Background:  India has approximately 1.1 million injection drug users (IDU) with HIV prevalence ranging from 1 to 64%; few estimates of incidence exist. We recently reported an estimated HIV incidence using the BED capture enzyme immunoassay (CEIA) of 4.0% per year using baseline samples of IDU enrolled in a longitudinal study in Chennai, India. The objectives of this analysis were to examine HIV incidence longitudinally, and changes in risk behaviors after HIV testing and risk-reduction counseling.

Methods:  A convenience sample of 1172 IDU (293 HIV⁺) from Chennai was recruited from March 2005 to June 2006. HIV incidence and changes in risk behaviors were examined in the first year after enrollment in 628 of 879 HIV^– with ≥1 semi-annual follow-up visit. An additional 82 of 293 HIV⁺ were also followed. Conditional logistic regression was used to examine visit-to-visit changes in risk behaviors.

Results:  Over 604 person-years, 4 HIV seroconversions were observed (HIV incidence = 0.66/100 person-years, 95%CI 0.18 to 1.7). Among HIV^–, reported injection drug use declined significantly from baseline to 6 months (odds ratio [OR] of injecting in the last month, 0.02; 95%CI 0.01 to 0.04). Similar declines were observed for non-injection drug use (OR 0.3, 95%CI 0.2 to 0.5). However, among persistent injectors, there was less of a decline in reported needle sharing (OR 0.7, 95%CI 0.3 to 1.7). In terms of sexual risk behavior, persons were less likely to report sex with a non-regular partner (OR 0.7, 95%CI 0.5 to 0.99) and more likely to report some condom use with these partners (OR 2.2, 95%CI 0.8 to 7.0). Less substantial changes in condom use with regular partners and exchange sex were observed. Similar trends were observed among HIV⁺ and at 1 year. Those who eventually ceased injection were less likely to be frequent injectors, use non-injection drugs, be educated, or have been incarcerated and were more likely to be drinking alcohol.

Conclusions:  The low observed HIV incidence in this population of IDU coupled with the dramatic reductions in risk behavior, particularly drug injection itself, are very encouraging and likely reflect the influence of risk-reduction counseling. However, we cannot rule out the possibility that declines in injection drug use reflect decreased availability of drugs due to police crackdowns and increased drug seizures beginning in India in 2006. Additional follow-up of this cohort may provide further insight into this issue.