Prevalence of Trichomonas vaginalis infection and protozoan load in South African women: a cross-sectional study

Objectives Trichomonas vaginalis is thought to be the most common non-viral sexually transmitted infection worldwide. We investigated the prevalence, risk factors and protozoan load of T. vaginalis infection in South African women. Methods A cross-sectional study of 604 women was conducted at 25 primary healthcare facilities in rural South Africa (Mopani district). T. vaginalis DNA was detected in vaginal and rectal swabs. In univariate and multivariate analyses, the T. vaginalis infection was investigated in relation to demographic characteristics, medical history and behavioural factors. The T. vaginalis load was determined as the logarithm of DNA copies per microlitre sample solution. Results Collected vaginal and rectal swabs were tested for T. vaginalis DNA. Prevalence of vaginal T. vaginalis was 20% (95% CI 17.0% to 23.4%) and rectal 1.2% (95% CI 0.6% to 2.4%). Most women (66%) with a vaginal infection were asymptomatic. Factors associated with T. vaginalis infection were a relationship status of single (OR 2.4; 95% CI 1.5 to 4.0; p<0.001) and HIV positive infection (OR 1.6; 95% CI 1.0 to 2.6; p=0.041). Women with vaginal T. vaginalis infection were more likely to have concurrent Chlamydia trachomatis rectal infection than those without vaginal infection (12%vs3%; p<0.001; OR 4.1). A higher median T. vaginalis load was observed among women with observed vaginal discharge compared with those without vaginal discharge (p=0.025). Conclusions Vaginal trichomoniasis is highly prevalent in rural South Africa, especially among single women and those with HIV infection, and often presents without symptoms.


IntrOductIOn
Trichomonas vaginalis is a protozoan parasite that causes trichomoniasis and is mostly sexually transmitted. It is the most common non-viral sexually transmitted infection (STI) worldwide. 1 The infection is asymptomatic in 85% of the women and in 77% of the men. 1 WHO describes that T. vaginalis in the African region occurs 10 times as often in women than in men. 2 Vaginal T. vaginalis infections in the entire African region are estimated to be 42.8 million. 2 If the infection is symptomatic, women may report a change in vaginal discharge, intermenstrual bleeding or vaginal blood loss during or after sexual contact. Other symptoms that can occur are vaginal itching, dysuria and abdominal pain. Also, upper reproductive tract disease syndromes can occur, including pelvic inflammatory disease (PID). Women with a T. vaginalis infection have 4.7-fold increase in the risk of PID [3][4][5] and tubal pathology. 6 Women infected with T. vaginalis also have a 1.3-fold increase in the risk of preterm labour. 5 In women, T. vaginalis infection can persist for months and, if left untreated, it can increase the risk for HIV acquisition if exposed. 6 7 T. vaginalis could lead to an increase in the vaginal HIV load and thus potentially increase the risk of HIV transmission to a sexual partner. 8 In addition, an association between T. vaginalis and Chlamydia trachomatis infection has been found. 9 Having a concurrent chlamydial infection was a predictor of both prevalent and incident T. vaginalis.
Despite the estimated large burden of T. vaginalis infection in the African region, data on clinical presentation, demographic and behavioural factors associated with infection, and microbiological factors are relatively limited. However, in other continents, multiple studies have been done in the context of T. vaginalis infections. [10][11][12][13] T. vaginalis prevalence has been reported for

Prevalence of Trichomonas vaginalis infection and protozoan load in South African women: a cross-sectional study
Dewi J de Waaij, 1,2 Jan Henk Dubbink, 1,2 Sander Ouburg, 1  Open Access a few African countries and ranges from 6.5% to 40%. 14 One of the explanations for the often high T. vaginalis prevalence is the lack of STI screening programmes and limited control measurements. The estimated burden of disease is significant but available data of symptomatology, coinfections and pathogen load are limited. C. trachomatis and Neisseria gonorrhoeae are other important causative micro-organisms for STI, showing an association with the pathogen load as compared with the clinical presentation and the development of long-term sequelae. 15 16 In this study, we present the epidemiology of T. vaginalis infection including the T. vaginalis load, as measured by a validated nucleic acid amplification technique. We studied a high HIV prevalence setting in rural South Africa.

Population and design
This study was part of a cross-sectional study of 604 women in rural Mopani District, South Africa, as described previously. 17 In brief, women aged 18-49 years who reported sexual activity during the prior 6 months were recruited at 25 primary healthcare facilities, regardless of the reason for visiting the facility that day. After informed consent, demographic characteristics, medical history and details of sexual behaviour were collected using a questionnaire. Healthcare workers collected vaginal and rectal swabs for molecular testing of T. vaginalis.
The Human Ethics Research Committee of the University of the Witwatersrand, South Africa, approved the study (Ref M110726).

Laboratory testing
Collected vaginal and rectal swabs were analysed at the Laboratory of Medical Microbiology and Infection Control, VU University Medical Center, Amsterdam, The Netherlands. High Pure PCR Template Preparation Kit (Roche Diagnostics, Basel, Switzerland) was used for the extraction of bacterial DNA. Detection of T. vaginalis DNA was done with the Presto plus assay (Microbiome, Amsterdam, The Netherlands) and the LightCycler II 480 (Roche Diagnostics) was used for the quantitative PCR for C. trachomatis, N. gonorrhoeae and T. vaginalis. 18 The Presto plus assay for C. trachomatis, N. gonorrhoeae and T. vaginalis is currently for inhouse use only and it will be transformed into a Presto-TV assay (Goffin Molecular Technologies, The Netherlands) in 2017 in a 200 reaction format as CE-IVD certified kit. PCR positivity was determined using the crossing point (Cp) value: we used a threshold Cp of <38 for a sample to be classified as positive. Samples with Cp values of ≥38 were retested and if again ≥38 were excluded from further analyses (indeterminate samples).

data analyses
The following risk factors were used in our analyses: age, being single, unemployment, current pregnancy, history of vaginal discharge syndrome, currently being pregnant, the use of hormonal contraceptives, administering intravaginal cleansing, visiting bars, alcohol use, concurrent partners, condom use during last sex act, partner >10 years older, experiencing coercion and/or force, and sex for money or for other benefits. 17 Coercion reflects sex inequality in relationships which is associated with increased risk. 19 Data were double entered in EpiData (Epi Info V.3.5.3) by two researchers independently, were compared for inconsistencies, cleaned and verified. The data were analysed using SPSS Statistics V.20.0 (IBM). Dichotomous data were compared using χ 2 test or Fisher's exact test and continuous data using the Mann-Whitney U test. ORs with 95% CI were calculated. Vaginal T. vaginalis infection was defined symptomatic when either abnormal vaginal discharge during physical examination, intermenstrual bleeding or vaginal blood loss during or after sexual intercourse was reported. Univariate analysis was performed to examine possible associations of demographic variables and clinical and behavioural factors in relation to T. vaginalis. Multivariate analysis was conducted using a backward selection procedure and likelihood ratio tests. As a proxy for the T. vaginalis load, we used the median Cp value, which was calculated for each risk factor and compared between infected versus non-infected women. resuLts demographics of study population A quarter of all women (n=154; 25.5%) had an age 18-24, over half of the women were single (334; 55.3%), 26 women reported anal intercourse (4.3%), a high proportion was unemployed and almost 30% reported to be infected with HIV (177; 29.3%). For more details, see table 1.

determinants of T. vaginalis load
The median Cp value for vaginal T. vaginalis load of infection was 28.0 (range 17.0-37.9). We did not observe any significant associations between the T. vaginalis load and demographic factors (table 2). Also, no association was found between the T. vaginalis load and symptoms in general. However, in the group of women with T. vaginalis infection who had vaginal discharge observed during examination, a significantly higher load was observed as compared with those without vaginal discharge (p=0.025).

dIscussIOn
In this study, we describe the prevalence, risk factors and the load of T. vaginalis infection. This is one of a few studies to report these aspects for T. vaginalis in a cohort from South Africa. We observed a vaginal T. vaginalis prevalence of 20% (95% CI 17.0% to 23.4%) and a rectal prevalence of 1.2% (95% CI 0.6% to 2.4%).
In multivariate analysis, two risk factors for vaginal T. vaginalis infection were identified: having a relationship status that is single and the presence of HIV infection. In addition, within women with vaginal T. vaginalis infection, a rectal C. trachomatis infection is significantly more present. Finally, a high T. vaginalis load appears to be associated with vaginal discharge observed during examination.
The prevalence of vaginal T. vaginalis infection in our study corresponds with the range reported by other studies in Africa (3.2%-40%). 14 20 Rectal T. vaginalis infections are rarely described. One study from San Francisco, USA, describes a rectal T. vaginalis prevalence of 0.6%, based on 'men who have sex with men' who reported receptive anal sex within 6 months prior to their clinic visit. 21 22 We show two demographic factors that may contribute to the high vaginal T. vaginalis prevalence in our study. First, we observed that a T. vaginalis infection is associated with a single relationship status. Joffe et al described that a single marital status is often associated with having more than one sexual partner and thereby an increased risk for STI in general. 23 To be more specific, it has been demonstrated that women having more than two concurrent male sexual partners are more likely to have a T. vaginalis infection. 24 Although we were not able to confirm this, we hypothesise that a single marital status is associated with having more than one concurrent sexual partner. It has been suggested that T. vaginalis infection is age tied. 25 26 Women, who were older than 20 years of age, were more likely to be infected with T. vaginalis. 27 In our study, we did not observe an association between age and T. vaginalis infections.
We confirm an association of positive HIV status with presence of T. vaginalis infection, which has been reported by several other studies. 4 8 28-31 T. vaginalis infection was present in almost a quarter of HIV-infected women in our study. This is in line with the previously described higher Trichomonas prevalence in high HIV prevalence cities compared with low HIV prevalence cities. 14 20 The biological link between HIV and TV is yet unclear; however, bacterial vaginosis is described to be an independent risk factor for acquisition of STI, and acquisition and transmission of HIV. 32 Although we did not include bacterial vaginosis, it is described to be associated with a 1.5-fold to 2-fold increased risk for incident trichomonal infection. 33 This could suggest an association between HIV and T. vaginalis where an asymptomatic T. vaginalis infection could increase the risk of HIV infection.
It has been described that non-sexual transmission of T. vaginalis occurs in girls who did not have sexual intercourse. 34 35 A striking prevalence of 43% was found in a study among women from Ndola, Zambia, who denied ever had sexual intercourse. 14 In those cases, T. vaginalis would be transmitted by wet clothes and water, since some studies suggest that T. vaginalis can survive up to days in warm mineral water and also for hours in urine. 34 T. vaginalis is described to have some kind of robustness, which may explain its different transmission ways compared with other STIs. This may be an explanation of many hypotheses for the high prevalence; however, they need to be further explored.
Twenty-six per cent of all women reported symptoms, without being diagnosed with a T. vaginalis infection. Although 20% of these women were diagnosed with another vaginal STI than T. vaginalis, symptoms are evidently a poor predictor for a T. vaginalis infection. We did not find a significant difference in symptomatology and T. vaginalis infection status. More knowledge about T. vaginalis disease would provide valuable information for potential targeted treatment approaches versus syndromic management.
In general, self-reported symptoms are not reliable in predicting an STI, like T. vaginalis. 36 37 However, 31 (16%) of the women vaginally infected with T. vaginalis were coinfected with either C. trachomatis or N. gonorrhoeae (data not shown). Symptoms of these frequent occurring STI may show comparable symptoms and should be taken into consideration. Although other studies revealed significant rates for coinfection between T. vaginalis and other STIs, 38 in the current study we did not observe an increased risk for coinfection with C. trachomatis and/or N. gonorrhoeae in patients infected with T. vaginalis (data not shown). However, in a previous study from our group, we describe an increased risk of obtaining a T. vaginalis coinfection when infected with either C. trachomatis or N. gonorrhoeae. Concurrent STIs may be due to common risk factors for these infections, such as sexual activity, contraceptive use and new partners. This is one of a few studies to report aspects of rectal T. vaginalis in a cohort from South Africa. We observed that rectal C. trachomatis infections are significantly more present in combination with a vaginal T. vaginalis infection. This may be due to several factors. First, sexual behaviour including anal sex is an important factor that may play a role in the presence of coinfections. Second, there is the ability to get reinfected via the gastrointestinal tract by autoinoculation, after being cured of the vaginal infection for Chlamydia, but it is unclear if this route of infection is also present for T. vaginalis infections and thus if the rectal site could be the reservoir for T. vaginalis infection. 39 These potential transmission routes may result from vaginal wiping techniques, poor personal hygiene, using fingers during sex or for cleansing afterwards and use of sex toys. 17 The most likely explanation is that the rectal samples may be contaminated by cervicovaginal fluid due to the close anatomical proximity 40 41 and that T. vaginalis positive rectal samples in women were contaminated with T. vaginalis positive vaginal fluid of the perianal surface. 42 It is conceivable that micro-organism load may affect the clinical course of infection. It has been shown in studies of another STI, C. trachomatis, that there is an association between the bacterial load and genital manifestations of infection. 43 44 A higher C. trachomatis bacterial load is associated with the presence of at least two specific symptoms (vaginal discharge, irregular bleeding, or pelvic or abdominal pain). The impact of T. vaginalis load has not yet been described. We hypothesised that T. vaginalis load may affect the clinical course of the infection and compared several factors with the load of T. vaginalis. We did not find any significant association between the load and reported symptoms of a T. vaginalis infection. However, within women with observed vaginal discharge during examination, a significantly higher load was found. The load determination was performed using Cp values of the T. vaginalis PCR. Another technique to determine the average load per human is to incorporate the number of human cells. Our results prove that there is an association between T. vaginalis load and discharge. However, the literature on T. vaginalis load is scarce, and the effect of the load has to be further explored.
Although this is a unique study about the epidemiology and microbiological characteristics of T. vaginalis in rural South Africa with a relatively good sample size, a limitation of our study is that the number of rectal T. vaginalis infections is rather small for further analyses. As a consequence, this could have played a role in our statistical analyses. In addition, there can be a selection bias based on for instance the geographical inclusion, and self-reported symptoms can be more subjective than clinician-based symptomatology. We cannot exclude potential cross-contamination between the cervicovaginal and anorectal sites. The results of this may be generalised to similar populations; however, these results might not be extended to other populations due to the aforementioned potential geographical and reporting biases.
In conclusion, T. vaginalis infections are highly prevalent in the rural South African population investigated. The asymptomatic nature, potential for non-sexual transmission especially in socioeconomic poor settings and the syndromic management approach to STI control in this region are potential factors contributing to the prevalence. We observed a low rectal prevalence, part of which might be contamination with T. vaginalis positive vaginal fluid of the perianal surface. In multivariate analyses being single or HIV positive was identified as a risk factor for vaginal T. vaginalis infection. Finally, a high T. vaginalis load seems to be associated with vaginal discharge observed during examination.
contributors DJdW: laboratory analyses, statistical analyses and writing of the manuscript. JHD: sample collection and writing of the manuscript. SO: statistical analyses, reviewing of the manuscript and study supervision. RPHP: sample collection, reviewing of the manuscript and study supervision. SAM: reviewing of the manuscript and study supervision. Provenance and peer review Not commissioned; externally peer reviewed. data sharing statement All data of this study have been described in the paper. No additional data for this study are available.
Open Access This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http:// creativecommons. org/ licenses/ by/ 4. 0/