To determine the incidence of and risk factors for human T-lymphotropic virus, type II (HTLV-II) seroconversion among injecting drug users (IDUs), specimens from IDUs recruited into the ALIVE Study in 1988/1989 were assayed at baseline for antibody to HTLV with use of enzyme immunoassay and Western blot. Participants were monitored semiannually with venipuncture and interviews. In 1992, the most recent sera of HTLV-negative participants were tested for HTLV with use of enzyme immunoassay and confirmed and typed by Western blot. For positive cases, assays were then performed for all intervening visits to determine the calendar time of seroconversion. Incidence rates were estimated using person-time. Risk factor analysis used a nested case-control design, with up to seven controls per case matched by time of study entry and duration of follow-up. At baseline, 251 HTLV-positive, 22 indeterminate, and 2,574 HTLV-seronegative IDUs were identified. Follow-up of the seronegative IDUs identified 38 seroconverters (all HTLV-II) over 5,813.6 person-years, for a rate of 0.7/100 person-years. Median lag time for seroconversion was 6.8 months. Factors associated with HTLV-II seroconversion included a specific needle-sharing practice called "backloading" within the previous 6 months [odds ratio (OR) = 6.52; 95% confidence interval (CI) = 1.94-21.95] and a baseline history of receiving money for sex (OR = 3.36; 95% CI = 1.32-8.57). Of those with more than one sex partner in the past 6 months, women were more likely than men to seroconvert (OR = 5.77; 95% CI = 1.33-25.05).(ABSTRACT TRUNCATED AT 250 WORDS)