The vaginal microbiota impact of antibiotic and probiotic compounds and point-of-care testing for vaginal dysbiosis and sexually transmitted infections

Abstract

THE VAGINAL MICROBIOTA IMPACT OF ANTIBIOTIC AND PROBIOTIC COMPOUNDS AND POINT-OF-CARE TESTING FOR VAGINAL DYSBIOSIS AND SEXUALLY TRANSMITTED INFECTIONS by Marijn Verwijs Studies employing modern molecular techniques have shown that the vaginal microbiota (VMB) of most women is dominated by beneficial Lactobacillus species. However, about one-third of women worldwide at any given time point have a non-optimal VMB, referred to as vaginal dysbiosis. The most common bacterial vaginal dysbiosis is bacterial vaginosis (BV), which is characterised by a decrease in lactobacilli and an increase in BV-associated anaerobic bacteria. The most common type of fungal vaginal dysbiosis is vulvovaginal candidiasis (VVC), which is characterised by the presence of Candida yeasts, usually in the presence of lactobacilli. BV and VVC have been associated with HIV acquisition, and BV also with adverse pregnancy outcomes. BV can be treated with the antibiotics metronidazole or clindamycin, but recurrence is common. Better therapies to treat BV and prevent its recurrence are needed. The prevalences of BV and sexually transmitted infections (STIs) are high worldwide but are highest in sub-Saharan Africa. Many sub-Saharan African countries rely on the World Health Organisation (WHO) syndromic management algorithms to treat symptomatic women given their limited clinic and laboratory infrastructure. These algorithms provide guidance on which urogenital infections to treat when certain urogenital symptoms (referred to as ‘syndromes’) are reported by the patient, without the need for diagnostic testing (and with pelvic examinations being optional). A syndrome is treated for all infections that might cause that syndrome. The performances of the WHO algorithms for vaginal discharge syndrome and lower abdominal pain syndrome in women are known to be poor, and asymptomatic infections are missed by definition. Support for the introduction of point-of-care testing (POCT) is mounting but has been hampered by the lack of ASSURED (‘Affordable, Sensitive, Specific, User-friendly, Rapid and Robust, Equipment-free, and Deliverable to end users’) POCTs for some urogenital infections. This thesis describes the results of two studies that we conducted at a research clinic in Kigali, Rwanda, and the results of a systematic review. The first study (referred to as the Rwanda VMB study) investigated the impact of oral metronidazole and two different vaginal probiotics on the VMB and on BV recurrence. Sixty-eight HIV-negative and non-pregnant women with laboratory-confirmed BV and/or Trichomonas vaginalis (TV) received a seven-day oral metronidazole treatment course. We assessed their VMB before and after treatment by Gram stain Nugent scoring and 16S rRNA gene HiSeq sequencing (relative abundances) combined with 16S rRNA gene qPCR (to estimate bacterial concentrations). As expected, metronidazole decreased the overall bacterial load, and the relative abundances and concentrations of BV-anaerobes (all BV-anaerobes combined, as well as those of key BV-anaerobe taxa). However, metronidazole also increased the relative abundances and concentrations of lactobacilli (all lactobacilli combined, as well as increases in all key Lactobacillus taxa). Pre-treatment pathobionts concentration (defined as Proteobacteria, streptococci, staphylococci, enterococci, and a few others), and having a pre-treatment vaginal microbiota type containing more than 50% Gardnerella vaginalis (compared to BV-like VMB types with less than 50% G. vaginalis), were associated with increased likelihood of treatment failure, although the latter did not reach statistical significance (p=0.044 and p=0.084, respectively). All 68 women were subsequently randomised to four groups (n=17 each): behavioural counselling only (negative control group), intermittent use of oral metronidazole, intermittent use of Ecologic Femi+ vaginal probiotic capsule (containing multiple Lactobacillus and one Bifidobacterium species; Winclove Probiotics, Amsterdam, Netherlands), or intermittent use of Gynophilus LP vaginal probiotic tablet (containing L. rhamnosus 35; Biose, Arpajon-sur-Cere, France). Participants used the interventions for two months and VMB assessments were done at baseline, Day 7, Month 1, Month 2, and Month 6. Adherence was assessed by triangulating multiple adherence assessments and was taken into account in our analyses. All three interventions were safe and the preliminary efficacies were promising. BV (Nugent 7-10) incidence was 10.18 per person-year at risk in the control group, and lower in the metronidazole (1.41/person-year; p=0.004), Ecologic Femi+ (3.58/person-year; p=0.043), and Gynophilus LP groups (5.36/person-year; p=0.220). In mixed effects models adjusted for hormonal contraception/pregnancy, sexual risk-taking, and age, metronidazole and Ecologic Femi+ users, each compared to controls, had higher Lactobacillus and lower BV-anaerobes concentrations and/or relative abundances, and were less likely to have a dysbiotic VMB type by sequencing. However, inter-individual variability was high and effects disappeared soon after intervention cessation. We also assessed the acceptability of the interventions using structured questionnaires and semi-structured focus group discussions and in-depth interviews and found that acceptability was high. Finally, we identified a need for information, education and counselling campaigns to inform at-risk women about steps they can take to prevent BV, VVC, and STIs. In a systematic review, we evaluated the impact of lactobacilli-containing vaginal probiotics on BV and VVC cure and/or recurrence, as well as VMB composition and vaginal detection of probiotic strains, to compare the results of our own study with the published literature. We identified 34 studies evaluating 22 different vaginal probiotics. Unfortunately, most identified studies were under-powered and of poor methodological quality (for example, only two studies – including our own – differentiated between probiotic and autologous lactobacilli), and methodological heterogeneity between studies was high. Our conclusions were that vaginal lactobacilli-containing probiotics are safe and hold promise for BV cure and prevention, but are much less promising for VVC cure and prevention. Vaginal detection of probiotic strains never lasted long beyond the dosing period, suggesting that the probiotic lactobacilli used to date do not colonise the vagina. The aim of the second study (referred to as the WISH study) was to compare the performances of vaginal discharge/lower abdominal pain algorithms incorporating POCTs (the WISH algorithms), with traditional WHO syndromic management algorithms, and with nucleic acid amplification test (NAAT; gold standard) results. The POCTs we evaluated included the EcoCare vaginal pH swab (Merete Medical, Luckenwalde, Germany; pH≥5.0 was considered BV) and the TV OSOM assay (Sekisui Diagnostics, Lexington, USA) for BV and TV, both regardless of symptom-reporting. Women with a positive risk score were POC-tested for Chlamydia trachomatis and Neisseria gonorrhoeae (CT/NG; by GeneXpert, Cepheid, Sunnyvale, USA). VVC was treated presumptively based on reported symptoms. We enrolled 705 Rwandan women at risk of STIs, regardless of HIV and pregnancy status, and regardless of the presence or absence of urogenital symptoms. The study showed that POCT integration was feasible in a Rwandan setting, and acceptable to participants and staff. NAAT-based urogenital infection prevalences were: CT 8.5%, NG 7.1%, TV 16.1%, BV 18.1%, and VVC 8.6%. Infection-specific sensitivities of the WHO syndromic management algorithms compared to the gold standards ranged from 58.3-74.6%, and specificities from 44.7-50.6%. WISH POCT-based algorithms had good sensitivity (68.5-76.0%) and specificity (97.4-100%) for CT, NG, and TV but low specificity for BV (41.2%; sensitivity 95.2%), and modest sensitivity (64.4%) and specificity (69.4%) for VVC. Sensitivity (73.6%) and specificity (100%) for BV can be improved by screening all women for vaginal pH and conduct confirmatory testing (such as Nugent scoring or a molecular test) in women with a vaginal pH≥5.5 (n=275 in the WISH study). Based on these data, we concluded that existing POCTs should be introduced when feasible, and that POCT development should continue, including the development of ASSURED POCTs with multiple testing targets. Taken together, the data in this thesis suggest that urogenital infection epidemics in sub-Saharan Africa could potentially be reduced by the introduction of existing POCTs, the development of additional ASSURED POCTs, the expansion of the evidence-base for lactobacilli-containing vaginal probiotics using improved trial designs that incorporate molecular techniques, and the development of improved vaginal probiotics and other therapies to cure and/or prevent BV. Such new tools may have the potential to ultimately lower incidence of HIV and adverse pregnancy outcomes in at-risk women.