HIV testing and counseling has been associated with reductions in risk behaviors in some populations. This study examined whether involvement in a long-term study, including exposure to repeated HIV testing and counseling, was associated with increased condom use among injection drug users (IDUs) through a retrospective analysis of an IDU cohort from Baltimore, MD, during 1994-1998. Eligibility included being aged 18 years or older, injecting within 10 years and not having initiated antiretroviral therapy. A logistic model of high versus low risk, based on condom use, was used. Of 322 eligible IDUs, most were male (66%) and African-American (94%). No significant change in the odds of inconsistent condom use was observed with continued study exposure. Condom use remained low, indicating a need for interventions to reduce sexual risks and HIV transmission in this population.