Abstract
Certain key populations have a high risk of hepatitis C virus (HCV) reinfection, which includes men who have sex with men (MSM) with HIV who continue to engage in behaviors associated with HCV acquisition following clearance. Among MSM with HIV, we aimed to identify longitudinal sexual behavior patterns and estimate reinfection risk within identified patterns. MSM with HIV from the longitudinal, prospective Dutch MOSAIC study (2009-2018) at risk for HCV reinfection were included. Follow-up started following HCV clearance. Risk behavior was assessed using the HCV-MOSAIC score (range = 0.0-7.0), where ≥2 indicates high risk of reinfection. Classes were inferred from the mean HCV-MOSAIC score over time using a latent process mixed-effects model with the covariates age, group sex and casual partnership. The association between classes and HCV reinfection risk was assessed using a joint survival model. In total, 123 MSM were included with a median follow-up of 2.7 years [interquartile range (IQR) = 1.2-4.7]. Two classes were identified: one with mostly lower (C1, n = 67) and one with mostly higher risk behavior (C2, n = 56). During follow-up, both classes had considerable variation in HCV-MOSAIC scores (C1, median = 1.1, IQR = 0.0-2.1 and C2, median = 3.0, IQR = 2.0-3.5). HCV reinfection probability was similar between both classes at year 3 of follow-up [C1, 17%, 95% confidence interval (CI) = 11%-35% and C2, 18%, 95%CI = 15%-47%], but became higher in C2 than C1 at year 5 (C1, 22%, 95%CI = 13%-39% and C2, 37%, 95%CI = 28%-69%). The variation in risk over time suggests that behavioral assessment is continually needed for early testing, treatment and offering behavioral inventions.