Variability of High-Risk Human Papillomavirus and Associated Factors among Women in Sub-Saharan Africa: A Systematic Review and Meta-Analysis

Background: Sub-Saharan Africa (SSA) carries the highest burden of high-risk human papillomavirus (HR-HPV) in the world, driven by, and together with, HIV infection. This systematic review aimed to identify HR-HPV genotypes and their associated factors among women in SSA. Methods: A systematic review and meta-analysis of studies conducted in SSA on HR-HPV was conducted. Standard electronic databases were searched. R software version 3.6.0 was used for meta-analysis, with p < 0.05 considered statistically significant. Results: We included 28 articles with a total of 22,652 participants. The overall pooled prevalence of HR-HPV genotypes was 55.13%, albeit high heterogeneity between studies. The overall pooled prevalence of HR-HPV genotypes in HIV-positive individuals was 75.51%, compared to 52.97% in HIV-negatives (OR = 4.68 (0.71–30.76)). HPV 16 (18%), 35 (10.12%), 52 (9.98%), 18 (9.7%) and 45 (6.82%) genotypes were the most prevalent. Twelve studies identified the most frequently reported risk factors associated with HR-HPV, with HIV infection (66.66%), multiple sexual partners (41.66%) and young age (41.66%) being the most reported risk factors. Conclusions: The combined prevalence of HR-HPV genotypes among women in general and HIV-infected women in particular remains high in SSA. The presence of several genotypes not covered by the vaccine is remarkable and suggests the need for revision of current vaccination policies to prevent HR-HPV infections.


Background
Human papillomaviruses (HPVs) are viruses responsible for the occurrence of lesions in the anogenital and oropharyngeal regions, which can result in condyloma acuminata, warts that can develop into precancerous lesions and then some years later into cervical cancer [1,2]. Of note, HPV is the most common sexually transmitted infection (STI) in the world, with 660 million people infected according to the World Health Organization [3,4]. Even though the prevalence of cervical cancer varies across geographical settings in the world, in sub-Saharan Africa (SSA) it represents the leading cause of cancer deaths in women [1,5]. Moreover, resource-limited settings (RLS), mainly within SSA, account for cancer", "human papillomavirus", "high-risk human papillomavirus", "HPV", "HPV-HR", "papillomavirus", "HIV", "AIDS", "human immunodeficiency virus", "acquired immune deficiency syndrome". In addition, we also included in our search strategy all countries in sub-Saharan Africa, as defined by the United Nations Statistics Division (UNSD) [24]. The search was developed using the Boolean operators "AND" and "OR". Details of the search are presented in Figure 1. The databases were searched from 1996 to 2021 for studies published in English and French. The reference lists of included articles were also manually searched.

Data Sources and Search Strategy
A comprehensive search strategy was applied to PubMed/Medline, Science Direct, African Journals Online, Academic Medical Education and Google Scholar databases to allow for exhaustive identification of relevant studies, using the search terms: "cervical cancer", "human papillomavirus", "high-risk human papillomavirus", "HPV", "HPV-HR", "papillomavirus", "HIV", "AIDS", "human immunodeficiency virus", "acquired immune deficiency syndrome". In addition, we also included in our search strategy all countries in sub-Saharan Africa, as defined by the United Nations Statistics Division (UNSD) [24]. The search was developed using the Boolean operators "AND" and "OR". Details of the search are presented in Figure 1. The databases were searched from 1996 to 2021 for studies published in English and French. The reference lists of included articles were also manually searched.

Inclusion and Exclusion Criteria
For this systematic review with meta-analysis, we included: (a) interventional and observational studies published in peer-reviewed journals, including epidemiological surveys, cross-sectional studies, cohorts, case series, case-controls and reports; (b) studies that reported HPV prevalence and/or risk factors; (c) studies with complete HPV genome sequencing results; (d) studies that reported HPV genotyping in HIV-positive/negative women or men and all cases of cervical lesions/cancers; (e) studies that reported the prevalence of HR-HPV or data to calculate this estimate. We defined HR-HPV as the proportion of the number of women with any HPV genotype divided by the total number of women visiting hospitals for lesions of the genital apparatus. For cohort studies, the cumulative incidence was taken as the prevalence, in which the number of new HPV cases was divided by the overall sample size. Case reports, reviews, systematic reviews and meta-analyses, commentaries, studies without full text, sham approach and duplicates were excluded.

Inclusion and Exclusion Criteria
For this systematic review with meta-analysis, we included: (a) interventional and observational studies published in peer-reviewed journals, including epidemiological surveys, cross-sectional studies, cohorts, case series, case-controls and reports; (b) studies that reported HPV prevalence and/or risk factors; (c) studies with complete HPV genome sequencing results; (d) studies that reported HPV genotyping in HIV-positive/negative women or men and all cases of cervical lesions/cancers; (e) studies that reported the prevalence of HR-HPV or data to calculate this estimate. We defined HR-HPV as the proportion of the number of women with any HPV genotype divided by the total number of women visiting hospitals for lesions of the genital apparatus. For cohort studies, the cumulative incidence was taken as the prevalence, in which the number of new HPV cases was divided by the overall sample size. Case reports, reviews, systematic reviews and meta-analyses, commentaries, studies without full text, sham approach and duplicates were excluded.

Patient and Public Involvement
No patients were involved.

Study Selection and Quality Assessment
An Excel spreadsheet was used to record the studies exported from the online databases. Duplicates identified from the full list of studies were removed. Titles and abstracts of eligi- ble studies were independently reviewed by two authors (M.C.T.T. and A.C.K.) for selection of relevant studies. Differing opinions among investigators regarding study selection were resolved by discussion, consensus and intervention of a third referee if necessary.
The quality of each study was assessed independently by all study authors using a scale dedicated to prevalence studies which is based on 10 components divided into two groups: internal and external validity of the study [25]. Scores of 0 or 1 were assigned to each question in the assessment tool for a total score of 10 per study. Scores of 0-3, 4-6 and 7-10 represented a high, moderate, and low risk of bias, respectively.

Data Extraction
Data from included studies were extracted using a Google form by three authors and verified by M.C.T.T. Data extracted were name of first author, year of publication, study design, inclusion criteria, country, sampling method, study period, age, sex, HIV status, sample size, HPV-HR rate and/or risk factors when identified, and cervical lesion types were extracted. Disagreements observed during data extraction were resolved by discussion and consensus.

Data Analysis
Heterogeneity between studies were estimated by I 2 and H statistics [25], where I 2 values indicated the degree of heterogeneity; values of 0%, 18%, 45% and 75% represented none, low, moderate and high heterogeneity, respectively [26]. Lack of evidence on heterogeneity between studies was indicated by obtaining H values close to 1, and these values were inversely correlated with the degree of heterogeneity. Prevalence, 95% confidence intervals (95% CI) and prediction intervals were estimated by random effect models [23,27]. Subgroup analyses according to country and HIV status were employed to adjust variations in the pooled estimate of prevalence. The statistical significance threshold was 0.05. The publication bias was assessed by visual inspection of the asymmetry of the funnel plot [28]. R version 3.6.0 software (packages "meta" and "metafor") through the RStudio interface was used to perform all meta-analyses [29,30].

Study Characteristics
We selected 28 studies for this systematic review (Figures 1 and 2). One study was conducted in Kenya [31], three in Uganda [3,32,33], one in Rwanda [34], one in Chad [35], three in Togo [36][37][38] (we included one study conducted in men as it met our eligibility criteria), one in Zambia [39], four in Zimbabwe [40][41][42][43], two in Burkina Faso [44,45] and finally 14 in South Africa [44,[46][47][48][49][50][51][52][53]. It should be noted that two studies were conducted simultaneously in two countries, namely Burkina Faso and South Africa in one study and Kenya and South Africa in the other. Participants in these studies were recruited from urban and rural areas, and samples were collected and analysed in health facilities (hospital, clinic, research laboratory).
According to the type of study found, we were able to retain 28 studies for our review, distributed as follows: 18 cross-sectional studies; 9 cohort studies and 1 case-control study. The articles included in this systematic review with meta-analysis were published between 2008 and 2022. Based on the sample size of the included studies, the minimum sample size was 85 participants and the maximum sample size was 8622 participants. The age of the participants in the different studies ranged from 15 to 90 years. Thirteen studies did not indicate the age range of the participants (Table 1).  According to the type of study found, we were able to retain 28 studies for our review, distributed as follows: 18 cross-sectional studies; 9 cohort studies and 1 case-control study. The articles included in this systematic review with meta-analysis were published between 2008 and 2022. Based on the sample size of the included studies, the minimum sample size was 85 participants and the maximum sample size was 8622 participants. The age of the participants in the different studies ranged from 15 to 90 years. Thirteen studies did not indicate the age range of the participants (Table 1).

Laboratory Methods to Detect HR-HPV Infection from the Included Studies
The molecular genotyping and HR-HPV detection techniques applied in the selected studies were as follows: next generation sequencing (95% sensitivity and 99.99% specificity; storage temperature: −20 • C to 25 • C; manual interpretation of results, two days for analysis delay), Linear Array HPV Genotyping Test (97% sensitivity and 90.5% specificity; storage temperature: −20 • C to 25 • C; automatic interpretation of results, 8 h for analysis delay), HPV care, genotyping by PCR restriction fragment length polymorphism analysis (with 70% sensitivity and 90% specificity; storage temperature: −20 • C to 25 • C; manual interpretation of results, 12 h for analysis delay), slide-based reverse hybridisation, multiplex RT-PCR, Sacace Biotechnologies, Anyplex II HPV28 real-time PCR (70% sensitivity and 90% specificity; storage temperature: −20 • C to 25 • C; automatic interpretation of results, 12 h for analysis delay), The Hybrid Capture DNA 2 (HC2) test (97% sensitivity and 97% specificity), INNO-LiPA HPV genotyping Extra ® test (97% sensitivity and 90.5% specificity; storage temperature: −20 • C to 25 • C; automatic interpretation of results, 8 h for analysis delay) ( Table 1). The varying performances of these HPV diagnostic tools suggest some inter-assay disparities and the need to develop adapted algorithms according to available tools and circulating HPV-genotypes within the setting of interest. Furthermore, these tests need a certain number of devices to be installed, they use expensive kits that cannot be stored at room temperature, interpretation of results is automatic for some platforms and manual for others, and use may require qualified personnel [55].

Factors Associated with HR-HPV Infection from the Included Studies
HIV infection, a history of having sex with three or more sexual partners, young age, condom use, sexually transmitted infection, single marital status, precarious employment, use of hormonal contraceptives, alcohol consumption and tobacco use were the most frequently reported risk factors associated with HR-HPV infection (Table 1).

Cervical Lesions and HR-HPV Infection in the Included Studies
Nine studies reported data on the cervical lesion status of participants in their studies; in total, 3814 tests were performed, and 459 participants had low grade squamous intraepithelial lesion (LSIL) compared to 534 who had high grade squamous intraepithelial lesion (HSIL). Two studies reported that the HR-HPV genotypes most commonly found in cervical lesions in the selected studies were 16 and 35 (Table 3).

Discussion
The objective of this study was to conduct a systematic review with meta-analysis to identify the different HR-HPV genotypes and their relationship with HIV, to verify their association with high-grade cervical dysplasia in sub-Saharan Africa, and to verify their adequacy with available vaccines. Thus, the synthesis of studies with detailed data on HIV-infected women showed that the prevalence of HR-HPV is high but similar to the prevalence of HR-HPV in HIV-uninfected women; this observed difference seems surprising. However, this result may also reflect the success of antiretroviral treatment (ART) among people living with HIV(PLHIV) due to scale-up of ART, which in turn ensures immune reconstitution to normalcy at level similar to that of HIV-negative individuals [56].
We decided to include studies that used a complete diagnostic kit capable of demonstrating at least eight HPV-HR genotypes. Although these kits are very costly for resourcelimited countries and require state-of-the-art technical facilities with qualified technicians, they have the advantage of demonstrating the HPV-HR genotypes that are prevalent in a region, and this information could have an impact on the appropriate vaccine for sub-Saharan Africa to eradicate this virus responsible for several public health problems [54].
HIV infection, a history of sex with three or more partners, young age of the woman, smoking, use of hormonal contraceptives were found to be the most frequently associated risk factors in this study. Smoking is known to act as an immunosuppressant, which decreases the host immune response to invasion of an HR-HPV genotype. With regard to sexuality, early sexual debut is an important risk factor, partly because of the immaturity of the cervix during adolescence, which makes it more susceptible to infection and damage by HR-HPV, and several authors have already shown that having sex with multiple partners significantly increases the possibility of exposure to HR-HPV [7,27].
With a positivity rate of 55.13% across all included studies, several high-risk oncogenic genotypes were revealed (genotypes HPV16, 18,31,33,35,39,45,51,52,56,58,59,61,62,66 and 68), with the most prevalent being HPV16 (the highest, 18.00%), followed by HPV35 (10.12%), HPV52 (9.98%) and HPV18 (9.87%). With the variability of genotypes identified from country to country, according to data generated by researchers in sub-Saharan Africa, it is essential to have data from each region in order to develop a good policy for the prevention of genital HPV infections which are significantly prevalent in the general population and in PLHIV in particular, who remain a vulnerable population when not put on ART [56]. Similar results were reported in a meta-analysis conducted in Ethiopian women, where genotypes HPV16, 52,35,18,58,51,45,31,53 and 56 (in decreasing order of prevalence) were the ten most common genotypes in this sub-Saharan African country. This dense genetic diversity in a context of non-adherence of the population to vaccination could be one of the reasons for the increase in the positivity rate of circulating HR-HPV in resource-limited countries [57]. The persistently high rate of HR-HPV positivity means that transmission of the infection continues to increase, highlighting the unsuspected high burden of cervical HR-HPV infection among women in sub-Saharan Africa, and underscoring the importance of implementing preventive measures such as vaccination to prepare the immune systems of these women to fight these infections [58]. Based on the information collected, our observations suggest that a large majority of these HR-HPV infections could a priori have been prevented by the use of Gardasil-9 vaccine (covering nine different genotypes, namely HPV6, 11,16,18,31,33,45,52,58), which could provide much better vaccine coverage in sub-Saharan Africa. However, for a variety of reasons (which could also be budgetary or scarcity of evidence-based findings), the current vaccination strategy available in most sub-Saharan countries that have adopted the vaccine is based on Gardasil 4 (targeting genotypes 6, 11, 16 and 18) or Cervarix (targeting genotypes 16 and 18). Hence, our current finings strongly support the introduction of Gardasil-9 vaccine into the national immunization program of sub-Saharan African countries for an optimal control of the increasing prevalence of HR-HPV genotypes in Africa. At the same time, and despite this great implementation of coverage by Gardasil-9, our study shows that nine other high-risk genotypes are not covered by Gardasil-9 vaccine, with genotypes such as HPV-35, HPV-51 and HPV-56 sometimes having a higher prevalence than the vaccine genotypes, as reported in Chad [35,59].
The association between HIV and HPV is well documented, with HIV increasing the risk of HPV infection and persistence [34]. Furthermore, summary estimates generated within this work revealed that HIV-infected women are more than four times at risk to be infected by HR-HPV in sub-Saharan Africa than their HIV-negative peers, in line with previous findings in this sub-region and worldwide [60,61]. Although adolescents have a low risk of cervical cancer, HIV-positive adolescents have a high risk of abnormal cervical cytology and are more likely to have persistent HPV infections [39]. In addition, the immune response induced by the quadrivalent HPV vaccine also appears to be robust in HIV-positive women, even as vaccine recommendations are identical in both groups. Modelling studies estimate that with adequate coverage, vaccination with the HPV16/18 vaccine could prevent 36-45% of invasive cervical cancers in countries such as Zambia [36].
All of the genotypes identified, especially those not yet included in the literature review, were made possible by the use of comprehensive molecular diagnostic techniques, which are increasingly available to laboratories in sub-Saharan Africa, thanks to collaboration between researchers, institutions and organisations in the North and South [22]. Authors have joined forces to promote science, which has led to the transfer of knowledge and skills, the reduction of boundaries, the mitigation of costs, the enhancement of the benefits of research results and the measurement of attributes. This justifies the interest of researchers to do more work on this topic of great importance to public health in sub-Saharan Africa [62].
Very few studies have reported data on the status of cervical lesions during infection and other coinfections. The available data nonetheless confirm that HIV is a major risk factor for the development of LSILs and HSILs, although the relative impact on HSILs is smaller [63]. This is consistent with cohort studies which show that, although HIV-positive women have a high number of cytological abnormalities, the vast majority of lesions are low grade, with only a slight increase in the prevalence of HSIL [64]. Therefore, HIV-positive women may be overtreated in screening programs due to the high number of low-grade lesions that may progress to HSIL [65].
In brief, it appears from our findings that accessibility to HPV screening tools needs to be expanded both within health facilities and at the community level. An optimal diagnostic algorithm, adapted to the context also needs to be developed and validated in low-income settings like sub-Saharan Africa, accompanied with training of health professionals, timely diagnosis and appropriate monitoring of women living with HIV/HPV co-infection as a vulnerable population. We must also enhance the sensitisation of stakeholders and community engagement for adherence to HPV screening and prevention strategies.

Conclusions
This systematic review indicates that the rate of HR-HPV genotype positivity among women in general and in HIV-infected women in particular remains high in sub-Saharan Africa, with no significant disparity compared to their HIV-negative peers, probably due to widespread access to antiretroviral treatment ("test and treat" strategy) aimed at eradicating this pandemic. The study also showed that an alarming number of HR-HPV genotypes are circulating in sub-Saharan Africa, with a considerable presence of genotypes not covered by the vaccination policy implemented by the health systems of certain countries, calling into question existing public health policies against cervical cancer. Cervical cancer diagnosis and treatment programs therefore need to be strengthened, while vaccination policies against new emerging HR-HPV genotypes need to be implemented in these contexts.