PEDIATRICS Vol. 103 No. 6 June 1999, p. e78

ELECTRONIC ARTICLE:
Antimicrobial Prevention of Early-onset Group B Streptococcal Sepsis: Estimates of Risk Reduction Based on a Critical Literature Review

William E. Benitz, MD^*, Jeffrey B. Gould, MD^§, and Maurice L. Druzin, MD

From the Departments of ^* Pediatrics and  Gynecology and Obstetrics, Stanford University School of Medicine, Stanford, California; and ^§ Maternal and Child Health Program, School of Public Health, University of California, Berkeley, California.

	ABSTRACT

Top Abstract Methods Results Discussion References

Objective.  To identify interventions that reduce the attack rate for early-onset group B streptococcal (GBS) sepsis in neonates.

Study Design.  Literature review and reanalysis of published data.

Results.  The rate of early-onset GBS sepsis in high-risk neonates can be reduced by administration of antibiotics. Treatment during pregnancy (antepartum prophylaxis) fails to reduce maternal GBS colonization at delivery. With the administration of intravenous ampicillin, the risk of early-onset infection in infants born to women with preterm premature rupture of membranes is reduced by 56% and the risk of GBS infection is reduced by 36%; addition of gentamicin may increase the efficacy of ampicillin. Treatment of women with chorioamnionitis with ampicillin and gentamicin during labor reduces the likelihood of neonatal sepsis by 82% and reduces the likelihood of GBS infection by 86%. Universal administration of penicillin to neonates shortly after birth (postpartum prophylaxis) reduces the early-onset GBS attack rate by 68% but is associated with a 40% increase in overall mortality and therefore is contraindicated. Intrapartum prophylaxis, alone or combined with postnatal prophylaxis for the infants, reduces the early-onset GBS attack rate by 80% or 95%, respectively.

Conclusions.  Women with chorioamnionitis or preterm premature rupture of membranes and their infants should be treated with intravenous ampicillin and gentamicin. Intrapartum antimicrobial prophylaxis may be appropriate for other women whose infants are at increased but less extreme risk, and supplemental postpartum prophylaxis may be indicated for some of their infants. Selection of appropriate candidates and prophylaxis strategies requires careful consideration of costs and benefits for each patient. group B streptococcus, neonatal sepsis, early-onset sepsis, prevention, prophylaxis.

Prevention of early-onset neonatal group B streptococcal (EOGBS) sepsis requires the identification of infants who are at high risk and the use of effective preventive intervention. The American Academy of Pediatrics (AAP),^1,2 American College of Obstetricians and Gynecologists (ACOG),^3,4 and Centers for Disease Control and Prevention (CDC)⁵ agree that this intervention should be intrapartum antibiotic treatment. The commentaries that accompany those recommendations and several recent cost-benefit or efficacy analyses implicitly^1-7 or explicitly^8-11 assume that intrapartum antimicrobial prophylaxis is 100% effective, but reports of at least 50 cases of EOGBS disease in infants whose mothers received intrapartum prophylaxis^12-15 demonstrate that this cannot be correct. The decision analysis cited by the CDC recommendations¹⁶ estimated that intrapartum prophylaxis reduces the attack rate by 93.75%, from 16 to 1/1000 live births. A critical review^17,18 has questioned whether intrapartum prophylaxis is effective at all. Reviews that support efficacy have not distinguished consistently treatment of mothers during labor alone from treatment of both mothers during labor and infants after delivery.^5,19,20 Amstey and Gibbs,²¹ as well as the CDC,⁵ have advocated the use of penicillin instead of ampicillin for intrapartum prophylaxis,²¹ but this regimen has not been evaluated extensively. Because of uncertainty about the efficacy of intrapartum prophylaxis, Wedgwood et al²² and Siegel and Cushion²³ have advocated reconsideration of universal postpartum prophylaxis, and Gotoff and Boyer²⁴ have proposed a strategy based on a combination of intrapartum maternal prophylaxis and postpartum neonatal prophylaxis. These proposals indicate that the AAP, ACOG, and CDC recommendations have not been universally accepted. To better understand the scientific basis for proposed prevention strategies, we have reevaluated published trials of methods for prevention of neonatal group B streptococcus (GBS) infection to establish estimates for the effectiveness of each prophylaxis regimen in reducing the rate of EOGBS infection.

	METHODS

Top Abstract Methods Results Discussion References

Trials of regimens for prevention of EOGBS infection were identified using a MEDLINE search, as well as reference lists for those articles and recent reviews.^25-29 Studies were considered relevant only if enrollment criteria were explicit and applied equally to both treatment and nontreatment (control) subjects, and if GBS disease was observed in at least 1 subject. Neonatal sepsis or GBS sepsis was diagnosed only if pathogenic bacteria or GBS, respectively, were recovered from cultures of blood, cerebrospinal fluid, or a tracheal aspirate. Infants who had only clinical findings consistent with sepsis and/or GBS antigen in their urine were not considered to have sepsis. Only early-onset cases (diagnosed within the first 7 days after birth) were included. Because spontaneous waxing and waning of GBS attack rates make sequential designs inappropriate for studies of this disease,^30,31 studies that relied on historical controls were evaluated primarily for evidence of adverse effects. Effects on neonatal mortality were also evaluated if possible. Because of heterogeneity in study designs, differences in therapeutic regimens, inconsistent criteria for diagnosis of sepsis, and implementation flaws, formal metaanalysis of treatment and prevention strategies is not practical.^17,18,32 We included all studies that meet these relevancy and diagnostic criteria, but doing so may overestimate treatment efficacy¹⁷ because of ascertainment, diagnostic, or reporting biases.^33,34

Differences in outcomes were evaluated using the G test of independence, an alternative to the ² or Fisher exact tests preferred for contingency tables in which marginal totals are not fixed by experimental design.³⁵ CIs for ORs were calculated using the method of Cornfield (see Armitage and Berry³⁶), typical ORs and confidence limits for data pooled from multiple studies were calculated by the method of Mantel and Haenszel after testing for heterogeneity by the method of Woolf (see reference³⁶), and relative risks were calculated by the method of Rothman and Boice (see reference³⁷). If no events were observed, we calculated one-tailed confidence limits for ORs using the method of Cornfield;³⁶ we calculated OR point estimates by adding .5 to each entry in the contingency table,³⁷ and we calculated CIs for attack rates using the normal³⁸ or binomial distributions.³⁹ Probabilities of type II errors were calculated as described by Sokol and Rohlf.⁴⁰ All P values are two-tailed.

	RESULTS

Top Abstract Methods Results Discussion References

Measures intended to prevent early-onset neonatal sepsis fall into two groups: treatment with broad-spectrum antibiotics with the intent of preventing any early-onset infection and prophylaxis using monotherapy specifically to prevent GBS transmission. The first group, in which antepartum or intrapartum antibiotic use is the primary intervention, has been categorized according to the clinical indication for treatment (preterm premature rupture of membranes or chorioamnionitis). The second group includes antepartum, intrapartum, postpartum, and combined intrapartum and postpartum prophylaxis regimens.

Intrapartum Treatment for Preterm Premature Rupture of Membranes

Thirteen studies describing the effects of intrapartum intravenous antibiotics and two studies describing the effects of oral antibiotic therapy on the prevalence of neonatal sepsis and/or GBS sepsis after preterm (<37 weeks gestation) premature (before the onset of labor) rupture of membranes were identified. Neither study of oral therapy demonstrated any impact on neonatal infection.^41,42 Three studies of intravenous therapy that did not report the number of infants with culture-proven bacteremia^43-45 were excluded. The remaining studies^46-55 are summarized in Table 1. Two individual studies^46,47 (one a randomized prospective trial⁴⁷), as well the pooled data, indicate that antibiotic treatment of women with preterm premature rupture of membranes before delivery reduced the rate of culture-proven sepsis among their infants. In several of these studies, GBS-colonized women were excluded^48,51,52 or were given intrapartum prophylaxis.^53,55 In three trials in which intervention was not contingent on GBS colonization status,^46,50,54 there were no cases of early-onset GBS infection in 104 infants whose mothers received intrapartum antibiotics and 7 cases in 152 infants born to women who were not given antibiotics (P > .05). When combined with a recent study from which GBS-colonized women were excluded,⁵⁵ these studies imply that administration of antibiotics to women with preterm premature rupture of membranes reduces the risk of early-onset sepsis by 56% and reduces the risk of GBS sepsis by 36% (Table 1). Because the latter study may overlook benefits in GBS-colonized patients, benefits of intrapartum treatment of women with preterm premature rupture of membranes may be underestimated. The observation that intrapartum ampicillin reduced the rate of early-onset clinical sepsis in infants born to GBS-colonized women but not in infants born to women who were not GBS-colonized⁴³ suggests that broad-spectrum antimicrobial coverage may be required for these patients.

Intrapartum Treatment of Chorioamnionitis

The effects of the intrapartum treatment of women with chorioamnionitis on the risk of neonatal infection are addressed by several studies (Table 2). In a prospective, nonrandomized study, Sperling et al⁹⁸ evaluated rates of early-onset sepsis in infants born to women with chorioamnionitis that was diagnosed based on intrapartum fever >100°F (37.8°C) and at least two additional clinical findings: maternal tachycardia, uterine tenderness, fetal tachycardia, purulent amniotic fluid, or leukocytosis. Bacteremia was less common (P < .001) in infants whose mothers received penicillin and gentamicin before delivery (Table 2). All infants were treated with ampicillin and gentamicin until infection was excluded. Treatment was not random, the antepartum treatment group included fewer low birth weight infants and had a longer time from diagnosis of chorioamnionitis to delivery, and there were no significant effects on neonatal sepsis in infants with birth weights <2000 g nor any difference in EOGBS rates between early and delayed treatment groups; therefore, it is not clear that the observed difference in sepsis rates was attributable to treatment. Gilstrap et al retrospectively examined outcomes of infants born to women with chorioamnionitis diagnosed using similar criteria.⁵⁶ Infants born to women who had been given antibiotics before delivery were less likely to have GBS bacteremia, but there was no significant effect on the total bacteremia rate (Table 2). Gibbs et al prospectively randomized women with chorioamnionitis, diagnosed as above, to treatment during labor or only after cord clamping.⁵⁷ Infants in both groups were treated with ampicillin and gentamicin until infection was excluded by negative blood cultures at 72 hours or for 10 days in those with confirmed sepsis. This series was too small to detect a significant effect on the rate of EOGBS infection, but treatment before delivery was associated with a significant reduction in neonatal sepsis in general (Table 2), providing the most compelling evidence that intrapartum treatment of such women is beneficial to their infants. In the above studies, treatment consisted of penicillin or ampicillin in combination with gentamicin. Treatment with a penicillin alone cannot be recommended, because 4 cases of infection attributable to ampicillin-resistant Enterobacteriaceae after prophylactic treatment of women with chorioamnionitis using ampicillin or amoxicillin alone have been reported.⁵⁸ Considered collectively, these studies indicate that intrapartum treatment of women with chorioamnionitis, using ampicillin and gentamicin, reduces the risk of EOGBS sepsis by 86% and reduces the risk of early-onset sepsis in general by 82% (Table 2).

Antepartum Prophylaxis

As the relationship between maternal GBS colonization and early-onset neonatal infection became apparent in the early 1970s, the potential for prevention of neonatal disease by identification and treatment of GBS carriers during pregnancy was recognized quickly.^59,60 In promising but uncontrolled observations, Franciosi et al described eradication of GBS colonization in 13 of 14 colonized women and their husbands after treatment with intramuscular benzathine penicillin G.⁶⁰ In a randomized controlled trial of treatment of GBS-colonized women with oral ampicillin for 1 week during pregnancy, Hall et al⁶¹ demonstrated reduced maternal colonization 3 weeks after treatment, but neither maternal (Table 3) nor neonatal colonization rates were reduced at delivery. Treatment of GBS-colonized women and their husbands with oral penicillin for 14 days during the third trimester also did not reduce maternal colonization rates at delivery.⁶² In a unique trial in which treatment of GBS-colonized women and their husbands was continuous from 38 weeks gestation until delivery, Merenstein et al⁶³ described a marked reduction in maternal colonization at delivery, but they recommended additional studies before these observations were applied to clinical practice and emphasized that this approach did not address the high-risk population of infants who deliver before screening at 38 weeks gestation. Lewin and Amstey⁶⁴ reported that treatment of colonized women and their husbands with benzathine penicillin/procaine penicillin injection in the third trimester was effective in reducing the colonization rate at delivery. Weeks et al⁶⁵ have recently reported a 72% reduction in GBS colonization rate and eradication of heavy GBS colonization at delivery after administration of benzathine penicillin G to women with GBS colonization at various times between 16 and 37 weeks gestation, but follow-up was unavailable for nearly a third of their subjects and no control group was studied. Now, it is generally accepted that it is extremely difficult or impossible to eradicate GBS from mucosal surfaces,⁶⁶ especially from the lower intestinal tract,⁶⁷ and that vaginal recolonization after a course of antibiotic therapy is exceedingly common, as initially reported by Hall et al.⁶¹ No studies have demonstrated an effect on neonatal infections.

Intrapartum Prophylaxis

A total of 17 reports describing intrapartum prophylaxis were identified. Of the reports, 3, in which both intrapartum maternal and postpartum neonatal prophylaxis were administered,^68-70 are considered separately. Of the reports, 9 were excluded: 4 had no control subjects^71,72 or only historical controls,^73,74 2 used clinical⁷⁵ or unspecified⁷⁶ criteria to diagnose sepsis, 1 did not report the number of recipients of prophylaxis,⁷⁷ and 2 were duplicate reports.^78,79 The 5 remaining studies^79-83 are summarized in Table 4. In the study by Allardice et al,⁸⁰ 28 women who were prospectively identified as GBS carriers by vaginal culture at 28 to 34 weeks gestation were confirmed to have GBS colonization at delivery and were given intrapartum ampicillin prophylaxis; none delivered infants with invasive GBS disease. Among 136 women who had vaginal GBS colonization at delivery but who did not receive intrapartum prophylaxis because they had not been identified previously as GBS carriers, 9 had infants with invasive GBS disease (P = .08). Morales et al⁷⁵ randomly assigned women with positive coagglutination screening tests for GBS colonization to treatment with intravenous ampicillin or no intervention and reported elimination of superficial colonization and invasive GBS disease in infants born to treated mothers (P = .13). Among infants born to women with heavy GBS colonization, ascertained using a rapid latex agglutination test on admission with labor, Tuppurainen and Hallman reported significantly fewer cases of clinical sepsis in those infants who were assigned randomly to treatment with intravenous penicillin.⁸² There was 1 case of culture-proven GBS infection in 88 infants born to treated women and 5 cases of culture-proven GBS infection among 111 infants born to untreated women (P = .16). Among 157 women with positive latex agglutination tests who could not be randomized because delivery occurred before screening test results were available, only 1 woman had an infant with GBS bacteremia; therefore, the relatively high attack rate in the group randomized to no treatment may have been anomalous. Matorrás et al⁷⁹ observed no invasive GBS disease in 60 infants born to GBS-colonized women assigned randomly to treatment with intrapartum intravenous ampicillin and 3 cases in 65 infants whose mothers were randomized to no treatment (95% CI for OR 0-.90, but P = .06 by G test). Pylipow et al⁸³ documented GBS bacteremia in 5 of 54 infants whose mothers were GBS colonized, had at least one intrapartum risk factor (fever or amnionitis, preterm labor, or rupture of membranes for >6 hours), and did not receive intrapartum antibiotics. Only 1 infant born to 70 such mothers who received intrapartum ampicillin had a blood culture positive for GBS. This difference is significant (P = .05), but treatment assignment was not random. It is notable that nontreatment subjects were enrolled preferentially in the last 6 months of this 21-month study, which was conducted during a period when attack rates were not stable at that hospital.⁸³ Differences between attack rates in treatment and nontreatment groups are statistically significant only in the last of these studies,⁸³ which was not a randomized trial. Studies in which treatment was randomized^79,81,82 did demonstrate a significant reduction in EOGBS (OR: .11; 95% CI: .02-.62). Studies of intrapartum prophylaxis have been criticized because of nonrandom assignment of penicillin-allergic subjects to the control group⁸¹ (although no biological mechanism by which maternal penicillin allergy could increase the risk of neonatal infection is apparent⁸⁴), failure to provide comparison data on the study groups or to describe fully the randomization procedure,^79,82 omission of statistical methods,⁸² and multiple outcome analyses.⁸² Although there are many potential sources of error, including bias and selective publication as well as those noted above, these studies provide reasonable evidence that intrapartum antibiotic prophylaxis reduces the risk of EOGBS sepsis by 80%.

The timing or duration of intrapartum prophylaxis is an additional concern. The 1992 AAP guidelines suggested beginning intrapartum prophylaxis at least 4 hours before delivery,¹ and the CDC⁵ recommends diagnostic evaluation for infants whose mothers were treated with antibiotics for 4 hours before delivery. The revised AAP guidelines indicate that treatment is not necessary for infants whose mothers received at least 2 doses of prophylactic antibiotics before delivery.² We are unable to identify data to support these recommendations. The prevalence of GBS colonization of infants born to colonized mothers decreases with increasing duration of intrapartum therapy, from nearly 50% without therapy or within 1 hour of initiation of therapy to 28% after 1 to 2 hours, 2.9% after 2 to 4 hours, and 1.2% after >4 hours (P < .005).⁷⁶ Treatment for <1 hour did not affect the rate of infant colonization, and colonization rates after >4 hours of treatment were not different statistically from those after 2 to 4 hours of treatment. In that study, there were no cases of early-onset GBS disease in infants whose mothers received antibiotics at any time before delivery and only 1 case among 253 infants whose mothers did not receive antepartum antibiotics; therefore, no relationship between the risk of disease and duration of therapy can be established. Pylipow et al⁸³ noted that the 2 infants in their survey who developed sepsis despite intrapartum prophylaxis had mothers who received antibiotics <2.5 hours before delivery. They also described a reduction in neonatal colonization after intrapartum prophylaxis (P < .05), and reported that colonization (6.9% vs 9%) and neonatal sepsis (4.7% vs 7%) rates were not significantly higher in infants whose mothers received only 1 intrapartum antibiotic dose than in those whose mothers received 2 doses (P > .5). Others have reported prophylaxis failures after antibiotic administration for >4 hours before delivery. Weisman et al reported prophylaxis failures in preterm infants even after antibiotic administration for up to 48 hours before delivery, as well as in term infants who received antibiotics for <4 hours before delivery.^12,14 In a series of intrapartum prophylaxis failures reported by Ascher et al,¹³ 6 of 16 occurred in term infants treated for 4 to 7 hours before delivery. No studies report either the number of women who received intrapartum treatment for longer intervals or the number of woman who were treated for a short time and whose infants were not infected. The 1992 AAP recommendation is supported by the rationale that >4 hours may be required to allow achievement of optimal ampicillin levels in the amniotic fluid and placental circulation, presumably producing a reduction in amniotic fluid bacterial counts.¹ Although bactericidal levels of ampicillin are achieved in fetal serum and amniotic fluid as soon as 5 minutes after maternal ampicillin administration,⁸⁵ there are no data relating these ampicillin levels to neonatal outcomes, and Silver et al found "no significant relationship between quantitative amniotic fluid GBS cultures and neonatal outcome."⁸⁶ Because the available data do not permit stratification of residual risk based on either the number of doses given or the duration of treatment before delivery, we conclude that it is impossible to determine the minimum duration of intrapartum treatment required to achieve detectable or optimal efficacy.

Postpartum Prophylaxis

The serendipitous observation that GBS infections were rare at centers in which newborn infants were given penicillin injections routinely to prevent gonococcal ophthalmia suggested that this intervention might also prevent EOGBS disease.^87-89 Several studies tested this hypothesis (Table 5). Lloyd et al described a significant reduction in the GBS attack rate after institution of routine penicillin prophylaxis for premature (<35 weeks) or low birth weight (<2500 g) infants.⁹⁰ In a prospective trial, Siegel et al demonstrated that postpartum penicillin prophylaxis significantly reduced the prevalence of EOGBS disease.^91,92 Patel et al reported a lower rate of EOGBS disease during an 18-month period when intramuscular penicillin prophylaxis at birth was routine, compared with the 18 months after the discontinuation of that practice.⁹³ Siegel and Cushion²³ and Wedgwood et al²² have recently advocated reconsideration of universal penicillin prophylaxis at birth. Siegel and Cushion²³ presented comprehensive data on GBS attack and mortality rates for infants delivered at Parkland Memorial Hospital between the start of their previously reported clinical trial on December 4, 1977 and the end of 1994.²³ Data gathered retrospectively for the period before the clinical trial were not considered in this analysis. Data from the clinical trial (December 1977 through May 1981), for infants who received penicillin prophylaxis after birth in June 1981 through October 1986, and for infants delivered in November 1986 through December 1994, who did not receive penicillin prophylaxis, are summarized in Table 6. The GBS attack rate was significantly lower during the penicillin prophylaxis periods both during and after the prospective trial (Table 5). Attack rates for invasive disease attributable to penicillin-sensitive bacteria were lower in penicillin-treated infants (Table 6). Mortality associated with EOGBS disease or with penicillin-sensitive organisms during the first month of life was not reduced significantly. These authors were "unable to demonstrate a significant increase in mortality associated with penicillin-resistant pathogens in infants who received single-dose penicillin prophylaxis,"²³ because their two-tailed 95% CIs for the ORs for mortality associated with penicillin-resistant organisms included 1 for both the 1977 through 1981 and 1982 through 1994 cohorts. However, CIs calculated by the method of Cornfield³⁶ do not include 1; the OR for the pooled data from both periods is 1.8 (95% CI: 1.2-2.8; Table 6), so the observed increase in mortality attributable to penicillin-resistant pathogens after penicillin prophylaxis is significant. In addition, a one-tailed test, which is more appropriate for evaluating the safety of an intervention, shows that the total mortality attributable to bacterial disease was also significantly greater in penicillin-treated infants (pooled OR: 1.4; lower 95% confidence limit: 1.05). Pyati et al found that postpartum penicillin prophylaxis was ineffective for infants with birth weights 2000 g⁹⁴ and suggested that fetal infection is established before delivery, because symptoms are present soon after birth in most infants with EOGBS infection.⁹⁵ Universal postpartum prophylaxis fails to improve overall outcomes, is associated with a higher mortality rate, and, therefore, should be contraindicated.

Combined Intrapartum and Postpartum Prophylaxis

Four reports of combined intrapartum and postpartum prophylaxis were identified.^68-70,96 In the only prospective, randomized trial of GBS prophylaxis, Boyer et al demonstrated a significant reduction in the GBS attack rate for infants of GBS-colonized women with either preterm labor (<37 weeks gestation) or prolonged rupture of membranes (>12 hours) if ampicillin prophylaxis was given to the women during labor and to the babies after birth⁶⁹ (Table 7). This trial has been criticized, because it did not use an intention-to-treat analysis and had a post-randomization dropout rate of 11%, attributable to exclusion of mothers who developed fever (6 control subjects and 7 treatment subjects, who received int-rapartum ampicillin; none of their babies had GBS infection) or for whom there were randomization errors or missing data (3 control subjects and 4 treatment subjects).³² Inclusion of the 13 subjects who were excluded because of fever in an intention-to-treat analysis does not alter the conclusion that this prophylaxis regimen was effective; the resulting OR is .084 (95% CI: 0-.51; P = .011). Attrition of 7 of 180 initially enrolled subjects (4%) is well below the recommended 10% threshold for tolerance of withdrawals from short-term trials.⁹⁷ It is possible that intrapartum prophylaxis merely suppressed recovery of GBS in blood cultures obtained immediately after birth before postpartum prophylaxis was administered⁶⁹, but benefits of treatment were not limited to effects on laboratory diagnosis of bacteremia (Table 7). Ampicillin-treated infants were also significantly less likely to have GBS colonization, to be heavily colonized, or to exhibit clinical signs of pneumonia. Although fewer infants died from infection, this was not significant statistically. Because blood cultures were obtained before ampicillin was given to the neonates, postpartum prophylaxis was not required to suppress immediate postpartum bacteremia, but the number of babies who might have developed bacteremia subsequently without supplemental prophylaxis is a matter of speculation. There were no cases of EOGBS disease among 320 nonrandomized subjects⁶⁹ (many included in previous reports^68,96) who received combined intrapartum and neonatal ampicillin prophylaxis (Table 8; upper 95% confidence limit for the OR: .81 by the binomial distribution).

Garland and Fliegner compared neonatal outcomes for 30 197 public clinic patients who had screening cultures at 32 weeks gestation to those for 26 915 private clinic patients who were not screened for GBS colonization.⁷⁰ GBS-colonized public patients were given penicillin during labor, and their infants received a single injection of penicillin after birth.⁷⁰ There were 25 cases of EOGBS disease in infants of untreated asymptomatic women (5 cases in public patients who were GBS-colonized women but did not receive prophylaxis attributable to protocol deviations and 20 cases in private patients). There were no EOGBS cases in infants born to GBS-colonized women who received prophylaxis. The number of women who were GBS-colonized or who received intrapartum treatment were not specified, but the OR (Table 8) can be estimated by assuming that rates of delivery at 32 weeks gestation (~2%) and GBS colonization (~14%; S. Garland, personal communication) were the same in treated and untreated women. Remarkably, neither the OR nor its confidence limits are sensitive to assumed values for these rates. From the binomial distribution,³⁹ the upper 95% confidence limit for the OR for GBS cases among infants born to women who received intrapartum prophylaxis, compared with those who did not, is .20.

The data of Boyer et al⁶⁹ and Garland and Fliegner⁷⁰ strongly support the inference that combined intrapartum and postpartum prophylaxis reduces the risk of early-onset GBS disease in neonates by 95% (Table 8). Because these studies have fewer methodologic weaknesses than those studies that used intrapartum prophylaxis alone, the efficacy of combined intrapartum and postpartum prophylaxis may be both greater and better established than that of intrapartum prophylaxis alone.

	DISCUSSION

Top Abstract Methods Results Discussion References

To better understand the scientific basis for the recommendations of the CDC,⁵ ACOG,⁴ and AAP,² we have systematically evaluated all studies that met criteria for clinical relevancy and diagnostic specificity. There are few data from randomized controlled trials, and no studies directly address relative efficacies of different regimens. Because of heterogeneous study designs, different therapeutic regimens, inconsistent diagnostic criteria, and implementation flaws, reports of strategies for prevention of EOGBS disease are not suitable for formal metaanalysis.^17,18,32 Despite these shortcomings, we do not believe that the available data are compromised beyond utility, and we have attempted to use as much of the data as feasible to quantitate the extent to which treatment or prophylaxis regimens reduce attack rates for early-onset neonatal sepsis in general and for EOGBS sepsis in particular. This inclusive approach increases the potential for ascertainment, diagnostic, and reporting biases^33,34 and, consequently, may overestimate efficacy.¹⁷ Ascertainment error is of particular concern for retrospective studies, of which only three are considered in these analyses.^46,56,93 Errors attributable to ascertainment bias are likely to be few for the studies in which blood cultures were performed uniformly on all enrolled infants^{48,57,69,94,98}. In one study, blood cultures were obtained routinely from infants who received postpartum penicillin prophylaxis, but not from infants reported as historical control subjects,⁹⁰ creating the potential for overascertainment in the penicillin-treated group. The potential for ascertainment errors in prospective studies in which blood cultures were obtained only if the infant was judged to be clinically ill^{47,53,79,81,83}, at the discretion of the attending neonatologist,⁴⁹ or according to unspecified criteria^{23,51,52,63,70,80,82,91,92} depends on the duration of observation and the proportion of infants for whom complete outcome data were available. All but two of these studies were prospective trials with apparently complete data on early-onset neonatal bacteremia; the other two studies were hospital-based studies that included data for the entire birth cohort. Because most infants with EOGBS sepsis are symptomatic within the first 24 hours,⁹⁹ few are likely to have been discharged before clinical signs prompted evaluation with blood cultures. To minimize the effects of preferential diagnosis of clinical sepsis in untreated infants, the diagnosis of sepsis was restricted to infants from whom pathogenic bacteria were recovered in nonpermissive cultures. Positive culture results can be ascertained reliably, ie, by review of laboratory summaries,⁹¹ so there should have been few failures to identify bacteremic infants with EOGBS infection. In making objective determinations of bacterial growth, laboratory personnel are unlikely to be influenced by knowledge of antimicrobial interventions, but clinicians are likely to use such knowledge to reach or exclude a diagnosis of sepsis, which is often based on subjective and nonspecific clinical observations. This criterion may lead to overestimation of the benefits of treatment because of the suppression of bacterial growth in cultures without concomitant reduction of other morbidities of bacterial infection. There are no reliable strategies for recognition of underreporting of negative trials. The effects of interventions on neonatal mortality secondary to early-onset sepsis are described in only a few studies including those of treatment of chorioamnionitis,^56,57,98 preterm premature rupture of membranes,^4749-52 postpartum prophylaxis,^23,91,92 and combined intrapartum and postpartum prophylaxis.⁶⁹ Only the very large dataset of Siegel et al included enough subjects to detect an effect on mortality rate,²³ and only one study gathered empirical data on infection-related morbidity and cost.¹⁰⁰ Future studies, including postimplementation surveillance of GBS prevention strategies, should report complete outcome data, particularly relating to neonatal mortality and hospitalization experiences.

This review identified two types of interventions that may reduce the rate of EOGBS. The first type consists of interventions initiated in response to relatively low prevalence conditions associated with a very high risk of neonatal GBS infection, including preterm premature rupture of membranes (particularly in GBS-colonized women), chorioamnionitis, GBS bacteriuria during pregnancy, and invasive GBS disease in a twin or sibling.¹⁰¹ These interventions could be considered treatment for apparent disease. The second group includes interventions provoked by more common conditions associated with a moderately increased risk of neonatal GBS infection. These interventions are more appropriately considered to be prophylaxis, because few infants with these conditions would become infected even without intervention. Such conditions include maternal vaginal colonization with GBS at delivery, prematurity and low birth weight, prolonged rupture of membranes, and intrapartum maternal fever.¹⁰¹ We considered these two categories of interventions separately, because infants in the highest risk category may require aggressive treatment rather than prophylactic measures.

We found that interventions in the first category are effective in reducing the risk of sepsis in general and GBS sepsis in particular. In a recent metaanalysis of strategies for management of preterm premature rupture of membranes, Egarter et al¹⁰² also concluded that antibiotic therapy is effective in reducing the risk of clinical sepsis defined as fever, leukocytosis, and C-reactive protein elevation (a positive blood culture was not considered sufficient and apparently was not required). Most of the reported regimens continued treatment either until delivery or until cultures excluded GBS colonization. We recommend treating women with preterm premature rupture of membranes with ampicillin and gentamicin beginning at diagnosis and continuing until delivery occurs or at least until maternal vaginal cultures exclude GBS colonization. Antibiotic treatment of women with chorioamnionitis reduces the risk of early-onset sepsis in general as well as GBS sepsis in particular, so these patients also should be treated with intravenous ampicillin and gentamicin, beginning at diagnosis and continuing until delivery. Nearly all (44 of 50) reported cases in which intrapartum antibiotic treatment failed to prevent EOGBS sepsis occurred in infants born to mothers with chorioamnionitis,^12-15 so complete evaluation and empiric treatment of infants of women with chorioamnionitis is also a necessity until neonatal sepsis is excluded.

There have been no studies of GBS prevention in infants born to women with GBS bacteriuria during pregnancy. Patients with asymptomatic GBS bacteriuria reported by Møller et al and by Persson et al apparently did not receive antibiotic therapy before delivery,^103,104 and Wood and Dillon were explicit that the mother of the infant with GBS infection in their series did not receive antibiotic therapy before delivery; therefore, potential effects of treatment at the time GBS bacteriuria is diagnosed are unclear. However, antepartum treatment of women with GBS bacteriuria is unlikely to prevent neonatal infection reliably because bacteriuria may reflect heavy colonization,¹⁰⁴ it is extremely difficult to eradicate GBS from the lower intestinal tract,^66,67 and recolonization of vagina after antibiotic therapy is common.⁶¹ Because intrapartum treatment for preterm premature rupture of membranes and chorioamnionitis and intrapartum prophylaxis both are effective, we recommend intrapartum administration of a penicillin to women with a history of GBS bacteriuria during pregnancy. Because the neonatal risk is apparently very high, and neither intrapartum therapy nor prophylaxis is completely effective, we also recommend empiric treatment with a penicillin for their infants until a diagnosis of invasive GBS disease can be confidently excluded.

When 1 infant from a multiple birth is diagnosed with invasive GBS disease, the apparent very high risk in the others justifies empiric treatment. Until there are empiric data to guide management, the minimum intervention should include a complete diagnostic evaluation and initiation of empiric therapy, using ampicillin or penicillin if GBS infection has been specifically diagnosed, and adding gentamicin if the sepsis is suspected only because of other clinical findings. Because of the unavoidable delay in availability of results and the potential for false-negative results, it would appear ill-advised to require confirmation of the diagnosis by positive cultures in the index case before initiating treatment of the other infant(s). Subsequent infants born to mothers of neonates with EOGBS disease also may be at very high risk,¹⁰¹ so intrapartum prophylaxis in subsequent pregnancies is apparently prudent. There are no data to address the question of whether it is sufficient to treat an at-risk twin or sibling only until cultures are found to be negative or whether a longer course of therapy is needed. Given the very high attack rate in initially asymptomatic twins, a full course of treatment may be appropriate. Extended treatment may not be necessary for siblings of previously affected infants, because the magnitude of their risk is unknown.

We also found that interventions in the second category are effective, providing opportunities to prevent infection in infants at moderately increased risk. Antepartum prophylaxis has appropriately been abandoned. Because the largest reported ex-perience shows an increased mortality rate with postpartum prophylaxis, this strategy should be abandoned, as well. Ohlsson and Myhr detailed deficiencies of intrapartum prophylaxis trials and concluded that "intrapartum chemoprophylaxis ... is not supported by conclusive evidence from well designed and conducted randomized, controlled trials".^17,18 Although published trials are imperfect, we have concluded that intrapartum prophylaxis is effective, in concurrence with several other reviews^25-28,71,105 and metaanalyses.^19,20 The necessity of distinguishing between intrapartum prophylaxis alone and in tandem with postpartum prophylaxis in the neonate has been almost universally overlooked, having been noted in print only by Ohlsson and Myhr.^17,32 Intrapartum prophylaxis alone reduces the risk of EOGBS sepsis by 80% (Tables 4, 7 and 8), and addition of postnatal prophylaxis increases this reduction in GBS disease to 95% (Table 8). The evidence that combined prophylaxis is superior to intrapartum prophylaxis alone is not conclusive (P = .14), but evidence supporting the former is arguably stronger than that supporting the latter. Considering the reported failures^12-15 of intrapartum prophylaxis, we concur with Boyer and Gotoff^24,69 that postpartum prophylaxis may be a useful supplement to intrapartum prophylaxis. Because neither Boyer et al⁶⁹ nor Garland and Fliegner⁷⁰ reported any cases of early-onset GBS disease in infants given 4 doses of ampicillin⁶⁹ or 1 dose of penicillin⁷⁰ after intrapartum prophylaxis, a single postpartum antibiotic dose may be sufficient. As the length of hospital stay after vaginal deliveries decreases,¹⁰⁶ more neonates may be discharged before a scheduled 4th antibiotic dose at 36 hours of age. Reported cases of intrapartum prophylaxis failure have been characterized by a high prevalence of chorioamnionitis (44 of 50 cases),^12-15 onset of symptoms before 1 hour of age in 15 of 16 patients reported by Ascher et al,¹³ admission to the NICU at .4 ± .3 hours of age in the 25 patients reported by Weisman et al,¹² and at least one of these findings in all the infants described by Ascher et al.¹³ Because continued treatment is recommended for infants whose mothers had chorioamnionitis and for those who are symptomatic, missed GBS cases will be unlikely if asymptomatic infants whose mothers received intrapartum prophylaxis are given intramuscular ampicillin every 12 hours until discharge from the hospital, up to a maximum of 4 doses.

The choice of agent for intrapartum prophylaxis remains controversial. The CDC recommends use of penicillin "because it has a narrower spectrum and thus is less likely to select for antibiotic resistant organisms".⁵ Amstey and Gibbs also advocated use of penicillin rather than ampicillin, because of its favorable pharmacokinetics and narrower spectrum of antimicrobial activity,²¹ citing a report of 4 infants who developed infections with ampicillin-resistant strains of Escherichia coli (3 cases) or Klebsiella pneumoniae (1 case) after prophylactic treatment with ampicillin or amoxicillin.⁵⁸ The mothers of those infants had clinical chorioamnionitis 3 to 12 days after initiation of antibiotic therapy for premature rupture of membranes, but only 1 received intravenous therapy and none received an aminoglycoside. Had these women been treated with intravenous ampicillin and gentamicin, either for preterm premature rupture of membranes while cultures were pending or for chorioamnionitis, it is much less likely that their infants would have developed infection with resistant organisms, so these observations may not be relevant to the use of intravenous ampicillin for GBS prophylaxis. Trials of intrapartum prophylaxis^79-83 or combined intrapartum and postpartum prophylaxis⁶⁹ have not included sufficient patient numbers to detect increased infections with resistant organisms or higher mortality, and data on these outcomes are rarely reported. Joseph et al¹⁰⁷ recently reported increases in the prevalence of ampicillin resistance, the frequency of maternal ampicillin therapy, and the virulence of disease among infants with early-onset E coli sepsis in the 5 years after implementation of routine intrapartum ampicillin prophylaxis. Although death from infection was more likely among infants with resistant organisms than in those whose organism was ampicillin-sensitive, there was no increase in the overall rate of E coli infection or neonatal mortality. Until more data are available, extrapolation from postpartum penicillin prophylaxis suggests that ampicillin should be the preferred agent for intrapartum prophylaxis, as well as for adjunctive postpartum prophylaxis. The AAP, ACOG, and CDC recommendations indicate that erythromycin or clindamycin may be appropriate alternatives, but the efficacy of these agents has not been studied.

The current AAP guidelines advise that "neonates with signs of septicemia should have a complete diagnostic evaluation and initiation of empiric antimicrobial therapy",² and the CDC concurs that antibiotic therapy "is appropriate for those infants suspected clinically of having sepsis".⁵ Management of asymptomatic infants is more problematic.¹⁰⁸ Pediatric practice has been inconsistent¹⁰⁹ and at times illogical. For example, pediatricians may be more rather than less likely to prescribe antibiotic therapy for infants whose mothers have received intrapartum prophylaxis, even though prophylaxis reduces the risk of disease.¹¹⁰ The revised AAP guidelines recommend limited evaluation (blood count, differential, and blood culture) if fewer than 2 doses of intrapartum chemoprophylaxis were given, or the infant is <35 weeks gestation² and has received no treatment otherwise.⁵ The rationale for these recommendations is not apparent, because no published data permit risk stratification based on the number of intrapartum antibiotic doses given or on a week-by-week for gestational age basis. The strategy of limited evaluation for infants treated for 4 hours or with fewer than 2 doses of antibiotic before delivery is also compromised by the low sensitivity of blood counts in diagnosis of neonatal sepsis.¹¹¹ Until additional data are available, management of asymptomatic infants whose mothers have received intrapartum prophylaxis will have to be individualized. For some, observation alone may be sufficient. Others may be managed appropriately with intramuscular prophylaxis alone or in concert with screening tests for infection, and those at highest risk may need full diagnostic evaluations and empiric therapy. These choices may be influenced by the expected marginal costs and benefits of interventions for each infant and are considered in greater detail in the third part this series.¹¹²

	ACKNOWLEDGMENTS

We thank Jerry Halpern, PhD, for advice regarding application of the binomial distribution to calculation of CIs for GBS attack rates and Byron W. Brown, PhD, for assistance in calculation of the probabilities of type II errors ( values).

	FOOTNOTES

Received for publication Oct 1, 1997; accepted Jan 27, 1999.

Reprint requests to (W.E.B.) Division of Neonatal and Developmental Medicine, 750 Welch Rd, Suite 315, Palo Alto, CA 94304. E-mail: benitzwe{at}leland.stanford.edu

	ABBREVIATIONS

EOGBS, early-onset neonatal group B streptococcus infection; AAP, American Academy of Pediatrics; ACOG, American College of Obstetricians and Gynecologists; CDC, Centers for Disease Control and Prevention; GBS, group B streptococcus.

	REFERENCES

Top Abstract Methods Results Discussion References

1. American Academy of Pediatrics, Committee on Infectious Diseases and Committee on Fetus and Newborn Guidelines for prevention of group B streptococcal (GBS) infection by chemoprophylaxis. Pediatrics 1992; 90:775-778 [Abstract/Free Full Text]
2. American Academy of Pediatrics, Committee on Infectious Diseases and Committee on Fetus and Newborn Revised guidelines for prevention of early-onset group B streptococcal (GBS) infection. Pediatrics 1997; 99:489-496 [Abstract/Free Full Text]
3. Group B, streptococcal infections in pregnancy ACOG Technical Bulletin Number 170-July 1992. Int J Gynaecol Obstet 1993; 42:55-59 [CrossRef][Medline]
4. Committee on Obstetric Practice American College of Obstetricians and Gynecologists. ACOG committee opinion. Prevention of early-onset group B streptococcal disease in newborns. Number 173, June 1996. Int J Gynaecol Obstet 1996; 54:197-205 [CrossRef][Medline]
5. Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease: a public health perspective. MMWR Morb Mortal Wkly Rep. 1996;45(No RR-7):1-24
6. Yancey MK, Duff P An analysis of the cost-effectiveness of selected protocols for the prevention of neonatal group B streptococcal infection. Obstet Gynecol 1994; 83:367-371 [Medline]
7. Gilbert GL, Isaacs D, Burgess MA, Prevention of neonatal group B streptococcal sepsis: is routine antenatal screening appropriate. Aust N Z J Obstet Gynaecol 1995; 35:120-126 [Medline]
8. Strickland DM, Yeomans ER, Hankins GD Cost-effectiveness of intrapartum screening and treatment for maternal group B streptococci colonization. Am J Obstet Gynecol 1990; 163:4-8 [Medline]
9. 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; 270:1442-1448 [Abstract/Free Full Text]
10. Landon MB, Harger J, McNellis D, Mercer B, Thom EA Prevention of neonatal group B streptococcal infection. Obstet Gynecol 1994; 84:460-462 [Medline]
11. Garland SM, Kelly N Early-onset neonatal group B streptococcal sepsis: economics of various prevention strategies. Med J Aust 1995; 162:413-417 [Medline]
12. Weisman LE, Stoll BJ, Cruess DF, Early-onset group B streptococcal sepsis: a current assessment. J Pediatr 1992; 121:428-433 [CrossRef][Medline]
13. Ascher DP, Becker JA, Yoder BA, Failure of intrapartum antibiotics to prevent culture-proved neonatal group B streptococcal sepsis. J Perinatol 1993; 13:212-216 [Medline]
14. Merenstein GB, Gibbs RE, Weisman LE Failure of maternal chemoprophylaxis to prevent neonatal group B streptococcal sepsis. Pediatr Res 1996; 39:298A
15. Yancey MK, Duff P, Kubilis P, Clark P, Frentzen BH Risk factors for neonatal sepsis. Obstet Gynecol 1996; 87:188-194 [CrossRef][Medline]
16. Rouse DJ, Goldenberg RL, Cliver SP, Cutter GR, Mennemeyer ST, Fargason CA Strategies for the prevention of early-onset neonatal group B streptococcal sepsis: a decision analysis. Obstet Gynecol 1994; 83:483-494 [Medline]
17. Ohlsson A, Myhr TL Intrapartum chemoprophylaxis of perinatal group B streptococcal infections: a critical review of randomized controlled trials. Am J Obstet Gynecol 1994; 170:910-917 [Medline]
18. Ohlsson A, Myhr TL. Group B streptococcal prophylaxis randomized trials. Am J Obstet Gynecol. 1995;172:1326-1327. Reply letter
19. Baley J, Fanaroff AA. Neonatal infections. Part 2. Specific infectious diseases and therapies. In: Sinclair JC, Bracken MB, eds. Effective Care of the Newborn Infant. Oxford, UK: Oxford University Press; 1992:480-483
20. Allen UD, Navas L, King SM Effectiveness of intrapartum penicillin prophylaxis in preventing early-onset group B streptococcal infection: results of a metaanalysis. Can Med Assoc J 1993; 149:1659-1665 [Abstract]
21. Amstey MS, Gibbs RS Is penicillin G a better choice than ampicillin for prophylaxis of neonatal group B streptococcal infections? Obstet Gynecol 1994; 84:1058-1059 [Medline]
22. Wedgwood JF, Carlin EB, Benjamin BL, Penicillin at birth can help prevent early-onset group B streptococcal disease. Pediatrics 1997; 99:651-652 Letter[Free Full Text]
23. Siegel JD, Cushion NB Prevention of early-onset group B streptococcal disease: another look at single-dose penicillin at birth. Obstet Gynecol 1996; 87:692-698 [CrossRef][Medline]
24. Gotoff SP, Boyer KM Prevention of early-onset neonatal group B streptococcal disease. Pediatrics 1997; 99:866-869 [Free Full Text]
25. Noya FJ, Baker CJ Prevention of group B streptococcal infection. Infect Dis Clin North Am 1992; 6:41-55 [Medline]
26. Katz VL Management of group B streptococcal disease in pregnancy. Clin Obstet Gynecol 1993; 36:832-842 [Medline]
27. Steele RW Control of neonatal group B streptococcal infection. J R Soc Med 1993; 86:712-715 [Abstract]
28. Schuchat A, Wenger JD Epidemiology of group B streptococcal disease. Risk factors, prevention strategies, and vaccine development. Epidemiol Rev 1994; 16:374-402 [Free Full Text]
29. Baker C, Edwards M. Group B streptococcal infections. In: Remington J, Klein J, eds. Infectious Diseases of the Fetus and Newborn Infant. Philadelphia, PA: WB Saunders Co, 1995:980-1054
30. Ramamurthy RS, Pyati SP, Pildes RS Penicillin prophylaxis for neonatal group-B streptococcal infection. Lancet 1979; 2:246-247 Letter[CrossRef][Medline]
31. Pyati SP, Pildes RS, Ramamurthy RS, Jacobs N Decreasing mortality in neonates with early-onset group B streptococcal infection: reality or artifact? J Pediatr 1981; 98:625-627 [CrossRef][Medline]
32. Ohlsson A, Myhr TL Intrapartum penicillin prophylaxis of early-onset streptococcal infection. Can Med Assoc J 1994; 150:1197-1198 Letter
33. Thacker SB Meta-analysis: a quantitative approach to research integration. JAMA 1988; 259:1685-1689 [Abstract/Free Full Text]
34. Elashoff JD Combining results of clinical trials. Gastroenterology 1978; 75:1170-1172 [Medline]
35. Sokol R, Rohlf F. Biometry. New York, NY: WH Freeman; 1995:698,724-760
36. Armitage P, Berry G. Statistical Methods in Medical Research. London, UK: Blackwell Science; 1994:508-519
37. Bracken MB. Statistical methods for analysis of effects of treatment in overviews of randomized trials. In: Sinclair JC, Bracken MB, eds. Effective Care of the Newborn Infant. Oxford, UK: Oxford University Press; 1992:13-20
38. Glantz S. Primer of Biostatistics. New York, NY: McGraw-Hill; 1992:114-121
39. Hanley JA, Lippman-Hand A If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA 1983; 249:1743-1745 [Abstract/Free Full Text]
40. Sokol R, Rohlf F. Biometry. New York, NY: WH Freeman; 1995:159-169
41. McGregor JA, French JI, Seo K Antimicrobial therapy in preterm premature rupture of membranes: results of a prospective, double-blind, placebo-controlled trial of erythromycin. Am J Obstet Gynecol 1991; 165:632-640 [Medline]
42. Mercer BM, Moretti ML, Prevost RR, Sibai BM Erythromycin therapy in preterm premature rupture of the membranes: a prospective, randomized trial of 220 patients. Am J Obstet Gynecol 1992; 166:794-802 [Medline]
43. Morales WJ, Angel JL, O'Brien WF, Knuppel RA Use of ampicillin and corticosteroids in premature rupture of membranes: a randomized study. Obstet Gynecol 1989; 73:721-726 [Medline]
44. Fortunato SJ, Welt SI, Eggleston M, Cole J, Bryant EC, Dodson MG Prolongation of the latency period in preterm premature rupture of the membranes using prophylactic antibiotics and tocolysis. J Perinatol 1990; 10:252-256 [Medline]
45. McCaul JF, Perry KG, Moore JL, Martin RW, Bucovaz ET, Morrison JC Adjunctive antibiotic treatment of women with preterm rupture of membranes or preterm labor. Int J Gynaecol Obstet 1992; 38:19-24 [CrossRef][Medline]
46. Miller JM, Brazy JE, Gall SA, Crenshaw MC, Jelovsek FR Premature rupture of the membranes: maternal and neonatal infectious morbidity related to betamethasone and antibiotic therapy. J Reprod Med 1980; 25:173-177 [Medline]
47. Amon E, Lewis SV, Sibai BM, Villar MA, Arheart KL Ampicillin prophylaxis in preterm premature rupture of the membranes: a prospective randomized study. Am J Obstet Gynecol 1988; 159:539-543 [Medline]
48. Johnston MM, Sanchez-Ramos L, Vaughn AJ, Todd MW, Benrubi GI Antibiotic therapy in preterm premature rupture of membranes: a randomized, prospective, double-blind trial. Am J Obstet Gynecol 1990; 163:743-747 [Medline]
49. Christmas JT, Cox SM, Andrews W, Dax J, Leveno KJ, Gilstrap LC Expectant management of preterm ruptured membranes: effects of antimicrobial therapy. Obstet Gynecol 1992; 80:759-762 [Medline]
50. Kurki T, Hallman M, Zilliacus R, Teramo K, Ylikorkala O Premature rupture of the membranes: effect of penicillin prophylaxis and long-term outcome of the children. Am J Perinatol 1992; 9:11-16 [Medline]
51. Matsuda Y, Ikenoue T, Hokanishi H Premature rupture of the membranesaggressive versus conservative approach: effect of tocolytic and antibiotic therapy. Gynecol Obstet Invest 1993; 36:102-107 [Medline]
52. Owen J, Groome LJ, Hauth JC Randomized trial of prophylactic antibiotic therapy after preterm amnion rupture. Am J Obstet Gynecol 1993; 169:976-981 [Medline]
53. Ernest JM, Givner LB A prospective, randomized, placebo-controlled trial of penicillin in preterm premature rupture of membranes. Am J Obstet Gynecol 1994; 170:516-521 [Medline]
54. Grable IA, Garcia PM, Perry D, Socol ML Group B Streptococcus and preterm premature rupture of membranes: a randomized, double-blind clinical trial of antepartum ampicillin. Am J Obstet Gynecol 1996; 175:1036-1042 [CrossRef][Medline]
55. Mercer BM, Miodovnik M, Thurnau GR, 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; 278:989-995 [Abstract/Free Full Text]
56. Gilstrap LC, Leveno KJ, Cox SM, Burris JS, Mashburn M, Rosenfeld CR Intrapartum treatment of acute chorioamnionitis: impact on neonatal sepsis. Am J Obstet Gynecol 1988; 159:579-583 [Medline]
57. Gibbs RS, Dinsmoor MJ, Newton ER, Ramamurthy RS A randomized trial of intrapartum versus immediate postpartum treatment of women with intra-amniotic infection. Obstet Gynecol 1988; 72:823-828 [Medline]
58. McDuffie RS, McGregor JA, Gibbs RS Adverse perinatal outcome and resistant Enterobacteriaceae after antibiotic usage for premature rupture of the membranes and group B streptococcus carriage. Obstet Gynecol 1993; 82:487-489 [Medline]
59. Baker CJ, Barrett FF Transmission of group B streptococci among parturient women and their neonates. J Pediatr 1973; 83:919-925 [CrossRef][Medline]
60. Franciosi RA, Knostman JD, Zimmerman RA Group B streptococcal neonatal and infant infections. J Pediatr 1973; 82:707-718 [CrossRef][Medline]
61. Hall RT, Barnes W, Krishnan L, Antibiotic treatment of parturient women colonized with group B streptococci. Am J Obstet Gynecol 1976; 124:630-634 [Medline]
62. Gardner SE, Yow MD, Leeds LJ, Thompson PK, Mason EO, Clark DJ Failure of penicillin to eradicate group B streptococcal colonization in the pregnant woman: a couple study. Am J Obstet Gynecol 1979; 135:1062-1065 [Medline]
63. Merenstein GB, Todd WA, Brown G, Yost CC, Luzier T Group B -hemolytic streptococcus: randomized controlled treatment study at term. Obstet Gynecol 1980; 55:315-318 [Medline]
64. Lewin EB, Amstey MS Natural history of group B streptococcus colonization and its therapy during pregnancy. Am J Obstet Gynecol 1981; 139:512-515 [Medline]
65. Weeks JW, Myers SR, Lasher L, Goldsmith J, Watkins C, Gall SA Persistence of penicillin G benzathine in pregnant group B streptococcus carriers. Obstet Gynecol 1997; 90:240-243 [CrossRef][Medline]
66. Paredes A, Wong P, Yow MD Failure of penicillin to eradicate the carrier state of group B streptococcus in infants. J Pediatr 1976; 89:191-193 [CrossRef][Medline]
67. Badri MS, Zawaneh S, Cruz AC, Rectal colonization with group B Streptococcus: relation to vaginal colonization of pregnant women. J Infect Dis 1977; 135:308-312 [Medline]
68. Boyer KM, Gadzala CA, Kelly PD, Gotoff SP Selective intrapartum chemoprophylaxis of neonatal group B streptococcal early-onset disease. III. Interruption of mother-to-infant transmission. J Infect Dis 1983; 148:810-816 [Medline]
69. Boyer KM, Gotoff SP Prevention of early-onset neonatal group B streptococcal disease with selective intrapartum chemoprophylaxis. N Engl J Med 1986; 314:1665-1669 [Abstract]
70. Garland SM, Fliegner JR Group B streptococcus (GBS) and neonatal infections: the case for intrapartum chemoprophylaxis. Aust N Z J Obstet Gynaecol 1991; 31:119-122 [Medline]
71. Dashefsky B, Wald ER, Green M Prevention of early-onset group B streptococcal sepsis. J Pediatr 1988; 112:1039-1042 [CrossRef][Medline]
72. Gibbs RS, McDuffie RS, McNabb F, Fryer GE, Miyoshi T, Merenstein G Neonatal group B streptococcal sepsis during 2 years of a universal screening program. Obstet Gynecol 1994; 84:496-500 [Medline]
73. Jeffery HE, McIntosh ED Antepartum screening and non-selective intrapartum chemoprophylaxis for group B streptococcus. Aust N Z J Obstet Gynaecol 1994; 34:14-19 [Medline]
74. Poulain P, Betremieux P, Donnio PY, Proudhon JF, Karege G, Giraud JR Selective intrapartum anti-bioprophylaxy of group B streptococci infection of neonates: a prospective study in 2454 subsequent deliveries. Eur J Obstet Gynecol Reprod Biol 1997; 72:137-140 [CrossRef][Medline]
75. Morales WJ, Lim D Reduction of group B streptococcal maternal and neonatal infections in preterm pregnancies with premature rupture of membranes through a rapid identification test. Am J Obstet Gynecol 1987; 157:13-16 [Medline]
76. de Cueto M, Sanchez MJ, Sampedro A, Miranda JA, Herruzo AJ, Rosa-Fraile M Timing of intrapartum ampicillin and prevention of vertical transmission of group B streptococcus. Obstet Gynecol 1998; 91:112-114 [CrossRef][Medline]
77. Katz VL, Moos MK, Cefalo RC, Thorp JM, Bowes WA, Wells SD Group B streptococci: results of a protocol of antepartum screening and intrapartum treatment. Am J Obstet Gynecol 1994; 170:521-526 [Medline]
78. Omenaca Teres F, Matorras R, Garcia Perea A, Elorza MD Prevention of neonatal group B streptococcal sepsis. Pediatr Infect Dis J 1987; 6:874 [Medline]
79. Matorrás R, Garcia-Perea A, Omenaca F, Diez-Enciso M, Madero R, Usandizaga JA Intrapartum chemoprophylaxis of early-onset group B streptococcal disease. Eur J Obstet Gynecol Reprod Biol 1991; 40:57-62 [CrossRef][Medline]
80. Allardice JG, Baskett TF, Seshia MM, Bowman N, Malazdrewicz R Perinatal group B streptococcal colonization and infection. Am J Obstet Gynecol 1982; 142:617-620 [Medline]
81. 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; 155:979-983 [Medline]
82. Tuppurainen N, Hallman M Prevention of neonatal group B streptococcal disease: intrapartum detection and chemoprophylaxis of heavily colonized parturients. Obstet Gynecol 1989; 73:583-587 [Medline]
83. Pylipow M, Gaddis M, Kinney JS Selective intrapartum prophylaxis for group B streptococcus colonization: management and outcome of newborns. Pediatrics 1994; 93:631-635 [Abstract/Free Full Text]
84. Allen UD, King SM Intrapartum penicillin prophylaxis of early-onset streptococcal infection. Can Med Assoc J 1993; 150:1200-1201 Letter
85. Bloom SL, Cox SM, Bawdon RE, Gilstrap LC Ampicillin for neonatal group B streptococcal prophylaxis: how rapidly can bactericidal concentrations be achieved? Am J Obstet Gynecol 1996; 175:974-976 [CrossRef][Medline]
86. Silver HM, Gibbs RS, Gray BM, Dillon HC Risk factors for perinatal group B streptococcal disease after amniotic fluid colonization. Am J Obstet Gynecol 1990; 163:19-25 [Medline]
87. Steigman AJ, Bottone EJ, Hanna BA Does intramuscular penicillin at delivery prevent group B beta hemolytic streptococcal disease of the newborn infant? J Pediatr 1975; 87:496 Letter[Medline]
88. Steigman AJ, Bottone EJ, Hanna BA Intramuscular penicillin administration at birth: prevention of early-onset group B streptococcal disease. Pediatrics 1978; 62:842-844 [Abstract/Free Full Text]
89. Steigman AJ, Bottone EJ, Hanna BA Control of perinatal group B streptococcal sepsis: efficacy of single injection of aqueous penicillin at birth. Mount Sinai J Med 1978; 45:685-693 [Medline]
90. Lloyd DJ, Belgaumkar TK, Scott KE, Wort AJ, Aterman K, Krause VW Prevention of group-B -haemolytic streptococcal septicaemia in low-birth-weight neonates by penicillin administered within two hours of birth. Lancet 1979; 1:713-715 [Medline]
91. Siegel JD, McCracken GH, Threlkeld N, Milvenan B, Rosenfeld CR Single-dose penicillin prophylaxis against neonatal group B streptococcal infections: a controlled trial in 18 738 newborn infants. N Engl J Med 1980; 303:769-775 [Abstract]
92. Siegel JD, McCracken GH, Threlkeld N, DePasse BM, Rosenfeld CR Single-dose penicillin prophylaxis of neonatal group-B-streptococcal disease. Lancet 1982; 1:1426-1430 [Medline]
93. Patel DM, Leblanc MH, Morrison JC, Postnatal penicillin prophylaxis and the incidence of group B streptococcal sepsis in neonates. South Med J 1994; 87:1117-1120 [CrossRef][Medline]
94. Pyati SP, Pildes RS, Jacobs NM, Penicillin in infants weighing two kilograms or less with early-onset Group B streptococcal disease. N Engl J Med 1983; 308:1383-1389 [Abstract]
95. Garland SM Early onset neonatal group B streptococcus (GBS) infection: associated obstetric risk factors. Aust N Z J Obstet Gynaecol 1991; 31:117-118 [Medline]
96. Boyer KM, Gotoff SP Strategies for chemoprophylaxis of GBS early-onset infections. Antibiot Chemother 1985; 35:267-280 [Medline]
97. Chalmers TC, Smith H, Blackburn B, A method for assessing the quality of a randomized control trial. Control Clin Trials 1981; 2:31-49 [CrossRef][Medline]
98. Sperling RS, Ramamurthy RS, Gibbs RS A comparison of intrapartum versus immediate postpartum treatment of intra-amniotic infection. Obstet Gynecol 1987; 70:861-865 [Medline]
99. Yagupsky P, Menegus MA, Powell KR The changing spectrum of group B streptococcal disease in infants: an eleven-year experience in a tertiary care hospital. Pediatr Infect Dis J 1991; 10:801-808 [Medline]
100. Lieu TA, Mohle-Boetani JC, Ray GT, Ackerson LM, Walton DL Neonatal group B streptococcal infection in a managed care population. Perinatal Group B Streptococcal Infection Study Group. Obstet Gynecol 1998; 92:21-27 [CrossRef][Medline]
101. Benitz WE, Gould JB, Druzin ML. Risk factors for early-onset group B streptococcal sepsis: estimation of odds ratios by critical literature review. Pediatrics. 1999;103(6). URL: http://www.pediatrics.org/cgi/content/full/103/6/e77
102. Egarter C, Leitich H, Karas H, Antibiotic treatment in preterm premature rupture of membranes and neonatal morbidity: a metaanalysis. Am J Obstet Gynecol 1996; 174:589-597 [CrossRef][Medline]
103. Møller M, Thomsen AC, Borch K, Dinesen K, Zdravkovic M Rupture of fetal membranes and premature delivery associated with group B streptococci in urine of pregnant women. Lancet 1984; 2:69-70 [Medline]
104. Persson K, Christensen KK, Christensen P, Forsgren A, Jorgensen C, Persson PH Asymptomatic bacteriuria during pregnancy with special reference to group B streptococci. Scand J Infect Dis 1985; 17:195-199 [Medline]
105. Gotoff SP, Boyer KM Prevention of group B streptococcal early onset sepsis: 1989. Pediatr Infect Dis J 1989; 8:268-270 [Medline]
106. Trends in length of stay for hospital deliveries, United States, 1970-1992. MMWR Morb Mortal Wkly Rep. 1995;44:335-337
107. 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 [Abstract/Free Full Text]
108. Alkalay AL, Brunell PA, Greenspon JS, Pomerance JJ Management of neonates born to mothers with group B streptococcus colonization. J Perinatol 1996; 16:470-477 [Medline]
109. Mercer BM, Ramsey RD, Sibai BM Prenatal screening for group B Streptococcus. I. Impact of antepartum screening on antenatal prophylaxis and intrapartum care. Am J Obstet Gynecol 1995; 173:837-841 [CrossRef][Medline]
110. Peralta-Carcelen M, Fargason CA, Cliver SP, Cutter GR, Gigante J, Goldenberg RL Impact of maternal group B streptococcal screening on pediatric management in full-term newborns. Arch Pediatr Adolesc Med 1996; 150:802-808 [Abstract/Free Full Text]
111. Greenberg DN, Yoder BA Changes in the differential white blood cell count in screening for group B streptococcal sepsis. Pediatr Infect Dis J 1990; 9:886-889 [Medline]
112. Benitz WE, Gould JB, Druzin ML. Preventing early-onset group B streptococcal sepsis: strategy development using decision analysis. Pediatrics. 1999;103:(6). URL: http://www.pediatrics.org/cgi/content/full/103/6/e76