Objective: Adding bevacizumab to cisplatin-paclitaxel for advanced cervical cancer significantly improves overall and progression-free survival. We evaluated bevacizumab with a widely used carboplatin-paclitaxel backbone. Methods: Patients with metastatic/recurrent/persistent cervical cancer not amenable to curative surgery and/or radiotherapy received 3-weekly bevacizumab 15 mg/kg, paclitaxel 175 mg/m2, and carboplatin AUC 5 until progression or unacceptable toxicity. Maintenance bevacizumab was allowed. Patients with ongoing bladder/rectal involvement, prior cobalt radiotherapy, a history of fistula/gastrointestinal perforation, or recent bowel resection/chemoradiation were excluded. The primary objective was to determine incidences of gastrointestinal perforation/fistula, gastrointestinal-vaginal fistula, and genitourinary fistula. Results: Among 150 treated patients, disease at study entry was persistent in 21%, recurrent in 56%, and newly diagnosed metastatic in 23%. After 27.8 months' median follow-up, median bevacizumab duration was 6.7 months; 57% received maintenance bevacizumab. Seventeen patients (11.3%; 95% CI: 6.7-17.5%) experienced ≥1 perforation/fistula event: gastrointestinal perforation/fistula in 4.7% (1.9-9.4%), gastrointestinal-vaginal fistula in 4.0% (1.5-8.5%), and genitourinary fistula in 4.7% (1.9-9.4%). Of these, 16 were previously irradiated, several with ongoing radiation effects. The most common grade 3/4 adverse events were neutropenia (25%), anemia (19%), and hypertension (14%). Five patients (3%) had fatal adverse events. Objective response rate was 61% (95% CI: 52-69%), median progression-free survival was 10.9 (10.1-13.7) months, and median overall survival was 25.0 (20.9-30.4) months. Conclusions: Bevacizumab can be combined with carboplatin-paclitaxel in the CECILIA study population. The fistula/gastrointestinal perforation incidence is in line with GOG-0240; efficacy results are encouraging.