably a little more aspart at dinner too. As a matter of fact, can we just redo the whole thing?” In essence, slow, retrospective thinking grants users the ability to do less, and less urgent, fast thinking. We are also fans of Dr. Kahneman’s book, Thinking, Fast and Slow, and love that Dr. Bergenstal would bring it into this talk.
• Again digging into Nobel-level behavioral economics literature, Dr. Bergenstal suggested that sometimes patients require a little “Nudge” to help them with their fast and slow thinking. Any framework to help patients and providers make rational, unbiased decisions, would be of great help – this could include set approaches for downloading data, education, and decision-making. We’re seeing Nudges creep into many areas of diabetes: the upcoming Medtronic/IBM Watson Sugar.IQ app with Guardian Connect (described at DTM as “simple, judgement-free nudges can lead to sustained behavior improvement”), phone notifications from apps like One Drop (“Glucose: Looks like you had some carbs 2h ago. Time for a glucose check!”), encouragement to walk or stand from Fitbit and Apple Watch, etc.
• We found it noteworthy that the example anchoring bias Dr. Bergenstal alluded to in diabetes was, “I learned that sulfonylureas were really good when I trained.” Sulfonylureas are among the most controversial drugs these days over the hypoglycemia risk and concerns that they drive earlier burnout of beta cells. In the latest chapter of this story, we heard former IDF President Dr. Jean-Claude Mbanya defend their use in type 2 diabetes, particularly gliclazide, which he considers safer and more effective.
• In passing, Dr. Bergenstal hoped that at this time next year, “CGM is all going to be factory calibrated.” Abbott can already check this box with FreeStyle Libre, and Dexcom is on the doorstep with a zero-cal G6 expected to launch in the US in 2H18 (currently under FDA review). Medtronic’s Guardian Connect standalone mobile CGM was just approved by FDA with two calibrations per day, and Senseonics Eversense is under review for the same. We would be surprised if both Medtronic and Senseonics delivered factory calibration in the next 12 months, but it’s not impossible, and perhaps Dr. Bergenstal has inside scoops.
• Dr. Bergenstal made an interesting plea for researchers and clinician to speak the same language when it comes to assessing a person’s overall glycemic management. Researchers, he believes, like the top bar of the AGP display the best, since it contains all of the metrics – time-in-ranges, average glucose, coefficient of variation, etc. Clinicians, on the other hand, prefer the AGP graph, since it is easiest to see where variability needs to be tightened, where hypos need to be addressed, and so forth. We think both are valuable and great parts of the one-page AGP graph, depending on how numerical or graphical someone is – and especially when pattern with automatic pattern recognition like Medtronic’s CareLink and Dexcom Clarity do (e.g., “high pattern between 8-10 am”). In Dr. Bergenstal’s words referring to the glucose profile, “the goal is to make everyone flat line. I know that doesn’t sound great.”
• Prior to concluding his remarks, Dr. Bergenstal simply said: “I am convinced that CGM – it would help if regulators and payers got on board – will transform the management of diabetes.” We too believe that far more progress needs to happen on the payer side than on the product development side, as CGM has matured nicely on the latter. We see regulators (at least the FDA) as pretty big CGM proponents and not the biggest barrier to uptake, relative to access and cost.

2. FreeStyle Libre’s (real-time) 1st US Product Theater: Touts Affordability, HCPs Owe it to Patients to Put Biases Aside and Introduce CGM

In Abbott’s first product theater on US soil since FreeStyle Libre was FDA-approved in September and launched in November, Marian University’s enthusiastic Dr. Ernest Asamoah argued that (i) providers can play a big role in accelerating CGM adoption by pushing their personal biases aside. He was also adamant that it is unacceptable for patients on insulin to not be using CGM, and that Libre’s affordability makes it a game-changer for reducing barriers to uptake. 

• Dr. Asamoah suggested that providers, whether they are CGM fans or not, owe it to patients to show them the device and let them decide if they want to use it. This goes without saying on principle, but it is easy to imagine how provider preconceptions or biases could interfere with their recommendations – especially for people with type 2 diabetes or those that seem less tech savvy. For example, they may think that CGM accuracy hasn’t approached that of SMBG, that CGM is just for people on intensive insulin therapy, or that a particular patient is “not engaged enough” to use CGM effectively. All of these beliefs could be projected, such that patients never learn about CGM in the first place or are dissuaded from trying it. (We are reminded of stories of insulin therapy being delayed because the prescriber views fear of needles as a barrier.) It’s hard to know how many people who could benefit from CGM and other tools/medications are not on them because of sub-optimal provider interactions, but Dr. Asamoah reiterated the point that that provider education and awareness is critical.
  
  • Informing patients about CGM doesn’t have to be cumbersome – Dr. Asamoah simply sets them up with a 3.5-minute YouTube video and leaves the room so he can catch up on work. Not only is this a huge time-saver for him, but by the time he returns, patients know a fair bit about the system and are “asking when they can get it.” There is something to “human touch” when helping people make important clinical decisions, but we love how he saves the information part for a video, and then returns to address questions and concerns. Providers have too much on their plate, so every three-minute block that can be salvaged helps! YouTube seems like a brilliant way to expand awareness of CGM and get more patients interested in trying it. (We wonder if it is this FreeStyle Libre video, which has ~325,000 views and is the third-most-viewed FreeStyle Libre video on YouTube; #1 is a 15-second Australia commercial (1+ million views), and #2 is an 11-minute demo of FreeStyle Libre in Portuguese (~450,000 views).)
• Dr. Asamoah pointed out that Libre addresses a litany of barriers to CGM adoption, including calibration, multiple components, and especially out-of-pocket cost. On other systems, he estimated, patients might pay between $200-$700 per month (an exaggeration on the top-end), and for professional CGM, providers often pay $700-$1,400 per system. Meanwhile, paying cash for FreeStyle Libre in the pharmacy for 24/7/365 wear (excluding 12-hour warmup) and a one-time reader purchase comes out to ~$135 per month. The FreeStyle Libre US site additionally suggests that private coverage is favorable, as “Most patients pay no more than $75 per month on eligible FreeStyle Libre prescriptions at major retail pharmacies.” The cash price for other CGM systems is certainly higher than ~$135 per month, but for someone with solid reimbursement, copays can be on par with that for Libre. For example, with our Aetna insurance, a monthly Dexcom copay is just over $70. The overall pricing comparison is complex, but generally, Dr. Asamoah is correct that FreeStyle Libre attenuates the cost factor, especially for trying the system upfront. (We’d emphasize that cost is the top barrier to CGM adoption, according to a recent paper from Dr. Korey Hood’s Stanford group.) And for him, Libre’s affordability makes it a “no-brainer.” He added that he would’ve liked to use CGM three or four years ago, but it was too expensive. He said the same for his patients: “I’ve had many patients say they’d willingly pay out of pocket. They wouldn’t do that with other systems, because that’s too expensive – thousands and thousands of dollars. To me, FreeStyle Libre is a game-changer.”
• Dr. Asamoah said that some of his cash-paying patients wear Libre intermittently – for example 20 days out of every month – though he wouldn’t encourage this “in a perfect world” because people on insulin should always wear CGM. This is an intriguing use case for Libre, and we’d be very interested to see blinded data from those 10 days a month when the patient cannot see continuous data. How much worse is glucose management, if at all? How much benefit in those 10 days without CGM does having CGM for the other 20 days grant? This answer was unexpected, as we believe the questioner was getting at whether Dr. Asamoah ever prescribes Libre for short spurts with long breaks in between (e.g., for 10 days, once per quarter). This is a major question for “intermittent” CGM – what is the right dosing for maximum benefit? For CGM to really scale in type 2 diabetes, this question will need to be explored in studies.
• Dr. Asamoah likened SMBG to old-time flip phones and “the dinosaur age” on multiple occasions (among other metaphors).
  
  • “Do we need CGM to monitor diabetic patients? Yes. Yes, yes. In 2018, if not using CGM to manage diabetes, it’s like using a flip phone. Does anyone in here use a flip phone? [One hand goes up.] You are living in the dinosaur ages. See, everyone is laughing at you. In 2018, it’s like having the iPhone X vs. a flip phone. Remember those old Verizon commercials? ‘Can you hear me now?’ No! We shouldn’t be doing that anymore.”
  • “Imagine you drive from here to the airport with a blindfold on, and you take it off twice or three times on way. That’s what diabetics do. They are literally blindfolded while driving, and they take the blindfold off four times per day. You want your eyes open, and ideally with GPS. CGM gives you that GPS.”
  • “In 2018, CGM is prime time. If you’re an endocrinologist who manages patients with insulin, you should strongly, strongly encourage patients to move to CGM.”

3. Dr. Anne Peters on Exercising with T1D: CGM Helps, 670G and DIY Can’t Keep Up – Bihormonal Might Be Able to?

USC’s Dr. Anne Peters asserted that CGM helps type 1s manage diabetes while exercising, but finds that single hormone hybrid closed loop – be it 670G or DIY – “can’t keep up.” In her mind, CGM has made exercise safer and more possible. This is a no-brainer, as seeing continuous data allows for real-time, minute adjustments to food intake and/or insulin (especially basal rates on pumps) to stay in range. Whereas technology is delivering in a major way for CGM, Dr. Peters finds hybrid closed loop to be less effective: “On the 670G, anyone exercising for any amount of time has to go out of auto mode into manual mode. I have yet to figure out how to keep people in closed loop for more than a few hours. It’s slave to either being at a target of 120 or 150 mg/dl. For less intense exercise, just set to 150 mg/dl for two hours before you exercise, and you should be fine on the temp target. But like anything, you need to individually adjust. I have a lot of patients on Loop DIY systems. They can adjust their target. But even when those patients do prolonged exercise, their systems still can’t keep up. I don’t think until we get to full-on bihormonal that it’ll be able to keep up with them – it’s still in brain, not in the system.” (This point on DIY definitely aligns with Adam’s experience on Loop – the speed of rapid-acting insulin is too slow to compensate for exercise-induced reductions in glucose.) At this point, the marginal benefit of a bihormonal system over insulin-only must be proven, but exercise is one area where it should have a definite advantage. Based on the last update from ATTD, Beta Bionics’ bihormonal pivotal trial will begin in June 2019, with a subsequent launch date to be determined, depending on stable glucagon and other factors. Dr. Peters currently considers exercise guidance to be “like a prescription,” and said that she spends a lot of time discussing it during every clinic visit. An effective closed loop system could go a long way for type 1s who exercise frequently, eliminating the need for intensive preparation and monitoring, and possibly reducing the fear that often prevents people from exercising in the first place.

• Numerous times throughout her talk, Dr. Peters exclaimed (with exasperation) that she cannot figure out how to help patients avoid the post-exercise dip in blood sugar. When someone with Dr. Peters’ expertise experiences this frustration, you can bet it’s a problem faced by most endocrinologists and patients around the world. She noted that this second phase of exercise-induced hypoglycemia – caused by continual uptake of glucose by the muscles as they replete glycogen stores – is difficult to address because (i) it is hard to predict and (ii) it can be difficult to get people to eat carbs after exercise. She even showed the trace of an athlete who kept blood glucose in range for an entire Valentine’s Day marathon (save for one adrenaline-fueled blip above 180 mg/dl when he proposed to his girlfriend), yet he still went low after. Said Dr. Peters, “I’m telling you I can’t get this to go away.” Monday morning armchair quarterbacks can say, “it’s simple, just dose carbs three hours after exercise,” but it’s difficult to predict exactly when a dip will happen and how deep it’ll go. We’d assume 670G and DIY closed loop systems will handle this post-exercise hypoglycemia fairly well, since it is generally a gradual downward drift that basal attenuation should be able to address; it is the rapid drops that are outside closed-loop’s purview. 

• Of the plentiful literature on exercise with type 1 diabetes, Dr. Peters recommended two articles that are useful to have on file: Physical Activity/Exercise and Diabetes: A Position Statement of the American Diabetes Association and Exercise management in type 1 diabetes: a consensus statement. She even gives the latter to some of her more “sophisticated” patients, who use it as a guide. (Adam has also written a lot on this specific topic, including a diaTribe article on staying in zone during exercise and the exercise chapter of Bright Spots & Landmines.)
• Mr. Craig Stubing, co-founder of Type One Run and one of Dr. Peters’ patients, came along for the trip to talk about his experiences with exercise: “When I am exercising, all I am doing is thinking about exercise and my blood sugars – it makes it easy to control then.  When I am done exercising, afterwards, I am living my life and eating and doing other things and forget to adjust my basal rates and doses, so I may go low.” Mr. Stubing founded Type One Run (now a Beyond Type 1 program), an organization that has established 69 chapters around the world to connect type 1 athletes with others. Not only do they find friends, but they can also learn strategies for exercising with diabetes from their peers. 

4. OpT2mise Post-Hoc : C-Peptide Levels DO NOT Impact Success on Pump Therapy; Could it Shift T2D Pump Reimbursement Criteria?

Medtronic’s Dr. Robert Vigersky presented findings from a post-hoc analysis of the OpT2mise study (n=331) demonstrating that the significant A1c reductions achieved in type 2 diabetes patients treated with pumps vs. MDI were observed regardless of baseline C-peptide levels. This result could have important implications for pump reimbursement. Currently, many commercial insurers and Medicare often deny type 2 patients pump therapy if their C-peptide levels are deemed too high (>10% of the upper limit of “normal”). Since this large, randomized, well-controlled, multinational study showed A1c improvements regardless of C-peptide status, it suggests this criterion is not a useful predictor of pump success. Will this resonate with payers, allowing for increased coverage and penetration of pump therapy in type 2? Thus far, this field has been hampered by reimbursement, and arguably, devices not suited for the type 2 population.

• In the OpT2mise study, 10% of pumpers with C-peptide levels ≤183 pmol/L and 12% of pumpers with C-peptide levels >183 pmol/L achieved an A1c <7.0%. On average, pump therapy resulted in an A1c reduction of 1.1% (baseline: 9.0%) at six months, with no differences observed by C-peptide level. Importantly, the outcomes were not obtained at the expense of hypoglycemia, as professional CGM performed at baseline, six months, and 12 months revealed no difference in time <70 mg/dl for either group. Ideally, the percentage of patients in either group achieving A1c <7.0% would be much greater than 10% or 12%, though as we noted back in 2014, this was a very tough population. The stats looked much better for pumps getting people to an A1c under 8%: 43%-55% of pumpers achieved an A1c <8.0% vs. 18%-29% of the MDI group. A more nuanced “responders analysis” from the trial might be useful in helping payers determine which factors predict benefit from pump use. 
• As a reminder, the year-long, multi-national OpT2Mise study is the largest RCT comparing pump to MDI in type 2 – see our coverage of The Lancet publication (six months) and the Diabetes Obes Metab publication (12 months). Eligible participants had a baseline A1c of 8%-12% and performed fingersticks at least 2.5 times/day. Following a two-month run-in period, participants were randomized to either receive pump therapy or continue with MDI. MDIs subsequently crossed over to pump therapy at six months. Upon switching to pumps, both groups achieved significant A1c reductions that were sustained over the full study period.
• We’re hopeful that pumps will expand in type 2s, especially given a meta-analysis presented at ATTD showing that pumps are associated with favorable reductions in A1c (~0.4%) and insulin dose (~24 units/day) vs. MDI. Dr. John Pickup’s analyses also revealed that pump therapy is cost-effective compared to MDI for people with type 2 diabetes and high baseline A1c. Cheaper, disposable (yet fully-featured) pumps developed for the type 2 population like Insulet/Lilly’s U500 and U200 Omnipods, as well as BD’s Swatch patch pump, may help chip away at costs while retaining benefits like connectivity – those are expected to launch in 2019 (BD Swatch, Omnipod U500) and 2020 (Omnipod U200). Type 2 patch insulin delivery devices like Valeritas’s V-Go, Cequr’s PAQ, and J&J’s bolus-only OneTouch Via have either scaled slowly (Valeritas) or not launched at all (PAQ, OneTouch Via). 

5. Dr. Grazia Aleppo Calls Tidepool a “Great Tool,” Cites Study Showing it Increases Data Downloading and Pump Adjusting; Urges Attendees to Teach Fellows How to Read AGP!

Northwestern’s Dr. Grazia Aleppo highlighted a study of 65 pump users/caregivers showing that use of Tidepool, a “great tool,” decreases the proportion that never downloaded device data by 60% after three months. The proportion of participants who downloaded data at least once a month also rose from just 9% to 38%. Perhaps more importantly, results indicate that users actually reviewed their data, as the proportion of participants who changed their pump settings at least monthly nearly doubled, from 11% to 21%. Dr. Aleppo underscored that one of the major challenges of remote pump monitoring is getting patients to actually upload and review their data: According to T1D Exchange data (n=4,000), 69% of pump users never download pump data, with only 16% of adults and 43% of children routinely doing so. Any tool that can improve this number, giving both the provider and patient better visibility into why glucose varies, is worthwhile. Dr. Aleppo commended Tidepool on being very intuitive, allowing patients to easily interpret a lot of the data themselves. She urged attendees to teach their fellows how to read AGP, emphasizing software standardization. Nice! This will become more important as AGP becomes more prevalent and more people use connected devices. She also described Diasend-Glooko, appreciating that it shows highly detailed data from “every possible component,” as well as Medtronic’s CareLink software, advising the audience to focus on time-in-range and time in auto-mode (at least when reviewing 670G data). Given the Tidepool study results, there’s clearly an opportunity for delightful, usable software to spur more engagement.  

• Dr. Aleppo recommended a step-wise approach with specific goals for remote monitoring of insulin pump therapy, emphasizing that it is most successful when accompanied by CGM data. She reviewed a three-month study showing that a five-step approach to pump data review in a type 1 pediatric population leads to a significant A1c reduction of 0.6% (baseline: 7.9%) plus reduced time spent on data interpretation (-8.8 minutes). 

Diabetes Therapy Highlights

1. AZ’s DERIVE Shows Farxiga’s Safety/Efficacy in Type 2s with Moderate (Stage 3A) CKD; Potential for New Indication Down to eGFR 45 ml/min/1.73m²; Evidence Accumulates for SGLT-2s in CKD

Dr. Paola Fioretto presented full results from AZ’s DERIVE study of SGLT-2 inhibitor Farxiga (dapagliflozin) in type 2 diabetes/moderate CKD. With this positive data, AZ could seek a label change that extends the use of Farxiga for people with eGFR down to 45 ml/min/1.73m². DERIVE (n=321) randomized adult type 2s with stage 3A CKD (eGFR between 45-59 ml/min/1.73m²) to dapagliflozin 10 mg daily or to placebo for 24 weeks. The primary endpoint was change in A1c, which fell from a baseline 8.3% to 7.9% in the Farxiga group (0.37% drop) and which stayed more or less constant at 8% in the placebo group (0.03% drop). Dr. Fioretto reported an estimated treatment difference of 0.34% for A1c favoring Farxiga (95% CI: 0.53%-0.15%, p<0.001). While this supports the glucose-lowering efficacy of the SGLT-2 inhibitor in patients with moderately impaired kidney function (small in magnitude, but highly statistically significant), we do wonder about the 0.3% between-group difference in baseline A1c, and we’ll be curious to see if this comes up in further discussions and thought leader analysis of DERIVE. Continuing on, Dr. Fioretto shared results on the secondary endpoint of body weight, which fell from a baseline 203 lbs to 196 lbs with Farxiga (7 lbs drop) and from a baseline 194 lbs to 190 lbs with placebo (4 lbs drop). She reported an estimated treatment difference of 2.8 lbs for body weight change, once again favoring Farxiga (p<0.001). Dr. Fioretto acknowledged the between-group difference for baseline body weight, and we’ll keep our ears peeled for thought leader commentary on this as well. It does seem slightly strange that end-of-trial body weight, on average, was still higher for patients on Farxiga (196 lbs) vs. starting body weight for patients on placebo (194 lbs). Systolic blood pressure declined very rapidly in the dapagliflozin arm – Dr. Fioretto’s slides depicted a 4.8 mmHg drop after just one week (baseline 136 mmHg), which was then maintained throughout the rest of the trial – and decreased by only 2 mmHg in the placebo arm from a baseline 135 mmHg (estimated treatment difference 3.1 mmHg, p<0.05). Change in eGFR was another important secondary endpoint. Mean eGFR at baseline was 53 ml/min/1.73m² for dapagliflozin patients, 54 ml/min/1.73m² for placebo patients. There was an initial decrease for participants on the study drug, which Dr. Fioretto attributed to a “well-known hemodynamic effect of all SGLT-2 inhibitors,” and indeed, this eGFR was recovered by the end of the trial such that there was no significant difference from the placebo group. Renal impairment/failure (eGFR <30 ml/min/1.73m²) was collected as an adverse event of interest, and was actually less common in the Farxiga arm (one patient) vs. the placebo arm (two patients). The one case of renal failure with dapagliflozin resolved during the three additional weeks of follow-up after stopping trial medication. Dr. Fioretto reviewed additional safety findings including genital infection (three patients on dapa vs. two patients on placebo), UTI (four vs. six), bone fractures (none in either group), and DKA (none in either group), concluding that Farxiga was effective, safe, and well-tolerated in this patient population with stage 3A CKD.

• The implications here are substantial (and to this end, we’ve been eagerly awaiting this read out from DERIVE!), because Farxiga is currently not recommended for eGFR <60 ml/min/1.73m². Now that the SGLT-2 has demonstrated safety/efficacy in the 45-59 ml/min/1.73m² range, we’ll be looking out for AZ to request a label update from FDA. Notably, J&J’s Invokana (canagliflozin) is not recommended for eGFR <45 ml/min/1.73m², but only the lower dose (100 mg) is recommended in the 45-59 ml/min/1.73m² range. Lilly/BI’s Jardiance (empagliflozin) is not recommended for eGFR <45 ml/min/1.73m², while Merck/Pfizer’s Steglatro (ertugliflozin) is contraindicated <30 ml/min/1.73m², but is not recommended <60 ml/min/1.73m². With support from DERIVE, Farxiga’s kidney-related stipulation could be relaxed somewhat so that the label reads more like Jardiance’s or Steglatro’s; since investigators used the higher available dose of dapagliflozin, which also comes in 5 mg tablets, we imagine FDA will consider a revised indication for both doses.
• What’s more, these findings bode well for the Dapa-CKD outcomes study of dapagliflozin in people with chronic kidney disease (with or without diabetes). AZ launched this dedicated CKD investigation in February 2017, after diabetes CVOTs for two other SGLT-2 inhibitors (EMPA-REG for Jardiance and CANVAS for Invokana) showed renal protective effects. AZ’s own DECLARE CVOT for Farxiga in type 2 diabetes is scheduled to complete in July, and management has announced that results will be reported in 2H18 – while the CV data will headline, we’re also eager to see how dapagliflozin affected renal risk in the trial. To be sure, there’s growing interest in SGLT-2 inhibitors for renal protection, both within and outside the context of diabetes. Lilly/BI have announced plans for a study to parallel Dapa-CKD, investigating Jardiance in a population with CKD, with or without diabetes. This will likely begin by year-end, and together with Dapa-CKD, could be quite telling about the extent of SGLT-2 efficacy not only for A1c reduction, but also cardioprotection and improving renal outcomes in a diversity of patient populations. The current eGFR contraindications for various SGLT-2 products are based on a presumed loss of efficacy, since the mechanism of glucose-lowering relies on kidney filtration, but this story may be changing soon: For one, DERIVE has demonstrated Farxiga’s A1c-lowering efficacy in moderate CKD. Moreover, Dr. David Fitchett and other experts have speculated that the CV/renal benefits of SGLT-2s may be independent of glycemic effects, and that people with kidney impairment may actually benefit the most from SGLT-2 inhibitors in terms of cardio- and renal protection. We’ll be watching closely as the SGLT-2/CKD story continues to unfold – obviously, a renal indication for this advanced diabetes drug class would be a major win for patients and for outcomes-based medicine (DERIVE showing positive effects on A1c, body weight, and blood pressure is one thing, but Dapa-CKD showing improvements on hard endpoints would be even better, and more meaningful for the field).

Questions and Answers

Q: Do you anticipate that this is a class effect for these medications?

A: Yes. The other SGLT-2 inhibitors on the market, meaning empagliflozin and canagliflozin, are already allowed for use down to eGFR of 45 ml/min/1.73m². That’s because of studies done years ago demonstrating a smaller effect on A1c compared to people with normal renal function, but still a significant benefit compared to not taking the SGLT-2 inhibitor at all. In the earlier studies on dapagliflozin, the agent showed a similar magnitude of effect on A1c, around 0.4%, but there was also a meaningful A1c reduction in the placebo group. For that reason, the data did not meet statistical significance. That was the cause for a label saying you shouldn’t use dapagliflozin in patients with eGFR below 60 ml/min/1.73m². DERIVE is important to extend the use, because we now know that dapagliflozin still has important benefits on blood pressure and body weight in patients with renal dysfunction, even if the effect on A1c is small. Large CVOTs have found very strong nephro-protective effects of these drugs. These might be the patients who benefit the most in terms of CV and nephro-protection.

Q: Do you have data on albuminuria change?

A: There was a 40%-50% reduction in UACR present after four weeks, which was maintained for 24 weeks (p=0.454). We found a similar reduction UACR in another study, which was maintained for one year. We also found this reduction was independent of effects on A1c, blood pressure, or body weight, so it looks like there is a direct effect on UACR. We also measured UACR after we washed out dapagliflozin, but unfortunately, I don’t have that data yet because we just finished the study. That will be provided later on. I imagine albumin would increase again, as seen in EMPA-REG OUTCOME once you stopped the drug, although it didn’t quite return to baseline value.

Q: One caveat I’d like to point out is that your baseline BMI was not very high, so these result might not apply to patients with diabetes and high BMI. Second, white Caucasians do apply, but African Americans or Asians will not apply, so we don’t know. You can’t generalize your findings to other ethnicities. Third, my sense is that the weight loss was not very significant – 3 kg (6.6 lbs) is nothing when I’m looking at the real-world type 2 diabetes population.

A: Obviously, these were predominantly Caucasian patients, and certainly this data should be confirmed in other ethnic groups. The BMI was not that high, but that’s the way patients are in Europe and Canada. A weight loss around 3 kg is what you find in patients with normal renal function as well. We are not claiming that this drug causes weight loss – it’s like something that comes for free. In addition to glucose-lowering and other benefits, there’s also this weight reduction, and the important thing is that you lose it and you keep it, you don’t regain. If you look at subgroup analyses from EMPA-REG and CANVAS, patients who started with renal impairment experienced the same benefit in terms of renal and CV protection, if not more. Of course, that was empa and cana, not dapa, so we’ll see.

2. Diabetes Thought Leaders Take on ACP Guidelines: Dr. Hirsh Criticizes Over-Reliance on A1c as Clinical Endpoint; Dr. Inzucchi Points Out Lack of Endos on Writing Committee

Endocrinologists were not shy about criticizing the new ACP guidelines, which wrongly recommend an A1c goal between 7%-8% for the majority of type 2 diabetes patients – this may be appropriate for some (and we appreciated ACP’s emphasis on individualizing targets), but it’s certainly not best practice for all. During several case presentations, speakers joked about ACP’s endorsement of a higher A1c target, even though ADA and AACE recommend ≤7% and ≤6.5%, respectively, also taking safety and cost into consideration. Tulane’s Dr. Vivian Fonseca presented a case of starting A1c ~7.8%, and remarked, “if you have an ACP membership, you’d be just delighted with this A1c!” Denmark’s Dr. Peter Gaede passed a similar comment during his meet-the-professor session: “How many people would follow the ACP guidelines and just not treat this patient at all, since he’s well below an A1c of 9% right now?” This was followed by chuckles from the audience. Dr. Robert Hood (Endocrine Clinic of Southeast Texas, Beaumont, TX) used a patient case to frame his product theater talk on Lilly’s Humulin: “He’s never had an A1c below 8%. That might be okay with ACP, but it’s not okay with me.” Said IDC’s Dr. Rich Bergenstal, “now here’s a patient with average fasting plasma glucose of 112 mg/dl and A1c of 7.8%. I’m not happy with that A1c – perhaps ACP is.” While many of these statements were perhaps made in jest, the underlying sentiment was quite serious – endocrinology experts are dissatisfied with ACP’s latest guidance, and they continue to defend a target A1c of at most ≤7%. We think this is an important message to spread among HCPs (as well as patients), and we hope the diabetes provider community doesn’t hastily embrace ACP’s guidelines, besides the piece encouraging personalized treatment plans.

• UW’s Dr. Irl Hirsch called out ACP for over-emphasizing A1c as a clinical endpoint (and indeed, there’s no mention of time-in-range or CGM anywhere in the new guidelines). “The ACP guidelines did not pay any attention to this, but I don’t know what the glucose level is for someone with an A1c of 7.5% or 8%. For a population, A1c makes sense, but for the individual, it’s not helpful to guide treatment. Average blood glucose can vary greatly for a constant A1c.” This is an extremely valid argument, because ACP’s aim is to guide internists in real-world diabetes care – treating individual patients, not populations. With its latest guidance statements, ACP seems to be responding to major challenges in diabetes management, including hypoglycemia. Good intentions, wrong decision. Avoiding hypoglycemia is certainly important in treating diabetes, and we understand that it’s not easy, especially not for busy PCPs. But rather than throwing their hands up and raising A1c goals, we wish the ACP writing committee had taken advice from Dr. Hirsch and focused on time-in-range and other glycemic outcomes beyond A1c instead. For patients who can access/afford it, CGM is an excellent tool for hypoglycemia risk reduction. There are also more advanced diabetes therapy classes today that come with less hypo risk. ACP repeatedly mentions insulin and SUs, known to cause hypoglycemia as well as weight gain, without any information on SGLT-2s, GLP-1s, combination products, or even DPP-4s, except to say at the very beginning of the document that they won’t be discussing those drug classes.
• Thought leaders also pointed out that very few (maybe even zero) endocrinologists were consulted in the writing of the new ACP guidelines. As Yale’s Dr. Silvio Inzucchi put it, “next time, I suggest we write HIV guidelines.” Dr. Fonseca echoed this view, urging HCPs to read the statements from ADA, AACE, AADE, and the Endocrine Society rejecting ACP’s discordant recommendations. “These professional societies aren’t the dissenting opinion,” he emphasized, “they’re the REAL opinion.” As we understand it, ACP has strict restrictions on who can contribute to a guideline-writing committee based on dualities. This unfortunately means that many of the smartest experts aren’t allowed at the table (i.e. thought leaders with expertise are also hired – because of their expertise – as consultants for diabetes drug and device companies).

3. Dr. Mark Molitch on Nephropathy: Improving Glucose Control to Minimize End-Stage Renal Disease and Associated CV Risk; Highlights eGFR/Albuminuria Heterogeneity

In discussing the diagnosis and management of diabetes-related kidney disease (DKD), Dr. Mark Molitch argued that the overarching goal of renal care in diabetes should be to reduce the number of people progressing to end-stage renal disease. While the overall rate of end-stage renal disease continues to rise due to ever-increasing diabetes prevalence, the rate at which diabetes patients develop end-stage renal disease has been declining since the mid-1990s, according to Dr. Molitch. Still, he showed how that favorable drop is now starting to plateau. Dr. Molitch outlined a three-pronged approach for renal care in diabetes: (i) prevent diabetes, (ii) prevent nephropathy after diabetes onset, and (iii) slow nephropathy progression once it sets in. He emphasized that screening is critical to success at every stage, and he reinforced the need to measure both albuminuria and eGFR at least annually. Turning to the data, he described how patients with type 2 diabetes in UKPDS progressed from no nephropathy to microalbuminuria at an annual rate of 2%, to albuminuria at 2.8%, and to elevated creatinine or renal replacement therapy at 2.3%. Each stage of kidney dysfunction increases mortality risk: Annual rates of death were 1.4% in UKPDS participants with no nephropathy, 3% in participants with microalbuminuria, 5% in participants with albuminuria, and 19% in participants with elevated creatinine/on renal replacement therapy. Dr. Molitch established end-stage renal disease as most morbid and mortal, and he encouraged endocrinologists to be vigilant in screening and intervening swiftly.

• Days before at Endo Fellows, Dr. Janet McGill explained that declines in eGFR are not always matched by albuminuria progression (and vice versa). In one study cited by Dr. Molitch, ~30% of people with eGFR <60 ml/min/1.73 m² had no albuminuria. As such, annual measurement of urinary albumin/creatinine ratio (UACR) in addition to an annual eGFR calculation based on serum creatinine, age, and race are equally critical.
• That said, both declining eGFR and increased albuminuria are independent but additive risks for major CV events in type 2 diabetes, each conferring two-three-fold higher risk in the most severe 10% of cases (eGFR <~50 ml/min/1.73 m² and albumin >~115 mg/24 hours). Indeed, NHANES data shows that 10-year cumulative mortality in type 2 diabetes is 4% with no kidney disease, 18% with albuminuria, 24% with impaired eGFR (<60 ml/min/1.73 m²), and 47% with both albuminuria and impaired eGFR on top of type 2 diabetes. Scientists are still trying to puzzle apart the link between declining renal function and CV risk, but this overlap has been the subject of many recent talks at diabetes conferences. Potential mechanisms include common genetic predisposition for CV and renal complications, endothelial dysfunction, hypertension, insulin resistance, atherogenic dyslipidemia, hyperglycemia, and volume overload. The good news, Dr. Molitch said, is that we’ve shown in rigorous studies how long-term glycemic control works to sustain eGFR and prevent albuminuria both. In fact, strong glucose control is the only thing conclusively shown to prevent (or delay) the development of DKD. Certainly, SGLT-2 inhibitors are interesting on this front, and secondary endpoints from CANVAS and EMPA-REG OUTCOME indicate a renal protective effect of J&J’s Invokana (canagliflozin) and Lilly/BI’s Jardiance (empagliflozin), respectively. These agents are now under investigation for DKD in CREDENCE and for CKD (with or without diabetes) in a planned study from Lilly/BI, plus AZ has launched a dedicated kidney outcomes study (Dapa-CKD) for SGLT-2 Farxiga (dapagliflozin). Excitingly, in the next few years, providers could have a greatly improved ability to prevent the progression of nephropathy in their patients with diabetes. 

4. The E-LIFT Trial: Empagliflozin Reduces Liver Fat vs. Placebo in Type 2s with NAFLD; Potential Applications in NASH?

A small in-human study (n=42) hinted at the potential for SGLT-2 inhibitor empagliflozin (Lilly/BI’s Jardiance) in NASH. Dr. Mohammad Kuchay presented these findings from E-LIFT, which was an investigator-initiated trial supported by a grant from the Endocrine and Diabetes Foundation in India. Patients with type 2 diabetes and non-alcoholic fatty liver disease (NAFLD), on a background of metformin, sulfonylureas, DPP-4 inhibitors, and/or insulin in any combination, were randomized to 10 mg empagliflozin or to placebo for 20 weeks. By the end of the treatment period, empagliflozin had significantly reduced liver fat as measured by MRI-derived proton density fat fraction (MRI-PDFF): This score dropped from a baseline 16.2% to 11.3% in the empagliflozin group and from 16.4% to 15.5% in the placebo group (p<0.0001). Empagliflozin was superior to placebo in lowering fat concentration across all segments of the liver scanned by MRI (p-values ranging from 0.002-0.041). Moreover, Dr. Kuchay emphasized that none of the 22 empagliflozin participants saw an increase in liver fat, while 23% experienced no change and 77% saw a decrease. In contrast, 25% of the 20 placebo participants experienced an increase in liver fat, 50% saw no change, and 25% experienced a decrease. Dr. Kuchay highlighted greater weight loss with empagliflozin (6.8 lbs) vs. placebo (3.5 lbs), though this didn’t reach statistical significance (p=0.154); baseline body weight was ~178 lbs in both groups. Importantly, E-LIFT was designed to maintain glycemic equipoise, so that any benefits to liver fat could be attributed to the empagliflozin molecule rather than glucose-lowering. As such, the A1c difference between groups was not statistically significant after 20 weeks (p=0.88), though patients experienced some A1c decline on average regardless of randomization. From a baseline A1c of 9%, Dr. Kuchay reported a drop to 7.2% in the empagliflozin arm and to 7.1% in the placebo arm (again, these participants continued on their regimen of metformin, SUs, DPP-4s, and/or insulin). He also explained that while this was an open-label study, radiologists were blinded. Despite its small size, E-LIFT has piqued our curiosity about possible applications of empagliflozin (and other SGLT-2s) in NAFLD/NASH, which remains a therapeutic area of high unmet need (no FDA-approved therapies to-date), and which is commonly comorbid with type 2 diabetes. We note that Lilly has a GLP-1/glucagon dual agonist in phase 1 for NASH (as well as type 2 diabetes), so this is apparently a disease area that the company is interested in exploring.

5. In Basic Science, Dr. Shulman Positions Hepatic Accumulation of Diacylglycerol as the Cause of Insulin Resistance; Suggests Widely-Applicable Drug for Diabetes, Obesity, NASH, Hyperlipidemia in an Agent That Elevates Mitochondrial Fat Oxidation in Liver

Yale’s Dr. Gerald Shulman discussed an emerging paradigm of insulin resistance, in which intracellular diacylglycerol (DAG) levels are directly related to insulin resistance via protein kinase C-epsilon (PKCe) signaling. As Dr. Shulman reviewed, the initial physiological defect in diabetes is moving glucose into the cell (leading to increased gluconeogenesis), and the rate-controlling step in this process is the insulin-dependent GLUT4 transporter. Dr. Shulman suggested that correcting insulin resistance will require GLUT4 activation. His team has found that intracellular fat in muscle cells is the best predictor of insulin resistance across diverse samples, theoretically due to accumulation of lipids blocking the insulin signaling cascade. Importantly, high fatty acid flux into the cell increases DAGs, which in turn upregulate PKCe. Downstream, PKCe is responsible for decreased GLUT4 translocation to the membrane, but things go one step further in fatty liver disease: PKCe may also inhibit the receptor tyrosine kinase activity of the insulin receptor, further exacerbating insulin resistance. In this sense, “fatty liver” actually precedes the development of diabetes. These results imply that we might need to modify our traditional understanding of insulin signaling in the liver: While the direct signaling pathway in which insulin stimulates glycogen synthase stands, there’s also indirect, inhibitory allosteric regulation via lipolysis. Indeed, knocking down the canonical insulin signaling pathway (Akt2 and FOXO) still gives insulin-dependent inhibition of hepatic gluconeogenesis. These observations have been duplicated across a wide range of studies by different researchers, including in humans. Particularly interesting to us was one of Dr. Shulman’s own projects: In a sample of 400 healthy, lean undergraduates, there was a naturally wide spread in insulin sensitivity, and those with lower insulin sensitivity (lowest quartile) had more intracellular lipid storage than the top quartile. As such, insulin sensitivity is implicated in the way ingested glucose is stored: Highly insulin sensitive people store glucose as glycogen, but insulin resistant folks become hyperinsulinemic and convert glucose to fat, leading to dyslipidemia and predisposing patients to NAFLD/NASH. That said, exercise does seem to activate glucose transport, increasing glycogen synthesis and decreasing lipogenesis after a single session. We see interesting implications in this work for identifying individuals at high risk for cardiometabolic diseases – could insulin resistance be monitored over time as an early warning sign? Turning to treatment, Dr. Shulman suggested that drugs should strive to ramp up mitochondrial fat oxidation in the liver to burn fat and lower gluconeogenesis – one could imagine the wide applicability of an agent like this in NASH, obesity, diabetes, and hyperlipidemia, and Dr. Shulman hinted that various pharmaceutical companies are working on a number of candidates on this front.

Cost and Epidemiology Highlights

1. US Diabetes Spending Up 7% Per Year for Each of Past 15 Years; More Spend on Supplies/Meds, Less on Hospitalizations; Insulin Price Inflation; Drives of Cost

University of Michigan’s Dr. William Herman provided an in-depth review of diabetes spend in the US, highlighting that since 2002, total US diabetes costs have consistently risen ~7% per year. According to the most recent IDF Diabetes Atlas, the US spent $348 billion on diabetes in 2017. Of course, this is not just a US problem (though the US did account for 48% of global diabetes dollars last year) – Dr. Herman cautioned that the projected increase in global healthcare expenditures attributable to diabetes and its complications will be unsustainable. Dr. Herman reviewed a number of recent trends in diabetes spending (more on supplies and medications, less on institutional care), insulin price inflation, and drivers of cost (half due to increase in prevalence, half due to increase in cost of care).

• In 2017, the US spent significantly more per capita on diabetes medications and supplies, and less on institutional care, than it did in 2007. Remarkably, the proportion of US per capita diabetes costs attributed to medications and supplies skyrocketed by 88%! In 2007, medications and supplies were responsible for just 24% of per capita healthcare costs – this figure rose to 45% in 2017. Meanwhile, institutional care, including hospitalizations for complications, comprised 57% of diabetes costs in 2007 and 32% in 2017. Interestingly, outpatient expenditure has remained fairly consistent, but these costs include prevention efforts. Grounds to make claims of causation are unstable, but it seems that the country’s increased spending on diabetes management could be leading to less inpatient utilization (for complications, hypoglycemia) – this is an exceedingly positive trend! These findings align nicely with a recent JAMA article finding that US healthcare spending is nearly twice that of other high-income countries and that the spending difference is driven not by greater utilization, but rather by higher price of services.
• Dr. Herman attributes greater US spend on medications and supplies to the rising costs of insulin – also a major focus of Dr. Irl Hirsch’s annual DT&T rant. He presented data showing that dollars spent on insulin rose six-fold in the US between 2002 and 2012, increasing four-fold on a per capita basis. In one particularly telling example, the price of Lantus rose from $44/vial when it was first introduced in May 2001 to almost $300/vial in November 2014. Dr. Herman offered a couple explanations for the rising costs – more patients are being treated, the total daily dose is higher – but also questioned whether we’re really getting value in return for money spent on newer insulins, especially in type 2. To this end, he cited data from a meta-analysis of type 2s showing short-acting analog insulin drives only a 0.03% A1c reduction and 0.17 fewer hypoglycemia episodes/month as compared to regular human insulin. Indeed, we have heard many a reference to patients in the US switching off of their expensive insulins and on to human varieties in the interest of saving cash.
• In the last five years, the ADA estimates that half of the increases in diabetes costs are due to growth of the diagnosed population, with the remaining half due to the rising average cost of care per person. Given these findings, Dr. Herman stated the obvious, but no less difficult truth: We must either prevent diabetes or control the costs (ideally both!). Luckily, cost-controlling follows from prevention. Yet, despite DPP data showing that lifestyle and pharmacologic interventions both work, <4% of continuously-insured patients with diagnosed prediabetes are prescribed metformin over three years, and only one in 10,000 (~8,400) Americans with prediabetes has participated in the National DPP. This up-front investment is a must, considering that we are starting down a diabetes avalanche and prevention is so much cheaper than treatment. 

2. CDC’s Dr. Ed Gregg: Decreased US DM Incidence, No Increase in Prevalence in Last 7 Years, but Still Nothing to Jump for Joy About

CDC’s Dr. Edward Gregg noted that in the last seven years, the US has seen a decrease in diabetes incidence and no increase in diabetes prevalence, but that’s still “nothing to jump for joy about.” He questioned whether increased testing in 1997 might have driven the rise in incidence, or whether the shift to using A1c to diagnose in 2010 might be responsible. However, for the latter, he stipulated that using A1c might actually underestimate prevalence and incidence, so it’s still unclear. Unfortunately, all racial/ethnic groups except non-Hispanic whites have seen increases in type 2 diabetes in youth, particularly American Indians. He found recent amputation trends (see below) to be “quite curious,” as there appears to be a significant spike in toe amputations around 2009-2011. He isn’t certain as to the main drivers, but reasons that it might be due to practice or provider changes in the way ulcers are managed, healthcare reform, increasing obesity, or the financial recession. In a separate talk, Michigan’s Dr. William Herman also mentioned a “curious plateau” for end-stage renal disease in 2009-2010, which may indicate that the previously declining trends might be on their way up. Dr. Herman predicts that diabetes complications will be observed at greater rates in both older and younger adults in the future, as mortality is decreasing and type 2 diabetes in youth is increasing.

Obesity Highlights

1. Dr. Tricia Tan’s Cocktail of Three Gut Hormones Mimics Roux-en-Y Gastric Bypass in Decreasing Caloric Intake, Improving Postprandial Glucose + Time-in-Range

Imperial College London’s Dr. Tricia Tan reviewed her work using gut hormone combination therapy to mimic the hormonal effects of bariatric surgery in people with obesity. Her “medical gastric bypass” cocktail consists of a GLP-1 agonist, an oxyntomodulin analog, and a PYY analog, dosed via pump to mirror the elevated levels of these hormones in people who have recently undergone roux-en-Y gastric bypass. In a small study (n=10), participants without diabetes who received the “medical gastric bypass” infusion experienced near-immediate changes in eating behavior and glycemic control, which is also what happens following bariatric surgery. Dr. Tan showed that calorie intake fell 32% with her gut hormone cocktail. In a longer study over 28 days in people with obesity and diabetes, the treatment significantly reduced body weight (people lost ~4 kg or ~9 lbs on average), improved postprandial glucose excursions, and increased time-in-range according to CGM data. This was a short-term study, so it’s unclear whether long-term use of the triple hormone infusion is also associated with enduring weight loss and improvements in glucose control. Still, the prospect of an injectable drug that offers weight loss and glycemic benefits on par with bariatric surgery is extremely intriguing. Enhanced gut hormone secretion is one of several mechanisms at play in bariatric surgery (alongside gut restriction and malabsorption, shifts in the gut microbiome, increase in bile acid levels, etc.), so it’s unlikely that “medical gastric bypass” can achieve exactly the same impressive weight loss and diabetes remission as surgery, but this approach is compelling nonetheless, as it would be a less-invasive (and potentially more affordable) option, implementable in more patients than surgery. Dr. Tan framed this early research as proof-of-concept that the hormonal changes in bariatric surgery can be replicated with pharmacotherapy. She also highlighted the synergistic effects of these gut hormones, explaining that it took several-fold higher doses to achieve the same reduction in food intake with single vs. triple hormone therapy (3x the dose for GLP-1 agonist monotherapy, 7.5x the dose for oxyntomodulin monotherapy, and 2.5x the dose for PYY monotherapy). From a drug development perspective, this points to combination therapy as a way to achieve better efficacy at smaller doses of the gut hormones (i.e. fewer or milder side-effects). Dr. Tan suggested that her multi-hormone therapy could be even more effective if it was tweaked to also reflect the post-surgery changes in ghrelin and glucagon, though this is a more difficult proposition since the precise nature of how these hormones levels change after surgery is less well-understood.

• Although Dr. Tan’s work involves the continuous infusion of gut hormones, industry has several injectable multi-hormone combinations in development. The GLP-1/glucagon dual agonist competitive landscape has been heating up of late, and features several high-profile candidates. Sanofi will advance its GLP-1/glucagon dual agonist into phase 3 for obesity sometime this year, and management has suggested potential weight loss efficacy on par with Novo Nordisk’s Saxenda (liraglutide 3.0 mg) or semaglutide. Novo Nordisk also has a GLP-1/GIP/glucagon tri-agonist in phase 1 for obesity, alongside other early-stage obesity candidates like a PYY analog, glucagon analog, and FGF21 analog.

2. Dr. Caroline Apovian Calls for “Paradigm Shift” in Obesity Care; Refers to Orexigen Bankruptcy as a “Tragedy”

Obesity Society President Dr. Caroline Apovian issued a passionate call for increased attention to the growing obesity crisis. “We need a paradigm shift in obesity care, and so far, it’s not happening.” Dr. Apovian underscored the need for a greater workforce trained in obesity management, especially in the trenches of primary care and among endocrinologists (where obesity care is synergistic with diabetes care). Among the many challenges in the obesity field – deep-seated social stigma, weight-based bias from providers, under-diagnosis of the condition, and a history of safety concerns surrounding previous obesity therapies – Dr. Apovian identified obesity’s lack of recognition as a true, biological disease as the major systemic barrier to adequate medical care. She likened this to stents for chest pain: “When the cardiologist says you need a stent, there’s no argument. But there is when an obesity management specialist recommends medications or bariatric surgery. Why? We don’t take obesity seriously as a disease.” Beyond this reticence on the part of patients and providers alike about medical management of obesity, Dr. Apovian alluded to access/reimbursement issues. As we know all too well, reimbursement for obesity drugs is extremely poor, leading to low prescription volume and in turn, placing added stress on companies investing in obesity R&D. This problem came to the forefront recently when Orexigen – manufacturer of the obesity drug Contrave (naltrexone/bupropion extended-release) – filed for bankruptcy. Dr. Apovian declared this a “tragedy,” noting that “we should have been seeing a deluge of prescriptions” given staggering obesity prevalence and Contrave’s strong safety/efficacy profile. We couldn’t agree more with Dr. Apovian’s commentary, and we’re left wondering where the momentum will come from for the paradigm shift she describes. In addition to the mounting strength of anti-stigma campaigns, we have a great deal of hope for Novo Nordisk’s upcoming SELECT trial for GLP-1 agonist semaglutide, which will be the first-ever obesity CVOT. No RCT to-date has “proven” that obesity and its complications are responsive to pharmacotherapy, and SELECT could be that trial. Moreover, results could help reinforce obesity’s rightful status as a serious medical condition, thereby boosting provider willingness to prescribe obesity pharmacotherapy and perhaps nudging PBMs/payers toward more generous reimbursement of these agents.

3. Mathematical Modeling Study Maps the Causes Underlying Variation in Body Weight; Implications for Personalized Obesity Therapy

Mayo Clinic’s Dr. Michael Camilleri discussed the extent to which a variety of GI and psychological traits are associated with obesity. Dr. Camilleri’s research team analyzed a barrage of anatomical measures, psychological traits, and physiological features (gastric emptying, fasting and postprandial gastric volume, satiation in response to various meals, gut hormone levels) in 328 adults across the spectrum of normal weight, overweight, and obesity. Using mathematical modeling, the team attributed 80% of the study group’s variation in body weight to the following factors: differences in satiety/satiation (21%), sensorimotor qualities of the GI tract (17%), psychological and behavioral traits (19%), stomach capacity (14%), and peak levels of GLP-1 and other gut hormones (9%). Dr. Camilleri speculated that differences in resting energy expenditure may account for a large part of the remaining 20% of variation in body weight. Beyond reinforcing the heterogeneity of obesity pathophysiology, this work has interesting implications for the personalization of obesity therapy. If HCPs can identify the dominant causes of an individual’s obesity (what Dr. Camilleri termed an “obesity phenotype”), they can correspondingly prescribe treatments that specifically target those mechanisms. For instance, people with a form of obesity that is driven by psychological and behavioral factors may benefit most from pharmacotherapies like Vivus’ Qsymia (phentermine/topiramate), Arena/Eisai’s Belviq (lorcaserin), and Orexigen’s Contrave (naltrexone/bupropion), which work on the brain’s reward system to reduce food cravings. People whose obesity derives from sensorimotor features of the stomach like gastric emptying may benefit most from a GLP-1 agonist – Novo Nordisk’s Saxenda (high-dose liraglutide) is already indicated for obesity, and the company is also investigating second-gen GLP-1 agonist semaglutide for obesity, with a robust phase 3 program slated to begin this year. As another example, people whose obesity comes primarily from their difficulty feeling satiated after a meal may benefit most from a device like vBloc, which directly stimulates the nerves mediating satiety; people who have a large stomach capacity may benefit most from gastric balloons or bariatric surgery. While this was a small study and certainly doesn’t represent the definitive breakdown of all the factors influencing body weight, we found this framework very useful for encouraging a more systematic approach to personalized obesity care.

4. Is Energy Intake the Most Important Factor in Weight Loss? Science Says Yes, But Clinical Experience Says No

In an amicable debate, NIDDK’s Dr. Kevin Hall argued that energy intake is the most important factor in weight loss, while University of Colorado’s Dr. Daniel Bessesen asserted that there is no one “most important” factor in weight loss, because patients need personalized and comprehensive care for the best outcomes. There was little disagreement that overcoming the physiological adaptations that pushes body weight higher does take caloric restriction; in other words, calorie deficit has a leading role to play in weight loss, and possibly in weight maintenance as well. The experts agreed, however, that obesity management is much more complicated than that: Simplifying the problem to “calories in, calories out” does patients more of a disservice than anything. Below, we highlight the argument presented for each side.

• Dr. Hall, a physicist by training, showed how weight loss is predicated on expending more calories than are taken in, adding that this has little to do with the source of those calories. Emphasizing that the endpoint in this debate was weight loss and not improvement in health, Dr. Hall primarily aimed to dismantle the alternative view that carbohydrate consumption drives fat storage by promoting insulin secretion, pulling calories out of circulation, and slowing down metabolism. From this theory arises the notion that fat loss requires carb reduction, and that carb intake is more important than caloric intake for weight management. Dr. Hall cited his own research, including one study in which 19 individuals with obesity were housed in the NIH for two weeks with completely controlled diets. While a low-carb diet reduced insulin secretion relative to a low-fat diet (~20% decrease in 24-hour C-peptide, p=0.001 for comparison), significantly more cumulative body fat loss was seen with the low-fat diet (p<0.001). Moreover, the low-carb diet actually decreased energy expenditure over 24 hours (p<0.005 vs. baseline, p=0.099 vs. low-fat diet), while the low-fat diet did not. That said, Dr. Hall called these data highly statistically significant but clinically meaningless, concluding that calorie source is not relevant in weight loss (though low-carb diets can hold glycemic benefit for patients with diabetes). To reinforce these results, a similar two-month crossover study (n=17 men with overweight/obesity) compared a high-carb baseline diet to an isocaloric ketogenic diet: Despite an impressive ~50% reduction in insulin secretion, there was only a slight transient increase in energy expenditure, and body fat loss actually slowed down following transition to the ketogenic diet. When asked by Dr. Bessesen whether a “calorie is a calorie,” Dr. Hall replied, “with respect to body fat loss and energy expenditure, when it comes to dietary carbs and fat, a calorie is almost a calorie. Dietary protein does seem to have a preferential effect on calorie expenditure and sparing fat-free mass during weight loss.” Further, Dr. Hall clarified that weight loss should not be the sole endpoint in obesity treatment in a real-world clinical setting and that the best diet for anybody, low-carb or low-fat, is what they can adhere to. In the end, Dr. Hall recognized the trend toward weight regain, acknowledging that you can’t just keep telling people to cut more calories as it gets harder and harder to lose weight.
• Rather than directly refuting Dr. Hall’s argument (in fact, he praised Dr. Hall’s research), Dr. Bessesen added critical nuance. In the real world, he argued, talking about calories is not the most important thing a clinician can do. Since the 1970s, scientists have noted the lifelong trend toward weight gain. Evidence is only accumulating to support that maintaining long-term weight loss is incredibly difficult. “If it was just about calories, we wouldn’t have this problem. Weight is regulated,” Dr. Bessesen explained. Indeed, the body doesn’t seem to “get used to” a reduced weight, and human physiology has been designed to prevent weight loss: When you lose weight, you burn fewer calories, improved insulin sensitivity promotes fat storage, and hunger goes up while satiety goes down. A recent statement from the Endocrine Society argues that we can’t even treat obesity because we don’t understand the basic pathophysiology of weight regain. But, as Dr. Bessesen says, we have to be able to offer patients something: We need to oppose the tendency for weight regain, but calorie restriction is not enough. Current tools include behavioral treatment, pharmacotherapy, and bariatric surgery. When it comes to effective behavioral change, readiness, motivation, and ability to implement change are paramount; that said, most providers aren’t really trained in behavioral therapy, and many patients’ circumstances aren’t ideal. Moreover, Dr. Bessesen argued that obesity pharmacotherapy is seriously under-prescribed: “We think of diabetes as a biological disease and prescribe lots of medication; we don’t prescribe drugs for obesity because we’re stuck on it as a behavioral problem” – we couldn’t agree more. Dr. Bessesen advocated that we shouldn’t think of these strategies as calorie restriction tools, but rather, as physiological tools to oppose weight regain. In the end, a flexible, personalized, and comprehensive approach will work best for any given patient, and providers also need to consider diet, exercise, environment, social support, and insurance factors when choosing a weight loss strategy for their patients. Ultimately, Dr. Bessesen maintained that he can’t name any one factor as the “most important” in weight loss, because these will vary from person to person, and treating obesity in any individual will require a multi-pronged approach.

5. Childhood Obesity Expert Dr. Ellen Lancon Connor on Bariatric Surgery in Adolescents: Filling an Unmet Need

In a meet-the-professor session on childhood obesity, Dr. Ellen Lancon Connor positioned surgery as one of the more promising strategies to treat severe obesity in adolescents. Currently, gastric bypass and sleeve gastrectomy are recommended for adolescents, but the lap band is not, even though the latter has been shown to ameliorate or resolve CV risk factors. We’d think CV risk reduction is all the more important in teens affected by obesity, considering the cumulative danger of having elevated CV risk across the lifespan. Dr. Lancon Connor reviewed the Endocrine Society’s guidelines for bariatric surgery, which limit its use to teenagers who are near their final height with a BMI >40 kg/m² or >35 kg/m² with severe comorbidities; surgery is only indicated after intensive lifestyle intervention, and in patients with no untreated psychiatric illnesses. Importantly, she clarified that psychological distress around obesity is not a contraindication for surgery, but providers do need to make sure teens can comply with pre-op and post-op requirements plus long-term follow-up for surgical complications and nutritional supplementation. Additionally, she underscored that the operation should occur in a dedicated adolescent bariatric center capable of providing adolescent-specific multidisciplinary surgical care. Despite her positive opinion on bariatric surgery for adolescent obesity, Dr. Lancon Connor also acknowledged its limitations for this patient population: In cases of extreme obesity, adolescents will likely still be affected by obesity/high body weight post-surgery. The problem will not disappear, and CV risk will remain, in spite of resolved type 2 diabetes, hypertension, and hyperlipidemia. Dr. Lancon Connor thus emphasized the importance of setting realistic expectations, for both individuals and families, about what bariatric surgery entails, what additional actions will be required on the part of the patient/family (surgery is by no means “an easy way out”), and possible outcomes, positive, negative, and neutral. She elaborated that many teens don’t want to make dietary changes alongside the surgery. GI upset is the most common side-effect, but the rate of severe complications (0.4%) does seem to be similarly low in teens, as in adults. All in all, this was a valuable session that encouraged HCPs to at least consider bariatric surgery for their younger patients with obesity, and Dr. Lancon Connor shared very interesting insights. We do hope that bariatric surgery becomes less stigmatized in adolescents who have struggled to lose weight with lifestyle and pharmacological intervention, especially as a preponderance of data builds to indicate that adolescent obesity typically becomes adult obesity. That said, given the rapidly rising prevalence of obesity among youth (according to NHANES data, childhood obesity prevalence in the US grew 33% from ~14% in 1999-2000 to ~19% in 2015-2016), expensive bariatric surgery procedures likely aren’t the best population-level solution. Like Dr. Lancon Connor, we’d love to see bariatric surgery more closely considered for individuals who might benefit, but we also see a tremendous need to improve lifestyle intervention and to boost uptake of obesity pharmacotherapy.

• Dr. Lancon Connor touched on obesity drugs briefly. Currently, the only pharmaceutical treatment available for people with obesity age 12-16 is orlistat, but Dr. Lancon Connor pointed out that only ~10% of patients maintain orlistat treatment beyond six months, and only half continue after one month. While patients who are able to continue orlistat fare relatively well, most high school students don’t find the treatment compatible with their lifestyle, particularly due to GI complaints. Moreover, pharmacotherapy is only recommended after a formal program of lifestyle modification. Looking to the future, GLP-1 agonist exenatide (AZ’s Bydureon), GLP-1 agonist liraglutide (Novo Nordisk’s Saxenda), and phentermine/topiramate (Vivus’ Qsymia) are all under or have been under investigation in childhood obesity, but there has been no regulatory action to-date, so using these therapies for weight loss in younger patients would be off-label for now.

-- by Ann Carracher, Abigail Dove, Brian Levine, Payal Marathe, Maeve Serino, Adam Brown, and Kelly Close