Trichomonas vaginalis infection and the diagnostic significance of detection tests among Ghanaian outpatients

Background There is little data on Trichomonas vaginalis infection in Ghana. This study evaluated the prevalence of trichomoniasis using different diagnostic methods and determined the risk factors for infection in patients. Methods A structured questionnaire was administered. Vaginal swabs, urethral swabs and urine specimens were obtained from consenting patients; and the samples processed following standard protocols. The presence of T. vaginalis was determined using wet mount microscopy and polymerase chain reaction (PCR) as gold standard. We also assessed the diagnostic performance the JD’s Trichomonas V® rapid antigen test to inform clinical practice. Results The PCR assay detected T. vaginalis positivity in 64 of 150 patients (42.6, 95%CI:35.0, 50.6) including all positive samples of wet mount microscopy and JD’s Trichomonas V® test. Wet mount microscopy showed low sensitivity (31.6%), high specificity (100%), moderate positive predictive value (75.0%), moderate positive likelihood ratio (3.0), and weak agreement (Cohen’s kappa, 0.283) with PCR assay. The JD’s Trichomonas V® test displayed lower sensitivity (25.0%), specificity (83.3%), and weaker measure of agreement (Cohen’s kappa, 0.233) with PCR. In multivariate analysis, the strongest independent predictor for T. vaginalis was female gender [adjusted odds ratio (AOR), 24.89; 95% confidence interval (CI): 10.58, 51.21; P-value< 0.001]. Knowledge of STI showed a protective effect against infection with the parasite (AOR, 0.13; 95%CI: 0.07, 0.29; P-value< 0.017). Conclusion The sensitivity of wet mount microscopy was low for T. vaginalis screening in our region. The JD’s Trichomonas V® test should not be considered as an alternative test. We recommend mandatory PCR assay for confirmation of negative wet mount results.


Background
Trichomoniasis is a common sexually transmitted infection (STI) caused by the protozoan parasite Trichomonas vaginalis. The highest prevalence of trichomoniasis has been reported from resource-limited countries and among deprived populations in developed countries [1,2]. Of the estimated 153 million cases of trichomoniasis, 25.64% are from the African [1,3]. Trichomoniasis is mainly transmitted by sexual intercourse with an infected person and affects the urogenital system, vagina, and urethra [4,5]. The clinical presentations of the infection in women include yellowish-green frothy vaginal discharge with vulvar irritation which may be confused with bacterial vaginitis, whereas infected men may briefly have urethral irritation, mild discharge, or slight burning after urination or ejaculation [3,[6][7][8][9]]. Factors such as low level of education, poor hygiene, low socio-economic status, and multiple sexual partners have been associated with high prevalence of trichomoniasis [10,11].
There is paucity of data on T. vaginalis infection in Ghana [12,13]. The few available studies were conducted among women. And most of these studies were performed with wet mount microscopy. Wet mount microscopy is reported to have low sensitivity of 45-60% compared with polymerase chain reaction (PCR) tests [14,15]. Consequently, the occurrence of trichomoniasis and its associated public health implications in Ghana may be underestimated. Other methods such as culture are recommended but it takes a long time for diagnosis and still requires expert microscopy. Simple-to-perform rapid diagnostic tests are emerging globally as the preferred point-of-care diagnosis for protozoan infections albeit with varying sensitivities compared to wet-mount. The JD's Trichomonas V® antigen-based rapid diagnostic strip test (Biotech™ Corp., China) that was recently introduced in our district but with no data on its reliability. To inform practice, we assessed the diagnostic performance of JD's Trichomonas V® test for T. vaginalis infections, compared to wet mount microscopy, and using PCR as the gold standard. We also report on the associated risk factors for trichomoniasis in our setting.

Study settings
This study was conducted at the Comboni Hospital in the South Tongu District of Ghana. The District has an estimated population of 86,000 with an annual growth rate of 2.5% [16]. The hospital has an annual outpatient attendance of 47,528. It has a bed capacity of 50 with 8 doctors. The Comboni hospital has a laboratory that provides some microbiological services, including routine parasitological examinations, but does not perform bacterial culture and antibiotic sensitivity testing. The laboratory offers immunology, biochemistry, as well as haematology and transfusion services. The hospital receives referral cases from other health centres in the district.

Study design
We conducted a hospital-based prospective cross-sectional study from May to June 2014. Sampling involved two steps: (i) recruiting patients and interviewing them; and (ii) collecting study samples for laboratory investigations. The study participants comprised patients referred to the Comboni hospital for laboratory investigation of any type of sexually transmitted diseases (STD) whether or not they were sexually active. From this cohort, we investigated patients for T. vaginalis infections. Prior to the commencement of work, the research team conducted meetings with the Comboni hospital authorities, staff, and patients to explain the objectives of our study including benefits, potential risks and discomforts. Patients were recruited based on willingness to fully participate and provision of written consent. Informed written consent were provided by patients before enrollment into the study. For patients aged < 18 years, we sought for informed consent from parents or legal guardians as well as verbal consent by the patient to willingly partake in the study. Patients or legal guardians who provided informed consent also permitted that the data to be collected could be published in a peer-reviewed journal. A structured questionnaire (Additional file 1) was administered through interview to collect information on demographics, knowledge, attitudes and patients' practices that predispose them to T. vaginalis infections. Because patients were recruited at the laboratory, we conducted a physician-assisted work through of participants' hospital folders to ascertain signs and symptoms, confirm diagnosis and record treatment history. Patients on past (within 2 weeks) or current treatment for suspected sexually transmitted infections were excluded from study. Approval for the study (ID: MS-Et/M.5-P.3.3/2013-2014) was granted by the Ethical and Protocol Review Committee, University of Ghana Medical School, Korle-Bu.

Sample collection and processing
The primary outcome of the study was the detection of T. vaginalis in patients coming for STD checkup. Specimen collection and investigations were conducted following standard protocols but independent of the routine standard operative procedures for laboratory diagnosis of STD at the Comboni hospital. In our study, vaginal swabs were requested from females. Males provided urethral swabs. At the same time, about 15-20 mL of first catch urine samples were collected from both female and male participants. Swab specimen from each patient were agitated vigorously in 1 ml of physiological saline (0.85%). The urine samples were centrifuged at 1000 rpm for 5 min and decanted to ensure the viability of trichomonads for wet mount microscopy. All patients who provided samples and accompanying data were de-identified and allotted arbitrary numbers to ensure anonymity. Patients found infected with parasites as well as those found with ailments not targeted by the project were referred to hospital for treatment.

Wet mount microscopy
Aliquot of the swab elute and decanted urine samples per patient were subjected to wet mount microscopy within 20 min after specimen collection. They were examined for motile trichomonads using × 100 and × 400 magnification. Three slides were prepared for each specimen to increase chances of recovering T. vaginalis. Trichomonas vaginalis was recorded as present if ≥1motile trichomonads were seen at any magnification.

Rapid antigen test
The JD Trichomonas vaginalis rapid tests were performed according to manufacturer's (Biotech™ Corp., China). Briefly, 0.5 mL of the swab elute was mixed in 0.5 mL of test buffer (0.01% Tris-Hcl and 0.05% NaN3, PH 7.5) for 10 s. The JD's Trichomonas V® test strip was placed in each mixed test buffer, and the result read visually after 15 min. For urine sediments, 0.5 ml of sample buffer was added to 0.5 mL of the sediments and mixed well for about 10 s. For each urine-buffer test buffer, a JD's Trichomonas V® test strip test strip was placed in the mixture, and the result read visually after 15 min.

Polymerase chain reaction tests
Aliquots of swab elutes and urine specimens from each patient were separately placed in DNA guard (Biometrica, Co. USA) until further work-up. Polymerase chain reaction (PCR) targeting the beta-tubulin genes of T. vaginalis was used for the detection of the organism. The targeted genes encode the amino acid sequence of beta-tubulin protein, a major component of the T. vaginalis flagella and cytoskeleton [17]. The Qiagen DNA extraction kit (column method) was used to extract DNA. The primer set used was BTUB 9/2 [18] which targeted a conserved region of the beta-tubulin genes btub1, − 2, and − 3 (GenBank accession numbers: btub1, L05468; btub2, L05469; and btub3, L05470) of T. vaginalis. The primer set was designed to amplify a DNA product of 112 bp from the three beta-tubulin genes. We used the sequence of forward primer (BTUB 9, 5' CAT TGA TAA CGA AGC TCTTTA CGA T 3: positions 850 to 874) and reverse primer (BTUB 2, 5' GCA TGT TGT GCC GGA CAT AAC CAT 3′: positions 961 to 938). A final volume of 25 μl was used for PCR amplification reaction. This contained 10 μl of DNA template, 0.625 μl of each primer, 0.5 μl of each of the four dNTPs (Sigma, Co. USA), 0.125 μl of Taq polymerase (Biopioneers, Co. USA), 5 μl of PCR buffer (Biopioneers, Co. USA) and 6.625 μl nuclease-free water. The following PCR conditions were used: an initial denaturing step of one minute at 95°C followed by 40 cycles of denaturation at 95°C for 45 s, annealing at 52°C for 45 s and extension at 72°C for 1 min and then a final extension at 72°C for 2 min. The PCR products were run on 2% agarose gel with ethidium bromide and visualized in visualized under ultraviolet trans-illuminator (JL Berker, Germany).

Data analysis
Data were entered into a Microsoft Excel spreadsheet and summarized using graphs and tables. Statistical analyses were performed using Epi info™7 statistical software package (CDC, Atlanta. The definitive consideration for T. vaginalis infection was positivity by PCR in any of two included patient specimens. Discrepant specimens (those with different results outcome for vaginal/urethral swab negative and urine samples) were determined to be T. vaginalis positive by performing PCR. The gold standard was PCR positivity. The diagnostic performance of detection methods was calculated by using the defined gold standard with confidence intervals based on normal approximation to binomial distribution. Agreement measures between detection methods was determined with Cohen's kappa tests. For descriptive statistics we calculated percentages with 95% confidence interval (95% CI), mean ± standard deviation (SD). Univariate comparisons were computed with Chi-square tests and unadjusted Odds ratios (OR) at 95% confidence interval (CI). In multivariate analysis, logistic regression was conducted with adjusted odds ratios (AOR) using all variables that had p < 0.5 in the univariate comparisons. Point estimates of statistical significance were indicated with 2 tailed P-values < 0.05. Association between age and T. vaginalis was demonstrated with Locally Weighted Scatterplot Smoothing (LOESS) was generated using Tricube kernel smooth fit line with 60% of points to fit.

Results
Overall, 203 patients were eligible for inclusion. Of these, 150 provided informed consent and were enrolled in the study. There were 110 females and 40 male patients with a mean age of 28.67 years (ranging from 16 to 55 years). Out of this, 110 (73.33%) were females with a mean age of 27.81 years (ranging from 16 to 50 years), and 40 (26.67%) were males with a mean age of 31.05 years (ranging from 20 to 55 years). All included participants were sexually active. Among the female subjects, 45 (40.91%) were pregnant. Genitourinary symptoms were observed in 81 of 150 patients (females,71; male, 10). The commonest symptoms associated with the included cases was genital discharge (70.4%, n = 57/81, females, 54; males, 3), followed by itching in 61.7% (females, 49; males, 1) of the cases and painful micturition (females, 18; males, 8) in 32.1 of cases.

Sample type and test positivity
We compared tests positivity of vaginal/urethral swabs and urine by wet mount microscopy, the JD's Trichomonas V® test and PCR assay. The PCR detected could

Performance of diagnostic techniques
The diagnostic performance of wet mount microscopy and the JD's Trichomonas V® test were compared using results of the PCR assay as gold standard (Table 2). Wet mount microscopy showed low sensitivity (31.6%) but high specificity (100%) with moderate to high positive predictive values (75.0%) and positive likelihood ratio (3.0) for all patients. The microscopy recorded its best sensitivity score among asymptomatic (40.0%) and female (39.2%) patients. The least sensitivity (0%) was noted for male patients. In comparison, there was weak agreement (Cohen's kappa, 0.283) between wet mount microscopy and PCR assay. The kappa improved to 0.533 only when asymptomatic cases were considered. Across all patients, the JD's Trichomonas V® test displayed lower sensitivity (25.0%) and specificity (83.3%) compared to wet mount microscopy but with better positive predictive values (83.3%) and positive likelihood ratios (5.0). A weak agreement (Cohen's kappa, 0.233) was noted between the rapid antigen test and PCR assay-and this was evident across all patient groups. Interestingly, although wet mount microscopy and the JD's Trichomonas V® test had moderate to high specificities, their agreement was moderate (Cohen's kappa, 0.541).

Univariate comparisons
We compared patients' characteristics and their odds ratios (OR) by univariate analysis for potential risk factors of T. vaginalis infection. Trichomonas vaginalis infection was defined as test positivity to PCR assay for any of included patient's sample. The unadjusted odds  Age was considered as a continuous variable from 16 to 55 years significant risk factor for T. vaginalis infection. The multivariate analysis show that with every 1 year increase in age doubled the risk of infection with T. vaginalis (AOR = 1.93, 95%CI: 1.11, 3.90, p = 0.011). The relationship between T. vaginalis infection and age was best described with a LOESS Tricube Kernel smooth fit line with local polynomial regression at 60% (Fig. 1). The correlation between T. vaginalis and subject's age was more apparent in females and less dominant among the elderly.

Discussion
Routine attempts towards the detection and treatment of Trichomoniasis has gained prominence owing to its recent association with HIV, adverse gynaecological and obstetric outcomes and infertility [19][20][21][22]. Trichomonas infection attracts little attention in sub-Sahara Africa perhaps because most infected people do not show any symptoms and when left untreated serve as asymptomatic carriers [22,23]. Poor diagnostic techniques for T. vaginalis also compounds the burden. In this study, 150 patients visiting the laboratory for STI check-up were examined for T. vaginalis and among these, 64 (42.6%) were positive by PCR assay. There were two major findings. First, the prevalence of T. vaginalis infection is indeed high. Although this could be due to our select patient group, it is also an indication of the existence of the infection in the study region. This high prevalence is probably due to the fact that, the T. vaginalis problem has long existed in our community, and the lack of awareness and routine laboratory detection may have compounded the problem. Second, the results of the present study point to T. vaginalis as being prevalent among even males with no  genitourinary symptoms and not just females. It is very difficult to predict the T. vaginalis infection without clinical findings, as in our study only 33% (n = 10/30) and none (n = 0/39) respectively of the male and female patients with no genitourinary findings were positive for T. vaginalis. Overall, T. vaginalis positivity was high (48.3%) among female. The female gender also constituted significant risk factor for T. vaginalis with increased infection odds of 24.9 ( Table 5). The prevalence obtained for females in this study is slightly above the WHO estimate (18.12%) in the African region [1] although high sensitive detection methods were included in the latter. The prevalence of T. vaginalis has generally been considered to be low among men, with a global prevalence of 1-5% [1,24,25]. The variation in prevalence observed between males and females is attributed to biological differences, particularly oestrogen, between gender. Also availability of iron [26] may facilitate persistent T. vaginalis infection among females during reproductive years.
In communities with prevalent T. vaginalis infections, an ideal detection test must have the greatest ability to pick true positives without selecting false negatives. Our results concerning T. vaginalis diagnosis demonstrated poor agreement between PCR assay as gold standard and the wet mount microscopy or JD's Trichomonas V® test. Two observations merit attention. First, both wet mount microscopy and JD's Trichomonas V® test tests recorded low sensitivities albeit with moderate to high specificities. It is noteworthy that, a certain proportion of false positives is often acceptable if the sensitivity margins are signicantly high. Yet, our wet mount microscopy could only detect 25 of 64 T. vaginalis positives and the rapid tests, 28 of 64 T. vaginalis positives. The negative likelihood ratios for both tests ranged between 0.1 through 0.52 depending on the patient population. Such low detection rates are more apt to frequent misdiagnosis and underestimation of the T. vaginalis problem. Second, in the search for less invasive human samples for detection of T. vaginalis, researchers have evaluated the possible use of urine specimens [14,27]. In this current study, we found no discrimination in PCR test results between vaginal/urethral swabs and urine samples. Rather, extremely poor diagnostic outcomes were observed for wet mount microscopy and JD's Trichomonas V® test performed with urine samples as opposed to vaginal/urethral swabs. The addition of urine did not significantly improve the T. vaginalis detection.
It has been emphasized that routine implementation of the PCR test should be used for T. vaginalis diagnosis. But the detection of T. vaginalis by PCR in microbiology laboratories remains a contentious issue with some experts questioning its clinical relevance in view of the financial demands involved. More so the strategy involves the use of skilled expertise. In contrast, expert Fig. 1 Profile of Trichomonas vaginalis prevalence across patients' age with Tricube Kernel smooth fit line generated with local polynomial regression at 60% of points to fit. Over the period of study, T. vaginalis cases were predominant in females. There is apparent relationship between age and infection prevalence among females that peaks at approximately age 26 years microscopy is abundant sub-Saharan Africa and easily adapted in microbiology laboratories for diverse routine work. It remains the conventional detection method for routine diagnosis of parasites in STI across many sub-Saharan African countries. The results however suggest that wet mount microscopy may have low ability in our subset population to discriminate between those individuals with the disease and without the disease. The finding agrees well with submissions by other researchers [14,28]. The JD's Trichomonas V® test did not appear to be a better alternative either. Suffice to say that the poor performance of the rapid antigen test contrasts the higher sensitivity (100%) reported elsewhere [29] but the basis if this discrepancy is unknown.
The spectrum of selected patient population affects the prevalence of studies. In multivariate logistic analysis, we noted knowledge of STI to be protective against T. vaginalis and that this pattern was associated with increasing age and is prevalent among the female gender. Nearly a third of participants had no knowledge of STI. Perhaps the high occurrence of T. vaginalis could be attributed to lack of awareness and ignorance of the public health implication of the infection. This observation should however be interpreted with the awareness that a greater proportion of females than males consented to participate in the study. This may have overemphasized the T. vaginalis burden.

Conclusion
In conclusion, the sensitivity of wet mount microscopy was low, thus it is unreliable for Trichomoniasis screening among patients in our region. However, given that a negative wet mount does not rule out the infection with T. vaginalis, we suggest a mandatory PCR assay as gold standard laboratory method for confirmation of negative results. The JD's Trichomonas V® test should not be considered as an alternative test.