Weigh risk vs. benefit when using bevacizumab combined with chemotherapy

Issue: February 25, 2011

ByAndrew H. Ko, MD

One of an oncologist’s primary responsibilities involves discussions with patients and their families regarding the potential benefits and risks associated with any therapy that is being recommended.

In the Feb. 2 issue of the Journal of the American Medical Association, Ranpura and colleagues report on a meta-analysis in which they examined published randomized controlled trials that compared bevacizumab (Avastin, Genentech) combined with chemotherapy or biologic therapy with chemotherapy or biologic therapy alone. Specifically, they wanted to determine the overall risk for fatal adverse events (FAEs) associated with the use of bevacizumab as part of combination regimens.

In total, the researchers analyzed data from 16 clinical trials encompassing more than 10,000 patients with diagnoses that included colorectal cancer (5 studies); non–small cell lung cancer (4 studies); breast cancer (3 studies); kidney cancer (2 studies); pancreatic cancer and prostate cancer (1 study each). They report that the overall incidence of FAEs in patients is significantly higher in patients who received bevacizumab-containing regimens, with a relative risk of 1.46 (incidence, 2.5% vs. 1.7%; P=.01). The most common specific causes of FAEs included hemorrhage, especially pulmonary and gastrointestinal (23.5%); neutropenia (12.2%); gastrointestinal perforation (7.1%); pulmonary embolism (5.1%); and stroke (5.1%).

Andrew H. Ko, MD
Andrew H. Ko

The researchers also attempted to dissect out specific subgroups of patients who may be at particularly higher risk for FAEs when treated with bevacizumab. There appeared to be an association with the specific chemotherapy regimen used: combining bevacizumab with platinum analogues or taxanes was associated with a significantly increased risk for FAEs (RR= 3.49; P<.001), whereas this association was not found with other chemotherapy or biologic platforms. No significant difference in FAE incidence was detected according to tumor type, although there was a wide range of relative risks, ranging from 0.69 in breast cancer to 3.85 in prostate cancer. Although higher-dose administration of bevacizumab (defined as 5 mg/kg per week) was associated with a significantly increased risk for FAEs (RR=2.55; P=.001) and lower-dose administration (2.5 mg/kg per week) was not (RR 1.36; P=0.17), this finding may have simply been a function of sample size, as there was, in fact, no significant difference in FAE rate between these higher and lower bevacizumab doses.

Interpreting data

How should these data be interpreted? Here, the difference between describing these results according to relative and absolute risk becomes an important distinction. One could appropriately say that use of bevacizumab increases the risk for a fatal event by almost 50% (relative risk), and this would be correct, but also, in some ways, perhaps unnecessarily alarmist. Conversely, informing a patient that the addition of bevacizumab to his/her chemotherapy regimen may increase the risk for treatment-related fatality by 0.8% (absolute risk) may provide some more reassuring context to the risk vs. benefit ratio discussion. When countered by the explanation of the potential benefits accorded by the use of bevacizumab (although this, itself, is the subject of an entirely separate and controversial debate, depending on disease type), most patients may be very willing to accept that less than 1% increased risk.

Equally or more importantly, these findings represent a bird’s-eye view that a meta-analysis provides, with all its inherent limitations. Broadly noting that the risk for FAEs increases by 0.8% (or 46%) with use of bevacizumab does not account for inter-individual differences that may increase, or attenuate, that fatality risk. The point here is that assigning risk vs. benefit in treatment decisions is very much made at an individual patient level based on many variables, some accounted for in this article, others not.

Oncologists who prescribe this drug frequently, who should not be surprised in any way by the types of FAEs reported here, will typically avoid bevacizumab in certain patients: those who have centrally located pulmonary lesions (primary or metastatic); a recent history of bleeding ulcers; or significant cardiovascular disease, for example. An octogenarian with known coronary artery disease has a different level of risk associated with use of bevacizumab than an otherwise healthy 45-year-old, even if both carry a diagnosis of metastatic colorectal cancer. In the long run, this meta-analysis is unlikely to change practice patterns in any meaningful way in terms of who receives bevacizumab; simply put, those individuals in whom the risk for serious adverse events, including fatality, is judged to outweigh any incremental benefit with the addition of this drug, will and should not get it.

Advanced disease

Finally, it is important to note that all of the studies included in this meta-analysis comprised patients with advanced stages of cancer, primarily those with metastatic disease, in whom treatment was rarely, if ever, curative in intent. The willingness to tolerate risk of treatment-related FAEs in this patient population (by providers, as well as by patients themselves) may differ to some extent from that in which therapy is administered in the adjuvant setting. In this latter context, patients may already be cured of their disease, and hence the risk of treatment-related death may be deemed far more unacceptable.

In two recently reported trials evaluating bevacizumab in the adjuvant setting for resected colon cancer (NSABP C08 and AVANT), no increased risk of FAEs was reported in the bevacizumab-containing arms compared with the chemotherapy-alone arms, even with the bevacizumab continuing for a full 12 months in both studies. However, it should be noted, there was no evidence of improvement in DFS in either study.

In summary, this JAMA article offers some valuable, if largely unsurprising, insight into the risks associated with bevacizumab use in the treatment of patients with different solid tumor types. Such information must be balanced against the short- and long-term benefits of this agent, the non-fatal complications that may affect quality of life and ability to deliver treatment in a timely fashion, and — certainly not to be ignored — cost-effective considerations. Above all, patient-specific context and helping individual patients take part in, and understand the nuances that go into, this decision-making process remain paramount.

Andrew H. Ko, MD, associate professor of clinical medicine at the University of California, San Francisco Comprehensive Cancer Center, is a HemOnc Today Editorial Board member. Disclosure: Dr. Ko reports no relevant financial disclosures.

For more information:

Ranpura V. JAMA. 2011;305:487-494.