KwaZulu-Natal is heavily affected by HPV and HIV. As these viruses are potentially important drivers in each other’s epidemic, interventions targeting one of them might have spillover benefits to the other. Mathematical simulation models can be used to explore the dynamics of HPV and HIV in this high endemic area. These explorations were performed as part of a comparative modeling exercise of three mathematical models of the cervical cancer working group of the Cancer Intervention and Surveillance Modeling Network (CISNET).
One of the models included in the comparative modeling exercise is the microsimulation model STDSIM. We used a previous application of STDSIM for HIV in KwaZulu-Natal, and incorporated the transmission of HPV. We calibrated natural history parameters of HPV, and co-factor effects between HIV and HPV to reproduce observed relationships between infections. We determined the impact of HPV vaccination (90% efficacy, 60% coverage) and changing eligibility for HIV treatment on HPV and HIV prevalence over a period of 20 years since the implementation of HPV vaccination in our model.
We were able to reproduce the observed HIV prevalence, the proportions of HIV positive and negative women with HPV, and the association between HPV and HIV. Preliminary results from the STDSIM model show that offering HPV vaccination reduces HPV prevalence from 26.8% to 26.0% and HIV prevalence from 22.1% to 20.8% after 20 years. However, spillover effects of HPV vaccination in our model diminished under guidelines of ART for all HIV-infected people. Finally, our analyses show that HIV interventions substantially impact HPV burden, as ART for all HIV-infected people is expected to decrease HPV prevalence from 26.8% to 23.5% after 20 years.
Our model suggests that interventions targeting HIV also impact HPV and thus the cervical cancer epidemic. In contrast, the impact of HPV vaccination on HIV incidence is limited. Future research should explore different intervention combinations to optimize HPV and HIV prevention in high endemic settings, and identify structural differences between the mathematical models.