Background: Inclusion body myositis is the most frequent myositis in patients older than 50 years. Classical immunosuppressants are ineffective in treating inclusion body myositis, and to date there are no recommendations for pharmacological approaches to treatment. When used after organ transplantation, sirolimus can block the proliferation of effector T cells, while preserving T regulatory cells, and induce autophagy, all of which are processes that are impaired in inclusion body myositis. In this pilot study, we aimed to test the efficacy of sirolimus in patients with inclusion body myositis. Methods: This randomised, double-blind, placebo-controlled, proof-of-concept, phase 2b trial was done at a single hospital in Paris, France. The study included men and women (aged 45–80 years) who had a defined diagnosis of inclusion body myositis according to established criteria. Eligible participants were randomly assigned (1:1) to receive once-daily oral sirolimus 2 mg or placebo. Centralised balanced block randomisation (blocks of four) was computer generated without stratification. The study comprised a 15-day screening period (days −15 to 0) and a 52-week treatment period (day 0 to month 12). The primary endpoint was the relative percentage change from baseline to month 12 in maximal voluntary isometric knee extension strength. Secondary endpoints included the following assessments at months 6 and 12: 6-min walking distance, isometric muscle strength for hand grip (finger flexors), knee flexion and elbow flexion and extension, forced vital capacity, muscle replacement with fat measured by quantitative nuclear MRI, Inclusion Body Myositis Weakness Composite Index (IBMWCI), Inclusion Body Myositis Functional Rating Scale (IBMFRS), Health Assessment Questionnaire without Disability Index (HAQ-DI), and analyses of T-cell subpopulations by mass cytometry. The primary analysis was done on the intention-to-treat population. The trial is registered at ClinicalTrials.gov, NCT02481453. Findings: Between July 15, 2015, and May 13, 2016, we screened 285 patients, 44 of whom were randomly allocated to sirolimus (22 patients) or placebo (22 patients). We observed no difference in the primary outcome of relative percentage change from baseline to month 12 of the maximal voluntary isometric knee extension strength (median difference 3·78, 95% CI −10·61 to 17·31; p=0·85). For secondary outcomes, differences between the groups were not significant for changes in strength of other muscle groups (grip, elbow flexion and extension, or knee flexion), IBMWCI, IBMFRS, and lower limb muscle fat fraction. However, we observed significant differences in favour of sirolimus between the study groups for HAQ-DI, forced vital capacity, thigh fat fraction, and 6-min walking distance. Ten (45%) of 22 patients in the sirolimus group had a serious adverse event compared with six (27%) of 22 patients in the placebo group. Four (18%) patients in the sirolimus group stopped their treatment because of adverse events (severe mouth ulcers, aseptic pneumonia, renal insufficiency, and peripheral lower limb oedema), which resolved after treatment discontinuation. Canker sores were the most frequent side-effect and were mainly mild or moderate in ten patients. Interpretation: We found no evidence for efficacy of sirolimus for treating inclusion body myositis based on maximal voluntary isometric knee extension strength and other muscle strength measures, and the side-effects of treatment were substantial for some patients. However, we believe there was enough evidence of benefit in certain secondary outcomes to pursue a multicentre phase 3 trial to further assess the safety and efficacy of sirolimus.