Question
Can clinical predictors be used to identify women at high risk for trichomoniasis who were negative for T. vaginalis by wet mount preparation but may benefit from additional testing by T. vaginalis culture?  

Design
The symptoms and signs of trichomoniasis in wet mount positive women were compared to those in wet mount negative/culture positive women to develop a predictive model for wet mount negative women who may benefit from additional specimen collection for T. vaginalis culture.   

Participants
Women (n = 2194) presenting for STD evaluation at two urban health departments located in Durham County and Wake County, North Carolina, were tested.  Women were excluded if they were younger than 18 years, used metronidazole in the previous 4 weeks, did not speak English, or were coming for a return visit.  Seventy-nine percent were African-American and 68.5% were between 18 and 29 years old.   

Description of Tests and Diagnostic Standard
Swabs of the posterior fornix vaginal secretions were obtained from each woman during a speculum examination.  A swab for wet mount microscopy was placed in normal saline at the time of collection.  In the laboratory the tube was agitated and 1 drop was placed on a glass slide, which was examined within 10 min of collection at 200X magnification.  A positive result was defined as the presence of motile trichomonads.  For T. vaginalis culture, a swab was immediately inoculated into an InPouch TV culture envelope (BioMed Diagnostics, San Jose, CA) and incubated according to standard procedure.  Positive cultures were defined as the presence of motile trichomonads.  A multivariate logistic regression model was used to assess associations among demographic, clinical, laboratory, and physical examination data using culture as the dichotomous outcome.      

Main Outcome Measures
Predictors for positive T. vaginalis cultures among wet mount negative women were identified.  

Main Results
The prevalence of T. vaginalis among the 2194 women using a combined standard of wet mount and culture results was 17.5%.  The results of wet mount preparation and T. vaginalis culture are shown in Table 1.  Adjusted analysis among wet mount negative/culture positive women in a full multivariate logistic regression model identified African-American race (odds ratio (OR) = 9.1, 95% confidence interval (CI) = 2.8, 29.5), contact to trichomoniasis (OR = 7.1, 95% CI = 1.3, 37.7), and any drug use (OR = 2.3, 95% CI = 1.4, 4.1) as significantly associated with trichomoniasis.  The reduced model performance of several combinations of these 3 factors for targeted culture screening among wet mount negative women is shown in Table 2.  The presence of any one of the factors identified 90 (91%) of the 99 wet mount negative women who were culture positive.  

Table 1.  Results of wet mount preparation and culture for T. vaginalis among 2194 women

Table 2.  Reduced model performance using combinations of 3 factors for targeted culture screening of wet mount negative women

Authors’ Conclusions
Three factors were predictive of culture positive trichomoniasis in wet mount preparation negative women.  We propose a 2-step procedure for detecting trichomoniasis based on these predictors.  However, these factors may only be generalizable to urban STD clinics or populations with high rates of trichomoniasis.  In addition, theoretical application of the screening criteria using cost-benefit analysis should be considered before implementation of this predictive model.  The number of cultures to be performed using the model may be cost-prohibitive in settings which serve predominantly African-American women or women with high rates of drug use.    

Source of funding: None given.

For correspondence:  Heidi Swygard, University of North Carolina at Chapel Hill, CB#7030, 130 Mason Farm Road, Chapel Hill, NC 27599-7030.  E-mail address:  heidi.swygard@med.unc.edu.