The rate of human immunodeficiency virus (HIV) disease progression or death of individuals coinfected with hepatitis C virus (HCV) is conflicting. The complete-case analysis systematically used, excludes patients unscreened for HCV. Our objective was to assess if rate of survival differed between HIV-infected patients screened and unscreened for HCV in a hospital-based prospective cohort study. Patients were enrolled in the Lyon section of the French Hospital Database on HIV between 1 July 1992 and 31 May 2005. A multivariate Cox regression model was used to analyse the association of HCV screening with survival. Of 3244 patients, 299 (9.2%) were not screened for HCV. The populations screened and unscreened differed by the proportion of acquired immune deficiency syndrome at baseline, presumed route of infection, CD4 cell count category at baseline, mean duration of follow-up, mean number of visits per year, type of antiretroviral therapy and survival. The rate of progression to death was higher for non-HCV-screened vs HCV-screened patients: the incidence rate among HCV-screened patients was 22.9/1000 patient-years; the incidence rate among HCV-unscreened patients was 52.4/1000 patient-years. The adjusted hazards ratio of death was 2.48 [95% confidence interval (1.83-3.35); P < 0.001] for patients with unknown HCV status compared with others. In conclusion, unscreened or unknown HCV status was associated with an increased risk of death in our hospital cohort. Important prognostic factors are related to, or confounded by the practice of HCV screening.