BACKGROUNDIn ARIEL3, rucaparib maintenance treatment significantly improved progression-free survival versus placebo. Here, we report prespecified, investigator-assessed, exploratory post-progression endpoints and updated safety data.METHODSIn this ongoing (enrolment complete) randomised, placebo-controlled, phase 3 trial, patients aged 18 years or older who had platinum-sensitive, high-grade serous or endometrioid ovarian, primary peritoneal, or fallopian tube carcinoma and an Eastern Cooperative Oncology Group performance status of 0 or 1 who had received at least two previous platinum-based chemotherapy regimens and responded to their last platinum-based regimen were randomly assigned (2:1) to oral rucaparib (600 mg twice daily) or placebo in 28-day cycles using a computer-generated sequence (block size of six with stratification based on homologous recombination repair gene mutation status, progression-free interval following penultimate platinum-based regimen, and best response to most recent platinum-based regimen). Patients, investigators, site staff, assessors, and the funder were masked to assignments. The primary endpoint of investigator-assessed progression-free survival has been previously reported. Prespecified, exploratory outcomes of chemotherapy-free interval (CFI), time to start of first subsequent therapy (TFST), time to disease progression on subsequent therapy or death (PFS2), and time to start of second subsequent therapy (TSST) and updated safety were analysed (visit cutoff Dec 31, 2017). Efficacy analyses were done in all patients randomised to three nested cohorts: patients with BRCA mutations, patients with homologous recombination deficiencies, and the intention-to-treat population. Safety analyses included all patients who received at least one dose of study treatment. This trial is registered with ClinicalTrials.gov, NCT01968213.FINDINGSBetween April 7, 2014, and July 19, 2016, 564 patients were enrolled and randomly assigned to rucaparib (n=375) or placebo (n=189). Median follow-up was 28·1 months (IQR 22·0-33·6). In the intention-to-treat population, median CFI was 14·3 months (95% CI 13·0-17·4) in the rucaparib group versus 8·8 months (8·0-10·3) in the placebo group (hazard ratio [HR] 0·43 [95% CI 0·35-0·53]; p<0·0001), median TFST was 12·4 months (11·1-15·2) versus 7·2 months (6·4-8·6; HR 0·43 [0·35-0·52]; p<0·0001), median PFS2 was 21·0 months (18·9-23·6) versus 16·5 months (15·2-18·4; HR 0·66 [0·53-0·82]; p=0·0002), and median TSST was 22·4 months (19·1-24·5) versus 17·3 months (14·9-19·4; HR 0·68 [0·54-0·85]; p=0·0007). CFI, TFST, PFS2, and TSST were also significantly longer with rucaparib than placebo in the BRCA-mutant and homologous recombination-deficient cohorts. The most frequent treatment-emergent adverse event of grade 3 or higher was anaemia or decreased haemoglobin (80 [22%] patients in the rucaparib group vs one [1%] patient in the placebo group). Serious treatment-emergent adverse events were reported in 83 (22%) patients in the rucaparib group and 20 (11%) patients in the placebo group. Two treatment-related deaths have been previously reported in this trial; there were no new treatment-related deaths.INTERPRETATIONIn these exploratory analyses over a median follow-up of more than 2 years, rucaparib maintenance treatment led to a clinically meaningful delay in starting subsequent therapy and provided lasting clinical benefits versus placebo in all three analysis cohorts. Updated safety data were consistent with previous reports.FUNDINGClovis Oncology.