Description:
<jats:title>Abstract</jats:title><jats:p> <jats:bold>Aims.</jats:bold> To determine whether young or recent‐onset injecting drug users (IDUs) are at increased risk of HIV infection, and to compare trends in risk behaviours and HIV incidence among subgroups of IDUs. <jats:bold>Design.</jats:bold> Associations of age and onset of injecting with HIV prevalence and injecting risk behaviours were determined among current IDUs who entered an Amsterdam cohort between 1989 and 1995, and compared with current IDUs recruited in two street surveys in 1990 and 1993. In the cohort, trends from 1986 through 1995 in injecting risk behaviour (as reported at entry) and in HIV seroconversion (among current IDUs during follow‐up) were determined using logistic and Poisson regression. <jats:bold>Findings.</jats:bold> Sizable portions of IDUs were young (≤ 25 years: 14–30%) or had recently started injecting (within the last 3 years: 17–21%). Between 37% and 50% of young IDUs recently started injecting. HIV prevalence was 12–24% among young IDUs, and 5–12% among recent‐onset IDUs. Young IDUs more often reported current borrowing than older IDUs. Borrowing declined significantly in all subgroups, except young IDUs. The crude HIV incidence declined from 18/100 person‐years (PY) in 1986 to 5/100 PY in 1995. Trends in HIV incidence were not significantly different for subgroups of age or onset of injecting. In a multivariate analysis, recent onset of injecting was an independent predictor (p = 0.04) for HIV seroconversion, but age was not (p = 0.68). <jats:bold>Conclusions.</jats:bold> Important proportions of drug users are young or have recently started injecting. HIV prevalence is relatively high among young and recent‐onset IDUs. Recent onset of injection is an independent predictor for HIV seroconversion. Our observations may be explained by non‐random patterns of borrowing used needles/syringes. Recent‐onset IDUs should be approached more actively for HIV counselling and testing.</jats:p>