Author Affiliations: Division of Epidemiology, Chinese University of Hong Kong, Hong Kong SAR, China (firstname.lastname@example.org).
To the Editor: We found some errors in the results in a recent systematic review and meta-analysis of treatment-related mortality with bevacizumab by Dr Ranpura and colleagues.1
The numbers of fatal adverse events (FAEs) were extracted incorrectly or by using different extraction criteria from some of the included studies. In the study by Hurwitz et al,2 the incidence of FAEs (“adverse event leading to death”) was 10 of 393 in the bevacizumab group and 11 of 397 in the placebo group, as opposed to 1 of 393 and 0 of 397 as reported by Ranpura et al.1 In the study by Escudier et al,3 8 of 337 in the treatment group and 7 of 304 in the control group had a “death not due to disease progression,” in contrast to 3 of 337 and 0 of 304 reported by Ranpura et al.1 In addition, the authors used the incorrect denominator for the incidence of FAEs in the study by Saltz et al4: it should be 694 in the bevacizumab group and 675 in the placebo group, instead of 675 in the treatment group and 694 in the control group, as reported in the review.1
Yang Z, Mao C, Tang J. Bevacizumab and Cancer Treatment-Related Mortality. JAMA. 2011;305(22):2291-2293. doi:10.1001/jama.2011.757