Background: Injection drug users (IDU) are at high risk for HIV and hepatitis C virus (HCV) infections. Here we report the HIV and HCV incidence and temporal changes in risk behavior over a period of 18 years in a prospective cohort of drug users (DU).

Methods: DU participating in the ongoing Amsterdam Cohort Study return every 4–6 months since 1985. Every visit participants complete a standardized questionnaire and blood is drawn. We prospectively tested for HIV antibodies and retrospectively for HCV antibodies.

HIV transmission routes were assessed using detailed standardized questionnaires. Trends in risk behavior were evaluated using the Generalized Estimating Equations method.

To compare HIV and HCV incidence, a smooth trend was assumed for both incidence curves over calendar time. If the trends have the same pattern, then the difference between the curves is a constant on a logarithmic scale.

Results: 1315 DU were at risk for HIV. 93 DU seroconverted during follow up. The HIV incidence was 7 per 100/person years (PY) in 1986 and 0-0.5/100 PY after 1999. Compared with the period before 1996, heterosexual risk behavior played the most important role in HIV transmission among DU seroconverting after 1996: OR 15.6 (95% CI: 2.6-94.6). Over time, both injecting and borrowing needles has significantly declined. Sexual risk behavior at follow-up visits decreased before 1996, but it did not decrease further after this period.

Among 960 IDU who ever injected (ever-IDU) 82.2% had HCV antibodies at study entry and 58 seroconverted for HCV during follow up. 26.2% of ever-IDU were HIV positive at study entry and 89 seroconverted for HIV.

The HCV incidence rate peaked in 1989 at 27.5/100 PY and dropped to 3/100 PY from 1996 onwards. The HCV incidence is estimated to be 4.4 times the HIV incidence. Testing the shape of both incidence curves revealed no statistically significant differences in the incidence pattern (fig).

Conclusion: The incidence rates of HIV and HCV were very high in the early 1980’s. Both HIV and HCV incidence among DU from the Amsterdam Cohort Study have declined since 1985. We found a similar pattern of the HCV and HIV incidence over nearly two decades; the decline is accompanied by a reduction in injecting risk behavior and occurred despite continued sexual risk behavior. At present, new HIV seroconversions are mainly related to unprotected heterosexual contacts, indicating that prevention programs for DU should also pay specific attention to the importance of safe sex practices.

The HCV incidence was on average 4 to 5 times the HIV incidence, most likely due to the higher background prevalence of HCV when compared to HIV, the higher efficiency of HCV transmission, and more potential sources of HCV exposure in the injection setting (like cookers). HIV risk reduction measures including needle exchange programs (NEP) appear to have had an important impact on the sharp decline of both HIV and HCV incidence.

Figure: Observed and fitted HIV (red) and HCV (black) incidence and 95% confidence intervals.