High burden of Human Papillomavirus and pre-malignant cervical lesions among women starting HIV treatment in KwaZulu-Natal, South Africa

Abstract

Objectives:

Human Papillomavirus (HPV) can cause cervical cancer, a leading cause of female cancer mortality in South Africa and worldwide. We assessed the burden of, and factors associated with, HPV infection using a molecular assay and cervical cytology among women living with HIV (WLHIV) in South Africa.

Methods:

In this cross-sectional analysis, WLHIV initiating antiretroviral therapy (ART) underwent cervical Xpert HPV testing and liquid-based cytology. The Xpert test detects 14 high-risk (hr) HPV types reported in a pooled qualitative result. We evaluated clinical and socio-demographic variables and proportions, between women testing positive or negative for hr-HPV, compared cytology with hr-HPV results and assessed associations with HPV positivity.

Results:

We enrolled 260 WLHIV, median age 31.0 [Interquartile range (IQR) 26.0 – 38.0] years. Overall, 91.3% of women were never screened for cervical cancer previously and none received HPV vaccination. In total, 67.3% (175/260) of women tested positive for any hr-HPV type, of which HPV16 and HPV18/45 were detected in 17.3% (45/260) and 22.7% (59/260) of women respectively and 56.5% (147/260) tested positive for 11 other hr-types. Of 258 WLHIV, 33.3% (86/258) had abnormal cytology: high-grade squamous intra-epithelial lesion (SIL) 7.8% (20/258), low-grade SIL 24.0% (62/258) and atypical squamous cells of undetermined significance 1.6% (4/258). Of these, 93.0% (80/86) tested positive for hr-HPV: 30.0% (24/80) for HPV16, 31.3% (25/80) for HPV18/45 and 92.5% (74/80) for other hr-HPV types. Having a CD4 count <200 cells/μl was associated with hr-HPV infection [adjusted Prevalence Ratio 2.24; 95% Confidence Interval 1.69 – 2.99 (p<0.001)].

Conclusions:

hr-HPV infection and cervical abnormalities are common among WLHIV starting ART, especially those with low CD4 counts, highlighting that early HIV testing and treatment initiation must be prioritised together with cervical cancer screening. The diversity of hr-HPV types suggests a need for vaccines with expanded HPV type coverage in this setting.

Introduction

Human Papillomavirus (HPV) infection remains a common sexually transmitted infection and the main cause of cervical cancer, which has a global incidence of 15.6 per 100,000 women per year (wpy)(1). South Africa’s cervical cancer incidence exceeds the global average at 35.6 per 100,000 wpy, with a high mortality rate of 19.5 per 100,000 wpy, making it the leading cause of cancer mortality among women aged 15–44 years(1). The high background prevalence of Human Immunodeficiency Virus (HIV) infection among women is a major contributing factor to cervical cancer morbidity and mortality in this setting(2).

HPV is a DNA virus with over 200 known genotypes, of which 12 are considered high-risk (hr) oncogenic types (3). HPV16 and 18, targeted by all commercially available HPV vaccines, together account for the majority of cervical cancers globally with other oncogenic types thought to play a lesser role(3). Women living with HIV (WLHIV) are known to carry a high burden of HPV infection and often harbour multiple infections, which are poorly cleared due to a suboptimal immune response(2). Clinical disease progression to invasive cervical cancer (ICC) is also known to occur more rapidly in WLHIV, not receiving antiretroviral treatment (ART) and with low CD4 counts at ART initiation(2). Overall, WLHIV are diagnosed with ICC almost 5–10 years earlier than their HIV-negative counterparts(2).

In South Africa primary prevention of cervical cancer is through targeted HPV vaccination of grade 4 female learners aged 9 years and older, with bivalent vaccines, since 2014. Unvaccinated, older women, in particular WLHIV, remain at high risk of cervical cancer and have relied on cytology screening for secondary prevention. The country has recently transitioned to HPV testing which is currently being implemented in the public health sector.

As part of an ongoing clinical trial, we assessed the burden of HPV infection among WLHIV initiating ART in an urban primary health clinic (PHC), using a molecular assay, with a view to inform clinical care for this high-risk group.

Methods

Study design and Study setting

This study is a cross-sectional analysis at enrolment into the Simplifying TREAtment and Monitoring for HIV (STREAM HIV) trial, a randomised controlled trial assessing point-of-care (POC) tenofovir adherence testing and POC viral load testing compared with standard of care in people living with HIV(4). Between February 2021 and June 2023, patients attending the Prince Cyril Zulu Communicable Disease Centre, a PHC in the eThekwini central business district, were recruited for enrolment if ART-eligible(4).

Study procedures

Participants provided written informed consent at study entry, followed by a medical history, physical examination, including pelvic examination and cervical cancer screening for women, screening for tuberculosis and World Health Organisation HIV staging performed by trained clinical staff(4). Clinical and laboratory data were collected and entered electronically in the DFexplore database (Version 5.1.4, 2018).

Cervical screening

Cytology samples were collected with an endocervical broom-like device and placed in a PreservCyt medium using a ThinPreP (Hologic, Marlborough, Massachusetts, United States) liquid-based cytology kit. In addition, HPV testing was performed at the on-site clinic laboratory by trained technicians using GeneXpert polymerase chain reaction technology in accordance with manufacturers specification(5). Of the 20 ml sample collected for cytology, 2.5ml was aliquoted into a cryovial and 1.1 ml was used for the Xpert HPV assay (Cepheid, Sunnyvale, California, Unites States). The test is designed to detect HPV16, HPV18/45 and 11 other hr-types, reported as a pooled qualitative result (P3: 31, 33, 35, 52 or 58; P4: 51 or 59 and P5: 39, 56, 66 or 68) with a turnaround time of 60 minutes(5). Cytology specimens were processed at the South African National Health Laboratory Service laboratory by trained cytologists, as part of the public health sector service, with a turnaround time of 4 to 6 weeks. Cytology results were classified in accordance with the 2001 Bethesda System.

Statistical Analysis

Baseline characteristics were summarized as means with standard deviation (SD) or medians with interquartile ranges (IQR) for continuous variables, and proportions for categorical variables. The Wilcoxon-Mann-Whitney test was used to compare continuous variables and Fisher’s exact test was used to compare proportions. Associations between sociodemographic, clinical exposures and HPV positivity were assessed using Poisson regression models with robust variance. Analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, North Carolina, US).

Results

Baseline characteristics

Of 317 women enrolled, 260 (82.0%) women who underwent HPV testing, based on test availability, contributed to the HPV sub-study. The median age was 31.0 [Interquartile range (IQR) 26.0 – 38.0)] years and median age at sexual debut was 18.0 (IQR 17.0 – 20.0) years. Almost 80% of women reported having at least 3 or more lifetime sex partners and all women had one sex partner in the past 6 months. The overall mean CD4 count was 436 cells/μl, higher among participants testing HPV negative than positive (576 vs 367 cells/μl, p <0.001). Overall, 91.3% of women were never screened for cervical cancer previously and none had ever received an HPV vaccine.

HPV frequency measured by the Xpert HPV test

Overall, 67.3% (175/260) of women tested positive for any hr-HPV type with 17.3% (45/260) testing positive for HPV16, 22.7% (59/260) for HPV18/45 and 56.5% (147/260) for 11 other hr-HPV types. HPV positivity was not mutually exclusive between groups on the Xpert assay. Of the other hr-HPV types, P3 types (HPV 31, 33, 35, 52 or 58) predominated at 39.6% (103/260), followed by P5 types (HPV 39, 56, 66 or 68) at 25.8% (67/260) and P4 types at 17.3% (45/260). HPV infection exclusively due to other hr-HPV types was 31.8%. Multiple HPV types were found to be common, with 150 women testing positive for at least two or more HPV pools.

Cervical cytology results by HPV status

Of 258 women with available cytology results 86 abnormal results were observed: 20/258 (7.8%) high-grade squamous intra-epithelial lesion (HSIL), 62/258 (24.0%) low-grade squamous intra-epithelial lesion (LSIL) and 4/258 (1.6%) atypical squamous cells of undetermined significance (ASCUS). A further 42/258 (16.3%) specimens were inadequate for evaluation due to an absent endocervical component, possibly due to inadequate sampling. Overall, 93.0% (80/86) of abnormal cytology was associated with a positive HPV result. Of women with either ASCUS, LSIL or HSIL and a positive HPV result, 30.0% (24/80) tested positive for HPV16; 31.3% (25/80) for HPV18/45 and 92.5% (74/80) for one or more of 11 other hr-HPV types. In total 43.7% (35/80) of women with abnormal cytology tested positive exclusively for other hr-HPV types. Notably, 47.7% (61/128) of normal cytology results were associated with a positive hr-HPV result.

Factors associated with HPV infection

In multivariable analysis, CD4 count 200 to <500 [adjusted Prevalence Ratio (aPR)1.71; 95% Confidence Interval (CI) 1.27 – 2.29 (p<0.001)] and CD4 <200 cells/μl [aPR 2.24; 95% CI 1.69 – 2.99 (p<0.001)] was associated with hr-HPV infection.

Discussion

In this cross-sectional study we demonstrate that more than two thirds of WLHIV presented at ART initiation with hr-HPV types and high rates of premalignant cervical lesions. Co-infection with multiple hr-HPV types was found to be common, with women above 25 years bearing the highest proportion of HPV infection overall. hr-HPV types other than HPV16 and HPV18/45 were found to be common in these women, not previously vaccinated against HPV. Almost 40% of women tested positive for P3 (HPV31, 33, 35, 52 or 58) types which are common to types targeted by the nonavalent HPV vaccine, apart from HPV35 (Figure 1a and 1b). This finding concurs with other prevalence studies from South Africa, which report that these hr-HPV types are commonly detected in high-grade cervical lesions, irrespective of HIV status(6, 7). Access to HPV vaccines with expanded HPV type coverage is needed to ensure vaccination has a substantial impact on ICC incidence and disease burden in this setting.

Figure 1a:

HPV frequency by cytology

Figure 1b:

Cytology findings by HPV type

In addition to the high burden of HPV, a high proportion of women presented with uterine cervical disease at ART initiation. Almost 8.0% (20/258) of women had a high-grade pre-malignant lesion requiring immediate colposcopy referral and treatment. When including women with ASCUS/LSIL and a positive HPV result, 31.4% of women required referral. With an estimated 8.4 million people requiring ART in South Africa in 2022, and over 50% being women, this study highlights the burden of HPV disease among WLHIV and the extensive resources required to prevent and treat ICC(8). Studies have shown that despite ART initiation, women are still at risk of developing ICC where ART is initiated at lower CD4 thresholds, due to an inability to completely restore immunity against HPV infection at the mucosal level(9). This is of particular concern in South Africa as persistently high rates of ICC are reported at two- and five-years post-ART initiation(10). In this study, having a low CD4 count at ART initiation was strongly associated with hr-HPV infection.

This study underscores the burden of HPV disease in WLHIV who did not benefit from HPV vaccination. Early HIV testing, ART initiation and scale-up of high-performance screening tests for WLHIV must be prioritised. Pragmatic interventions such as decentralised colposcopy services are needed in parallel to shift care for women with cervical disease to a PHC level.

What is already known on this topic:

Several studies have described the high burden of Human Papillomavirus (HPV) among women living with HIV (WLHIV) in South Africa. However, few studies report on HPV prevalence among WLHIV in KwaZulu-Natal province, the epicentre of the HIV epidemic in South Africa with an HIV prevalence of 21.8%.

What this study adds:

This study underscores the burden of HPV infection among women newly initiating antiretroviral therapy (ART). Using a point-of-care molecular assay with extended genotyping we demonstrate the diversity of high-risk HPV types in this population in relation to premalignant lesions of the cervix. In this study, having a low CD4 count at ART initiation was associated with high-risk HPV infection.

How this study might affect research, practice or policy:

Together with cervical cancer screening, HIV counselling and testing and timeous antiretroviral treatment initiation must remain a priority for women at highest risk for cervical cancer. This study highlights a need for HPV vaccines with expanded HPV type coverage for primary prophylaxis in this setting.