peer reviewed ; AIMS: To estimate viral seroprevalences for HCV, HBV and HIV among belgian intravenous (IVDU) and non intravenous (non-IVDU) drug users; to assess risk factors for HCV infection in IVDU; to assess feasibility of chronic hepatitis C follow-up in this population. DESIGN: Cross-sectional study. Demographic and behavioural characteristics were obtained by a standardized questionnaire. Serum samples were tested for HCV, HBV and HIV. SUBJECTS AND SETTING: 329 patients (244 IVDU and 85 non-IVDU) attending ten general practitioners in 1995. RESULTS: HCV seroprevalence was 78.3%; it was 35.7% for HBV and 0.9% for HIV in IVDU, vs 2.4%, 8.3% and 0%, respectively, in non-IVDU. In logistic regression analysis, independent risk factors for HCV infection were: 1/sharing of syringes and/or of "cottons" used as filters (adjusted prevalence odds ratio [POR] = 31.7; 95% confidence interval [CI] = 9.8-102.5), 2/duration of injecting upper than one month (adjusted POR = 8.6; CI = 3.0-24.7) and 3/age (adjusted POR = 1.2 by year of difference; CI = 1.0-1.3). A biochemical follow-up was obtained in 70% of HCV seropositive users; 79.5% of them had chronic hepatitis C (mean value of ALT = 3.5 times upper normal value, range 1.1-23.0). Among these, 24.7% went through liver biopsy during the three years follow-up period of the study. CONCLUSIONS: HCV seroprevalence is very high among belgian IVDU. Prevention strategies have to focus on neophytes injectors. They must be urgently revisited for what concern needles/syringes exchange programs: "cottons" must be included. Follow-up and treatment of chronic hepatitis C seem to be poorly effective among drug users.