J van de Wijgert

Vaginal practices, microbicides and HIV: what do we need to know?

The goal of a safer vaginal environment could be reached by identifying harmful vaginal practices and an effective microbicide, thereby increasing options for HIV prevention The global burden of HIV, its increasing feminisation, and chronic difficulties with development of options for HIV prevention all argue for an intensified re-examination of factors influencing the efficiency of heterosexual HIV transmission. This includes vaginal practices and products used by large numbers of women worldwide to tighten, dry, warm and clean their vagina. Women’s efforts to change their genital environment can undermine each component of innate defences against pathogens.1 In particular, vaginal practices have been linked with loss of lactobacilli and disruption of the vaginal epithelium.2–4 These practices may therefore be an important mediator in acquisition of STI, including HIV, or worsen pre-existing infections. Despite this, surprisingly little is known about the effects of specific vaginal practices on HIV transmission dynamics. In past decades, both cross-sectional and longitudinal studies have found an association between intravaginal cleansing and adverse reproductive outcomes, including pelvic inflammatory disease, ectopic pregnancy and bacterial vaginosis (BV).1,5,6 BV could be an intermediary factor between vaginal practices and HIV infection (fig 1). Though much uncertainty remains about the pathophysiology of BV, its accompanying inflammatory milieu, characterised by pro-inflammatory cytokines and immune cell changes, is likely to enhance HIV transmission.7 Also, immunological changes with BV can stimulate HIV expression, raising HIV levels in the genital tract virus and likely the infectivity of women.7 Acquisition of HSV-2 may also be higher among women with BV.8 Figure 1  Hypothesised causal pathway of vaginal practices and HIV. Results of the GSVP …

Home-based versus clinic-based self-sampling and testing for sexually transmitted infections in Gugulethu, South Africa: randomised controlled trial.

Objectives: To test whether more women are screened for sexually transmitted infections when offered home-based versus clinic-based testing and to evaluate the feasibility and acceptability of self-sampling and self-testing in home and clinic settings in a resource-poor community. Methods: Women aged 14–25 were randomised to receive a home kit with a pre-paid addressed envelope for mailing specimens or a clinic appointment, in Gugulethu, South Africa. Self-collected vaginal swabs were tested for gonorrhoea, chlamydia and trichomoniasis using PCR and self-tested for trichomoniasis using a rapid dipstick test. All women were interviewed at enrolment on sociodemographic and sexual history, and at the 6-week follow-up on feasibility and acceptability. Results: 626 women were enrolled in the study, with 313 in each group; 569 (91%) completed their 6-week follow-up visit. Forty-seven per cent of the women in the home group successfully mailed their packages, and 13% reported performing the rapid test and/or mailing the kit (partial responders), versus 42% of women in the clinic group who kept their appointment. Excluding partial responders, women in the home group were 1.3 (95% CI 1.1 to 1.5) times as likely to respond to the initiative as women in the clinic group. Among the 44% who were tested, 22% tested positive for chlamydia, 10% for trichomoniasis, and 8% for gonorrhoea. Conclusions: Self-sampling and self-testing are feasible and acceptable options in low-income communities such as Gugulethu. As rapid diagnostic tests become available and laboratory infrastructure improves, these methodologies should be integrated into services, especially services aimed at young women.

Vaginal high-risk human papillomavirus infection in a cross-sectional study among women of six different ethnicities in Amsterdam, the Netherlands: the HELIUS study

Objective In the Netherlands the incidence of cervical cancer is higher among ethnic minority populations compared with the general Dutch population. We investigated the prevalence of, and risk factors associated with, vaginal high-risk human papillomavirus (hrHPV) infection in women of six different ethnicities living in Amsterdam. Methods For this cross-sectional study we selected women aged 18–34 years old of six ethnicities from the large-scale multiethnic HEalthy LIfe in an Urban Setting study. Self-collected vaginal swabs were tested for HPV DNA and genotyped using a highly sensitive PCR and reverse line blot assay (short PCR fragment (SPF)10-PCR DNA enzyme immunoassay/LiPA25-system version-1, delft diagnostic laboratory (DDL)). Participants completed a questionnaire regarding demographics and sexual behaviour. Logistic regression using generalised estimating equations was used to assess risk factors of hrHPV, and to investigate whether prevalence of hrHPV differed among ethnicities. Results The study population consisted of 592 women with a median age of 27 (IQR: 23–31) years. Dutch and African Surinamese women reported the highest sexual risk behaviour. HrHPV prevalence was highest in the Dutch (40%) followed by the African Surinamese (32%), Turkish (29%), Ghanaian (26%), Moroccan (26%) and South-Asian Surinamese (18%). When correcting for sexual risk behaviour, the odds to be hrHPV-positive were similar for all non-Dutch groups when compared with that of the Dutch group. Conclusions We found an overall higher hrHPV prevalence and higher sexual risk behaviour in the native Dutch population. Further research is needed to unravel the complex problem concerning cervical cancer disparities, such as differences in participation in the cervical cancer screening programme, or differences in clearance and persistence of hrHPV.

P3.060 There is a Need For Multipurpose Prevention Technologies Targeting HIV and Common Reproductive Tract Infections: Data from the Microbicide Safety Biomarkers Study Team

Background The ideal vaginal microbicide should reduce the risk of HIV infection and other reproductive tract infections (RTIs) while preserving the integrity of the cervicovaginal epithelium. Future microbicides and multipurpose prevention technologies (MPT) could improve maternal reproductive health and prevent multiple sexually transmitted infections. Objectives and Methods The Microbicide Safety Biomarkers Study is a prospective cohort study of 110 adults, 30 adolescents and 30 pregnant women in Kenya and South-Africa, 30 women engaging in vaginal practises in South-Africa and 30 high-risk and 30 HIV-positive women in Rwanda. RTIs and biomarkers of the vaginal microbiome and inflammation were studied. Results Baseline prevalence RTI data are presented in the table. A significant difference (p = 0.027 to 0.001) between the study groups was present for all RTIs except for Trichomonas vaginalis (TV). Neisseria gonorrhoeae (NG), syphilis and HSV-2 were associated (p = < 0.001) with sexual risk taking behaviour (sex worker OR at least 3 partners last year OR at least one sexual partner with HIV in the past 3 months OR age first sex less than 15 years). HSV-2 was detected in 51.5% of the high risk-takers compared to 28.6% of the low risk-takers. For women with bacterial vaginosis (Nugent 7–10) Chlamydia trachomatis (CT) (p = < 0.028) was present in 14.9% and TV (p = < 0.001) in 9% compared to 6.3% and 1.5% in women without BV (Nugent 0–3), respectively. Abstract P3.060 Table 1 Group HSV-2 CT NG Syphilis TV Candida Adults 34.0% 10% 0.9% 0% 3.7% 19% Pregnant 26.7% 10% 0% 1.7% 6.8% 40% Adolescents 6.7% 6.7% 0% 0% 6.8% 20% Vaginal Practices 45.2% 26.7% 3.3% 0% 14.3% 33.3% High risk 46.6% 10% 6.7% 6.7% 10% 10% HIV-positive 82.8% 0% 13.3% 20% 10% 13.3% Conclusion RTIs are common among African women targeted for microbicide trials. The introduction of a MPT targeting a combined prevention of HIV and HSV-2 is warranted in these populations.

P3.221 Oral and Injectable Hormonal Contraception Decrease Risk of Bacterial Vaginosis But Oral Contraception May Increase Risk of Vaginal Candidiasis: A Systematic Review of Published and Unpublished Data

Background A recent World Health Organization (WHO) technical consultation concluded that combined oral contraception (COC) does not increase HIV acquisition in women, but the evidence for depot medroxyprogesterone acetate (DMPA) is conflicting. Significant evidence suggests that bacterial vaginosis (BV) and vaginal candidiasis, both representing an ‘unhealthy’ vaginal microbiome, increase HIV acquisition in women. Methods We conducted a systematic review using the PRISMA 2009 guidelines, and re-analysed the Hormonal Contraception and HIV Acquisition (HC-HIV) study, to evaluate the effect of HC use on the vaginal microbiome. Vaginal microbiome outcomes included BV by Nugent scoring, vaginal candidiasis by culture or KOH wet mount, and microbiome compositions as characterised by molecular techniques. Results Our review of 36 eligible studies found that COC and DMPA use reduce BV by 10–20% and 18–30%, respectively. The HC-HIV data showed that COC and DMPA use also reduce intermediate microbiota (Nugent score of 4–6) by 11% for each. In contrast, COC use (but not DMPA use) may increase vaginal candidiasis; 7 of 12 studies reported a statistically significant increase in vaginal candidiasis, 2 reported a positive association approaching significance, 2 reported no association, and one reported a statistically significant reduction. Evidence for a reduction of BV risk in HC users is much stronger than evidence for a potential increased candidiasis risk in COC users: the quality of the BV studies was higher and the results more consistent. Molecular vaginal microbiome studies (n = 4) confirm that high oestrogen levels favour a vaginal microbiome composition dominated by ‘healthy’ Lactobacillus species; the effects of progesterone on the microbiome are less clear. Conclusions The hypothesis that DMPA use may increase HIV risk by increasing BV or vaginal candidiasis risk is not supported by the evidence. COC use may predispose for vaginal candidiasis, but is not believed to be associated with increased HIV acquisition

O04.1 The Influence of Hormonal Contraception and Pregnancy on the Vaginal Microbiome, Sexually Transmitted Infections, and Cytokine Responses in a Cohort of Rwandan Sex Workers

Background The effects of hormonal contraception and pregnancy on the vaginal microbiome (by molecular methods), acquisition and persistence of sexually transmitted infections (STIs), and genitourinary mucosal immunology are still largely unknown. Methods HIV-negative, non-pregnant female sex workers (n = 397) in Kigali, Rwanda, were followed for two years. Demographic, behavioural, clinical, STI and pregnancy data were collected at regular intervals. The vaginal microbiome was cross-sectionally determined using a phylogenetic microarray (n = 174). Women with STIs were purposefully oversampled in this subsample. Inflammatory cytokines were measured in cervicovaginal fluid using Luminex and ELISA methodology (n = 343). Hormonal exposure was defined as use of hormonal contraception (oral or injectable) or a positive urine pregnancy test. Women in the exposure groups were compared to non-pregnant women who did not use hormonal contraception. Adjustments were made for demographic data and sexual risk taking. Results At baseline, 12% of the women used hormonal injectables, and 6% oral contraceptives (OC); 7.7% was pregnant. OC use was associated with higher HPV prevalence (aOR 3.09; 95% CI 1.42–7.72), higher Chlamydia trachomatis incidence (aOR 7.13; 95% CI 1.40–36.30), and lower syphilis prevalence (0% vs 7.2% in controls) and incidence (0% vs 1.2%). Hormonal injectables were associated with higher HSV-2 prevalence (aOR 2.08; 95% CI 1.23–3.50). Pregnancy was weakly associated with higher Trichomonas vaginalis(aOR 1.67; 95% CI 0.97–2.88) and vaginal yeast (aOR 1.95; 95% CI 0.99–3.82) incidence. Six vaginal microbiome clusters were identified. No associations between hormonal exposure status and vaginal microbiome clusters were found; however, pregnant women had lower Gardnerella vaginalis levels. Pregnant women had higher IL-8 levels in cervicovaginal fluids than non-exposed women. Conclusions Both hormonal contraception and pregnancy were associated with higher STI incidence. Overall, vaginal inflammation and microbiome composition were similar among groups, but pregnant women had lower Gardnerella and higher IL-8 levels.

P3.117 Intravaginal Practices and HIV Acquisition Among Women at High Risk For Infection in Tanzania and Uganda

Background Intravaginal practises (IVP) are highly prevalent in sub-Saharan Africa and may facilitate HIV transmission. In two microbicide feasibility cohorts in North-West Tanzania and Kampala, Uganda, we describe baseline prevalence of IVP and investigate associations between IVP and HIV acquisition prospectively. Methods We enrolled HIV-negative women who worked in bars, guesthouses and similar facilities in Tanzania and sex workers and bar workers in Uganda, and followed them quarterly for 12 and 18 months, respectively. At each visit, participants were tested for HIV and interviewed about IVP in the past 3 months. We assessed the association between IVP at each follow-up visit and HIV acquisition using Poisson regression in a combined analysis of both cohorts, controlling for potential confounders. Results 1611 participants were enrolled (966 Tanzanians; 645 Ugandans). At enrolment, 92% of Tanzanians and 93% of Ugandans reported ever cleansing inside the vagina (Table 1); the majority who cleansed used soap/soapy water at least once (75% of Tanzanians and 58% of Ugandans). Nearly half (49%) of Ugandans, but only 13% of Tanzanians, reported ever inserting a substance inside the vagina. The most common substances inserted were herbs, petroleum-based jelly, detergents, aerated drinks, honey, and salt. 1472 participants contributed 71 seroconversions/1,780 pyrs. HIV incidence was 3.99/100 pyrs. Incidence was lower among women who reported cleansing in the past 3 months (aRR:0.44, 95% CI: 0.21–0.93). HIV incidence was similar in women cleansing with soap and those not cleansing/cleansing with water only. Inserting detergent (aRR:3.05, 95% CI: 1.30–7.18) or petroleum-based jelly (aRR:2.52, 95% CI: 1.08–5.89) were associated with HIV incidence. Abstract P3.117 Table 1 Reported IVP at enrolment among women at increased risk for HIV in Tanzania and Uganda Type of IVP used in the last 3 months Tanzania(N = 966) Uganda(N = 645) Intravaginal Cleansing 891 (91.9%) 601 (93.2%) Of those who cleansed, substances ever used Water only 352 (39.5%) 250 (41.6%) Water and Soap 664 (74.5%) 351 (58.2%) Other 1 (0.1%) 2 (0.3%) Intravaginal Insertion 122 (12.6%) 504 (49.0%) Of those who inserted, substances ever used Herbs or other traditional substance 55 (45.1%) 139 (44.1%) Washing powder or detergent 27 (22.1%) 63 (20.0%) Petroleum-based jelly or lotion(e.g Vaseline) 40 (32.8%) 99 (31.4%) Lemon 15 (12.3%) 7 (2.2%) Aerated drinks (e.g. Coca Cola) 0 (0.0%) 113 (35.9%) Honey 4 (3.3%) 64 (20.3%) Salt 1 (0.8%) 61 (19.4%) Other 19 (15.6%) 1 (0.3%) Conclusions Intravaginal cleansing was highly prevalent in both cohorts; however, insertion was more common among Ugandans. Cleansing was not a predictor of HIV in this study, and may be protective; however, some substances used for insertion may be harmful. These rarer and more harmful types of IVP warrant further investigation.

P1.033 The Vaginal Microbiome and Its Clinical Correlates in a Cohort of African Sex Workers

Background Although Sub-Saharan Africa is one of the most important areas in the world to study the complex relationships between the vaginal microbiome and reproductive health outcomes, data are limited. Methods Endocervical samples of 174 female sex workers in Kigali, Rwanda, were analysed cross-sectionally using a phylogenetic microarray specifically designed for the cervicovaginal microbiome. Women with sexually transmitted infections (STI) were purposefully oversampled. Two hundred fifty one probes were used for co-regularised spectral clustering analysis and 123 probes (specific at species or genus level) to describe the vaginal microbiome clusters. Demographic, behavioural, and clinical correlates of the clusters were also determined. Results The prevalence of HIV (36%) and other STIs (bacterial STI 46%, HPV 48%, and HSV-2 78%) in the analysis sample were high by design. Six distinct vaginal microbiome clusters were identified. Two clusters were dominated by Lactobacillus crispatus and L. iners, respectively, and were associated with a Nugent score of 0–3. Three clusters were dominated by Gardnerella vaginalis, Atopobium spp. and Prevotella spp in different compositions, and were associated with a Nugent score of 7–10. The sixth cluster, also dominated by anaerobic bacteria, was not associated with a particular Nugent score category. Women belonging to the L. crispatus cluster were significantly less likely to have bacterial (0% compared to 32–67%) and viral STIs (36% compared to 89–100%) than women in the other 5 clusters. Conclusion In this sample of African sex workers with a high prevalence of HIV and STIs, six vaginal microbiome clusters were identified. Sex workers with a vaginal microbiome dominated by L. crispatus (but not L. iners) did not have bacterial STIs and were less likely to have viral STIs than women with other microbiome compositions. Longitudinal studies are needed to determine the temporal relationships between the vaginal microbiome and various STIs.