Background: Angiopoietin 1 and 2 regulate angiogenesis and vascular remodelling by interacting with the tyrosine kinase receptor Tie2, and inhibition of angiogenesis has shown promise in the treatment of ovarian cancer. We aimed to assess whether trebananib, a peptibody that inhibits binding of angiopoietin 1 and 2 to Tie2, improved progression-free survival when added to carboplatin and paclitaxel as first-line therapy in advanced epithelial ovarian, primary fallopian tube, or peritoneal cancer in a phase 3 clinical trial. Methods: TRINOVA-3, a multicentre, multinational, phase 3, double-blind study, was done at 206 investigational sites (hospitals and cancer centres)in 14 countries. Eligible patients were aged 18 years or older with biopsy-confirmed International Federation of Gynecology and Obstetrics (FIGO)stage III to IV epithelial ovarian, primary peritoneal, or fallopian tube cancers, and an ECOG performance status of 0 or 1. Eligible patients were randomly assigned (2:1)using a permuted block method (block size of six patients)to receive six cycles of paclitaxel (175 mg/m2)and carboplatin (area under the serum concentration-time curve 5 or 6)every 3 weeks, plus weekly intravenous trebananib 15 mg/kg or placebo. Maintenance therapy with trebananib or placebo continued for up to 18 additional months. The primary endpoint was progression-free survival, as assessed by the investigators, in 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 NCT01493505, and is complete. Findings: Between Jan 30, 2012, and Feb 25, 2014, 1164 patients were screened and 1015 eligible patients were randomly allocated to treatment (678 to trebananib and 337 to placebo). After a median follow-up of 27·4 months (IQR 17·7–34·2), 626 patients had progression-free survival events (405 [60%]of 678 in the trebananib group and 221 [66%]of 337 in the placebo group). Median progression-free survival did not differ between the trebananib group (15·9 months [15·0–17·6])and the placebo group (15·0 months [12·6–16·1])groups (hazard ratio 0·93 [95% CI 0·79–1·09]; p=0·36). 512 (76%)of 675 patients in the trebananib group and 237 (71%)of 336 in the placebo group had grade 3 or worse treatment-emergent adverse events; of which the most common events were neutropenia (trebananib 238 [35%]vs placebo 126 [38%])anaemia (76 [11%]vs 40 [12%]), and leucopenia (81 [12%]vs 35 [10%]). 269 (40%)patients in the trebananib group and 104 (31%)in the placebo group had serious adverse events. Two fatal adverse events in the trebananib group were considered related to trebananib, paclitaxel, and carboplatin (lung infection and neutropenic colitis); two were considered to be related to paclitaxel and carboplatin (general physical health deterioration and platelet count decreased). No treatment-related fatal adverse events occurred in the placebo group. Interpretation: Trebananib plus carboplatin and paclitaxel did not improve progression-free survival as first-line treatment for advanced ovarian cancer. The combination of trebananib plus carboplatin and paclitaxel did not produce new safety signals. These results show that trebananib in combination with carboplatin and paclitaxel is minimally effective in this patient population. Funding: Amgen.