This article is a collaboration between 鶹ý and:
SILVER SPRING, MD -- An FDA advisory committee voted 17-2 on Tuesday that the diabetes drug liraglutide (Victoza) appeared to reduce cardiovascular risk in patients with type 2 diabetes, but expressed concern about which groups the finding applied to.
The Endocrinologic and Metabolic Drugs Advisory Committee also voted 19-0 that the drug's use in type 2 diabetes patients did not result in excess cardiovascular risk.
Liraglutide's maker, Novo Nordisk, is seeking an additional indication for the drug's use as an adjunct to standard treatment of cardiovascular risk factors to reduce the risk of major adverse cardiovascular events (cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke) in adults with type 2 diabetes and high cardiovascular risk.
Novo Nordisk's application was based entirely on a single trial known as the Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial. Although the FDA typically requires at least two trials to support a new efficacy claim, the agency review documents noted that there is precedent to rely on a single trial if it shows strong evidence of a benefit.
Its case in point was empagliflozin (Jardiance), which last year became the first diabetes drug to gain a supplemental indication for prevention of cardiovascular death in patients with type 2 diabetes and cardiovascular disease based on the results of the EMPA-REG outcomes trial alone.
The LEADER trial involved 9,340 patients randomized to either liraglutide or placebo; the trial had a minimum duration of 42 months and a targeted minimum number of 611 major adverse cardiovascular events (MACE). Patients who were randomized to liraglutide started with 0.6 mg of the drug and could be escalated weekly to up to 1.8 mg as needed. Inclusion criteria included a hemoglobin A1c of at least 7% and high cardiovascular risk, which was defined in two ways:
- Patients ages 50 to 59 who had one of several risk factors including prior myocardial infarction, more than a 50% stenosis on angiography, chronic kidney disease, or chronic heart failure NYHA class II-III, or
- Patients ages 60 or greater with microalbuminuria or macroalbuminuria, hypertension and left ventricular dysfunction, or ankle/brachial index greater than 0.9
The trial found that 13% of the patients on liraglutide experienced a MACE, compared with 14.9% of patients on placebo. In addition, 3.9% of patients in the liraglutide group died from cardiovascular disease, compared with 4.9% of the placebo group.
However, when the FDA analysts stratified the results by geographic area, they found that the U.S. patients on liraglutide -- who represented 27% of the patients on the drug -- actually showed a 3% increase in MACE compared with those from outside the U.S. (HR 1.03, 95% CI 0.84-1.25) while those outside the U.S. had a 19% decrease; similar results were found for cardiovascular deaths in the U.S. liraglutide group (HR 1.04, 95% CI .0.75-1.45), while the non-U.S. group showed a 30% decrease. These findings disturbed many committee members.
"I was very much concerned and swayed by the subgroup analysis," said Carmen Allegra, MD, chief of hematology and oncology at the University of Florida, in Gainesville, who voted "no" on the issue of whether the drug reduced cardiovascular disease risk. "I think the U.S. target population is a pretty darn important population to consider. We saw a significant interaction with outcomes versus the region by the FDA's analysis. It wasn't a small amount of patients, it was a substantial population given it was a significant percentage of the total."
Cecilia Wang, MD, associate director of endocrinology, diabetes, and metabolism at the University of Colorado, disagreed. "I have to say the primary endpoint, the results of that were very convincing especially because all of the components were also significant including all kinds of mortality. I have some concerns about the U.S. subgroup, but I think that adherence may explain that; it's hard to know."
"One of the things that this trial highlights is the heterogeneity in patients with diabetes," Wang added. "Even though we divide patients into Type 1, Type 2, 3, 4, the Type 2 patients are extremely heterogeneous, so I think we need to keep that in mind as we're thinking about drug labels."
In looking at secondary endpoints, the FDA researchers noted a small increase in pancreatic cancer among liraglutide patients, although reviewer Julie Golden, MD, noted that "it is not known if that is attributable to liraglutide, other factors, or chance. The uncertainty decreases our confidence somewhat in the findings."
The agency also studied incidents of thyroid cancer and found no excess events of overall thyroid neoplasms, C-cell hyperplasia, or medullary thyroid cancer in the liraglutide group compared with placebo, although staff member Shannon Sullivan, MD, PhD, noted that there were "small overall event rates for thyroid neoplasms (0.1%) and a short duration of follow-up (median 3.8 years), which is a relatively short time to observe thyroid events."
The data from LEADER could potentially impact the current boxed warning for risk of dose-dependent and treatment-duration-dependent thyroid C-cell tumors and precautions about pancreatitis risk on the product label.
The FDA does not have to follow the recommendations of its advisory committees, but it often does.
Thomas Walsh, 鶹ý intern, contributed to this story.