The FDA has posted briefing documents for Thursday’s meeting of the Endocrinologic and Metabolic Drugs Advisory Committee to reconsider the new drug application for lorcaserin (Lorqess, Arena). The same panel recommended against approval of the drug in September 2010, citing weak efficacy and safety concerns.
The FDA reviewers do not appear to have substantially altered their view of lorcaserin. There are no major new safety concerns, but the limited efficacy of the drug gives little reason for panel members to take a risk.
The same safety signals– in particular, lingering concerns about heart valve problems– remain. It is unclear whether new, previously unpublicized data submitted by Arena on the effect of lorcaserin on the serotonin receptor (the source of previous heart valve problems with phen/fen) will move the dial on this issue. One important note of caution here: although no statistically significant increase in risk of valve disease was found by the FDA (relative risk 1.16, 95% CI 0.81-1.67), the reviewer noted that the “upper bound exceeds the 1.5 upper bound requested by FDA to rule out an excess risk of VHD.”
The cancer issue seems less important this time, although its earlier significance may have been overstated.
Although the draft discussion questions for the committee mostly focus on the safety issues, I suspect that the efficacy data will actually play a more important role in determining how the committee votes. The efficacy data has not substantially changed since the previous meeting. A new trial, BLOOM-DM, is consistent with the previous trials, demonstrating a modest effect that meets some but not all of the FDA predefined criteria for efficacy.
Here’s how the FDA’s statistical reviewer summarized the efficacy findings:
All three studies had similar estimates of the placebo-adjusted effect of lorcaserin 10 mg bid at 52 weeks (TABLE 1). The consistency of the efficacy results across Studies 010, 009 and 011 supports the collective evidence for the efficacy of lorcaserin 10 mg bid. However, the efficacy endpoints, while statistically significant, do not fully meet the benchmarks for clinical significance that are described in the Agency’s Weight Management Guidance (2007):
• For the continuous endpoint, the guidance states that the difference in mean weight loss between the active product and placebo-treated groups should be at least 5% and the difference should be statistically significant. For all three studies, the placebo-adjusted percentage change from baseline at week 52 was statistically significant. However, in each of the three studies, the placebo-adjusted effect of lorcaserin was statistically significantly less than 5%.
• For the categorical endpoint, at least 5% of weight loss at week 52, the guidance states that the observed percentage of responders should be at least 35% and at least double the percentage in the placebo-treated group. These criteria are met in all three studies, when the last observation carried forward (LOCF) method was used to impute the 52week results from subjects who discontinued early. However, these results are somewhat sensitive to the imputation method. When early dropouts are classified as non-responders, Studies 009 and 011 meet the criteria for the categorical endpoint but Study 010 does not.
In my opinion, the 5% responder endpoint is a key endpoint because of the substantial percentage of early withdrawals in all three studies. Because of the relationship between dropping out and being less successful at weight loss in these studies, I believe it is reasonable to classify dropouts as non-responders. This approach may be a reasonable way to extend the study results to the intended target population.