Sexually transmitted infections, bacterial vaginosis and genital inflammation as risk factors of HIV acquisition in adolescent girls and young women in South Africa

Barnabas, Shaun Lawrence

Sexually transmitted infections, bacterial vaginosis and genital inflammation as risk factors of HIV acquisition in adolescent girls and young women in South Africa

Doctoral Thesis

2019

Abstract

Background: In South Africa, young women are at increased risk of HIV infection, predominantly through heterosexual contact. Socio-behavioural and biological factors most likely play an important role at increasing risk. BV, STIs and genital yeast infections are known to influence HIV risk, although few studies have been conducted in African adolescent girls and young women (AGYW). Understanding the role of behavioural and biological risk factors in this key population is essential to inform on new prevention strategies. Aims: (1) To define the prevalence of sexual risk behaviour, bacterial vaginosis (BV), sexually transmitted infections (STIs), and genital fungal infections in South African AGYW; (2) to assess the relationship between symptoms and an etiological STI diagnosis; (3) to evaluate the influence of electronic and paper-based data collection methods on reported demographics and sexual risk behavior; (4) to examine the impact of non-infectious and infectious causes on genital inflammatory cytokine profiles; and (5) to investigate the use of cytokine biomarkers to identify young women with asymptomatic vaginal dysbiosis and STIs. Approach: To address these aims, the multi-center Women's Initiative in Sexual Health (WISH) study was conducted as a longitudinal observational trial that included HIV-negative, South African AGYW between the ages of 16-22 (EDCTP Strategic Primer 2013-2015), of which 149 from Masiphumelele, Cape Town were enrolled longitudinally for three visits and the 149 from Soweto, Johannesburg were seen cross-sectionally. Genital samples were collected to test for STIs (including Chlamydia trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis, Mycoplasma genitalium, herpes simplex virus (HSV)-1 and -2, Haemophilus ducreyi, Treponema pallidum and Lymphogranuloma venerum), BV (Nugent scoring) and yeast infections, to assess vaginal pH, and to measure prostate specific antigen (PSA) as a marker for the presence of semen. Multi-locus sequence typing (MLST) was performed on samples positive for C. trachomatis. In addition, 44 inflammatory, adaptive, regulatory cytokines, chemokines and growth factors were measured in genital secretions byLuminexR. Concentrations of sex hormones (estrogen, progesterone and luteinizing hormone), HSV-serology and cotinine (smoking) were measured in blood plasma. A HIV test was done at screening and each visit. Only HIV negative women were enrolled in the study Results: The prevalence of BV (Nugent 7-10) in South African AGYW were similarly high in Masiphumelele and Soweto (48% and 45%, respectively), as was intermediate microbiota (12% and 16%, respectively). In addition, 40% of South African AGYW were infected with any STI. C. trachomatis and N. gonorrhoeae were more prevalent in Masiphumelele compared to Soweto (42% vs. 18% for C. trachomatis and 11% vs. 5% for N. gonorrhoeae, respectively), while the prevalence of the other STIs tested for, including T. vaginalis, M. genitalium and HSV-2, was similar at both sites. Two-thirds (67%) of AGYW had HPV infection while 10% had a fungal infection. The indicators of risky behavior (including partner's HIV status being unknown [57%], having unprotected vaginal sex in the last three months [40%], and being in HIV discordant relationships [20%]) were high in this population, and tended to be more prevalent in Soweto than Masiphumelele. It was striking that only 24% of women with a diagnosed STI or BV were symptomatic, underlining the inadequacy of syndromic management in this population. The most sensitive symptom in this cohort was dysuria, which had 50% sensitivity in diagnosing a STI, and the least sensitive was abnormal vaginal discharge with a sensitivity of 22%. Having multiple, concurrent conditions (multiple STIs or a STI and BV) did not increase the symptom frequency or severity. In Chapter 3, two methods were compared to obtain demographic and sexual risk behaviour data – an electronic tablet device and a paper-based questionnaire. This was done to explore the presence of reporting bias and to improve the frequency of it in the cohort. There were no obvious differences seen in the reporting of risk between the two arms. It was striking that geographical differences in risk reporting were observed, with AGYW from Soweto reporting a higher frequency of HIV discordant relationships and intergenerational relationships than women from Masiphumelele, independently of the data collection method used. As genital inflammatory cytokines are known to influence HIV risk, physiological and pathological factors that may influence the genital cytokine profiles of AGYW were examined in Chapter 4 and 5. A change of vaginal pH appeared to play a crucial role, as it was associated with a significant increase in half (22/44) of the cytokines measured in AGYW with no STIs, no yeast infections and a Nugent score ≤3. Further, in these adolescents that one would consider healthy, a pronounced effect was seen with hormonal contraception choice, with Net-En significantly increasing the median concentrations of 31/44 cytokines, and DMPA upregulating 27/44 cytokine concentrations. Recent sex (being PSA positive) did not influence genital inflammation. Next, the impact of BV, viral STIs (HPV and HSV-2) and yeast infections on genital inflammation in adolescents was investigated. BV was the most inflammatory condition seen with 34/44 cytokines being upregulated. Interestingly, BV was also associated with downregulation of chemokines at both sites, including GRO-α, IP-10 and MIG. BV alone was more inflammatory than coinfection of BV with C. trachomatis, BV with another STI, or BV, C. trachomatis and another STI. In this cohort, HPV and HSV-2 shedding did not influence genital inflammation. In Chapter 5, the influence of bacterial and parasitic STIs (including C. trachomatis, N. gonorrhoeae, T. vaginalis and M. genitalium) on the genital inflammatory profile was explored. C. trachomatis and T. vaginalis were equally inflammatory, with 23/44 cytokines being elevated. M. genitalium infections had a very little impact with only GM-CSF being significantly downregulated in Masiphumelele, while N. gonorrhoea infection did not influence the genital inflammatory milieu. Geographical variation in genital cytokine responses were observed in adolescents infected with T. vaginalis, which was moderately inflammatory in AGYW from Masiphumelele but not in AGYW from Soweto. With C. trachomatis infection, however, there was a more pronounced inflammatory response seen in adolescents from Soweto (where 23/44 genital cytokines were significantly increased) than those from Masiphumelele (no cytokines were changed), compared to adolescents with no STIs, no yeast infections and a Nugent score ≤3. Using MLST, further differences were associated with C. trachomatis infection: sequence type (ST) 3 was the most inflammatory C. trachomatis ST detected with 14/44 cytokines upregulated; ST137 caused no significant changes in cytokine markers, and ST100b was the least inflammatory with four cytokines being down regulated. Exploring the relationship between years of sexual experience and C. trachomatis infection showed low levels of inflammation in women with only with <2 year of sexual experience (only MIG was upregulated), while C. trachomatis infection in AGYW with ≥2 years sexual experience was associated with an increase in 19/44 cytokines. Finally, in Chapter 6 the cytokine biomarkers IL-1α, IL-1β and IP-10 were evaluated in the classification of asymptomatic STIs, BV and intermediate microbiota. These biomarkers correctly classified 76% of women using their SoftcupR samples, 76% of women using lateral vaginal wall swabs and 73% using vulvovaginal swab. The addition of pH to the biomarkers model improved the classification (82%) and sensitivity (87%) of results, but reduced specificity (64%). The validation of these biomarkers could lead to the development of a point-of-care test that could be used to triage women into risk categories independently of symptoms. Conclusion: The results show a high prevalence of risky behavior, STIs and BV in this vulnerable population, with poor correlation between clinical symptoms and diagnosis of STIs or BV, with most cases in adolescents being asymptomatic. Injectable hormonal contraceptive use led to increased genital inflammation in adolescents. Despite the lack of symptoms, most cases of BV and STIs were associated with significantly elevated concentrations of genital cytokines in the genital mucosa of these at-risk adolescents compared to their uninfected counterparts, with variation in levels of inflammation seen by geographic location, type of infection, and years of sexual experience. These results supported the use of biomarkers as a potential method to identify at risk women. Overall, these findings highlight the urgent need to include adolescents in the design and testing of new HIV prevention strategies and the importance of active identification and management of any infection or dysbiosis in this key at risk population.

Keywords

Medical Virology

Reference:

Collections

PhD / Doctoral

Full item page