American Diabetes Association 76th Scientific Sessions

June 10-14, 2016; New Orleans, LA; Full Report – Glucose Monitoring – Draft

Executive Highlights

This document contains our coverage of glucose monitoring at ADA 2016. Immediately below, we enclose our themes on the category, followed by detailed discussion and commentary. Talk titles highlighted in yellow were among our favorites from ADA 2016; those highlighted in blue are new full report additions from our daily coverage.

For comprehensiveness, we have included some talks in this report that also overlap with our ADA 2016 Insulin Delivery and Digital Health Full Reports

Themes

  • ADA 2016 was a BIG meeting for sensor outcomes data: Dexcom’s DIaMonD study (testing CGM in MDI) and Abbott’s IMPACT study (testing FreeStyle Libre in well-controlled type 1s) both impressed. Both studies represented strong and positive results for Dexcom, Abbott and the entire field, and a signal of how far industry has come and where things are going in the future: proving outcomes.
    • Dexcom’s DIaMonD study randomized MDI users to six months of CGM (n=105) or six months of usual care (n=53). A1c declined a strong 0.9% with CGM at six months vs. -0.4% with usual care (baseline: 8.6%), for an adjusted mean difference of -0.6% in favor of CGM (p<0.001). The advantage for CGM was impressively consistent across age, baseline hypoglycemia, education, and diabetes numeracy – 60+ year-old CGM users saw the same benefit as 25-60 year-old users in this study. At the same time A1c declined, hypoglycemia significantly improved with CGM: a 30% improvement in time <70 mg/dl (-23 mins/day; p=0.006) and a strong 50% improvement in time <50 mg/dl (-11 mins per day; p=0.005), both outperforming 17% and 21% improvements with usual care (-15 mins, -6 mins). While the absolute reductions are not huge here, the high A1c baseline patients were not experiencing an overwhelming amount of hypoglycemia at baseline. On the high end, CGM users were spending 83 fewer minutes per day above range (>180 mg/dl) at 24 weeks, while the usual care group was spending nine more minutes per day above range (p=0.04). That translated to CGM users spending an hour more per day in range (70-180 mg/dl) at 24 weeks, while the usual care group spent 15 fewer minutes per day in range (p=0.006). CGM trended towards less severe hypoglycemia: a 2% rate (two out of 105 patients) vs. a 4% rate in usual care (two out of 53 patients). Glycemic variability also improved a bit with CGM (median CV: 42% to 38%), but did not change in usual care (42% to 42%) (p<0.001). Daily SMBG tests declined as expected in the CGM group (from 5.1/day to 3.6/day), but stayed roughly similar in the usual care group (5.1/day to 4.6/day) (p<0.001). CGM wear >6 days per week was seen in an impressive 89% of patients at six months, a testament to the better technology and the tight adherence criteria (>85% wear) patients had to demonstrate during the blinded CGM phase before randomization.
      • DIaMonD shows that MDI users not at glycemic target can definitely benefit from CGM – getting a meaningful reduction in A1c (-0.9% from baseline), shaving off highs, cutting their time in mild and dangerous hypoglycemia, and improving variability. We hope this large randomized study can help influence more CGM prescribing in MDIs, countering the “pump first” mentality that Dexcom has always battled. More importantly, we hope DIaMonD can influence professional guidelines and further improvement CGM reimbursement. DIaMonD is also a milestone for Dexcom, who has never run an outcomes study, and will need to do more to keep up with Medtronic’s and Abbott’s growing lists. Phase 2 of the study will cross some of the MDI patients over to pumps, so we’ll eventually see if insulin delivery method makes a difference. We want to see tools driving therapeutic change and creating “higher quality” A1cs and this certainly seemed to happen here.
    • Abbott’s IMPACT study compared FreeStyle Libre to SMBG in type 1 patients in very good control (baseline A1c: 6.7 %). The study met its primary endpoint at six months – relative to the control group, patients using FreeStyle Libre spent a striking ~74 minutes fewer per day <70 mg/dl (a 38% reduction; p<0.001). Pre-specified secondary endpoints were particularly compelling – patients using Libre spent ~49 minutes fewer per day <55 mg/dl (a 50% reduction; p<0.0001) and ~33 minutes fewer per day <45 mg/dl (a 60% reduction; p<0.0001). Measures of nocturnal hypoglycemia were also significantly lower with FreeStyle Libre as patients spent ~28 minutes fewer per night (a 40% reduction; p<0.0001) in hypoglycemia. Patients using Libre spent ~22 minutes fewer per day in extreme hyperglycemia > 240 mg/dl (p=0.02) and spent ~60 minutes greater/day between 70-180 mg/dl (p=0.0006). There was not a significant difference in A1c between the groups, and both saw a marginal ~0.15% increase by the end of the study – a positive given that these patients were spending three hours per day in hypoglycemia at baseline! In other words, FreeStyle Libre prompted a higher quality A1c, with one hour less hypoglycemia per day. The factory-calibrated sensor pretty much completely replaced fingerstick testing, suggesting a high level of confidence in its accuracy: SMBG frequency with FreeStyle Libre fell from a mean of ~5.5 tests/day at baseline to 0.5 tests/day (one every two days) at six months, a testament to the real-world accuracy in patients on insulin therapy. This is very good news as it seeks FDA approval, with now two large RCTs (REPLACE at ATTD and now IMPACT) backing up this finding.
      • IMPACT highlights the truly scary amount of hypoglycemia type 1s on insulin therapy are experiencing every day, and the tremendous challenges of dosing insulin as A1c approaches goal – all patients at baseline were spending ~200 minutes per day <70 mg/dl!!! The hypoglycemia data discussed above is very clinically meaningful (-74 minutes per day), and there is still room to improve therapeutic approaches: patients on Libre were still spending two hours <70 mg/dl per day at six months. Was the residual hypoglycemia driven by over-treating hyperglycemia while on FreeStyle Libre (so-called “hyper avoidance” in intensively managed patients)? Could that be managed with better education? Meanwhile, standard of care patients were still spending over three hours <70 mg/dl per day, no big change from baseline. In that sense, the results tell us as much about FreeStyle Libre’s ability to reduce hypoglycemia as they do about the real-world dangers of insulin therapy, especially in “well controlled” patients skating close to the edge; three hours per day in hypoglycemia is downright dangerous, at the same time these patients would be congratulated for getting below 7%. Avoiding lows on insulin therapy is truly difficult as A1c gets below 7%, and we’re not sure that delicate balance is appreciated enough. Libre and other sensors can help quantify that, and we hope clinical decision support software will help HCPs and patients start to titrate insulin in a data-driven way.
  • Where are glucose sensors going? Accuracy and reliability remain absolutely necessary, but they are no longer sufficient in this cost competitive, outcomes-driven, increasingly digital healthcare environment. It was just three years ago that ADA 2013 featured an entire oral session devoted to CGM accuracy and reliability. With multiple companies boasting available or upcoming sensors with MARDs of ~9-12%, the entire field has clearly picked up its game. Now, accuracy and reliability are the minimum criteria for any new sensor, as there is much more to consider: cost; clinical outcomes data to drive reimbursement; fingerstick calibrations; on-body size; prescribing hassle; connectivity, mobile apps, and clinical decision support software; etc. There is still tremendous room to expand the market, but the race is on to offer the most cost-effective and clinically impactful sensor system and to convince payers and patients of the added value. What will payers think of Dexcom’s DiaMond and Abbott’s REPLACE/IMPACT studies? Will they see this technology as very positive return-on-investment and standard-of-care therapy in type 1 diabetes? Will they make it less of a hassle to get on CGM, even in the US (e.g., no more prior authorizations, appeals, documentation)? Will outcomes and healtheconomic studies be the key frontier on which sensor battles are fought?
    • Automated insulin delivery will be a net positive for the CGM field, though with strong standalone sensor outcomes, will payers do a double take? “Wait, do we really need to cover an automated insulin delivery system (pump and CGM) if patients can get good outcomes with a sensor on MDI?” In that sense, do IMPACT and DIaMonD represent a threat to automated insulin delivery reimbursement? It’s hard to know, but these studies put some pressure on AID systems to show an additional, incremental advantage over best-in-class standalone sensors – particularly because improved apps, pattern recognition, and decision support software are going to make standalone sensors much better.
  • Will the FDA approve a non-adjunctive (BGM replacement) claim for CGM? Dexcom gave a persuasive preview of what we might expect at the July 21 FDA meeting. In short, the risk of an insulin overdose with Dexcom CGM was very low (~0-3%) based on the frequency of erroneously reading 20% or more above YSI in its pivotal trials. The retrospective risk analysis analyzed the two accuracy trials of the G4 Platinum + Software 505 (the same algorithm as in G5) in patients 2-17 years old (n=79) and 18+ years (n=51) vs. YSI. Dr. Walker concluded that G5 mobile is safe for diabetes decision-making, the risk for overdosing is likely minimal, and alerts and alarms further reduce the risk associated with non-adjunctive use of CGM. By contrast, patients using BGM for decision making get point in time snapshots, with no reference to direction, rate of change, or alerts and alarms, plus numerous interfering substances (not to mention hand washing and questionable meter accuracy in the post competitive bidding era). We thought it was a persuasive presentation and look forward to the full case Dexcom presents in a few weeks. Insulin dosing isn’t included on any BGM label, and we wonder how the FDA will approach that irony on July 21.

Table of Contents 

Detailed Discussion and Commentary

Symposium: The Value of Continuous Glucose Monitoring in Patients with Type 1 Diabetes Using Multiple Daily Injections – DIaMonD Study

Results

Elena Toschi, MD, PhD (Joslin Diabetes Center, Boston, MA)

Joslin’s Dr. Elena Toschi reported very positive results from Dexcom’s DIaMonD study, which randomized MDI users not at goal (baseline A1c: 8.6%) to six months of CGM (n=105) or six months of usual care (n=53). A1c declined a strong 0.9% with CGM at six months vs. -0.4% with usual care, for an adjusted mean difference of -0.6% in favor of CGM (p<0.001). The advantage for CGM was impressively consistent across age, baseline hypoglycemia, education, and diabetes numeracy – 60+ year-old CGM users saw the same benefit as 25-60 year-old users! At the same time A1c declined, hypoglycemia significantly improved with CGM: a 30% improvement in time <70 mg/dl (-23 mins/day; p=0.006) and a strong 50% improvement in time <50 mg/dl (-11 mins per day; p=0.005), both outperforming 17% and 21% improvements with usual care (-15 mins, -6 mins). While the absolute reductions are not huge here, we’d note the high baseline patients were not experiencing an overwhelming amount of hypoglycemia at baseline. On the high end, CGM users were spending 83 fewer minutes per day above range (>180 mg/dl) at 24 weeks, while the usual care group was spending nine more minutes per day above range (p=0.04). That translated to CGM users spending an hour more per day in range (70-180 mg/dl ) at 24 weeks, while the usual care group spent 15 fewer minutes per day in range (p=0.006). CGM trended towards less severe hypoglycemia: a 2% rate (two out of 105 patients) vs. a 4% rate in usual care (two out of 53 patients). Glycemic variability also significantly improved with CGM (median CV: 42% to 38%), but did not change in usual care (42% to 42%) (p<0.001). Daily SMBG tests declined as expected in the CGM group (from 5.1/day to 3.6/day), but stayed roughly similar in the usual care group (5.1/day to 4.6/day) (p<0.001). CGM wear >6 days per week was seen in an impressive 89% of patients at six months, a testament to the better technology and the tight adherence criteria (>85% wear) patients had to demonstrate during the blinded CGM phase before randomization – that ensured, of course, that patients would actually wear CGM and see benefit during the study. The presentation only said the “latest” Dexcom CGM was used, which we assume means the G4 sensor with Software 505. More study details below!

  • Overall, these are very strong and positive results for Dexcom and the entire CGM field, which has primarily shown data over the years in patients using pumps – clearly the group that can benefit from CGM is much larger. DIaMonD shows that MDI users not at glycemic target can get a meaningful reduction in A1c (-0.9% from baseline), shave off highs, cut their time in mild and dangerous hypoglycemia, and improve variability. We hope this large randomized study can help influence more prescribing of CGM in MDIs, and more importantly, influence professional guidelines. DIaMonD is also a milestone for Dexcom, who has never run an outcomes study, and will need to do more to keep up with Medtronic’s and Abbott’s growing lists. This study enrolled some patients with type 2 and collected healtheconomic data, but neither was presented today – we look forward to seeing much more on this front. Phase 2 of the study will cross some of the MDI patients over to pumps, so we’ll eventually see if insulin delivery method makes a difference. The power of CGM to drive therapeutic change is obviously very high and we also look forward to learning more about what is the best method to identify and drive change.

A1c Results

  • A1c declined 0.9% with CGM at six months vs. -0.4% with usual care, for an adjusted mean difference of -0.6% in favor of CGM (baseline: 8.6%; p<0.001). The effect was similar at 12 and 24 weeks, suggesting the benefits occurred quickly and were maintained at six months.

 

Baseline

Week 12

Week 24

 

CGM

8.6%

7.6%

7.7%

P<0.001

Usual Care

8.6%

8.1%

8.2%

 
  • CGM’s A1c advantage rose to -0.8% vs. usual care in those with a baseline A1c >8.5% (-1.3% vs. -0.5% from baseline-24 weeks), and a -0.4% A1c advantage in those with a baseline A1c <8.5% (-0.6 vs. -0.2% from baseline-24 weeks).
  • Notably, the A1c results were consistent across age: 60+ year-old CGM users saw a -1.0% A1c improvement from baseline, the same as 25-60 year-old users. The usual care group also saw the same -0.4% A1c change in both age groups.
  • A notable 52% of CGM users saw an A1c reduction of >1% compared to only 19% of the usual care group (p<0.001). Only 18% of the CGM group got to an A1c <7%, though that quadrupled the 4% of patients that did so in the usual care group (p=0.02). This was not too surprising, as patients started at a very high baseline (8.6%), making an A1c <7% difficult to achieve. We’ll be very eager in the future to better understand what drives success once patients have a full picture of their glycemic profile.

Time In Range Results

  • At 24 weeks, CGM users were spending an hour more per day in range (70-180 mg/dl ) from baseline, while the usual care group spent 15 minutes fewer per day in range (p=0.006). This translated to an 11% improvement in time-in-range for the CGM group vs. a 2% decrement in the usual care group.

Minutes Per Day in 70-180 mg/dl
(median)

 

Baseline

Week 24

Change

CGM

662

734

+72 minutes

+11%

Usual Care

648

633

-15 minutes

-2%

Hypoglycemia Results

  • Hypoglycemia significantly improved with CGM, including a 30% improvement in time <70 mg/dl (p=0.006) and a strong 50% improvement in time <50 mg/dl (p=0.005); both outperformed 17% and 21% improvements, respectively, in the usual care group. Time in hypoglycemia was relatively low in both groups at baseline (~5% of the day – an artifact of the high A1c baseline), so there wasn’t a huge runway to improve here. [For comparison, Abbott’s IMPACT study saw three hours per day in hypoglycemia at baseline stemming from the 6.7% starting A1c, while this study saw ~1-1.5 hours per day.] There was a slight imbalance at baseline, as the usual care group saw more hypoglycemia, and thus, had more room to improve during the study. Still, the relative improvement was nearly twice as big with CGM for <70 mg/dl, and more than twice as big for <50 mg/dl.

Minutes Per Day <70 mg/dl
(median)

 

Baseline

24 weeks

Change

CGM

76

53

-23 minutes

-30%

Usual Care

90

75

-15 minutes

-17%

 

Minutes Per Day <50 mg/dl
(median)

 

Baseline

24 weeks

Change

CGM

22

11

-11 minutes

-50%

Usual Care

29

23

-6 minutes

-21%

  • CGM drove a 79% reduction in nocturnal hypoglycemia (<60 mg/dl) vs. a 40% improvement for usual care (p=0.005). Daytime hypoglycemia was halved with CGM vs. no change in usual care (p=0.02). Again, the amount of hypoglycemia was pretty low here, but like automated insulin delivery, CGM really offers benefits at night.

Nighttime % of Readings <60 mg/dl

 

Baseline

24 weeks

Change

CGM

2.9%

0.6%

-79%

Usual Care

4.0%

2.4%

-40%

Hyperglycemia Results

  • At 24 weeks, CGM users were spending 83 fewer minutes per day above range (>180 mg/dl ) from baseline, while the usual care group spent nine more minutes per day above range (p=0.04). This translated to a 12% improvement in time-above-range for the CGM group vs. a 1% decrement in the usual care group.

Minutes Per Day >180 mg/dl
(median)

 

Baseline

24 weeks

Change

CGM

687

604

-83 minutes

-12%

Usual Care

725

734

+9 minutes

+1%

Other Results

  • Glycemic variability significantly improved with CGM (median CV: 42% to 38%), but did not change in usual care (42% to 42%) (p<0.001). We wish the investigators had shown modal day plots, but perhaps the difference would have been difficult to detect graphically.
  • CGM trended towards less severe hypoglycemia: a 2% rate (two out of 105 patients) vs. a 4% rate in usual care (two out of 53 patients). This was not a focus of the commentary or conclusions, though it implies less severe hypoglycemia in the CGM group – a long sought after end point, and a clear coup on the payer front. There was no p-value on the slide, and presumably a larger and longer study with a hypoglycemia-enriched population would be needed to show a significant benefit on severe hypoglycemia.
    • There were no DKA events in either group.
  • Daily SMBG test declined by 24 weeks in the CGM group (from 5.1/day to 3.6/day), but stayed roughly similar in the usual care group (5.1/day to 4.6/day) (p<0.001). This is to be expected with CGM, and though insulin-dosing data was not collected, we assume many patients were using readings to dose insulin without a confirmatory fingerstick.
  • DIaMonD saw very good CGM adherence: at week 24, 89% of patients were using CGM >6 days per week. This was fairly consistent with week 4 (94%), down very slightly from week 12 (96%), but better than other studies like STAR-3 (23% adherence on Dr. Wolpert’s slide).
    • In hallway chatter, some pointed out that the pre-randomization criteria required >85% adherence to the blinded CGM during the run-in, ensuring patients would stick with the technology once in the study. This could detract from the “real-world” aspect of the study, though we’ll be interested to hear more commentary.
  • All study results were intention to treat, and DIaMonD saw outstanding retention: of 105 enrolled in the CGM arm, 103 completed the study (98%). Of 53 enrolled in usual care, 53 completed the study (100%). What a way to run a trial!

Study Background and Baseline Characteristics

  • The 24-site US study included patients with a mean age of 48 years, a mean A1c of 8.6%, and a mean BMI of 28 kg/m2 at baseline. Eleven percent of patients had a severe hypoglycemia event in the last 12 months, and 1% had DKA in the last 12 months. The study population was 44% female, 84% white, and 44% had less than a bachelor’s degree.
  • The DIaMonD study randomized MDI users (n=158) not at A1c goal (baseline: 8.6%) to six months of CGM (n=105) vs. six months of usual care (n=53). After a screening period, a two-week run-in with blinded CGM established baseline glycemia in both groups. Patients were then randomized to 24 weeks of CGM vs. usual care (continued SMBG) for 24 weeks. Two cohorts were included – type 1 and type 2 – though only type 1 data was reported today. Healtheconomic data was also collected in phase 1, though not included today. In phase 2 of the study, patients on MDI will switchover to a pump to see if there are any benefits to change insulin delivery method.
    • The study design limited clinical encounters to ensure real-world outcomes. Weeks 1-3 included device initiation, and the CGM group had a clinic visit at week 1 to troubleshoot the device. Both groups received phone calls in weeks 2-3. At weeks 4 and 12, diabetes management visits occurred, and the clinician downloaded devices and reviewed glucose data (either CGM or SMBG) and made insulin adjustments per usual care in both groups. The usual care group had clinic visits to place a blinded CGM placed at weeks 11 and 23. All patients received one initial session of general diabetes education, and the CGM group received some basic advice on how to use CGM data, adjust insulin doses based on trends, etc.

Discussion

Howard Wolpert, MD (Joslin Diabetes Center, Boston, MA)

Dr. Howard Wolpert succinctly summarized the positive implications of the DIaMonD study results: “Clinicians should consider recommending CGM to all patients with type 1 diabetes who have not attained their glycemic goals.” Presumably not at glycemic goal means A1cs over 7%, or A1cs below 7% with too much hypoglycemia, though this was not discussed at length. He noted the consistency of these A1c outcomes with the JDRF CGM trial and STAR-3 trials (~0.5% reduction). However, this trial was more real world, with fewer visits and phone calls than in either the JDRF or STAR-3 trials – see the picture below. He also pointed out the high sensor adherence (>6 days per week) in DIaMonD, which exceeded usage in other CGM studies (in STAR 3, only 23% used the CGM >80% of the time) – this is expected as sensors improve, and would have exceeded adherence using the older Dexcom STS and Seven. CGM compliance, said Dr. Wolpert, is all about the tradeoff between benefits and demands – with better technology now (accurate, reliable, easier to use), the benefits are starting to outweigh the hassles for more patients. Indeed, Dr. Wolpert characterized the glucose-monitoring field as “at an inflection point.” The transition from urine testing to intermittent fingersticks in the DCCT era reduced A1c, but increased hypoglycemia. Now we’re making the transition from intermittent fingersticks to CGM, which reduces A1c and brings fewer hypoglycemia events. To close his remarks, Dr. Wolpert noted that among MDI users in the T1D exchange, 93% are using SMBG alone, and 7% are using CGM. This large randomized study supports the benefits of CGM in those on MDI, and we expect it to drive further penetration of the technology in MDIs.

Panel Discussion

Ms. Davida Kruger: Can you comment on insulin dose changes from baseline to the end of the study?

Dr. Wolpert: There was no difference between the two groups. It speaks to the fact that when utilizing CGM, benefit is gained from the additional insight into glycemic fluctuations, changes in food choices, and bolus timing rather than in insulin adjustments per se.

Dr. Zachary Bloomgarden: Although you said the hypoglycemia differences post-treatment were significant, it looked like the non-CGM group had higher frequencies of hypoglycemia at baseline, can you comment?

Dr. Wolpert: It’s a bit of a statistical artifact related to low-range glucose detection in CGM, which has been seen in other studies as well. So there wasn’t a baseline imbalance, at least not a statistically significant one.

Q: Were there quality of life indices at the start of the study and following six months. How did CGM change lives?

Dr. Janet McGill: Yes. They have not been completely analyzed or looked at, so we don’t have that for you today. But yes, quality of life was done.

Q: Brilliant study. Thanks. Cost-effectiveness data…when is that coming? And do you think that you will stratify the results by duration of prior diabetes? Clearly with the memory (legacy) effect, long-term patients might have to wait quite some time to derive benefit. This study might apply more to newer patients.

Dr. McGill: Health economics will also be looked at. The second phase is ongoing. To the second question – would this be better in newer onset patients to help them with their glycemic memory? This is strictly my opinion: Absolutely, new onset patients should be in tighter control. Long-term patients face safety issues. For them, CGM might have as much to do with safety and variability as it does with regulating exact A1c numbers. Different groups will benefit differently.

Dr. Ahmann: This opens the door for new considerations and paradigms. Now we’re looking at the majority of patients that can benefit from CGM. The earlier patients are exposed to this, the more they learn from it – what happens with activity, foods, etc. shifts over time. It’s going to make a big difference for me and for payers. Age didn’t make a difference. We have this issue in Medicare, and we have more evidence now – this is really important. Hypoglycemia reduction is so important. Hopefully we’ll hear more studies coming on that.

Q: My main issue is whether the patients in this CGM trial were representative of patients in my clinic with poor glycemic control. What was the percentage of patients who did not succeed 14 days of CGM and could not enter study?

Dr. McGill: We only had 15 that did not proceed, who did not complete everything in the first two weeks. These patients were required to monitor fingersticks, so it was just getting patients to do fingerstick monitoring.

Dr. Ahmann: 44% of patients did not have a college degree. This was not a population that had tremendous social opportunities. My experience, I think more patients that didn’t make it past the first phase because they didn’t check their glucoses enough as opposed to the number who couldn’t use CGM.

Dr. Mark Evans: Very exciting data. I guess the question we all, including yourselves, want answered is: What is it that patients did differently as a consequence of the CGM information? Not that you have an answer to that. The one thing you asked of patients is for them to adjust meal-time insulin, can you talk about that?

Dr. Ahmann: Carb counting was identical between the groups, but one group was guided by CGM, so that’s the difference. But if you’re talking about adjustments according to arrows, of course that’s specific to CGM. Individual HCPs had opportunities to prescribe for individuals in terms of how they interpreted and responded to arrows.

Q: What specific algorithms were provided within the two arms? Do you give recipes for dose adjustment – 10%-20% for one or two trends arrows? And what about the control arm – how much postprandial testing did they do? Was it all individualized?

Dr. Wolpert: We gave them guidelines used in other studies, but not evidence based in terms of arrows – a 10% or 20% increase. The control group was no given prescription around postprandial monitoring to optimize dosing. They did more fingersticks in the actual study.

Q: What do you tell the patients to make them better able to deal with hyperglycemia and also avoid hypoglycemia? What’s the advice-linkage based on interpretation of hypoglycemia data, and can we standardize that?

Dr. Wolpert: You alluded to some important concepts. When people use CGM, insulin dose timing is crucial. The other big area in our approach is to look at composition (glycemic index) of food. We have seen people change food choices due to CGM. We all reiterate to patients that they should avoid dose stacking, and we highlight the pharmacodynamics of insulin, which can be dangerous. We need to ensure that people don’t over-treat lows because when they go low, if they treat, the CGM lags and they think they are still hypoglycemic. Patients need to be prompted around that so they don’t over-treat their lows. Those are the key differences between CGM and fingerstick users.

Dr. McGill: You have carb counting, dose adjustment, sensitivity factors, and the arrows. The CGM taught them perhaps as much or more. We saw the usual care group, and the effect of that type of education, which is good. It’s that delta, and it’s not a particular algorithm. The CGM taught the patients. There wasn’t a manual that told educators to do exactly this. It was very standard education.

Q: Do you have any data on the number of injections at the end of treatment for both groups? How many additional rapid-acting injections did patients on CGM perform? Do you have any data on the basal/bolus ratio at the end for each group?

A: Unfortunately since we don’t have devices like Bluetooth pens, we don’t have specific dosing information.

Posters

Using Novel Flash Glucose-Sensing Technology Reduces Hypoglycemia in Individuals with Type 1 Diabetes (868-P)

Jan Bolinder, Ramiro Antuna, Nel Geelhoed-Duijvestijn, Stephan Matthaei, Raimund Weitgasser

In a poster presentation, Abbott shared long-awaited topline results from the FreeStyle Libre IMPACT study, a randomized six-month trial comparing Abbott’s FreeStyle Libre to SMBG in type 1 patients on MDI/pump therapy in very good control (baseline A1c: 6.7 %). The study randomized 241 type 1 patients 1:1 to use either capillary blood glucose testing (n=120; FreeStyle Lite) or real-time use of FreeStyle Libre (n=121). The data will be presented in full tomorrow (1 PM poster presentation; 7 PM symposium), but we wanted to bring you our analysis in a timely fashion following the poster’s appearance this morning.

After REPLACE, we were pleased to see that IMPACT’s primary endpoint was met at six months – relative to the control group, patients using FreeStyle Libre spent ~74 minutes fewer per day <70 mg/dl (a 38% reduction; p<0.001). There was not a significant difference in A1c between the groups by study end, though both saw a 0.15% increase from baseline to six months (6.7% to 6.9%). Certainly, the “quality” of A1c was improved immeasurably. Pre-specified secondary endpoints were particularly compelling – patients using Libre spent ~49 minutes fewer per day <55 mg/dl (a 50% reduction; p<0.0001) and ~33 minutes fewer per day <45 mg/dl (a 60% reduction; p<0.0001). These are remarkably important indicators since changes in a small number of minutes spent in that dangerously low range could mean many healthcare dollars saved annually; we have to imagine that that is one of the most compelling takeaways for Abbott and would form the crux of any reimbursement strategy. Measures of nocturnal hypoglycemia were also significantly lower with FreeStyle Libre as patients spent ~28 minutes fewer per night (a 40% reduction; p<0.0001) in hypoglycemia. The data here help counter the argument that the device’s lack of alarms poses a nighttime danger; we now have two sets of solid evidence that data from Libre can help patients and providers identify some nocturnal hypoglycemia and make change accordingly – good news for Abbott. Patients using Libre spent 22 minutes fewer per day in extreme hyperglycemia > 240 mg/dl (p=0.02) and spent ~60 minutes greater/day between 70-180 mg/dl (p=0.0006). FreeStyle Libre pretty much completely replaced fingerstick testing, suggesting a high level of confidence in the factory-calibrated sensor: SMBG frequency with FreeStyle Libre fell from a mean of ~5.5 tests/day at baseline to 0.5 tests/day (one every two days) at six months, a testament to the real-world accuracy in patients on insulin therapy. Quality of life data was significantly in favor of FreeStyle Libre. There were no device-related serious adverse events and 13 instances of minimal adverse events (e.g., infection, allergy) reported by 10 subjects.

  • The positive results of IMPACT are not a surprise after REPLACE, and highlight the high and very unsettling degree of hypoglycemia type 1 patients on insulin therapy are experiencing every day. The hypoglycemia data discussed above is very clinically meaningful (-74 minutes per day) – notably, there is still much room to improve time spent in hypoglycemia, given that patients on Libre were still spending two hours (!) <70 mg/dl per day at six months. Presumably when patients are receiving active advice from HCPs on how to change therapy (particularly to reduce hyperglycemia), this would improve further. Meanwhile, control group patients were still spending over three hours <70 mg/dl per day! In that sense, the results tell us as much about Libre as they do about the real-world dangers of insulin therapy, particularly in these well controlled, motivated patients who are clearly still skating close to the edge. Avoiding hypoglycemia on insulin therapy is truly, truly difficult as A1c gets below 7%, and we’re not sure that delicate balance is appreciated enough. The message here is to give patients and providers better tools to detect hypoglycemia patterns and titrate insulin therapy better, accordingly.
  • These data also underscore the limitations in using A1c as an endpoint for diabetes technology (this study did not, unlike REPLACE). Both groups saw a 0.15% increase in A1c in this study (6.7% to 6.9%), though the quality of A1c improved markedly in the FreeStyle Libre group – 74 fewer minutes per day <70 mg/dl (a 38% reduction), and ~33 fewer minutes per day <45 mg/dl (a 60% reduction in time spent at a highly dangerous level). Meanwhile, the control group was still spending three hours per day in hypoglycemia with an A1c 6.9%! Do payers appreciate the gravity of that change? How could that change impact short and long-term costs? Some may criticize that Libre did not improve A1c in this study, though we think it would have been unlikely: the population had a very low A1c to start (6.7%), there was no mandatory per-protocol insulin adjustment, and the baseline hypoglycemia was simply too high to drive things down further. While we do not suggest comparing results since the technologies were not used head-to-head, some will note that the MiniMed 670G pivotal reduced A1c 0.5% from a higher 7.4% baseline and drove a 40% reduction in hypoglycemia – that Medtronic study did not have a control group, and time spent in hypoglycemia was 1.4 hours per day during the 670G baseline compared to 3.4 hours per day in this study. Overall, it’s terrific to see multiple technologies focused on creating greater safety and higher-quality A1cs for patients – patients will choose technologies based on cost, convenience, etc.  
  • Presumably, Abbott can secure some FreeStyle Libre reimbursement with these outcomes studies. Even nearer term, they could help secure FDA approval of the real-time version of FreeStyle Libre, with now two robust, large RCTs in the bank showing patients on Libre basically stop doing fingersticks.
  • We highly commend Abbott for conducting these two long-term outcomes studies to put its glucose monitoring technology to the test in a real-world setting – REPLACE focused on reducing A1c in poorly controlled type 2s, and IMPACT focused on cutting hypoglycemia in well-controlled type 1s. Notably, neither study mandated insulin adjustment, leaving it up to clinicians and patients to make therapy changes as they saw fit. That ensured a real-world study, though perhaps left some efficacy on the table – it also reinforces in the “real world” how often patients are told they are in “good control” even with a significant amount of hypoglycemia.

Study Design

  • IMPACT (ClinicalTrials.gov Identifier: NCT02232698) randomly assigned 241 type 1 patients 1:1 to use either capillary blood glucose testing (n=120; FreeStyle Lite) or sensor glucose data (n=121; FreeStyle Libre) for optimization of glucose levels. All patients entered the study as regular blood glucose testers (~5 fingersticks/day). During the study phase, patients in the intervention arm reviewed FreeStyle Libre Software summary reports (Ambulatory Glucose Profiles) with their clinician at regular intervals in order to make therapy adjustments, while those in the control arm reviewed diary readings with their clinician at similar intervals.
  • Notably, insulin dose adjustments were made on an intention-to-treat basis. Providers were instructed to optimize therapy as they saw fit, but there were no A1c targets or previously mandated dose adjustments. This made it a very real-world study, and what we might expect from FreeStyle Libre in routine use.
  • Patients at baseline had a mean age of 44 years, a mean A1c of 6.7%, a mean 22 year duration of diabetes, a mean BMI of 25 kg/m2, and a mean self-reported blood glucose frequency of ~5.5 per day. Roughly 67% of patients (n=161) were on MDI. In short, this was a population with high potential to reduce hypoglycemia meaningfully if they were acting on data, but a challenging one in which to show an A1c benefit.

Results

  • All measures of hypoglycemia were significantly lower following intervention with FreeStyle Libre vs. SMBG. Specifically, patients: (i) spent ~74 minutes fewer/day <70 mg/dl (p<0.0001); spent ~49 minutes fewer/day <55 mg/dl (p=0.001); (iii) spent ~28 minutes fewer/day <45 mg/dl (p<0.0001); and (iv) spent ~33 minutes fewer/night (p<0.0001). These are very important data given the costs associated with hypoglycemia, and the true dangers of spending any time at all <45 mg/dl.
    • Despite the encouraging trend above, we were struck by the prevalence of hypoglycemia at all levels. All patients at baseline experienced far-too-high ~200 minutes per day < 70 mg/dl, while patients on Libre experienced a still-far-too-high ~120 minutes per day < 70 mg/dl at six months.

Table: Hypoglycemia Data

Glucose Level

Difference vs. control in change from baseline (minutes)

Difference (vs. control) in change from baseline (%)

P-value

Time < 70 mg/dl

~74  mins

-38%

p<0.0001

Nocturnal time < 70 mg/dl

~33 mins

-40%

p<0.0001

Time < 55 mg/dl

~49 mins

-60%

p<0.0001

Time < 45 mg/dl

~33 mins

-50%

p<0.0001

  • There was no significant difference in A1c between the groups, though both saw a slight numerical increase (0.15%). This is actually a good thing, given the incredibly low average and the ridiculous amount of hypoglycemia all patients saw at baseline. By the end of this study, the FreeStyle Libre group clearly had a higher quality A1c!

Table: A1c Data

Intervention

Control

Adjusted Mean

p-value

Baseline

Final

Baseline

Final

6.79%

6.94%

6.74%

6.95

0.00

0.96

  • Patients using Libre spent ~22 minutes fewer/day > 240 mg/dl (p=0.02) and spent ~60 minutes greater/day between 70-180 mg/dl (p=0.0006). We were glad to see time-in-range and time above range improve, suggesting the reduction in hypoglycemia was going to more quality time spent in an ideal glucose range.  

Table: Time-in-Range / Hyperglycemia Data

Glucose Level

Difference vs. control in change from baseline (minutes)

P-value

Time > 240 mg/dl

~22 mins

p<0.0001

Time between 70 – 180 mg/dl

~60 mins

p<0.0001

 
  • As we saw in REPLACE, FreeStyle Libre pretty much completely replaced blood glucose testing, suggesting a high level of confidence in the factory-calibrated sensor. SMBG frequency with FreeStyle Libre fell from a mean of ~5.5 tests/day at baseline to 0.5 tests/day (one every two days) at six months. The trend is a testament to the real-world accuracy of FreeStyle Libre in patients on insulin therapy.

Figure: Number of FreeStyle Libre Scans and Blood Glucose Tests Per Day

  • Patients in the FreeStyle Libre arm scanned for glucose 15.1 times per day, which is more than once every two waking hours. This is certainly more real-time glucose data than they were getting with SMBG! As a reminder, FreeStyle Libre’s label recommends confirmatory fingersticks in the EU: (i) during times of rapidly changing glucose; (ii) when hypoglycemia or impending hypoglycemia is reported by the system; or (iii) when symptoms do not match the system readings. However, these data continue to remind us that patients do not do fingersticks with FreeStyle Libre in the real world, something we’ve been hearing since the system launched.
    • The control group maintained their level of blood glucose testing through the study – baseline: ~5.5 test/day; six months: 5.6 tests/day.
  • The poster also shared positive user experience data of FreeStyle Libre from the study. There was no background on how these questions were asked – we assume Yes/No – and we’re not how large the absolute improvements were on these scales. Overall, patients reported significantly improved satisfaction, convenience, and flexibility associated with Libre, and reported significantly greater satisfaction with their understanding of diabetes. They were significantly more likely to recommend Libre and to be satisfied continuing Libre use.
    • DQol satisfaction with treatment score and total worry score significantly improved with FreeStyle Libre vs. control (p<0.0001).
    • Perceived frequency of hyperglycemia was significantly improved with Libre vs. control (p<0.0001).
    • Diabetes Treatment Satisfaction Questionnaire showed increased treatment satisfaction with Libre vs. control (p<0.0001). We’ve broken down the DTSQ results below:
      • Patients reported significantly improved satisfaction with Libre than with their control baseline therapy (p<0.0001).
      • Patients reported significantly greater convenience with Libre than with their baseline therapy (p<0.0001).
      • Patients reported significantly greater flexibility with Libre than with their baseline therapy (p<0.0001).
      • Patients reported significantly greater satisfaction with their understanding of diabetes when using Libre than when using their baseline therapy (p<0.0018).
      • Patients were significantly more likely to recommend Libre therapy to someone else with diabetes than their baseline therapy (p=0.001).
      • Patients were significantly more likely to be satisfied continuing Libre treatment than they would be continuing their baseline therapy (p<0.0001).

Close Concerns’ Analysis

  • The hypoglycemia improvement definitely lived up to our expectations. All measures of hypoglycemia (day+night and nocturnal) were significantly lower with FreeStyle Libre, and it’s quite evident that these results are meaningful – minutes of dangerous hypoglycemia saved daily could translate to ER visits avoided. The 50% reduction in time spent < 45 mg/dl is particularly striking.
    • Yikes, hypoglycemia is a real concern, and we finally have a tool to measure it. The results document not just the VERY high prevalence of hypoglycemia at baseline (> ~200 minutes per day < 70 mg/dl) but even after six months on Libre (~120 minutes per day < 70 mg/dl). Ultimately, the numbers speak to just how hard it is to dose insulin appropriately AND avoid hypoglycemia while getting A1c under control, a fact that is so often underappreciated. Looking at these numbers, we were reminded that devices do not improve diabetes management in isolation and – from a patient perspective – that accessible, smarter methods of insulin delivery are just as key.
    • IMPACT was designed with reimbursement in mind, and we wonder if payers will appreciate the value of the “higher quality” A1cs achieved here. This is a place where we believe patients can make a difference with advocacy, if they get organized. We’ve heard time and again that payers view A1c as the ultimate endpoint. We hope the time <45 mg/dl data is compelling for showing short-term cost savings. We also wonder if Abbott might be able to leverage its new data management platform LibreView to answer payer’s concerns. Abbott’s long-term plan is to allow Libre glucose data to automatically populate LibreView upon scanning, and we imagine it would provide Abbott the ability to collect real-world personal uploads from Libre users in the same way that Medtronic and Dexcom collect personal uploads from CareLink and Clarity downloads. Medtronic has shown similar Big Data on cost reductions associated with the MiniMed 530G driving less hypoglycemia, and we imagine that Abbott could do the same. If the UHC/Medtronic deal shows one thing, it’s that payers need to be persuaded with data to reimburse technology.
    • Libre has seen impressive uptake, even out-of-pocket. What could these results add? Sales of Libre outside the US have driven four consecutive quarter of operational growth in the International business, though we do believe that reimbursement could be the key that will unlock the floodgates for Abbott. Could these results drive further uptake?
  • All in all, we commend Abbott for breaking the mold in glucose monitoring by conducting now two long-term outcomes study of its device. There is certainly a wealth of evidence here that more frequent, actionable glucose data is beneficial for patients and that A1c in isolation just isn’t the best metric for clinical studies. As the evidence mounts, we hope that this will be reinforced.

Accuracy of a Fourth-Generation glucose Sensor Throughout Its functional Life (897-P)

R Brazg, M Christiansen, B Bode, T Bailey, S Garg, R Slover, S Huang, J Shin, S Lee

Medtronic presented accuracy data from a pivotal study of its fourth-generation sensor (Enlite 3), to be used with the MiniMed 670G or the Guardian Connect mobile app. The new sensor demonstrated an overall MARD of ~10.5% vs. YSI values measured on days one, three, and seven at 12-hour in-clinic visits (an impressive 23,709 total paired CGM-YSI points). MARD was ~13% on day 1, ~9% on day 3, and ~10% on day 7, with an equal balance of data points collected on each day. Eight-nine participants took part in the study, each wearing two sensors on the abdomen (one paired to the 640G pump and one paired to Guardian Connect mobile app; we have averaged the data for brevity). Enlite 3 was calibrated once at the start of every 12-hour in-clinic visit, and not again unless the device asked for a smart calibration. A separate poster showed that 13% of glucose values (YSI) were <75 mg/dl, 56% were 76-180 mg/dl, and 31% were >180 mg/dl.  Of the 89 study participants, 26 had type 2 diabetes (of whom 16 did not require insulin). Overall, this sensor is a clear improvement over Enlite and Enlite Enhanced, and we’ve heard from 670G trial participants that it is a big upgrade. How it compares to Dexcom and Abbott’s accuracy and reliability is an unknown, but we’re glad to see Medtronic making strides. At some point, further improvements in accuracy (for any company) will offer diminishing marginal value, and we continue to believe the future of CGM innovation will be in dramatically cutting cost, reducing calibrations, improving on-body wearability, and offering valuable software that augments the data. Obviously, it needs to give the data people expect too! Getting the right balance of all these factors is the tricky part, and Dexcom and Medtronic are both talking about multiple product lines with different indications (e.g., the fully disposable Medtronic/Qualcomm professional CGM, Dexcom/Verily). 

  • As we reported at ATTD, the seven-day wear Enlite 3 sensor has an improved algorithm with intelligent diagnostics that determine if it is safe to enter closed loop. The algorithm will also request a calibration when the system detects the overall performance can be improved, and data is not displayed when it detects poor sensor performance

Fifth Generation Continuous Glucose Monitoring Sensor System with Extended Wear and Fewer calibrations (905-P)

J Ulloa, A Varsavsky, R Gautham, I Premakumar

Following its ATTD poster hall debut, a Medtronic poster shared a larger data set on its fifth-generation sensor (i.e., Harmony 1), featuring one calibration per day, 10-day wear, and an overall MARD of 11.4% vs. the Bayer Contour Next Link meter (n=142 sensors, 12,602 evaluation points). This accuracy study included 37 participants with diabetes who wore up to four sensors on the abdomen or arm for 10 days. Meal challenges were administered at three in-clinic session (days 1, 7, 10), and blood glucose measurements were recorded every 15 minutes for three to four hours with the Bayer Contour Next Link meter. Participants were also asked to take 8-10 blood glucose measurements daily when at home. Overall MARD was 11.4% (12.3% on the abdomen and 10.5% on the arm), including a day #1 MARD of 13.3%. Mean absolute difference (MAD) in hypoglycemia (<70 mg/dl) was 11 mg/dl, and 85% of overall points were within 20 mg/dl or 20%. Sensors lasted a mean of 9.5 days, though an unspecified number were removed from analysis early due to adhesiveness or battery failures – the percentage was not specified, and both are critical question marks for Medtronic’s clamshell transmitter design (larger on the body and less secure than Dexcom and Abbott sensors). Roughly 45% of sensors had a MARD of 10% or less. The fifth-gen CGM includes a 90-minute warm up, redundancy via two sensor flexes, a proprietary fusion algorithm to combine the two outputs, and intelligent diagnostics to assist with fault detection and sensor health. Overall, this feasibility data looks encouraging, though the accuracy is behind what Dexcom has said it expects for its ten-day wear, one calibration per day G6 (similar to the current MARD of 9.0%). As of Medtronic’s Analyst Meeting before ADA, a launch of Harmony 1 is expected by April 2019.

Accuracy and Longevity of an Implantable Continuous Glucose Sensor in the PRECISE Study: A 180-Day, Prospective, Multicenter, Pivotal Trial (892-P)

A Senseonics poster shared full accuracy data from the 180-day EU pivotal study of its Eversense implantable CGM system, showing an encouraging MARD of 11.6% vs. YSI, though only 40% of sensors successfully reported data over the 180-day period. The single-arm, multicenter investigation enrolled 71 patients with type 1 diabetes, who had two sensors inserted bilaterally into their upper arm (Clinical Trials Identifier: NCT02154126). At first glance, six-month accuracy was relatively encouraging and consistent with preliminary 90-day data first seen at DTM 2015 – overall MARD was 11.6% with Clarke Error Grid analysis showing 84% of measurements in Zone A and 15% in Zone B (# of paired points unreported). Pre-specified secondary endpoints showed accuracy diminished in the hypoglycemic range (<70 mg/dl), where overall MARD was 22%. The median sensor life was just shy of five months, and only 40% of sensors successfully reported continuous glucose data out to 180 days. Naturally, the findings beg the question of how much this adds to Eversense’s product profile, and how it will stand up to next-gen offerings from Abbott, Dexcom, and Medtronic. We're always glad to see expanded options and form factors, since CGM penetration is far too low in Europe, in the US, and globally. We look forward to seeing how initial commercialization goes this year in Europe, and what Senseonics can get in its label with this updated long-term data (e.g., 150-day wear? 180-day wear?).

  • As a reminder, Senseonics does have plans to submit a CE Mark amendment for its second-gen transmitter in 3Q16 that features a number of form factor improvements. We wonder whether the updated design will include improvements in durability.
  • Senseonics has long discussed the extension of its sensor life from 90 to 180 days as a priority, though we wonder if these results will support that label. This was previously slated for a 2Q16 CE Mark submission, and we’re not sure if this has happened yet. We assume the 180-day results will not materially impact ongoing launches in Sweden and Norway/Denmark or upcoming launches in Germany, Italy, and the Netherlands. For more details on Senseonics upcoming plans, see our coverage of the 1Q16 update.

Glucommander Outpatient, a Cloud-based Insulin Management Solution Adjusted Insulin Doses and Achieved 2.7% Drop in A1c Percentage Points (84-LB)

John G. Clarke, Bruce W. Bode

A Glytec poster (84-LB) showcased very impressive results from a 41-patient, uncontrolled, 3-month, outpatient study testing its Glucommander insulin dosing clinical decision support software – from a high baseline A1c of 10.3%, patients ended three months with an estimated average A1c of 7.6% (p<0.000001). The study enrolled 41 type 1 and type 2 patients (mean age: 38 years, BMI: 32 kg/m2) at Dr. Bruce Bode’s clinic in Atlanta, who were treated for 12 weeks with Glucommander Outpatient. The cloud-based software provided periodic insulin dose titration recommendations to a provider based on analysis of a patient’s SMBG glucose data, communicated wirelessly via the cellular-enabled Telcare meter. The provider then communicated the new insulin doses to patients via text message or email. The topline findings from this small study are very impressive – patients using Glucommander saw a 2.7% reduction in A1c (baseline: 10.3%) at three months, and only 1.6% of blood glucose values were <70 mg/dl. Strikingly, no values were <40 mg/dl and, on the human factors side, patients satisfaction results indicated that 96% of patients would recommend the service to family and friends. The poster hinted at Glytec’s strong long-term data as well, citing a smaller cohort of patients that have continued on Glucommander for six (n=14) and nine (n=5) months and have maintained this 2.7% reductions. Small cohorts, but still, this is a whopping improvement. The outcomes are encouraging given the challenges of titrating insulin and the potential for this software to scale expertise, though larger prospective randomized clinical trials are needed to confirm these positive early findings from an uncontrolled study. The company does plan to begin a larger study that includes cost-related metrics such as readmissions, emergency room visits, medication adherence, and healthcare provider productivity, and we’re hopeful that data will show this kind of clinical decision support is very warranted (a “no-brainer” many say). We’re not sure what the business model looks like going forward, but assume Glytec’s in-hospital experience will be very valuable as it thinks about going outpatient. As a reminder, Glucommander Outpatient is already FDA-cleared and is in the process of being deployed across the US. See our previous in-depth coverage here.

  • How could the Glucommander software be scaled? Could it be packaged with existing devices or even drugs? We’ve long thought that insulin-dose titration is a missing piece in the diabetes data ecosystem, and this early data shows how much can be done (and parallels what Hygieia has shown in Europe). We wonder how this Clinical Decision Support software could be packaged with existing devices or even drugs on the market to enhance their effectiveness in the hands of providers. We also have to assume this product saved tremendous provider time, and we look forward to seeing larger studies showing cost-effectiveness. This is where we see digital health really driving better outcomes: collecting data seamlessly and making valuable recommendations that drive seriously better outcomes with less effort.

Barriers to Device Uptake in Adults with Type 1 Diabetes (914-P)

M Tanenbaum, S Hanes, K Miller, D Naranjo, and K Hood

This study invited 1,503 adult patients (mean age=35 years) in the T1D Exchange to take a 30-minute web-based survey in order to understand barriers to insulin pump and CGM uptake. Coming into the survey, 32% of patients reported using both a pump and CGM, while 5% used just a CGM, 38% used just a pump, and 25% used neither. Unsurprisingly, a majority of patients cited financial burden as the biggest barrier to the use of either device – 60% of patients expressed concern about insurance coverage, the cost of the device, and the cost of supplies. Other popular barriers appeared far more modifiable: 35% did not like having diabetes devices on their bodies, 47% did not like the hassle of having to wear the device all of the time, 26% did not like how the diabetes devices looked on their bodies, and 20% were nervous that the device might not work. The survey also asked patients who had discontinued use of their devices for their rationale – patients reported abandoning CGM because of too many alarms, inaccurate data, a distaste for the device on their body, time requirements, or discomfort, while a majority of patients discontinued use of pumps because they didn’t like the device on their bodies or because the device was uncomfortable. We wonder how attrition breaks down by manufacturer. Younger adults (18-25 years old) were less likely to use devices than older adults, and this younger population had higher levels of diabetes distress and higher A1cs. Overall, findings suggest that cost remains the biggest barrier to address, but size on the body is not far behind. We wonder if many of the CGM quitters were on earlier systems that were less accurate (e.g., Seven Plus), and perhaps they would be less frustrated with the more accurate out now or coming soon. Of course, with self-reported data, there is always some question about the reliability of results, though the data echo much of what we hear about the real-world barriers to device uptake anecdotally.

  • We’d note that CGM users in this study were five times more likely to be on a pump (38% used pump+CGM vs. 5% used MDI+CGM), echoing what Dexcom has long said – patients are more likely to be prescribed a CGM if they are already on a pump. We hope the positive results from Dexcom’s DIaMonD study can change that (see Drs. Howard Wolpert and Elena Toschi’s talks elsewhere in this report).
  • The survey also compared the differences between CGM users and non-users. CGM users were, on average, five years older than non-users (38 years vs. 33 years; p <0.001), viewed technology more favorably, and had significantly lower A1cs (7.3% vs 7.7%; p =0.003).

Comparison of the New Factory Calibrated Sensor with Existing CGM Sensor (896-P)

J Kesavedev, P Pillai, A Shankar, A Ashok, G Sanal, R Iyer, G Krishnan, S Jothydev

This small, poorly detailed study compared the accuracy of Abbott’s factory-calibrated FreeStyle Libre Pro blinded sensor (approved in India in March 2015) with that of Medtronic’s blinded iPro2, finding that accuracy from the two sensors do “not appear comparable.” Researchers simultaneously tested the devices in type 1 diabetes (n=2), type 2 diabetes (n=2), gestational diabetes (n=2), prediabetes (n=2), and non-diabetic (n=2) subjects for approximately a week. Both sensors provided similar trends, but the difference in readings was apparently “substantial enough to alter clinical management.” Disappointingly, there wasn’t any quantification of the accuracy of either device relative to the reference SMBG device, and there were not many SMBG-paired points based on the chart. The poster shared a trace from one patient (see below), suggesting that variability in readings occurred near maximum and minimum glucose values that would impact real-world decision-making. To our eyes, it looks like the sensors generally moved in tandem, with Libre Pro dipping lower than iPro2 at some points. The poster noted that similar variation was seen in six out of ten patients, though didn’t break this out by cohort – we wonder whether there were patterns based on stage of type of diabetes. Ultimately, it’s extremely difficult to read too far into the results given the tiny sample size and limited insight into study design: How many paired points? What was the actual quantified accuracy and how does it compare in hypoglycemia, euglycemia, and hyperglycemia? Was the iPro 2 calibrated accurately? Were differences in accuracy related to the delay associated with the wear location of the sensors (Libre Pro on arm vs. iPro 2 on abdomen)?

Oral Presentations: Closing the Loop on Insulin Management – Are We There Yet?

Similar Estimated A1c Results Reported between Patients with Diabetes Using CGM whether on Multiple Daily Injection (MDI) or Continuous Subcutaneous Insulin Infusion (CSII) (81-OR)

David Price, MD (Dexcom, San Diego, CA)

Dexcom’s Dr. David Price shared a retrospective database evaluation from six months of G4 Share users, showing no differences in mean glucose, estimated A1c, or glucose variability between pump (n=939) or MDI (n=648) users. The de-identified data on glucose values were supplemented by customer info (age, insulin delivery) that Dexcom collects. The pump and MDI groups had identical average glucose values and variability across every age group (from 2-6 year-olds all the way to 65+ year olds), with just a single small exception: glucose variability was statistically significantly lower in adolescents (13-18 years) using injections. As would be expected, adult CGM users had better average glucose values vs. pediatric CGM users by ~30 mg/dl (see tables below estimated from the charts shown). This analysis also aligns with results from the T1D Exchange showing that in every age group, the same pattern holds – similar A1cs for CGM users on MDI or a pump. Dr. Price noted that CGM use is increasing (now up to 16% in the T1D Exchange), but is overwhelmingly prescribed to pumpers: of all Exchange CGM users, 85% are on pumps vs. just 15% on MDI. Studies like this also underscore the inherent value in data streaming from devices to the cloud automatically – companies can use it to put data behind their arguments, drive study design, and inform marketing. Medtronic has written the book on this with CareLink, and we expect Dexcom will begin driving this too. 

All Pediatrics 2-18 years

 

MDI
(n=300)

Pump
(n=301)

P-value

Mean CGM Glucose

~180 mg/dl

~180 mg/dl

0.92

Estimated A1c

~7.9%

~7.9%

--

Standard Deviation

~65 mg/dl

~60 mg/dl

0.39

~ Estimated from Bar Graphs

Adults >18 years

 

MDI
(n=403)

Pump
(n=369)

P-value

Mean CGM Glucose

~158 mg/dl

~159 mg/dl

0.55

Estimated A1c

~7.1%

~7.1%

--

Standard Deviation

~57 mg/dl

~59 mg/dl

<0.23

~ Estimated from Bar Graphs

Oral Presentations: Hypoglycemia Potpourri

Risk Assessment of Using the New Continuous Glucose Monitoring (CGM) System for Treatment Decisions (Patients 2 YO+) (349-OR)

Tomas Walker, DNP (Director of Clinical Project, Dexcom, San Diego, CA)

Dexcom’s Dr. Tomas Walker gave a preview of what we might expect at the July 21 FDA meeting to obtain a non-adjunctive (BGM replacement) claim for the G5 CGM. In short, the risk of an insulin overdose with Dexcom CGM seems very low (~0-3%) based on the frequency of reading 20% or more above YSI in its pivotal trials. Dexcom ran a retrospective risk analysis of its two pivotal trials testing the accuracy of the G4 Platinum + Software 505 (the same algorithm as in G5) in patients 2-17 years old (n=79) and 18+ years (n=51) vs. YSI. The analysis looked at erroneously high CGM values relative to YSI (20-30% and 30%+ over) that posed a risk of excessive insulin doses. These events were very rare in the trials, regardless of what glucose range was looked at (200-249, 250-299, 300-349, 350-400 mg/dl). See the table below for how infrequent these events were. Dr. Walker emphasized G5’s strong overall accuracy – MARD of 10% in pediatrics, 9% in adults – which was a bit higher on day one (~11%-13%) in the pivotal trials, but drops to ~8%-9% on days four and seven. He concluded that G5 mobile is safe for diabetes decision making, the risk for overdosing is likely minimal, and alerts and alarms further reduce the risk associated with non-adjunctive use of CGM. By contrast, patients using BGM for decision making get point in time snapshots, with no reference to direction, rate of change, or alerts and alarms, plus numerous interfering substances (not to mention hand washing and questionable meter accuracy). We thought it was a persuasive presentation and look forward to the full case Dexcom presents in a few weeks. We wish that so much focus had been given to patients’ risk of taking insulin using BGM – insulin dosing isn’t included on the BGM label. See our Dexcom 1Q16 and FDA announcement coverage for more on this topic.

Oral Presentations: Management of Hyperglycemia in the Hospitalized Patient

State-of-the-Art Lecture – The Future of Technology in the Management of Inpatient Diabetes

David Klonoff, MD (Mills Peninsula Health Services, San Mateo, CA)

During his state-of-the-art lecture on the future of technology in the management of inpatient diabetes, Dr. David Klonoff briefly referenced the Diabetes Technology Society’s (DTS) new BGM Surveillance Program and recently published DTS Cybersecurity Standard for Connected Diabetes Devices. He shared that the DTS has started testing its BGM Surveillance Program to assess the accuracy of off-the-shelf, FDA-cleared BGMs – we first reported on this yesterday – though he added a new detail that a total of 18 BGMs will be tested whose sales make up 90% of all meter sales in the US. (We still do not know which meters these are, but assume its Abbott, Ascensia, LifeScan, Roche, and popular store brand meters.) Dr. Klonoff also stated that popular hospital BGMs will be tested with the new DTS Surveillance Program to improve the quality of point of care capillary blood glucose testing. We wonder if there is enough overlap between popular point-of-care and hospital meters that this will not require too significant an additional investment – especially considering how long it took the organization to amass the requisite funding for the point-of-care program. In addition, Dr. Klonoff shared that the DTS’s Cybersecurity Standard for Connected Diabetes Devices was published last month, consistent with timing first shared at DTM 2015. As expected, the document contains a set of performance requirements to improve cybersecurity by describing the scope of the cybersecurity challenge, providing a generic framework for how devices need to be protected, and creating an assurance plan for how to get there (e.g., accrediting labs to test devices). He stated his belief that manufacturers who voluntarily test with DTSec will be able to increase their sales and profits, though we’re not entirely positive how results are going to be interpreted, whether industry will buy in to the process, and if government agencies will pay attention. We agree that cybersecurity is an important component of device design, but we’re not sure if DTS is uniquely qualified to perform such testing, or whether extra security standards will negatively impact device design (e.g., passwords, two-actor authentication, etc.). How much will it cost manufacturers to test their devices? Will standards constantly need to be updated to reflect technological progress? If results expose devices with poor security, will FDA take action, or will payers eliminate coverage of those products? For more coverage of the DTS cybersecurity standard, please see our coverage from DTM 2015.

Symposium: The Role of Continuous Glucose Monitoring in Diabetes Management

Is There a Racial Difference in the Mean Continuous Glucose Monitoring Glucose in Relation to the A1c? – Study Background

Rose Gubitosi-Klug, MD (Case Western Reserve University, Cleveland, OH)

Dr. Rose Gubitosi-Klug had the responsibility of introducing the T1D Exchange/Jaeb Center’s recently completed study, funded by the Helmsley Charitable Trust, examining the relationship between mean glucose and A1c in African Americans and non-Hispanic whites. She began by providing an overview of the current evidence that glycation gaps (i.e., A1c levels that do not reflect true mean blood glucose) are much more common than are widely appreciated in practice, noting that there are a number of reasons for this divergence: hemoglobin variants, decreased red blood cell survival, drug interference, etc. Importantly, she pointed to race as an underappreciated variable that may affect this relationship, too, pulling from many studies that have linked race and the glycation gap – e.g., the landmark A1c-Derived Average Glucose (ADAG) trial that first hinted that African Americans might have a lower average blood glucose vs. non-Hispanic whites for the same A1c. She stressed that while underlying genetic variation may account for up to 62% of the variability in the association of mean glucose and A1c, the factors affecting A1c and race have been relatively understudied. Especially considering the consistency of higher A1cs across studies of African Americans and non-Hispanic whites with diabetes, she introduced the Jaeb Center’s specific research question – Does a racial difference exist in the association of mean glucose with A1c between non-Hispanic African Americans and non-Hispanic whites? – as a critical step toward thinking about diabetes management in racial subpopulations.

Is There a Racial Difference in the Mean Continuous Glucose Monitoring Glucose in Relation to the A1c? —Design, Methods, and Baseline Demographics

Robin Gal, MSPH (Jaeb Center for Health Research, Tampa, FL)

Ms. Robin Gal provided an overview of the Jaeb Center’s study design that examined racial differences in mean glucose vs. A1c. The multicenter (five adult sites; five pediatric sties) observational study enrolled 208 patients with type 1 diabetes who wore FreeStyle Libre Pro [blinded] sensor for 12 weeks. Patients were asked to visit their provider every two weeks, with central laboratory measurements (A1c, frustosamine, glycated albumin) at study entry, four, eight, and 12 weeks.

  • The study recruited approximately equal numbers of non-Hispanic African American and non-Hispanic white patients, though the age and gender distributions were slightly imbalanced – driven by a higher number of African American adults and females who enrolled (see below). The groups, though, were well matched in terms of stability of glycemic control during the study as neither group saw any change in glycemia over the course of the study.
    • We would also point out that the white cohort was significantly wealthier, which translated into higher pump and CGM use. These differences strike us as particularly relevant when thinking about differences in A1c between these cohorts, as it is critical to separate these non-glycemic from biological variables.

Table: Baseline Characteristics

 

African Americans (n=104)

Whites (n=104)

Patients with age < 18

34

51

Patients with age > 18

70

53

Female

62%

53%

Mean Baseline A1c

9.1%

8.3%

Mean Final A1c

9.1%

8.3%

T1D Duration

15 years

16 years

Pump use

36%

66%

CGM use

8%

23%

Annual Household Income

 

 

<$50,000

42%

18%

$50,000 - <$100,000

23%

22%

>$100,000

7%

35%

Missing

28%

26%

 

Is There a Racial Difference in the Mean Continuous Glucose Monitoring Glucose in Relation to the A1c? – Results and Conclusions

Roy Beck, MD, PhD (Jaeb Center for Health Research) and Richard Bergenstal, MD (International Diabetes Center, Minneapolis, MN)

Drs. Rich Bergenstal and Roy Beck presented (on behalf of the T1D  Exchange) results confirming that racial differences exist in the relationship between mean glucose and A1c. Specifically, results showed that African Americans had an average 0.3% higher A1c for a given mean glucose level, which varied with glucose level (smaller differences at lower glucoses; greater differences at higher glucoses). Dr. Bergenstal shared that African Americans’ increased rate of glycation is “unlikely to be clinically meaningful.” More importantly, however, he highlighted that there is considerable variation in the mean-glucose A1c relationship irrespective of race: in this study, mean glucose ranged a remarkable 80 mg/dl for a given A1c – e.g., an estimated A1c of 8.0% could correspond to a mean glucose of 120 mg/dl or 200 mg/dl for either cohort. Yet again, it suggests that looking at A1c alone could be dangerously misleading. With this in mind, these results tell more about the importance of variation in A1c with average glucose between individuals than between races.

Dr. Bergenstal emphasized that we would be wise to personalize treatment decisions based on glucose values, something Dr. Irl Hirsch echoed strongly in Q&A: “...there is SO MUCH variability between what any one A1c means in terms of glucose...very few people do fingerstick glucose testing until they go on insulin. We’re making decisions about whether or not a patient should go on a certain therapy based on a number that could be 80 or 90 mg/dl off. The paradigm by which we treat type 2 diabetes is wrong, and we’ve been doing it wrong for the past 30 years. As we read about hypoglycemia in the elderly, it’s not A1c I’m interested in. It’s the glucose. Simply put, I believe we need to use more fingerstick testing in type 2 diabetes.” Discussion during Q&A also focused on the fascinating question of whether glycation differences matter when thinking about outcomes: do African Americans develop complications at the same glucose level that non-Hispanic whites do?

  • Topline results were consistent with the prevailing hypothesis – data showed a 0.8% higher A1c in African Americans vs. Whites that reflected only a 10 mg/dl separation in mean glucose (i.e., instead of the ~23 mg/dl difference that would be expected for a 0.8% difference in A1c). For the same given mean glucose level, African Americans had an average 0.3% higher A1c level. The glycemic difference between the two groups was very clear though Dr. Bergenstal broke this elevated A1c into two possible components: (i) non-glycemic factors [e.g., genetics, biology]; and (ii) glycemic factors [e.g., access to care, technology, social determinants of health]. Importantly, he stressed that the study group has now shown a specific impact of the first component, thereby confirming that biological racial differences in the A1c-mean glucose relationship exist. Dr. Bergenstal termed this a “small but real” overestimation. Notably, this relationship was not seen for fructosamine or glycated albumin, suggesting that those metrics are resistant to this bias.
  • Dr. Bergenstal shared that African Americans’ increased rate of glycation is “unlikely to be clinically meaningful.” He cited that the 0.3% A1c increase (biological) represents less than half of the overall baseline A1c difference (0.8%) between the two cohorts and – more importantly – that there is considerable variation in mean-glucose A1c relationship irrespective of race.
    • Dr. Bergenstal shared findings that mean glucose could range 80 mg/dl for a given A1c – e.g., an estimated A1c of 8.0% could correspond to a mean glucose of 120 mg/dl or 200 mg/dl for either cohort – suggesting that looking at A1c alone could be dangerously misleading. [We’ve seen similar data from Dr. Irl Hirsch before.] With this in mind, the results ultimately tell more about variation in A1c with average glucose than race. Indeed, as Dr. Bergenstal noted, the implication is that while A1c is an established measure of risk of developing complications, we would be wise to personalize diabetes and treatment decisions based on glucose values. This is not groundbreaking by any means – we’ve long known that A1c in isolation simply is not the best metric. However, Dr. Bergenstal went further, stressing that A1c is not simply inadequate but can be dangerously misleading and clinically ignorant (e.g., treating a patient with mean glucose of 120 mg/dl and 200 mg/dl the same because the A1c reads 8.0%).
    • Dr. Beck added valuable commentary to this discussion, emphasizing that A1c does remain a very valuable endpoint in clinical trials as a metric for hyperglycemia reduction. He argued that A1c is a “fabulous” and reliable outcome when groups are randomized to eliminate difference, but that for individual patients, looking at glucose is critical. He also added that an individual’s mean glucose-A1c bias is relatively steady over time, meaning that if you can identify a specific individual’s glycation bias, then you can actually preserve A1c as a valuable clinical metric [e.g., If an individual’s average glucose reads 10 mg/dl lower than what would be predicted based on A1c, then providers can assume that relationship will hold in the future as well.]
  • Our most compelling takeaway from the presentation was the need to move to glucose as a primary metric for an individual’s glycemic control rather than A1c. Providers have been making therapeutic decisions for years based on A1c alone, but these data really hammered home the variance in this metric. Indeed, following the presentation of these data, it is clearer than ever that A1c measurements as a guidepost for therapy titration present risks for clinical decision-making in individuals. UW’s Dr. Irl Hirsch spoke convincingly during Q&A:
    • “Historically, we started using A1c because we didn’t have access to glucose. Over time, we learned all these limitations to A1c and all these other biomarkers. In my practice, we don’t know about the difference between Asian Americans and Hispanics, but in the big scheme of things, there is SO MUCH variability between what any one A1c means in terms of glucose. The bigger concerns I have in term of clinical medicine is that very few people do fingerstick glucose testing until they go on insulin. We’re making decisions about whether or not a patient should go on a certain therapy based on a number that could be 80 or 90 mg/dl off. The paradigm by which we treat type 2 diabetes is wrong, and we’ve been doing it wrong for the past 30 years. As we read about hypoglycemia in the elderly, it’s not A1c I’m interested in. It’s the glucose. Simply put, I believe we need to use more fingerstick testing in type 2 diabetes.” – Dr. Irl Hirsch
  • Discussion during Q&A also focused on the fascinating question of whether glycation differences matter when thinking about outcomes. In other words, do African Americans develop complications at the same glucose level that non-Hispanic whites do? Or are African Americans actually protected by some unknown mechanism and do they develop complications only at higher glucose levels? Speakers pointed to emerging epidemiological evidence that the threshold for complications is not different between races, but stressed that this remains an important unknown at present.
  • Ultimately, we are hopeful that these findings will start a dialogue about the validity of A1c – as a threshold for complications, for defining prediabetes and diabetes, and as a metric in and of itself. Granted, the findings are not new news per se since the insufficiencies of A1c have long been known – however, we loved that the discussion took a very patient-centered turn that we believe has the potential to spur real action. As a reminder, the FDA is holding a workshop on August 29 focused on outcomes beyond A1c (register here), and JDRF has an ongoing program to evaluate measures beyond A1c for assessing clinical outcomes and potential therapies (outcomes expected in mid-2017).
    • Multiple audience members expressed disbelief that we may be misdiagnosing patients in “both directions” – e.g., diagnosing prediabetes in those whose glucose remains in a very normal range and missing diagnosis of folks who have diabetes but whose glucose is elevated above what their A1c would suggest – and we’re hopeful that this indignation like this can spark momentum in the right direction moving forward.
    • Especially in this era of pharmacotherapies and technologies that may not impact A1c but do reduce hypoglycemia and improve time-in-range, the panel’s commentary also made us appreciate the need for more widespread glucose monitoring and tools that capture glucose more accurately – we need more CGM in clinical trials! A big question, of course, is how to validate such a paradigm shift in the eyes of payers and the FDA. Will anyone fund a modern-day prospective study to validate this new framework?

Panel Discussion

Q: I would ask everyone to look at Poster 1446-P. We did an observation study using a VA database that looked at the same thing you have here in a slightly more outcome-focused way. We looked at incident retinopathy and found that Blacks developed retinopathy at around 0.4-0.5% higher A1c than Whites. Those clinical outcome corresponds well with difference in glucose that you have shown.

Dr. Beck: That’s incredibly important. Thank you.

Q: How are we going to use this new information is very important. If we are going to use it as a major diagnostic metric, then the question is where to put the threshold. What if it’s different for women and men? What do we take away from this?

Dr. Bergenstal: I think this session is going to start that dialogue. You are helping us to think about how important this is.

Dr. Beck: It’s very hard to come up with a cut point with A1c unless that cut point is very high. You want the cut point to be valuable and capture at least 95% of the population of interest and have to remember that some of these folks with an A1c of 6.5% might have a mean glucose of 100 mg/dl. So on takeaway here is that setting cut points for diabetes diagnosis with A1c is very challenging.

Q: You mentioned that we should really start to look at more CGM data. But I think we should look at CGM data and A1c because some people glycate more efficiently than others. That’s what we’re learning here and that’s important because we know that there is going to be an impact on other proteins, even if there isn’t on fructosamine and glycated albumin.

Dr. Beck: We actually think that if you measure glucose once with CGM and establish an individual’s relationship to his/her A1c, then you could just use A1c after that. Dr. Darrell Wilson (Stanford University, Palo Alto, CA) has shown this – he looked at the relationship between CGM mean glucose and A1c and demonstrated that this bias was constant six months apart.

Q: How long do you need glucose data to get accurate assessment of someone’s long-term glucose profile?

Dr. Beck: We found that up to 14 days, you got better precision each day in predicting A1c. After 14 days, it didn’t change. Seven is good but after 14 days, you can pretty much have the three-month perspective.

Dr. Bergenstal: Fourteen days is an incredible marker of predicting an individual’s glucose profile.

Dr. Irl Hirsch (University of Washington, Seattle, WA): Historically, we started using A1c because we didn’t have access to glucose. Over time, we learned all these limitations to A1c and all these other biomarkers. In my practice, we don’t know about the difference between Asian Americans and Hispanics, but in the big scheme of things, there is SO MUCH variability between what any one A1c means in terms of glucose. The bigger concerns I have in term of clinical medicine is that very few people do fingerstick glucose testing until they go on insulin. We’re making decisions about whether or not a patient should go on a certainly therapy based on a number that could be 80 or 90 mg/dl off. The paradigm by which we treat type 2 diabetes is wrong, and we’ve been doing it wrong for the past 30 years. As we read about hypoglycemia in elderly, it’s not A1c I’m interested in. It’s the glucose. Simply put, I believe we need to use more fingerstick testing in type 2 diabetes.

Dr. Bergenstal: I wholeheartedly agree.

Dr. Beck: Just to play Devil’s Advocate and stand up for A1c a bit, I would point out that when you are looking at reducing hyperglycemia in clinical trials, A1c is still a fabulous endpoint because groups get randomized. For individual patients, I agree with you that looking at glucose is critical. But for groups, not so much.

Dr. Stephanie Amiel (King’s College London, UK): The epidemiological data suggests that the A1c threshold for complications is not different between races, but we need to look at this for glucose. What are we going to do with the patient who has A1c of 8.0% but a mean glucose of 110 mg/dl? They are NOT going to have same risk as a patient with an A1c of 8.0% and a mean glucose 200 mg/dl. But on the other hand, as far as I know, we simply don’t know that that’s not true. We need data.

Dr. Beck: It seems to be true, but all good studies end up bringing up other questions that need to be answered. Maybe there are data from the DCCT that can answer that.

Dr. Bergenstal: Just think if we had had foresight to use CGMs in all those CVOTs that everyone got so mad about. We currently know about people who got complications and developed complications, but we could know so much more. But that’ll be for the next series of studies.

Dr. Howard Wolpert (Joslin Diabetes Center, Boston, MA): Did you look at all at differences in glycemic variability? Or the relationship between glycation and variability?

Dr. Beck: We didn’t look at that.

Dr. Bergenstal: We would assume that with higher glucoses there is higher glycation but there’s a lot more data to analyze out of this data set.

Dr. Wolpert: My main question is whether advanced glycators are going to have more variability than lesser glycators.

Q: This individual variation is the most important message here if you ask me. Individuals vary in their mean glucose-A1c relationship between each other but tend to be consistent relative to themselves over time. I wonder, if you take glucose and divide by A1c, what that population distribution would look like. It’s supposed to be ~22 but I bet you would get two different distributions for the two races with a different mean.

Dr. Bergenstal: We welcome the advice, and we’ll look at it.

Dr. Beck: You’re right.

Symposium: Digital Health in Diabetes – Hope or Hype?

Is Transforming Diabetes Possible? Lessons from Participatory Design of Digital Health Solutions

Joyce Lee, MD (University of Michigan, Ann Arbor, MI)

Dr. Joyce Lee provided an overview of her team's fascinating Nightscout user survey, highlighting the value of lead-user innovation for driving progress in patient-centered technology – “People in the community will not wait for healthcare providers to come up with technology. They won’t wait for regulators to regulate, or companies to innovate. They need solutions now.” She referenced the strong traction the Nightscout movement has gained over the past few years, noting that since launch in 2013, over 25,000 patients from around the world have reported using the system. According to Dr. Lee, the private Facebook group for CGM in the Cloud now has 18,000 members, making it one of the largest groups representing diabetes on Facebook. Dr. Lee summarized findings from a University of Michigan-led survey of a subset of the Facebook group (n=1,276) – we first saw these results at the 2015 DiabetesMine Innovation Summit and D-Data Exchange, where we were impressed by Nightscout's positive impact on CGM adoption, improved quality of life, and reduction in self-reported A1c. Notably, a striking ~20% of CGM in the Cloud members started CGM as a result of Nightscout and ~9% restarted CGM as a result of Nightscout. Survey respondents reported a significant improvement in quality of life metrics, and a majority stated that they found the technology "extremely empowering" and "not at all prying.” Self-reported A1c also indicated that glycemic control improved meaningfully (0.7%-1.2%) pre- vs. post-Nightscout. In addition, survey results suggested that users largely replaced their BGMs with CGM data for insulin dosing; respondents reported taking fewer blood sugar checks per day and giving a greater number of insulin boluses without a meter blood sugar on Nightscout. This reinforces our impression that many patients use CGM for insulin dosing regularly, a fact that has become increasingly pertinent given the upcoming FDA Advisory Committee meeting to vote on a replacement label claim for Dexcom’s G5 CGM.

  • Dr. Lee also discussed the “maker movement” mentality, stressing its role in unleashing a cascade of creativity and innovation. For Nightscout, this has resulted in a plethora of devices, codes, and functionality, providing opportunities for customizing the glucose monitoring experience. [The same is true of the DIY OpenAPS community, which has a lot of overlap with the Nightscout developers.] Dr. Lee noted that patients are also “tinkering” with the system’s hardware, creating solutions such as xDrip, a tiny alternate CGM receiver that can be used to receive data from the Dexcom G4 transmitter and send it to other devices.
  • According to Dr. Lee, the current issue is whether or not Nightscout will disappear now that the FDA has approved commercial solutions that display CGM data on personal devices (Dexcom G4/G5, MiniMed Connect). She believes that Nightscout’s popularity will persist, given the high rate of new members entering the community. We believe the influx speaks to the community that Nightscout has created; the wider device compatibility and more “glanceable” and customizable displays on smartwatches; and to the overall message that Nightscout brings to the diabetes community: an empowering solution to help patients and caregivers monitor glucose with increased safety and more peace of mind.
  • Dr. Lee’s presentation echoed a valuable JAMA viewpoint she recently coauthored, which details Nightscout’s patient-driven founding, the system’s rapid growth, and its impact on changing the current definitions of health production and patient engagement.

Symposium: Technology and Diabetes Care – Your Patients Are Moving Forward. Are You Observing or Moving?

Make the Data Work for You―Increasing Accessibility, Integration, and Usability of Ambulatory Glucose Profiles

Deborah Mullen, PhD (Park Nicollet Health Services, Minneapolis, MN)

Dr. Deborah Mullen opened her presentation with a simple message: “Diabetes is complicated. Your data shouldn’t be.” With that context, she took attendees through an overview of the International Diabetes Center’s Ambulatory Glucose Profile (AGP) with a specific focus on how the one-page glucose data download report can facilitate personalized medicine. She stressed that one of the biggest gating factors to improved diabetes management is lack of feedback, arguing that AGP software both helps patients understand their own data at home and helps providers provide more individualized feedback in clinic. We’ve heard lots of enthusiasm for AGP over the past year and Dr. Mullen’s commentary echoed much of what we’ve heard before – however, she did remark on new news that the IDC has signed two agreements in the past week to license its AGP to two diabetes device makers (Roche and Abbott) and two diabetes data management companies (Diasend and Glooko), calling the partnerships a huge step in the right direction. We would agree! (See our coverage of the announcement here for more.) Ultimately, she pushed for even more consensus on this front, emphasizing perfect should not be the enemy of the good – an agreed-upon one-pager will go a long way for clinicians, and she urged industry not to be dissuaded by “not-invented-here” syndrome (plus, as she noted, it would not be too difficult for industry to provide proprietary reports on top of the standardized one-pager).

Questions and Answers

Q: Can you provide a comparison of your product vs. Glooko?

A: Glooko is a licensed AGP partner now. Their newest mobile app actually has the AGP graph on it. We are doing work together. We have partnerships, too, with Roche and Abbott and diasend who all have plans to come to market with AGP soon. That too, we have other partners coming.

Symposium: Inpatient Management of Diabetes and Hyperglycemia – Novel Insights and Effective Approaches

Glucometrics for Measuring Inpatient Glycemic Outcomes – Definition, Implementation, and Benchmarks

Curtiss Cook, MD (Mayo Clinic College of Medicine, Scottsdale, AZ)

Dr. Curtiss Cook elaborated on the role of glucometrics (i.e., metrics measuring the success of hospital glucose management) in improving inpatient outcomes. He noted that glucometrics consist largely of three measures – (i) glycemic exposure; (ii) efficacy of control; and (iii) rates of adverse events – noting that hospitals that adopt proven glucometrics programs gain a much better understanding of their weaknesses and strengths. However, despite the emergence of glucometrics as a subject of high interest, Dr. Cook noted that there remains a lack of standardization across the US in how glucose parameters are measured and reported. This issue, in his view, must be addressed before national benchmarking processes can be created that will allow institutions to compare inpatient glucose control performance against established guidelines. In this vein, he emphasized a need for standardization at all levels in order to maximize the utility of glucometrics programs, noting that standards must include consensus on which measures to report, the unit of analysis, definitions of targets for hyperglycemia treatment, a definition of hypoglycemia, etc. Dr. Cook concluded that hospitals need to be included in these discussions and that further dialogue on this topic is sorely needed.

  • According to Dr. Cook, the most important next steps for glucometrics programs involves standardization at all levels, from definitions to data reports. A 2008 survey of 269 hospitals conducted by Cook et al. showed target blood glucose ranges for ICU vs. non-ICU patients varied widely, such that there was no comparable standard from hospital to hospital. In addition to standardizing metrics, Dr. Cook stressed the importance of standardizing how the data is collected, calculated, and reported.
  • We assume greater hospital use of CGM and even automated insulin delivery will change this area significantly over the next five years, hopefully moving the field toward standardized time-in-range, hypoglycemia, hyperglycemia, and variability metrics. Could the upcoming artificial pancreas metrics paper in Diabetes Care help inform this discussion (more on that here)?

Questions and Answers

Q: Many hospital studies show much higher rates of hypoglycemia in inpatients. You have not seen that here. How do you account for this underrepresentation of hypoglycemia?

A: This study was just looking at patients with known, diagnosed diabetes as opposed to the overall population of inpatients.

Q: Have you looked at conducting studies using CGMs? The data could show a much granular picture with regard to highs, lows, and variability than we see here.

A: We’ve recently had a consensus conference where we discussed the potential of using CGMs. If we do implement them, the number of highs and lows and the amount of glucose variability we document could actually be a lot higher.

Joint ADA/JDRF Symposium: Optimizing Use of Technology and Therapeutics in Pediatric Diabetes

Status of Insulin Pump and Continuous Glucose Monitoring Use in Pediatric Diabetes

Jenise Wong, MD, PhD (UCSF, San Francisco, CA)

Dr. Jenise Wong provided a comprehensive overview of device utilization in pediatric type 1 diabetes. Beginning with pumps, she noted that there are a host of studies documenting the glycemic and quality-of-life benefits in children and adolescent but noted that use remains very variable by country. Dr. Wong attributed the variability to a number of factors from reimbursement and guidelines around pump use to the availability of pump reps and supplies – she cited T1D Exchange data suggesting that pump penetration is as high as 60% in the US, more than three or four times the penetration in EU countries. She contrasted that level of penetration with CGM, noting that the penetration of this technology remains VERY sparse (~15% in the Exchange). She was very positive on the benefits of sensor technology, though noted that a significant discontinuation rate prevents serious uptake. We do think that part of this has to do with patients’ sky-high expectations for CGM technology and the reality that the devices – while improving in accuracy and usability – are still not perfect, particularly on day one. She noted that even among CGM users, few patients take full advantage of the technology since such a minority of patients routinely review device data (see below) – we think this is because retrospective device data does not offer nearly enough benefits relative to the hassle of obtaining it and making sense of it. Once powerful, automated pattern recognition and compelling insulin titration algorithms are in place, we think device downloading will improve, or perhaps move to a real-time notification model. Ultimately, she suggested that education on the importance of device data (both on a provider and patient level) is a key step toward greater penetration in the future. We agree with her, though believe that CGM reimbursement, factory calibration, on-body form factor, a BGM replacement claim, and clinical decision support are more important for expanding uptake.

  • Dr. Wong presented valuable data on the lack of device data utilization from the T1D Exchange. Unsurprisingly, utilization is lowest among young adults with higher use among children and adolescents (where we assume parents are very involved with care). This does not surprise us one bit – CGM download software does not yet provide enough benefits to make the hassle of obtaining it and interpreting it worth it. Further, we believe the retrospective data mindset must change to a real-time model that gives patients more in-the-moment pattern recognition. What is more useful – a “morning high” pattern alert that is triggered every few months someone downloads, or a real-time notification at 10 am (“High pattern after breakfast observed for the past six days”)? In our view, personal CGM is a real-time technology, and patients stand to benefit the most from real-time, in-the-moment data analysis.

CGM Download Frequency

Never

1-3x per month

1x per week

Children

24%

37%

8%

Adolescents

36%

22%

12%

Young Adults

45%

16%

4%

Adults

42%

17%

5%

  • Dr. Wong also shared broader data comparing data downloading frequency (BGMs + CGMs) among adult patients and caregivers. We assume these numbers are self-reported (meaning they are probably overestimates), and they still fall woefully short of where we would like them to be. This will unquestionably improve as devices become connected to the cloud and stream data automatically, such as Dexcom’s G5, Abbott’s LibreLink, Medtronic’s MiniMed Connect, and a host of cellular- and Bluetooth-enabled meters. However, it still begs the philosophical question – is the future of diabetes data in real-time decision support and automatic pattern recognition? Is it realistic to ever expect patients to download and review their historical data?

Device Download Frequency

Never

Sometimes

Routinely

Adults

69%

20%

12%

Caregivers

44%

40%

27%

Questions and Answers

Q: What is surprising to me is the lack of penetration of pumps in pediatric patients considering the low discontinuation rate. That must mean we’re not getting people onto pumps in the first place?

A: Yes, that’s the first barrier – getting people onto pumps. We talked about the 60% on a pump; what we didn’t talk about is what is keeping those other 40% from starting. There are social factors to be sure - racial disparities, socioeconomic disparities. There are very real ways we could change the system and address cultural barriers that would help. So you’re right that discontinuation isn’t as much of a factor in pump use. The factor is getting people to start.

Q: Is discomfort a limiting factor with CGM? I’ve heard with Libre in the EU that skin reactions stops some kids from using the device.

A: Skin reactions are a common reason for discontinuation for all CGMs. I don’t know if the solution is better troubleshooting from an educator and clinician standpoint or a device innovation that needs to happen. However, it is a common cause of discontinuation and people who are using it will still complain.

Industry Updates

Victory For Data Standardization: Roche, Abbott, Diasend, Glooko Say Yes To AGP

In a victory for data standardization, we learned during ADA that the International Diabetes Center signed two agreements to license its one-page, standardized Ambulatory Glucose Profile (AGP) to two diabetes device makers (Roche and Abbott) and two diabetes data management companies (Diasend and Glooko). This was terrific news to hear – we’re eager to see greater standardization of data and use of that data to drive therapeutic change. The partnerships give the companies the right to use the AGP in all their diabetes devices and existing software; the agreement with Abbott extends the groups’ existing partnership to other devices since Abbott already uses the AGP report to visualize downloaded FreeStyle Libre glucose data. Clinicians have for a long time told us how much they like this standardized report and it’s fantastic to see that companies are beginning to sign on, especially after hearing for years the laundry list of reasons why industry was hesitant: (i) a desire to maintain control of data due to liability concerns; (ii) a desire to protect against competition by building their own proprietary software; (ii) a desire to preserve their direct relationship with patients; and (iv) the antiquated view that patients do not need access to their data (though we’d note that this is almost gone now.) Indeed, on all these levels, the willingness of Abbott, Roche, Diasend, and Glooko to license the report is a real win for patients and providers alike – after all, not every company is going to get what they want with a standardized report, but the field will benefit significantly from consensus. We’re hopeful this creates momentum and these partnerships set the stage for this standardization movement to reach a critical mass – will Dexcom and Medtronic sign on? Along these lines, we learned that AGP partnerships with three additional device companies and aggregators are slated for “the next month,” and we’ll be watching closely for updates on the IDC team’s recently launched website: AGPreport.org.

Corporate Symposium: Two Years of Flash Glucose Monitoring: The Global Clinical Experience (Supported by Abbott)

FreeStyle Libre T1DM Clinical Outcomes Trial Results (IMPACT)

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

Dr. Raimund Weitgasser shared full results from the FreeStyle Libre IMPACT study, comparing Abbott’s FreeStyle Libre to SMBG in type 1 patients in very good control (baseline A1c: 6.7%). The study met its primary endpoint at six months – relative to the control group, patients using FreeStyle Libre spent ~74 minutes fewer per day <70 mg/dl (a 38% reduction; p<0.001). These results were presented in detail in poster presentation 868-P (see detailed coverage above), though Dr. Weitgasser did add in a few additional details regarding: (i) the pattern of hypoglycemia behavior change; (ii) sensor use; and (iii) glycemic variability.

  • Dr. Weitgasser noted that time spent in hypoglycemia (both <70 mg/dl and <55 mg/dl) decreased as soon as sensor results became visible to patients and remained reduced throughout the study. Though it was only documented qualitatively, we were impressed to see how dramatic, immediate, and sustainable these changes were.
  • Patients collected 93% (6.5 days/week) of all possible data assuming they wore the sensor continuously for six months. Clearly, patients were scanning often enough to capture near-24/7 data, even though the FreeStyle Libre sensor can only store eight hours of data at a time. Capturing 6.5 days of sensor data per week means patients were missing about 12 hours of glucose data per week, or less than two hours per day. Presumably most of that lost data was overnight, when some probably didn’t scan twice within an eight-hour period.
  • Glycemic variability improved significantly in the intervention arm vs. control arm according to a variety of different metrics: SD, MAGE, CONGA, and low blood glucose index (LBGI). The actual values were not reported on the slide, though the improvement with FreeStyle Libre was in the ~10%-30% range. MODD and high blood glucose index (HBGI) were the only variability metrics that did not reach statistically significance.

Clinical Utilization of the FreeStyle Libre Flash Glucose Monitoring System in Europe

Ramiro Antuña DeAlaiz, MD (Asturias Medical Center, Oviedo, Spain)

“In my two years with FreeStyle Libre, it has been amazing how it has changed people’s lives!” This kind of striking enthusiasm defined Dr. Ramiro Antuña DeAlaiz’s presentation on his experience with Libre in the EU. The bulk of his talk shared data from a remarkable Spanish survey in which patients with both type 1 and type 2 diabetes reported complete elimination of nighttime hypoglycemia events within two weeks of starting Libre. The main benefit, Dr. Antuña DeAlaiz stressed, is that FreeStyle Libre empowers patients, giving them the impetus to interpret and respond to their own data. In addition, Dr. Antuña echoed positive sentiments we heard from other speakers on Abbott’s Ambulatory Glucose Profile that, in his words, “provides insight into aspects of glucose that are invisible with A1c.”

  • On who should be using FreeStyle Libre, Dr. Antuña DeAlaiz shared his belief that patients with prediabetes present an important target population for future expansion – “There will come a day when payers convince themselves that it is cheaper to prevent diabetes than to treat diabetes.” As this paradigm shift takes place, Dr. Antuña DeAlaiz stressed that the value of technology earlier in disease progression will become more widely appreciated.

Review of FreeStyle Libre Pro Glycemic Variability Study Results

Eugene Wright, Jr., MD (Cape Fear Valley Health Affairs, Fayetteville, NC)

Dr. Eugene Wright presented results from a study of FreeStyle Libre Pro demonstrating that the system is effective for capturing glycemic variability profiles in type 2 diabetes. The multicenter, prospective, single arm trial recruited 114 patients to wear two Libre Pro (blinded) sensors for two weeks while going about daily activities and adhering to their previously-established diabetes treatment plan. Inclusion criteria were quite broad, with patients recruited across eight therapy groups (sulfonylureas, basal insulin, premix insulin, lifestyle alone, metformin, GLP-1s agonists, DPP-4s inhibitors, SGLT-2s inhibitors) and with A1cs ranging from 6.0%-12.0%. Data was consistent with our expectations – greater variability was associated with the use of insulins and sulfonylureas, while less variability was observed in patients whose diabetes had not yet progressed as far (metformin, lifestyle alone). Unsurprisingly, data also indicated a greater prevalence of hypoglycemia (time spent <70 mg/dl) in patients using insulins and sulfonylureas. Individual group sample sizes were too small to read into further differences (and mean glucoses were not matched either). There were no device-related serious adverse events and nine instances of minimal adverse events (e.g., infection, allergy). Overall, Dr. Wright concluded that FreeStyle Libre Pro is both effective and safe in capturing glucose profiles in type 2 diabetes; we wonder how it can be used for therapy titration, and perhaps as a companion diagnostic to aid therapy selection in type 2 diabetes (e.g., “Oh, you have a postprandial problem; you need a GLP-1 or SGLT-2”).

  • Below, we have summarized the glycemic variability data collected with FreeStyle Libre Pro in this study. A few things jump out in the data:
    • Hypoglycemia is prevalent in all groups, but alarmingly high in the A1c 6.0-7.4% group (2+ hours per day);
    • SGLT-2s and GLP-1s seem to keep patients more in range relative to other therapies in the 7.5-12.0% group;
    • For patients in the 7.5%-12.0% A1c group, sulfonylureas are highly ineffective, keeping patients in range just 10 hours per day with an average blood glucose of 192 mg/dl.

A1c

Therapy

N

Mean Glucose (mg/dl)

CV (%)

Time (h/day) 70-180 mg/dl

Time (h/day) < 70 mg/dl

6.0-7.4%

Sulfonylurea

15

118

29.8

20.1

2.2

Basal insulin

10

118

33.4

19.7

2.2

Premix insulin

8

138

35.8

16.2

2.4

Metformin

15

144

22.6

17.1

0.8

7.5-12.0%

Sulfonylurea

12

192

26.7

10.2

0.5

GLP-1 agonist

9

159

23.8

14.8

0.8

DPP-4 inhibitor

6

201

21.9

11.7

0.1

SGLT-2 inhibitor

4

157

27.3

17.5

0.2

Basal insulin

12

182

32.8

12.2

0.7

Premix insulin

10

170

33.9

13.5

1.4

 
  • Dr. Wright shared positive data from user experience studies of FreeStyle Libre Pro in the study:
    • 96% agreed that the sensor was comfortable to wear.
    • 94% agreed that the sensor was easy to wear due to its small size.
    • 77% agreed that other people did not notice that they were wearing a sensor.
    • 88% agreed that they did not feel any discomfort under their skin while wearing the sensor.
    • 91% agreed that the sensor did not get in the way of daily activities.
    • 90% agreed that the sensor fit in well with their lives.

Clinical Utilization of the FreeStyle Libre Pro Flash Glucose Monitoring System in India

KM Prasanna Kumar, MD (Bangalore Diabetes Hospital, Bengaluru/India)

Dr. KM Prasanna Kumar offered glowing praise for FreeStyle Libre Pro as an aid for patients and HCPs in India. He spoke enthusiastically about the pattern recognition in Abbott’s Ambulatory Glucose Profile, the value in optimizing patients’ therapy by helping providers feel more successful, and – most importantly – making the invisibility of diabetes more tangible to patients. Especially for patients in India who cannot afford CGM technology, he impressed upon the audience that the insight provides a valuable retrospective look to identify trends and can even inform smarter allocation of limited fingersticks. Though Dr. Kumar acknowledged that the device holds value for anyone on insulin, he stressed that such an approach is not feasible in India. He explained that India is an “out-of-pocket” market, where 89% of the population has no insurance coverage – as a result, he shared that the average diabetes patient spends only ~$120 (!) on diabetes medication and monitoring annually. [According to Dr. Kumar, this expenditure buys just one SMBG per week, one venous blood test per three months, and one A1c test per six months.] In such a market, he noted that it would be irrational for payers to purchase Libre Pro frequently, regardless of the magnitude of the benefit it confers. As such, Dr. Kumar typically uses the Pro in his patients once, for 5-6 days every three months, to calibrate and assess the prescribed treatment - a great workaround to a tough situation, and far better than one test strip per week and an occasional A1c. As he concluded: “In this way, we keep treatment cheap. The AGP assesses the extent and cause of glycemic variance as well as the compliance of the patient in diet, medication, exercise, etc…This is an excellent education tool – every patient can understand how his HCP is trying to achieve euglycemia.” 

Panel Discussion

Ramiro Antuña DeAlaiz, MD (Asturias Medical Center, Oviedo, Spain), KM Prasanna Kumar, MD (Bangalore Diabetes Hospital, Bengaluru/India), Raimund Weitgasser, MD (Paracelsus Medical University, Salzburg, Austria), Eugene Wright, Jr., MD (Cape Fear Valley Health Affairs, Fayetteville, NC)

Q: How was glucose variability measured in IMPACT?

Dr. Weitgasser: We used a lot of different parameters. That was depicted in one of the slides I showed.

Q: How can you reconcile concerns that the lack of hypoglycemia alarm is a danger?

Dr. Antuña DeAlaiz: Many people with Libre scan before going to bed, which allows them to prepare for the nighttime and take insulin or food as necessary. In my experience, some people need an alarm because they are completely hypoglycemia unaware, but others are very happy without alarms. One size doesn’t fit all.

Dr. Weitgasser: This is not a system designed for people with hypoglycemia unawareness. We excluded these people in our study.

Dr. Wright: In addition, as an educational tool, Libre gives you the opportunity to frame conversation with patients about choices with respect to lifestyle and medications. Those choices have an impact on the glycemic profile.

Q: Given the very early changes in hypoglycemia in patients using Libre, do you believe this is purely a behavioral effect or were HCPs involved in some specific counseling?

Dr. Antuña DeAlaiz: I think patients were interested in education. They wanted to maximize the benefit they were getting from system and, as a result, the patient wanted to learn how to manage and respond to the arrows.

Dr. Kumar: Remember that patients are able to send their AGP to HCPs. When you look at the whole profile, you can adjust habits and insulin quickly. You get this comprehensive update, so can fine-tune diet, exercise, and medications.

Dr. Antuña DeAlaiz: This is a great education tool – an eye-opener. People’s lives change completely.

Dr. Gavin: There is a lot more information to talk about with patients when you use Libre and the AGP, plus of course, the contextual comments from patients.

Q: Can you talk about the science of this sensor? What’s different from the Dexcom and Enlite sensors? There is no calibration required for Libre.

Mr. Steve Scott (Abbott Diabetes Care, Alameda, CA): The sensor is very stable over time and we get very little sensor-to-sensor variation. We have key processes to ensure this minimized variance. This allows for the 14-day wearing period. We have in-vitro calibration and apply this in vivo.

Joint ADA/JDRF Symposium – Optimizing Use of Technology and Therapeutics in Pediatric Diabetes

Information Overload―Or Is It?

Saleh Adi, MD (Madison Clinic for Pediatric Diabetes, San Francisco, CA/USA)

In a persuasive, case-based talk, Dr. Saleh Adi championed the use of CGM over intermittent fingersticks and potentially misleading A1c values. He highlighted the benefits of CGM in MDI users, called for better insurance coverage of data interpretation, asked for industry to develop decision-support software, and requested EMR integration to take some of the burden off of the HCPs. Dr. Adi noted the educational, eye-opening value of CGM: “Look at this chart,” he said to one repeatedly skeptical mom, “you are over-treating your child’s lows and highs.” Dr. Adi urged everyone to recognize that A1c does not provide enough information for clinical decision making, and that every patient should have access to CGM because everyone can benefit. He concluded by calling for a DCCT 2.0 that could demonstrate the benefits of maximizing time in range and reducing glycemic variability on short- and long-term outcomes. He even envisioned new standards of care one day in which A1c is a secondary measure and treatment is individualized for each patient. Wouldn’t that be something?!

  • Dr. Adi shared two case studies illustrating how CGM completely alters understanding of glucose profiles, particularly in patients with near-goal A1c’s. Both patients’ multiple daily fingersticks typically yielded readings in the 100-200 mg/dl range and A1cs of 7.3% and 7.9% - in other words, fairly well-managed glucose levels in pediatrics. However, upon using CGM for a period, they (and Dr. Adi) learned that one patient was experiencing severe post-prandial hyperglycemia, and the other was experiencing mid-sleep excursions <50 mg/dl and post-dinner hyperglycemic excursions >400 mg/dl. “We can all agree that this is not good enough,” said Dr. Adi, despite what their A1c values and intermittent fingersticks say.
  • Dr. Adi highlighted one of his MDI patients that has still benefited greatly from CGM. There is a false misconception, he said, that CGM is only for patients who use pumps. This is clearly not the case, as we saw at this meeting with Dexcom’s DIaMonD study: A1c declined a strong 0.9% with CGM at six months vs. -0.4% with usual care (baseline: 8.6%), for an adjusted mean difference of -0.6% in favor of CGM (p<0.001). At the same time A1c declined, hypoglycemia significantly improved with CGM: a 30% improvement in time <70 mg/dl (-23 mins/day; p=0.006) and a strong 50% improvement in time <50 mg/dl (-11 mins per day; p=0.005), both outperforming 17% and 21% improvements with usual care (-15 mins, -6 mins).
  • Dr. Adi said HCPs should focus on one thing at a time with CGM downloads; otherwise, “information overload” is too easy! “Find the most obvious thing and solve it.” This approach will not only avoid overwhelming patients with too much information, but also empower them by making them realize that they are capable of managing their condition. In addition, Dr. Adi advocated for insurance coverage of data interpretation counseling, the development of decision-support software, and EMR integration to take some of the burden off of the HCPs.
  • CGM marks the end of the “I’ll believe it when I see it” excuse. Dr. Adi reports that with just a few fingersticks per day, patients and caregivers deny that they are mistreating their lows and highs. CGM worn 24/7 can document what is actually happening, driving a much feedback loop than fingersticks alone.

Questions and Answers

Q: I don’t disagree with anything you said. I would reformulate the question: What you don’t know is that DCCT/EDIC have analyzed data on DCCT subjects who maintained an A1c of 8.8% over 32 years of follow-up. They contrasted with the intensive group who maintained a 7.2% A1c over 32 years. There was a huge difference in complications rate between groups. Getting A1c down to 7.2% or lower is enough. However, the purpose of treatments as we go forward is to do it safer (less hypo), smoother (less glycemic variability), and more automatic.

A: I agree with everything you say.

Q: I would be thrilled if my adolescents took insulin. I have patients who SMBG test 2-3 times a day and look at CGM 2-3 times a day… there is no difference. DCCT had selection bias – these technologies are awesome, including closed loop. I suspect my patients would wear and finally we’d see some better control. But not all patients should be on CGM if they won’t use it.

A: Nothing will solve 100% of the problems of all people. But they should all be given the option to wear CGM.

-- by Adam Brown, Samiul Haque, Varun Iyengar, Brian Levine, Ava Runge, and Kelly Close