Parts of this report have been reproduced from Daniels et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.
Group B Streptococcus disease
Group B Streptococcus (GBS) (Streptococcus agalactiae) is a ubiquitous bacterium that forms part of the normal bacterial flora of the gut and genital tract. In adults, GBS is an occasional cause of serious systemic infections in immunocompromised people, but is more commonly seen as an opportunistic pathogen of the female urogenital tract.2 However, if a neonate, whose immune system is immature, is exposed to GBS then it can lead to sepsis and death. Most systemic GBS infections usually present within 24 hours of delivery as rapidly progressing septicaemia. However, early-onset GBS disease of the newborn is defined by the National Institute for Health and Care Excellence (NICE) as occurring within the first 72 hours of life,3 and by the Royal College of Obstetricians and Gynaecologists (RCOG) as occurring before the first 7 days of age.4,5 Exposure to GBS present in the genital tract of the mother during birth is thought to be the most common route for early-onset colonisation in the neonate. The incidence of early-onset GBS sepsis can be decreased if women colonised with GBS or who have risk factors associated with early-onset GBS infection are given intrapartum antibiotic prophylaxis (IAP) during labour. Any infection with GBS in babies between 7 days and 3 months of age is deemed late onset and is more often associated with localised infections (especially meningitis and pneumonia). Colonisation from environmental sources is thought to be the most common cause of late-onset GBS infection and is beyond the remit of the Group B Streptococcus 2 (GBS2) trial.
Early-onset group B Streptococcus infection
Maternal colonisation
The gastrointestinal tract is the natural reservoir of GBS in humans and is the likely source of vaginal colonisation. Asymptomatic maternal colonisation with GBS has been reported at a global level of 15%,6 although this figure can vary with race, region and the method of laboratory culture used for its detection.6,7 A previous review of UK studies indicated an overall colonisation rate of 18% from vaginal/rectal swabs [95% confidence interval (CI) 16% to 21%], higher than from vaginal swabs alone.8
Transmission
Neonates with early-onset GBS infection show initial colonisation mainly in the mucous membranes of their respiratory tract, and the major route of vertical transmission at the time of birth is thought to be through aspiration of vaginal, rectal and amniotic aerosols during birth. Vertical transmission in utero is thought to occur as a consequence of prolonged rupture of membranes (although GBS can cross intact membranes9) and is regarded as one of the causes of stillbirth.10 Colonisation of the mother is less predictive for late-onset GBS infection, with prematurity being the major risk factor.11
The association between the rates of maternal colonisation, transmission and infection has been established (Figure 1). A meta-analysis of six studies of the maternal and baby colonisation rates in an untreated general population showed a transmission rate between the colonised mother and her baby of 36.4% (95% CI 28.1% to 45.0%). A further analysis in the same report gave an average incidence of 3.0% (95% CI 1.6% to 4.7%) of babies born to colonised mothers who went on to develop early-onset GBS disease.8 There is some evidence that maternal colonisation with GBS is associated with preterm birth, particularly ascending infection due to GBS bacteriuria.12
FIGURE 1 Model of colonisation, transmission and early-onset GBS disease. (Figure not reproduced here; see the original article.)
Burden of early-onset group B Streptococcus infection
Group B Streptococcus remains one of the most important causes of severe early-onset infection in newborn infants. A systematic review estimated that the global incidence of early-onset GBS infection was 0.41 (95% CI 0.36 to 0.47) per 1000 live births, with the highest burden found in the Caribbean and South Africa.13 Enhanced surveillance in the UK and Ireland from 2014 to 2015 showed an incidence of culture-proven GBS infection in babies aged < 90 days of 0.94 (95% CI 0.88 to 1.00) per 1000 live births.14 In this study, 60% of cases were early-onset infection, a rate of 0.57 (95% CI 0.52 to 0.62) per 1000 live births, an increase from 0.48 per 1000 live births in 2000. In the USA, the incidence of early-onset GBS infection fell from 1.7 per 1000 live births in the early 1990s to 0.22 per 1000 live births in 2014, without any concurrent rise in Gram-negative sepsis.15
The global case–fatality ratio is estimated at 8.4%, with a fourfold variation between Africa and high-income countries,13 whereas in the UK it was 6.2% in 2014, an improvement from 9.6% in 2001, possibly because of the improvements in neonatal care. Mortality is much higher in preterm babies. Oddie and Embleton16 found that preterm infants accounted for 38% of all cases and 83% of the deaths from early-onset GBS infection. Information on morbidity among survivors is less clear, but significant long-term morbidity, including impaired psychomotor development, has been reported in up to 30% of survivors.17
It is highly likely that some cases of serious early-onset sepsis caused by GBS are unrecognised because cultures of blood and cerebrospinal fluid are negative. By taking into account superficial swab culture results from all neonates who underwent a septic screen in the first 72 hours of life, Luck et al.18 concluded that the true incidence of early-onset GBS disease in the UK may be as high as 3.6 per 1000 live births (i.e. over seven times higher than previously estimated). Epidemiological studies have suggested that various factors present at the time of birth are associated with the neonate having an increased risk of developing GBS disease, presenting as either an early- or late-onset infection. UK surveillance data suggest that only 35% of infants with early-onset GBS infection were born to mothers with one or more risk factors,14 as defined by the RCOG guidelines implemented at that time.19
Maternal risk factor for neonatal disease
Maternal risk factors for vertical GBS transmission to the newborn disease have been suggested to include the following.
Prematurity
Colonised premature babies are at a high risk of developing early-onset GBS infection, as their immune system is immature and they are less likely to have received passive immunity transplacentally. The pooled odds of maternal colonisation in preterm births, compared with births at > 37 weeks’ gestation, was 1.53 (95% CI 1.14 to 2.05). Birth weight is highly correlated with prematurity and inversely related to developing early-onset GBS infection. The UK surveillance study indicated an incidence of 2.24 (95% CI 1.31 to 3.59) early-onset GBS infection cases per 1000 live births in babies weighing < 1500 g at birth, compared with 0.43 (95% CI 0.38 to 0.49) per 1000 live births in those weighing ≥ 2500 g at birth.14
Prelabour rupture of membranes
Prelabour rupture of membranes (PROM) with a delay in progress to established labour would be expected to lead to an increased likelihood of ascending infection and baby colonisation in utero, although there is debate as to what, if any, role the presence of GBS plays in the induction of PROM. Rupture of the membranes > 18 hours before birth is significantly associated with early-onset GBS infection (odds ratio 25.8, 95% CI 10.2 to 64.8) compared with non-infected infants.17 Therefore, babies born to mothers who experience preterm labour with PROM of any duration, or preterm labour if there is suspected or confirmed intrapartum rupture of membranes lasting > 18 hours, are especially thought to be at risk of developing early-onset GBS infection. UK NICE guidelines recommend that induction of labour at term is offered at 24 hours after PROM.3
Maternal fever
Intrapartum fever is also highly associated with the development of early-onset GBS infection (odds ratio 10.0, 95% CI 2.4 to 40.8).16 In a UK surveillance study, 19% of babies with early-onset GBS infection were born to mothers with intrapartum pyrexia and 31% to mothers with suspected chorioamnionitis.14
Previous baby with group B Streptococcus disease
The proportion of women giving birth at term who have had a previous baby with early-onset GBS infection was estimated at 0.08%. These women are considered to be at higher risk of another infected baby, than multiparous women whose previous babies were not affected, although the data are insufficient to describe the size of this increased risk.20
Group B Streptococcus detected in current pregnancy
Group B Streptococcus bacteriuria is associated with higher risk of chorioamnionitis and neonatal disease, although the data are insufficient to quantify the increased risk.21,22 In the 2014 UK surveillance study, 9% of mothers of GBS-infected babies were known to be colonised with GBS, although a figure of 5% was used in a model of the effectiveness of screening.23
Testing for risk of group B Streptococcus disease
The aim of maternal GBS testing is to prevent early-onset GBS infection. However, no tests discriminate between colonised mothers who will or will not transmit GBS to their babies, or between babies who will or will not develop early-onset GBS infection. Instead, there are several methods for identifying GBS maternal colonisation in late pregnancy or during labour.
Tests for maternal group B Streptococcus colonisation
Bacteriological culture
The sensitivity of testing methods based on the culture identification of maternal colonisation of GBS depends on the timing of specimen collection, the source of the specimen and the culture technique used by the microbiology laboratory. A systematic review of prospective studies showed that a single vaginal/rectal swab has a positive predictive value (i.e. the proportion positive at 35–37 weeks’ gestation who remain positive at term labour) of 70.2% and a negative predictive value of 95.2%, but there were variations in testing methods.24 Overall, using retrospective and prospective studies and different types of swabs and culture methods, sensitivities of the antenatal tests ranged from 51% to 71% and specificities were consistently around 85%.24
Molecular tests for group B Streptococcus colonisation
Development of molecular methods that allow rapid bedside detection of microorganisms, such as polymerase chain reaction (PCR) methods, offers the potential to target use of antibiotics more specifically than was previously possible. Previous experience shows that implementation of complex point-of-care (POC) tests for GBS colonisation are technically feasible.25 However, the practical value of any POC test depends on accurate results being reliably available within the clinically required time frame. To this end, careful consideration is required of a number of factors, including the expected frequency of testing, achievable results turnaround times, the amount of hands-on test time, strategies to deal with test failure and assurance of the ongoing availability of sufficient trained staff able to undertake testing when required.
The majority of the commercially available test systems required multiple preparation and incubation steps. What is required, is a system that can use swabs directly and be operated by midwives on a maternity unit. In the light of these limitations, the technology used in the GBS2 trial was the Cepheid GeneXpert® system (Cepheid, Maurens-Scopont France), which is feasible as a POC test. The Xpert® GBS test (Cepheid) for this platform allows detection of GBS within 35 minutes of placing a swab into a cartridge and loading one machine, with a hands-on preparation time of < 2 minutes.
Accuracy of GeneXpert rapid test to detect group B Streptococcus colonisation
The Cepheid GeneXpert GBS system has been available since 2008 and several groups have assessed its accuracy in studies, although with different swabbing strategies and reference standard comparators. The manufacturer cites a sensitivity (1 – false-positive rate) of 91.9% (95% CI 84.7% to 96.5%) and a specificity of 95.6% (95% CI 95.0% to 98.9%) for intrapartum vaginal/rectal samples, compared with selective enrichment culture of a second intrapartum swab.26 A meta-analysis, undertaken in 2014, of nine studies, which used an enrichment culture method for the reference standard, obtained estimates of the average accuracy of the test of sensitivity 96.4% (95% CI 90.8% to 98.6%) and specificity 98.9% (95% CI 97.5% to 99.5%) (Jane Daniels, University of Nottingham, 2014, personal communication). Overall, the quality of the studies was rated as being adequate, but the potential for biases remained, from the lack of blinding, the use of only vaginal swabs and the inappropriate use of non-enrichment culture as the reference standard, which could underestimate colonisation rates. A subsequent meta-analysis of 15 studies using various test PCR platforms produced a pooled sensitivity of 93.7% (95% CI 92.1% to 95.3%) and specificity of 97.6% (95% CI 97.0 % to 98.1%).27 These accuracy parameters, together with the turnaround time, suggest that the GeneXpert system meets the proposed criteria set by the US Centers for Disease Control and Prevention for a clinically useful GBS test, namely that it should consist of a simple bedside kit that can be used by delivery suite staff, have a turnaround time of < 30 minutes, and have a sensitivity and specificity of ≥ 90%.26
Current policy for prevention of early-onset group B Streptococcus infection
The strategy recommended by the UK’s RCOG is to consider and offer IAP to women who have risk factors identified during the pregnancy or in labour for having a baby with early-onset GBS infection.4 Reviews by the UK National Screening Committee (NSC)24,28 and NICE3 have endorsed this approach.
The risk factors to consider in this approach, and the management options available in the original,29 the 201219 and the 20174 revised guidance are summarised below.
Women with a previous baby with group B Streptococcus disease
- 2003, 2012, 2017 guidelines: offer IAP.
Women with group B Streptococcus bacteriuria in the current pregnancy
- 2003 guidelines: consider IAP.
- 2012 and 2017 guidelines: offer IAP.
Women with an incidental finding of vaginal or rectal group B Streptococcus colonisation, or from an intentional test, in the current pregnancy
- 2003 guidelines: consider IAP.
- 2012 and 2017 guidelines: offer IAP.
Prematurity of < 37 weeks’ gestation without known group B Streptococcus colonisation
- 2003 guidelines: discuss IAP.
- 2012 guidelines: do not offer IAP in women presenting in established preterm labour with intact membranes and with no other risk factors for GBS, unless they are known to be colonised with GBS.
- 2017 guidelines: recommend IAP to all women in confirmed preterm labour.
Prolonged rupture of membranes > 18 hours
- 2003 guidelines: consider IAP.
- 2012 guidelines: states that the evidence for IAP is unclear for women at term with PROM.
- 2017 guidelines: if a woman is known to be colonised then she should be offered IAP and the induction of labour. If colonisation status is negative or unknown, offer the induction of labour immediately or by 24 hours.
Fever in labour > 38 °C
- 2003 guidelines: discuss IAP.
- 2012 and 2017 guidelines: offer IAP.
The 2017 revision recommends that (1) antenatal antibiotics (before labour starts or membranes rupture) are not offered to women, even if proven to be colonised with GBS by a vaginal or rectal swab, and (2) women undergoing planned caesarean in absence of labour and with intact membranes, whether at term or preterm, do not require GBS-specific antibiotics. A further recommendation is that women who were diagnosed as GBS carriers in a previous pregnancy, the opportunity for culture-based testing in the late third trimester or IAP without testing should be offered in subsequent pregnancies.
NICE issued guidance in 2012 on antibiotics for the prevention and treatment of early-onset neonatal infection.3 The guidance recommends that IAP should be offered to women who have had:
- a previous baby with an invasive GBS infection
- GBS colonisation, bacteriuria or infection in the current pregnancy.
The guidance suggests that IAP is considered for women in preterm labour if there is:
- PROM of any duration
- suspected or confirmed intrapartum rupture of membranes lasting > 18 hours.
For women with PROM at term, including prolonged (> 24 hours) rupture, the use of prophylactic antibiotics is not recommended.30
Evidence for group B Streptococcus testing strategies
Introducing universal testing for maternal GBS colonisation into the UK health-care system has been considered alongside other testing and vaccination strategies from the perspectives of both cost-effectiveness and overtreatment. The UK NSC reviewed the evidence for universal and risk factor-based screening in 2012 and again in 2017, and at both times concluded that there was insufficient evidence against its standardised criteria to justify a change from the current risk factor-based screening approach to guide administration of IAP.24,28 The NSC estimated that the number of women needing to receive IAP on the basis of culture-based testing at 35–37 weeks to prevent one case of early-onset GBS infection missed by the risk factor approach was 1000–1500 in the 2012 analysis and 1675–1854 in the 2017 analysis. In the 2019 hypothetical cohort, an additional 96,260 women would receive IAP under a testing policy; however, with the test having a positive predictive power of 0.2%, the overwhelming majority of infants born to women receiving IAP would never have been at risk.23 However, the model’s input parameters have been called into question, as the risk factor strategy emerged with an early-onset GBS infection rate of 0.49 per 1000 live births, which is significantly lower than UK surveillance data suggest.14,23
A French study31 assessed whether or not rapid intrapartum testing for GBS reduced the proportion of women inappropriately given IAP, compared with a hypothetical situation in which the national screening programme standard of antenatal testing prevailed. The French investigators assessed the diagnostic performance of both the GeneXpert test and microbiological screening at 34–38 weeks’ gestation compared with a reference standard of microbiological culture of intrapartum swabs. IAP was directed during the study by the rapid test, and the adequacy of this strategy compared with what hypothetically would have been followed by use of the 35–37 weeks’ gestation culture result. The authors31 concluded that the universal rapid intrapartum test would correspond to an absolute risk reduction of 0.925%, equivalent to 108 more women needing to be tested in labour and provided IAP to prevent a single case of early-onset GBS infection that would be missed by the culture-based testing.31
Acceptability and implementation of testing for maternal group B Streptococcus colonisation
Even if the trial shows the superiority of one test over the other, the likelihood of successful implementation will depend on the acceptability of the test for childbearing women and for maternity care professionals. The previous UK GBS rapid testing study24 did examine the acceptability of the swabbing process and the information provided, but in a hypothetical situation in which the results were not being used to direct care. Although study participation was less than half of those approached, among the women who did agree to participate, there was a high level of satisfaction with the process, with 80.5% of women satisfied or very satisfied with the information provided, 94.3% of women happy or very happy with the way the swabs were taken and 94.1% of women confident in its use in routine care. The NSC consider evidence that the criterion, namely that any test should be acceptable to the target population, has been met as uncertain.24
Cost-effectiveness of group B Streptococcus testing for maternal group B Streptococcus colonisation
At the time of the initiation of the GBS2 trial, the most relevant cost-effectiveness evidence was a 2010 decision model, which suggested that a policy of routine culture testing at 35–37 weeks was considered the most cost-effective once routine untargeted IAP was eliminated as a potential strategy.32 Intrapartum testing using first-generation PCR tests was not cost-effective unless the average cost of testing was reduced from £29.95 to £7.00. The 2017 NSC review did not identify any new cost-effectiveness estimates relevant to testing in a UK setting.4 Economic models do not yet have the data to incorporate aspects such as the effect of widespread use of antibiotics on the development of antibiotic resistance and the impact this will have, the impact of maternal and neonatal microbiomes, and the effect of very rare but potentially catastrophic anaphylaxis in labour.
Antibiotic resistance
Antibiotic resistance is considered an imminent threat to human health and the impact of antibiotic resistance spans all ages, including infants.33–35 Children may be colonised with antibiotic-resistant bacteria early in life.36 Antibiotic (treatment)-resistant bacteria have been increasingly shown to cause early- and late-onset neonatal sepsis and neonatal intensive care unit sepsis outbreaks.37,38 Studies of transmission of bacteria from mothers to their infants have predominantly focused on the causative agents of early neonatal sepsis: Streptococcus agalactiae (GBS) and Escherichia coli.39,40 Risk factors for neonatal colonisation with multiantibiotic-resistant strains of E. coli, for example extended-spectrum β-lactamase (ESBL)-producing strains, have been described, but the relative contribution of maternal carriage at the time of birth is uncertain.41 As part of an Olympics surveillance project and in collaboration with Public Health England in 2014, ESBL-producing Enterobacteriaceae were isolated from 20% of women of childbearing age (i.e. 15–45 years) who submitted a faecal sample for laboratory examinations in the north-east of London.42 At the same time, colonisation with ESBL-producing Enterobacteriaceae has been demonstrated in 7% of neonatal intensive care unit infants of < 31 weeks’ gestational age recruited into a multicentre double-blind placebo-controlled randomised probiotic feeding study in the south-east of England.43
Strategies for control of antibiotic resistance have focused on antibiotic stewardship programmes designed to reduce the selection pressure for resistance and early detection of carriers. Carriers of antibiotic-resistant microbes can be identified by screening and actions can be taken to prevent the spread from carriers to others, including the implementation of contact precautions and decontamination of colonised individuals, when this is feasible. By contrast with methicillin-resistant Staphylococcus aureus (MRSA), there is currently no reliable method of decontamination of individuals colonised with antibiotic-resistant Enterobacteriaceae, including those with ESβLs. ESβLs have recently emerged in community-acquired E. coli and Klebsiella pneumoniae, and their identification as causal agents of infections in neonatal units and the lack of effective therapeutic options is a worrying development.36
Prevalence of antibiotic resistant group B Streptococcus
Penicillin remains the first choice of antibiotic for GBS prophylaxis and treatment.44–47 However, since 2008, strains with reduced susceptibility have emerged. Studies in Japan have detected alarming rates of resistance, with 14.7% of GBS isolates testing resistant to penicillin, a marked increase from 2.3% in 2005–6, albeit from a range of samples and populations and not vaginal/rectal swabs from pregnant women.48 Although resistance to penicillin appears to be low outside Japan, resistance to the secondary antibiotics of choice, clindamycin and erythromycin, is high, and so it is possible that these will not forever be available as an alternative for penicillin-allergic women. Clindamycin was removed as a recommended alternative to benzylpenicillin in the 2017 RCOG guidelines, as the resistance rate in the UK is reported as 16%.14 Other research highlights substantial risks of adverse consequences arising from exposure of the fetus and newborn infant to unnecessary antibiotics, including necrotising enterocolitis, inflammatory bowel disease, fungal infection49 and cerebral palsy.50