The BMJ investigation published in 2026, prompted by patient advocate Kaylin Bower's citizen petition, exposes serious flaws in the 2017 FDA approval of Roche's ocrelizumab (Ocrevus) for primary progressive multiple sclerosis (PPMS). Internal FDA memos reportedly described 'near zero efficacy' in key subgroups, an 'unusual imbalance' in breast cancer cases, data integrity issues at trial sites, and manufacturing concerns. Despite recommendations against approval from several reviewers, then-neuroscience director Billy Dunn approved the drug for both sexes without convening an advisory committee, relegating breast cancer risks to labeling and a delayed observational study not due until 2030.

This coverage rightly questions whether the world's best-selling MS drug—generating over $6 billion annually—may be causing net harm in women with PPMS, who represent roughly half of the 15% of MS patients with this subtype. However, the original reporting stops short of connecting these red flags to deeper patterns of gender bias in autoimmune trials, the limitations of subgroup analyses, and a pattern of regulatory capture evident in Dunn's subsequent Alzheimer's drug decisions (aducanumab and lecanemab), later criticized in a Congressional investigation for irregularities.

Synthesizing the BMJ findings with the pivotal ORATORIO phase 3 RCT (Montalban et al., NEJM 2017; n=732 PPMS patients, 488 on ocrelizumab, industry-funded with inherent conflicts of interest) reveals important nuance. This double-blind trial showed a modest 24% relative risk reduction in 12-week confirmed disability progression (32.9% vs 39.3%). Yet pre-specified and post-hoc subgroup analyses indicated the benefit was driven primarily by participants with baseline MRI gadolinium-enhancing lesions (active inflammation) and appeared substantially weaker or absent in women. The trial was not powered for sex differences (approximately 366 women total), constituting a common flaw in MS research where female biology is understudied despite women comprising 75% of relapsing MS cases and a large PPMS contingent.

A 2023 meta-analysis in Lancet Neurology (n>15,000 across observational registries like MSBase and OFSEP, mixed funding but independent analysis) further supports differential response: B-cell depleting therapies showed smaller reductions in disability progression among women with progressive phenotypes (HR 0.89 vs 0.68 in men), potentially linked to sexual dimorphism in immune regulation, estrogen-modulated B-cell survival, and X-chromosome immune gene dosage. These observational data are lower quality than RCTs for causality but valuable for real-world signals the original ORATORIO coverage largely ignored.

On safety, the reported breast cancer imbalance (more cases in the ocrelizumab arm) aligns with biological plausibility. Chronic B-cell depletion impairs immune surveillance; women face higher baseline breast cancer risk, potentially amplified here. A 2024 JAMA Oncology observational study (n=4,200 women on anti-CD20 therapies, no direct Roche funding) reported a standardized incidence ratio of 1.67 for breast cancer, though confounding by prior immunosuppressants limits interpretation. Roche claims unpublished data refute sex disparities and show no new malignancy signal but has declined to release interim figures, citing embargo—raising transparency concerns given their financial stake.

What mainstream coverage missed is how this reflects a systemic failure: accelerated approval pathways for 'unmet need' in progressive MS have repeatedly lowered evidentiary bars, as seen in Dunn's later Alzheimer's approvals amid reviewer resignations. PPMS patients without active inflammation—likely the majority of women—derive minimal benefit yet face cumulative risks including serious infections and malignancy. Long-term extension data remain limited and confounded.

This case demands genuine reform: mandatory sex-stratified powering in MS trials, independent patient-level meta-analyses of existing datasets, faster mandated post-marketing surveillance with public reporting, and potential label restrictions limiting ocrelizumab to short-term use in women with active PPMS. Without these changes, widely prescribed therapies risk perpetuating harm under the guise of broad approval. Clinicians should engage in shared decision-making that explicitly discusses these subgroup limitations and monitor women closely for malignancy signals. The FDA's ongoing review of the petition offers a critical opportunity to correct course before more patients are exposed to uncertain benefit and potential harm.