A Meta-analysis of the Association Between Needle Exchange Programs and HIV Seroconversion Among Injection Drug Users

We assessed the association between different levels of needle exchange program (NEP) use and human immunodeficiency virus (HIV) seroconversion in the injecting drug user (IDU) population using meta-analysis to aggregate risk estimates from any reported cohort studies or randomized controlled trials (RCTs) in the literature. We searched the literature for articles published from January 1990 to August 2014 using Medical Subject Headings and other terms from MEDLINE® (using Ovid), Embase, ProQuest, the International Aids Society Abstract database, and the European AIDS clinical society database, and the European Conferences Abstract Archive. Articles were included if data were reported from an original study; the study was a prospective cohort or RCT design; estimates for seroconversion (hazard ratios [HRs]) for drug-users with different levels of NEP-use, as well as variance data, or the information to calculate these were reported; risk estimates were adjusted for unstable housing, risky sexual behaviors, frequency of injections, cocaine use, and risky needle sharing practices; and the study was published between January 1990 and August 2014. Information abstracted was general study information (i.e., study name, authors, publication year, study site, sample size, length of follow-up, and follow-up intervals, incentives to IDUs for improving compliance/enrollment), outcomes variable measures (seroconversion HR estimates, variance figures, and factors adjusted for), description of the study population (inclusion and exclusion criteria and definitions of comparison groups), type of intervention (NEP program), statistical methods used, and sub-group information. Two prospective cohort studies with a total of 3,172 IDUs were eligible for inclusion. Comparison groups had different levels of NEP-use (e.g., daily use vs. non-daily use and NEP-users vs. non-users) from fixed-site NEPs. Our Q-statistic was insignificant with a p-value of 0.401 while the I2 value was 0.0%. A random-effects model was used to aggregate the estimates, and we found an overall significant positive association between NEP-use and HIV seroconversion with an HR estimate of 1.59 (95% confidence interval [CI]: 1.2 to 2.1). According to our results, higher usage of NEPs is associated with a higher risk of HIV seroconversion in the IDU population. The observed association aligns with previous findings of NEP programs being inadequate for HIV control in IDUs. Further research on the topic needs to be done including studies on different NEP designs and how they can be made more effective by combining with other strategies, including the study of IDU characteristics which make them more likely to use safe syringes when they inject.