Am Fam Physician. 2008;77(5):686-691
Background: Although infection with Trichomonas vaginalis is nearly as common as chlamydia and more common than gonorrhea among adolescent women, traditional testing measures are not accurate. Wet mount slide preparation is the most common testing method and can be performed in the office, but its sensitivity may be as low as 36 to 70 percent. Culture for T. vaginalis is better, but its sensitivity may still be only 70 to 85 percent, and it can take days to receive results. Two newer options include nucleic acid amplification and a point-of-care rapid antigen detection test. Huppert and colleagues assessed the sensitivity and specificity of the four tests and whether accuracy was affected by vaginal symptoms, bleeding, or other sexually transmitted infections (STIs).
The Study: The study included women 14 to 21 years of age who presented to a local health center and emergency department. Eligible women included those who presented with genitourinary symptoms such as pelvic pain, dysuria, abnormal vaginal bleeding, or vaginal discharge or itching. High-risk asymptomatic women (those who had unprotected vaginal sex, two or more partners in the previous three months, or contact with a person with an STI within the previous six months) were also included. Patients were excluded if they had used metronidazole (Flagyl) within the previous two weeks.
Pelvic examinations and gonorrhea and chlamydia testing were performed on all participants. Four vaginal swabs were collected and then randomly assigned to a testing method: wet mount, culture, rapid antigen testing, or nucleic acid (transcription-mediated) amplification. Sensitivities of the respective tests were initially calculated by the traditional standard (positive culture or wet mount, which initially assumed 100 percent sensitivity and specificity for these tests) and a composite reference standard (considered positive if any test detected T. vaginalis, and negative if no test detected it). The data from these two methods were also used for latent class analysis, which estimated overall sensitivity and specificity for each test.
Results: Of the 330 women who had complete laboratory data for the study, 63 percent reported at least one vaginal symptom. A total of 269 patients (81.5 percent) tested negative for T. vaginalis infection with all four tests, and 28 patients (8.5 percent) tested positive on all four tests. Sixty-one patients (18.5 percent) tested positive on at least one test; wet mount was positive in 31 patients (9.4 percent). Slightly more than one half of the patients with T. vaginalis infection also had chlamydia or gonorrhea.
When latent class analysis was performed, the sensitivity of wet mount testing for detecting T. vaginalis was found to be significantly lower than for the other tests (see accompanying table). The sensitivity of nucleic acid amplification was significantly greater than traditional methods and was equivalent to rapid antigen testing. In asymptomatic women, nucleic acid amplification was more sensitive than the other tests. However, the testing technique was less effective if the patient had vaginal bleeding or a concurrent STI.
|Test||Sensitivity (%)||Specificity (%)|
|Rapid antigen test||90.3||100|
|Nucleic acid (transcription-mediated) amplification||98.2||98.0|
Conclusion: Like wet mount preparations, the rapid antigen test is a convenient method that can be performed in the office or at the bedside. However, rapid antigen testing is significantly more sensitive than wet mount for detecting T. vaginalis in symptomatic women, and it performed as well as the more time-consuming culture and nucleic acid amplification methods. Nucleic acid amplification was more sensitive for diagnosing T. vaginalis infections in asymptomatic women.