When preclinical toxicology studies show a clear and dose-related association between a medication exposure and a specific tumor type, an obvious question is, what will happen in humans? Regulatory agencies often respond by requiring additional studies before approval or a safety study to address the possible medication-outcome association in the postmarketing environment. However, if the medication is not commonly used, and the tumor rarely occurs in the general population, the probability of being able to confirm anything other than a very large increase in risk in a treated population is low. For the symposium described in the Pinheiro et al. article in this issue, we illustrated this type of challenge using a program of research relating to teriparatide; we offer some observations here.
