A 12-month study was conducted to identify risk factors for human immunodeficiency virus (HIV) infections among intravenous drug users (IDU) attending drug rehabilitation clinic of the Psychiatric Hospital, Manama, Bahrain. Patients provided demographic and behavioural information based on a questionnaire. Two hundred and forty male IDUs participated in the study on voluntary basis. The seroprevalence of HIV was 21.1 per cent. The presence of HIV antibody was associated with educational status, frequency of injecting drugs and needle sharing.
PIP: During January 1st to December 31st, 1991, a total of 242 male intravenous drug users (IDUs) attending the drug rehabilitation clinic of the Psychiatric Hospital, Manama, Bahrain, for treatment were interviewed by trained interviewers using a structured questionnaire to gather information on age, sex, marital status, nationality, religion, income, educational status, history of drug use, sexual practices, and blood and blood product transfusions. The sera were tested for antibodies to HIV-1 and HIV-2 using a second generation enzyme linked immunoassay (ELISA) and confirmed by Western Blot. 51 (21.1%) IDUs were positive for anti-HIV-1 by ELISA as confirmed by Western blot. 5 of 62 (8.1%) cases with college education were positive for HIV compared with 35 of 138 (25.4%) with secondary level of education and 11 of 42 (26.2%) with primary or lower level of education (p 0.01). Among the behavioral characteristics, only needle sharing showed an association of borderline statistical significance (p = 0.06). Frequency of intravenous drug use and sexual practice showed 95% confidence limits close to 1.0 (the null value) and higher limits of 4.61 and 2.71, respectively. Only educational status showed a significantly lower risk of HIV positivity in persons with college education with an odds ratio of 0.25. Stepdown logistic regression analysis was performed on 219 IDUs using the five potential predictor variables: age, education, intravenous drug use, needle sharing, and sexual practice. This showed that sexual practice was not an independent predictor variable of HIV positivity. Hence a second logistic regression analysis was carried out using age as a continuous variable and education, intravenous drug use, and needle sharing as dichotomous variables. Data on all 242 drug users were available for this analysis. Only education and needle sharing emerged as significant predictor variables. The odds ratios were: education 0.250 and needle sharing 2.46.