OBJECTIVE: The administration of antibiotic prophylaxis to laboring women who harbor Group B streptococci (GBS) depends on identification of carriers. We sought to evaluate the diagnostic accuracy of real-time polymerase chain reaction (PCR) for detection of GBS using a more stringent culture method.
METHODS: Two swabs were used simultaneously to obtain rectovaginal GBS samples from consenting women. One swab was analyzed using a stringent, validated culture technology, which included direct plating onto selective agar and inoculation of a selective broth. The other swab was used for a commercial real-time PCR assay, which uses amplification to detect the presence of the cfb gene sequence of GBS DNA. We calculated the assay accuracy using sensitivity and specificity.
RESULTS: A total of 233 samples were available. Both the culture and PCR methods were positive for 59 and negative for 157 patients. The culture method was positive and PCR was negative in 9 patients. The culture was negative and the PCR positive for 8 patients. The sensitivity of the PCR assay was 86.8% and specificity was 95.2%. The positive predictive value was 88.1% and the negative predictive value was 94.6%.
CONCLUSION: Although a rapid PCR assay may be useful to determine GBS status in the urgent intrapartum setting, the false-negative rate of 13.2% for the real-time PCR assay prohibits its use for standard GBS screening in the office.
LEVEL OF EVIDENCE: II-1
Because the polymerase chain reaction&#x2013;based Group B Streptococcus assay is less sensitive than previously reported, it may be used in acute settings, but not for standard care.
From the Departments of 1Obstetrics & Gynecology and 2Pathology & Immunology, Washington University in St. Louis, Saint Louis, Missouri.
Corresponding author: Kristin Atkins, Washington University School of Medicine, Obstetrics and Gynecology, 4566 Scott Avenue, Campus Box 8064, Saint Louis, MO 63110; e-mail: firstname.lastname@example.org.
The authors thank Infectio Diagnostic (I.D.I.) Inc. for providing the kits used for the polymerase chain reaction.
This work was presented at the 26th Annual Meeting of the Society for Maternal–Fetal Medicine, Miami, Florida, February 1–4, 2006.
Group B streptococci (GBS, Streptococcus agalactiae) was recognized in the 1970s as a leading cause of neonatal infection and death.1–5 Intrapartum chemoprophylaxis has greatly reduced the risk of neonatal infection in women who carry this bacterium. Despite these measures, GBS remains a leading cause of infectious neonatal morbidity in the United States. Before universal screening, the incidence of early onset GBS disease was 0.47 cases per 1,000 live births.6 In 2002, the Center for Disease Control and Prevention (CDC) published guidelines for the prevention of perinatal GBS disease, and recommended routine culture for all pregnant women between 35 and 37 weeks of gestation.7 These recommendations were prompted by the recognition that a culture-based screening strategy is greater than 50% more effective in preventing neonatal infection than the previously recommended risk-based screening.8 Implementation of this screening protocol has decreased the incidence of early-onset GBS disease to 0.32 cases per 1,000 live births in 2003.6
The CDC currently recommends screening cultures using selective broth medium for the recovery of GBS from rectovaginal specimens. This technique requires inoculation of broth with the specimen and incubation for 24 hours before plating. A study comparing the selective broth medium culture method with direct plating using selective agar medium alone determined that a combination of these methods resulted in a 15% improved detection rate for GBS.9 Therefore, the combined method is currently the standard at our institution and is available for use in any laboratory.
Rapid detection methods for GBS may be advantageous for women who present in labor without any information on their GBS colonization status. Bergeron et al10 reported on the efficacy of a rapid polymerase chain reaction (PCR) assay for the detection of GBS, prompting the development of the PCR assay for commercial use. Infectio Diagnostic (I.D.I.) Inc. (Quebec, Canada) has recently received approval from the Food and Drug Administration for either antepartum or intrapartum use of a PCR assay for rapid detection of the cfb gene sequence of the GBS genome.
We performed a prospective evaluation of diagnostic accuracy of the IDI-Strep B real-time PCR assay using combined culture method as standard. Given the improved detection rate of the combined culture method, we speculated that the sensitivity of the PCR assay would be significantly less than previously reported.
MATERIALS AND METHODS
The Human Studies Committee of Washington University School of Medicine approved this study. Between November 1, 2004, and June 30, 2005, all women for whom a GBS culture was ordered at Barnes Jewish Hospital labor and delivery, the antepartum clinics, or private offices were eligible for the study. After informed consent was obtained, information was collected about participant’s demographics as well as complications of pregnancy and past medical history. A previous positive culture for GBS was also recorded.
The vaginal and rectal specimens were collected using the technique recommended by the CDC.7 The lower one-third of the vagina, perineal body, and rectum beyond the anal sphincter, were sampled with two swabs simultaneously. Each swab was placed in a separate, labeled tube for transport to the laboratory where one was used for culture, and the other for the IDI-Strep B assay. Each collected specimen was labeled with a unique study identification number.
All culture swabs were processed as described by Dunne and Holland-Staley9 with modification. This technique uses a combination of selective broth and agar media. For this study, LIM broth was used in place of selective broth medium. All cultures were performed and evaluated by trained laboratory personnel. The swabs were first inoculated onto blood agar plates containing 30 mg/L neomycin and 15 mg/L nalidixic acid (NNA, Becton Dickinson, Sparks, MD) and then submersed in LIM Broth (BBL, Becton Dickinson) which contained 10 mg/L colistin and 15 mg/L nalidixic acid.11 Both cultures were incubated overnight at 35°C, with 5% CO2 included for the plated specimens. The NNA plates were then examined for characteristic GBS colonies; if such colonies were present, they were identified by Gram stain, catalase reaction, and either the CAMP test or a particle agglutination assay (Streptococcal Grouping Kit, OXOID, Basingstoke, Hants, United Kingdom), depending on the number of isolated colonies. Negative cultures were incubated for an additional 18–24 hours. Any suspicious colonies were handled as above, and cultures that remained negative were finalized. The LIM Broth cultures were then subcultured onto blood and colistin-nalidixic acid (CNA) agar plates (Becton Dickinson) and incubated for 18–24 hours at 35°C with 5% CO2 before examination. As with the NNA plates, all LIM Broth subcultures were examined for colonies typical of GBS, and the colonies were identified by the methods described above. Negative subcultures were reincubated and were examined again the following day. The results were reported and used for the management of laboring women.
The investigator (R.M.A.) who performed the PCR assay was blinded to the results of the culture. Swabs submitted for the IDI-StrepB assay were processed according to manufacturer’s instructions for amplification of the cfb gene of the GBS. The PCR results were not reported to the physicians who provided care to participants.
Results of the combined culture and PCR methods were compared for each specimen. For discrepant results, PCR was repeated twice to confirm the initial findings. Sensitivity, specificity, and positive and negative predictive values along with their confidence intervals were calculated for the PCR assay. The prevalence of GBS carriers during pregnancy at Barnes Jewish Hospital is 28.5%. Using this prevalence, 225 samples were required for the study assuming that the actual sensitivity of the PCR was at least 90% with an acceptable 95% confidence interval of ± 7.5%.
One thousand two hundred forty-three GBS samples were processed during this period. A total of 433 women who gave consent for the study. The PCR assay must be run within 24 hours of collection. Two hundred thirty-four of the samples were able to be processed within 24 hours. One patient was excluded because only one swab was obtained, leaving 233 paired specimens for comparison.
The demographic information is presented in Table 1. The range of gestational age for these patients reflects the culture of three patients who were outside of the normal recommended gestational age for sampling. The husband of one patient had previously had a positive culture for GBS on infertility evaluation. The other two patients had experienced previable rupture of membranes in a previous pregnancy and were being assessed for the presence of multiple bacteria.
The distribution of participants’ ethnic background reflects that of women receiving obstetric care at our facility. Fourteen of the multiparous patients were aware that they were GBS positive in their prior pregnancy. There were no infants born to these mothers who developed GBS infection. All other patients were either negative or unaware of their prior status.
Both the culture and PCR were positive for 59 and negative for 157 patients. Culture was positive and PCR was negative for 9 patients. Culture was negative and PCR positive for 8 patients. Sensitivity of the PCR assay was 86.8% (95% confidence interval [CI] 76.4–93.8%) and specificity was 95.2% (95% CI 90.7–97.9%). The positive predictive value was 88.1% (95% CI 77.8–94.7%) and the negative predictive value was 94.6% (95% CI 90.0–97.5%). The PCR was correct for 92.7% (95% CI 88.6–95.7) overall.
We noted that eight specimens were culture negative and PCR positive. The PCR was repeated two additional times confirming the positive result of these eight samples. For five of these eight specimens, the specimen elution buffer was cultured. Three of these had a positive culture from the elution buffer and two were negative. For the nine specimens that were culture positive and PCR negative, the PCR specimens were repeated two more times to confirm the result. Two of these specimens were reported as having the rare presence of GBS colonies by culture.
A rapid, sensitive and accurate method would provide the best means of GBS detection, allowing the highest percentage of women to be treated, and reducing the rate of GBS transmission to the newborn. This method would be highly beneficial in the case of preterm labor, preterm premature rupture of membranes, and in patients with limited or no prenatal care. It would also be helpful when a previous culture result is unavailable. Using a combined culture approach as the standard, we found that a PCR-based methodology had comparable sensitivity, specificity, and predictive values. Notwithstanding, these values fell short of the double-culture method, previously shown to be superior to the selective agar medium or selective broth medium alone.9 The advantage of the combined culture approach is due to the competitive overgrowth of GBS by enterococci in the broth medium.
Additional assays have been developed in an attempt to provide a rapid, sensitive assay for GBS detection. Among these, optical immunoassay has shown limited clinical usefulness with a sensitivity of 14% with light GBS colonization and 41.5% with heavy colonization.12 Probe hybridization methods, most popularly the AccuProbe Group B Streptococcus Test from Gen-Probe (San Diego, CA), demonstrate sensitivity, specificity, positive predictive vale and negative predictive value of 94.7–100%, 99.5–100%, 97.8–100%, and 98.8–100%, respectively.13–15 Although these results may be obtained more rapidly than a culture, the technique requires 18 hour incubation before analysis and renders this assay suboptimal in an acute setting.
Two studies have previously compared PCR detection of the GBS cfb gene to broth culture. Bergeron et al10 reported 100% sensitivity and 100% specificity of PCR method, while Davies et al16 reported PCR sensitivity of 94%, with a specificity of 95.9%. These results are different from ours likely because our standard was superior to that used in previous reports.8
It is notable that we cannot calculate the true sensitivity of the LIM broth alone or perform discrepant analysis in our study because according to our protocol if the NNA was positive for GBS, no further work was performed with the LIM broth. Given the previously published sensitivity of the selective broth medium of 86%, our results demonstrate that the sensitivity of the PCR assay is comparable (86.8%) and suggests that the combined culture method provides improved sensitivity.9
Eight women in our study exhibited positive GBS by PCR and not by culture. For three of these cases, GBS was subsequently cultured from the cell suspension used to prepare DNA for amplification. These discrepancies could have resulted from the presence of nonviable organisms in the specimen or unequal sample collection between the two swabs obtained for the study. In the case where the bacteria are nonviable, bacterial DNA would be available for amplification in the absence of a positive culture. It should be noted that false-positive PCR results may not result in any harm to the patient, because the only implication might be an unnecessary administration of antimicrobial agents. If management were based solely on the PCR result, eight patients in our cohort who would have received antimicrobial prophylaxis unnecessarily.
If the PCR technique alone were used, the specimens that were culture positive and PCR negative would result in no treatment for women who need intrapartum antibiotic prophylaxis. Two of these discrepant results can be explained by the presence of minimal growth on the culture. In these cases, the amplification may not have met the threshold set by the manufacturer to determine a positive PCR result. Given the ramifications of early neonatal GBS infection, we believe that a false-negative rate of 13.2% (9/58) is unacceptable.
The results of our study using a stringent reference method indicate that although the rapid GBS assessment using the IDI-StrepB assay is a viable option when GBS status assessment is urgently needed, the unacceptable false negative rate renders this technology substandard in the office setting. Therefore, in the office setting, a culture should be obtained.
1. Baker CJ, Barrett FF. Transmission of group B streptococci among parturient women and their neonates. J Pediatr 1973;83:919–25.
2. Baker CJ, Barrett FF, Gordon RC, Yow MD. Suppurative meningitis due to streptococci of Lancefield group B: a study of 33 infants. J Pediatr 1973;82:724–9.
3. Barton LL, Feigin RD, Lins R. Group B beta hemolytic streptococcal meningitis in infants. J Pediatr 1973;82:719–23.
4. Franciosi RA, Knostman JD, Zimmerman RA. Group B streptococcal neonatal and infant infections. J Pediatr 1973;82:707–18.
5. McCracken GH Jr. Group B streptococci: the new challenge in neonatal infections. J Pediatr 1973;82:703–6.
6. Early-onset and late-onset neonatal group B streptococcal disease—United States, 1996–2004. MMWR Morb Mortal Wkly Rep 2005;54:1205–8.
7. Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of perinatal group B streptococcal disease. Revised guidelines from CDC. MMWR Recomm Rep 51(RR-11):1–22, 2002.
8. Schrag SJ, Zell ER, Lynfield R, Roome A, Arnold KE, Craig AS, et al. A population-based comparison of strategies to prevent early-onset group B streptococcal disease in neonates. N Engl J Med 2002;347:233–9.
9. Dunne WM Jr, Holland-Staley CA. Comparison of NNA agar culture and selective broth culture for detection of group B streptococcal colonization in women. Journal of Clinical Microbiology 1998;36:2298–300.
10. Bergeron MG, Ke D, Menard C, Picard FJ, Gagnon M, Bernier M, et al. Rapid detection of group B streptococci in pregnant women at delivery. N Engl J Med 2000;343:175–9.
11. Jones DE, Friedl EM, Kanarek KS, Williams JK, Lim DV. Rapid identification of pregnant women heavily colonized with group B streptococci. J Clin Microbiol 1983;18:558–60.
12. Thinkhamrop J, Limpongsanurak S, Festin MR, Daly S, Schuchat A, Lumbiganon P, et al. Infections in international pregnancy study: performance of the optical immunoassay test for detection of group B streptococcus. J Clin Microbiol 2003;41:5288–90.
13. Bourbeau PP, Heiter BJ, Figdore M. Use of Gen-Probe AccuProbe Group B streptococcus test to detect group B streptococci in broth cultures of vaginal-anorectal specimens from pregnant women: comparison with traditional culture method. J Clin Microbiol 1997;35:144–7.
14. Daly JA, Clifton NL, Seskin KC, Gooch WM. Use of rapid, nonradioactive DNA probes in culture confirmation tests to detect Streptococcus agalactiae, Haemophilus influenzae, and Enterococcus spp. from pediatric patients with significant infections. J Clin Microbiol 1991;29:80–2.
15. Williams-Bouyer N, Reisner BS, Woods GL. Comparison of gen-probe AccuProbe group B streptococcus culture identification test with conventional culture for the detection of group B streptococci in broth cultures of vaginal-anorectal specimens from pregnant women. Diagn Microbiol Infect Dis 2000;36:159–62.
16. Davies HD, Miller MA, Faro S, Gregson D, Kehl SC, Jordan JA. Multicenter study of a rapid molecular-based assay for the diagnosis of group B Streptococcus colonization in pregnant women. Clin Infect Dis 2004;39:1129–35.