Comparison of Aptima Trichomonas vaginalis Transcription-Mediated Amplification Assay and BD Affirm VPIII for Detection of T. vaginalis in Symptomatic Women: Performance Parameters and Epidemiological Implications

Abstract

Trichomonas vaginalis is an underestimated sexually transmitted infection (STI) associated with numerous clinical sequelae. The true prevalence and clinical impact of trichomoniasis are unknown, as current methods of detection exhibit poor sensitivity compared to molecular amplification methods. Limited data exist comparing the BD Affirm VPIII hybridization assay to the Gen-Probe Aptima T. vaginalis (ATV) transcription-mediated amplification (TMA) assay for detection of T. vaginalis. In this study, specimens from 766 patients were evaluated. Specimens were retrieved consecutively from patients with vaginal complaints and/or with histories suggestive of STI. Study inclusion was dependent upon the request for and collection of both a vaginal swab for Affirm and a specimen for Aptima Combo 2 by the health care provider during the same office visit. Affirm was performed using the specific collection swab and the transport provided for the test. The ATV assay was performed on remnant Aptima Combo 2 specimens. A second ATV TMA assay, utilizing an alternate T. vaginalis primer and probe set, was performed on all specimens positive by the initial TMA and/or the Affirm assay. Infected-patient status was defined as positive T. vaginalis test results by at least 2 assays. Overall, 5.1% of subjects were positive for T. vaginalis. T. vaginalis was most prevalent in women who were 36 to 45 (11.9%), 51 to 60 (7.7%), and 16 to 25 (4.2%) years of age. The ATV assay was statistically more sensitive than the Affirm assay (100% versus 63.4%, P < 0.0001), identifying 36.6% more positive patients.

INTRODUCTION

Despite a worldwide prevalence rate likely to be double that of Neisseria gonorrhea and Chlamydia trachomatis combined, Trichomonas vaginalis is not currently a reportable disease in the United States (1, 7). Recent literature suggests an association between T. vaginalis infection and the sequelae of other sexually transmitted infections (STIs), including low birth weight, premature rupture of membranes, preterm labor, atypical pelvic inflammatory disease, infertility, prolonged carriage of human papillomavirus (HPV), and an increased risk for acquiring HIV (9, 10, 13, 18, 21, 31, 32). Current methods used for the diagnosis of T. vaginalis, including wet mount microscopy, rapid antigen testing, and culture, have been shown to have poor sensitivity compared to molecular amplification methods (6, 14, 17, 19, 24, 25, 27). In addition, variable performance occurs with antigen testing and wet mount microscopy, depending on whether patients are symptomatic or asymptomatic (17, 26). The BD Affirm VPIII (Affirm) assay (Becton Dickinson, Sparks, MD) is the only Food and Drug Administration (FDA)-cleared, direct-specimen, RNA probe-based diagnostic test designed to differentiate and identify pathogens associated with bacterial vaginosis (Gardnerella vaginalis) and vaginitis (T. vaginalis and Candida species). The Affirm test is widely used, yet limited data exist comparing this test with other molecular methods (5).

The purpose of this study was to evaluate the performance parameters of the BD Affirm assay with the Gen-Probe Aptima Trichomonas vaginalis (ATV) assay, a molecular diagnostic assay that uses transcription-mediated amplification (TMA) for the detection of T. vaginalis. The ATV assay recently completed clinical trials and has been submitted to the FDA for clearance. Samples were collected from a female population for whom Chlamydia trachomatis, Neisseria gonorrhea, and T. vaginalis testing was being performed due to presentation of symptoms of G. vaginalis or because of a history of STIs. T. vaginalis prevalence by age in the population tested was compared to those of C. trachomatis and N. gonorrhea. The population assessed is considered a low-prevalence population for both C. trachomatis and N. gonorrhea (28).

(Preliminary findings were presented as a poster at the 110th General Meeting of the American Society for Microbiology, San Diego, CA, 23 to 27 May 2010 [3].)

MATERIALS AND METHODS

Subjects.

A total of 781 consecutive specimen pairs, consisting of one vaginal specimen and one cervical or urine specimen, were collected from 766 women from July 2009 to August 2010 during office visits with their health care provider. Only patients with specimen pairs or who requested tests for both T. vaginalis by the Affirm assay and N. gonorrhea/C. trachomatis by the Aptima assay were included in the study. The study design was approved by the Lifespan Institutional Review Board (approval number 0065-08).

Specimen collection and testing.

One vaginal specimen for the BD Affirm VPIII assay was collected from women during their office visit and transported by the assay's ambient-temperature transport system. A cervical specimen and/or a urine sample was collected during the same office visit with an Aptima cervical swab or urine collection kit and tested by the Aptima Combo N. gonorrhea/C. trachomatis assay and subsequently by the ATV assay.

Affirm assay.

Vaginal specimens collected and transported with the Affirm ambient-temperature transport system were used for testing by the Affirm assay. Affirm is an FDA-cleared hybridization assay that detects rRNA for the three organisms associated with vaginosis/vaginitis: G. vaginalis (vaginosis), Candida albicans (yeast vaginitis), and T. vaginalis (trichomoniasis vaginitis). Specimens were tested and results were interpreted according to the manufacturer's instructions.

Aptima Combo 2 assay.

The Aptima Combo 2 TMA-based assay is FDA cleared to detect both C. trachomatis and N. gonorrhea in one sample. The Aptima Combo 2 assay was conducted as previously described (4, 20). Results of the Aptima Combo N. gonorrhea/C. trachomatis assay were used in an epidemiological assessment of the population and to determine the prevalence of STIs.

ATV assay.

The ATV assay targets T. vaginalis-specific 16S rRNA. The ATV assay was performed on the remnant Aptima Combo 2 specimens collected during the same office visit as the specimens for N. gonorrhea/C. trachomatis testing. The ATV assay was performed by using the Aptima general-purpose reagent (GPR) and ATV analyte-specific reagent (ASR) oligonucleotides (Gen-Probe), which were reconstituted as described previously (24). Before 1 March 2010, specimens were tested on the semiautomated direct tube sampling (DTS) platform (as described in the Aptima Combo 2 package insert; Gen-Probe Inc., San Diego, CA); after that date, specimens were tested on the fully automated Tigris DTS platform. Positive specimens were retested to determine assay reproducibility. Samples with results of >50,000 relative light units (RLU) were considered positive. Samples with RLU of 10,000 to 50,000 were considered indeterminate and retested in duplicate. If the specimen retested within this range, the final result was called indeterminate. RLU below 10,000 were considered negative, and specimens were not retested.

Confirmation of an initial positive ATV result and discrepant analysis.

All initially positive ATV results were analyzed by an alternate GPR-based TMA assay that targets different T. vaginalis rRNA sequences (Alt-ATV assay). Specimens with a result of >50,000 RLU were considered positive. Performance parameters for the Alt-ATV assay at our institution were previously determined from 266 T. vaginalis-negative specimen pairs evaluated by both the ATV assay and PCR (8). The Alt-ATV assay sensitivity and specificity were 100 and 98.5%, respectively.

Data analysis.

A true-positive result for T. vaginalis infection was defined as a specimen with two positive test results for T. vaginalis by two different molecular assays (Affirm and ATV, Affirm and Alt-ATV, and/or ATV and Alt-ATV assays). Age prevalence percentages were determined by dividing the number of unique infected patients by the total number of unique patients within an age group.

Fisher's exact test was performed to calculate the statistical significance of general demographics and sensitivity/specificity using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA). The null hypothesis was rejected if the P value was <0.05.

RESULTS

Subject population.

The study population consisted of 766 females, with a mean age of 29.2 years (ranging from 13 to 88 years). Thirty-four point nine percent of women were negative for all organisms tested; G. vaginalis was the most prevalent organism found (41.3%), followed by Candida species (14.8%), T. vaginalis (5.1%), C. trachomatis (3.4%), and N. gonorrhea (0.7%) (Fig. 1). While C. trachomatis and N. gonorrhea infections were found only in women under 30 years of age, T. vaginalis was the most prevalent in women who were 36 to 45 (11.9%; 13/109) and 51 to 60 (7.7%; 2/26) years of age, in addition to women under 30 (4.0%; 20/496) (Fig. 2); the median age of T. vaginalis-positive women was 32 years.

Fig. 1.

Infections found in 766 symptomatic females using the Affirm and Aptima assays. BV, bacterial vaginosis (caused by G. vaginalis); CA, C. albicans; TV, T. vaginalis; CT, C. trachomatis; GC, gonococci (N. gonorrhoeae). GV and CA, tested by the Affirm assay; CT, GC, and TV, tested by the Aptima assay.

Fig. 2.

Age-dependent STI prevalences in 766 symptomatic females. TV, T. vaginalis; CT, C. trachomatis; GC, gonococci (N. gonorrhoeae).

ATV and Affirm assay results.

Results and performance parameters for the Affirm and the ATV assays are shown in Table 1. Of the 781 samples analyzed with both assays, 41 were considered T. vaginalis positive and 740 were T. vaginalis negative (according to the infected status definition). The ATV assay detected all 41 T. vaginalis-positive samples and produced no false-positive or false-negative results, yielding a sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 100%. The Affirm assay, however, produced 15 false-negative samples and 1 false-positive sample, yielding a sensitivity of 63.4% (26/41), a specificity of 99.9% (739/740), a PPV of 96.3% (26/27), and an NPV of 98.0% (739/754).

Table 1.

Comparison of molecular methods for detection of T. vaginalis in symptomatic females^a

Diagnostic method	No. of true-positive results	No. of false-positive results	No. of false-negative results	No. of true-negative results	Sensitivity		Specificity		Predictive value (%)
					%	95% CI	%	95% CI	Positive	Negative
Affirm VPIII	26	1	15	739	63.4	55.0–65.4	99.9	99.4–100	96.3	98.0
ATV	41	0	0	740	100	95.4–100	100	99.7–100	100	100

^aSeven hundred eighty-seven specimen pairs were analyzed by the Affirm VPIII and ATV assays.

Specimens submitted for the Aptima assay were 74.5% (590/781) cervical and 24.5% (191/781) urinary, with 19 and 22 ATV-positive results for these specimen types, respectively. There were no discrepant ATV results for patients with more than one Aptima specimen type submitted during an office visit. All specimens that tested positive upon initial ATV screening retested as positive in a second analysis of the same specimen, with RLU clearly demarcated as >50,000. Additionally, no positive samples exhibited an RLU of <100,000. Only 1 specimen exhibited an initial RLU in the indeterminate range (26,000); the specimen was retested twice, with subsequent RLU of <10,000 each time, and determined to be negative.

DISCUSSION

We show herein that the ATV assay was statistically (P < 0.0001) more sensitive than the Affirm assay (100% versus 63.4%), detecting 36% more T. vaginalis-infected women in the population tested. Previous studies compared the ATV assay with nonmolecular assays for T. vaginalis detection, such as wet mount microscopy and culture, and showed that the ATV assay exhibited superior performance (15, 17, 22–24). Other studies have shown that ATV performance is comparable to that of noncommercial PCR assays (12, 24). This study is the first comparing the ATV assay side by side with another commercially available molecular assay (BD Affirm VPIII assay) for T. vaginalis detection. One previously published abstract comparing the Affirm and ATV assay performances in which T. vaginalis was detected in both urine and vaginal specimens (30) confirms the results present in this study. This is important, as both genital and urine specimens are useful specimens for T. vaginalis detection (24).

Not surprisingly and in accordance with published data (13), T. vaginalis was more prevalent than C. trachomatis and N. gonorrhea in the population tested. What was somewhat surprising was the high prevalence of T. vaginalis in women 36 to 45 years of age in addition to in the typical at-risk age group for STIs, 16 to 25 years of age. Interestingly, women who were 36 to 45 years old had no C. trachomatis or N. gonorrhea infections but had significantly more T. vaginalis infections than women who were 16 to 25 years old (P = 0.005, odds ratio [OR] = 3.1, 95% confidence interval [95% CI] = 1.4 to 6.6). While this statistically significant double peak is consistent with symptomatic- and asymptomatic-general-population data from our institution (7a), other studies have also shown the unique prevalence of T. vaginalis in an age group outside the recommended STI screening age groups (11, 16). While women aged 51 to 60 years also exhibited an increased T. vaginalis prevalence, this finding was not statistically significant compared to findings for any other age group. Lack of significance may be due to an inadequate number of women sampled within this age group. No wide-range prevalence studies have been conducted to date with amplification methods to determine the actual prevalence of T. vaginalis by age. Determining T. vaginalis prevalence in different age groups should help to optimize T. vaginalis testing resources in high-prevalence age groups. For example, for women who are 16 to 25 years old, performing T. vaginalis testing in conjunction with C. trachomatis/N. gonorrhea testing may be appropriate, while for women ≥30 years old, performing T. vaginalis testing in conjunction with a cervical cancer screening with liquid cytology may be an optimal approach.

A total of 316 (41.3%) patient specimens were found to be positive for G. vaginalis by the Affirm assay. This seems quite high given the low prevalence of STIs in the population studied; however, all patients were symptomatic with vaginal complaints which may have selected for a more highly infected patient group. In addition, the validity of these results is unknown, since no comparative method was used for the diagnosis of G. vaginalis. Among specimens identified as both G. vaginalis positive and T. vaginalis negative by the Affirm assay were 7 specimens that were in fact true-positive T. vaginalis specimens. These specimens make up 46.7% (7/15) of Affirm false-negative T. vaginalis results. Thus, it is possible that some patients ended up being treated for the wrong etiology and/or with a less-than-optimal regimen. Issues related to the recurrence of symptoms, continued transmission of T. vaginalis, and development of resistance are potential concerns if an incorrect diagnosis of G. vaginalis is made in lieu of a correct diagnosis of T. vaginalis. The lack of sensitivity in clinically diagnosing overlapping vaginosis/vaginitis symptoms supports the need to have a definitive test for T. vaginalis vaginitis (2, 29).

In this population with a low prevalence of STIs, the prevalence of T. vaginalis infection (5.1%) was higher than that for N. gonorrhea and C. trachomatis infections combined (4.1%), supporting routine testing for T. vaginalis. The ATV assay provided highly reliable data for T. vaginalis detection from a sample already routinely collected for C. trachomatis/N. gonorrhea testing. Therefore, the ATV assay could be considered a useful companion assay to C. trachomatis/N. gonorrhea testing in young women (<30 years of age) and/or in conjunction with liquid cytology collected for cervical cancer screening in women who are ≥30 years of age.