Background:  Since 2000 outbreaks of acute hepatitis C virus (HCV) infection among HIV-infected men having sex with men (MSM) have been described. However, recent phylogenetic analysis indicated that the spread of HCV started earlier, in 1996. Our study aimed to estimate the incidence of HCV in 3 calendar periods in HIV-infected MSM in the last 20 years using data from 12 cohorts of the CASCADE collaboration.

Methods:  CASCADE is a collaboration of cohorts, which collect date of HIV seroconversion. Systematic HCV data collection in each individual cohort started at different calendar years, often several years after cohort initiation because the first commercial HCV test became available in 1991. Incidence of HCV was estimated with methods for interval censored data in MSM. To explore possible selection bias (testing might be selective before the date of systematic HCV data collection), 4 different strategies were used with respect to inclusion of individuals who were only tested before that date (e.g. including or ignoring HCV test results before systematic data collection). Incidence estimates were obtained separately for 3calendar periods: 1990 to 1995, 1995 to 2000, and 2000 to 2007. Only individuals who were at risk in the respective periods contributed to the analysis.

Results:  Of 3020 MSM, 222 (7%) had a positive HCV result and 2888 (96%) had a last negative HCV result, and 3% had both results. Among those who tested HCV positive, 14% had their first positive test before 1995 and 27%, between 1995 to 2000. In the period 1990 to 1995 HCV incidence was estimated 3.8 to 7.0 per 1000 patient-years, depending on the strategy applied. In the period 1995 to 2000, HCV incidence already substantially increased and was estimated to be 11.2 to 15.7 per1000 patient-years, whereas, in the period 2000 to 2007 HCV incidence remained high at 8.6 to 14.4 per 1000 person-years.

Conclusions:  Regardless of the data handling strategy applied, our data supports the phylogenetic findings that HCV incidence increased substantially among HIV-infected MSM in the mid-1990s. Using interval censored methods, incidences can be estimated even when the majority of the individuals either have only a last negative or a first positive test result. Therefore, large multi-centre studies with sufficient power and timely supply of data are necessary. The incidence estimates might help to earlier identify changes in the spread of important co-infections and start raising awareness and routine testing in time.