American Diabetes Association 75th Scientific Sessions

June 5-9, 2015; Boston, MA; Full Report – SGLT-2 Inhibitors – Draft

Executive Highlights

As SGLT-2 inhibitors have now taken their place as a more established drug class in type 2 diabetes, much of the new data at this year’s ADA focused on novel indications or combinations. Given that phase 3 results for SGLT-2 inhibitor/DPP-4 inhibitor fixed-dose combinations (FDCs) were a major highlight of last year’s ADA, it was exciting to see one such combination (Lilly/BI’s Glyxambi [empagliflozin/linagliptin]) featured in the exhibit hall at this year’s meeting – Lilly disclosed that the combination has been a driver of Jardiance (empagliflozin) franchise growth of late. We also saw positive data from two AZ-sponsored studies of triple SGLT-2 inhibitor/DPP-4 inhibitor/metformin combination therapy, which could potentially support a triple oral FDC in the future. This year’s agenda also featured several encouraging studies on the use of SGLT-2 inhibitors in type 1 diabetes, which we expect to remain a hot topic in the next few years as larger randomized controlled trials begin to report results.

Of course, much of the recent discussion around SGLT-2 inhibitors has focused on concerns about euglycemic DKA, and ADA was no exception. Speakers generally framed the phenomenon as more of a manageable risk than a deal-breaker for the SGLT-2 inhibitor class though not everyone – at the TCOYD/The diaTribe Foundation event, Dr. James Gavin (Emory University, Atlanta, GA) urged a greater degree of caution. Dr. Julio Rosenstock (University of Texas Southwestern Medical Center, Dallas, TX), Dr. John Buse (University of North Carolina, Durham, NC), and Dr. Anne Peters (USC, Los Angeles, CA) all emphasized the importance of continuing to increase awareness of the risk, which was largely unknown until Dr. Peters first began drawing attention to it toward the beginning of this year. Much of the concern at this point is focused on off-label use of SGLT-2 inhibitors in type 1 diabetes, as the risk/benefit profile is not as clear in that population as in type 2 diabetes (where the phenomenon appears to be more rare and influenced by clearer risk factors like surgery or insulin deficiency). Dr. Rosenstock predicted that the ongoing trials of SGLT-2 inhibitors in type 1 diabetes will likely conclude that the benefits (insulin-independent efficacy, reduced glucose variability and insulin doses, and improvements in weight and blood pressure) outweigh the risks for most patients. Dr. Buse similarly suggested that the phenomenon is unlikely to be a “showstopper” for SGLT-2 inhibitors as long as people are aware of it and know how to treat it. Dr. Peters took a somewhat more cautious tone, but she suggested that the class can still be used safely in some cases as long as appropriate precautions (i.e. low doses, frequent ketone monitoring, and good patient-provider communication) are taken.

The discussion of euglycemic DKA has continued to evolve since ADA. Dr. Peters, Dr. Buse, and Dr. Irl Hirsch (University of Washington, Seattle, WA) are co-authors on a case series that was recently published in Diabetes Care, marking the first major publication we’ve seen on the issue. It should certainly continue to increase awareness in the medical community and we applaud the trio for staying on top of the developing issue. The FDA and EMA are both currently investigating the phenomenon, and we imagine that some sort of label additions are likely in the future. The impact on the class’ prospects in type 1 diabetes will depend on the results from ongoing trials, which should provide much more clarity on the broader risk/benefit profile.

Talk titles highlighted in yellow were among our favorites from ADA 2015; those highlighted in blue are new or expanded full report additions from our daily coverage.

Table of Contents 

Detailed Discussion and Commentary

Oral Presentations: Experience with SGLT2 Inhibitors

A Randomized, Double-Blind, Phase 3 Trial of Dapagliflozin Add-On to Saxagliptin + Metformin in Type 2 Diabetes (105-OR)

Chantal Mathieu, PhD (University Hospital of Leuven, Leuven, Belgium)

Dr. Chantal Mathieu reported on results from a trial (n=320; Identifier: NCT01646320) comparing the safety and efficacy of dapagliflozin (AZ’s Farxiga) vs. placebo as an add-on to saxagliptin (AZ’s Onglyza) and metformin in patients with type 2 diabetes. Participants received open-label saxagliptin plus metformin for a period of either eight or 16 weeks, depending on trial stratum. Those who had inadequate glycemic control (A1c of 7%-10.5%) after this open-label period were then randomized to placebo or dapagliflozin plus open-label saxagliptin and metformin. The primary endpoint was change from baseline A1c at week 24, with secondary endpoints including fasting plasma glucose, 2-hour postprandial glucose, body weight, and proportion of participants achieving A1c <7%. Overall, the combination of dapagliflozin, saxagliptin, and metformin led to greater A1c reductions (-0.82% vs. -0.10%), greater improvements in the secondary endpoints, and more participants achieving an A1c <7% than saxagliptin and metformin with placebo. Adverse effects were generally similar in the two groups with a low overall risk of hypoglycemia, though more participants developed genital infections in the dapagliflozin group compared to placebo (5% vs. 0.6%). Dr. Mathieu concluded that the addition of dapagliflozin to saxagliptin and metformin improved glycemic control and was well tolerated in patients with type 2 diabetes inadequately controlled on saxagliptin and metformin.

  • AZ highlighted this trial during an Investor Day last November as part of an overall push toward early, aggressive combination therapy for type 2 diabetes. Management suggested during that presentation that results from this trial and another evaluating saxagliptin vs. placebo as an add-on to dapagliflozin and metformin could potentially support a triple fixed-dose combination, though such a product would not be submitted until 2017 at the earliest.

Questions and Answers

Q: Regarding data on adverse events, what about heart failure? The dapagliflozin arm has two heart failure events. What do you read from that?

A: We looked carefully at these patients. The first patient had heart failure very early on and had many other problems in addition to that. I don’t really remember more details about the other one. But if you’re asking if dapagliflozin can protect against heart failure, the numbers are too small to say anything.

Q: Should we be thinking about the sequence of medications?

A: I don’t know. But I think studies like this one can approach that question. It’s interesting to see what would happen if we switched the order in the same study. We’ll try to compare but there are different populations so that always makes it more difficult.

Q: Did you analyze the two strata differently? Were there any outcomes that were different?

A: They were not different.

Effect of Longer-Term Canagliflozin Treatment on eGFR in Patients with Type 2 Diabetes Mellitus and Various Degrees of Baseline Renal Function (107-OR)

Matthew Weir, MD (University of Maryland Medical Center, Baltimore, MD)

Dr. Matthew Weir presented results of three interrelated, long-term studies examining the effect of canagliflozin alone or compared with glimepiride and placebo on estimated glomerular filtration rate (eGFR) in people with type 2 diabetes. In the first two-year study, patients between 18 and 80 years old with a mean eGFR of 90 ml/min/1.73 m2 received canagliflozin at 100 mg or 300 mg, or glimepiride. The second two-year study studied patients aged 55 to 80 years old with mean eGFR of 77.5 ml/min/1.73 m2 who received canagliflozin at 100 mg or 300 mg, or placebo. The third study was one year in duration and featured patients older than 25 with stage 3 chronic kidney disease (eGFR of 39 ml/min/1.73 m2), who received canagliflozin at 100 mg or 300 mg, or placebo. Findings showed that participants’ eGFR initially decreased with canagliflozin but then stabilized or moved back toward the baseline within the first three to six weeks, remaining more or less unchanged over the next two years. eGFR progressively declined with glimepiride and did not change at all with placebo. Thus, participants on canagliflozin saw transient eGFR changes that attenuated over time and all study groups saw relatively consistent absolute effects, regardless of baseline renal function. The urine albumin to creatinine ratio (UACR) decreased with both the 100 mg and 300 mg canagliflozin doses, but it increased with glimepiride and placebo. Therefore, Dr. Weir concluded that eGFR changes with canagliflozin were overall not permanent and generally fairly weak. Of course, the most definitive answers regarding the effect of canagliflozin on renal function will come from the ongoing CREDENCE renal protection study, scheduled to complete in 2020.

Questions and Answers

Q: We sometimes see patients initiated on SGLT-2 inhibitors and then we see a drop in eGFR. Is there anything we should do?

A: I would do the same thing as I do with an ACE inhibitor or ARB. Check that the patient isn’t taking other anti-inflammatory or diuretic drugs.

Q: Did you have uric acid levels measured in your study?

A: We don’t have that. We know there aren’t significant changes.

Q: Are there any concerns with these drugs being used in patients on diuretics?

A: No, a lot of our patients were on diuretics.

Q: Thoughts on DKA?

A: The majority of reports of DKA occurred in type 1 diabetes. Very few are reported in people with type 2 diabetes. We don’t have any data to indicate the cause of ketoacidosis.

Oral Presentations: New Insights Into the Effects of Oral Agents

No Increased Risk of Cardiovascular Events with Dapagliflozin in Elderly Patients With Type 2 Diabetes Mellitus, Cardiovascular Disease, and Hypertension (15-OR)

Ingrid Gause-Nilsson, MD, PhD (AstraZeneca, Mölndal, Sweden)

AstraZeneca scientist Dr. Ingrid Gause-Nilsson presented the results of a post-hoc analysis of cardiovascular events in 1,263 elderly, high-cardiovascular-risk patients enrolled in 19 phase 2b/3 clinical trials (duration up to four years) for AZ’s Farxiga (dapagliflozin). This was a follow-up analysis to one presented at AHA last year (abstract 16682) for the full trial population, not just elderly high-CV-risk patients – that analysis found a hazard ratio of 0.79 (95% CI: 0.58-1.07) for MACE + unstable angina with dapagliflozin. The analysis presented at ADA was of interest because the elderly high-CV-risk subpopulation is arguably closer to the population that would be enrolled in a CV outcomes trial. Fortunately, there was no cause for worry in this analysis: the hazard ratio for MACE was 0.92 (95% CI: 0.51-1.64) for MACE and 0.82 (95% CI: 0.50-1.37) for MACE + unstable angina, with no imbalance in CV death, MI, or stroke. Analyses like these are somewhat reassuring from a safety standpoint but are nothing compared to the outcomes data we will have from the DECLARE-TIMI outcomes trial, though that is only expected to end in 2019.

  • Dr. Gause-Nilsson cited a very slight imbalance in renal impairment/failure (13.7% with dapagliflozin vs. 11.3% with comparator therapy), although eGFRs appeared fairly stable – we didn’t see this as very worrying given that most diagnoses were based on laboratory values. There was a 5.8% incidence of genital infections with dapagliflozin vs. 0.8% with comparator therapy, as well as a very modest imbalance in urinary tract infections that emerged in longer studies. It was hard to interpret the glucose lowering efficacy results (mean difference of 0.4%) because some studies had active comparators.

Questions and Answers

Q: From the data here, I’m not surprised that there was no increase in cardiovascular risk. I’m surprised that this drug does not decrease cardiovascular events due to the effects on blood pressure, weight, and A1c. Is it just a matter of power?

A: The possible positive effects of these risk factors may need a longer follow up beyond four years.

Q: Were all of these events adjudicated?

A: The cardiovascular endpoints were adjudicated, but adverse events of special interest were not.


Glycemic Effects of SGLT2 Inhibitor Canagliflozin in Type 1 Diabetes Using Dexcom G4 Platinum CGM (932-P)

N Argeno and K Nakamura

Results of a retrospective review of 27 patients with longstanding type 1 diabetes using Dexcom’s G4 Platinum CGM (PCGM) showed that addition of J&J’s Invokana (canagliflozin) 100 mg to insulin provided statistically significant glycemic benefits. The study examined patient records, including 30-day CGM downloads, at baseline and after at least one month on canagliflozin (mean exposure of 3.7 months). 27 patients with similar baseline characteristics who did not undergo any change in therapy other than insulin adjustment served as a control group. Results showed that the addition of canagliflozin reduced mean CGM glucose, reduced blood glucose variability, improved time spent in target blood glucose range, and reduced A1c, systolic blood pressure, weight, and total insulin dose from baseline, with no significant change in renal function (serum creatinine or eGFR). The control group did not experience significant changes in any parameters from baseline. Canagliflozin therapy was generally well tolerated; there were three episodes of DKA in the canagliflozin group, two of which were attributed to pump site failure and one to a patient running out of basal insulin. These encouraging results illustrate the many potential benefits of SGLT-2 inhibitors in type 1 diabetes, though it remains to be seen what effect the concerns about euglycemic DKA will have on the class’ overall risk/benefit profile for type 1 – ongoing randomized controlled trials will hopefully offer more clarity.

  • Baseline demographics for the canagliflozin vs. control group were well matched. The canagliflozin group (n=27) had a mean age of 52 years, duration of type 1 diabetes of at least 10 years (mean 34 years), and duration of CGM use >1 year (mean 4.8 years). The control group consisted of 27 patients (mean age 51) with “similar indications” who were judged to be eligible for canagliflozin therapy and did not undergo any treatment changes other than insulin adjustment.
  • The study compared two sets of patient records, including 30-day CGM downloads, for each group. In the canagliflozin group, investigators examined records at baseline and after at least one month on canagliflozin (mean exposure of 3.7 months). For the control group, investigators examined the earliest most complete sequential pair of downloads between May 2013 and January 2015.
  • Addition of canagliflozin led to significant improvements from baseline in a number of glycemic and non-glycemic parameters. Patients in the canagliflozin group experienced significant improvements in mean glucose (168 mg/dl to 147 mg/dl; p<0.0001), glucose variability (measured by standard deviation of CGM glucose; 68 mg/dl to 56 mg/dl; p<0.0001) and percentage of time in range (70-180 mg/dl) (57% to 70%; p<0.0001) from baseline. Canagliflozin also led to significant reductions in A1c (7.6% to 7.2%; p<0.0001), systolic blood pressure (122 mmHg to 114 mmHg; p=0.0089), total daily insulin dose (51 units to 45 units; p=0.0002), and body weight (89 kg to 87 kg; p=0.0007) from baseline. The poster did not directly compare final results in the canagliflozin vs. control groups, but the two groups were well matched at baseline and the control group did not experience significant changes in any parameters.
  • Canagliflozin appeared to be well tolerated, with no significant changes in renal function and no cases of DKA attributed to the drug. Serum creatinine and eGFR were comparable at baseline and after canagliflozin treatment. Three cases of DKA occurred in the canagliflozin group: one at baseline attributed to pump site failure and two mild episodes in the follow-up period, one attributed to pump site failure and one to the patient running out of basal insulin and attempting to treat with rapid-acting insulin alone. There were no episodes of DKA in the control group.

Empagliflozin (EMPA) Decreases Glucose Exposure and Variability in Patients with Type 1 Diabetes (T1DM): Continuous Glucose Monitoring (CGM) Data (EASE-1) (1241-P)

S Famulla, T Pieber, J Eilbracht, D Neubacher, N Soleymanlou, H Woerle, U Broedl, and S Kaspers

This randomized, double-blind, placebo-controlled phase 2 trial examined the effect of Lilly/BI’s SGLT-2 inhibitor Jardiance (empagliflozin), as an adjunct to insulin on glucose exposure (measured by area under the median glucose curve), interquartile range of glucose (IQR), and mean amplitude of glucose excursions (MAGE) in type 1 diabetes patients. For four weeks, 75 patients were randomized to placebo or one of three doses of empagliflozin (2.5 mg, 10 mg, or 25 mg), in addition to insulin (stable insulin dose during the first week and freely adjustable dose thereafter). There were significant improvements in glucose exposure and variability with all doses of empagliflozin vs. placebo following the one-week stable insulin period (p<0.05). However, all improvements were diminished at four weeks; significance was retained for glucose variability at all doses (p<0.001) and for glucose exposure only at the highest dose (p<0.05). Based on these results, it is somewhat challenging to tease apart the effect of the empagliflozin on glucose exposure and variability from the effect of stable insulin dosing in the first week of the trial. Time in range (>70 mg/dl to ≤180 mg/dl) was significantly improved with all doses of empagliflozin vs. placebo after one week (p<0.05) but only with the highest dose after four weeks (p<0.001).

  • All groups had similar baseline characteristics. Average age was ~39-42 years and all patients were white. Mean A1c was ~8.2%-8.3%, average body weight was ~75-87 kg (165-192 lbs), and average systolic blood pressure was ~122-124 mm Hg in all groups. 7-day ambulatory glucose profiles (AGP) were generated based on CGM data collected over 24-hour periods.
  • Glucose exposure and time in range (>70 mg/dl to ≤180 mg/dl) improved significantly from baseline with all doses of empagliflozin vs. placebo after one week, but the improvements remained significant (albeit diminished) after four weeks only at the highest (25 mg) dose. At one week, area under the median glucose curve decreased by 12.2, 30.2, and 32.9 mg/dl*hr more than the reduction observed in the placebo group (-0.9 mg/dl*hr) for the 2.5, 10, and 25 mg doses (p=0.04, p<0.001, p<0.001), respectively. However, at four weeks, a significant placebo-adjusted reduction was observed only with the 25 mg dose (-15.9 mg/dl*hr, p=0.03). Similarly, at one week, time in range increased by 2.6, 3.7, and 4.3 more hours per day with the 2.5, 10 and 25 mg doses of empagliflozin, respectively, vs. placebo (p<0.01, p<0.001, p<0.001), but the only significant result observed at four weeks was with the 25 mg dose (2.7 more hours per day than placebo, p<0.001).
  • Glucose variability (IQR and MAGE) decreased significantly with all doses of empagliflozin compared to placebo at weeks one and four, though the reduction from baseline was less dramatic at four weeks. After one week of stable insulin dosing, the IQR for the placebo group decreased from baseline by a staggering 21.6 mg/dl. The 2.5, 10, and 25 mg doses of empagliflozin produced additional reductions of 14.0, 20.0, and 21.1 mg/dl, respectively (p<0.001 for all). After four weeks, placebo IQR had increased from baseline by 6.5 mg/dl. For the 2.5, 10, and 25 mg empagliflozin groups, IQR decreased from baseline by 15.4, 15.0, and 20.7 mg/dl, respectively (p<0.001 vs. placebo for all groups). Because placebo IQR increased from baseline at four weeks, the placebo-adjusted reductions with empagliflozin were larger than at one week, but the absolute differences from baseline were smaller.

The Combination of Insulin, Liraglutide, and Dapagliflozin as Triple Therapy for Type 1 Diabetes (130-LB)

N Kuhadiya, A Mehta, H Ghanim, J Hejna, A Makdissi, A Chaudhuri, M Batra, and P Dandona

Continuing their group’s pioneering work on liraglutide in type 1 diabetes, Dr. Kuhadiya and colleagues conducted a small (n=10), 12-week retrospective study of dapagliflozin in people with type 1 diabetes who had already been using liraglutide in addition to insulin for a mean of 11 months. At baseline, study participants had mean A1c 8.01%, mean glucose 172 mg/dl, age 56 years, age at diabetes diagnosis 29 years, BMI 29 kg/m2, body weight 86 kg (189 lbs), daily carbohydrate intake 196 g, and blood pressure 125/75 mm Hg. Patients began dapagliflozin at 5 mg daily and, after 7±1 days, increased the dose to 10 mg daily. After 12±1 weeks of dapagliflozin therapy, patients had mean A1c reduction of 0.66%, body weight reduction of roughly 1.5 kg (3 lbs), mean glucose reduction of 28 mg/dl as monitored by continuous glucose monitoring, and daily carbohydrate increase 0f 30 g while insulin dose stayed constant (on average) at 0.7 U/kg. No additional hypoglycemia was observed, but one patient underwent diabetic ketoacidosis (DKA) despite normal glucose levels, within two days after up titrating their dapagliflozin to 10 mg daily. This patient had notably decreased their total daily dose of insulin from 32.9 to 28.5 U (0.45 to 0.39 U/kg) while also increasing their daily carbohydrate intake from 50 to 95 g. We look forward with great interest to randomized controlled trials of SGLT-2 inhibitors plus GLP-1 agonists in type 1 diabetes. As to the documented instances of euglycemic DKA associated with SGLT-2 inhibitors in type 1 diabetes, we hope that the risk can be minimized with education of patients and clinicians and with selective prescribing patterns – see our coverage of a recent Diabetes Care case series for more.

  • This was a 12-week, n=10 retrospective study of dapagliflozin added to baseline therapy of liraglutide and insulin in type 1 diabetes.
  • The study concluded that addition of dapagliflozin to insulin and liraglutide in patients with type 1 diabetes significantly improved glycemia. After 12 weeks of treatment, mean A1c fell by 0.66% (p=0.0004) and mean glucose decreased from 172 mg/dl to 144 mg/dl  (p=0.016), though  total insulin dose (0.7 u/kg) did not change. Time in range (70-160 mg/dl) increased from 45% to 56% (p<0.05) and time with glucose >160 mg/dl decreased from 48% to 37% (p<0.05). Daily carbohydrate intake increased from 166 g before the addition of dapagliflozin to 196 g after (p=0.04). Both body weight and BMI fell after dapagliflozin was added, from 87 kg to 85 kg (p=0.02) and 29 kg/m2 to 28 kg/m2, respectively.



12 wks

A1c (%)



Mean glucose (mg/dl)



Body weight (kg)



Body weight (lbs)



Body mass index (kg/m2)



Daily carbohydrate intake (g)



Daily insulin dose (U/kg)



                           * p<0.05

  • Dr. Kuhadiya and his team advised caution when reducing insulin doses to avoid euglycemic DKA. One patient in the retrospective analysis developed DKA despite normal blood glucose concentrations within 48 hours of increasing the dapagliflozin dose to 10 mg/day. This patient’s dose of insulin had declined from 0.45 to 0.39 u/kg and carbohydrate intake had increased from 50 to 95g/day. With regard to other adverse events, one female patient developed a vaginal yeast infection that was successfully treated.

Symposium: Non-Insulin Adjunct Therapies in Type 1 Diabetes

The Potential Use of SGLT-2 Inhibitors in Type 1 Diabetes

Julio Rosenstock, MD (University of Texas Southwestern Medical Center, Dallas, TX)

Dr. Julio Rosenstock provided a nuanced but generally positive perspective on the potential of SGLT-2 inhibitors in type 1 diabetes. The key benefits include insulin-independent efficacy, reduced glucose variability and insulin doses, and improvements in weight and blood pressure. He strongly encouraged the audience to attend the presentation of full results from a four-week study of Lexicon’s SGLT-1/SGLT-2 dual inhibitor sotagliflozin in type 1 diabetes – the results were promising across the board, showing reductions in A1c (0.5% placebo-adjusted), postprandial glucose, and bolus insulin doses, improvements in glycemic variability, and a ~2 kg weight benefit. With regard to the recent concerns about euglycemic DKA, Dr. Rosenstock suggested that it is important for patients and physicians to be aware of the phenomenon but he framed it more as a risk to be managed rather than a deal-breaker. He also noted that a substantial portion of the reported cases occurred after reductions in insulin doses, which can contribute to the vicious cycle initiated when the body switches to a fasting mode (increased glucagon production, lipid oxidation, etc.) in response to increased glucose excretion; other speakers throughout the meeting expressed similar sentiments regarding the role of low insulin levels. We heard a great deal of commentary on euglycemic DKA at AACE last month as well, most of it similarly balanced. On the whole, we agree with Dr. Rosenstock’s belief that the ongoing trials of SGLT-2 inhibitors in type 1 diabetes will likely conclude that the benefits outweigh the risks for most patients, though it will be very important for patients to have a ketone meter and know when to use it.

Symposium: SGLT2 Blockers in 2015 (Supported by BI/Lilly)

Similarities and Differences Between SGLT2 Blockers

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

Dr. Robert Henry reviewed the mechanism of action and clinical effects of SGLT-2 inhibitors, though he refrained from making claims about the relative benefits of specific agents given the lack of comparative data. He did discuss the potential added benefits of SGLT-1/SGLT-2 dual inhibitors like Lexicon’s sotagliflozin over pure SGLT-2 selective inhibitors due to increases in gut hormones like GLP-1 and PYY – he suggested that this may account for a small portion of the efficacy of J&J’s Invokana (canagliflozin), which has some slight SGLT-1 activity. He presented data demonstrating that all three currently approved SGLT-2 inhibitors offer A1c reductions of approximately 0.5%-1%, durable ~3% weight loss, blood pressure reductions, and positive changes in HDL cholesterol and triglycerides. On the negative side, Dr. Henry also discussed the increases in LDL cholesterol and hepatic glucose production seen with the class and the initial drop in estimated glomerular filtration rate (eGFR) that has prompted “lots of speculation” about harmful renal consequences, but he stressed that any impairment in kidney function appears to be transient. He briefly acknowledged the concerns about euglycemic DKA but did not offer further commentary – we were somewhat surprised that this did not come up during Q&A, as it has appeared to be top of mind for many providers at recent conferences. Dr. Henry closed with a discussion of SGLT-2 inhibitors’ broad potential in combination with other agents, particularly incretins, and of the need for long-term studies to elucidate their effects on bone health and renal and cardiovascular outcomes.

Questions and Answers

Q: With the rise in glucagon, we know after one month in the Ferrannini study that it was no longer significant. In your month-long study, were fasting glucagon levels increased?

A: Yes, they were. We did another study with dapagliflozin that also saw an increase in glucagon.

Q: The compensatory hepatic glucose production is found continuously after two to four weeks but A1c stays down. Do you really feel it attenuates the effect?

A: A1c comes down and stabilizes. In all preclinical studies, endogenous glucose production decreases, and in clamps the ability of insulin to suppress glucose production is improved, probably due to a reduction in glucotoxicity.

Q: You were clever not to conclude anything about whether one is superior to another. Looking into the trials with canagliflozin 300 mg, would you propose that this higher dose of canagliflozin is marginally superior to the others?

A: It is superior to the 100 mg dose, probably due to greater effects on glucosuria. What you’re probably asking is whether there is a significant effect of SGLT-1 inhibition that translates in people with type 2 diabetes. We haven’t done the studies completely, not with tracer uptake. My personal view is that it would be a small effect. I doubt we would be able to detect it as clinically significant. If it does, it would be in the postprandial period.

Q: What is your hypothesis for the elevation in LDL?

A: It wasn’t even noted initially in the individual studies. It wasn’t until all the data was together from all the phase 3 trials that people started to look more carefully. The bottom line is there are lots of postulates, and it appears from observation that glucose ingestion goes up with SGLT-2 inhibitors, so it makes it hard to explain. It remains unknown. It definitely is a phenomenon. At least the three approved agents all show it. Whether it’s different among the three, only head to head studies can answer.

Q: Do you think we have to worry about it?

A: I think the increase in LDL is small enough, zero to 8-9%, that it generally can be handled with statins and diet, and it probably won’t have an effect, though we have to wait for the cardiovascular outcomes trial.

Effects on Sodium Balance and Hypertension

George Bakris, MD (University of Chicago, Chicago, IL)

Dr. George Bakris focused on the modest blood pressure reductions observed with the SGLT-2 inhibitor class, questioning the underlying physiology. Clinical trials with the class have demonstrated a roughly 4-mmHg decline in systolic blood pressure. As most patients included in the trials were not hypertensive or were in good control, Dr. Bakris suggested that studies indicate this decline could be as high as 11 mmHg in hypertensive patients (he noted that based on FDA guidelines the drugs could technically qualify as antihypertensives). In a review of animal studies with the class, he confirmed that the drugs produce an acute natriuresis (urinary sodium excretion), though this effect becomes minimal with chronic use – suggesting it would not account for the entirety of the blood pressure reductions. Likewise, given that blood pressure reductions are the same in patients with advanced kidney disease in whom the osmotic diuresis induced by the class is reduced, he indicated this effect would also be insufficient to explain the blood pressure results. This is complicated as well by the rise in glucagon and increased sympathetic tone caused by SGLT-2 inhibitors, which should theoretically increase blood pressure. In the end, Dr. Bakris was unable to provide an answer for how the drugs produce blood pressure reductions – during Q&A, he hazarded a guess that another gut hormone may be involved, though this would require further study.

Questions and Answers

Q: I understand when these agents have been studied in those with reduced renal function they still see a very good decrease in blood pressure. How can you explain that?

A: I honestly don’t know; that’s the big question in my mind. The natriuresis and osmotic diuresis are proportional to kidney function, so it’s unclear to me what is going on. We have these countervailing mechanisms of sympathetic tone going against you. It’s pure speculation, but is it possible there is another gut hormone involved that is vasodilatory, through SGLT1 or leptin or some other mechanism? This would require further study.

Q: Is there reduced bicarbonate reabsorption with the class?

A: I’ve heard this. I will say in the studies we’ve done I have not seen a change in bicarbonate at all. Nor have I seen a change in potassium. It may be interesting but not clinically relevant.

Comment: In my practice I see greater declines in blood pressure than mentioned.

A: One of the things I do routinely is I stop the diuretic when I start an SGLT2. That’s exactly what you’ll find – these studies usually people are on one antihypertensive and it’s not a diuretic.

Which Patients Might Benefit the Most?

Vincent Woo, MD (University of Manitoba, Winnipeg, Manitoba, Canada)

After briefly mentioning the SGLT-2 inhibitors currently on the market or in development and discussing where SGLT-2 inhibitors are currently positioned in diabetes guidelines, Dr. Vincent Woo provided his opinions on which individuals with type 2 diabetes may benefit most from SGLT-2 inhibitor therapy. He believes that SGLT-2 inhibitors would be particularly fitting for those who are trying to achieve early intensive glucose lowering, those with a high baseline A1c who need combination therapy, and insulin users. Dr. Woo noted that the best predictors of response to SGLT-2 inhibitor therapy are high baseline A1c and high eGFR; meanwhile, baseline BMI has no predictive value for A1c or weight. In closing, Dr. Woo mentioned several other potential areas of application for SGLT-2 inhibitors in the future: prediabetes, type 1 diabetes, obesity (in combination with other agents), and diabetic nephropathy.

Questions and Answers

Q: Which patients are most likely not to benefit from the class and have the most side effects?

A: Obviously, we have a number of patients who have incredibly poor adherence. If you don’t take the drug, you won’t benefit. People who might have side effects to the medication, as has been alluded to earlier, are those who experience reductions in blood pressure and do not monitor it correctly. They can get hypotension and dizziness. I have heard that those on low-carb diets may potentially not do as well either. I haven’t seen data on this, but it makes sense. Obviously, renal function is key, and as A1c is lower, A1c lowering with SGLT-2 inhibitors is lower too. 

Q: Some of my colleagues mentioned anecdotally that balanitis is more common in uncircumcised males.

A: The clinical trials support that those who are circumcised experienced less balanitis.

Nonglycemic Effects of SGLT-2 Inhibition: Long-Term Risks vs. Potential Benefits

Richard Gilbert, MD, PhD (University of Toronto, Toronto, Ontario, Canada)

Dr. Richard Gilbert provided a great overview of a complex area of diabetes science: why SGLT-2 inhibitors might (or might not) have renal protective effects. The theory caught our attention in a big way when we learned about J&J’s CREDENCE outcomes trial for Invokana (canagliflozin); before then, the prevailing opinion was that – if anything – SGLT-2 inhibitors may have slight renal safety risk due to a slight reduction in eGFR in patients initiating therapy. However, that decrease may not be what it initially appears: Dr. Gilbert drew strong parallels to ACE inhibitors, which are renoprotective but show an initial drop in eGFR. Dr. Gilbert proposed a few potential mechanisms that could underlie a benefit. Increased sodium excretion causes a group of cells within the kidney called the macula densa to constrict afferent arterioles, reducing eGFR – this is the body’s mechanism to prevent sodium loss, but in this case it could help by reducing pressure in the glomerulus, where blood is filtered. SGLT-2 inhibitors may also increase uric acid excretion, and effect which is being examined more broadly as a way to improve chronic kidney disease (allopurinol is an example). Dr. Gilbert did note with caution that SGLT-2 inhibitors do not have as wide a range of beneficial effects in the kidneys as ACE inhibitors, meaning that a benefit is far from certain. For now, he is eagerly awaiting results from the CREDENCE outcomes study, which are expected in 2019.

Symposium: ADA Diabetes Care Symposium — Novel Clinical Interventions in Therapy That Impact the Management of Diabetes

Sotagliflozin, a Dual SGLT1 and SGLT2 Inhibitor, as Adjunct Therapy to Insulin in Type 1 Diabetes

John Buse, MD, PhD (UNC, Chapel Hill, NC)

Phase 2 data on Lexicon’s SGLT-1/SGLT-2 dual inhibitor sotagliflozin (LX4211), presented by investigator Dr. John Buse, snagged a spot at the massively well-attended Diabetes Care symposium. Although topline data was released over a year ago, making it into this session was a big deal – this was a standing-room-only event in the largest presentation hall of the conference and the abstract was concomitantly published online in Diabetes Care. The phase 2 study (n=33) ran for 30 days, finding improvements across the board, including: (i) a roughly 0.5% placebo-adjusted reduction in A1c from a baseline of around 8% and reductions in postprandial glucose; (ii) a significant improvement in time-in-range and multiple metrics of glycemic variability; (iii) a 32% decrease in bolus insulin dose (6% reduction with placebo) in the context of better glycemic control; and (iv) a relative weight benefit of slightly over 2 kg (~4 lbs) in a short time period. (We wouldn’t leave out the convenience of a pill for type 1 diabetes.) Improvements in GLP-1, PYY, and urinary glucose excretion suggest that both the SGLT-1 inhibition in the gut and SGLT-2 inhibition in the kidney contributed to efficacy. The presentation likely stokes interest in sotagliflozin for type 1 diabetes, which excitingly has moved into phase 3.

  • We were glad to see CGM used for the duration of the 30-day study and applaud the way Lexicon has focused on “time in zone” to support sotagliflozin’s value in type 1 diabetes. With the room for A1c improvements narrowing with each new class of drugs, we see non-A1c parameters such as glycemic variability as increasingly important for novel products to prove their value for patients (and potentially also to payers if there are parallel improvements in insulin dose and hypoglycemia).
  • The most commonly reported adverse effect with sotagliflozin was nausea, consistent with an increase in GLP-1. Around 25% of patients on sotagliflozin experienced nausea vs. 6% on placebo. However, Dr. Buse noted that nausea was generally of mild intensity, within five days of onset of treatment, and of short duration (1-2 days).
  • Of interest given the euglycemic ketoacidosis issue with SGLT-2 inhibitors, Dr. Buse disclosed that there were two cases of DKA with sotagliflozin. However, he noted that both cases were assessed as pump related and explicitly stated that neither was “euglycemic DKA.” We have heard theories that SGLT-1/2 dual inhibitors may have a lower risk of euglycemic DKA, as they cause less volume depletion and also increase GLP-1 levels, both of which should help reduce the risk of excess ketones. With outcomes data soon to arrive (at least one) for SGLT-2 selective inhibitors, we wonder if Lexicon may look into ways to more conclusively back up their argument on SGLT dual inhibitors and DKA.
  • In addition to the recently initiated phase 3 program in type 1 diabetes, Lexicon is also conducting a second phase 2 study (with JDRF support) studying sotagliflozin in younger, more poorly controlled type 1 diabetes patients.

Questions and Answers

Q: What is the impact of SGLT-1 inhibition on fasting vs. postprandial glucose?

A: I think that’s a little hard to estimate. I think the relative effect of the SGLT-2 transporter is about 20-fold greater than the SGLT-1 transporter using classic kinetics. What the relative contribution on the overall glucose profile is would be harder to estimate, but we believe the SGLT-1 contribution is real and perhaps as much as 50% of the effect.

Q: You said that the reason for the nausea in the study was the increase in GLP-1, but GLP-1 didn’t look like it was increased all that much in the study.

A: It is speculation, but as there was an increase in the two enteroendocrine hormones we looked at, it’s likely that there were increases in others. I’ll admit that that is speculation. More importantly, as opposed to previous hypotheses that SGLT-1 inhibition would be associated with intolerable GI events, the GI adverse event profile was really very mild.

Q: Why would sotagliflozin cause nausea when endogenous GLP-1 does not cause nausea?

A: I really can’t speculate further on that.

Q: Can you comment on the mechanism of the increase in GLP-1?

A: The notion is that by inhibiting proximal glucose absorption and shifting glucose further down the intestinal tract, it changes fatty acid metabolism by bacteria. Exactly the contribution of each of the changes in the enteroendocrine hormone changes is unknown.

Q: Could there be a long-term problem with the change in PYY acting as a vasoconstrictor?

A: That would be speculation. My personal belief is that it’s unlikely to be an issue, but there will be a full development program for the compound in type 1 diabetes, and we’ll be careful to examine cardiovascular endpoints. There is a modest decrease in blood pressure associated with sotagliflozin.

Q: With Ramadan coming up, how safe is it to continue SGLT-2 inhibitors during the fasting month?

A: I saw a paper or abstract that addresses that directly, and my recollection was that things looked pretty good. I had the same thought that a period of prolonged fasting, including no water, could limit the efficacy of these agents due to a loss of urine. But there is something published on that.

Best of Diabetes Care 2014: Diabetic Nephropathy – Biomarkers, Progression, and SGLT2 Inhibitors

George Bakris, MD (University of Chicago, Chicago, IL)

Dr. George Bakris selected three standout papers published in Diabetes Care in 2014 related to diabetes and kidney disease. Notably, Dr. Bakris praised the investigators of BI/Lilly’s SGLT-2 inhibitor Jardiance (empagliflozin) for running the only “proper” study of the effects of SGLT-2 inhibitors on blood pressure (according to Dr. Bakris, EMPA-REG BP was the only SGLT-2 trial that measured 24-hour ambulatory blood pressure rather than simply office blood pressure). Impressively, despite participants’ relatively normal baseline 24-hour ambulatory systolic blood pressure (SBP; ~131 mm Hg), empagliflozin treatment provided a placebo-adjusted reduction of ~3 mm Hg (10 mg dose) or ~4 mm Hg (25 mg dose) (Tikkanen et al., Diabetes Care 2014). For context, the FDA requires only a 5 mm Hg SBP reduction to approve an agent for anti-hypertensive indications, and patients enrolled in trials for anti-hypertensive agents would have much higher SBP at baseline. Therefore, this trial demonstrated that diabetes patients who also have hypertension control can expect significant blood pressure benefits from SGLT-2 inhibitors. The other two papers that Dr. Bakris selected identified promising biomarkers for monitoring diabetic nephropathy. First, urinary adiponectin (uADP) was found to be better than albuminuria, and equally as good as eGFR, in predicting progression from kidney damage to end-stage renal disease in type 1 diabetes (Panduru et al., Diabetes Care 2014). Second, serum TNFR1 concentration was identified as a strong predictor of all-cause mortality in patients with type 2 diabetes and chronic kidney disease (Saulnier et al., Diabetes Care 2014).

Corporate Symposium: Comprehensive Management of DM – Focus on SGLT-2 Inhibitors and Combination Therapy – Emerging Concepts of Renal Effects (Supported by AstraZeneca)

Contribution of the Kidney to Whole Body Glucose Homeostasis: A Critical Re-Evaluation

Zachary Bloomgarden, MD (Mount Sinai Hospital, New York, NY)

Dr. Zachary Bloomgarden provided a mechanistic overview of the kidney’s contributions to glucose homeostasis in people with and without diabetes, a topic that had been somewhat overlooked by the medical community prior to the arrival of SGLT-2 inhibitors. He explained that the kidneys play an important role in gluconeogenesis and appear to account for approximately half of the increase in glucose production seen in diabetes. He also presented evidence showing that the kidneys respond potently to insulin and catecholamines and potentially to glucagon, which likely contributes to much of the altered physiology in diabetes as well. Setting the stage for the subsequent presentations on SGLT-2 inhibitors, Dr. Bloomgarden closed with a review of the data showing a paradoxical increase in renal glucose reabsorption in people with diabetes and a direct relationship between hyperglycemia and the threshold for urinary glucose excretion.

SGLT-2 Inhibitors and the Kidney: Mechanisms and Clinical Efficacy

Lawrence Leiter, MD (University of Toronto, Toronto, Ontario, Canada)

Dr. Lawrence Leiter discussed the mechanism of action of SGLT-2 inhibitors in people with diabetes with both normal and impaired renal function. He began by highlighting the recently updated ADA/ EASD guidelines for type 2 diabetes treatment that include a key role for SGLT-2 inhibitors. He then walked attendees through the intestinal and renal roles of SGLT-1 and SGLT-2 in glucose absorption and discussed the rationale for development of SGLT-1 and/or SGLT-2 inhibitors. He commented that these classes might be able to be used in type 1 diabetes as well as type 2, focusing on the clinical data for Lexicon’s sotagliflozin (LX4211); he did not comment on the recent concerns about euglycemic DKA in this population. Dr. Leiter also emphasized that SGLT-2 inhibitors can still be efficacious in patients with impaired renal function, showing data demonstrating that canagliflozin and empagliflozin still have beneficial effects on A1c, body weight, and blood pressure in that setting. He acknowledged that there is a small risk of volume-related adverse events in certain patients but cautioned against writing off these drugs entirely in patients without normal renal function, even suggesting that the class may prove to be reno-protective.

SGLT-2 Inhibitors and the Kidney: Potential for Renoprotection

Matthew Weir, MD (University of Maryland, Baltimore, MD)

Dr. Matthew Weir provided a nephrologist’s perspective on the factors that contribute to diabetic kidney disease and the potential for SGLT-2 inhibitors to improve renal outcomes. He cautioned that there have not been prospective randomized controlled trials with RAAS blockers examining the effect of reductions in blood pressure or proteinuria on renal outcomes, though many secondary analyses have shown a benefit. There is strong evidence to support the use of renin-angiotensin system (RAAS) blockers like ACE inhibitors and ARBs, but Dr. Weir noted that they only lead to a ~20% reduction in adverse renal outcomes – “a nice start, but nothing to cheer about.” Turning to SGLT-2 inhibitors, he explained that at least in animal models, the drugs appear to reverse the impairment in tubular glomerular feedback (a mechanism by which the kidney corrects hyperfiltration) seen in early stages of diabetic kidney disease. Other effects of SGLT-2 inhibitors, particularly blood pressure reduction, could also potentially have a positive impact on kidney function. Dr. Weir also argued that the initial drop in estimated glomerular filtration rate (eGFR) seen with SGLT-2 inhibitors may actually be a positive sign, as a greater initial drop in eGFR was associated with better long-term outcomes in clinical trials of losartan. As Dr. Weir noted, the CREDENCE renal outcomes trial of J&J’s Invokana (canagliflozin), expected to complete in 2020, should provide much more insight in this area – this is one of the most exciting ongoing studies in diabetes in our view, and a positive outcome would be an enormous win for patients.

Exploring Mono- and Dual Therapy: Implications for Kidney Function

Stefano Del Prato, MD (University of Pisa, Pisa, Italy)

Dr. Stefano Del Prato discussed the safety of DPP-4 inhibitors in type 2 diabetes patients with renal impairment and emphasized the utility of combination therapies. He reviewed data showing that the efficacy of DPP-4 inhibitors in type 2 diabetes patients with moderate to severe renal impairment is the same as in those with normal renal function. In fact, DPP-4 inhibitors in type 2 diabetes have been associated with reductions in hsCRP (a marker of CV risk), albuminuria, and microalbuminuria, perhaps suggesting some nephron-protective effects. Dr. Del Prato continued by stressing that early achievement and durable maintenance of good glycemic control remains the most effective approach to reduce the risk of microvascular complications, suggesting that combination therapy that can simultaneously correct different pathogenic mechanisms is optimal. He emphatically stated that the lesson from UKPDS was that “monotherapy is going to fail no matter what treatment you’re going to use” and that “you will ultimately require a second intervention.” Dr. Del Prato thus pushed for starting with combinations as soon as possible, pointing to DPP-4 inhibitor/metformin and SGLT-2 inhibitor/DPP-4 inhibitor combinations as examples. However, he ended by noting that more studies are needed to provide greater clarity on which combinations to use and whether or not the effects are cumulative.

Panel Discussion

Q: What is the mechanism of the up-regulation of SGLT-2 in diabetes? Is it reversible in any other interventions?

Dr. Lawrence Leiter (University of Toronto, Toronto, Canada):  We talked about upregulation of SGLT-2 for those who are hyperglycemic and why it occurs. How? We don’t know. It’s an adaptive change so that you don’t pee out all your glucose and can preserve energy; at the same time, it may be maladaptive as well as it will preserve hyperglycemia. Will anything reduce the up-regulation? I am not aware of any studies on that. I would assume with chronic euglycemia, the up-regulation should come down to normal. The original data was based on only five patients.

Dr. Stefano Del Prato (University of Pisa, Pisa, Italy): We have experimental data showing that hyperglycemia associated with a lack of regulation of SGLT-2 was normalization of expression. It’s a secondary phenomenon. One hypothesis is that hyperglycemia is the inability to use glucose.

Dr. Yehuda Handelsman (Metabolic Institute of America, Tarzana, CA): Why do we have fat on us even though there’s so much food available? The body doesn’t know; it knows it should keep fat as energy for long-term need when there might be famine. I’m more simplistic. I think SGLT-2 is there to make sure it brings every glucose molecule it sees back to the system, and it tries to do so until it cannot. It looks simple: you bring down glucose and it stops doing it because it’s not exposed to as much glucose. That’s a simplistic view. The question is can we prove it.

Dr. Zachary Bloomgarden (Mount Sinai Hospital, New York, NY): It’s fascinating to realize there’s no physiologic way we can imagine the situation of diabetes having evolutionary pressure to cause increased expression of SGLT-2. It’s probably a more epi-phenomenon of something that evolved in response to dietary sodium and secondarily to glucose loads, but it’s a detriment.

Dr. Handelsman: It points out the fact that what we think is simple never is. Some of us were involved when we used to think the effects of bariatric surgery were mechanical, but it’s actually a very complex biological phenomenon involved with gut action. When SGLT-2 inhibitors came out, we looked at them as putting a block on a gate, when in reality it’s a quite complex metabolic phenomenon. We’re changing how SGLT-2 works. Is maladaptation going on? I asked Dr. Weir before whether glucagon works in the kidney? He’s been doing this forever, but no one’s studied this. How many of you remembered that the kidney does gluconeogenesis? I bet the majority of you forgot until five years ago. The body does have energy management that’s involved in insulin resistance. It’s a protective mechanism, though that’s a different story. The body tries to protect itself.

Q: Could you please comment on the side effects of hypokalemia and the warning risk of bladder cancer in SGLT-2 inhibitors? How clinically relevant are these?

Dr. Weir: On potassium, the potassium signal is miniscule at best especially if you compare it to clinical practice and put the patient on anti-inflammatory drugs or ARBs. The latter two tend to boost potassium. If you see any change, you might see 0.1 or 0.2 with SGLT-2 and that’s under conditions that are pretty unusual. I’ve looked at databases and written papers. This signal is minimal compared to other commonly used drugs. In any other patients with diabetes, monitor potassium as you feel comfortable. If they’re taking other herbal supplements, just pay attention to it. But I don’t think twice about it.

Dr. Leiter: With regard to bladder cancer, in the dapagliflozin clinical trials, there was an imbalance in the number of bladder cancers in those subjects on dapagliflozin, although the numbers were small. There is no plausible mechanism for it. The vast majority of these subjects had hematuria at baseline or shortly thereafter so it was likely just pre-existing bladder cancer that was unmasked during the clinical trial. Furthermore, there are no SGLT-2 receptors found in the bladder and no animal models have shown increased risk for it or any other cancer. The large, ongoing trial with dapagliflozin is the DECLARE study which is designed to look not only at cardiovascular safety but also to have a sufficient number of bladder cancers to assess risk; there are over 17,000 patients in the study. Importantly, this is monitored by a DSMB and no concerns have been expressed to date about the risk of bladder cancer.

Dr. Handelsman: AACE put together a panel looking at diabetes and cancer and medications and how they may affect it, and we did use a basic scientist to look at the time to effect. In none of the drugs, including dapagliflozin, did we find anything relevant. It’s possible that the numbers will change but right now it’s good news. Has anyone felt that potassium elevation is an issue managing patients on SGLT-2 inhibitors? When you give a diuretic, sometimes you think there will be less potassium, not more. There’s some data in some patients, specifically those with renal impairment, to be careful.

Q: Is volume status related to the mechanism of weight loss?

Dr. Del Prato: It’s most likely the diuretic effect that’s responsible for the very early reduction, but maintenance over four years is unlikely due to the diuretic effect. If you calculate it, the amount of sugar you lose through the urine accounts for a good deal of the reduction in body weight. There have also been investigations looking at body composition showing a reduction in all components of adipose tissue, both visceral and subcutaneous. There’s a real reduction in adipose tissue, which accounts for the 3-4 kg body weight reduction. The diuretic effect is early, the same as you see with a diet. On a diet, the initial drop is liquid loss and then it’s fat mass.

Dr. Handelsman: Any study with any drug compared to sulfonylureas showed similar results. It’s pretty much line in with everything.

Q: What are the effects on heart failure?

Dr. Leiter: In the five ongoing cardiovascular outcomes studies with SGLT-2 inhibitors, heart failure will be assessed and adjudicated as one of the clinical endpoints. To my knowledge, no study has been initiated specifically in patients with heart failure.

Dr. Handelsman: It’ll be interesting to see what the long-term effect will be. There’s a very small reduction in blood pressure. I don’t think it’s direct one to one but it’s possible.

Q: Initially based on its mechanism, we expected that SGLT-2 inhibitors should be more capable at lowering postprandial glucose. The recently announced guideline didn’t endorse them as a good choice for reducing postprandial glucose. In general, the overall data showed they were more reliable in reducing fasting glucose. What explains the lack of a postprandial effect?

Dr. Del Prato: I think the system works as a function, which has been explained: the more glucose the tubule sees, the more is lost through the urine. If you consider that the postprandial period is limited in duration compared to the non-postprandial period, the cumulative effect may be greater on fasting glucose, but the effect on postprandial glucose is there as well. I’m not sure I understand what you mean by a lack of effect.

Q: I mean it’s supposed to be more reliably dose-dependent on lowering postprandial glucose, but the published trials suggest that’s not the case.

Dr. Del Prato: Dr. Leiter alluded to that. We think that may be a compensatory mechanism from SGLT-1. Any time you increase the tubular load, more glucose is not reabsorbed by SGLT-2 but is available for SGLT-1. It accounts for more than 10% of the reabsorption, and now with a larger volume, it may reduce the effect.

Dr. Handelsman: The system works the moment it gets glucose. It’s the fastest thing it’ll respond to. I caution you not to rely primarily on phase 3 studies to assess postprandial effects. That’s true for all drugs. The FDA designed trials not to capture prandial glucose. That’s why in publications, you see differently. The clinical effects are quite impressive on both fasting and prandial. DPP-4 inhibitors have more effects on prandial than on fasting. GLP-1 works on the whole story. SGLT-2s also see prandial over fasting.

Q: Do we currently know how this drug’s mechanism increases LDL-cholesterol? Any mechanism?

Dr. Leiter: We don’t have a good explanation. These drugs do tend to raise LDL by a small amount of about 2%-3%. The higher dose of canagliflozin may raise it a little more by about 4%-5%. We don’t know why. Other diuretics also lead to a small increase in LDL. Another hypothesis is that we know there is compensatory overeating in patients on SGLT-2 inhibitors that limit the extent of weight loss; perhaps there is also a change in macronutrient consumption. But that’s totally speculative. There are no studies on food selection.

Dr. Handelsman: HDL also goes up. LDL barely changes. Clinically, hopefully long-term studies will show relative efficacy. ADA wants everyone on statins.

Q: Is there any relevant information about reduction in uric acid for renal protection and as a cardiovascular risk protector?

Dr. Weir: I’m not aware of consistent effects of SGLT-2 inhibitors on uric acid. If there are, they’re modest at best. I’m not aware of data tying uric acid to cardiovascular events. There is a directionality in the relationship to kidney function, but we don’t know which association drives what. That area needs to be explored. There may be interest in hypouricemic therapies, but they’re not prime time.

Satellite Corporate Event

Lexicon Pharmaceuticals Meet the Expert Session

John Buse, MD, PhD (UNC, Chapel Hill, NC), Paul Strumph, MD (Lexicon, The Woodlands, TX), David Powell, MD (Lexicon, The Woodlands, TX)

Lexicon hosted an evening meet-the-experts session following the presentation of positive phase 2 data on its SGLT-1/SGLT-2 dual inhibitor sotagliflozin earlier in the day. The panel discussion at the evening event touched upon the recent concerns about euglycemic ketoacidosis with selective SGLT-2 inhibitors, partly in response to questions about the two cases of DKA observed in the sotagliflozin trial. Dr. John Buse noted that both cases were assessed as pump related and were “clearly” not euglycemic. More broadly, in remarks reminiscent of his presentation on this subject at AACE, Dr. Buse suggested that the phenomenon is unlikely to be a “showstopper” for SGLT-2 inhibitors as long as people are aware of it and know how to treat it. Dr. Paul Strumph expressed hope that future trials will demonstrate a lower risk of DKA with SGLT-1/SGLT-2 dual inhibitors than selective SGLT-2 inhibitors, as the dual inhibitors cause less volume depletion and increase GLP-1 levels, both of which should help reduce the risk of excess ketones. With outcomes data soon to arrive for SGLT-2 selective inhibitors, we wonder if Lexicon may look into ways to more conclusively prove this suspected advantage. See below for a full transcription of the panel discussion, which also addressed topics including sotagliflozin’s effects on blood pressure, LDL cholesterol, and the GI tract as well as Lexicon’s development plans – Dr. Strumph confirmed that phase 3 is ongoing but declined to speculate on a launch timeline.

Panel Discussion

Q: Do you think DKA will be a showstopper in type 1 diabetes?

Dr. Strumph: I assume you mean with compounds in the SGLT-2 inhibitor class. It will be a risk-benefit decision. We now know that DKA with SGLT-2 selective inhibitors can be masked or present differently than practitioners are used to. It’s a burden on everyone in the field. We are a dual inhibitor, but we still have a burden to generate data to indicate that with appropriate protocols. We’re informing investigators and patients that DKA may be present and we will come to the regulatory authorities with a complete dataset showing risk and benefit. We believe the risk of DKA is still to be determined but with appropriate protocols, data will be available and it’s worth moving forward. So far, DKA is treatable if recognized early. DKA occurs in type 1 diabetes. It occurs in 5% of people every year on average. If the A1c is above 10%, it’s 20% per year. We’ve trained people to check ketones if they feel sick and their sugar is above a certain level. Now we have to train them to just check ketones if they feel sick. If it’s positive, implement “sick day rules” for treatment of ketones, which include frequent injections of short-acting insulin, and contact your doctor for further instructions. Ketosis can be treated and is usually reversible.

Dr. Buse: I’m pretty certain that in the Wild Wild West of people taking SGLT-2 inhibitors off label, for those therapeutic zealots, the rate of DKA is probably 10%. It’s maybe twice what you’d expect. The problem is only the euglycemic part. If we had a way of figuring that out, there would be no major issue. And we do, we tell people to check ketones if they don’t feel well, period. I suspect it will work out well. I’ve counseled patients not to take SGLT-2 inhibitors if they’ve had DKA, but if they want to, lower doses are less likely to be associated with DKA than the full dose that many take for weight loss. I don’t think it’s a showstopper or deal-breaker.

Q: What are the odds of the drug not getting to market?

Dr. Strumph: We’re coming to the market.

Q: Has the safety committee reviewed the decision that the DKA cases were due to the pump rather than the drug in view of recent reports?

Dr. Buse: The short answer is no because there was no DSMB. The longer answer is that it’s arguable who started the hoopla about this, but I think it was me, and the first case of DKA I saw was in this trial. One of the two cases was in our site and was clearly a pump issue. There was other stuff going on. There was nothing euglycemic about this. The person’s sugar was like 560 mg/dl. It was regular old DKA. I’m confident about that case. I’ve seen the details of the other and it also seems like a pump malfunction and was not euglycemic. It’s not like the cases that have been reported.

Q: The euglycemic cases might stand out because they’re unusual but if it’s happening to others, people might just be shrugging their shoulders.

Dr. Buse: We’ve done a database search. The index case for me was a person who, having seen the results in this trial, decided to do it on her own with canagliflozin and her personal physician and she got DKA. I would say there are a number of things SGLT-2 inhibitors do that might tip the balance. Less insulin exposure, more glucagon, volume constriction – a number of things make it easier. Some are less likely with an SGLT-1/SGLT-2 dual inhibitor. I’ve done a database search of 1,560 patients prescribed an SGLT-2 inhibitor and ten cases seem to be ketoacidosis in the setting of SGLT-2 inhibitor therapy. Nine out of ten have a glucose under 300 mg/dl. It’s mostly euglycemic. There are other cases of real DKA as well.

Dr. Strumph: It may be different with a dual inhibitor and I would like to provide data that it is different. First, I believe we will at least meet what’s out there on the market in terms of efficacy and based on lower urinary glucose excretion. With less glucuresis, you have less likelihood of volume depletion. If that’s true, and we will have to demonstrate it, but that would be less of a known risk factor for ketoacidosis. Additionally, we’ve seen a published preclinical study and since then some clinical studies showing that GLP-1 agonism decreases ketosis and decreases glucagon. We have provided data in this study, in type 2 diabetes, in healthy volunteers, and preclinically that GLP-1 increases with our drug and not with the selective SGLT-2 inhibitors. So two risk factors for ketosis may be less with our drug. Again, that’s forward-looking and we need data, but I look forward to being in the race.

Q: When are you planning to launch in the EU and the US?

Dr. Strumph: We’re in phase 3 and the completion date is public. I wouldn’t have a job if I talked about the launch date.

Q: What was the 24-hour increase in urinary glucose excretion?

Dr. Paul Strumph (Lexicon, The Woodlands, TX): We don’t have the 24-hour urinary glucose excretion data in type 1 diabetes, but we have data indicating a three-fold increase over three hours vs. placebo. We need to provide that data at a later time. In our type 2 diabetes dose-ranging study, the 24-hour range was between 50 and 60 mg with sotagliflozin vs. 15-20 g with placebo. Urinary excretion plateaued at 200 mg/day and didn’t increase with the maximum dose of 400 mg/day yet glycemic efficacy was better, consistent with the difference being the local effect on the GI tract of SGLT-1.

Q: How much ingested carbohydrate was not absorbed with sotagliflozin vs. control?

Dr. Strumph: We don’t have data in humans.

Q: Is GLP-1 increased?

Dr. Strumph: It is increased. In the type 1 diabetes study we only measured it at the mixed meal at breakfast, so we don’t have full data in type 1 diabetes.

Dr. David Powell (Lexicon, The Woodlands, TX): When we treated diabetic mice with sotagliflozin, we saw a marked increase in glucose in the small intestine and colon along with a decrease in the pH of colon contents, which we assume is due to fermentation of glucose to short-chain fatty acids in the colon. This dramatically increased circulating GLP-1 and PYY levels. The delayed glucose absorption, accompanied by blunted blood glucose excursions and increased GLP-1 levels, are all due to SGLT-1 inhibition and contribute to the improved glycemic control. We did not perform a study in these mice that quantitated the extent of glucose absorption.

Dr. John Buse (UNC, Chapel Hill, NC): I bet you won’t find much glucose in the stool. Malabsorption doesn’t really happen.

Dr. Powell: I agree, what we see is delayed glucose absorption. Some of the glucose is ultimately absorbed as short-chain fatty acids. This is an advantage because short-chain fatty acids are potent inducers of GLP-1 and PYY release by the colon.

Q: Should I be concerned about GI side effects?

Dr. Buse: It seems simple. The answer is yes, you should be concerned given the discussion about SGLT-1 inhibitors, but now you should be relieved because we did the experiment and there were no GI adverse events. There was mild nausea in three patients for one day early in therapy. It’s not an unreasonable concern, but it’s largely been mitigated.

Dr. Powell: If you look at humans treated with SGLT-1 inhibitors or people who lack SGLT-1 and eat a diet that contains some glucose, they may not have diarrhea from the dietary glucose but they may have abdominal pain and nausea. The mechanism depends on delayed glucose absorption in the small intestine. If enough glucose is eaten, they have symptoms. Humans lacking SGLT-1 tolerate glucose-containing diets over time. Mice fed a glucose-containing diet have diarrhea after receiving a high enough dose of a selective SGLT-1 inhibitor, but they adapt quickly. After a few days the diarrhea often completely resolves on the same diet and same dose of the inhibitor.

Dr. Buse: If you swallow the tablet, SGLT-1 is on the luminal side, and you have liquid sotagliflozin flowing past SGLT-1 and inhibiting those. Then the liquid keeps heading south and some is absorbed. How much of the SGLT-1 inhibition is first-pass on the luminal side and how much is in a steady state between the plasma and free exchange?

Dr. Powell: I don’t know if there’s much SGLT-1 inhibition from the systemic side. The SGLT-1 inhibition appears to occur on the luminal side. In the laboratory, we incubated sotagliflozin with cells expressing SGLT-1 and then tried to wash the sotagliflozin off. We found that the cells were unable to transport glucose for many hours after washing. Either sotagliflozin has a slow off-rate from SGLT-1 or it gets into the cells, can’t be washed off, and somehow inhibits SGLT-1 this way. I don’t totally understand the mechanism, but this inhibition may be more prolonged for sotagliflozin than for most SGLT-1 inhibitors that have been studied.

Dr. Buse: You have numerical insulin dose reductions that are similar across all meals and statistically positive for breakfast. How much of that is the tablet tumbling in front of the meal?

Dr. Powell: I was surprised how long glucose transport mediated by SGLT-1 was inhibited in cultured cells after sotagliflozin was washed off. I believe this effect explains why sotagliflozin has a prolonged ability to inhibit intestinal SGLT-1 in individuals with diabetes.

Q: What happens to LDL?

Dr. Strumph: There’s a small increase of 2-3 mg/dl, similar to SGLT-2 inhibitors. We don’t have much data in type 1 diabetes, but in type 2 diabetes it’s consistent with the SGLT-2 selective inhibitors.

Q: What has given you the confidence to run phase 3 trials in parallel?

Dr. Strumph: It’s the same confidence that has allowed all of the type 1 and 2 diabetes programs I’m aware of to not do trials serially by the time you get to phase 3. I can’t speak to what other SGLT-2 inhibitor companies are doing in type 1 diabetes, but it’s fairly common in diabetes, in type 1 diabetes insulin studies and type 2 diabetes insulin and molecule studies to do phase 3 in parallel. It’s too prolonged a process with a limited patent life to do long studies in series.

Q: What are the effects, if any, of sotagliflozin on blood pressure?

Dr. Strumph: It has been published in Diabetes Care that sotagliflozin demonstrated a reduction in blood pressure in type 2 diabetes. It was a dose-ranging study with 60 patients per arm, and with systolic blood pressure greater than 130 mm Hg, there was a reduction of 14 mm Hg. For normotensive people, a reduction was not seen. That’s very encouraging. The SGLT class tends to lower blood pressure, and we may be able to define better utility in type 1 diabetes lowering blood pressure. In the current study, placebo lowered blood pressure 3.9 mm Hg and sotagliflozin lowered it 4.9 mm Hg, so there was no difference.Q: What would happen if you added an alpha-glucosidase inhibitor?

Dr. Powell: Alpha-glucosidase inhibitors block absorption of glucose that is present in complex sugars. Unlike sotagliflozin, they do not inhibit the absorption of simple glucose molecules. The effect of both together would be additive at best in terms of efficacy.

Product Theater

Invokana: An Individualized Approach to the Treatment of Adults with Type 2 Diabetes (Sponsored by Janssen)

Serge Jabbour, MD (Jefferson University, Philadelphia, PA)

In this packed product theater, Dr. Serge Jabbour discussed J&J’s Invokana’s (canagliflozin) mechanism of action, its utility across a broad range of adult patients with type 2 diabetes, its dosing and administration, and its use in those with moderate renal impairment. He noted that in clinical trials, Invokana was effective and well tolerated across various circumstances – as monotherapy, in dual therapy, in triple therapy, in older patients, and with insulin. In particular, Dr. Jabbour highlighted that Invokana produced greater A1c reductions, less hypoglycemia, and more weight loss compared to glimepiride. He pointed out that in clinical trials, the top four most common adverse events were female genital mycotic infections (10-11%), urinary tract infections (4.3-5.9%), increased urination (4-5%), and male mycotic infections (~4%). He commented that if his patients experienced mycotic infections as a result of Invokana treatment, he would not discontinue Invokana, but rather, treat the infection. In closing, Dr. Jabbour mentioned that Invokana has been prescribed over 3 million times since launch, and has 75% preferred commercial and Medicare access – in a product theater at AACE, Dr. James Gavin (Emory University, Atlanta, GA) noted that these statistics are already out of date, as Invokana has now reached over 4 million prescriptions and achieved 80% coverage. During the subsequent panel discussion, Dr. Carol Wysham (Rockwood Clinic, Spokane, WA) and Dr. Jabbour both commented that ketoacidosis has been a rare occurrence among individuals taking SGLT-2 inhibitors; though this may be the case, Dr. Wysham intimated that these agents should be used with caution for patients with type 1 diabetes. 

Panel Discussion

Serge Jabbour, MD (Jefferson University, Philadelphia, PA), Carol Wysham, MD (Rockwood Clinic, Spokane, WA), and Greg Mitchell, MD (Anne Arundel Medical Center, Annapolis, MD)

Q: Can you discuss the potential risk of ketoacidosis with Invokana?

Dr. Wysham: There were 20 reports of ketoacidosis [captured by the FDA adverse event reporting system] among those who had been started on SGLT-2 inhibitor therapy. There is very little description of the cases themselves. Some were individuals with type 1 diabetes. Some had a precipitating cause – surgery, trauma, infection, dehydration. It appears to be a rare complication – over 400,000 patients or so received treatment in the reporting timeframe. It does appear that things that affect dehydration/food intake may be the precipitating cause. The problem is that we don’t understand the patient characteristics in detail. We suspect a significant proportion of them were on insulin, and that their doses were reduced upon starting the medication. Furthermore, when they get normal glucose values, they feel like they don’t need to take (as much) insulin, so I suspect some patients made additional adjustments.

Dr. Jabbour: Janssen looked at 17,000 patients – 10,000 in their phase 3 trials, and 7,000 in other trials. The incidence of ketoacidosis was less than 0.1%. It was very rare. There were only a small number of patients who experienced ketoacidosis. I wouldn’t change the way I practice on a daily basis because of this risk.

Dr. Wysham: We have patients with longstanding diabetes on MDI, and some are on insulin pumps. For those individuals, I would contemplate measuring C-peptide before they use an SGLT-2 inhibitor, just to make sure we are not dealing with someone who is so insulin deficient that they are at risk of ketoacidosis. 

Q: Is there any difference between the three SGLT-2 inhibitors?

Dr. Wysham: We don’t have any head-to-head studies so it’s really difficult to project any differences. There was a small study that looked specifically at the differences in urinary glucose excretion with dapagliflozin and canagliflozin. There was more urinary glucose excretion with canagliflozin. That’s the only head-to-head study I’m aware of, and it was in individuals without diabetes.

Dr. Mitchell: Canagliflozin is indicated for GFRs as low as 45, whereas dapagliflozin is only indicated for use in patients with eGFRs of 60 or above. In comparing selectivity, dapagliflozin may be a bit more selective. Canagliflozin has a bit of SGLT-1 inhibition, which may be the mechanism by which the 300 mg dose reduces postprandial glucose absorption.

Q: Is there a benefit of SGLT-2 inhibitors versus GLP-1 agonists?

Dr. Wysham: There are no head-to-head studies comparing the agents in terms of A1c. Both are effective, and both are associated with reductions in weight. Often times patients prefer orals over injectables.

Q: Are there any studies in type 1 diabetes?

Dr. Wysham: Phase 2 studies are in progress, but in light of the new reports of ketoacidosis, we should tread very lightly on that therapy [for use in type 1 diabetes].

Product Theater

SGLT2 and DPP-4 Inhibition: Multiple Options for Improved Glycemic Control in Adults with Type 2 Diabetes (Sponsored by Lilly/BI)

James Gavin, MD, PhD (Emory University, Atlanta, GA)

Dr. James Gavin provided an overview of clinical considerations for prescribing BI/Lilly’s SGLT-2 inhibitor Jardiance (empagliflozin) and SGLT-2/DPP-4 dual inhibitor Glyxambi (empagliflozin/linagliptin). He placed a strong emphasis on the need for early combination therapy and extra-glycemic effects. At one point in his remarks he made an implicit contrast to J&J’s Invokana, stating that with Jardiance and Glyxambi, there is no need to adjust dosing based on GFR, just a hard cutoff of 45 ml/min/1.73 m2 below which the drug is contraindicated. Conspicuously missing from the presentation was discussion of the cases of SGLT-2 inhibitor-associated ketoacidosis recently addressed by FDA. This was by far the audience’s most significant concern during Q&A. Dr. Gavin framed the unexpected ketoacidosis as a problem seen more in people already under some other sort of physiological stress, people using the drug off label (e.g., people with type 1 diabetes), and as an issue that providers can teach patients to monitor and avoid.

Questions and Answers

Q: I see that the risk of ketoacidosis is on many people’s minds.

A: The FDA announced a case series of some 20 cases or so. It is important to contextualize this. I don’t want to trivialize the fact that this was an important clinical observation in patients treated with SGLT-2 inhibitors. These patients were mostly type 1 patients, so they were using the drug off label. There were a few type 2 patients as well. In most instances, again to contextualize this, these are patients who were already under some form of stress. For example, surgical patients, hospitalized patients, patients who had reasons to be less hydrated than they should have been, and patients who were probably under-insulinized in the face of high glucagon levels, which would set up them up for increased ketone formation. Historically, the way we have taught patients is to use their blood glucose as a signal for being aware of ketone levels. If the blood glucose was high and they felt sick, that’s when we’ve told them to check ketones. We know now with these SGLT-2 inhibitors, we’ve removed the hyperglycemic signal as an alert for checking ketones. So we have to revise our old advice. Now we have to tell patients that if they are unable to remain hydrated or if they feel sick they may need to start checking ketones independently of what their glucose level is. That’s the only way we’ll know.

Q: Was there any evidence of synergism in therapeutic efficacy for Glyxambi?

A: One thing we’ve been puzzled about is the absence of a dose response relationship as we go from a certain dose to double or triple that dose. And then when we add different agents, if one agent has a 0.7% lowering of A1c and the other has a 0.8% ability, we don’t predictably always see 1.5% (additive) reductions or greater. That is simply not the physiology of these agents. And this is not new. We see this over and over. The only explanation I can offer is the closer you get to normal, the more difficult it is to see huge leaps beyond that. Which is why I’m encouraged by what I see with this fixed dose combination in the form of empagliflozin/linagliptin. We’re seeing people actually get to goal. This is 6.7-6.9% A1c. We see a level of efficacy that is still highly impressive even if not synergistic. One other principle – we’ve talked about this for years. There is much more benefit of submaximal doses of different agents working in concert than maximizing the dose of a single agent. We’re also seeing this come to light here.

Q: Has there been any trial of these drugs in teenagers?

A: I don’t think there’s a trial underway. There are some pediatric trials under way with some therapies including a DPP-4 inhibitor. I am not aware of any SGLT-2 inhibitors and certainly nothing with a combination product.


-- by Melissa An, Eric Chang, Jessica Dong, Stephanie Kahn, Emily Regier, Joseph Shivers, Jenny Tan, Tony Thaweethai, Manu Venkat, Vincent Wu, and Kelly Close