Targeted point-of-care testing compared with syndromic management of urogenital infections in women (WISH): a cross-sectional screening and diagnostic accuracy study

Summary

Background

Sexually transmitted and urogenital infections are typically managed by WHO-recommended syndromic algorithms in resource-poor countries, and presumptively in Europe. However, algorithms for vaginal discharge and lower abdominal pain perform poorly in women. The women's improvement of sexual and reproductive health (WISH) study in Kigali, Rwanda, sought to improve case-finding and infection management in women by introducing point-of-care tests. The main aim was to compare the performance of the WISH algorithms and the WHO vaginal discharge and lower abdominal pain algorithms with gold standard testing.

Methods

This cross-sectional screening and diagnostic accuracy study recruited women aged 18 years or older with or without urogenital symptoms at risk of acquiring sexually transmitted infections in Kigali, Rwanda. Recruitment activities were implemented by study staff with the help of community mobilisers at health centres, pharmacies, markets, women's organisations, and at “umuganda” community meetings. At the study visit, participants had a face-to-face interview that included questions about current urogenital symptoms. Participants were first asked without prompting (spontaneous reporting), followed by questions about 14 specific symptoms (structural reporting). Next, the WISH algorithms were implemented. All participants had point-of-care tests for bacterial vaginosis (vaginal pH of 5·0 or above) and Trichomonas vaginalis (immunoassay) regardless of symptom reporting. Women with a positive risk score had point-of-care tests for Chlamydia trachomatis and Neisseria gonorrhoea (nucleic acid amplification tests). Vulvovaginal candidiasis was treated presumptively. Nucleic acid amplification tests for C trachomatis, N gonorrhoeae, T vaginalis, bacterial vaginosis, and vulvovaginal candidiasis were the gold standard, and all patients provided swabs for these.

Findings

Participants were recruited between July 5, 2016, and March 14, 2017. 705 participants were enrolled in the study and completed a study visit, and 51 attended 53 additional visits. Prevalence by gold standard testing was 8·5% for C trachomatis, 7·1% for N gonorrhoeae, 16·1% for T vaginalis, 18·1% for bacterial vaginosis, and 8·6% for vulvovaginal candidiasis. The WISH algorithms identified similar numbers of C trachomatis, N gonorrhoeae, and T vaginalis infections, but much higher numbers of bacterial vaginosis and vulvovaginal candidiasis infections. Compared with gold standard testing, the WISH algorithms had a good sensitivity and high specificity for C trachomatis (sensitivity 71·7%, specificity 100%), N gonorrhoeae (sensitivity 76·0%, specificity 100%), and T vaginalis (sensitivity 68·5%, specificity 97·4%), high sensitivity but low specificity for bacterial vaginosis (sensitivity 95·2%, specificity 41·2%), and moderate sensitivity and specificity for vulvovaginal candidiasis (sensitivity 64·4%, specificity 69·4%). The performance of vaginal pH testing for bacterial vaginosis improved by increasing the cutoff to 5·5, followed by confirmatory testing (sensitivity 73·6%, specificity 100%). The WHO algorithms had moderate sensitivity and poor specificity for all infections compared with gold standard testing: C trachomatis sensitivity 58·3%, specificity 44·7%; N gonorrhoeae sensitivity 66·0%, specificity 45·2%; T vaginalis sensitivity 60·4%, specificity 45·6%; bacterial vaginosis sensitivity 61·6%, specificity 46·0%; and vulvovaginal candidiasis sensitivity 74·6%, specificity 50·6%. Two participants attended additional visits because they had a mild allergic reaction to metronidazole. Staff and participants considered point-of-care testing feasible and acceptable.

Interpretation

Point-of-care testing for urogenital infections might improve case-finding and infection management and is feasible in resource-poor settings. Point-of-care tests should be further developed, including those targeting multiple conditions. Additional studies in other populations, including populations with low prevalence of sexually transmitted and urogenital infections, are warranted.

Funding

European and Developing Countries Clinical Trials Partnership.

Introduction

Sexually transmitted infections and other urogenital infections cause a major burden of disease worldwide.¹ WHO estimated that 357 million new curable infections caused by Chlamydia trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis, and Treponema pallidum occurred in 2012.² In women, bacterial vaginosis, vulvovaginal candidiasis, and urinary tract infections are also common.3, 4 Long-term sequelae include increased risk of HIV acquisition and transmission, pelvic inflammatory disease, pregnancy complications, and invasive neonatal infections.4, 5, 6

Most resource-poor countries diagnose genital infections syndromically, using local guidelines that are based on WHO guidelines for the management of sexually transmitted infections.7, 8 Each patient-reported symptom, potentially augmented by clinician-observed signs during a physical examination, is treated for all organisms that might cause that symptom.⁷ In women, the four main syndromes on which the WHO algorithms are based are vaginal discharge without lower abdominal pain, lower abdominal pain with or without vaginal discharge, genital ulcers with or without inguinal buboes, and inguinal buboes without genital ulcers.⁷

The vaginal discharge syndrome is the most common.8, 9 WHO recommends that women with vaginal discharge should always be treated for bacterial vaginosis and trichomoniasis, and for chlamydia and gonorrhoea if local prevalence is high or if locally designed risk assessments are positive. WHO recommends additional treatment for vulvovaginal candidiasis if the discharge is curd-like or is accompanied by vulval oedema, erythema, or excoriations. In Europe, most sexually transmitted infections and urogenital infections in women are treated presumptively by primary care physicians.

By definition, syndromic and presumptive approaches miss all asymptomatic infections. Asymptomatic infections in women are common and are associated with the complications outlined above.4, 10 Furthermore, many studies in different countries have shown that the performance of algorithms for vaginal discharge and lower abdominal pain in symptomatic women are suboptimal, leading to undertreatment, overtreatment, or inadequate treatment of patients.9, 11, 12, 13

The women's improvement of sexual and reproductive health (WISH) study in Kigali, Rwanda, sought to improve case-finding and infection management in women by introducing point-of-care tests (POCTs).¹⁴ Our aim was to use POCTs that comply with WHO ASSURED criteria (affordable, sensitive, specific, user-friendly, rapid and robust, equipment-free, and deliverable to end users) as much as possible.15, 16 Participants were first asked about urogenital symptoms as if we were to provide them with syndromic care, but were then offered the POCT-based WISH algorithms. Stored clinical samples from all women were also tested by gold standard nucleic acid amplification tests (NAATs). This study design allowed us to compare the performance of the WISH algorithms and the WHO vaginal discharge and lower abdominal pain algorithms with gold standard results, to evaluate the feasibility and acceptability of the WISH algorithms, recommend optimal algorithms given currently available POCTs, and identify POCT development gaps.