Lung cancer (LC) is now the leading cause of cancer mortality in the world, therefore it would be useful to identify prognostic factors to determine patient outcome. The objective of this study is to evaluate the usefulness of platelet counts at the time of diagnosis as a prognostic factor. A retrospective study of patients with histological diagnostic evidence of LC was carried in our catchment area over a 3-year period. Survival adjusted for other factors was assessed according to the platelet count at the time of diagnosis. Patients with platelet levels within the reference range (RR) (135000-381000/microl) were divided into two groups, between 135000-258000/microl and 258000-381000/microl. A third group was made up of patients with platelet counts over 381000/microl. Adjusted survival was analysed using Cox regression models. Patients with high platelets have a 37% worse survival than those with a platelet level within the RR, but lower than 258000/microl. When tumour stage is included in the covariates, platelet levels are no longer an independent survival factor. In conclusion, platelet levels at the time of diagnosis could be a useful prognostic factor in LC.