gms | German Medical Science

53. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie

15. bis 18.09.2008, Stuttgart

Safety of Bevacizumab, an Antibody against Vascular Endothelial Growth Factor, in the Management of Cancer: A Meta-Analysis of Randomized Controlled Trials

Meeting Abstract

  • Sabine Geiger-Gritsch - Department of Public Health, Medical Decision Making and Health Technology Assessment, UMIT – University of Health Sciences, Medical Informatics and Technology und Ludwig Boltzmann Institut für HTA, Hall in Tirol, Österreich
  • Bjoern Stollenwerk - Department of Public Health, Medical Decision Making and Health Technology Assessment, UMIT – University of Health Sciences, Medical Informatics and Technology, Hall in Tirol, Österreich
  • Rebecca Miksad - Institute for Technology Assessment and Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, USA
  • Beate Guba - Ludwig Boltzmann Institut für HTA, Wien, Österreich
  • Claudia Wild - Ludwig Boltzmann Institut für HTA, Wien, Österreich
  • Uwe Siebert - Department of Public Health, Medical Decision Making and Health Technology Assessment, UMIT – University of Health Sciences, Medical Informatics and Technology, Hall in Tirol, Österreich

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 53. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie (gmds). Stuttgart, 15.-19.09.2008. Düsseldorf: German Medical Science GMS Publishing House; 2008. DocMBIO1-3

The electronic version of this article is the complete one and can be found online at:

Published: September 10, 2008

© 2008 Geiger-Gritsch et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( You are free: to Share – to copy, distribute and transmit the work, provided the original author and source are credited.




Bevacizumab, a humanized recombinant monoclonal antibody against vascular endothelial growth factor (VEGF), was the first angiogenesis inhibitor used for cancer treatment. It is approved for the treatment of patients with metastatic colorectal cancer, advanced non-small-cell lung cancer, metastatic breast cancer or metastatic renal cell carcinoma, and its application in other types of cancer is also undergoing extensive clinical research. In phase III randomized controlled clinical trials the combination of bevacizumab with chemotherapy has been shown to improve survival in patients with metastatic colorectal and non-small-cell lung cancer or to increase the progression-free survival time in patients with breast and renal cell cancer. Compared to other cancer treatments, bevacizumab is usually well-tolerated. However, in some trials bevacizumab was associated with an increased risk of adverse events such as hypertension, proteinuria or thromboembolism [1], [2]. Other potential side-effects (e.g., GI perforation, bleeding, leukopenia) have not yet been systematically assessed and bevacizumab-related toxicity still remains a controversial issue. The aim of this study was to perform a comprehensive systematic meta-analysis on the safety of bevacizumab in patients with cancer.


We conducted a systematic literature search in electronic databases including Medline (1950 to December 2007), EMBASE (1988 to August 2007) and Web of Science (1990 to June 2007) complemented by a manual search in reference lists. We only included full publications of randomized controlled trials (RCTs) in patients with cancer that compared bevacizumab therapy with chemotherapy alone or placebo and reported comprehensive data on adverse events. We performed a series of meta-analyses for different safety outcome. The primary endpoint of our analysis was “any severe adverse event” (grade 3 or 4). In addition, we explored further secondary endpoints as reported in the identified studies. For each endpoint, we calculated relative adverse event risks (RR), risk differences (RD), and number needed to harm (NNH) with 95% confidence interval (95%CI) using random-effects or fixed-effects model, depending on heterogeneity. We generated funnel plots to assess publication bias.


We included 12 RCTs published between July 2003 and December 2007 comprising a total of 5,826 patients with different types of cancer from 610 articles screened for analysis. Included studies reported safety data on hypertension, thrombotic events and bleeding (n=12), proteinuria (n=11), hematological events (i.e., leukopenia, neutropenia or thrombocytopenia, n=9) gastrointestinal hemorrhage/perforation (n=8). Seven studies reported our primary outcome “any grade 3/4 adverse event”. Insufficient data were available for hemoptysis, wound-healing complications, and asthenia. Compared with controls, bevacizumab was associated with a slightly higher risk of any severe (grade 3/4) events. However this difference was not statistically significant. The pooled RR was 1.07 (95%CI: 0.99 to 1.16) and the pooled risk difference was 5% (95%CI: -1% to 12%), which translates into 20 treated patients required to harm one (NNH). Further explorative analysis showed an increased risk regarding hypertension, proteinuria, and any bleeding and gastrointestinal hemorrhage/perforation. Bevacizumab was not associated with an increased risk for thrombotic or hematological events. Some funnel plots indicated the potential of a moderate publication bias.


We systematically summarized and quantified the risk of several important adverse events observed in RCTs using bevacizumab. Our study has several limitations. First, we pooled the results of clinical trials that were not originally designed to explore safety outcomes, and therefore, have limited power to detect adverse events. Second, the quality of safety outcome reporting and the type of reported adverse events differed substantially between included studies. Third, besides our primary endpoint overall severe adverse event (grade 3 or 4), we explored several secondary more specific endpoints using data from the same patients which has the problem of multiple testing. As these analyses were of exploratory nature, we did not correct p-values and 95%CI for multiple testing. Forth, there was substantial heterogeneity in patient characteristics, study design, treatment combination and doses of bevacizumab, concurrent chemotherapy, and reporting quality. However, these factors only partially explained the differences between study results. Finally, as in most meta-analyses, results may be affected by publication bias. Despite these limitations, our results provide comprehensive information for clinicians who use bevacizumab in cancer therapy and underline the importance of individual benefit-risk assessment and personalized decision making. Furthermore, as bevacizumab is costly and probably extended to further indications, further studies should evaluate the cost-effectiveness of bevacizumab in cancer therapy. Finally, our meta-analytic results should be considered in clinical guidelines, bedside decision tools, and decision-analytic health technology assessment models to assess the trade-off between efficacy, safety and costs of bevacizumab-based therapy in comparison to other cancer treatment regimens or palliative care. These models should be tailored to the specific type and stage of cancer, the patient characteristics as well as the patient preferences in order to guide personalized treatment decisions [3].


Scappaticci FA, Skillings JR, Holden SN, Gerber HP, Miller K, Kabbinavar F, et al. Arterial thromboembolic events in patients with metastatic carcinoma treated with chemotherapy and bevacizumab. J Natl Cancer Inst. 2007 Aug 15;99(16):1232-9.
Zhu X, Wu S, Dahut WL, Parikh CR. Risks of proteinuria and hypertension with bevacizumab, an antibody against vascular endothelial growth factor: systematic review and meta-analysis. Am J Kidney Dis. 2007 Feb;49(2):186-93.
Siebert U. When should decision-analytic modeling be used in the economic evaluation of health care? [Editorial]. The European Journal of Health Economics. 2003;4(3):143-50.