The role of continued drug use in development of AIDS has been investigated, using mainly CD4(+) cells as outcome variable. The objective of this study was to verify whether continued drug use can influence HIV-1 plasma viral load.We used cross-sectional enrolment data of the HIV-infected IDUs cohort study Manif 2000 (October 1995-October 1996), recruiting patients in hospital departments of Marseilles, Nice and Paris suburbs. To minimize biases due to differential access to health care, only patients receiving outpatient care for at least 2 years but with no antiretroviral treatment were selected for analysis (n = 108). Available information regarding clinical and laboratory information from medical records and clinical examination as well as drug use and HIV-related risk practices were obtained by a face-to-face and a self-administered questionnaire. Patients denying recent heroin injection were cross-validated by a serological assay to detect morphine. Forty-two patients (39%) reported recent heroin injection; among those denying recent use (n=66), eight (positive for serum morphine assay) were re-classified as IDUs. A difference of 0.35 log in viral load was observed between active and ex-IDUs which increases (0.6 log, p=0.03) in those who have been using drugs for 10 years or more.This result persisted when adjusted for CD4(+) counts, clinical stage or years since diagnosis. Continued drug use may have a significant but limited impact on HIV viral load only in patients with a longer history of drug use. Consequences of persisting drug use on long-term progression to AIDS has to be investigated further.