*For more information on using chlorhexidine (Hibiclens) instead of antibiotics to treat GBS during labor see this article.
Prenatal screening for group B streptococcal infection: gaps in the evidence
Centre for Paediatric Epidemiology and Biostatistics, Institute of Child Health, University College London, 30 Guilford Street, London, WC1N 1EH, UK. E-mail: firstname.lastname@example.org
Weighing the benefits and harms of antibiotic prophylaxis
The 1990s saw the widespread adoption of antibiotic prophylaxis during labour in many western industrialized countries. By the end of the decade, 30% of women delivering in US maternity units involved in a large multicentre study received intravenous (IV) antibiotics during labour.1 In 24% of women, antibiotics were administered for vaginal carriage of group B streptococci, in order to reduce the risk of early onset neonatal group B streptococcal (GBS) disease. In the remaining 6%, antibiotics were given for risk factors related to preterm delivery, such as preterm prolonged rupture of membranes or preterm onset of labour.1 Although early onset neonatal GBS disease has undoubtedly decreased in association with the uptake of universal screening for maternal GBS colonization, there are concerns that antibiotic prophylaxis might select for more virulent, or antibiotic resistant pathogens.2 This question is as pertinent to the UK and other countries where universal screening for GBS is not yet offered, as it is to centres in the US where GBS screening was introduced a decade ago. Much of the evidence is non-randomized, and is drawn from time trends in the incidence of neonatal infection during a period when, in addition to GBS screening, many other aspects of obstetric and neonatal infection control have changed. What are the strengths, weaknesses, and gaps in the evidence on the effectiveness of antibiotic prophylaxis during labour for GBS infection?
Why do some neonates get early onset GBS disease?
GBS infection is an important cause of morbidity and mortality, particularly in the first 3 months of life. Early onset (within the first week of life), and late onset disease (7–89 days), are usually considered separately as risk factors as the clinical presentation and the prognosis differ with age.3–5 Furthermore, treatment strategies differ: early onset disease may only be treatable intrapartum owing to very rapid progression.
The causal pathway from maternal colonization to early onset GBS disease in the infant is summarized in Figure 1. The primary reservoir for GBS is the lower gastrointestinal tract. Maternal colonization is strongly associated with ethnic group (higher in African-American), young age, use of an intrauterine device or tampons, and diabetes. Although GBS can be sexually transmitted, colonization has not been associated with frequency of sexual activity or number of partners.3 The prevalence of maternal colonization is strongly affected by sampling and culture methods, but a review of studies with adequate techniques produced regional estimates which were highest in the US (26%), and lowest (12%) in India and Pakistan (19% in Asia, 19% in sub-Saharan Africa, 22% in the Middle East and North Africa, 14% in Central and South America).6 In the UK, there is a lack of contemporary data, but in the early 1980s studies reported prevalence rates between 15% and 28%.7–9
In colonized women, the risk of giving birth to a colonized neonate is directly related to the intensity of maternal colonization (inoculum size),3 which in the US is highest in African-American women.10 No other maternal or paripartum factors have been related to the risk of early neonatal colonization.3 The risk of early onset disease in colonized neonates is increased if there has been prolonged membrane rupture, maternal signs of infection, amnionitis, intrapartum fetal monitoring, or if the baby has a low birthweight or is born preterm.11,12
Changes in the incidence of early onset GBS disease
In the US, the risk of mother to child transmission and the incidence of early onset GBS disease were declining before screening for maternal GBS infection was introduced in the early 1990s. The incidence of early onset GBS disease fell from 2–3/1000 live births in the 1970s−1980s, to 1.4–1.8/1000 in 1990.13 Although the prevalence of maternal colonization remained stable at 20–25% during this period,3 it is not known whether the intensity of maternal colonization was diminishing, as might be expected with improved access to health care for high risk groups and improved management of urinary tract or other infections during pregnancy. Screening for GBS and intrapartum prophylactic antibiotics undoubtedly contributed to further decline in the incidence of early onset GBS disease during the 1990s, which stabilized at 0.2 to 0.5/1000 live births in the mid to late 1990s.14–25 In contrast, the incidence rate for late onset disease did not change over the 1990s remaining at approximately 0.4/1000 live births.13,26 In the UK, a study in Oxford from 1985 to 1996,27 and a recent national surveillance study in 2000–2001,28 reported incidence rates of early onset GBS disease (defined as septicaemia, pneumonia, or meningitis before 7 days of age) of 0.5/1000 live births, similar to the US, despite the lack of screening in the UK.
Case-fatality of GBS
The mortality due to early onset GBS disease has declined over time but remains higher than for late onset disease.3 A large US multicentre study conducted in 1993–1998 reported a mortality rate of 4.7% for early onset disease (defined as positive culture from blood/cerebrospinal fluid), and 2.8% for late onset disease.13 The recent UK national surveillance study reported a mortality rate of 10.6% for early onset disease.28 This may be high relative to the US study because of delayed diagnosis in the absence of screening, or because less severely affected babies with GBS disease were not notified. In the US study, 2% of term babies with early onset GBS disease died, compared with 21% of babies born preterm.13 Overall, 17% of early onset GBS disease occurred in preterm babies, but they accounted for 68% of the deaths. UK studies have found that 83–100% of deaths due to GBS were in preterm babies.27,29
Maternal screening to prevent early onset GBS disease: guidelines and cost effectiveness
During the early 1990s, a variety of screening algorithms were proposed and adopted in North America, Australasia, and Europe.2,21,23,30 These attempted to identify women at high risk of chorioamnionitis or neonatal GBS disease, or those culture-positive for GBS using vaginal and/or rectal swabs taken before or during admission in labour. The majority of women culture-positive for GBS do not have any other underlying obstetric risk, thus the two strategies identify very different women.1 Women identified by these methods would then be treated with intrapartum prophylactic antibiotics usually consisting of IV ampicillin or penicillin during labour, given 4 hourly until delivery. The 1996 Centers for Disease Control consensus guidelines led to a rapid increase in the adoption of screening by culture of vaginal or rectal swabs in North America, Australia, and parts of Europe.2,21,23,31 In contrast, a recent national survey found that screening for GBS has barely been adopted at all in the UK (personal communication, Sara Kenyon).
Several studies have analysed the cost effectiveness of screening. One analysis, from the early 1990s, found that the cost of introducing screening based on risk factors was greater than the calculated cost of the disease until the incidence of early onset neonatal GBS infection exceeded 0.6 per 1000 live births. The cost of introducing the screening by culture strategy was greater than the cost of disease until the incidence of infection exceeded 1.2 per 1000 live births.32 None of these studies considered the adverse effects of antibiotics on competing or resistant pathogens.
Does antibiotic prophylaxis reduce neonatal mortality or morbidity? A review of the evidence
To determine whether antibiotic prophylaxis for GBS colonization reduces neonatal morbidity and mortality due to any cause, we searched the Cochrane Library, Medline, and reference lists for systematic reviews and randomized controlled trials (RCT) of the effect of maternal antibiotic prophylaxis on neonatal bacteraemia or sepsis. A more limited search, based on reference lists from primary studies and review articles, sought time trend data on neonatal bacteraemia from 1988 onwards, in order to examine changes in cause-specific incidence rates during the implementation of GBS screening.
Prophylactic antibiotics for women colonized with GBS
One review of four RCT, and one further RCT were found.33,34 Four RCT compared antibiotic prophylaxis versus no treatment in women receiving no other treatment4,5,35,36 but only GBS sepsis or bacteraemia were considered outcomes, and effects on other infections were not evaluated systematically. The results showed a highly significant reduction in the risk of GBS sepsis or pneumonia (pooled odds ratio [OR] = 0.17; 95% CI: 0.07, 0.39) with none of the 368 babies born to treated mothers suffering GBS bacteraemia and only one suffering clinical signs of sepsis/pneumonia. As none of the studies were placebo controlled, detection bias due to non-blind assessment is possible.4,5,35,36
One further trial, not included in the review, measured all-cause bacteraemia in GBS colonized women. IV ampicillin and erythromycin were compared with placebo for women with preterm rupture of the membranes.33 Women who were colonized with GBS were randomized and analysed separately from women without GBS colonization and all of them received 5 days of oral ampicillin in addition to the IV treatment or placebo. Women who received IV ampicillin and erythromycin as well as oral ampicillin had a higher risk of neonatal sepsis than those who received only oral ampicillin: 23% (14/61) versus 13% (7/55) respectively (OR = 1.77; 95% CI: 0.77, 4.1). Although this difference was not significant, and no information was given on the type of organism, it raises the possibility that intensive antibiotic treatment may lead to selection of resistant organisms or to selective growth of competing pathogens.
All the RCT of antibiotic prophylaxis in GBS colonized women were based on women with other risk factors for early onset neonatal GBS infection, such as preterm labour. The treatment effect may not be the same in the majority of women without such risk factors who screen positive using culture methods.1
Prophylactic antibiotics for non-GBS indications
GBS colonization is just one of the indications for prophylactic antibiotics during or immediately prior to labour. Prophylactic antibiotics have been investigated most extensively for women in preterm labour or with preterm rupture of membranes, based on evidence that preterm labour and preterm rupture of the membranes may be related to sub-clinical chorioamnionitis.37 The primary aim of prophylaxis in these conditions is to delay delivery and prevent neonatal infection and associated complications. We reviewed the effect of antibiotic prophylaxis on all-cause bacteraemia.
Antibiotic prophylaxis for preterm labour with intact membranes
One review of antibiotic prophylaxis for preterm labour with intact membranes was found, based on nine RCT, all of reasonable quality, including the large ORACLE II trial (n = 6295).37,38 There was no evidence that prophylaxis increased or decreased the 2% risk of all-cause neonatal bacteraemia or sepsis (OR = 0.86; 95% CI: 0.64, 1.16). Consequently, any benefit for the subgroup of GBS-positive women, assumed to be between 15% to 25%, is likely to be small. Prophylaxis did not increase the time to delivery or improve other neonatal outcomes.38
Antibiotic prophylaxis for preterm rupture of the membranes
The review of prophylactic antibiotics for women with preterm rupture of membranes involved four RCT of good quality, including the ORACLE I trial involving 4826 women.39,40 The review found that antibiotics (erythromycin, amoxil-clavulanic acid, or both) reduced the risk of neonatal bacteraemia from 8.4% to 6.4% but the effect was not significant (OR = 0.50; 95% CI: 0.21, 1.19). However, when a broader, more subjective definition was used for infection, which did not require positive blood culture, 11 RCT could be included. ORACLE I was excluded but a large (614 women) US trial funded by the National Institute of Child Health and Development was included.33 All 11 trials were placebo controlled and 9 had adequate concealment of randomization. The results showed a significant reduction in the antibiotic treated group with a 6% risk difference (OR = 0.68; 95% CI: 0.53, 0.87). No studies found an increase in the risk of sepsis associated with antibiotic prophylaxis, apart from the subgroup of intensively treated GBS-positive women already mentioned.33 However, two studies found that amoxil-clavulanic acid significantly increased the risk of necrotizing enterocolitis in the neonate (treated 29/1236 versus untreated 6/1256; relative risk [RR] = 4.60, 95% CI: 1.98, 10.72).
The studies in this review did not report bacteraemia outcomes according to the type of organism, although such data are currently being analysed for the ORACLE trials. Such analyses will be important to determine whether there is evidence for an increase in gram negative bacteraemia in the treated group, and to what extent the 2% reduction in bacteraemia is due to elimination of GBS disease. Prophylactic antibiotics delayed delivery, reduced the need for respiratory support, and the risk of an abnormal cranial ultrasound scan in neonates. Consequently, screening this population for GBS is unlikely to be justified as all should be treated—unless further studies show that benefits are restricted to GBS colonized women. Should IV ampicillin or penicillin be given to GBS positive women at the onset of labour in addition to oral erythromycin at the time of membrane rupture? Head to head studies comparing erythromycin with penicillin are lacking, although erythromycin is recommended for penicillin allergic women.14 In addition, information is lacking on the effect of timing of treatment on mother to child transmission of GBS. However, intensive combination treatment may increase the risk of neonatal bacteraemia.33
Antibiotics for pre-labour rupture of membranes near term
One review of the effect of antibiotics on pre-labour rupture of the membranes near term41 found two poor quality studies. There was no evidence for a reduction or increase in neonatal bacteraemia, but given the quality and precision of these studies, considerable uncertainty remains.
Time trend studies
Changes in specific bacteraemia rates before and after the introduction of antibiotic prophylaxis for GBS colonization currently provide the best available information on whether the adoption of GBS screening has been associated with an increase in gram negative bacteraemia, particularly Escherichia coli. A subset of time trend studies that report early onset bacteraemia rates for GBS, E. coli (or gram negative bacteria), and total bacteraemia are shown in Figure 2. (Results for other studies that reported rates for GBS bacteraemia are available from the author.) Eight studies reported rates for E. coli (or gram negative bacteraemia), GBS, and all-cause bacteraemia. Seven studies were based on all deliveries, and one on very low birthweight neonates. All except one study were conducted in the US. The results show a stable or declining rate of GBS bacteraemia during the 1990s, which reached a plateau in the late 1990s.18,19,25,42 Rates for E. coli bacteraemia remained stable or increased, except in the Australian study which found a decline in E. coli bacteraemia.21 Where annual rates for E. coli bacteraemia are given18,19,25,38,42 trend analysis, using Poisson regression and adjusting for study as a random effect, found a significant linear increase over time (P = 0.001). In the study of very low birthweight infants, there was a significant decrease in the incidence of GBS bacteraemia and significant increase in bacteraemia due to E. coli (P <0.001).18,19,25,42 Several studies reported an increased prevalence of ampicillin resistance in E. coli isolates causing bacteraemia.18,19,25,42,43 Two studies19,25 reported mortality rates which confirmed established evidence that mortality due to gram negative bacteraemia is higher (9% for all babies,19 and 41% for very low birthweight babies25), compared with GBS bacteraemia (3% for all babies, and 26% for very low birthweight babies).
These findings need to be regarded with caution. Trends in cause-specific bacteraemia are based on very low numbers of events leading to chance fluctuations. In addition, reporting and publication bias are likely to be important in surveillance studies. Results of surveillance studies may not be published at all, and which specific bacteraemia rates are reported is likely to be influenced by whether trends have changed. Secondly, some studies group rates for several years, rather than reporting annual rates, which could allow selective grouping to exaggerate trends. Thirdly, multicentre studies18,19,21,25,38,42 may not detect real changes if results are combined from centres with differing timing of screening implementation. In addition, spurious trends can arise in multicentre studies if centres with high or low risk groups join the network during the surveillance period.21 Fifth, error may be introduced in the denominator, particularly in referral centres that combine inborn and out-born babies, as the latter will be at high risk of infection.
Causal inferences about the reduction in GBS bacteraemia and increase in E. coli bacteraemia should also be regarded with caution, as many factors could have changed in association with bacteraemia trends. Implementation and uptake of GBS screening may be a factor, but this is usually defined, if at all, in terms of policy which may not be an accurate reflection of practice, particularly for high risk groups.44 The surveillance system may become better organized and more sensitive over time, shifting from retrospective to prospective ascertainment, and definitions may change. The population of babies on neonatal intensive care units has changed, with more very low birthweight babies surviving for longer. At the same time, methods for monitoring and treating babies with suspected infection may have changed. There is evidence that the maternal flora is changing with an increased prevalence of ampicillin resistant E. coli,45 and in some neonatal intensive care units, there has been an increase in gram negative pathogens causing nosocomial infections.46
In conclusion, there is moderate evidence that antibiotic prophylaxis is beneficial for GBS colonized women at high risk of adverse neonatal outcomes. Evidence is lacking on the effectiveness of antibiotic prophylaxis for the vast majority of women identified by GBS screening who have no other risk factors, and any benefits are likely to be small. Screening for GBS would not change management for women with prolonged preterm rupture of the membranes, as all should be treated, but screening may be beneficial for women at moderate risk of adverse outcomes, for example in preterm labour with intact membranes, in whom universal prophylaxis is not justified. The decline in GBS bacteraemia has been associated with a small increase in early onset E. coli bacteraemia in neonates which is associated with a higher mortality. However, it is not known whether changes in E. coli rates are related to antibiotic prophylaxis, or changes in the maternal or hospital acquired flora. Time trend data from surveillance studies are critical for understanding the effect of this major change in the management of labour but are highly susceptible to reporting and publication biases. Systematic review of such data, combined with improved standardization of reporting, and web-based access to unreported and disaggregated data, could help to address these problems.
Website address for additional tables: http://www.ich.ucl.ac. uk/ich/html/academicunits/paed_epid/cebch/about.html
I am grateful to Mario Cortina-Borja for statistical analyses of time trend data and to Melissa Harden for performing the searches for this review.
1 Schrag SJ, Zell ER, Lynfield R et al. A population-based comparison of strategies to prevent early-onset group B streptococcal disease in neonates. N Engl J Med 2002;347:233–39.
2 Moore MR, Schrag SJ, Schuchat A. Effects of intrapartum antimicrobial prophylaxis for prevention of group-B-streptococcal disease on the incidence and ecology of early-onset neonatal sepsis. Lancet Infect Dis 2003;3:201–13.[CrossRef][Web of Science][Medline]
4 Tuppurainen N, Hallman M. Prevention of neonatal group B streptococcal disease: intrapartum detection and chemoprophylaxis of heavily colonized parturients. Obstet Gynecol 1989;73:583–87.[Web of Science][Medline]
5 Boyer KM, Gotoff SP. Prevention of early-onset neonatal group B streptococcal disease with selective intrapartum chemoprophylaxis. N Engl J Med 1986;314:1665–69.[Abstract]
9 Easmon CS, Hastings MJ, Deeley J, Bloxham B, Rivers RP, Marwood R. The effect of intrapartum chemoprophylaxis on the vertical transmission of group B streptococci. Br J Obstet Gynaecol 1983;90:633–35.[Web of Science][Medline]
10 Hickman ME, Rench MA, Ferrieri P, Baker CJ. Changing epidemiology of group B streptococcal colonization. Pediatrics 1999;104:203–09.
11 Schuchat A, Deaver-Robinson K, Plikaytis BD, Zangwill KM, Mohle-Boetani J, Wenger JD. Multistate case-control study of maternal risk factors for neonatal group B streptococcal disease. The Active Surveillance Study Group. Pediatr Infect Dis J 1994;13:623–29.[Web of Science][Medline]
12 Adair CE, Kowalsky L, Quon H et al. Risk factors for early-onset group B streptococcal disease in neonates: a population-based case-control study. Can Med Assoc J 2003;169:198–203.
13 Schrag SJ, Zywicki S, Farley MM et al. Group B streptococcal disease in the era of intrapartum antibiotic prophylaxis. N Engl J Med 2000;342:15–20.
14 Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of perinatal group B streptococcal disease. Revised guidelines from CDC. MMWR Recomm Rep 2002;51:1–22.[Medline]
16 Baltimore RS, Huie SM, Meek JI, Schuchat A, O'Brien KL. Early-onset neonatal sepsis in the era of group B streptococcal prevention. Pediatrics 2001;108:1094–98.
17 Chen KT, Tuomala RE, Cohen AP, Eichenwald EC, Lieberman E. No increase in rates of early-onset neonatal sepsis by non-group B Streptococcus or ampicillin-resistant organisms. Am J Obstet Gynecol 2001;185:854–58.[CrossRef][Web of Science][Medline]
19 Hyde TB, Hilger TM, Reingold A, Farley MM, O'Brien KL, Schuchat A. Trends in incidence and antimicrobial resistance of early-onset sepsis: population-based surveillance in San Francisco and Atlanta. Pediatrics 2002;110:690–95.
20 Wendel GD Jr, Leveno KJ, Sanchez PJ, Jackson GL, McIntire DD, Siegel JD. Prevention of neonatal group B streptococcal disease: A combined intrapartum and neonatal protocol. Am J Obstet Gynecol 2002;186:618–26.[CrossRef][Web of Science][Medline]
21 Isaacs D, Royle JA. Intrapartum antibiotics and early onset neonatal sepsis caused by group B Streptococcus and by other organisms in Australia. Australasian Study Group for Neonatal Infections. Pediatr Infect Dis J 1999;18:524–28.[CrossRef][Web of Science][Medline]
22 Cordero L, Sananes M, Ayers LW. Bloodstream infections in a neonatal intensive-care unit: 12 years‘ experience with an antibiotic control program. Infect Control Hosp Epidemiol 1999;20:242–46.[CrossRef][Web of Science][Medline]
23 Andreu A, Sanfeliu I, Vinas L et al. [Decreasing incidence of perinatal group B streptococcal disease (Barcelona 1994–2002). Relation with hospital prevention policies]. Enferm Infecc Microbiol Clin 2003;21:174–79.[CrossRef][Web of Science][Medline]
24 Davies HD, Raj S, Adair C, Robinson J, McGeer A. Population-based active surveillance for neonatal group B streptococcal infections in Alberta, Canada: implications for vaccine formulation. Pediatr Infect Dis J 2001;20:879–84.[CrossRef][Web of Science][Medline]
25 Stoll BJ, Hansen N, Fanaroff AA et al. Changes in pathogens causing early-onset sepsis in very-low-birth-weight infants. N Engl J Med 2002;347:240–47.
26 Zangwill KM, Schuchat A, Wenger JD. Group B streptococcal disease in the United States, 1990: report from a multistate active surveillance system. MMWR CDC Surveill Summ 1992;41:25–32.[Medline]
27 Moses LM, Heath PT, Wilkinson AR, Jeffery HE, Isaacs D. Early onset group B streptococcal neonatal infection in Oxford 1985–96. Arch Dis Child Fetal Neonatal Ed 1998;79:F148–49.
29 Oddie S, Embleton ND. Risk factors for early onset neonatal group B streptococcal sepsis: case-control study. BMJ 2002;325:308.
31 Davies HD, Adair CE, Schuchat A, Low DE, Sauve RS, McGeer A. Physicians’ prevention practices and incidence of neonatal group B streptococcal disease in 2 Canadian regions. Can Med Assoc J 2001;164:479–85.
32 Mohle-Boetani JC, Schuchat A, Plikaytis BD, Smith JD, Broome CV. Comparison of prevention strategies for neonatal group B streptococcal infection. A population-based economic analysis. JAMA 1993;270:1442–48.
33 Mercer BM, Miodovnik M, Thurnau GR et al. Antibiotic therapy for reduction of infant morbidity after preterm premature rupture of the membranes. A randomized controlled trial. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. JAMA 1997;278:989–95.
36 Morales WJ, Lim DV, Walsh AF. Prevention of neonatal group B streptococcal sepsis by the use of a rapid screening test and selective intrapartum chemoprophylaxis. Am J Obstet Gynecol 1986;155:979–83.[Web of Science][Medline]
37 Kenyon SL, Taylor DJ, Tarnow-Mordi W. Broad-spectrum antibiotics for spontaneous preterm labour: the ORACLE II randomised trial. ORACLE Collaborative Group. Lancet 2001;357:989–94.[CrossRef][Web of Science][Medline]
40 Kenyon SL, Taylor DJ, Tarnow-Mordi W. Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLE I randomised trial. ORACLE Collaborative Group. Lancet 2001;357:979–88.[CrossRef][Web of Science][Medline]
42 Main EK, Slagle T. Prevention of early-onset invasive neonatal group B streptococcal disease in a private hospital setting: the superiority of culture-based protocols. Am J Obstet Gynecol 2000;182:1344–54.[CrossRef][Web of Science][Medline]
43 Joseph TA, Pyati SP, Jacobs N. Neonatal early-onset Escherichia coli disease. The effect of intrapartum ampicillin. Arch Pediatr Adolesc Med 1998;152:35–40.
45 Gupta K, Scholes D, Stamm WE. Increasing prevalence of antimicrobial resistance among uropathogens causing acute uncomplicated cystitis in women. JAMA 1999;281:736–38.