PREDICTORS OF MORTALITY IN THE AMSTERDAM COHORT OF HUMAN-IMMUNODEFICIENCY-VIRUS (HIV)-POSITIVE AND HIV-NEGATIVE DRUG-USERS

The impact of human immunodeficiency virus (HIV) infection and other r isk factors on mortality was studied in a cohort of Dutch injection dr ug users and drug users who did not inject. Participants were recruite d between 1985 and 1992 and followed up through 1993. Vital status was ascertained through repeat visit information, supplemented by populat ion register data. A total of 77 deaths were recorded among 632 drug u sers, for a mortality rate per 1,000 person-years of 7 for HIV-negativ e noninjection drug users, 18 for HIV-negative injection drug users, a nd 64 for HIV-positive injection drug users. In multivariate analyses, limited to injection drug users, a positive HIV serostatus, age above 40 years, and using benzodiazepines several times daily were signific antly associated with an elevated risk of death, both for death from a ll causes and for death preceding acquired immunodeficiency syndrome ( AIDS) diagnosis (pre-AIDS). For pre-AIDS death, the adjusted relative risk associated with HIV infection was 2.2 (95% confidence interval 1. 3-3.7). Only 38% of HIV-infected injection drug users who died were di agnosed with AIDS. However, 76% of HIV-infected injection drug-users w ho died without AIDS diagnosis had evidence of immunosuppression (CD4 count < 500/mu l). Daily use of methadone and participation in needle and syringe exchange schemes were not associated with lower mortality rates. This study illustrates in a group of injection drug users with a 30% HIV seroprevalence and a high background mortality the profound influence on mortality that HIV infection has gained. Am J Epidemiol 1 996;143:380-91.