BACKGROUND: Progressive supranuclear palsy is a neurodegenerative disorder associated with tau protein aggregation. Tilavonemab (ABBV-8E12) is a monoclonal antibody that binds to the N-terminus of human tau. We assessed the safety and efficacy of tilavonemab for the treatment of progressive supranuclear palsy. METHODS: We did a phase 2, multicentre, randomised, placebo-controlled, double-blind study at 66 hospitals and clinics in Australia, Canada, France, Germany, Italy, Japan, Spain, and the USA. Participants (aged ≥40 years) diagnosed with possible or probable progressive supranuclear palsy who were symptomatic for less than 5 years, had a reliable study partner, and were able to walk five steps with minimal assistance, were randomly assigned (1:1:1) by interactive response technology to tilavonemab 2000 mg, tilavonemab 4000 mg, or matching placebo administered intravenously on days 1, 15, and 29, then every 28 days through to the end of the 52-week treatment period. Randomisation was done by the randomisation specialist of the study sponsor, who did not otherwise participate in the study. The sponsor, investigators, and participants were unaware of treatment allocations. The primary endpoint was the change from baseline to week 52 in the Progressive Supranuclear Palsy Rating Scale (PSPRS) total score in the intention-to-treat population. Adverse events were monitored in participants who received at least one dose of study drug. Prespecified interim futility criteria were based on a model-based effect size of 0 or lower when 60 participants had completed the 52-week treatment period and 0·12 or lower when 120 participants had completed the 52-week treatment period. This study is registered at ClinicalTrials.gov, number NCT02985879. FINDINGS: Between Dec 12, 2016, and Dec 31, 2018, 466 participants were screened, 378 were randomised. The study was terminated on July 3, 2019, after prespecified futility criteria were met at the second interim analysis. A total of 377 participants received at least one dose of study drug and were included in the efficacy and safety analyses (2000 mg, n=126; 4000 mg, n=125; placebo, n=126). Least squares mean change from baseline to week 52 in PSPRS was similar in all groups (between-group difference vs placebo: 2000 mg, 0·0 [95% CI -2·6 to 2·6], effect size 0·000, p>0·99; 4000 mg, 1·0 [-1·6 to 3·6], -0·105, p=0·46). Most participants reported at least one adverse event (2000 mg, 111 [88%]; 4000 mg, 111 [89%]; placebo, 108 [86%]). Fall was the most common adverse event (2000 mg, 42 [33%]; 4000 mg, 54 [43%]; placebo, 49 [39%]). Proportions of patients with serious adverse events were similar among groups (2000 mg, 29 [23%]; 4000 mg, 34 [27%]; placebo, 33 [26%]). Fall was the most common treatment-emergent serious adverse event (2000 mg, five [4%]; 4000 mg, six [5%]; placebo, six [5%]). 26 deaths occurred during the study (2000 mg, nine [7%]; 4000 mg, nine [7%]; placebo, eight [6%]) but none was drug related. INTERPRETATION: A similar safety profile was seen in all treatment groups. No beneficial treatment effects were recorded. Although this study did not provide evidence of efficacy in progressive supranuclear palsy, the findings provide potentially useful information for future investigations of passive immunisation using tau antibodies for progressive supranuclear palsy. FUNDING: AbbVie Inc.