Comparison of culture and rapid enzyme immunoassay for the detection of group B streptococcus in high-risk pregnancies.

OBJECTIVE
The purpose of this study was to evaluate the Equate Strep B(R) test for clinical use in patients at high risk for complications from group B streptococcus (GBS) disease.


METHODS
Vaginoperineal swabs were obtained from patients with preterm premature rupture of the membranes and/or preterm labor and semiquantitative GBS cultures and Equate(R) assay were performed.


RESULTS
From May 14, 1990, to April 30, 1992, 650 patients were enrolled; 626 had both culture and Equate(R) results available, of whom 24% were colonized with GBS. The sensitivity, specificity, positive predictive value, and negative predictive value of the rapid assay were 28%, 84%, 35%, and 79%, respectively. Although the prevalence of GBS was higher in patients with ruptured membranes compared with those with intact membranes, rupture of membranes did not affect test sensitivity or specificity.


CONCLUSIONS
We conclude that the Equate(R) rapid assay is not a sensitive method of GBS detection in high-risk patients.

cies. Although maternal disease due to GBS colonization is relatively rare, neonatal colonization occurs in 40-70% of infants born to colonized women. Approximately 1-2% of colonized infants ultimately develop early onset neonatal GBS sepsis, a disease with a high mortality rate and much morbidity. Factors that increase the risk for development of GBS sepsis include low birth weight and prematurity, prolonged membrane rupture, and maternal factors such as fever in labor and heavy vaginal colonization.
It has been shown that intrapartum administration of ampicillin to mothers colonized with GBS in the 3 rd trimester results in a decrease in neonatal colonization and neonatal GBS sepsis. ,2 The problem remains, however, in identifying those patients who would most benefit from ampicillin prophylaxis. Antepartum cultures ave not predictive of intrapartum colonization status, as only 55-67 % of patients with positive cultures in the 3rd trimester will have positive cultures at the time of labor, even in the absence of antibiotic treatment. 3'4 Moreover, approximately 5-10% of patients with negative antepartum cultures will be colonized when Our purpose was to evaluate one of the currently available rapid test kits for clinical use in a population at high risk for maternal GBS colonization and the subsequent development of neonatal GBS colonization and sepsis.

SUBJECTS AND METHODS
As part of an ongoing study of the use of a rapid GBS test in the management of high-risk pregnancies, all women admitted to the Medical College of Virginia Hospital with the diagnosis of preterm (<37 complete weeks gestation) labor and/or preterm premature rupture of the membranes (PTROM) were offered enrollment in the study. In addition, those patients who were undergoing a preterm induction of labor for maternal indications (usually severe preeclampsia) were eligible. Exclusion criteria included age < 18 years, antibiotic use in the previous 2 weeks, or a previous infant with GBS sepsis. The study was approved by the Committee on the Conduct of Human Research at Virginia Commonwealth University, and an informed consent was obtained from all patients.
Three vaginoperineal swabs were obtained from all study patients at the time of admission to labor and delivery. Two swabs were transported in culture media to the clinical microbiology laboratory, where was streaked onto non-selective medium (5% sheep blood agar) and the other was placed in selective enrichment broth (Todd-I--Iewitt broth supplemented with 5 % sheep blood, 0.01 mg/ml of colistin, and 0.005 mg/ml of oxolinic acid). Following 24 h of incubation, the broth was plated, again on non-selective medium. Colonies of GBS were confirmed using Phadebact(R) Strep B latex agglutination test (Remel, Lenexa, KS), in accordance with the manufacturer's recommended procedures.
The 3rd swab, provided in the Equate(R) kit, was used to perform the Equate(R) assay, as per the manufacturer's directions. Following an initial 11month trial period during which 37 assays were performed by 3 rd-year house officers, all Equate(R) assays were performed by of 4 designated laboratory technicians.
Results of the culture for GBS were reported semiquantitatively as containing many, moderate, or rare colonies or as negative based on the number of quadrants on the plate that contained growth. Those samples that resulted in growth only following broth incubation were classified as containing only rare colonies. Equate(R) assays were reported as positive or negative.
Chi-square analysis was used to compare discrete variables. P < 0.05 was considered significant. 14, 1990, andApril 30, 1992, 650 patients were enrolled in the study. Of these, 626 had both Equate(R) and vaginal culture results available. The mean age of the study participants was 24.6 5.7 years and the mean gestational age at enrollment was 32.2 + 3.3 weeks. Thirty-five percent of the patients were nulliparous, while 21% had a history of preterm birth. Three hundred twenty-eight (52%) of the patients were admitted with preterm labor, 231 (37%) with PTROM, and 67 (11%) were enrolled at the time of preterm induction. All patients in the latter group had intact membranes at enrollment.

Between May
One hundred forty-nine (24%) of the patients were culture positive for GBS. Of these, 77 (52%) had rare, 54 (36%) had moderate, and 18 (12%) had many colonies.  When patients with intact membranes were analyzed separately from those with PTROM, the results were similar. Although the sensitivity and specificity were essentially the same, the positive predictive value (PPV) was higher and the negative predictive value (NPV) was lower in those patients with PTROM. This finding is not surprising, given the significantly higher prevalence of GBS colonization in the patients with PTROM (32% vs. 19%, P 0.004). Because prior studies of rapid GBS detection systems have shown improved sensitivity when applied only to those patients with heavy vaginal colonization, 5we then analyzed our results considering only those patients with heavy colonization, defining moderate or many colonies (>2 quadrants on the initial plate or > 100,000 cfu/ml) as positive (N 72). As shown in Table 1, the sensitivity and NPV of the rapid test also improved in our study, although marginally, when only heavy maternal colonization was considered. The false negative rate, or the percentage of carriers not identified, increases with decreasing colony counts. Thirteen percent of those with rare colonies, 8% of those with moderate colonies, and 3 % of women with many colonies were not detected by the rapid test.

DISCUSSION
Although the first studies to utilize rapid detection tests reported high sensitivities for the detection of GBS (58% by Isada and  Using yet another enzyme immunoassay (Quidel(R)), Granato and Petosa 9 reported a sensitivity of 89%, a specificity of 99%, and an NPV of 99% in comparison with culture on non-selective medium. However, when lightly colonized women are included, the sensitivity decreases to 74%. It remains to be seen whether this rapid test will be adequately sensitive when compared with culture using selective medium.
Although many investigators have found increased sensitivity when the rapid tests were applied only to heavily colonized women, whether detection of only heavily colonized mothers is adequate is a matter of some discussion. 8, 0, 6 Isada and Grossman 2 found that, in a population with a GBS prevalence of 4.4%, all 9 cases of "clinically significant" perinatal GBS morbidity were identified using a rapid latex agglutination test, including 5 cases of neonatal GBS sepsis. Although heavy maternal colonization increases the risk for neonatal GBS sepsis, many cases of neonatal GBS sepsis have been reported following delivery of women who were only lightly colonized. 2'8'5-7 A prospective study by Tuppurainen and Hallman 8 found that using a latex agglutination test (Streptolatex, Orion Diagnostica, Finland) for selecting those intrapartum patients to be randomized to receive prophylactic penicillin or placebo resulted in neonatal GBS disease in 0.07% of the rapid-test negative (untreated) group. Nine percent of the neonates born to rapid-test positive patients who received placebo developed GBS disease, while 1.1% of the neonates born to treated women had GBS disease. In a similar study, Morales and Lim 7 performed rapid tests on 260 patients with PTROM. Eighty-four patients had positive rapid tests, 36 of whom were given antibiotic prophylaxis prior to delivery. Although patients with positive rapid tests who were untreated were more likely to develop chorioamnionitis (13% vs. 0) and to deliver neonates who developed neonatal sepsis (15% vs. 0), no results are reported for those women who were rapid-test negative. As a result, it is impossible to evaluate the effectiveness of this treatment protocol. Unless chemoprophylaxis protocols based on the use of rapid assays for the intrapartum detection of maternal GBS colonization can be shown to result in similar or lower neonatal GBS colonization and sepsis rates compared with current perhaps less selective management protocols, 9 the use of the Equate(R) assay in the intrapartum management of high-risk pregnancies is to be discouraged.