OBJECTIVE: To assess the relationship between platelet counts and risk of AIDS and non-AIDS-defining events.DESIGN: Prospective cohort.METHODS: EuroSIDA patients with at least one platelet count were followed from baseline (first platelet ≥ 1 January 2005) until last visit or death. Multivariate Poisson regression was used to assess the relationship between current platelet counts and the incidence of non-AIDS-defining (pancreatitis, end-stage liver/renal disease, cancer, cardiovascular disease) and AIDS-defining events.RESULTS: There were 62 898 person-years of follow-up (PYFU) among 12 279 patients, including 1168 non-AIDS-defining events [crude incidence 18.6/1000 PYFU, 95% confidence interval (CI) 17.5-19.6] and 735 AIDS-defining events (crude incidence 11.7/1000 PYFU, 95% CI 10.8-12.5). Patients with thrombocytopenia (platelet count ≤100 × 10/l) had a slightly increased incidence of AIDS-defining events [adjusted incidence rate ratio (aIRR) 1.42, 95% CI 1.07-1.86], when compared to those with platelet counts 101-200 × 10/l, whereas the incidence of non-AIDS-defining events was more than two-fold higher (aIRR 2.66, 95% CI 2.17-3.26). Among non-AIDS-defining events, the adjusted incidence of cancer (aIRR 2.20, 95% CI 1.61-3.01), but not cardiovascular disease (aIRR 0.66, 95% CI 0.32-1.34), was significantly higher in patients with thrombocytopenia. The association between thrombocytopenia and cancer remained unaltered in sensitivity analyses requiring repeated platelet counts to confirm thrombocytopenia and lagging platelets by 1 year prior to clinical events.CONCLUSION: Patients with thrombocytopenia had increased incidence of AIDS-defining and non-AIDS-defining events, but the association with the latter, in particular cancer, was stronger. Future studies should investigate whether the pathophysiological processes underlying thrombocytopenia are associated with the development of cancer during treated HIV disease.