Members of an FDA advisory panel were sharply divided as to whether avacopan, an oral small molecule inhibitor of the C5a receptor, was sufficiently safe and effective for treating antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis.
As to whether the efficacy data supported approval of the drug, the FDA’s Arthritis Advisory Committee was split 9 to 9; on whether the safety data were sufficient, the vote was 10 in favor and 8 against; and on whether the benefit-risk profile supported approval, again the vote was 10 in favor and 8 against.
Concerns about the efficacy of avacopan in ANCA-associated vasculitis focused primarily on the study design and reliance on a single pivotal study. The FDA usually requires two studies for an approval decision, but in some cases a single study can suffice, although the study results have to be persuasive and robust. Panelists voting “no” indicated that the efficacy results did not meet that standard.
The application by sponsor ChemoCentryx was based on a single phase III trial known as ADVOCATE, which enrolled 331 patients, randomizing them to avacopan, 30 mg twice daily, or a 20-week prednisone taper. Participants also received background rituximab or cyclophosphamide, and glucocorticoids as needed at the investigators’ discretion.
In the study, avacopan was found at 26 weeks to be noninferior to the prednisone taper, with 72.3% of avacopan-treated patients achieving remission compared with 70.1% of those given prednisone. This was statistically significant for noninferiority (P<0.001) though not for superiority (P=0.24). By week 52, sustained remission was observed in 65.7% of the avacopan group and in 54.9% of the prednisone-taper group, which was significant for both noninferiority (P<0.001) and superiority (P=0.007).
The lack of superiority seen at week 26 was the subject of much discussion, as was the use of a noninferiority study design and the determination of the noninferiority margin.
A further concern for the efficacy question was the fact that prednisone use outside of the study-defined 20-week taper was permitted in both groups, and was actually used in the vast majority of patients. The use of steroids beyond the study protocol made it difficult to interpret the effect of avacopan, but not allowing the administration of prednisone in the event of disease flare would not be ethical or practical, the panel noted.
A further efficacy question related to the background use of rituximab. At the time of the study initiation, rituximab had not been approved for repeated use, but since that time repeat dosing has been approved and has become more widely used. In the study, therefore, patients on background rituximab in effect were given that agent for induction but were given no maintenance therapy. In contrast, those whose background induction therapy was cyclophosphamide were given maintenance therapy with azathioprine, again making interpretation of the data challenging, the panel noted. Some panelists suggested that an additional study would be helpful, with patients on rituximab being given both induction and maintenance therapy.
Members of the committee who voted “yes” on the efficacy question suggested that this drug would provide a treatment option for patients at high risk for prednisone-associated toxicity.
In voting on whether the safety data supported approval, panelists in favor pointed out that the study population was small (166 patients were exposed to avacopan for 52 weeks), but that postmarketing surveillance and close monitoring could address potential long-term concerns, and that the safety concerns with prednisone are substantial, with some events having irreversible effects.
Those who voted “no” on whether the safety data were adequate voiced concerns about adverse events relating to hepatotoxicity and angioedema, and also long-term concerns in minority groups.
On the final question as to whether the benefit-risk profile was adequate to support approval, those in favor of approval suggested that the FDA could recommend judicious use and provide additional guidance as to which patients might be suitable candidates. Those who voted against approval noted that there were “too many unknowns” with the drug, and that additional confirmatory trials should be considered.
Although the FDA is not required to follow its advisory committees’ recommendations, it typically does. However, split votes are generally taken as no recommendation one way or the other.