This study aimed to compare the sensitivity of different culture methods from three different anatomic sites and to evaluate the sensitivity of polymerase chain reaction (PCR) assay targeting the 16 S ribosomal RNA gene in detection of group B streptococcus (GBS) colonization in pregnant women. From 100 pregnant women at 35–37 weeks of gestation, three cotton swabs were used to obtain vaginal, rectal, and rectovaginal (RV) specimens and plated onto Columbia agar with colistin and nalidixic (CNA), group B streptococcus differential agar (GBSDA), and chromID Strepto B agar (CA), with and without Lim broth enrichment. PCR assay was done on the RV swabs. The overall GBS colonization rate was 29 % by culture and 31 % by PCR. GBS positivity for RV sampling (100 %) was significantly higher than detection on the basis of vaginal sampling (50 %), but not significantly higher than for rectal sampling (82 %). Direct plating of the rectovaginal swab on CNA, GBSDA, and CA resulted in detection of 74, 58, and 100 % of the carriers, respectively, whereas subculturing of Lim broth yielded 65, 59, and 83 % positivity, respectively. Using GBS culture as the “gold standard,” the sensitivity of PCR was 100 %, and specificity was 97 %. We found that the inoculation of RV secretions directly onto CA is the most rapid, easy, and sensitive method than that of Lim broth enrichment. Also, we found that group B streptococci can be detected rapidly and reliably by a PCR assay of rectovaginal secretions from pregnant women.
