Validating the predicted impact of HPV vaccination on HPV prevalence, cervical lesions, and cervical cancer: A systematic review of population level data and modelling studies.

BACKGROUND

We compared model predictions with independently published primary data from population-based studies on the impact of HPV vaccination on HPV prevalence, cervical cancer and its precursors.

METHODS

We searched Cochrane Library, EMBASE, MEDLINE, Web of Science for studies concerning high-income countries published between 2005 to June 2, 2023. Relative risk (RR) for HPV-related outcomes comparing the pre-vaccination and post-vaccination periods were collected from observational and modelling studies. The relationship between vaccination coverage and observed relative reductions was determined using meta-regressions, and we compared model prediction to observations.

FINDINGS

We identified a total of 5649 potential articles, of which one systematic review, 14 observational studies and 32 modelling studies met our inclusion criteria. A clear relation was found between the RR of HPV diseases related outcomes in the pre- versus post-vaccination era and the vaccination coverage, with 23 out of 28 data points and 19 out of 20 data points showing significant reductions in HPV prevalence and CIN2+ prevalence respectively. Around 67 % (n/N = 12/18) of model predictions were more optimistic on HPV prevalence reductions compared to the 95 % CI of the meta-regression derived from observational studies. For CIN2+ lesions, 48 % (n/N = 31/64) of model predictions for CIN2+ outcomes fell within the 95 % CI.

INTERPRETATION

Model predictions and observational data agree that HPV vaccination can have a substantial impact on HPV related outcomes on a population level. Despite large heterogeneity in observational data and modelling studies, it is particularly encouraging that model predictions on the impact of HPV vaccination on CIN2+ model lesions align with observational studies.

FUNDING

Ontario Health (formerly known as Cancer Care Ontario).