There were twenty-three trials of 25,867 patients but only data from 15,263 were included in the meta-analysis. Eleven trials were double-blind. Median follow-up (where reported) ranged from six to 43.5 months. Progression-free survival ranged from 2.9 to 14.1 months.
There was a statistically significant increased risk of all-grade neutropenia (RR 1.15; 95% CI 1.01 to 1.30; seven studies; Ι²=0%) for the bevacizumab group compared with controls. There was also an increased risk of high-grade neutropenia (RR 1.08; 95% CI 1.02 to 1.13; 21 studies; Ι²=29.2%) and febrile neutropenia (RR 1.31; 95% CI 1.08 to 1.58; 16 studies; Ι²=0%) for the bevacizumab group compared with controls.
The bevacizumab group had an increased risk for all-grade thrombocytopenia (RR 1.22; 95% CI 1.00 to 1.48; seven studies; Ι²=0.1%). There did not appear to have been a difference in risk between the bevacizumab group and controls for high-grade thrombocytopenia (RR 1.10; 95% CI 0.79 to 1.54; 15 studies; Ι²=45.5%).
Bevacizumab was associated with a decreased risk of all-grade (RR 0.81; 95% CI 0.68 to 0.96; five studies; Ι²=0%) and high-grade anaemia (RR 0.73; 95% CI 0.60 to 0.89; 15 studies; Ι²=6.3%).
There was no evidence of differences in risk of haematological adverse events between groups when different doses of bevacizumab were compared, whether the trials used chemotherapy or immunotherapy or on the basis of quality assessment criteria.
There was some evidence of publication bias identified for high-grade incidence of anaemia, neutropenia and thrombocytopenia.