American Diabetes Association 76th Scientific Sessions

June 10-14, 2016; New Orleans, LA; Full LEADER Results – Draft

Executive Highlights

We’re back with our full coverage of the results from the LEADER trial for Novo Nordisk’s Victoza (liraglutide) presented to a packed auditorium at ADA on Monday. Read on below for our top themes related to the results, followed by detailed coverage of the results and discussant presentations. For those looking for more, see our coverage of the topline results, our initial report on the full results, and the full paper published in the NEJM.  

  

Themes

  • While LEADER was arguably one of the more likely diabetes drug CVOTs to demonstrate cardioprotection, the positive results were nonetheless surprisingly positive (from a historical sense) and monumental – cardioprotection and renal protection announced simultaneously. There has been speculation about the potential for cardioprotection with GLP-1 agonists for some time due to the class’ positive effects on weight, blood pressure, and lipids and possible direct effects on the heart and vasculature. Novo Nordisk had expressed cautious optimism in several quarterly updates (most recently in 3Q15) that LEADER would be more likely than most CVOTs to reveal any benefit that existed due to its greater individual patient exposure compared to other trials (mandated minimum exposure of 3.5 years per patient and total exposure of over 30,000 patient-years). That said, there were also many reasons to expect a neutral outcome: the trial was only powered to show non-inferiority, it enrolled a high-risk patient population, any effect of GLP-1 agonists on CVD was expected to be subtle and gradual, and the only other GLP-1 agonist CVOT to report (ELIXA for Sanofi’s lixisenatide) had demonstrated neutral results. In that context, these results were certainly a pleasant surprise and a groundbreaking moment for the type 2 diabetes field.
  • The early consensus seems to be that the benefits in LEADER were most likely related to atherosclerosis. This was the main hypothesized mechanism of CV benefit for GLP-1 agonists before the results were released, as the class positively affects a number of endpoints related to atherosclerosis (glucose, blood pressure, weight) and may have direct effects on reducing arterial plaque. Speakers at the press briefing for LEADER and in the main results presentation agreed that the results appeared consistent with an atherosclerotic mechanism: the effects took some time to appear and increased over the course of the trial, and they were relatively consistent across multiple atherosclerotic endpoints. This is in marked contrast with the EMPA-REG OUTCOME results for Lilly/BI’s Jardiance (empagliflozin), which demonstrated an immediate effect on heart failure and CV mortality and little to no signal of an effect on MI or stroke. While we expect plenty of further speculation and investigation of the exact mechanism of benefit in LEADER (Dr. Laurie Baggio offered a few specific hypotheses in her discussant presentation), the field appears to be closer to a consensus hypothesis in this case compared to EMPA-REG OUTCOME.
  • We have many remaining questions about the clinical implications of the results that could take years to resolve. Dr. John Buse (University of North Carolina, Chapel Hill, NC) emphasized in his concluding presentation that the results should only be applied to the specific patient population enrolled in LEADER – those with longstanding type 2 diabetes and high CV risk. We expect that any updates to the Victoza label and to type 2 diabetes treatment guidelines will only be applied to this population in the absence of what would be a very expensive trial in a lower-risk population. However, we also imagine that many clinicians may incorporate the results into their risk/benefit calculus for Victoza even for lower-risk patients. We are also very curious what the implications will be for Saxenda (liraglutide 3.0 mg for obesity), which is unlikely to undergo a CVOT of its own. We would love to see Novo Nordisk conduct a CVOT for the more potent GLP-1 agonist semaglutide in patients with obesity and without type 2 diabetes, though we understand such a trial would also be very expensive. The question of a GLP-1 agonist class effect on CV outcomes is sure to be a hot topic for the next several years as CVOTs for other agents report results. In the near term, we expect the results to bolster the class as a whole but to disproportionately benefit Victoza. It will be very interesting to see how clinicians weigh the demonstrated CV benefits with Victoza against advantages like once-weekly dosing with Lilly’s Trulicity (dulaglutide) or guaranteed adherence with Intarcia’s ITCA 650 (implantable exenatide mini-pump).
  • The results from LEADER and EMPA-REG OUTCOME will undoubtedly raise the already high bar for new type 2 diabetes drugs. We see little room in the current landscape for new drugs that offer nothing more than incremental glucose-lowering benefits, or those (other than insulin) that carry a risk of weight gain or hypoglycemia. We can even imagine a future in which new type 2 diabetes drugs will have to demonstrate a positive effect on cardiovascular and/or renal outcomes in order to be successful now that two agents in different classes have already done so. We are also very curious to see how the LEADER and EMPA-REG OUTCOME results are judged relative to each other. Will GLP-1 agonists and SGLT-2 inhibitors (or the specific agents in each class that have demonstrated CV benefit) both be considered preferred second-line treatments for all patients at high CV risk? Will the recommendations be different depending on the specific type of CV risk (i.e., empagliflozin for heart failure and liraglutide for atherosclerosis)? The prospect of combination therapy with two drugs that reduce CV risk by different mechanisms is also an intriguing one, though one attendee during a Tuesday session on EMPA-REG OUTCOME raised the possibility that the two mechanisms could actually work against each other if liraglutide reduces ketone body production that is contributing to the benefit with empagliflozin.
  • The significant 22% improvement in renal outcomes was a very positive surprise from the LEADER results. The implications of this strong renal benefit are substantial, perhaps even prompting consideration of a dedicated chronic kidney disease trial for liraglutide. While improvements in risk factors like glucose and blood pressure could explain some of the benefit, newer evidence suggests that there might be a direct effect at play, involving GLP-1 receptor-stimulated relaxation of muscles around glomeruli (the functional filtering units of the kidney). This finding, along with the positive renal outcomes with empagliflozin presented on the last day of the conference, is truly profound given the currently enormous unmet need for new therapies to treat chronic kidney disease. We wonder what the implications will be for ongoing and future trials of drugs being developed specifically for diabetic nephropathy – will those agents be required to demonstrate a benefit in addition to that provided by liraglutide and/or empagliflozin?
  • Large-scale changes to the FDA’s 2008 CV Guidance appear increasingly unlikely now that two CVOTs have reported positive results. The consensus opinion has certainly changed from a year ago, when the streak of four completely neutral trials (SAVOR, EXAMINE, TECOS, and ELIXA) had raised questions about whether the value of these trials was worth the massive investment. Now that two trials have revealed important benefits that might otherwise have remained unknown, we have heard several speakers (including Dr. Darren McGuire and Dr. Steve Nissen at AACE) argue strongly that the field is getting its money’s worth from these studies. Even some previous opponents of the FDA Guidance have changed their opinions over the past year – for example, Dr. Silvio Inzucchi (Yale University, New Haven, CT) stated at EASD that he was prepared to completely revise his previous assessment of the requirements in light of the EMPA-REG OUTCOME results. At the same time, the FDA officially eliminated the Risk Evaluation and Mitigation Strategy for rosiglitazone last December, effectively declaring the main rationale for the 2008 Guidance to be invalid. We also continue to question whether an across-the-board CVOT requirement with the same guidelines for all diabetes drugs is the most effective tool to assess the risks and benefits of new products. We believe that a more nuanced approach, in which drugs with a signal for CV risk would be required to conduct a safety trial and those with potential for benefit would be required or incentivized to conduct a superiority trial, would offer the most value to the field.

Detailed Discussion and Commentary

Symposium: Results of the Liraglutide Effect and Action in Diabetes – Evaluation of Cardiovascular Outcome Results (LEADER) Trial

Introduction, Study Rationale, and Design

John Buse, MD, PhD (University of North Carolina, Chapel Hill, NC)

Dr. John Buse (UNC, Chapel Hill, NC) kicked off the session with an introduction to the background and design of the LEADER trial. He acknowledged the enormous scope of the study and the people who made it possible with an impressive slide reviewing the key numbers for the trial: LEADER enrolled 9,340 patients in 32 countries across 410 sites over five years. In total, an astounding 11,000 people worked on the trial to collect 27.6 million data points over 13,500+ monitoring visits, resulting in information on almost 30,000 patient-years of exposure to liraglutide. LEADER was a double-blind, randomized, placebo-controlled, time- and event-driven trial: participants were treated for at least 3.5 years and no more than 5 years and the trial was to accumulate at least 611 events included in the primary endpoint (CV death, non-fatal MI, and non-fatal stroke). Key secondary outcomes included an expanded composite endpoint of CV death, non-fatal MI, non-fatal stroke, coronary revascularization, unstable angina pectoris requiring hospitalization, and hospitalization for heart failure. All-cause death and each individual component of the expanded CV composite were considered as secondary endpoints as well. Participants had type 2 diabetes with an A1c ≥7% and fell into one of two high cardiovascular risk categories: (i) at least 50 years old with established CVD or chronic renal failure or (ii) at least 60 years old with risk factors for CVD. Participants could be diabetes drug-naïve, on oral diabetes medications, or on basal or premix insulin. Patients on GLP-1 agonists, DPP-4 inhibitors, pramlintide, or rapid-acting insulin were excluded from the trial, as were those with type 1 diabetes or a familial or personal history of multiple endocrine neoplasia type 2 (MEN-2) or medullary thyroid cancer (MTC). Participants were treated according to standard of care guidelines with target A1c ≤7%, blood pressure 130/80 mmHg, and LDL cholesterol <100 mg/dl (<70 mg/dl in patients with previous CV events). Statins were recommended for all patients and aspirin or clopidogrel were recommended for patients with prior CV events. Events were adjudicated by an external committee made up of subcommittees each focused on cardiovascular, microvascular, pancreatitis-related, or neoplasm-related events.

Study Population

Neil Poulter, FMedSci (Imperial College London, London, UK)

Dr. Neil Poulter reviewed the characteristics of the LEADER study population, emphasizing that this was truly a global effort (32 countries, 30% of participants from North America, 35% from Europe, 8% from Asia, and 27% from the rest of the world). He also highlighted the trial’s very impressive retention rate: 97% of participants in both groups completed the trial and there were only 29 dropouts whose vital status was unknown at the end of the study. The study population was 64-65% male, with an average age of 64 and an average diabetes duration of 13 years, which Dr. Poulter noted is significantly longer than in most diabetes trials. Baseline A1c was fairly high at 8.7%, likely due to the lack of an upper A1c limit in the entry criteria. Baseline BMI was 32.5 kg/m2, baseline weight was 92 kg (~203 lbs), baseline blood pressure was fairly well controlled at 140/77 mmHg, and 18% of participants had heart failure at baseline. The liraglutide and placebo groups were well matched for all baseline characteristics. Over 80% of participants fell into the higher-risk group of patients ≥50 with existing CVD or CKD, and many had more than one qualifying condition. Consistent with this high level of risk, over 90% of patients were on antihypertensive therapy, 40% on diuretics, 75% on lipid-lowering medications (which Dr. Poulter described as suboptimal), and just under 70% on platelet inhibitors. Metformin was by far the most commonly used diabetes medication (used by approximately 75% of participants), followed by sulfonylureas at 50% and insulin at 45%. Dr. Poulter also highlighted the impressive treatment exposure in the trial, with a median exposure time of just over 3.5 years and patients spending an average of 83%-84% of the time on study drug.

Clinical and Metabolic Outcomes

Bernard Zinman, MD (Lunenfeld-Tanenbaum Research Institute, Toronto, Ontario, Canada)

Dr. Bernard Zinman (University of Toronto, Canada) presented the key clinical and metabolic outcomes from the study. At three years post-randomization (a pre-specified time point), mean A1c in the liraglutide-treated group was 0.4% lower than in the placebo-treated group (p<0.001), despite the instructions to treat to a target A1c of 7% in both groups.  The gap in A1c diminished throughout the trial, largely due to an intensification of other agents in the placebo group. The proportion of patients on metformin, sulfonylureas, alpha-glucosidase inhibitors, TZDs, glinides, and insulin was fairly evenly matched between the two groups at the start of the trial, but more participants in the placebo group initiated treatment with these classes (particularly insulin and sulfonylureas) throughout the trial. Other clinical and metabolic outcomes of interest include 2.3 kg weight loss with liraglutide compared to placebo at three years (p<0.001), a 1.2 mmHg decrease in systolic blood pressure (p<0.001), a 0.6 mmHg increase in diastolic blood pressure (p=0.004), a 3 beats/min increase in heart rate (p<0.001), a 1.6 mg/dl decrease in LDL cholesterol (p=0.02), and a non-significant 0.3 mg/dl increase in HDL cholesterol (p=0.07). Dr. Zinman generally characterized these outcomes as expected effects associated with liraglutide and suggested that the increase in diastolic blood pressure and decrease in LDL cholesterol were not clinically significant. During the trial, a greater proportion of participants in the placebo group initiated new cardiovascular medications such as antihypertensive therapies, diuretics, lipid-lowering drugs, platelet aggregation inhibitors, or other anti-thrombotic medications. The LEADER study also investigated health-related quality of life measures, as measured by the EQ-5D index score (which assesses mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and the EQ-5D VAS score (a visual scale of health state from 0 to 100 representing worst and best health, respectively). Based on the index score, both liraglutide and placebo reduced participants’ quality of life, but liraglutide offered a 0.018 index score improvement over placebo (p=0.034). Liraglutide also offered a 1.3 point advantage on the VAS score (p=0.03). We were extremely pleased to see quality of life measures included in the study and hope that this kind of patient-centered metric could help encourage greater coverage from payers.

Cardiovascular Outcomes

Steven Marso, MD (UT Southwestern, Dallas, TX)

Interrupted multiple times by applause, Dr. Steven Marso presented the primary results from the LEADER trial. He emphasized the consistency of the effect for all the outcomes measured, noting that every single endpoint had a point estimate <1 that fell within a narrow range from 0.6-0.98. Taken together, these results fit the pattern that most in the field had predicted any positive CVOTs for diabetes drugs would show: risk reduction for multiple CV outcomes that appeared gradually and increased over time. This is in marked contrast with the EMPA-REG OUTCOME results for Lilly/BI’s Jardiance (empagliflozin), which demonstrated an immediate benefit for CV death and hospitalization for heart failure and little to no signal of benefit for other endpoints. As several speakers noted, the LEADER results appear more likely to be mediated through a reduction in atherosclerosis, whether that is due to indirect effects (reductions in A1c, weight, blood pressure, etc.) or direct effects on the heart or blood vessels.

  • Primary outcome: The hazard ratio for the primary outcome of three-point MACE (non-fatal MI, non-fatal stroke, and CV death) was 0.87 (95% CI: 0.78-0.97; p<0.001 for non-inferiority; p=0.01 for superiority), translating to a significant 13% risk reduction. Dr. Marso noted that the event rate (3.7 events/100 patient-years) was two-fold higher than projected, perhaps due to the high percentage of participants with existing CVD. The magnitude of the effect was consistent and robust in a number of sensitivity analyses, with the point estimate ranging from 0.82-0.87 and the upper bound of the 95% confidence interval consistently <1.
  • Individual components: While all three components of the primary endpoint contributed to the benefit, the most robust effect was a 22% risk reduction for CV death (HR = 0.78; 95% CI: 0.66-0.93; p=0.007; 497 events). The Kaplan-Meier curves for the two groups began to separate 12-18 months into the trial and continued to diverge until the end of treatment. The results for non-fatal MI and non-fatal stroke trended in the right direction but were not statistically significant. The hazard ratio for non-fatal MI was 0.88 (95% CI: 0.75-1.03; p=0.11, 598 events) and the hazard ratio for non-fatal stroke was 0.89 (95% CI: 0.71-1.11; p=0.30; 336 events). Dr. Marso pointed out that the shape of the curves for non-fatal MI followed the same pattern as those for CV death and for the primary endpoint – initial separation between 12 and 18 months and continued divergence over the course of the trial – perhaps suggesting that the risk reduction might have become statistically significant over a longer time period.
  • Subgroup analyses: There was no significant heterogeneity of the effect on the primary endpoint in subgroups divided by sex, age, region, race, ethnicity, BMI, A1c, diabetes duration, heart failure status, or anti-diabetic therapy. The two subgroups with a p-value for interaction <0.05 (with the caveat that the analyses did not control for multiplicity of testing) were CVD status and renal function. The point estimate for the group with established CVD was 0.83, while the point estimate for the group with CV risk factors was 1.2. Dr. Marso attributed this to the fact that the vast majority of trial participants fell into the first category, which had a much higher event rate; the lower-risk group most likely did not accumulate enough events to show a benefit. By renal function, the point estimate for patients with an eGFR <60 ml/min/1.73 m2 was 0.69 while the point estimate for patients with an eGFR >60 was 0.94.
  • Key secondary outcome: Results for an expanded MACE endpoint (including coronary revascularization and hospitalization for heart failure or unstable angina in addition to the primary endpoint) demonstrated a significant 12% risk reduction (HR = 0.88; 95% CI: 0.81-0.96; p=0.005).
  • All-cause mortality: The liraglutide group experienced a significant 15% risk reduction in all-cause mortality (HR = 0.85; 95% CI: 0.74-0.97; p=0.02), accounted for by the significant reduction in CV death and comparable rates of death from non-CV causes. Dr. Marso noted that these curves also separated after 12-18 months (much later than in EMPA-REG OUTCOME) and continued to diverge throughout the trial.
  • Heart failure: Results for the much-watched endpoint of hospitalization for heart failure also trended in a positive direction but did not reach statistical significance (HR = 0.87; 95% CI: 0.73-1.05; p=0.14).

Microvascular Outcomes

Johannes Mann, MD (Friedrich Alexander University of Erlangen, Erlangen, Germany)

While microvascular outcomes usually don’t receive as much attention as cardiovascular results at CVOT sessions, the topic turned out to be one of the real positives of the entire LEADER session. We were very surprised to see a 16% statistically improvement in time to first microvascular event  (encompassing renal and ophthalmic adverse outcomes; 95% CI: 0.73-0.97, p=0.02), driven entirely by a 22% statistically significant improvement in renal outcomes (95% CI: 0.67-0.92, p=0.003). There was no statistically significant improvement in eye outcomes – the hazard ratio was 1.15 (95% CI: 0.87-1.52, p=0.33). The Kaplan-Meier curves for renal outcomes separated one year into the study and continued to diverge throughout the study. The renal benefit was driven primarily by a difference in the diagnosis of persistent macroalbuminuria (HR=0.74, 95% CI: 0.60-0.91).

  • The implications of this strong renal benefit are substantial, perhaps even prompting consideration of a dedicated chronic kidney disease trial for liraglutide. While improvements in risk factors like glucose and blood pressure could explain some of the benefit, newer evidence suggests that there might be a direct effect at play, involving GLP-1 receptor-stimulated relaxation of muscles around glomeruli (the functional filtering units of the kidney).
  • The patient population in LEADER had a fairly high level of microvascular disease at baseline: around 2.5% of patients had an eGFR below 30 ml/min/1.73m2, ~20% were between 30 and 59 ml/min/1.73m2, and ~41% were between ml/min/1.73m2.
  • The specific outcomes assessed in the category of microvascular disease were as follows:
    • Renal: (i) New onset of persistent macroalbuminuria – as mentioned above, this was the renal sub-outcome that drove the improvement, (ii) persistent doubling of serum creatinine, (iii) need for continuous renal replacement therapy, and (iv) death due to renal disease.
    • Eye: (i) Need for retinal photocoagulation or treatment with intravitreal agents, (ii) vitreous hemorrhage, and (iii) diabetes-related blindness.

Safety

Michael Nauck, MD, PhD (Diabeteszentrum Bad Lauterberg, Germany)

Dr. Michael Nauck (St. Josef Hospital, Bochum, Germany) presented safety and adverse event data from LEADER. Most notably, treatment with liraglutide was associated with a 20% risk reduction for confirmed hypoglycemia with a blood glucose ≤56 mg/dl (HR=0.80, CI: 0.74-0.88, p<0.001) and a 31% risk reduction for severe hypoglycemia requiring assistance (HR=0.69, CI: 0.51-0.93, p=0.016). We expect that this reduction in hypoglycemia risk is largely or at least somewhat attributable to the smaller number of participants in the liraglutide group initiating insulin or sulfonylurea therapy during the trial. The results demonstrate the real-world, long-term benefits of using GLP-1 agonists as an alternative to therapies that increase hypoglycemia risk, and we imagine they could further strengthen the case for GLP-1 agonists (or even GLP-1 agonist/basal insulin combinations) as a first injectable option ahead of insulin. Interestingly, Dr. Anne Peters suggested at The diaTribe Foundation/TCOYD forum later that evening that the overall cardiovascular benefit in LEADER may be at least partially attributable to the lower use of sulfonylureas/insulin and lower hypoglycemia rates in the active treatment group rather than to specific benefits associated with liraglutide – we wonder whether this opinion will be widely shared and whether it will become a significant part of the debate over the results.

  • Total, serious, and severe adverse event rates were generally similar between the liraglutide and placebo groups. As expected, a higher proportion of patients experienced nausea, vomiting, and diarrhea leading to discontinuation in the liraglutide group (p<0.001 for each). Abdominal pain and discomfort were also significant adverse events resulting in treatment discontinuation (p=0.03 for pain, p=0.002 for discomfort). Decreased appetite was also significantly more prevalent in the liraglutide-treated group (p=0.01). Liraglutide also produced a statistically significant increase in acute gallstone disease (145 events vs. 90 in placebo, p<0.001) and acute cholecystitis (36 vs. 21 in placebo, p=0.046). Injection site reactions were higher in the liraglutide-treated group as well (32 vs. 12 in placebo, p=0.002).
  • There was a signal toward increased risk of any, malignant, and benign neoplasms by 12%, 6% and 16%, respectively, but the increase was not significant for each. Interestingly, by type, liraglutide produced a statistically significantly reduced risk of prostate cancer (HR=0.54, CI:0.34-0.88) and leukemia (HR=0.36, CI:0.13-0.99), though Dr. Nauck acknowledged that the number of events was very small. Pancreatic carcinomas were numerically higher in the liraglutide group (13 vs. 5), but the difference was not statistically significant. There were no cases of medullary thyroid carcinoma in the liraglutide-treated group and one in the placebo group. Acute and chronic pancreatitis rates in the liraglutide group were not significantly different from the placebo group.
  • However, both all-cause serious and severe adverse event rates were lower in the liraglutide group than in the placebo group – in fact, serious adverse events overall were significantly lower in the liraglutide group (p=0.01). We expect at least part of this was driven by the reduced risk of hypoglycemia in the liraglutide-treated group compared to the placebo group.

Conclusion

John Buse, MD, PhD (University of North Carolina, Chapel Hill, NC)

Dr. John Buse contextualized the LEADER results by discussing potential mechanisms of benefit, clinical implications, and comparisons to results from other CVOTs for diabetes drugs. He stressed from the outset that directly comparing results from two separate trials is an “entirely inappropriate activity” due to differences in study population, definition and adjudication of events, secular trends in diabetes management, and the countries where the studies were conducted. With that in mind, Dr. Buse shared that he was “nevertheless going to engage in that activity,” suggesting that the divergent results from LEADER and ELIXA (CVOT for Sanofi’s lixisenatide) could have been due either to differences in study design or intrinsic differences between the molecules. He contrasted the benefits seen in LEADER with those in EMPA-REG OUTCOME, suggesting that the LEADER benefit was more likely mediated through a reduction in atherosclerosis while the EMPA-REG results may have been due to an osmotic effect or an impact on fuel energetics. Dr. Buse offered minimal commentary on the clinical implications of the results but stressed that the conclusions need to be limited to high-risk patients. Like previous speakers, he closed by highlighting the trial’s impressive retention and follow-up, suggesting that the trial conduct should inspire strong confidence in the results.

  • Dr. Buse believes that the “substantial differences” between the results from LEADER and ELIXA could be due either to differences in trial design or intrinsic drug-specific differences. He noted that ELIXA had a different primary endpoint (four-point MACE) and recruited an even higher-risk patient population (patients with a recent acute coronary syndrome event) than LEADER. The nature of the population appeared to have an impact on the pattern of events (the majority of events occurred at the beginning of the trial) and may have impacted the results as well. However, Dr. Buse also stressed that lixisenatide and liraglutide are very different molecules both in terms of structure and pharmacokinetics. Lixisenatide is an analog of the Gila monster-derived exenatide while liraglutide is an analog of human GLP-1, and liraglutide has a much longer half-life and provides 24-hour coverage. We imagine that the PK differences in particular could very likely have contributed to the discrepancy on CV outcomes; future CVOTs for other GLP-1 agonists should provide more insight on this question. 
  • Dr. Buse contrasted the gradual, consistent benefit in LEADER with the more immediate and unexpected benefit in EMPA-REG OUTCOME. He argued that while the hazard ratio for the primary endpoint was “remarkably similar” (13% vs. 14%) in the two trials, the overall picture was very different. EMPA-REG OUTCOME showed a rapid separation between groups followed by maintenance of the difference, whereas the curves in LEADER began to separate within the first 18 months and continued to diverge for the rest of the trial. Similarly, while both studies showed the most profound benefit on CV death, LEADER showed much more consistent results across other outcomes. Dr. Buse suggested that the pattern and timing of responses in LEADER is more consistent with an effect on atherosclerosis, while the main hypotheses at this point for the EMPA-REG OUTCOME results relate to osmotic diuresis and fuel energetics.  

Discussant

Laurie Baggio, PhD (Lunenfeld-Tanenbaum Research Institute, Toronto, Ontario, Canada)

Dr. Laurie Baggio’s discussant presentation explored possible mechanistic explanations for the effects observed in LEADER. Regarding the surprisingly large benefit seen on nephropathy, Dr. Baggio suggested that improvements in risk factors like glucose, oxidative stress, weight, blood pressure, and inflammation played an indirect role, but there also exists the tantalizing possibility that liraglutide could be acting directly to reduce filtration pressures in the glomerulus (the functional filtration unit of the kidney), preventing them from burning out. On the observed increase in gallstones, Dr. Baggio speculated on potential mechanisms (weight loss in known to cause an increase in gallstones), but concluded by highlighting the dearth of data in this area, characterizing gallstones as a major area for future investigation on GLP-1 agonists. In her conclusion, Dr. Baggio stated that it is a “huge advantage” for patients to know that liraglutide reduces mortality, cardiovascular death, MI, and kidney disease, though they must balance those benefits against increases in nausea, vomiting, diarrhea, and gallstone disease.

  • Improvements in cardiovascular risk factors likely explain the lion’s share of the MACE reduction seen with liraglutide, but Dr. Baggio did not rule out the possibility of direct GLP-1-receptor-mediated action directly on the heart. Evidence is mixed on whether GLP-1 receptors are present in cardiac ventricles. The evidence is clearer on the improvements in risk factors, including body weight, blood vessel health, diuresis, natriuresis, reducing inflammation, glucose control, postprandial lipids, coagulation, and reduced hypoglycemia. However some of these risk factors that were measured, such as blood pressure, did not change enough to explain much of the overall cardiovascular benefit seen in the trial.
  • Dr. Baggio appeared unconvinced that GLP-1 agonists are the best option for patients with advanced heart failure. She pointed that existing outcomes studies, including LEADER, do not suggest a benefit of GLP-1 agonists on heart failure. In her words, there is no reason to avoid using a GLP-1 agonists in a patient with early-stage heart failure, but that prescribers “may want to think twice” in patients with more advanced heart failure. In any case, SGLT-2 inhibitors (or at least empagliflozin from EMPA-REG OUTCOME data) appear to do better with reducing the incidence of heart failure.
  • Dr. Baggio briefly discussed the increase in heart rate seen with liraglutide, but suggested that LEADER does not provide sufficient information to assess the issue one way or the other.
  • The increases in amylase and lipase are hard to assess, given that their levels are usually variable and often elevated in type 2 diabetes patients who are otherwise asymptomatic. In fact, 23% of LEADER patients had elevated levels of the enzymes at baseline. Amylase and lipase levels are examined because they are diagnostic for pancreatitis, but overall the pancreatitis data were reassuring, as is experimental evidence that GLP-1 agonists do not promote acinar cell inflammation or proliferation.
  • LEADER was not powerful or long enough to provide a definitive answer on liraglutide and cancer. Dr. Baggio sees the GLP-1/cancer question as a valid one, in light of evidence that GLP-1/GLP-1 agonists can reduce apoptosis, increase intestinal mass, and increase beta cell mass. She suggested that it will be helpful to know when in LEADER the cancer cases occurred.

 

-- by Helen Gao, Emily Regier, Manu Venkat, and Kelly Close