Background: There is an unmet need to develop therapeutic interventions directed at the neurodegeneration that underlies progression in multiple sclerosis. High-dose, pharmaceutical-grade biotin (MD1003) might enhance neuronal and oligodendrocyte energetics, resulting in improved cell function, repair, or survival. The MS-SPI randomised, double-blind, placebo-controlled study found that MD1003 improved disability outcomes over 12 months in patients with progressive multiple sclerosis. The SPI2 study was designed to assess the safety and efficacy of MD1003 in progressive forms of multiple sclerosis in a larger, more representative patient cohort. Methods: SPI2 was a randomised, double-blind, parallel-group, placebo-controlled trial done at 90 academic and community multiple sclerosis clinics across 13 countries. Patients were aged 18–65 years, had a diagnosis of primary or secondary progressive multiple sclerosis fulfilling the revised International Panel criteria and Lublin criteria, a Kurtzke pyramidal functional subscore of at least 2 (defined as minimal disability), an expanded disability status scale (EDSS) score of 3·5–6·5, a timed 25-foot walk (TW25) of less than 40 s, evidence of clinical disability progression, and no relapses in the 2 years before enrolment. Concomitant disease-modifying therapies were allowed. Patients were randomly assigned (1:1) by an independent statistician using an interactive web response system, with stratification by study site and disease history, to receive MD1003 (oral biotin 100 mg three times daily) or placebo. Participants, investigators, and assessors were masked to treatment assignment. The primary endpoint was a composite of the proportion of participants with confirmed improvement in EDSS or TW25 at month 12, confirmed at month 15, versus baseline. The primary endpoint was assessed in the intention-to-treat analysis set, after all participants completed the month 15 visit. Safety analyses included all participants who received at least one dose of MD1003. This trial is registered with ClinicalTrials.gov (NCT02936037) and the EudraCT database (2016-000700-29). Findings: From Feb 22, 2017, to June 8, 2018, 642 participants were randomly assigned MD1003 (n=326) or placebo (n=316). The double-blind, placebo-controlled phase of the study ended when the primary endpoint for the last-entered participant was assessed on Nov 15, 2019. The mean time in the placebo-controlled phase was 20·1 months (SD 5·3; range 15–27). For the primary outcome, 39 (12%) of 326 patients in the MD1003 group compared with 29 (9%) of 316 in the placebo group improved at month 12, with confirmation at month 15 (odds ratio 1·35 [95% CI 0·81–2·26]). Treatment-emergent adverse events occurred in 277 (84%) of 331 participants in the MD1003 group and in 264 (85%) of 311 in the placebo group. 87 (26%) of 331 participants in the MD1003 group and 82 (26%) of 311 participants in the placebo group had at least one serious treatment-emergent adverse event. One (<1%) person died in the MD1003 group and there were no deaths in the placebo group. Despite use of mitigation strategies, MD1003 led to inaccurate laboratory results for tests using biotinylated antibodies. Interpretation: This study showed that MD1003 did not significantly improve disability or walking speed in patients with progressive multiple sclerosis and thus, in addition to the potential of MD1003 for deleterious health consequences from interference of laboratory tests, MD1003 cannot be recommended for treatment of progressive multiple sclerosis. Funding: MedDay Pharmaceuticals.