BACKGROUND Although several studies have evaluated the frequency of adverse drug reactions (ADRs) and drug-drug interactions (DDIs) in general medicine, few studies have looked at the epidemiology of adverse drug events (ADEs) in oncology. OBJECTIVE We sought to investigate how many hospital admissions in oncology are related to a DDI or an ADR. METHODS All cancer patients admitted to an oncology ward during an eight-month period had their charts retrospectively evaluated for reasons of hospitalization, using a 4-point scale (definitely, probably, possibly, or unlikely associated) to classify admissions by their probability of being associated with either a DDI or an ADR. RESULTS From September 2007 to May 2008, there were 550 hospital admissions and 458 were eligible. Among unplanned admissions (n=298), 39 (13.0%, 95% confidence interval [CI] 9.4%-17.4%) were considered to be associated with an ADE, 33 (11.0%, 95% CI 7.7%-15.2%) with an ADR, and six (2.0%, 95% CI 0.7%-4.3%) with a DDI. The most common DDIs involved warfarin, captopril, and anti-inflammatory agents, and the most frequent ADR was neutropenic fever post-chemotherapy. Most patients were discharged completely recovered, but two patients died. CONCLUSIONS Approximately one in 10 unplanned hospitalizations of cancer patients is associated with an ADE. Prospective and population-based studies are warranted to evaluate their magnitude in oncology.