This article is a collaboration between 鶹ý and:
SILVER SPRING, Md. -- An FDA advisory committee voted 16-0 Wednesday, with one abstention, in favor of recommending approval of semaglutide, a GLP-1 receptor agonist for once-weekly injection in type 2 diabetes along with lifestyle changes.
"The benefit with regard to hemoglobin A1c (HbA1c) was impressive, and there was no signal for cardiovascular harm," said Paul Palevsky, MD, chief of the renal section at the VA Pittsburgh Healthcare System and a temporary member of the FDA's , explaining his Yes vote. "The adverse events, other than [an increased risk of retinopathy-related events] were as would be expected for this class [of drugs] and not a concern. The retinopathy [findings were] a modest concern and outweighed by the benefit seen."
"I really wanted to vote yes," said Yves Rosenberg, MD, chief of the atherothrombosis and coronary artery disease branch at the National Heart, Lung, and Blood Institute, in Bethesda, Maryland, explaining why he abstained. "I think there are a number of reasons this drug should be approved, but rather than voting an uncomfortable yes, I really wanted to make a point that there should be follow-up studies" on both the retinopathy issue and to further explore the drug's cardiovascular benefit.
The FDA presented data from the six trials that sponsor Novo Nordisk had conducted for the drug, all under the name SUSTAIN; they included one cardiovascular outcomes trial known as SUSTAIN-6.
Overall, the trial data from the five non-cardiac trials showed positive results, with a mean change from baseline in HbA1c of -1.3% to -1.7% compared with placebo over a period of 30 weeks. Patients also lost weight, with a mean weight loss of 2.2 kg (about 5 lbs) to 4.7 kg compared with placebo. The drug also outperformed competitors sitagliptin (Januvia), insulin glargine (Lantus), and once-weekly exenatide (Bydureon).
The drug's safety profile was similar to that of other GLP-1 agonists, according to FDA staff. The most common side effects were gastrointestinal, including nausea, vomiting, diarrhea, and constipation; on the plus side, there was also a low inherent risk of hypoglycemia.
Also on the plus side for drugmaker Novo Nordisk, the 2-year SUSTAIN-6 trial demonstrated a 26% reduction in risk of major cardiovascular events with semaglutide, including a 35% decrease in rates of stroke and 18% lower rates of myocardial infarction (although, curiously, there was no reduction in cardiovascular death risk).
However, the FDA also highlighted an increase in the rate of complications related to diabetic retinopathy for study patients on semaglutide -- a hazard ratio of 1.76 (95% CI 1.11-2.78) with about 3.2% of patients experiencing such effects compared with 1.8% of the study's placebo group over the 2-year study.
Ya-Hui Hsueh, PhD, of the FDA's biometrics division, noted that the FDA's consulting ophthalmologist expressed several concerns about the way the retinopathy data were collected, including:
- No standardized approach to fundus evaluations -- dilation was not required, and the evaluation could be performed by the investigator or by a local ophthalmologist or optometrist
- No formal grading of retinopathy findings
- No uniform agreement on what retinopathy characteristics dictated a need for treatment; the independent committee evaluating the data required that treatment be administered, not just "needed," in order to count as a complication
One question raised during the trial was whether the increase in retinopathy was due to the "early worsening effect" of large drops in blood glucose, as seen in other major diabetes trials such as the Diabetes Control and Complications trial. That trial showed a cumulative increase in retinopathy progression during the first 2 years with intensive glucose-lowering treatment as compared with conventional treatment, but after that point, those undergoing intensive therapy had less progression of retinopathy than their conventionally treated counterparts.
"The sponsor argued that the total causal effect of semaglutide can be broken into two components -- a direct effect not associated with a change in HbA1c and an indirect effect mediated by [glucose decline]," said Hsueh. If the indirect affect is adjusted for, according to Novo Nordisk, the increased risk drops from 76% to 22%.
However, there are limitations to the sponsor's analysis, she continued. First, "the estimate of the total effect is unchanged ... unless the indirect effect can be removed or reduced." Also, the model might be incorrect, and there could be a "multiplicity" issue -- the change in HbA1c at week 16 was chosen post-hoc as an outcome, and it is unclear whether other variables and other visit times (in weeks) were considered as possible mediators, she said.
Frederick Ferris III, MD, director of epidemiology and clinical applications at the National Eye Institute, in Bethesda, Maryland, said he was "appalled" by the way some of the retinopathy data were collected, including the fact that exams could be performed undilated and by an optometrist. "I'm nervous about retina surgeons and [their] dilated exam being reproducible," he said. "This? I don't know what to make of it."
Panel members were asked about whether a retinopathy trial should be required of the sponsor; most said a study should be encouraged but not required. Many members felt that the drug's label should recommend a visit to an ophthalmologist for some patients, similar to the labeling for insulin.
Another concern raised during the meeting was the studies' lack of important subgroups. "The trial was particularly overwhelmingly white," said Suzanne Robotti, president and founder of the MedShadow Foundation, a New York City organization that facilitates discussion between patients and providers about prescription drugs.
"Cardiovascular disease is the number one killer of women, and a [large] killer of black people; it has a higher incidence rate in blacks and Hispanics and they're not well represented," said Robotti, the committee’s consumer representative. "The ability to do analysis on this concerns me. I'm going to feel pretty bad if we don't go on record saying there should be more trial participants, and [the trial] should reflect the people who need the drug."
The FDA is not obligated to follow recommendations of its advisory committees, though it often does.