Background Olaparib, a poly(ADP-ribose) polymerase (PARP) inhibitor, has previously shown efficacy in a phase 2 study when given in capsule formulation to all-comer patients with platinum-sensitive, relapsed high-grade serous ovarian cancer. We aimed to confirm these findings in patients with a BRCA1 or BRCA2 (BRCA1/ 2) mutation using a tablet formulation of olaparib. Methods This international, multicentre, double-blind, randomised, placebo-controlled, phase 3 trial evaluated olaparib tablet maintenance treatment in platinum-sensitive, relapsed ovarian cancer patients with a BRCA1/ 2 mutation who had received at least two lines of previous chemotherapy. Eligible patients were aged 18 years or older with an Eastern Cooperative Oncology Group performance status at baseline of 0-1 and histologically confirmed, relapsed, high-grade serous ovarian cancer or high-grade endometrioid cancer, including primary peritoneal or fallopian tube cancer. Patients were randomly assigned 2: 1 to olaparib (300 mg in two 150 mg tablets, twice daily) or matching placebo tablets using an interactive voice and web response system. Randomisation was stratified by response to previous platinum chemotherapy (complete vs partial) and length of platinum-free interval (> 6-12 months vs > 12 months) and treatment assignment was masked for patients, those giving the interventions, data collectors, and data analysers. The primary endpoint was investigator-assessed progression-free survival and we report the primary analysis from this ongoing study. The efficacy analyses were done on the intention-to-treat population; safety analyses included patients who received at least one dose of study treatment. This trial is registered with ClinicalTrials. gov, number NCT01874353, and is ongoing and no longer recruiting patients. Findings Between Sept 3, 2013, and Nov 21, 2014, we enrolled 295 eligible patients who were randomly assigned to receive olaparib (n=196) or placebo (n=99). One patient in the olaparib group was randomised in error and did not receive study treatment. Investigator-assessed median progression-free survival was significantly longer with olaparib (19.1 months [95% CI 16.3-25.7]) than with placebo (5.5 months [5.2-5.8]; hazard ratio [HR] 0.30 [95% CI 0.22-0.41], p< 0.0001). The most common adverse events of grade 3 or worse severity were anaemia (38 [19%] of 195 patients in the olaparib group vs two [2%] of 99 patients in the placebo group), fatigue or asthenia (eight [4%] vs two [2%]), and neutropenia (ten [5%] vs four [4%]). Serious adverse events were experienced by 35 (18%) patients in the olaparib group and eight (8%) patients in the placebo group. The most common in the olaparib group were anaemia (seven [4%] patients), abdominal pain (three [2%] patients), and intestinal obstruction (three [2%] patients). The most common in the placebo group were constipation (two [2%] patients) and intestinal obstruction (two [2%] patients). One (1%) patient in the olaparib group had a treatment-related adverse event (acute myeloid leukaemia) with an outcome of death. Interpretation Olaparib tablet maintenance treatment provided a significant progression-free survival improvement with no detrimental effect on quality of life in patients with platinum-sensitive, relapsed ovarian cancer and a BRCA1/2 mutation. Apart from anaemia, toxicities with olaparib were low grade and manageable.