Human papillomavirus vaccine effect against human papillomavirus infection in Rwanda: evidence from repeated cross-sectional cervical-cell-based surveys

Summary Background Rwanda was the first African country to implement national human papillomavirus (HPV) vaccination (against types HPV6, 11, 16, and 18). In 2011, a school-based catch-up programme was initiated to vaccinate girls aged younger than 15 years but it also reached older girls in schools. We aimed to estimate the population-level effect of HPV vaccination on HPV prevalence. Methods Cross-sectional surveys were done between July, 2013, and April, 2014 (baseline), and between March, 2019, and December, 2020 (repeat), in sexually active women aged 17–29 years at health centres in the Nyarugenge District of Kigali, Rwanda. HPV prevalence was assessed in cervical cell samples collected by a health-care worker in PreservCyt solution (Cytyc, Boxbourough, MA, USA) and tested using a general primer (GP5+ or GP6+)-mediated PCR. Adjusted overall, total, and indirect (herd immunity) vaccine effectiveness was computed as the percentage of HPV detection among all women and among unvaccinated women. Findings 1501 participants completed the baseline survey and 1639 completed the repeat survey. HPV vaccine-type prevalence in participants aged 17–29 years decreased from 12% (173 of 1501) in the baseline survey to 5% (89 of 1639) in the repeat survey, with an adjusted overall vaccine effectiveness of 47% (95% CI 31 to 60) and an adjusted indirect vaccine effectiveness of 32% (9 to 49). Among participants aged 17–23 years, who were eligible for catch-up vaccination, the adjusted overall vaccine effectiveness was 52% (35 to 65) and the adjusted indirect vaccine effectiveness was 36% (8 to 55), with important heterogeneity according to education (overall vaccine effectiveness was 68% [51 to 79] in participants with ≥6 years of school completed and 16% [–34 to 47] in those with <6 years) and HIV status (overall vaccine effectiveness was 55% [36 to 69] for HIV-negative participants and 24% [–62 to 64] for HIV-positive participants). Interpretation In Rwanda, the prevalence of vaccine-targeted HPV types has been significantly decreased by the HPV vaccine programme, most notably in women who were attending school during the catch-up programme in 2011. HPV vaccine coverage and population-level impact is expected to increase in future cohorts who are eligible for routine HPV vaccination at age 12 years. Funding Bill & Melinda Gates Foundation.

Methods S1: Methods adopted to assess the lifetime risk of cervical cancer.
Derivation of the formula for estimating the contribution to the cumulative incidence of cervical cancer for a particular age group taking account of the competing risk of dying from any cause before being diagnosed with cervical cancer.
For a given cohort of women, the expected cumulative number of cervical cancers between age 15 and 79 years in the absence of vaccination was calculated by using 5-year age group-specific cervical cancer incidence rate projections from the Global Cancer Observatory (GLOBOCAN) (http://gco.iar.fr) project 1,2 and 5-year age group-specific mortality rates were obtained from United Nations, Department of Economic and Social Affairs, Population Division. 3 Let 0 be the age group 15-19 years, …, 12 be the age group 75-79 years. Let us denote by and respectively cervical cancer incidence rate and mortality rate, both assumed to be piecewise constant functions on each five-year age group. That is, for = 0, … , 12, , and , denote respectively the constant -specific cervical cancer incidence and mortality rates. Let us also designate by and respectively the lower and upper bound of age interval , with for > 0, = −1 . Then, the survival free from event (i.e., cervical cancer diagnosis or death) by age for a cohort of women aged 15 at the start of followup can be calculated according to the formula: Now, the cumulative incidence of cervical cancer by age can be calculated as: We then have: Based on the assumption that the age group-specific incidence rates were stable across birth cohorts, the cumulative incidence between ages 15 and 79 was then obtained by summing the age group-specific contributions:CumI 15−79 = ∑ CumI i 12 =0 . Finally, the cumulative number of cervical cancers was calculated by multiplying this cumulative incidence by the size of the cohort at age 15 years, also estimated from the GLOBOCAN data. Uncertainty intervals were estimated with Monte-Carlo simulation combining uncertainty about cervical cancer incidence as reported in GLOBOCAN 2020 and HPV distribution in cervical cancers as reported by Guan et al. 4 Our estimates were obtained based on Monte Carlo simulation with 100 repetitions as this number of repetitions provides stable estimates. Of note, no burn-in is needed for the applied Monte Carlo simulation since this used to assess the uncertainty of our estimates rather than fitting our model to target statistics.
Of note, we recently validated GLOBOCAN estimates for Rwanda 5 using an independent method to quantify the cervical cancer incidence based on age-specific population-based HPV prevalence. The independent estimates are consistent with those reported in GLOBOCAN.
Finally, the procedures adopted to estimate the expected lifetime risk (LTR) of cervical cancer and the annual age-standardised incidence rate, are documented at the following website: https://iarc-miarc.gitlab.io/methis/methis.atlas/reference/atlas_ui.html#references     HPV6, 11, 16, and 18)     Note that, all mentally and physically competent women were eligible for the study, regardless of their marital status, with exception of those who were known to be pregnant. All participants were residents of Nyarugenge District, Kigali and were recruited in two ways: 1) A population-based invitation of women aged 18-69 years residing in pre-defined areas of Nyarugenge District was made through community workers and community meetings. Women were invited to attend the Reproductive Health Department of the district hospital (Muhima). Participation rates could not be estimated due to the uncertainty on the number of women who had been reached by the information campaign and because exhaustive lists of the population resident in the catchment areas of health centres are not available. 2) An additional opportunistic approach was used to recruit women aged ≤29 years who were spontaneously attending Muhima hospital or health centers in the Nyarugenge District. Principal reasons for consultation were family planning, childhood vaccination, post-natal care services and HIV-related services. Study procedures were undertaken in the centre where the woman was consulting and very few refusals were recorded. Following signature of an informed consent form, a structured risk questionnaire was administered to all study participants, and data transferred electronically to IARC via a specifically designed on-line webpage/application." 12 RWA HPV prevalence survey -Informed Consent Version 2.3

INFORMED CONSENT FORM
The Ministry of Health in Rwanda is working with the International Agency for Research on Cancer to conduct a study on the prevalence of human papillomavirus (a virus that is commonly found in the genital tract) in Rwanda.
It will help to assess the impact of the HPV vaccination program to prevent cervical cancer in the future. You are being invited to participate in this study, which has been approved by the Rwandan National Ethical Committee (contacts: Dr. ………………………. at ………….).

STUDY PROCEDURES
A trained health worker will place a speculum in your vagina and take a specimen of cells from the surface of your cervix using a soft brush. In the event that any gynecological abnormalities are seen during the sample taking, you will be referred for appropriate medical evaluation and treatment.
You will be asked questions about your background, education, pregnancies, family planning, health, smoking and lifestyle. Some of the questions will be personal. You can refuse to answer any question that bothers you.

TEST AND TEST RESULTS
The specimen of cells from your cervix will be used to detect human papillomavirus and other factors potentially associated with cervical cancer. The tests will be purely research tests, with the aim to improve cervical cancer prevention in the future.

BENEFITS AND RISKS
The HPV test is not informative for your clinical care as you are less than 30 years, and will be used for research purposes only.
The examination to collect these specimens is a routine procedure and considered safe. No complications or discomfort is expected except for the usual discomfort that you may feel from a routine gynecological exam.

CONFIDENTIALITY
All the information you provide and the results from your biological specimens will be kept in strict confidence. Your name will not be associated with the results of this study. Information from the interview and your specimens will be identified by a study number only. Data collected will only be used to develop statistical summaries to be presented in published reports.

PARTICIPATION
11. CAN YOU READ ANY LANGUAGE?