Objective

To find the prevalence of HIV infection and risk behaviors among injecting drug users (IDUs) in Karachi, Pakistan. Design

A cross-sectional study of IDUs conducted in Karachi, Pakistan from February through June 1996. Results

Of the 242 IDUs, 11 (4%) refused HIV testing. One (0.4%; 95% confidence interval (CI)=0.37–0.48%) was HIV positive. All subjects were male. Over the past 6 months 47% had engaged in receptive needle sharing, 38% had perceived a change in their social network, 22% had had sexual intercourse, of whom only 7% always used condoms, and none had washed their needles with bleach. Younger age (28 vs. 31 years; p=0.01), younger age at first injection (25 vs. 28 years; p=0.001), fewer years of schooling (3 vs. 5 years; p=0.001), lower monthly income ($70 vs. $80; p=0.03), inhaling fumes of heroin from a foil in the year before injecting (OR=4.8; CI=2.2–10.3), injecting first time with heroin (OR=3.6; CI=1.2–12.6), having a temporary job (OR=2.5; CI=1.2–5.2), and a perceived change in one's social network (OR=4.4; CI=2.4–7.9) were all associated with receptive needle sharing. IDUs who knew about HIV spread through contaminated needles were less likely to share (OR=0.4; CI 0.2–0.8). In the final logistic regression model receptive needle sharing was associated with inhaling of fumes of heroin on a foil in the year prior to injecting (adjusted OR=5.6; CI=2.6–12.0), a perceived change in one's social network (adjusted OR=4.0; CI=2.2–7.4), and inversely associated with age at first time of injection (β=−0.07; p=0.002). Conclusion

Background HIV prevalence was low among IDUs in Karachi despite high-risk behavior in 1996. In order to control HIV transmission among IDUs in Pakistan, continual HIV surveillance with well-coordinated and effective HIV risk reduction, and drug demand reduction programs need to be implemented among drug users.