Background The poly(ADP-ribose) polymerase inhibitor olaparib has shown antitumour activity in patients with platinum-sensitive, recurrent, high-grade serous ovarian cancer with or without BRCA1 or BRCA2 mutations. The aim of this study was to assess the efficacy and tolerability of olaparib in combination with chemotherapy, followed by olaparib maintenance monotherapy, versus chemotherapy alone in patients with platinum-sensitive, recurrent, high-grade serous ovarian cancer. Methods In this randomised, open-label, phase 2 study, adult patients with platinum-sensitive, recurrent, high-grade serous ovarian cancer who had received up to three previous courses of platinum-based chemotherapy and who were progression free for at least 6 months before randomisation received either olaparib (200 mg capsules twice daily, administered orally on days 1-10 of each 21-day cycle) plus paclitaxel (175 mg/m(2), administered intravenously on day 1) and carboplatin (area under the curve [AUC] 4 mg/mL per min, according to the Calvert formula, administered intravenously on day 1), then olaparib monotherapy (400 mg capsules twice daily, given continuously) until progression (the olaparib plus chemotherapy group), or paclitaxel (175 mg/m(2) on day 1) and carboplatin (AUC 6 mg/mL per min on day 1) then no further treatment (the chemotherapy alone group). Randomisation was done by an interactive voice response system, stratified by number of previous platinum-containing regimens received and time to disease progression after the previous platinum regimen. The primary endpoint was progression-free survival according to Response Evaluation Criteria in Solid Tumors version 1.1, analysed by intention to treat. Prespecified exploratory analyses included efficacy by BRCA mutation status, assessed retrospectively. This study is registered with ClinicalTrials.gov, number NCT01081951, and has been completed. Findings Between Feb 12 and July 30, 2010, 173 patients at 43 investigational sites in 12 countries were enrolled into the study, of whom 162 were eligible and were randomly assigned to the two treatment groups (81 to the olaparib plus chemotherapy group and 81 to the chemotherapy alone group). Of these randomised patients, 156 were treated in the combination phase (81 in the olaparib plus chemotherapy group and 75 in the chemotherapy alone group) and 121 continued to the maintenance or no further treatment phase (66 in the olaparib plus chemotherapy group and 55 in the chemotherapy alone group). BRCA mutation status was known for 107 patients (either at baseline or determined retrospectively): 41 (38%) of 107 had a BRCA mutation (20 in the olaparib plus chemotherapy group and 21 in the chemotherapy alone group). Progression-free survival was significantly longer in the olaparib plus chemotherapy group (median 12.2 months [95% CI 9.7-15.0]) than in the chemotherapy alone group (median 9.6 months [95% CI 9.1-9.7) (HR 0.51 [95% CI 0.34-0.77]; p=0.0012), especially in patients with BRCA mutations (HR 0.21 [0.08-0.55]; p=0.0015). In the combination phase, adverse events that were reported at least 10% more frequently with olaparib plus chemotherapy than with chemotherapy alone were alopecia (60 [74%] of 81 vs 44 [59%] of 75), nausea (56 [69%] vs 43 [57%]), neutropenia (40 [49%] vs 29 [39%]), diarrhoea (34 [42%] vs 20 [27%]), headache (27 [33%] vs seven [9%]), peripheral neuropathy (25 [31%] vs 14 [19%]), and dyspepsia (21 [26%] vs 9 [12%]); most were of mild-to-moderate intensity. The most common grade 3 or higher adverse events during the combination phase were neutropenia (in 35 [43%] of 81 patients in the olaparib plus chemotherapy group vs 26 [35%] of 75 in the chemotherapy alone group) and anaemia (seven [9%] vs five [7%]). Serious adverse events were reported in 12 (15%) of 81 patients in the olaparib plus chemotherapy group and 16 of 75 (21%) patients in the chemotherapy alone group. Interpretation Olaparib plus paclitaxel or carboplatin followed by maintenance monotherapy significantly improved progression-free survival versus paclitaxel plus carboplatin alone, with the greatest clinical benefit in BRCA-mutated patients, and had an acceptable and manageable tolerability profile.