Screening pregnant women for chlamydial and gonococcal infections allows treatment of organisms that could potentially be transmitted from mother to fetus and reduces the risk of postpartum endometritis. Enzyme immunoassay and DNA probe techniques have essentially replaced cervical cultures and have been shown in some studies to be more sensitive in identifying these organisms. Preliminary evidence suggests that the presence of low numbers of polymorphonuclear lymphocytes in cervical secretions is associated with the absence of chlamydial and gonococcal colonization of the cervix compared with a culture result. Bohmer and associates evaluated the utility and potential cost savings of wet mount preparations in gravid women to rule out the presence of gonococci and Chlamydia trachomatis.

The study consisted of 341 prenatal patients who had been screened for gonococci and C. trachomatis at the initial visit or at 36 weeks of gestation. A total of 106 patients had a history of a sexually transmitted disease (STD). A cervical specimen was obtained with the Gen-Probe PACE 2 DNA probe. Each specimen was also examined microscopically as a wet mount. Polymorphonuclear lymphocytes were counted per high-power field, and an average number was determined.

The following percentages of patients were identified with infections: C. trachomatis in 8.8 percent, gonorrhea in 2.9 percent, Trichomonas vaginalis in 4.5 percent and bacterial vaginosis in 40 percent. In evaluating the wet mounts, 32 percent were found to have fewer than 10 polymorphonuclear lymphocytes per high-power field, of which four were found (by DNA probe) to be infected with C. trachomatis. This results in a negative predictive value of 96 percent. One patient was found to be infected with gonorrhea, resulting in a negative predictive value of 99 percent.

The low prevalence of gonorrhea in this population makes it difficult to subdivide the patients on the basis of risk factors, but three variables were associated with an increased risk of C. trachomatis infection: age younger than 20 years, black race and a history of an STD. In patients with any one of these risk factors, the negative predictive value of fewer than 10 polymorphonuclear lymphocytes on a wet mount was 95.4 percent. None of these risk factors was present in 23.2 percent of patients.

Results of this study demonstrate that the cervical wet mount is an accurate predictor of the absence of gonococcal and C. trachomatis colonization of the cervix in prenatal patients. Accuracy was enhanced in patients without the three risk factors for cervical infection. The negative predictive values for gonorrhea and C. trachomatis were 99 and 100 percent, respectively.

Results of this study may be particularly useful in managing patients with minimal risk factors for STDs. The current DNA probes have a sensitivity and specificity for gonococci and C. trachomatis of 90 and 95 percent, respectively. If the prevalence of these infections is 4 percent in a particular population, the positive predictive value of the DNA probe is 48 percent. This low positive predictive value would result in 50 percent of patients being notified that they have chlamydial or gonorrheal infections when, in fact, they do not.

The authors conclude that use of the wet mount as an initial screening method could eliminate the need for DNA probe testing for gonococci and C. trachomatis and limit the use of DNA probe testing to pregnant women at higher risk for these infections. This strategy would improve the positive predictive value of the DNA probe. The reduction in initial DNA probe testing would decrease costs by 33 percent.

editor's note: In an accompanying discussion, Soper commented on the substantial cost savings demonstrated by the authors in avoiding the initial DNA probe testing in pregnant women with fewer than 10 white blood cells per high-power field on wet mount. A high diagnostic sensitivity and a high negative predictive value are the most important characteristics of a screening test. In this study, most women with gonococcal and chlamydial infections had more than 10 white blood cells per high-power field (sensitivity 90 percent and 87 percent, respectively). The negative predictive value was excellent. False-positive results were reduced with DNA probe testing.—b.a.