FDA panel focused on safety issues in rejecting roxadustat from FibroGen
Click Here to Manage Email Alerts
I was disappointed that the recent decision by the Cardiovascular and Renal Drugs Advisory Committee of the FDA will further delay the availability of an orally administered agent for anemia treatment in patients with chronic kidney disease.
Citing safety concerns, committee members voted at their meeting on July 15 (see story on page 20) against roxadustat (FibroGen Inc.), an oral hypoxia-inducible factor prolyl hydroxylase (HIF-PI) inhibitor, for patients with non-dialysis dependent chronic kidney disease (NDD-CKD) and for dialysis-dependent CKD (DD-CKD).
There is an unmet need for such an agent as was so passionately articulated by the patients and physicians who spoke during the public comment portion of the meeting. This is particularly true in the NDD-CKD population for whom travel to clinics to receive erythropoiesis-stimulating agents (ESAs) and IV iron therapy can be particularly burdensome.
Nonetheless, the Cardiovascular and Renal Drugs Advisory Committee (CRDAC) appropriately takes its duty to protect this population seriously and carefully scrutinized the safety data for roxadustat in the NDD-CKD and DD-CKD clinical trials. These trials are robust: the pooled safety data in the NDD-CKD studies exposed 4,270 patients (vs. placebo) and the pooled safety data in the DD-CKD studies exposed 3,880 patients (vs. epoetin) to roxadustat.
The data consistently demonstrated efficacy of roxadustat in raising hemoglobin (Hb) levels to target range. These results were not contested by the CRDAC. The presentations by the FDA and discussion by CRDAC members at the meeting focused exclusively on the safety data, a significant portion of which has not been presented at national and international meetings nor published in peer-reviewed journals.
Cardiovascular events
The primary safety outcome in the NDD-CKD and DD-CKD studies is major adverse cardiovascular events (MACE). In the NDD-CKD studies (vs. placebo), the primary prespecified outcome (agreed upon with the FDA) uses an ascertainment window of on-study exposure, meaning all MACE events are counted until the subject leaves the study (due to drop-out or end of follow-up).
This is also called intention to treat (ITT) and may capture events long after exposure to the study drug ends that may or may not be due to the drug itself.
Nonetheless, an ITT analysis is considered the gold standard for most clinical trials. Roxadustat met MACE safety non-inferiority for patients with NDD-CKD (hazard ratio 1.10, CI 0.96-1.27) using this analysis. As a sensitivity analysis for MACE safety in NDD-CKD, the FDA presented data for on-treatment (OT) plus 7 days. Roxadustat did not meet safety non-inferiority (HR 1.38, CI 1.11-1.70) using this analysis.
The patients receiving placebo had a higher drop-out rate than those receiving roxadustat because the former group became more symptomatic because their anemia was not being treated. As a result, the exposure time to roxadustat was greater than that to placebo and, because MACE is an event-driven analysis, this unfairly disadvantaged roxadustat in the OT plus 7 days analysis.
Vascular thrombosis
CRDAC members expressed concern about the increased frequency of non-MACE adverse events in the NDD-CKD population (vs. placebo), particularly vascular thrombosis, seizures and serious infections. The sponsors proposed a risk mitigation plan recommending lower doses of roxadustat with a post-approval clinical trial to demonstrate this results in a decrease in thrombotic events; the sponsor also proposed label warnings to use the drug with caution in patients with previous seizure history or active infection.
Nonetheless, the CRDAC voted against approval of roxadustat for the NDD-CKD patient indication; there was some discussion that this decision should be revisited should a future clinical trial demonstrate a lower incidence of adverse events using the risk mitigation strategy of lower roxadustat doses and a target Hb level of 10 g/dL to 11 g/dL rather than the 10.5 g/dL to 12 g/dL range used in phase 3 studies.
In the DD studies (vs. epoetin), the primary prespecified MACE safety outcome (agreed upon with the FDA) is the OT plus 7 days analysis, for which roxadustat met non-inferiority (HR 1.02, CI 0.88-1.20). The sensitivity analysis presented by the FDA for the DD-CKD pooled safety data is the on-study (ITT), which did not demonstrate non-inferiority (HR 1.14, CI 1.00-1.30).
The cardiologists on the panel expressed the opinion that because the ITT analysis is the gold-standard it should be taken seriously even though it was not the primary prespecified analysis. However, the major safety concern in the DD-CKD population was not MACE but vascular access thrombosis (VAT) which can compromise the “lifeline” for hemodialysis patients. Data were presented by the FDA demonstrating increased incidence of VAT (relative risk 1.5) and deep venous thrombosis (DVT, relative risk 3.9) among patients with DD-CKD treated with roxadustat vs. epoetin.
The sponsors presented data demonstrating the risk of VAT and DVT correlated with rate of Hb rise and roxadustat dose and proposed a risk mitigation strategy of both lower roxadustat dose (initial for non-ESA treated patients and conversion for ESA-treated patients) and lower-target Hb level that would decrease the rate of Hb rise and significantly lower thrombosis risk based on modeled data. The CRDAC remained sufficiently concerned regarding the VAT data that they voted against approval of roxadustat in the DD-CKD population.
Future reviews
The discussion by the CRDAC regarding roxadustat clearly sets a precedent for the review of other hypoxia inducible factor (HIF) stabilizer drugs in the U.S. approval pipeline, namely vadadustat (Akebia Pharmaceuticals Inc.) and daprodustat (GlaxoSmithKline). Top-line MACE safety data for vadadustat in global phase 3 trials failed to demonstrate non-inferiority to ESA in the NDD population. Top-line MACE safety data for daprodustat in global clinical trials announced recently showed non-inferiority to ESA in NDD-CKD and DD-CKD populations, but it is clear from the roxadustat proceedings the FDA will examine all safety data, some of which may not be presented at professional meetings or in publications.
Many comments by the FDA and CRDAC members reflect a perception that ESAs are dangerous drugs and that any alternative therapy for CKD anemia should not be less safe than ESAs. If we knew in 1989 what we know now about the safety of ESAs, there is the possibility that ESAs might not have been approved.
Transfusions sensitize patients to human antigens which limits the pool of potential transplant donors and increases time on deceased-donor transplant waitlists. The mean Hb at dialysis initiation in the United States is 9.3 g/dL, only 15% have received ESAs, 40% have received a blood transfusion. The prospect of an oral therapy to address this problem is exciting, and patients and providers should have a choice whether they wish to accept after full disclosure what may be a safety signal.
This brings up the possibility of a risk evaluation and mitigation strategy to require such discussion with demonstration that patients understand and accept the risks. It is likely that such a requirement for roxadustat and potentially other HIF stabilizers in the NDD-CKD space where there is the greatest unmet need will pose insurmountable barriers to adoption and leave the need unmet.
Nephrologists and our patients eagerly await the approval of a relatively safe (acknowledging that no drug is completely safe) oral option for treatment of CKD anemia.
For more information:
- Jay B. Wish, MD, is professor of clinical medicine at Indiana University School of Medicine and chief medical officer for outpatient dialysis at Indiana University Health. He is also vice-chair of the Editorial Advisory Board for Nephrology News & Issues.
Collapse
Read more about
Click Here to Manage Email Alerts