Much has already been written in recent months about avacopan / TAVNEOS, the oral complement C5a receptor antagonist. It is in the news because of the way the parent company, Chemocentryx, its Adjudication Board and the CRO handled the unblinded data and ‘re-adjudicated’ the results to achieve the desired outcome, an all-important significant P-value for FDA approval.[1],[2],[3]
This drug was developed in ANCA vasculitis for a reason. The market for C5 inhibitors is getting crowded; Alexion has dominated the field with eculizumab and ravulizumab for a long time. Both drugs are biologicals and injectables, of proven benefit in PNH and a few other rare diseases like aHUS. An oral drug like avacopan would be great to have in these indications, however, Chemocentryx decided to pursue a niche indication, ANCA vasculitis, an area of unmet medical need which held promise for blockbuster potential.
FDA participated in the design of the single pivotal Phase 3 ANCA vasculitis trial which the Chemocentryx team started in 2017. The design was complicated; below the FDA slide showing its overall outline.
The schematic points out an obvious imbalance between treatment arms. Even so it is not immediately obvious how flawed the trial design actually is. Let us analyze its key features:
The largest group (approx. 2/3 of patients) were those started on rituximab / RTX. They received the following treatment regimens:
Comparator Group A (“prednisone group”) N=107
ie, tapered prednisone (over 20-wks) then no Rx for the remaining 32 wks
plus a single dose of RTX
Active Group B (“avacopan group”) N=107
ie, avacopan (52-wks) for the entire year
plus a single dose of RTX
Everything else being equal, this means that the actual comparison was between a 20-wk prednisone taper and a 52-wk avacopan arm. In our view, this is a lopsided comparison, with the odds favoring avacopan.
There was another smaller cohort of patients (approx. 1/3 of patients) that were started on CYC –> AZA and received either of the following regimens:
Comparator group C (“prednisone group”) N=57
ie, tapered prednisone (over 20 wks)
plus CYC –> AZA (52-wks)
Active Group D (“avacopan group”) N=59
ie, avacopan (52-wks)
plus CYC –> AZA (52-wks)
Here we are dealing with background therapies lasting a full 52-wk BUT if we subtract the contribution of CYC –> AZA (as it is part of both arms) the comparison becomes one of tapered prednisone 20-wks vs 52-wks of avacopan. Therefore, anyway you look at it, this is also a lopsided comparison comparing short-term tapered prednisone (20-wks) with long-term avacopan (52-wks).[4]
It sure looks like the design of this Phase 3 trial called ADVOCATE was meant to make avacopan look good.
Well, despite its uneven design, a statistically superior treatment benefit was only seen in the (predominant) subgroup of patients that had no maintenance therapy at all. In the FDA’s own assessment, there was a “lack of robustness of the superiority assessment”. We wholeheartedly agree.
With such a slim margin of superiority, lack of confirmation in the ADVOCATE investigator analysis, and lack of a confirmatory 2nd trial, it is not surprising that FDA reviewers in-house were highly critical of the submission.[6] When discussed with outside experts at the now defunct FDA Arthritis Advisory Committee, the vote on efficacy was split 9:9.
So far, we have not even talked about DILI and VBDS yet, toxicities that became even more transparent since the drug was approved.
There are dark clouds collecting over avacopan.[9] There will be more news to share next month after the planned FDA Hearing with Amgen.
Let us conclude this blog on a positive note.
Avacopan may not be the best complement C5aR inhibitor for AAV but there is nothing wrong with its underlying rationale. Inhibition of C5aR1 is a logical next step for a new generation of C5a inhibitors.[7],[8] Existing C5 inhibitors are less specific and interfere with the membrane-attack complex (MAC) which can result in overwhelming meningococcal sepsis. The new classes of C5a and C5aR1 inhibitors are designed to leave the MAC intact. And that is a good thing. A more specific C5a agent, even if SC, would be welcome. An oral drug could be a game-changer.
ABBREVIATIONS
AAV ANCA-associated vasculitis
aHUS acquired hemorrhagic uremic syndrome
ANCA anti-neutrophil cytoplasmic antibody
AZA azathioprine
C5, C5a complement factors C5, C5a
C5aR, C5aR1 complement C5a / C5a1 receptor
CI confidence interval
CYC cyclophosphamide
DILI drug-induced liver injury
MAC membrane-attack complex (C5-C9)
PNH paroxysmal nocturnal hemoglobinuria
SC subcutaneous
VBDS vanishing bile duct syndrome
REFERENCES
[1] https://www.healio.com/news/rheumatology/20260515/fda-seeks-to-withdraw-avacopan-approval-over-alleged-data-manipulation-false-statements
[2] Jayne D. Avacopan for the Treatment of ANCA-Associated Vasculitis. NEJM 2021;384:599
[3] Rubin E. Retraction: Jayne DRW et al. Avacopan for the Treatment of ANCA-Associated Vasculitis. NEJM 2026. DOI: 10.1056/NEJMe2608684
[4] CI ranges from FDA review of the Avacopan NDA, based on ‘re-adjudicated’ data set
[5] FDA Arthritis Advisory Committee Meeting, May 6, 2021
[6] FDA CDER Multidisciplinary Review & Evaluation, Avacopan NDA 214487. 2021
[7] Horiuchi T. Complement-targeted therapy: development of C5- and C5a-targeted inhibition. Inflamm Regen 36:11, 2016
[8] Tesar V. Complement Inhibition in ANCA-Associated Vasculitis. Front. Immunol. 13:888816, 2022
[9] Editorial. The end for avacopan. Lancet Rheumatology Published online July 10, 2026 https://doi.org/10.1016/S2665-9913(26)00225-0
