SEARCH

SEARCH BY CITATION

Keywords:

  • Group B streptococcus;
  • intrapartum prophylactic antibiotics;
  • screening

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

Please cite this paper as: Daniels J, Gray J, Pattison H, Gray R, Hills R, Khan K on behalf of the GBS Collaborative Group. Intrapartum tests for group B streptococcus: accuracy and acceptability of screening. BJOG 2011;118:257–265.

Objective  To assess the accuracy and acceptability of polymerase chain reaction (PCR) and optical immunoassay (OIA) tests for the detection of maternal group B streptococcus (GBS) colonisation during labour, comparing their performance with the current UK policy of risk factor-based screening.

Design  Diagnostic test accuracy study.

Setting and population  Fourteen hundred women in labour at two large UK maternity units provided vaginal and rectal swabs for testing.

Methods  The PCR and OIA index tests were compared with the reference standard of selective enriched culture, assessed blind to index tests. Factors influencing neonatal GBS colonisation were assessed using multiple logistic regression, adjusting for antibiotic use. The acceptability of testing to participants was evaluated through a structured questionnaire administered after delivery.

Main outcome measures  The sensitivity and specificity of PCR, OIA and risk factor-based screening.

Results  Maternal GBS colonisation was 21% (19–24%) by combined vaginal and rectal swab enriched culture. PCR test of either vaginal or rectal swabs was more sensitive (84% [79–88%] versus 72% [65–77%]) and specific (87% [85–89%] versus 57% [53–60%]) than OIA (P < 0.001), and far more sensitive (84 versus 30% [25–35%]) and specific (87 versus 80% [77–82%]) than risk factor-based screening (P < 0.001). Maternal antibiotics (odds ratio, 0.22 [0.07–0.62]; P = 0.004) and a positive PCR test (odds ratio, 29.4 [15.8–54.8]; P < 0.001) were strongly related to neonatal GBS colonisation, whereas risk factors were not (odds ratio, 1.44 [0.80–2.62]; P = 0.2).

Conclusion  Intrapartum PCR screening is a more accurate predictor of maternal and neonatal GBS colonisation than is OIA or risk factor-based screening, and is acceptable to women.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

Early-onset group B streptococcus (EOGBS) disease, with an incidence of 1 per 2000 live births, is the leading cause of serious neonatal sepsis.1 GBS is an opportunistic pathogen, harboured in the vagina or rectum of an estimated 14% of women (range, 5–21%),2 and is transmitted from colonised mothers to neonates during birth. Intrapartum antibiotic prophylaxis (IAP) reduces the risk of EOGBS infection in the newborn,3 and so screening for maternal GBS colonisation to target IAP is advocated. However, the optimal screening strategy for preventing EOGBS disease is uncertain.

The UK is in the minority of developed countries in recommending IAP on the basis of certain maternal risk factors.4,5 However, one-third of EOGBS cases do not show any maternal risk factors2 and the use of IAP varies.6 UK policy contrasts with that of many other countries, where antenatal screening, with the culture of vaginal and rectal swabs, is performed. Culture at 35–37 weeks of gestation is also not a wholly accurate predictor of the colonisation status at labour, leading to unnecessary antibiotic treatment for a proportion of women.7,8 This strategy misses preterm births, a high-risk subgroup for EOGBS infection,4,5 and, inevitably, results are not available for all women in labour.9 In countries without the necessary means to implement a universal screening programme, preterm prelabour rupture of membranes is the likely trigger for empirical treatment with antibiotics.10

Rapid, i.e. nonculture-based, tests, if accurate, could be developed into point-of-care tools for intrapartum screening to allow the targeting of IAP during labour. A systematic review has shown that, of the many available intrapartum tests, real-time polymerase chain reaction (PCR) and optical immunoassay (OIA) are the most promising candidates for a head-to-head comparison.11 Manufacturers’ guidelines have suggested that test results are available within 100 and 30 minutes for PCR and OIA, respectively, allowing IAP to be administered to most colonised women. The main objective of this study was to reliably determine the accuracy of PCR and OIA technologies for maternal vaginal and rectal GBS colonisation at the onset of labor, and to compare their performance with risk factor-based screening,4,5 using selective enrichment culture as the reference standard. We also investigated how these tests, risk factors and IAP predicted neonatal GBS colonisation and the acceptability of intrapartum testing to women.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

We conducted a classical prospective test accuracy study12,13 in two UK hospitals (Birmingham Women’s, Birmingham and King George, Ilford, Essex) serving large, socioeconomically and ethnically diverse populations. All pregnant women, other than those for whom an elective caesarean delivery was planned, were invited to take part in the study at their antenatal booking visit or at the mid-trimester scan. Women who had agreed in principle at this stage were formally asked to consent to participation when they presented after 24 weeks of gestation to either the labour ward with anticipated delivery, or when they were admitted to the antenatal ward for the induction of labour. Women who had previously declined or were delivering very prematurely (≤24 weeks) were not re-approached for consent at labour.

Swabs were taken from the lower vagina and rectum, where possible before vaginal examination as the accuracy of tests for maternal GBS colonisation may be altered by vaginal examinations performed prior to obtaining swabs for testing.14,15 Women who first presented in false labour, defined as no delivery within 7 days, were re-approached when they returned in labour. Swabs from false labour were not included in the accuracy analysis.

Results were not provided to the midwives or doctors caring for the women. Individual clinicians made decisions concerning IAP based on one or more established risk factors after swabs had been taken. UK policy is that IAP is indicated for women who have had a previous baby with GBS disease, who have suspected or confirmed chorioamnionitis or who present with preterm, premature rupture of membranes, and should be considered for those with an incidental finding of maternal GBS colonisation or GBS bacteriuria during pregnancy, maternal fever (>38°C), prolonged rupture of membranes (≥18 hours at term) or preterm delivery (<37 weeks).4,5 IAP consisted of intravenous penicillin (3 g) given as soon as possible after the onset of labour and 1.5 g intravenously 4-hourly until delivery. Clindamycin, 900 mg given intravenously 8-hourly, was used for women who were allergic to penicillin. Recent antibiotic therapy (within 14 days of the onset of labour) was documented, but was not a contraindication for inclusion in the study.

To minimise the variation in the amount of bacteria on individual swabs, and to increase patient acceptance, a triple swab, consisting of three separate swabs bound together by two plastic bands (Medical Wire & Equipment Co., Corsham, Wiltshire, UK), was used to collect three samples simultaneously. Vaginal specimens for testing were obtained by gently rotating the swabs across the mucosa of the lower vagina. Rectal swabs were obtained by inserting the swabs through the anal sphincter and then gently rotating. The triple swab was then separated and each swab was used for either one of the intrapartum tests or the reference test. A swab from the neonate was also collected as soon as possible after birth from the external ear canal, the most sensitive indicator of neonatal colonisation,16–18 to determine transmission rates.

PCR, OIA and maternal risk factors4,5 were the index tests used. PCR and OIA were performed separately on rectal and vaginal swabs, following a standard operating procedure, using the Cepheid Smart GBS® Kit and SmartCycler® System (Cephid, Sunnyvale, CA, USA) and Inverness Medical BioStar OIA STREP B® Kit (Discontinued. Inverness Medical UK, Stockport, UK). The majority of tests were run in batches by the research staff, with results usually obtained after delivery to maintain the blinding of test results.

The reference standard for the verification of index test results was laboratory-based selective enrichment culture of maternal and neonatal swabs. Swabs were inoculated into Lim broth with subculture onto trypticase soy agar after overnight incubation at 37°C.17 Biomedical scientists independently interpreted the reference swab cultures and were blind to the intrapartum test results and neonatal outcomes through a password-restricted database. Neonatal swabs could not be used as a reference standard because of ‘treatment paradox’,19 as IAP on the basis of maternal risk factors would alter the rates of transmission from mother to baby.

The timing of sample collection, antibiotic administration, delivery and time taken to perform tests were recorded to assess the adequacy of the timeframe for antibiotic prophylaxis. The acceptability of intrapartum testing to women was evaluated through a study-specific structured questionnaire, developed using focus groups, administered as soon as possible after delivery. This included responses on five-point scales, for which a score of five represented the most positive answer, to questions concerning satisfaction with the information provided, the procedure for obtaining samples and perceptions of intrapartum swabbing as part of routine care. Standardised measures of anxiety, the six-item Spielberger State–Trait Anxiety Index20 and health anxiety Whitely-7,21 were incorporated. The age and ethnicity of women who declined to take part in the study when first approached were also recorded.

Sample size and statistical analysis

Given the potentially serious consequences of GBS infection and the accuracy of other screening strategies, we considered a sensitivity of at least 75% as a minimum for an intrapartum test. Assuming a ‘true’ sensitivity of 90%, 70 GBS-positive mothers would be needed to refute reliably a sensitivity of <75% with 90% power at P < 0.05. To allow for a prevalence of maternal GBS colonisation as low as 5%, a total of 1400 participants was required. The assumption regarding a disease prevalence of 5% was deliberately conservative,2 so as to ensure adequate power for exploratory subgroup analyses should the prevalence be higher.22 An independent monitoring committee reviewed confidential interim analyses, with the remit to decide whether there was both ‘proof beyond reasonable doubt’ that one test was sufficiently accurate and evidence that may reasonably have been expected to influence clinical practice at any stage.23 The committee recommended the continuation of recruitment at each of their two meetings.

The sensitivity and specificity of each index test in determining maternal GBS colonisation were calculated, with 95% confidence intervals (CIs). For the primary analysis, index test results from vaginal and/or rectal swabs were compared against the reference standard of vaginal and rectal laboratory culture, where a positive culture from either site was classed as confirmed colonisation, because GBS harboured in either of these sites could be transmitted to the neonate.24 Conversely, both vaginal and rectal tests had to be negative for the intrapartum test to be defined as negative for maternal GBS colonisation, and similarly for the reference standard. The extent of colonisation was described qualitatively as light, moderate or heavy, subjectively based on the number of bacterial colonies seen on cultured agar plates. The presence of at least one of the five risk factors described above4,5 was considered as a predictor for maternal GBS colonisation.

McNemar’s test was used to test whether the PCR result was identical to the reference standard more frequently than either OIA or risk factors, using a two-sided exact test. Subgroup analyses, planned a priori, were carried out to compare the performance of intrapartum tests according to the site of the swabs, and by the presence or absence of risk factors.

Multivariate logistic regression analyses25,26 were used to investigate whether intrapartum tests added value to risk factor-based screening in the prediction of maternal colonisation and mother to child transmission, and also to generate predictive probabilities for maternal colonisation from various combinations of maternal risk factors and PCR test results. A multivariate model was used, adjusting for the effect of IAP, to determine whether maternal risk factors or PCR test results predicted neonatal GBS colonisation. In this model, we also assessed whether the duration of antibiotic administration prior to birth had an impact on neonatal colonisation. Four hours was considered to be a sufficient duration based on current recommendations;27 any duration less than this was considered to be insufficient.

Proportions, with 95% CIs, of positive responses to questionnaire items were computed and compared with the chi-squared test. Mean scores were analysed using one-way analysis of variance (anova), with post hoc Bonferroni adjusted comparisons. Analyses were performed using stata version 8 (StataCorp, College Station, TX, USA) and spss version 16 (SPPS Inc., Chicago, IL, USA) statistical software.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

Between June 2005 and January 2007, 1418 women were swabbed: 945 from Birmingham and 473 from Essex (Figure 1). Of the 1394 women with history data for analysis, 308 (22%) had one or more risk factors (Table 1). Maternal GBS colonisation rates, as defined by a positive enriched culture from either the vagina or rectum, were 21% (95% CI, 19–24%) for all, 29% (95% CI, 24–34%) in the presence and 19% (95% CI, 17–21%) in the absence of risk factors (Table 2). Of those with a definite indication for IAP, 71% (36/51) received antibiotics; 50% (153/308) of women with any risk factor received IAP and 3% (30/1084) of women with no recorded risk factors received IAP. The neonatal colonisation rate was 8.5% (108/1276) of deliveries where both maternal and neonatal cultures were available, 36% (99/273) for colonised women and 1% (9/1003) for culture-negative women. Fifteen babies had an infection reported immediately postpartum, six of which were invasive. Of the invasive infections, three were diagnosed as EOGBS disease, all of whom recovered. The mothers of these three babies were colonised with GBS: two mothers exhibited fever in labour and were given IAP, but no risk factors were evident in the third. OIA (mean, 38.8 minutes; 95% CI, 38.3–39.3) produced a result considerably more rapidly than the PCR test (mean, 80.8 minutes; 95% CI, 80.4–81.3; P < 0.0001).

image

Figure 1.  Recruitment of women into the intrapartum tests for group B streptococcus diagnostic accuracy study. aNo reason recorded, 827 (44%); objected to rectal swab, 73 (3.9%); objected to swabbing, 241 (12.8%); planned caesarean, 132 (7.0%); delivery planned elsewhere/at home, 17 (0.9%); insufficient English, 111 (5.9%); did not want to participate in research, 199 (10.6%); wanted no added intervention, 77 (4.1%); family refused, 38 (2.0%); worried if things went wrong/anxious about pregnancy, 28 (1.5%); other reasons, 137 (7.3%). Nonparticipants did not have to give a reason, nor were midwives expected to record one. bNo time to collect swabs, 86 (5.5%); caesarean section, 94 (6.0%); delivered elsewhere or at home, 56 (3.6%); language barrier, 2 (0.1%); delivered after end of study, 12 (0.8%); midwife not trained, 8 (0.5%); no sticker on notes/denied giving provisional consent, 314 (19.9%); pregnancy loss, 8 (0.5%); mother too distressed, 4 (0.3%); clinical complications at admission, 58 (3.7%); missed for various unknown reasons, 933 (59.2%). Midwives were not required to record a reason. cThroughout the study, colonisation or test results from either site were only reported if both vaginal and rectal swab results were available. dThe proportion of optical immunoassay (OIA) test results was lower than for polymerase chain reaction (PCR) as the manufacturer of the OIA kit was unable to maintain a constant supply to the study in the later months. Vaginal swabs were prioritised over rectal swabs when supplies of the OIA kit were low.

Download figure to PowerPoint

Table 1.   Characteristics of women participating in an accuracy study of intrapartum polymerase chain reaction (PCR) and optical immunoassay (OIA) tests for maternal group B streptococcus (GBS) colonisation
Number of women with demographic data1400
  1. Parity was not recorded for 89 women; length and type of labour or delivery were not recorded for eight women.

Age (years), mean (SD)29.6 (5.9)
Parity, n (%)
0690 (52.6)
1366 (27.9)
2161 (12.3)
353 (4.0)
432 (2.4)
≥59 (0.7)
Ethnic group, n (%)
White868 (62.0)
Mixed29 (2.1)
Asian297 (21.2)
Black145 (10.4)
Other21 (1.5)
Not given40 (2.9)
Number of women with risk factor data1394
Single risk factor, n (%)
Previous GBS baby4 (0.3)
Incidental finding of GBS in current pregnancy58 (4.2)
Maternal fever12 (0.9)
Prolonged rupture of membranes (PROM)171 (12.3)
Premature18 (1.3)
Multiple risk factors, n (%)45 (3.2)
GBS in current pregnancy and PROM9 (0.6)
Premature and GBS in current pregnancy1 (0.1)
Premature and PROM17 (1.2)
Fever and PROM15 (1.1)
Premature, GBS in current pregnancy and PROM2 (0.1)
Premature, fever and PROM1 (0.1)
Any risk factor, n (%)308 (22.1)
No risk factor, n (%)1086 (77.9)
Type of labour, n (%)
Induction625 (44.6)
Spontaneous767 (54.8)
Not recorded8 (0.6)
Type of delivery, n (%)
Vaginal856 (61.1)
Assisted vaginal233 (16.6)
Caesarean303 (21.6)
Not recorded8 (0.6)
Length of labour (hours), n (%)
<4395 (28.4)
4–6325 (23.3)
7–9252 (18.1)
10–12185 (13.3)
≥13222 (15.9)
Not given21 (0.9)
Table 2.   Maternal group B streptococcus (GBS) colonisation rates in the presence or absence of risk factors
 Risk factor*Total
Present 307 (22%)Absent 1080 (78%)
  1. n = 1385 for both culture results and risk factor data. Risk factor data are missing for 15 participants and rectal culture data are missing for two participants (including one missing risk factor datum).

  2. *Any one of an incidental finding of maternal GBS colonisation or GBS bacteriuria during pregnancy, a previous baby with early-onset group B streptococcus (EOGBS) disease, maternal fever (>38°C), prolonged rupture of membranes (≥18 hours at term) and preterm delivery (<37 weeks).

  3. **Colonisation determined by selective enrichment culture.

Maternal GBS colonisation**
Vaginal
 Present67147214
 Absent2409331173
 Prevalence (%) (95% CI)21.8 (17.6–26.8)13.6 (11.7–15.8)15.4 (13.6–17.4)
Rectal
 Present79188267
 Absent2278921119
 Prevalence (%) (95% CI)25.8 (21.2–31.0)17.4 (15.3–19.8)19.3 (17.3–21.4)
Either
 Present89205294
 Absent2178741091
 Prevalence (%) (95% CI)29.1 (24.3–34.4)19.0 (16.8–21.5)21.2 (19.2–23.5)

Test accuracy and predictors of colonisation

The PCR test of either vaginal or rectal swabs was more sensitive (84% [95% CI, 79–88%] versus 72% [95% CI, 65–77%]) and specific (87% [95% CI, 85–89%] versus 57% [95% CI, 53–60%]) than OIA (P < 0.001) (Table 3). Rectal PCR provided the most sensitive single test for maternal GBS colonisation at 71% (95% CI, 66–76%) compared with 58% (95% CI, 52–64%) for vaginal PCR. The specificity was 92% (95% CI, 90–93%) for both rectal and vaginal PCR. The PCR test, using either swab, was far more sensitive (84 versus 31% [95% CI, 25–36%]) and specific (87 versus 80% [95% CI, 77–82%]) than screening based on the presence of at least one risk factor (P < 0.001).

Table 3.   Accuracy of intrapartum polymerase chain reaction (PCR) and optical immunoassay (OIA) tests and risk factor-based screening of maternal group B streptococcus (GBS) colonisation using either vaginal or rectal enriched culture as reference standard
Test site and testNumber of testsAccuracy measurement (95% confidence intervals)
TPFNTNFPSensitivity (%)Specificity (%)
  1. FN, false negative; FP, false positive; TN, true negative; TP, true positive.

  2. Culture, PCR or risk factor data missing for 60 participants, and so prior probabilities do not match exactly with Table 2.

  3. The sensitivity and specificity of each intrapartum test compared against the individual enriched cultures are shown in Table S1.

  4. *McNemar’s test used for comparison.

Vaginal PCR1711229758358 (52–64)92 (90–94)
Vaginal OIA881658607735 (29–41)92 (90–93)
Comparison PCR versus OIA*    P < 0.001P = 1.0
Rectal PCR208849748671 (66–76)92 (90–93)
Rectal OIA1508049834165 (59–71)59 (56–63)
Comparison PCR versus OIA*    P = 0.154P < 0.001
Either vaginal or rectal PCR2464791513484 (79–88)87 (85–89)
Either vaginal or rectal OIA1646547636372 (65–77)57(53–60)
Comparison PCR versus OIA*    P = 0.001P < 0.001
Risk factors8920687121730 (25–35)80 (77–82)
Comparison risk factors versus PCR*    P < 0.001P < 0.001

The strongest predictor of maternal GBS colonisation was a positive maternal PCR result from either the vaginal or rectal swab, increasing the probability from 21 to 65%. A negative PCR result reduced the probability to 5% (Table 4). The presence or absence of risk factors did not change the post-test probabilities substantially, although, as would be expected from the higher GBS prevalence in those with risk factors (29 versus 19% without risk factors), the tests were somewhat more sensitive in the former group and more specific in those without risk factors (Table S1). There was no evidence that women with risk factors had a greater GBS colonisation burden, based on a qualitative assessment of bacterial growth in culture. Of the 209 women with no risk factors with available vaginal or rectal culture growth data, 70 had light growth of GBS, compared with 19 with light growth and 66 with heavy growth in 85 women with at least one risk factor (χ2 test, P = 0.06). Maternal GBS colonisation rates were similar in women of different ages and ethnic origin (Table S2).

Table 4.   Estimated post-test probability of maternal group B streptococcus (GBS) colonisation combining results obtained from vaginal or rectal polymerase chain reaction (PCR) test with information available on maternal risk factors for GBS colonisation
 Prior probability (%)Estimated post-test probability (%)
  1. n = 1340 with colonisation, PCR and risk factor data, and so prior probabilities do not exactly match with Table 2.

  2. *Colonisation determined by selective enrichment culture.

  3. **Any one of an incidental finding of maternal GBS colonisation or GBS bacteriuria during pregnancy, a previous baby with early-onset group B streptococcus (EOGBS) disease, maternal fever (>38°C) or chorioamnionitis, prolonged rupture of membranes (≥18 hours at term) and preterm delivery (<37 weeks).

Maternal GBS colonisation*Maternal risk factor** Maternal PCR negativeMaternal PCR positive
Not considered21.34.864.5
Absent (n = 1041)19.44.662.6
Present (n = 299)29.75.968.8

After adjusting for the effect of antibiotics, the odds of having a colonised baby were far higher when an intrapartum vaginal or rectal PCR test was positive (odds ratio [OR], 29.4; 95% CI, 15.8–54.8; P < 0.001). The presence of maternal risk factors was not significantly related to neonatal GBS status (OR, 1.44; 95% CI, 0.80–2.62; P = 0.2) (Table 5). Neonatal GBS colonisation rates were reduced with ≥4 hours’ duration of antibiotic use prior to birth (OR, 0.22; 95% CI, 0.07–0.62; P = 0.004), whereas there was no statistically significant association of a reduction with <4 hours’ duration (OR, 0.49; 95% CI, 0.18–1.34; P = 0.2). Neonatal colonisation was 36% (5/14) in babies born to GBS-positive mothers in hours 0–2 after IAP administration, 25% (5/20) in hours 3–6 and 5% (1/20) in hours 7+ (Cochran–Armitage test for trend P = 0.02), compared with 40% (88/219) in babies whose GBS-positive mothers did not receive IAP.

Table 5.   Logistic regression estimates of the association between various factors and neonatal group B streptococcus (GBS) colonisation
Predictor variableMultiple regression*
Odds ratio (95% confidence interval)P
  1. *Multiple logistic regression model with neonatal GBS colonisation determined by selective enrichment culture of neonatal ear swab as the binary dependent variable and any maternal risk factor, intrapartum antibiotic prophylaxis (IAP) and intrapartum vaginal or rectal PCR result as independent variables.

Any maternal risk factor (present/absent)1.44 (0.80–2.62)0.23
Intrapartum antibiotics (≥4 hours IAP/<4 hours IAP or IAP not given)0.22 (0.07–0.62)0.004
Intrapartum antibiotics (any IAP/IAP not given)0.49 (0.18–1.34)0.164
Intrapartum vaginal or rectal polymerase chain reaction (PCR) (positive/negative)29.4 (15.8–54.8)<0.001

Acceptability of intrapartum testing

Overall levels of acceptability for intrapartum testing amongst participants were high, with 80.5% (95% CI, 80.0–81.3%) satisfied or very satisfied with the information provided, 94.3% (95% CI, 92.9–95.7%) happy or very happy with the way the swabs were taken and 94.1% (95% CI, 93.1–95.2%) confident in its use in routine care. However, the women who participated found vaginal swabbing to be more comfortable (82.4 versus 70.1%, P < 0.0001) and less embarrassing (96.2 versus 87.7%, P < 0.0001) than rectal swabbing. There was no evidence that screening raised anxiety. The mean state anxiety score (34.1; 95% CI, 33.3–34.9) was within the normal range for the test (normative value of 35.2 for working adult females).20,28

There was a significant association between a refusal to take part in the accuracy study and ethnicity (P < 0.001) and age (P < 0.001), with South Asian women and younger women less likely to participate. No reason was recorded for the majority of participants declining to participate in the study. However, white British and Irish and younger women were over-represented in those who declined because they objected to the swabs, particularly rectal swabs.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

PCR was more accurate than OIA, and far more accurate than risk factor-based screening as an intrapartum test for maternal GBS colonisation. Risk factors did not predict neonatal GBS colonisation, whereas results from PCR were highly predictive. The presence or absence of risk factors did not alter the post-test probabilities obtained with PCR tests substantially. The prevalence of maternal GBS colonisation and test accuracy varied with the site of swabbing, with the accuracy higher for rectal than vaginal swabs, and highest using both vaginal and rectal swabs, which may be related to the degree of colonisation. Although rectal swabbing is necessary to maximise sensitivity, participants found rectal swabs less comfortable and more embarrassing than vaginal swabs. The accuracy of PCR, when considering samples from both vagina and rectum, compares favourably with that of screening by culture of swabs taken at 35–37 weeks of gestation.8

Our study was larger and methodologically more robust than previous studies.11 We recruited our prespecified sample so that it had sufficient power to exclude a clinically unacceptable accuracy and to allow the investigation of heterogeneity across subgroups. We complied with the STARD statement13 and minimised bias as far as possible, ensuring that tests and reference standard were performed independently and interpreted blind to each other.29 A high proportion of index tests were verified by reference standard: over 96% for PCR. Our estimates of the sensitivity of PCR and OIA tests were not as high as those reported previously,11 but are more likely to be valid as meta-analysis of studies of variable quality can overestimate accuracy.29 A recent good quality study using the next generation of PCR technology and vaginal swabs only reported a high degree of accuracy (sensitivity, 98.5%; 95% CI, 94.8–99.8%; specificity, 99.6%; 95% CI, 98.8–99.9%); however, intrapartum PCR results were not available for 105 of the 968 women (10.8%) because of technical problems.7

The weak association between risk factors and maternal and neonatal GBS colonisation in our study contrasts with the previously observed associations with EOGBS disease in surveillance studies.30 Such studies can produce unreliable results as a consequence of bias in the retrospective ascertainment of risk factors,29 and a strength of our prospective study is that risk factors were identified in normal clinical practice and compared directly with the conditions targeted for treatment, namely maternal and neonatal GBS colonisation. Our model did, however, control for IAP usage, which, in this study, should have been directed by the presence of risk factors; therefore, the role of risk factors in predicting neonatal colonisation cannot truly be captured because of this causal relationship. That a positive PCR result increased significantly the risk of neonatal colonisation in this model cannot be disputed.

Our study population was selected. Only about one-third of all women delivering in the two centres were approached about the study and, of these, only 28% were recruited, representing about 10% of the total number of deliveries. Unlike the recent French study,7 we were unable to obtain a waiver of consent, which impeded recruitment and rectal swabbing was a probable reason. Compared with all women delivering at the two centres, proportionally more white women, fewer younger women, fewer premature deliveries and greater numbers of labour inductions or emergency caesarean sections were recruited. However, with GBS prevalence similar at different ages and in different ethnic groups, we consider that the study findings are probably generalisable to routine care, at least in the UK. A lower proportion of our study participants had risk factors than in other studies – 22% compared with 29% in a meta-analysis2– possibly because of fewer premature births. By contrast, the rate of maternal GBS, at 21% for combined vaginal or rectal colonisation, was higher than that reported previously (14%),2 partly explained by some studies reporting vaginal colonisation only. The risk of neonatal GBS colonisation given maternal colonisation was similar to the pooled estimate from other studies.2 The incidence of EOGBS disease, at 2.14 cases per 1000 live births, was higher than may have been anticipated, but from a small cohort and not inconsistent with UK national data.1 The anxiety levels of those who participated were slightly lower than the values obtained in a previous study of antenatal screening for GBS during pregnancy.31

The most striking finding that has implications for clinical practice and policy is the low sensitivity of risk factor-based screening compared with PCR or culture tests in predicting maternal and neonatal GBS colonisation – below that which we considered to be a minimally acceptable sensitivity for our study – which calls into question the validity of the current UK policy.4,5 Moreover, consistent with previous evidence of practice variation,6 the risk factor-based screening policy was poorly adhered to, with one-third of women with indications for IAP not treated.

PCR screening would be a far more sensitive test to guide IAP at labour, but a substantial limitation is the 80 minutes taken to obtain a result, which would have been too late to obtain the full effect of IAP for 50% of women in our study. Screening of women for GBS colonisation with the culture of vaginal and/or rectal swabs taken at 35–37 weeks of gestation – as practised in many countries32– is not quite as sensitive as the PCR test in detecting maternal intrapartum GBS colonisation (76 versus 86%),8 but does allow IAP to be initiated on admission to the labour ward. In our study, this would still have been shorter than recommended for the 28% of women who delivered within 4 hours of admission to the labour ward. Moreover, with screening at 35–37 weeks of gestation, substantial numbers of mothers with false-negative GBS tests would not receive IAP. Therefore, given the relatively low cost of IAP and the high provider and societal costs of EOGBS, offering IAP to all women at the onset of labour may be the most cost-effective strategy to minimise the risk of EOGBS. Clinicians would be reluctant to adopt such a policy, however, because of concerns that overtreatment with antibiotics could have delayed health consequences, particularly for infants.33 Widespread unselective use of antibiotics may also encourage antibiotic resistance, because, although GBS isolates appear to remain sensitive to β-lactams, resistance to clindamycin and erythromycin is increasing in prevalence.34 These poorly quantified risks are key factors in determining the optimum screening policy, and further research would be useful.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

If an intrapartum test is to be considered for practice, PCR is more accurate than OIA, which has a much lower sensitivity. However, the current point-of-care testing procedure is too cumbersome to be practical and is not sufficiently rapid to introduce antibiotics in sufficient time prior to birth to be effective. The ideal test would need to be sufficiently rapid to allow at least 4 hours IAP administration prior to delivery.27 The PCR technology will need to be simplified and made more rapid to be useful as a point-of-care test. New developments in PCR platforms that eliminate the preparatory steps may allow this in the future,7 but their cost-effectiveness compared with other strategies needs to be demonstrated.

Disclosure of interest

All authors declare that they have no conflict of interest.

Contribution to authorship

JPD and KSK had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept: KSK, JG. Study design: JPD, JG, HP, RG, RKH, KSK. Data acquisition and management: JPD, HP, KSK. Analysis of data: JPD, HP, KSK. Drafting the manuscript: JPD, JG, HP, KSK. Critical review and approval of the final version of the manuscript: all authors.

Details of ethics approval

Research ethical approval was obtained from East London and the City Research Ethics Committee on 8 December 2004 (reference P3/04/Q0605/37). NHS Trusts’ research governance approval was obtained at both sites.

Funding

This study was funded by a grant from the National Institute for Health Research Health Technology Assessment Programme http://www.hta.ac.uk (02/38/04). The sponsor, under the UK Research Governance Framework, was the University of Birmingham. The funder and the sponsor played no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. The opinions and conclusions expressed here are those of the authors and do not necessarily reflect those of the UK National Health Service or the Department of Health. The University of Birmingham Clinical Trials Unit receives core funding from the Department of Health.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

We thank all the delivery suite, antenatal clinic and community midwives for their commitment and dedication to the study. Finally, we wholeheartedly thank all the mothers who agreed to participate in this study.

Collaborative group

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

The GBS Collaborative Group consists of the authors, together with those involved with the recruitment of women and rapid testing: Elizabeth Edwards, Gemma Barnfield, Michelle Tenage, Philip Milner, Peter Thompson (Birmingham Women’s Healthcare NHS Foundation Trust), Lisa Spicer, Richard Howard (Barking, Redbridge and Havering Hospitals NHS Trust); central coordination and interim analyses: Elisabeth King (University of Warwick), Laura Magill, Edward Tyler, Leanne Fulcher (University of Birmingham Clinical Trials Unit), Laura Buckley (CRUK Clinical Trials Unit, University of Birmingham); the acceptability assessment: Nicola Elliman (Aston University); the parallel economic evaluation: Tracy Roberts, Billingsley Kaambwa, (University of Birmingham), Stirling Bryan (University of British Columbia). SB, RH and EK were employed by the University of Birmingham at the time of the study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information
  • 1
    Heath PT, Balfour G, Weisner AM, Efstratiou A, Lamagni TL, Tighe H, et al. Group B streptococcal disease in UK and Irish infants younger than 90 days. Lancet 2004;363:2924.
  • 2
    Colbourn T, Gilbert R. An overview of the natural history of early onset group B streptococcal disease in the UK. Early Hum Dev 2007;83:14956.
  • 3
    Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal Group B streptococcal colonization. Cochrane Database Syst Rev 2009;Issue 3.
  • 4
    Royal College of Obstetrics and Gynaecology. Prevention of Early Onset Neonatal Group B Streptococcal Disease. Guideline No. 36. London: Royal College of Obstetrics and Gynaecology, 2003.
  • 5
    Health Protection Agency. Interim ‘Good Practice’ Recommendations for the Prevention of Early Onset Neonatal Group B Streptococcal (GBS) Infections in the UK. London: Health Protection Agency, 2004.
  • 6
    Cromwell D, Joffe T, van der Meulen J, Dhillon C, Hughes R, Murphy D. The Prevention of Early-Onset Neonatal Group B Streptococcal Disease in UK Obstetric Units: an Audit of Reported Practice in England, Scotland, Wales and Northern Ireland. London: Royal College of Obstetrics and Gynaecology, 2007.
  • 7
    El Helali N, Nguyen JC, Ly A, Giovangrandi Y, Trinquart L. Diagnostic accuracy of a rapid real-time polymerase chain reaction assay for universal intrapartum group B streptococcus screening. Clin Infect Dis 2009;49:41723.
  • 8
    Colbourn T, Asseburg C, Bojke L, Philips Z, Claxton K, Gilbert R. Prenatal screening and treatment strategies to prevent group B streptococcal and other bacterial infections in early infancy: cost effectiveness and expected value of information analyses. Health Technol Assess 2007;11:1226.
  • 9
    Van Dyke MK, Phares CR, Lynfield R, Thomas AR, Arnold KE, Craig AS, et al. Evaluation of universal antenatal screening for group B streptococcus. N Engl J Med 2009;360:262636.
  • 10
    Festin M. Antibiotics for Preterm Rupture of Membranes: RHL Commentary. Geneva: World Health Organisation, 2003.
  • 11
    Honest H, Sharma S, Khan KS. Rapid tests for group B streptococcus colonization in laboring women: a systematic review. Pediatrics 2006;117:105566.
  • 12
    Knottnerus JA, Muris JW. Assessment of the accuracy of diagnostic tests: the cross-sectional study. J Clin Epidemiol 2003;56:111828.
  • 13
    Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Br Med J 2003;326:414.
  • 14
    Brady K, Duff P, Schilhab JC, Herd M. Reliability of a rapid latex fixation test for detecting group B streptococci in the genital tract of parturients at term. Obstet Gynecol 1989;73:67881.
  • 15
    Brady K, Sizemore KL, Duff P, Aamodt LW. The effect of bacteriostatic lubricant on group B streptococcal cultures of the female genital tract. Obstet Gynecol 1989;74:84850.
  • 16
    Fowlie PW, Schmidt B. Diagnostic tests for bacterial infection from birth to 90 days – a systematic review. Arch Dis Child Fetal Neonatal Ed 1998;78:F928.
  • 17
    Thompson PJ, Greenough A, Gamsu HR, Nicolaides KH, Philpott-Howard J. Congenital bacterial sepsis in very preterm infants. J Med Microbiol 1992;36:11720.
  • 18
    Baker CJ, Clark DJ, Barrett FF. Selective broth medium for isolation of group B streptococci. Appl Microbiol 1973;26:8845.
  • 19
    Whiting P, Rutjes AW, Reitsma JB, Glas AS, Bossuyt PM, Kleijnen J. Sources of variation and bias in studies of diagnostic accuracy: a systematic review. Ann Intern Med 2004;140:189202.
  • 20
    Marteau TM, Bekker H. The development of a six-item short-form of the state scale of the Spielberger State–Trait Anxiety Inventory (STAI). Br J Clin Psychol 1992;31:3016.
  • 21
    Fink P, Ewald H, Jensen J, Sorensen L, Engberg M, Holm M, et al. Screening for somatization and hypochondriasis in primary care and neurological inpatients. A seven-item scale for hypochondriasis and somatization. J Psychosom Res 1999;46:26173.
  • 22
    Bachmann LM, Puhan MA, Ter Riet G, Bossuyt PM. Sample sizes of studies on diagnostic accuracy: literature survey. Br Med J 2006;332:11279.
  • 23
    Pocock SJ. When to stop a clinical trial. Br Med J 1992;305:23540.
  • 24
    Dillon HC Jr, Gray E, Pass MA, Gray BM. Anorectal and vaginal carriage of group B streptococci during pregnancy. J Infect Dis 1982;145:7949.
  • 25
    Miettinen OS, Henschke CI, Yankelevitz DF. Evaluation of diagnostic imaging tests: diagnostic probability estimation. J Clin Epidemiol 1998;51:12938.
  • 26
    Bachmann LM, Khan KS, Ogah J, Owen P. Multivariable analysis of tests for the diagnosis of intrauterine growth restriction. Ultrasound Obstet Gynecol 2003;21:3704.
  • 27
    Illuzzi JL, Bracken MB. Duration of intrapartum prophylaxis for neonatal group B streptococcal disease: a systematic review. Obstet Gynecol 2006;108:125465.
  • 28
    Spielberger CD. State–Trait Anxiety Inventory Manual. Palo Alto, CA: Consulting Psychologists Press Inc., 1983.
  • 29
    Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, Prins MH, van der Meulen JH, et al. Empirical evidence of design-related bias in studies of diagnostic tests. J Am Med Assoc 1999;282:10616.
  • 30
    Heath PT, Schuchat A. Perinatal group B streptococcal disease. Best Pract Res Clin Obstet Gynaecol 2007;21:41124.
  • 31
    Cheng PJ, Shaw SW, Lin PY, Huang SY, Soong YK. Maternal anxiety about prenatal screening for Group B streptococcus disease and impact of positive colonization results. Eur J Obstet Gynecol Reprod Biol 2006;128:2933.
  • 32
    Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of perinatal group B streptococcal disease. Revised guidelines from CDC. MMWR Recomm Rep 2002;11:122.
  • 33
    Bedford Russell AR, Murch SH. Could peripartum antibiotics have delayed health consequences for the infant? BJOG 2006;113:75865.
  • 34
    Panda B, Iruretagoyena I, Stiller R, Panda A. Antibiotic resistance and penicillin tolerance in ano-vaginal group B streptococci. J Matern Fetal Neonatal Med 2009;22:1114.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgements
  9. Collaborative group
  10. Study monitoring committee
  11. References
  12. Supporting Information

Table S1. Sensitivity and specificity of intrapartum tests for maternal colonisation according to the presence or absence of risk factors.

Table S2. Maternal group B streptococcus (GBS) colonisation rates in different age and ethnic groups.

FilenameFormatSizeDescription
BJO_2725_sm_TableS1-2.doc75KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.