Objectives The BASHH guidelines recommend molecular tests to aid diagnosis of Trichomonas vaginalis (TV) infection; however many clinics continue to use relatively insensitive techniques (pH, wet-prep microscopy (WPM) and culture). Our objectives were to establish a laboratory pathway for TV testing with the Becton-Dickinson Qx (BDQx) molecular assay, to determine TV prevalence and to identify variables associated with TV detection.
Methods A prospective study of 901 women attending two urban sexual health services for STI testing was conducted. Women were offered TV BDQx testing in addition to standard of care. Data collected were demographics, symptoms, results of near-patient tests and BDQx results for TV, Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (GC). Women with any positive TV result were treated and followed up for test of cure (TOC).
Results 901 women had a TV BDQx test. 472 (53%) were white, 143 (16%) black and 499 (55%) were symptomatic. Infections detected by BDQx were: 11 TV (1.2%), three GC (0.3%) and 44 CT (4.9%). Of the 11 BDQx-detected TV infections, 8 (73%) were in patients of black ethnicity. Of these, four of seven cases (57%) were WPM-positive. All patients received treatment and nine of nine (100%) were BDQx-negative at TOC. In univariate analysis, only black ethnicity was associated with likelihood of a positive TV BDQx result (relative risk (RR) 10.2 (95% CI 2.15 to 48.4)).
Conclusions The use of the BDQ enhanced detection of TV in asymptomatic and symptomatic populations. Cost-effective implementation of the test will rely on further work to reliably detect demographic and clinical variables that predict positivity.
- EPIDEMIOLOGY (CLINICAL)
Statistics from Altmetric.com
Trichomonas vaginalis (TV) is a sexually transmitted protozoan infection. In 2012, there were 6600 incident cases reported in genitourinary (GU) medicine clinics in England, UK.1 At a population level, however, it is a rare infection with one recent study estimating a population prevalence of 0.3%.2 In this same study it is also noteworthy that prevalence varied greatly according to ethnicity with women of black and mixed ethnicity being disproportionately affected (prevalence 2.7%). In sexual health clinic attendees, prevalence is higher still with the infection affecting 1.4% of men and 3.6% of women (9% in black women) in a study of 3503 patients testing with a nucleic acid amplification technique (NAAT).3
The diagnosis of TV can be made using direct wet-prep microscopy (WPM), culture and, more recently, NAATs.4 A recent large prospective trial comparing TV NAAT with TV culture and WPM revealed sensitivities of 97.4%, 98.3% and 68.7%, respectively (p<0.001).5 Regarding clinic-based screening, a prospective cross-sectional study of Sydney sexual health clinic attendees published in 2010 revealed a similar increase in the identification of TV using NAAT testing when compared with WPM.6
Our STD clinics currently test for TV only in symptomatic women, despite serving an ethnically diverse West London population. At clinic A, testing comprises vaginal pH screening followed by WPM when pH is >4.5. In clinic B, screening is performed using WPM and TV culture for WPM-negative samples. We wished to evaluate the use of the Becton-Dickinson ProbeTec Qx assay (BDQx) TV NAAT in all women attending our services. This NAAT assay can detect infection with TV, Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG). This requires a single sample (eg, clinician-taken endocervical or high vaginal swab; a self-collected vulvovaginal swab or a urine sample). In the laboratory, separate kits are used for CT/NG detection and TV detection.
The aims of this evaluation study were as follows:
To determine the prevalence of TV among symptomatic and asymptomatic female patients attending our GU medicine services.
To identify epidemiological factors associated with TV infection in our clinic populations in order to inform our future testing strategies.
To determine the proportion of TV cases undiagnosed with current screening and testing strategies (including sensitivity of both pH testing and WPM).
To describe laboratory pathways for TV BDQx testing and to determine result turnaround time.
To describe the utility of BDQx test of cure (TOC) in patients treated with first-line treatments.
We conducted a prospective study of women attending clinics A and B over a period of 1 month in July 2015. The projected sample size was 900 women. To be included, women had to be aged ≥18 years and eligible for STI screening. We excluded women who reported any treatment with metronidazole, tinidazole or other treatment against TV in the 2 weeks prior to presentation. We also excluded women who had used vaginal foams or gels within 24 hours of presentation to clinic.
Verbal consent was obtained for participation and swabs were collected according to standard screening protocols. For asymptomatic women, this included a nurse-collected or self-collected vulvovaginal swab. Symptomatic women underwent examination and near-patient testing. All patients had a sample taken with a plastic loop for a vaginal gram stain. In clinic A, pH testing was performed on this sample and if the pH was >4.5, WPM was performed. In clinic B, WPM was performed without prior pH testing and negative samples were sent for TV culture (vaginal swab inoculated into Oxoid Trichomonas medium, incubated at 37°C for 48 hours and examined under light microscopy at ×40 power). In both clinics, a second swab was then used for vaginal CT/NG testing. No additional swab was required for TV BDQx testing as this was performed on the CT/NG sample. Women with any positive TV result were treated according to BASHH guidelines and all were invited to attend for TOC 7 days after treatment.4 Those women who were compliant with treatment and asymptomatic following treatment were not offered a TOC. Male partners of study participants were contacted and tested/treated appropriately.
We collected the following data prospectively: demographics, symptoms, results of near-patient tests and BDQx for TV, CT and NG. Data were collated in Excel and analysed using SPSS (V.22; SPSS, Chicago, Illinois, USA).
Over the study recruitment period (July–August 2015), 901 women were recruited to the study and underwent BDQx testing for TV. A total of 451 patients were recruited at clinic A and 450 at clinic B. Of these 901 women, 472 (53%) were white (NHS ethnicity codes A–C); 143 (16%) were black (codes D, E, M, N and P which includes mixed race black and any other ethnicity).7 The median age of study participants was 28 years (range 18–67 years) and 499 (55%) were symptomatic. One woman was living with HIV-1 infection.
Infections detected by the BDQx assay comprised 11 cases of TV (1.2% (95% CI 0.49% to 1.91%)), three cases of NG (0.3% (95% CI 0.00% to 0.66%)) and 44 cases of CT (4.9% (95% CI 3.49% to 6.31%)). The prevalence of TV detected by BDQx was 1.4% (7/499; (95% CI 0.03% to 1.97%)) in symptomatic women and 1.0% (4/401; (95% CI 0.37% to 2.43%)) in asymptomatic women.
Of the 11 TV BDQx-positive cases, eight (73%) were among women of black ethnicity and seven (64%) of the patients had presented with symptoms. Of these symptomatic cases, four of seven (57%) had received a diagnosis of TV at the time of screening following a positive WPM result. Three of these women were recruited at clinic A had a vaginal pH of >4.5. All seven symptomatic TV BDQx-positive women had a Hay-Ison grade 2 vaginal gram stain result, and one of the three patients recruited at clinic B had a positive vaginal TV culture result. Table 1 summarises the demographics and test results of the 11 TV cases.
There were no negative TV BDQx tests in the presence of positive WPM or TV culture.
In a univariable logistic regression model including the whole study population, the only significant association with TV BDQx positivity was with black versus non-black ethnicity (relative risk (RR): 10.2 (95% CI 2.15 to 48.4)). In a subanalysis of symptomatic women, black ethnicity remained predictive of TV BDQx positivity (RR 13.73 (95% CI 1.64 to 115.25)).
Mean turn-around time was 3.44 days (SD: 1.48 days). All women with a positive TV BDQx received treatment and nine of nine (100%) were BDQx-negative at TOC (mean time to TOC: 15 days (range: 7–42 days)).
We identified a prevalence of TV infection of 1.2% in our clinic populations. This exceeded that of NG infection in the same population by fourfold. The statistical power of the study was sufficient to demonstrate a clear association of TV infection with black ethnicity. However, relatively few events may have limited the ability to detect other associations.
While data concerning the uptake of testing were not collected prospectively, use of the BDQx triple NAAT meant the test was highly acceptable to patients and staff and integrated well within existing clinic and laboratory pathways. This was demonstrated by very quick (1 month) recruitment to the study target of 900 patients.
In symptomatic women, there were no cases of TV identified by WPM or culture that were not identified by the BDQx assay. Conversely, of TV BDQx-identified TV infections, only 57% and 33% of cases were identified by WPM and culture, respectively. We cannot make these comparisons in the asymptomatic population as WPM and TV culture were not performed. While a proportion of the women diagnosed with TV in this study would not have been diagnosed or treated due to poor sensitivity of routine diagnostics, a larger prospective study would be necessary to design clinically and economically effective TV screening algorithms.
In conclusion, the prevalence of TV infection was high in particular subgroups, but we are limited in making firm recommendations for screening by the relatively few TV cases identified. This is the main study limitation. Based on the data presented, however, we do recommend TV NAAT screening in all asymptomatic women of black and mixed black ethnicity and all symptomatic women with vaginal discharge attending our services.
The study team thank Lisa Simonds and Lee Pitty for their assistance in the laboratory.
Handling editor Jackie A Cassell
Contributors CT and MRa conceived the study, wrote the protocol, obtained regulatory approval, analysed study data (in conjunction with SM) and wrote the first draft of the manuscript. LW and SO’N co-wrote the study protocol, recruited patients to the study and collected clinical data during the conduct of the study. They provided edits for the draft manuscript. LH collected trial data and contributed to the manuscript edits. LH collected trial data and contributed to the manuscript edits. MRe, AO’R and JM conducted the laboratory analyses for the study and wrote the laboratory aspects of the manuscript. DW and RJ were the principal investigators for the two study sites and approved the final draft of the manuscript.
Funding Tests were provided at a reduced rate for evaluation by Becton Dickinson Ltd (New Jersey, USA).
Competing interests None declared.
Ethics approval Bloomsbury Research Ethics Committee (REC) (ref: 15LO 0457).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.