Bisphosphonates and esophageal cancer—a pathway through the confusion

Abstract

When two similar, credible analyses of the same patient database report seemingly contrasting risk estimates for the association with cancer of a widely-prescribed therapy, it is difficult for clinicians, never mind for patients, to make decisions about treatment options. Nevertheless, clinicians can—and must—interpret the available evidence for their patients, helping them to weigh the potential benefits and harms of their prescription. Two recent papers that examined the association between oral bisphosphonates and the risk of esophageal cancer reported seemingly discrepant findings, despite both studies using data from the UK General Practice Research Database, from similar time periods. Each paper set out brief conclusions within their abstracts, generating two opposing impressions: that the risk of incident esophageal cancer is not significantly increased with bisphosphonate use, or that the risk is increased when bisphosphonates have been prescribed ten or more times, or for longer than 5 years. Of course, the latter conclusion was more widely reported in the popular press, and with patients arriving at the clinic wanting to discuss the reports, clinicians must be able to interpret the existing evidence correctly. In this article we discuss important methodological issues regarding the two observational studies, define areas of consensus and discrepancy, examine possible bias and confounding, and condense the results into a clinically useful summary.