Difficulties in interpretation of complex study design
In the case of avacopan, though, the data from the key avacopan trial, Study CL010_168, known as, there were substantial uncertainties around the phase 3 study design and results, raising questions about the adequacy of this single trial to inform the benefit-risk assessment.
In the briefing document, the FDA staff noted that it had “communicated many of the concerns” about ChemoCentryx’s research earlier to the company.
“Complexities of the study design, as detailed in the briefing document, raise questions about the interpretability of the data to define a clinically meaningful benefit of avacopan and its role in the management of AAV,” the FDA staff wrote.
“We acknowledge that AAV is a rare and serious disease associated with high morbidity and increased mortality. It is also a disease with high unmet need for new therapies. However, FDA wants to ensure that new products have a defined context of use, i.e., how a product would be used, and a favorable benefit-risk assessment for patients,” the staff added.
In addition, there were differences in the assessments performed by investigators and the adjudication committee, most frequently related to the attribution of persistent vasculitis, the FDA staff noted.
Statistical analyses of the primary endpoint using investigators’ estimates “resulted in more conservative estimates of treatment effect, e.g., statistical significance for superiority would no longer be demonstrated,” the FDA staff noted. “While the prespecified analysis used the Adjudicator assessments, the assessment based on the Investigators, experienced in management of vasculitis, may better reflect real-world use.”
Imbalances in use of glucocorticoids and maintenance therapy
Also among the complications in assessing the ADVOCATE trial data were the glucocorticoids taken by patients in the study, the FDA staff said.
In the avacopan arm of the trial, 86% of patients received non–study-supplied glucocorticoids. In addition, more avacopan‐treated patients experienced adverse events and serious adverse events within the hepatobiliary system leading to discontinuation.
Subgroups given different treatments represented another challenge in interpreting ADVOCATE results for the FDA staff.
At week 26, the proportion of patients in disease remission in the avacopan group (72.3%) was noninferior to the prednisone group (70.1%), the FDA staff said in the briefing document.
But at week 52, a disparity was observed between subgroups that had received rituximab and cyclophosphamide (intravenous and oral) induction treatment. The estimated risk difference for disease remission at week 52 was 15.0% (95% CI, 2.2%-27.7%) in the subgroup receiving induction with rituximab and 3.3% (95% CI, –14.8% to 21.4%) in the cyclophosphamide plus maintenance azathioprine subgroup, the agency’s staff said.
“Based on the data, there is no evidence of clinically meaningful treatment effect in the cyclophosphamide induction subgroup,” the FDA staff wrote. “Further, the treatment comparison in the complementary rituximab induction subgroup may not be considered meaningful because these patients did not receive maintenance therapy, i.e., due to undertreating of patients, the effect observed in the rituximab subgroup may not represent a clinically meaningful treatment effect, compared to standard of care.”
Rachel L. Glaser, MD, clinical team leader in FDA’s division of rheumatology and transplant medicine, reiterated these concerns to the advisory committee at the May 6 meeting.
“Throughout the development program, FDA advised the applicant that a noninferiority comparison would not be sufficient to show that avacopan can replaced glucocorticoids as it would be difficult to establish whether avacopan is effective or whether an effect was due to the rituximab or cyclophosphamide administered to both treatment arms,” she said.
In its briefing for the meeting, ChemoCentryx noted the limits of treatments now available for AAV. It also emphasized the toll of the condition, ranging from skin manifestations to glomerulonephritis to life-threatening pulmonary hemorrhage. If untreated, 80% of patients with GPA or MPA die within 2 years of disease onset, ChemoCentryx said in its briefing materials for the meeting.
The side effects of glucocorticoids were well known to the FDA panelists and the ChemoCentryx presenters. Witnesses at an open public hearing told their own stories of depression, anxiety, and irritability caused by these medicines.
During the ChemoCentryx presentation, a presenter for the company,, of the University of Pennsylvania, Philadelphia, said avacopan would provide patients with AAV with an alternative allowing them “to go on a much lower glucocorticoids regimen.”
A similar view was presented in a February 2021 editorial in the New England Journal of Medicine, titled “” Written by , of the Mayo Clinic, Rochester, Minn., the opinion article called the ADVOCATE trial “a milestone in the treatment of ANCA-associated vasculitis; complement inhibition with avacopan has glucocorticoid-sparing effects and results in superior disease control.”
Dr. Warrington reported no conflicts in connection with his editorial nor payments from ChemoCentryx. He did report grants from other firms such as Eli Lilly.
, of Vanderbilt University, Nashville, Tenn., was among the more skeptical members of the FDA panel. She was among the “nays” in all three voting questions put to the panel. Still, she said there were signs of “clinically meaningful benefit” in the data presented, but noted that the nonstudy use of glucocorticoids made it difficult to interpret the ADVOCATE results.
Dr. Lewis noted that the FDA usually requires two studies for a drug approval, particularly with a compound not yet cleared for any use. While ANCA-associated vasculitis is rare, it would be possible to recruit patients for another trial of avacopan, adding to the results reported already for avacopan from ADVOCATE, she said.
“Were there to be another study, this would certainly be a supportive study and maybe qualify as two studies,” she said.