In this retrospective study, we investigated adverse events and outcomes in patients treated with bevacizumab for ovarian, fallopian tube, or primary peritoneal cancers at a single hospital. We determined the cumulative incidences of various bevacizumab-related adverse events and the correlation between dose and adverse event incidences. We analyzed data from 154 patients that received 251 rounds of bevacizumab as first-line, first salvage, >2 salvage treatments. Adverse events of any grade were observed in 121 (78.6%) patients; at least one grade 3 or 4 adverse event occurred in 32 (20.8%) patients. The two most common events were proteinuria (38.3%) and hypertension (33.8%). The first-line treatment group displayed significantly higher frequencies of hypertension (52.7% vs. 18.9% vs. 15.5%, < 0.001), wound complications (9.1% vs. 0% vs. 1.2%, = 0.010), arthralgia (29.1% vs. 11.3% vs. 8.3%, = 0.003), and reduced range of joint motion (14.5% vs. 5.7% vs. 3.6%, = 0.046), compared to those in the first and >2 lines salvage groups, respectively (Kruskal-Wallis test). The cumulative incidences of all grades and grades 3/4 of hypertension cumulative incidence plateaued at around 30% for all grades and 10% for grades 3 and 4, at bevacizumab doses above 8080 and 3510 mg, respectively. The proteinuria cumulative incidence plateaued at around 35% for all grades and 3% for grades 3 and 4, at bevacizumab doses above 11,190 and 4530 mg, respectively. We concluded that, in this realistic clinical population, different kinds and higher cumulative incidences of adverse events were observed compared to those reported in previous clinical trials. Moreover, bevacizumab doses showed cumulative toxicity and plateau effects on hypertension and proteinuria.