Background: Multiple stakeholders wish to know if medications increase the risk of cancer. Clinical trials and enhanced pharmacovigilance have limitations for studying cancer such as incomplete capture and high cost (for trials). Observational studies are used to characterize the risk of cancer but may be limited due to inadequate case identification, exposure assessment and data sources, particularly for rare cancers. The ability to link patients to existing national cancer outcome data could be an ideal solution.
Objective: Review postmarketing commitments (PMC) to identify ones that may benefit from collaboration with cancer registries in the US.
Methods: We reviewed the FDA PMC database to identify cancer outcomes under study in nononcological drugs. We reviewed drugs with a NDA/BLA approval date after 1994 and excluded drugs indicated for oncology treatments or supportive therapy or where an animal study or clinical trial was requested. We reviewed approval letters and other published material to characterize the therapeutic class, study design and method for identifying cancer.
Results: Forty-six PMCs for 33 different drug entities were identified from the following drug classes: Immunologic (n = 11), endocrine and metabolic (n = 8), dermatologic (n = 3) and other (n = 11). The most common cancer for the 12 entities that had a boxed warning for cancer was lymphoma (n = 7), followed by thyroid C-cell tumor (n = 4), other malignancies (n = 4), skin cancer (n = 2) and osteosarcoma (n = 1). Study designs were not well described for all PMCs. Of the 46 PMCs, the most common method for identifying cancer was active surveillance of patients (n = 10). Two studies mentioned cancer registries for long-term follow-up.
Conclusions: Postmarketing drug safety studies require the ability to properly identify and classify cancer outcomes over long periods. Linking treated cohorts from postapproval registries, database studies or clinical trials to cancer registries at a national level could provide a scientifically robust way to efficiently and accurately quantify cancer risk.
