Study objective: Adolescent females are disproportionately affected by Sexually Transmitted Infections (STIs). Endocervical Gram stain smears taken during gynecological examination are inexpensive, relatively easy procedures to perform and interpret. The purpose of this study was to evaluate the performance characteristics of Gram smears in the diagnosis of Chlamydia trachomatis (CT), Neisseria gonorrhoeae (GC) and Trichomonas vaginalis (TV) in a female adolescent population presenting to the Emergency Department (ED) with genitourinary complaints. Methods: This study was a retrospective, cohort analysis of consecutive females (ages 13 – 19) seen at three academic medical centers over a two-year study period. All patients underwent a pelvic exam with diagnostic testing for STIs. Positive criteria for a Gram stain included greater than ten white blood cells per high-power field, gram-negative intracellular/extracellular diplococci (suggesting GC), or direct visualization of TV organisms. Polymerase Chain Reaction (PCR) assays were used as the gold standard definition for CT/GC infection. Direct microscopic visualization of organisms on a separate wet mount prep was considered the gold standard for TV infection. Demographic information, clinical findings of cervicitis, and the results of diagnostic testing were obtained from ED records using standardized abstraction forms. Results: During the study period, 1303 adolescent females were evaluated for genitourinary complaints. A total of 181 adolescents (12.9%) had at least one documented STI. Overall, 298 patients (22.9%) had positive gram stains. The sensitivity, specificity, and positive likelihood ratio for Gram stain in the diagnosis of STI were 28.7% (95% CI, 22.2 to 35.9), 78.1% (95% CI, 75.5 to 80.5), and 1.31 (95% CI 1.0 to 1.7), respectively. The sensitivity of Gram stain to Chlamydia trachomatis, Neisseria gonorrhoeae, and Trichomonas vaginalis were 35.9% (95% CI 26.8 to 45.8), 34.5% (95% CI 18.0 to 54.3), and 5.7% (95% CI 1.3 to 15.7), respectively. The consistency of the data recording was excellent, with a median kappa statistic of 0.89. Conclusions: The positive likelihood ratio of inflammation on endocervical Gram stain is too low to recommend its use to direct empiric treatment of adolescents at risk for sexually transmitted infections. Diagnostic uncertainty or treatment failures should prompt specific laboratory testing.
Read full abstract