FDA Advisory Committee Meeting for BI/Lilly’s Empagliflozin in Type 1 Diabetes

November 13, 2019; Silver Spring, MD; Notes – Draft

Executive Highlights

  • FDA’s Advisory Committee meeting for BI/Lilly’s Jardiance (empagliflozin, 2.5 mg dose) in type 1 diabetes concluded with a 14-2 vote against approval. Consensus formed around the need for a larger evidence base in terms of empagliflozin’s efficacy and safety in the type 1 population, as the EASE-3 program contained just 241 participants over 26 weeks of drug exposure, and only five combined certain DKA events across drug and placebo groups.

  • Panelists overwhelmingly agreed on unmet need in type 1 and emphasized that they saw promise for the therapy; they stressed that longer-term follow up studies with more patients in order to develop a more definitive assessment of DKA risk was also needed. Meaningful interest was shown in the therapy’s potential to slow kidney related complications.

  • Today’s discussion made clear that there is a path forward for SGLT-2 inhibitors for type 1 diabetes. Multiple panelists expressed support for a probable “sweet spot” of efficacy and safety with the lower dose, and pointed to larger studies that may help confirm this profile.

  • Notably, panelists also expressed a strong interest in discussing broader endpoints in addition to A1c – time in range was raised multiple times throughout the discussion and in closing comments, FDA said it was open to more discussion of this endpoint. What is required for further validation is something that will likely be further discussed.

  • Lilly/BI gave a press release update following today’s proceedings – see here. Notably, the companies remain optimistic concerning their submission of the 2.5 mg dose: "We continue to believe the totality of data from the EASE program indicates a favorable benefit-risk profile for empagliflozin 2.5 mg in adults with type 1 diabetes and look forward to continuing to work with the FDA in this review process." 

  • See below for more detailed highlights from the day, including brief summaries of each of the eight speakers during the Open Public Hearings and a detailed breakdown of the final vote and panelist rationales for their voting decisions.

Top Four Highlights

1. Next Steps for Empagliflozin in Type 1: Cardio and Renal Outcomes Studies?

Given the nature of today’s discussion and ultimate tally, while it seems clear that FDA will not approve BI’s application in its current state, there did seem to be optimism that a low dose SGLT inhibitor could one day be approved given more expansive data to better understand its entire risk/benefit profile. That said, there was also doubt whether DKA risk in clinical trials could be assessed in a controlled setting. Moving forward, we sincerely hope to see outcomes trials done with SGLT inhibitors in type 1 diabetes to determine whether the powerful reductions in renal and cardiovascular events seen in type 2 diabetes translate to the type 1 population – Dr. John Buse said through an open comment speaker that he felt very strongly that those with type 1 and chronic kidney disease would have a very strong likelihood of seeing reduction of risk. We strongly believe that this is one of the most important considerations for type 1 patients who may consider starting an SGLT inhibitor. During today’s discussion, we heard a few panelists reference this call for such outcomes trials, especially in terms of including renal outcomes as secondary endpoints in any future trials of these agents. While many have been interested in a type 1 outcomes trial for quite some time, the elephant in the room remains where the funding for such a major investment would come from. CVOTs and renal outcomes studies are expensive to carry out, and many sponsors may look at the smaller type 1 diabetes market and deem the risk/reward ratio to not be favorable enough to actually carry out such a trial. Encouragingly, hearing KOLs hypothesize that it’s likely that CV and renal benefits extend to type 1 should help moving forward. We wonder whether the multiple SGLT sponsors—Lilly/BI, J&J, Merck, and AZ—might collaborate to carry out such an outcomes study for the entire class, perhaps with government and other funders. The savings to the area would surely be in the billions of dollars and the impact might well be seen earlier than anticipated though that is hard to speculate on – the trial may not be particularly easy to enroll, either. We hope that next steps for empagliflozin in type 1 include some sort of movement toward outcomes trials in the type 1 population looking at potential cardio and renal protective effects, although we do realize the barriers for such a trial are high.

2. Panelists Critique Small Sample Size and Short Duration of EASE-3; Call for Larger + Longer Follow-Up Studies; Positivity around Time in Range + Outcomes Beyond A1c

The majority of panelist critique of BI’s application surrounded what they deemed to be an inadequate study size and length for the sole phase 3 trial, EASE-3. We are a bit surprised this was not discussed earlier with FDA by the sponsor though it was said in the discussions that the sponsor originally planned for a higher dose to be submitted. The disconnect appeared to be around submitting a lower dose and only one trial. Notably, the study included just 241 patients on the 2.5 mg dose of empagliflozin and was carried out for just 26 weeks – we would have assumed this would’ve been discussed earlier whether this was “enough” but the panelists made clear they would like to see significantly more data. The panelists were also slightly blithe in our view of how easy this would be or whether the support would be there for this. Multiple Ad Comm members questioned the durability of the 0.28% drop in A1c seen at 26 weeks and surmised whether or not an upward drift would be seen if the study were carried out for a longer duration, particularly since part of the benefit was already placebo seeing a higher A1c. Although a set length was not decided upon, most panelists seemed to suggest a second study spanning 52 weeks to get a better sense of long-term changes and a more definitive evaluation of DKA risk (more on this below).

  • Generalizability to the real-world setting was also mentioned multiple times. BI/Lilly were complimented on the company’s dedication to carefully educating and keeping patients in the trial safe, and ironically, even though “real world data” is not typically sought, panelists pointed out that the level of care shown in the trial would not be as generalizable so may not be as relevant to the real world. It makes sense that the results seen would not be as applicable. Importantly, the sponsor gave each of its study participants a point of care device to track their blood glucose and ketone levels – multiple panelists noted how helpful this device was in helping participants mitigate DKA risk and make clinical decision in partnership with their physicians. While the point was raised that it’s not clear whether all patients in the real-world would be able to afford such monitors in practice, we believe its use could be easily mandated. Continuous kidney monitoring, similar to continuous glucose monitoring, was also raised briefly and we are glad this is being pursued on the device side.   

  • There was some debate regarding the reliability of the ~0.3% decline in A1c seen at 26-weeks on empagliflozin. Dr. Manjunath Pai pointed out that the majority of decline seen in empagliflozin compared to placebo was driven by increases in the placebo arm A1c rather than an absolute drop in A1c from the empagliflozin arm. Dr. Pai pointed out that pramlintide, the only currently approved adjunct in type 1 diabetes, also demonstrated a ~0.3% drop in A1c; however, both placebo and pramlintide were moving in the same direction, further fortifying the delta seen. Other panelists, such as Dr. Michael Blaha and Dr. Erica Brittain, however, were quick to push back on this assertion, claiming that change from placebo is indeed the right measure.

  • On a positive note, we sensed an increased enthusiasm over time in range and other beyond A1c metrics in the evaluation of empagliflozin in type 1. Multiple panelists pointed to the benefits seen in time in range as a major sticking point for patients considering taking the drug (one cited friends who were “obsessed” with optimizing their child’s time in range), and called for further studies to evaluate the robustness of this benefit over a longer period of time. While FDA does not currently consider time in range to be a valid endpoint due to what is characterized as an unclear link to complications over the long-run, there was clearly greater interest from panelists on the metric and Ms. Lisa Yanoff emphasized in closing comments that FDA was very open to discussing additional endponts. Panelists also repeatedly pointed to improvements in weight loss and blood pressure, along with overall patient well-being and satisfaction, as other points to consider as part of empagliflozin’s overall efficacy profile.

3. FDA Questions Sponsor’s DKA Adjudication Process; Finds a Number of “Clinically Significant Serious Events” Were Not Classified as “Certain DKA”

Much of the FDA’s clinical safety assessment was focused on concerns regarding the reliability of BI’s DKA adjudication system, consisting of the classifications:

  • Certain DKA: acidosis and ketosis, with or without DKA symptoms

  • Unlikely DKA but Ketosis: single high ketone without acidosis or DKA symptoms

  • Potential DKA: DKA symptoms and ketosis or ketosis, persistent high ketones

  • Unlikely DKA: normal pH or normal ketones, irrespective of DKA symptoms

While the sponsor did not find an imbalance in “certain DKA” events with empagliflozin 2.5 mg vs placebo (2 events vs. 3 events), FDA did not agree with some classifications funneled into the “unlikely ketoacidosis but ketosis” category instead of “certain DKA.” Despite not checking off BI’s list of qualifiers to make up “certain DKA,” FDA found these events to be “clinically significant serious events” that required hospitalization and prompt intervention, such as discontinuation of the drug. FDA presented three specific examples of patients adjudicated as “unlikely ketoacidosis but ketosis,” including (i) a 20-year-old woman, who was hospitalized for three days and permanently discontinued taking empagliflozin; (ii) a 51-year-old woman, who visited the ER twice after continued symptoms of nausea and vomiting; and (iii) a 32-year-old woman, who required insulin infusion and IVF following vomiting and dehydration. Instead of only focusing on “certain DKA” events, FDA asserted that total adjudicated DKA events would be a more relevant measure. This point was broadly agreed upon by the voting members of the panel as well, with several panelists noting that it might be more instructive to think of DKA along a spectrum of risk, rather than a simple binary category. Overall, 13 such events (certain + potential + unlikely ketoacidosis but ketosis) were demonstrated in the population on 2.5 mg empagliflozin compared to 6 events on placebo. See the breakdown below. We are curious how events might be perceived that could be moving “toward” DKA but where patients themselves moved from vomiting etc. at home toward full health – it seems like once a patient is vomiting and in the hospital, that is hard to imagine the patient is not in DKA. We would have assumed that questions like this would have been more standardized based on BHB levels etc. than perhaps they actually are – it was clear that the panelists did not all have full knowledge of ketosis vs. ketoacidosis.

4. Fueled by Shared Goal of Fulfilling Unmet Need, Panelists Optimistic About Potential Benefits of 2.5 mg Dose of Empagliflozin

Throughout today’s meeting, we felt a shared desire and sense of urgency to fill the current unmet need in type 1 diabetes adjunct therapies from all parties present. We applaud the FDA, Lilly/BI, and the panelists for committing to high-level treatments for an area in much need of innovation. Overall, the Ad Comm members seemed very optimistic about empagliflozin’s future (if not in full understanding of what drives forward the compound commercially), and many panelists did highlight the 2.5 mg dose as the probable “sweet spot” to balance both efficacy and safety, though additional confirmatory data would be needed. Despite the FDA noting that the -0.28% decline in A1c is less than 0.3%, which is the non-inferiority margin that’s been used in many glycemic control trials for diabetes, the panelists seemed prepared to accept the modest change due to the limited number of treatment options currently available. Dr. Brendan Everett lauded the drug for being able to treat hyperglycemia without an increased risk of hypoglycemia, resulting in ~one hour increase in TIR. Dr. Everett added that for many of his friends who have children with type 1 diabetes, tracking the amount of time their children spend in therapeutic range “borderlines obsession,” and therefore, such a change would be quite significant. Dr. Kashif Munir similarly picked up on this point, noting that one hour of TIR translates to an added ~15 days per year spent in range  (he helpfully walked through the math of what an hour a day times 365 days is)– a finding that certainly feels relevant for patient quality of life and productivity. Other panelists also calling for further investment into measures beyond A1c, including TIR, microvascular and macrovascular complications, and renal complications specifically. Excitingly, the FDA’s Dr. Lisa Yanoff seemed to suggest that the FDA is becoming increasingly open to accepting TIR and other traditionally “secondary” outcomes as validated measures. We’ll be curious to see if this paradigm could even shift for any future clinical trials that Lilly/BI plan to run for the current type 1 indication application.

  • Dr. Rita Kalyani also noted that a change in hypoglycemic events would have played an important role in building up empagliflozin’s positive profile. Dr. Kalyani underscored that the major driver of improved TIR seen in EASE-3 was less time in hyperglycemia rather than less time in hypoglycemia. Clinically, an adjunct therapy that reduces time in hypoglycemia remains elusive though multiple panelists also pointed out the high value of a therapy that did not increase time in hypoglcyemia.

Open Public Hearing

Today’s Open Public Hearing was comprised of eight speakers, including Ms. Kelly Close (The diaTribe Foundation), Dr. Sanjoy Dutta (JDRF), Dr. Simeon Taylor (University of Maryland and former head of dapagliflozin’s development at BMS), Dr. Helena Rodbard (former AACE president), Dr. Sydney Wolfe, and Close Concerns’ Mr. Martin Kurian and Ms. Rhea Teng. Presentations generally focused on similar concepts: unmet needs in type 1 diabetes management, the potential of the SGLT class, the seriousness of observed DKA risk, and lack of generalizable data supporting the sustained safety and efficacy of empagliflozin 2.5 mg. See brief summaries of each speaker’s remarks below, in the order that they were presented today:

  • Dr. Charles Alexander (he was unable to appear live and speech was presented by Close Concern’s Mr. Kurian) began the OPH by drawing attention to the need for therapies in type 1 that modify CV and renal risk. He noted that after the discovery of insulin, cardiovascular and renal complications of diabetes have become increasingly prominent. Given insulin’s lack of cardio- or renal- protection, there is great need for new drugs to treat type 1 and its complications. Results from EMPA-REG and CREDENCE support SGLT-2 inhibitors’ ability to offer cardiac and renal benefits with few effects on glycemic control. However, DKA is a potentially lethal side effect of this drug class, and scientific efforts need to be placed toward finding a drug that does not increase risk of DKA. Dr. Alexander expects that SGLT-2 inhibitors will continue to be used among patients with type 1, and that empagliflozin’s approval may increase the number of patients that can benefit.

  • Remarks and slides from Kelly Close of The diaTribe Foundation focused on an analysis of the benefits of off-label SGLT inhibitor in from real world patients with type 1 (see slides). Comments were derived from T1D Exchange and dQ&A data that assessed type 1 patient need, barriers to diabetes management, and reasons for starting an SGLT-2. According to T1D Exchange data, the ADA’s A1c target of 7.0% is only met by ~26% of patients age 26 years and up very few people across any age group. Top barriers to diabetes management include unexpected highs and lows (among all age groups) and difficulty eating a healthier diet, prompting the need for non-lifestyle diabetes interventions. Thus, 71% of type 1 patients surveyed by dQ&A initiated SGLT-2 use to lower their A1c, and most of these people did report positive outcomes in A1c, wellbeing, confidence, and ease of disease management.

    • Though DKA is the main risk that stems from SGLT use, over 75% of adult type 1 patients are willing to invest 15-30 minutes in a ketone test training program and many were happy to devote considerably more time. Ms. Close believes that FDA-mandated education programs can help alleviate the burden and fear of DKA when using SGLT-2s.

    • Strikingly, even among type 1 participants already using SGLT-2s, only 44% reported learning about the signs of DKA from their health care provider, and only 46% reported having been taught how and when to test for ketones by their provider. It is clear that DKA is a very real risk from using this drug class, but education can be formally integrated into standard clinical practice to support diabetes management with SGLT use.

    • There is great potential for type 1 patients to do better and be safer. Ms. Close emphasized that type 1 patients could benefit greatly from more choice in therapeutics, and that systemic changes can be made to support patients on SGLTs. These include:

      • Patient discussions focused on safety, risk, and risk reduction with HCPs

      • Mandated, detailed, and frequent ketone training and testing

      • Broader patient focus on disease intervention and complication prevention

      • Improved speed of training

      • More study of CV and CKD risk reduction in type 1

      • More investment in the translational approach (via therapy labels) to risk avoidance

    • Because people with type 1 are not doing well enough given the medical advancements present, preventing complications is needed to address unmet needs. Reducing complications would largely impact diabetes management, and patients should have the choice to assess the risk/benefit of taking a drug with their providers. One more hour in range a day leads to a 5% change in TIR – a lofty improvement in diabetes management that should not be ignored. One panelist later remarked in the day that this 1/hr day would translate to nearly 15 more days in range for patients over the course of one year – wow!

  • JDRF’s Dr. Sanjoy Dutta centered his talk on the unmet needs in type 1 care that SGLT inhibitors can help address. He explained that less than 1/3 of adults and 1/5 of children meet recommended glycemic targets according to T1D Exchange data. The crux of his speech focused on SGLT inhibitors’ clinical meaningfulness in and beyond A1c. Though DKA risk detracts from the overall efficacy profile, he stated this could be reinforced and reduced with proper clinical management. Overall, he stressed that empagliflozin may be a good adjunct therapy to improve glucose control and reduce risk of complications.

  • Dr. Helena Rodbard shared impactful off-label use stories from her patients in clinical practice. Dr. Rodbard believes that the drug class is becoming the new standard of care given its CV and renal benefits in the landscape of few treatment options beyond insulin. Her patients that have used SGLT’s off-label report improvements in A1c, lower glycemic variability, increased general wellbeing, improved treatment satisfaction, weight loss, reduced insulin doses, and reduced risk of hypoglycemia. According to Dr. Rodbard, these anecdotes are enough to demonstrate benefits outweighing risks. She added that DKA risk can be reduced by factors like patient selection, education, and clinical management, as well as new frameworks for DKA management and patient care. She hopes empagliflozin will be approved to benefit many more patients, reduce off-label prescriptions, and increase access with insurance coverage.

  • Close Concerns’ Mr. Martin Kurian and Ms. Rhea Teng emphasized that more data regarding real world DKA risk mitigation strategies is needed before approval. Among six DKA risk reduction protocol published since 2018, all indicate important unresolved questions in (i) patient selection, (ii) SGLT inhibitor use management, and (iii) DKA management. For example, it is unclear whether higher BMI increases risk of DKA, whether a specific insulin dose reduction is required when using SGLTs, whether blood ketone monitoring should be used exclusively over cheaper urine ketone monitoring methods, or how often ketones should be monitored. While empagliflozin may benefit a broader patient population, more research needs to be done in implementing these drugs safely. These points were echoed by panelists as important next steps for empagliflozin’s development – one panelist (Dr. de Lemos) remarked that BI/Lilly should conduct a trial that randomizes patients to different risk mitigation strategies to determine their effectiveness. Although we don’t know what the feasibility of such a trial may actually be, we’re glad to see this type of thinking in terms of determining the best risk mitigation strategy moving forward.

  • Dr. Simeon Taylor firmly expressed this belief that currently available data proves that SGLT inhibitors’ risk outweigh their benefit in type 1 diabetes. Available date on the drug’s efficacy only extends to six months, with only 0.11% or 0.16% A1c lowering benefit sustained past that time point. Studies also ignore the effects of red blood cell turnover lowering A1c; specifically, as empagliflozin promotes new red blood cell creation, the decrease in A1c seen can be a non-glycemic artifact unrelated to the drug’s glucose-lowering ability. According to Dr. Taylor, it will take months to achieve a steady state in red blood cells, so 26-week A1c data likely overestimates the drug’s glucose-lowering ability. Three to five-year long studies, and their findings, are likely more reflective of empagliflozin’s real world effects. An approval should not be made before this data is available, according to Dr. Taylor.

  • Mr. Sydney Wolfe also opposed approval, citing short study duration as the main lack of evidence for empagliflozin’s safety. He stated that SGLT inhibitors' benefits decrease after the first 8-12 weeks of use, while risk may last much longer and can even compound with longer exposure. Drug efficacy peaks at 4-12 weeks and declines 40-50% as early as week two post-study. He also argued that there is currently no FDA-tested risk modification program for the drug class, so it may be irresponsible to approve the drug and then perform risk modification studies afterward. He noted that over half of the cases of DKA happened in the second year of 2.5 mg use, which is outside the currently recorded study’s time period. Overall, he argued that empagliflozin should not be approved because there is a paucity of significant evidence that benefits outweigh risks, mainly due to underpowered studies that are too short, adjudication and problems defining risk. Approval may send an incorrect signal to doctors on the safety of off-label SGLT inhibitor use.

  • The final speaker in the OPH highlighted the lack of diversity in the EASE patient population. As the large majority of the sample was white (97%), diversity is lacking, and results cannot be generalized to a larger population. With this, only 30% of those enrolled in the study were from North America, with no specific numbers given on US patients. Differences in obesity, eating, drinking, and health habits, as well as access to healthcare among different North American countries, affect risk and benefits of drug use. Short term data is especially problematic because of the expectation that these drugs will be taken for many years, and no evidence of long-term risk cannot be omitted when considering approval. She did not acknowledge that type 1 is disproportionately experienced by Caucasians.  

Summary of Committee Votes

Panelist Name



Michael Blaha



Erica Brittain

Mathematical Statistics


Kenneth Burman



Elizabeth Chrischilles



James de Lemos



Brendan Everett



Rita Kalyani



Carling Lellock

Patient Representative


Cecilia Low Wang



Anna McCollister-Slipp

Consumer Representative


Kashif Munir



Martha Nason

Mathematical Statistics


Connie Newman



Manjunath Pai



Thomas Weber



Jack Yanovski

Growth and Obesity


Commentary from Voting Members


  • Ms. Anna McCollister-Slipp voted yes as the consumer representative of today’s Ad Comm despite “significant concerns” with the application (namely the small sample size seen in EASE-3). Ms. McCollister-Slipp focused on the fact that many people with type 1 are already taking SGLT-2 inhibitors off-label and stated that she would prefer having a manufacturer track and study DKA outcomes. In addition, the issue of approving the drug so that it could be covered by insurance was brought up. Ms. McCollister-Slipp shared a personal anecdote in which her insurance company stopped covering her prescription of GLP-1 Victoza because the drug is currently solely indicated for type 2 usage. She added that often only patients with the resources and education to “fight” insurance companies can keep taking these medications off-label, so approval is needed to increase access.

  • Dr. Kashif Munir was eventually compelled in the positive direction by the unsurprising nature of the data. He acknowledged that he too would have preferred a longer trial with more data but found that in the end, the data was what he had expected. In praise of the lower 2.5 mg dose, Dr. Munir also noted that he already currently prescribes SGLT-2 inhibitors to his patients with type 1 diabetes off-label, and those patients often cut the existing tablets to obtain a smaller dose.


  • Dr. Michael Blaha voted no and noted that he struggled between balancing the efficacy seen on A1c (leading to long term benefits on microvascular complications) and short term harm with acute DKA events. He noted that low dose SGLTs are promising in his mind, and he though the A1c efficacy was clinically meaningful, but would encourage the sponsors to do another large study – a safety study that uses broad DKA adjudications as its primary endpoint. He believes such a study would teach the community a lot about type 1 diabetes, DKA, SGLT-2 inhibitors, and the concept of adding on therapies for insulin reduction in type 1 diabetes.

  • Dr. Erica Brittain voted no, but remains optimistic that this low dose of empagliflozin may prove to be useful. She noted that one six month study is insufficient to assess the tradeoffs between benefit and risk. In designing the next trial, she said that it will be important to decide what an acceptable upper bound for DKA risk actually is.

  • Dr. Kenneth Burman voted no because he was unsure of how DKA risk might translate to the real world. He noted that practical issues around studying real world DKA risk are very real, but hopes that at least a one-year long study in the real world can be conducted to assess this risk.

  • Dr. Elizabeth Chrischilles voted no because she believes that there was a lack of efficacy evidence in the sponsor’s application. She would like to see a trial conducted with more patients and over a longer period of time.

  • Dr. James de Lemos voted no and noted that it was a very clear decision for him because the safety database in the submitted trial was not adequate to ensure safety. He believes a much larger trial needs to be conducted to get a precise signal of what the DKA risk actually is. He noted that this trial might even include a randomization to different risk mitigations strategies to determine their effectiveness.

  • Dr. Brendan Everett voted no due to the limited number of patients tested on the 2.5 mg dose. Although Dr. Everett stated that he would be “willing to accept” an 0.3% drop in A1c as a clinically meaningful change because of the substantial unmet medical need in type 1 therapies, he did not find sufficient evidence of the change in the evidence presented. Future recommendations included a larger and longer study with the standard outcomes of A1c, body weight, and blood pressure, as well as added measures such as episodes of hypoglycemia and time in range.

  • Dr. Rita Kalyani found issue with the expected safety to benefit ratio in a broader population. As a clinician, Dr. Kalyani voiced concerns over less intensive monitoring in a real-world setting, especially in regard to the increased signal for ketone-related events found with empagliflozin 2.5 mg over placebo. Similar to many other committee members, Dr. Kalyani also called for a longer study to test the durability of the drug’s proposed efficacy.

  • Carling Lellock voted no as the patient representative. Interestingly, most of Ms. Lellock’s qualms with the application came from a lack of strong benefit seen with the drug rather than its risk profile. She stated that as a patient with type 1 diabetes, she is already familiar with taking risks with her therapeutic regimen. In addition, Ms. Lellock recommended a second study of longer duration and with more focus on patient-relevant secondary outcomes.

  • Dr. Cecilia Low Wang echoed many of her colleagues, stating that sufficient evidence was not presented. Of note, Dr. Low Wang pointed out that further investigation into DKA mitigation risk strategies is necessary to flesh out simple concerns such as how often a patient should test ketones. Dr. Low Wang also called for a broader definition of adjudicated DKA, reflecting a sentiment widely shared amongst the Ad Comm members. 

  • Dr. Connie Newman voted no and cited concerns over insufficient evidence to accurately characterize benefits and risks that come from taking a 2.5 mg dose. She believes there should be at least one more RCT that follows patients for at least one year to test safety. This study should also have a year-long, open-label extension to paint a more complete picture of the overall profile. Risk-mitigation strategies should also be tested in clinical trials.

  • Dr. Manjunath Pai voted no because of a lack of benefit durability.

  • Dr. Thomas Weber voted no because of the lack of data that described the risk-benefit profile. While empagliflozin has shown an ability to lower hypoglycemic events and weight, the improvement in TIR needs to be studied further according to Dr. Weber. He argued that a two year-long clinical trial is the bare minimum needed to classify risk.

  • Dr. Jack Yanovski voted no due to lack of adequate investigation of risk mitigation. He noted that more studies need to be performed to determine empagliflozin’s efficacy in different subgroups. Moreover, researchers need to determine how to assure adequate ketone testing and monitoring to minimize the risk of DKA. Future studies should also include categories like aborted DKA and illness early on in order to best manage those potential side effects.

  • Dr. Martha Nason voted no because she deemed that there was not enough evidence in the current application for a favorable benefit/risk ratio. She noted that all of the confidence intervals for ketosis and DKA events in the EASE trial go to hazard ratios of three or four, which was too high in her mind to justify approval. She would like to see longer term safety data. Like multiple other panelists, she expressed optimism about the future of empagliflozin in type 1, however.


--by Rhea Teng, Ursula Biba, Martin Kurian, and Kelly Close