Modelling multi-site transmission of the human papillomavirus and its impact on vaccination effectiveness

Epidemics. 2017 Dec;21:80-87. doi: 10.1016/j.epidem.2017.08.001. Epub 2017 Aug 25.

Abstract

Objective: Previous HPV models have only included genital transmission, when evidence suggests that transmission between several anatomical sites occurs. We compared model predictions of population-level HPV vaccination effectiveness against genital HPV16 infection in women, using a 1) uni-site (genital site), and a 2) multi-site model (genital and one extragenital site).

Methods: We developed a uni-site and a multi-site deterministic HPV transmission model, assuming natural immunity was either site-specific or systemic. Both models were calibrated to genital HPV16 prevalence (5%-7.5%), whilst the multi-site model was calibrated to HPV16 prevalence representative of oral (0%-1%) and anal (1%-7.5%) sites. For each model, we identified 2500 parameter sets that fit endemic genital and extragenital prevalences within pre-specified target ranges. In the Base-case analysis, vaccination was girls-only with 40% coverage. Vaccine efficacy was 100% for all sites with lifetime protection. The outcome was the relative reduction in genital HPV16 prevalence among women at post-vaccination equilibrium (RRprev). RRprev was stratified by extragenital prevalence pre-vaccination.

Results: Under assumptions of site-specific immunity, RRprev with the multi-site model was generally greater than with the uni-site model. Differences between the uni-site and multi-site models were greater when transmission from the extragenital site to the genital site was high. Under assumptions of systemic immunity, the multi-site and uni-site models yielded similar RRprev in the scenario without immunity after extragenital infection. In the scenario with systemic immunity after extragenital infection, the multi-site model yielded lower predictions of RRprev than the uni-site model.

Conclusions: Modelling genital-site only transmission may overestimate vaccination impact if extragenital infections contribute to systemic natural immunity or underestimate vaccination impact if a high proportion of genital infections originate from extragenital infections. Under current understanding of heterosexual HPV transmission and immunity, a substantial bias from using uni-site models in predicting vaccination effectiveness against genital HPV infection is unlikely to occur.

Keywords: Human papillomavirus; Mathematical model; Vaccination.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Cost-Benefit Analysis
  • Female
  • Human papillomavirus 16*
  • Humans
  • Models, Theoretical
  • Papillomavirus Infections / epidemiology
  • Papillomavirus Infections / prevention & control*
  • Papillomavirus Infections / transmission
  • Papillomavirus Vaccines*
  • Prevalence
  • Vaccination
  • Young Adult

Substances

  • Papillomavirus Vaccines