Human Papillomavirus Type Replacement Following the Implementation of HPV Vaccination

Infection with high-risk human papillomavirus (HPV) is a necessary cause of cervical cancers and an associated cause of several other anogenital and oropharyngeal cancers. The beginning of the era of prophylactic human papillomavirus vaccination, via the implementation of efficacious first-generation vaccines targeting the two most high-risk oncogenic HPV types 16 and 18, offers the opportunity for the control of HPV16/18 infection and elimination of HPV16/18 associated cancers. The first-generation vaccines, however, target only 2 of 12 IARC classified high risk oncogenic HPV types. Previously vaccination of a targeted subset of strains of a pathogen has in some instances led to an increase in the non-targeted strains in a phenomenon known as serotype (or genotype) replacement. Therefore, there has been concern that targeted HPV vaccination might induce HPV genotype replacement by the non-vaccine targeted high risk HPV types, potentially undermining the impact of HPV vaccination.

In this dissertation, we have evaluated and compared the sustainability of vaccine-induced neutralising and cross-neutralising antibody response of the two first generation vaccines, the bivalent and the quadrivalent vaccine, by following up two cohorts of vaccinated Finnish trial participants in the population-representative Finnish maternity cohort serum biobank.

We have then evaluated the degree of HPV16/18 niche clearance, both the direct and indirect impact of vaccination, via active and passive follow-up of the population-based Finnish community randomised trial of HPV vaccination strategy with moderate vaccination coverage (gender-neutral vaccination versus girls only HPV vaccination using the bivalent HPV vaccine).

Subsequently we have evaluated the occurrence of the non-vaccine HPV types among both HPV vaccinated and unvaccinated women via the same active and passive follow-up of the community randomised trial, to assess whether the non- vaccine types increased in occurrence and took advantage of the cleared partially cleared HPV16/18 niche (measured as decreased occurrence) in a manner indicative of type replacement. We further investigated the same, among the high-risk taking core group, where transmission dynamics are greater and early indications of type replacement may be observed.

We found that both the bivalent and the quadrivalent vaccines induce a sustainable neutralising antibody response against HPV16 and HPV18 among women vaccinated with three doses of the respective vaccine. However, among the quadrivalent vaccine recipients 15% of the women were found to be seronegative to HPV18 with no detectable HPV18 neutralizing antibodies, whilst among the bivalent vaccine recipients 100% had seroconverted to HPV18. Notable differences in the cross-neutralising antibodies response were observed between the two vaccinated cohorts, with seroconversion to HPV31, 33, 45, 52, and 58 being higher among the bivalent vaccine recipients in comparison to the quadrivalent vaccine recipients.

Following the community randomised trial, we observed the greatest HPV16/18 niche clearance among the unvaccinated residents of the gender-neural trial arm communities. Significant reduction of both HPV16 and HPV18 prevalence was observed after gender-neutral HPV vaccination. No HPV16 niche clearance in unvaccinated residents was observed in the girls-only vaccination arm communities. Among the HPV vaccinated participants, the prevalence of vaccine targeted HPV16 and 18 was almost negligible, whilst the prevalence of HPV31, 33 and 45 were also markedly reduced in comparison to the HBV vaccinated controls.

When evaluating the occurrence of the non-vaccine HPV types via the active follow-up of the community randomised trial we observed consistent increases in HPV51, and some indications of increased HPV58 among the participants from the interventions. However, these increases generally stemmed from the older birth cohort. Additionally, some sporadic inconsistent increases in HPV39 and 66 were observed. However, during the passive follow up among the unvaccinated women resident in the trial communities pre- and post- vaccination era, none of these findings were replicated. We observed an increase post-vaccination in the seroprevalence (cumulative incidence) of HPV68 (a type not measured in the active follow-up) among the residents from the girls-only intervention communities. However, a similar HPV68 seroprevalence increase was also observed among the unvaccinated women from the control arm communities, suggesting that this increase was unlikely to be due to vaccine-induced type replacement.

Overall, we observed that both first generation vaccines induce a high level of sustainable HPV16/18 neutralising antibodies over time among women, whilst the bivalent vaccine provided a higher prevalence of cross-protective neutralising antibodies as compared to the quadrivalent vaccine. We further found significant niche clearance (also HPV16 niche clearance) following gender-neutral vaccination even with moderate vaccination coverage. However, despite this observed niche clearance, we found no clear, decisive signs of increased non-vaccine occurrence which could irrefutably be due to vaccine-induced type replacement. These findings were specific to the study setting (vaccination coverage, the vaccine used and the duration of follow-up), thereby, continued surveillance of the non-vaccine HPV types will remain crucial in the future.