Model-based estimation of viral transmissibility and infection-induced resistance from the age-dependent prevalence of infection for 14 high-risk types of human papillomavirus

Abstract

Viral transmissibility and natural resistance to infection are key determinants in assessing the population impact of human papillomavirus (HPV) vaccination, yet information on these parameters is scarce. Using data from 2 large-scale surveys on sexual behavior in the Netherlands (carried out in 2005-2006), the authors employed a Bayesian framework to fit a transmission model to the cross-sectional age-dependent prevalence of HPV DNA in cervical smears (data collected in 1992-2002), assuming that the prevaccine situation reflected an endemic equilibrium, and calculated type-specific estimates of transmissibility and infection-induced resistance. The posterior median transmission probability per heterosexual partnership covered a range of 0.43-0.94 among the 14 high-risk types of HPV. The transmission probability of HPV-16 was estimated at 0.80 (95% posterior interval: 0.60, 0.99) and that of HPV-18 at 0.93 (95% posterior interval: 0.72, 1). The model predicted that the decrease in HPV prevalence with age could not solely be explained by sexual activity and screening but also by resistance to reinfection, which is lost at a rate of 0.014-0.047 (1%-5%) per year. These results support the notion that HPV infection is highly transmissible, and they suggest a gradual loss of type-specific immunity over time. Because high transmission potential is associated with a low impact of herd immunity, extensive vaccination coverage will be required to substantially reduce cervical cancer incidence.