San Diego-based Capricor Therapeutics (Nasdaq: CAPR) faces a critical regulatory test on Tuesday when the US FDA's Cellular, Tissue, and Gene Therapies Advisory Committee convenes to vote on whether deramiocel — an allogeneic cardiosphere-derived cell therapy — has demonstrated substantial evidence of effectiveness for cardiomyopathy in Duchenne muscular dystrophy (DMD). The agency's CBER briefing document for BLA 125842, released ahead of the July 29 meeting, presents a comprehensively negative efficacy assessment and concludes that the benefit-risk profile appears unfavorable, signaling a high bar for the committee to clear.
The meeting marks Capricor's second attempt at approval. The original BLA, based on Phase II data from Studies HOPE-2 and HOPE-2-OLE, received a Complete Response Letter in July 2025 due to lack of substantial evidence of effectiveness. The resubmission rests on the pivotal Phase III HOPE-3 study, but the FDA briefing document indicates that study also failed to meet its pre-specified endpoints — and that subsequent statistical modifications by the applicant have further undermined the data's credibility.
HOPE-3 was a 12-month, randomized, double-blind, placebo-controlled trial enrolling 106 males with DMD, predominantly non-ambulatory, with a mean left ventricular ejection fraction (LVEF) of 57% at baseline. The study was designed to detect a 1.5-point difference in the change from baseline to Month 12 in PUL 2.0 total score, a measure of upper limb skeletal muscle function, with LVEF change as the key secondary endpoint. Under the pre-specified analysis in statistical analysis plan (SAP) version 1.1, the least-squares mean difference in PUL 2.0 total score was 0.66 points (95% CI: −0.45, 1.77; p=0.24) — not statistically significant. On LVEF, the between-group difference was −0.041 percentage points (95% CI: −2.58%, 2.49%; p=0.97), indicating no treatment effect.
The applicant subsequently generated at least two additional SAP versions after study completion. The final SAP (version 3.0), dated November 24, 2025, was not submitted to FDA prior to BLA resubmission and was not agreed upon by the agency. Key modifications included changing the primary endpoint from absolute change to percent change in PUL 2.0 total score, and converting the LVEF analysis from a mixed-model repeated measures approach to an analysis of covariance on ranked change — a metric FDA describes as clinically uninterpretable and not standard in cardiac disease assessment. Under the applicant's SAP 3.0 analyses, the PUL 2.0 result reached nominal significance (p=0.029) and the LVEF ranked change difference yielded p=0.041. FDA concluded that the nominally significant findings produced under the applicant's revised SAP were not robust, with statistical significance disappearing under alternative missing-data assumptions affecting as few as two placebo patients.
FDA identified an additional structural concern: a distinctive adverse event profile between treatment arms created conditions for potential functional unblinding. Hypersensitivity reactions occurred in 41.5% of deramiocel-treated subjects versus 15.4% of placebo subjects. Eleven placebo subjects who had not experienced hypersensitivity reactions during the blinded phase subsequently developed them upon receiving open-label deramiocel in the extension period, providing a mechanism by which blinded-phase treatment assignment could be retrospectively inferred. The agency concluded that SAP modifications made after data collection under these conditions carry a high risk of being data-driven, regardless of formal blinding status.