Objective. In January 1995, a combined maternal and neonatal protocol for prevention of early-onset group B streptococcal (GBS) infection was implemented that consisted of a risk factor-based approach for maternal intrapartum chemoprophylaxis using ampicillin combined with a single intramuscular dose of penicillin given to all newborns within 1 hour of delivery. The objective of this study was to review the cases of early-onset GBS infections that occurred from 1995 to 1999 to identify factors associated with their continued occurrence despite implementation of a GBS chemoprophylaxis protocol.
Methods. Infants ≤72 hours of age with early-onset GBS infection born at Parkland Memorial Hospital in Dallas from January 1995 to December 1999 were identified through a prospective laboratory-based surveillance system. Maternal and infant medical records were reviewed for clinical and demographic data.
Results. There were 32 cases (0.47/1000 live births) of early-onset GBS infection for the 5-year period. This represented a 76% reduction compared with the rate from 1986 to 1994 (1.95/1000), when there was no protocol for GBS chemoprophylaxis. Thirteen cases (41%) did not have any identifiable maternal risk factor. Of the 19 cases (59%) with risk factors, maternal intrapartum fever was the most frequent (15 [79%]), followed by prematurity (6 [32%]) and prolonged rupture of membranes (6 [32%]). Among the 19 mothers with risk factors, 15 (79%) mothers received intrapartum chemoprophylaxis, and 12 (80%) of the 15 mothers had intrapartum fever. Only 33% of mothers with risk factors received ≥2 doses of intrapartum chemoprophylaxis, and among those with intrapartum fever, 25% received ≥2 doses. None of the 32 infants with early-onset GBS infection received the combination of intrapartum ampicillin and postnatal penicillin.
Conclusions. A combined obstetric and neonatal chemoprophylaxis protocol significantly reduced early-onset GBS infection. Maternal intrapartum fever was the most frequent risk factor associated with failure of chemoprophylaxis.
Group B streptococcus (GBS) has been the leading cause of early-onset neonatal sepsis in the United States since the 1970s.1 Early-onset GBS infection is associated with maternal GBS urogenital and anorectal colonization, as well as chorioamnionitis, preterm delivery, prolonged rupture of fetal membranes, and previous delivery of a GBS-infected infant.1–3 Identification of these risk factors with subsequent administration of intrapartum chemoprophylaxis has resulted in a 65% reduction in cases of early-onset GBS infections in the United States.3,4
In January 1995, a combined maternal and neonatal protocol for prevention of early-onset GBS infection was implemented at Parkland Memorial Hospital (PMH), a large inner-city county hospital in Dallas with >13 000 deliveries per year.5 This strategy consisted of a risk-based approach for maternal intrapartum chemoprophylaxis using ampicillin combined with a single intramuscular (IM) dose of penicillin to all newborns within 1 hour of delivery. The reasons for developing this specific protocol are as follows: 1) the risk-based strategy was deemed more feasible in our large public hospital setting because of the expense and time required for obtaining and processing >13 000 maternal GBS cultures annually as well as the logistic challenge of having results available in the labor suite at the time of decision making; 2) the use of intrapartum ampicillin was based on the early intrapartum chemoprophylaxis studies that demonstrated significant reduction in the rates of early-onset GBS infection when ampicillin was provided as the chemoprophylactic agent6–9; and 3) the use of the neonatal component was based on the results of a randomized clinical study by Siegel et al10,11 that demonstrated a significant reduction in early-onset GBS infection from 1.19/1000 live births to 0.25/1000 live births with the administration of a single IM dose of penicillin to newborns within 1 hour after delivery with no adverse effect. Additional observations during a period of >5 years (June 1, 1981, through October 31, 1986), during which a single dose of penicillin G was administered to all newborns at PMH and another 8 years (November 1, 1986, through December 31, 1994) when no GBS prophylaxis was given, confirmed the protective effect of this regimen.12 Implementation of the combined maternal and neonatal chemoprophylaxis strategy at PMH resulted in a 76% reduction in the incidence of early-onset GBS disease, from 1.95 per 1000 live births between 1986 and 1994 when no prophylaxis was provided to 0.47/1000 live births between 1995 and 1999 (odds ratio: 0.24; 95% confidence interval [CI]: 0.16–0.35; P < .001).5,12,13 This study reviews the cases of early-onset GBS infections that occurred at PMH from 1995 to 1999 during the use of the combined protocol to identify factors that might be associated with its continued occurrence despite the implementation of a GBS chemoprophylaxis protocol.
All infants who were born at PMH from January 1, 1995, to December 31, 1999, and developed early-onset GBS infection were identified by prospective surveillance of the microbiology laboratory records. The medical records of the infants and their mothers were reviewed for pertinent demographic and clinical information. Early-onset GBS infection was defined by the isolation of GBS from blood or cerebrospinal fluid (CSF) from an infant ≤72 hours of age. Three days was used to define early-onset GBS infection because this is most compatible with the pathogenesis of GBS disease acquired in the peripartum period, and >95% of cases of GBS infection that occur within the first week of life occur at ≤72 hours of age.12,14,15
The GBS chemoprophylaxis protocol implemented in January 1995 (Fig 1) consisted of administration of intrapartum intravenous (IV) ampicillin (2 g every 6 hours) to all mothers with any of the following risk factors: 1) preterm labor or anticipated delivery before 37 weeks of gestation, 2) rupture of membranes for ≥18 hours (PROM), 3) intrapartum fever of 38°C or greater (chorioamnionitis), and 4) previous delivery of an infant with GBS infection. In addition to ampicillin, mothers with chorioamnionitis received gentamicin (loading dose of 120 mg, followed by 80 mg IV every 8 hours). Clindamycin (900 mg IV every 8 hours) was provided to women who reported allergy to β-lactam antibiotics. No routine maternal GBS screening cultures were performed.
The neonatal component consisted of a single IM dose of aqueous penicillin G administered in the delivery room within 1 hour of birth to all newborns with no clinical signs suggestive of possible infection. From January 1995 to September 1999, penicillin was administered at a dose of 50 000 U (1 mL) to full-term infants and 25 000 U (0.5 mL) to infants with birth weight <2 kg. Because of a change in the aqueous penicillin G preparation from 50 000 U/mL to 60 000 U/mL in 1999, the dose was changed to 60 000 U IM (1 mL) for full-term infants and 30 000 U (0.5 mL) if the birth weight was <2 kg to allow for greater ease of preparation and administration. In 1995, as part of a study evaluating the new combined protocol,5 infants without clinical signs of infection but born to mothers with risk factors had a blood culture obtained before administration of IM penicillin. From 1996 through 1999, these infants did not have a blood culture obtained routinely before receiving penicillin. Newborns who had clinical signs of infection or whose mothers had received ampicillin and gentamicin for intrapartum fever did not receive penicillin prophylaxis in the delivery room but were treated with ampicillin and gentamicin after a blood culture was obtained in the nursery (Fig 1).
Lumbar puncture was performed at the discretion of the provider. Meningitis was diagnosed when the CSF culture yielded GBS or the CSF indices were abnormal and the blood culture was positive for GBS.
χ2 analysis was used to compare the rates of GBS infection before and after implementation of the GBS prophylaxis strategy.
After implementation of the combined GBS chemoprophylaxis protocol in 1995, there were 32 cases of early-onset GBS infection among 68 764 live births at PMH, or 0.47 cases per 1000 live births (Table 1). The demographic and clinical characteristics of these infants and their mothers are presented in Table 2. All of the mothers were Hispanic or black. From 1995 to 1999, the ethnicity of mothers who delivered at PMH was as follows: Hispanic, 72%; black, 19%; white, 6%; and other, 3%. The rate of early-onset GBS infection in black infants, 1.07 per 1000 live births, was significantly higher than among Hispanics (0.36 per 1000 live births; P = .003). Of the 7 (22%) mothers who had a cesarean section performed, all were in labor, 1 had intact membranes, 4 had intrapartum fever, and 2 had PROM.
FSThe majority of the 32 infants with early-onset GBS infection were full-term and female, and all were singleton deliveries (Table 1). Only 6 (19%) of the infants had a gestational age <37 weeks; of these, 4 (67%) were between 33 and 36 weeks, 1 (17%) was 29 weeks, and another 1 was 31 weeks. Two (6%) infants had a birth weight ≤1500 g, and 1 of these weighed <1000 g at birth.
Of the 32 infants, 14 (44%; Table 2) had no clinical signs of infection but were evaluated for possible bacteremia because of maternal intrapartum fever (10 [71%]) or prematurity (4 [29%]). Eighteen (56%) infants had clinical signs of sepsis that consisted of respiratory depression (39%) or distress (56%) soon after birth. One additional infant was evaluated for sepsis only because of hypoglycemia. All of the infants received a diagnosis of GBS infection in the first 24 hours of age. The majority of infants (30 [94%] of 32) received a diagnosis within 4 hours after birth by having either a blood culture performed that subsequently yielded GBS or clinical signs of infection.
The majority of infants received a diagnosis of only bacteremia or sepsis (81%). Five (16%) infants had radiographic evidence of pneumonia in addition to having a positive blood culture for GBS. Thirty infants had a lumbar puncture performed; 28 had sterile CSF cultures, and in 2 infants, there was insufficient CSF for culture. Overall, 16 infants had normal CSF analyses, and 5 infants had abnormal CSF indices suggestive of meningitis. The CSF of 7 infants was unassessable secondary to a traumatic lumbar puncture or insufficient quantity for measurement of all indices. There were no deaths as a result of GBS since implementation of the combined chemoprophylaxis protocol.
Maternal Risk Factors and Chemoprophylaxis
A maternal risk factor for early-onset GBS infection was present in 19 (59%) cases. Maternal intrapartum fever was the most frequent (79%), followed by prematurity and PROM in 6 (32%) cases each. No maternal risk factor for early-onset GBS infection was identified in 13 (41%) of the 32 cases (Table 3).
Of the 19 mothers with risk factors, 15 (79%) received intrapartum chemoprophylaxis, 12 (80%) for intrapartum fever and 3 (20%) for preterm delivery and PROM (Table 4). Of these 15 mothers, only 5 (33%) received 2 or more doses of ampicillin for preterm delivery (n = 1), preterm delivery with PROM (n = 1), preterm delivery with PROM and intrapartum fever (n = 1), intrapartum fever (n = 1), and intrapartum fever with PROM (n = 1). Ten (67%) of the mothers received only 1 dose. One mother with intrapartum fever and PROM reported a penicillin allergy and received 1 dose of clindamycin. All GBS isolates were susceptible to both clindamycin and erythromycin. All 15 infants born to mothers who received intrapartum chemoprophylaxis received ampicillin and gentamicin. Six (40%) infants had clinical signs of sepsis and were either preterm (n = 3) or born to mothers with intrapartum fever (n = 3). The other 9 (60%) infants had no sign of infection but were born to mothers with intrapartum fever.
Four (21%) of the 19 mothers with risk factors (3 intrapartum fever; 1 prematurity) did not receive intrapartum chemoprophylaxis (Table 4). In 1996, 2 mothers with intrapartum fever delivered 18 and 24 minutes after the onset of fever. Another mother in 1996 had a preterm delivery at 33 to 34 weeks; her labor was induced secondary to severe pregnancy-induced hypertension, but she was later delivered by cesarean section with intact membranes. One mother in 1999 had intrapartum fever, but the time before delivery was not known. With respect to their infants, 2 (1 preterm and 1 full-term exposed to maternal intrapartum fever) infants received penicillin and 2 (both mothers with intrapartum fever) were treated with ampicillin and gentamicin.
Thirteen (41%) infants were born to mothers who had no identifiable risk factor at the time of delivery and therefore did not receive intrapartum chemoprophylaxis (Table 4). Four infants were assessed as preterm on physical examination and Ballard score, although by obstetric dates they were >36 weeks of gestation. These 4 newborns had no clinical signs of infection but were born in 1995 when, by protocol, a blood culture was obtained before penicillin prophylaxis was provided. Three of these infants remained well during the nursery stay, and 1 developed transient hypoglycemia only. Eight infants had clinical signs of sepsis, and another 1 was mistakenly identified by nursery personnel as being born to a mother with intrapartum fever; all of these newborns received ampicillin and gentamicin after blood cultures were obtained. The combination of intrapartum ampicillin and postnatal penicillin had not been administered to any of these 32 mother–infant pairs.
The implementation in 1995 of a GBS chemoprophylaxis strategy that combined a maternal risk factor–based approach with administration of IM penicillin to all newborns within 1 hour of delivery reduced significantly the rate of early-onset GBS disease from 1.95 cases per 1000 live births from 1986 to 1994 when no chemoprophylaxis was used to 0.47 cases per 1000 live births from 1995 to 1999. Specifically, the incidence of early-onset GBS infection was reduced significantly among infants born to Hispanic women, from 1.49 to 0.36 cases per 1000 live births (76% reduction), to black women, from 2.83 to 1.07 cases per 1000 live births (62% reduction), and to white women, from 1.89 to 0 per 1000 live births (100% reduction). In addition, there were no deaths as a result of GBS since implementation of the chemoprophylaxis protocol, compared with a case-fatality rate of 8% from 1986 to 1994.12 Among the 32 cases of early-onset GBS infection that occurred at PMH after implementation of the combined protocol, the most frequent maternal risk factor was intrapartum fever or presumed chorioamnionitis. However, 41% of cases were not associated with the presence of any maternal risk factor at delivery.
In 1996, the Centers for Disease Control and Prevention (CDC) published consensus guidelines for prevention of early-onset GBS infection.3 These guidelines recommended either a risk factor–based or screening-based approach for identification of obstetric patients for intrapartum prophylaxis. In a population-based surveillance conducted in selected counties of 8 states, Schrag et al4 found that implementation of these guidelines was associated with a 65% reduction in early-onset GBS infection. Similarly, in Connecticut, Baltimore et al16 found a reduction in early-onset GBS disease from 0.61 cases per 1000 live births in 1996 to 0.23 case per 1000 live births in 1999 after implementation of the CDC recommendations. Neither study, however, determined the contribution of each of the 2 prophylaxis strategies to the decrease in GBS cases.
Several studies have reported the efficacy of the screening-based strategy for prevention of early-onset GBS disease.17–19 Hafner et al17 observed a 5-fold reduction in rates of early-onset GBS disease, from 5.4 cases per 1000 live births to 1.1 per 1000 live births (P = .002), when a universal screening-based strategy supplanted a risk factor–based strategy. Similarly, Main and Slagle19 reported a significant decrease in early-onset GBS infection to 0.07 cases per 1000 live births after implementation of a screening-based protocol after a risk-based strategy did not reduce the rate of GBS infection (1.1 cases/1000 live births) from a baseline period of no chemoprophylaxis (1.1/1000 live births). Of the 15 cases of early-onset GBS disease that occurred when the risk-based protocol was in place, 40% had no maternal risk factor present at delivery. The proportion of cases without a maternal risk factor is similar to our findings and supports the potential benefit of postnatal prophylaxis when a risk-based approach is used. In a retrospective cohort study of 5848 women who delivered full-term infants between 1994 and 1996, Gilson et al18 reported no cases of early-onset GBS infection among infants whose mothers were managed using a screening-based protocol compared with 1.5 cases per 1000 live births (4 cases) when a risk-based approach was used (P = .04). Among these 4 cases of early-onset GBS disease, only 1 was associated with a maternal risk factor. However, adherence to the protocol for administering intrapartum antibiotics to those women with positive screening cultures was not optimal, because only 72% of GBS-colonized women received intrapartum antimicrobial prophylaxis and only 48% received appropriate prophylaxis >4 hours before delivery.
The effectiveness of risk-based intrapartum chemoprophylaxis for prevention of early-onset GBS disease also has been documented.5,13,20,21 Factor et al20 reported a 94% reduction in the incidence of early-onset GBS infection at the University of Miami-Jackson Memorial Medical Center, from 1.7 cases per 1000 live births in 1992 to 0.2 cases per 1000 live births in 1995, that was associated with increased intrapartum antibiotic use in women with risk factors (P = .002). Recently, in a retrospective, multicenter, case-control study, Lin et al21 demonstrated that the administration of intrapartum antibiotics based on the presence of maternal risk factors without reference to maternal GBS colonization was 86% effective in preventing early-onset GBS disease. On the basis of a 70% prevalence of maternal risk factors expected among cases in the absence of intrapartum prophylaxis, they estimated that the risk-based strategy could reduce early-onset GBS disease by 60%. In contrast, the CDC, in a multistate surveillance study performed in 1995, estimated that the proportion of preventable early-onset GBS disease was only 41% for the risk-based strategy but 78% for the screening-based approach.22 Locksmith et al,23 however, failed to show a statistically significant difference in early-onset GBS infection between the time periods when the 2 protocols were implemented at their institution. We achieved a reduction of 76% in early-onset GBS infection after implementation of a combined GBS chemoprophylactic strategy that was based on a maternal risk-based approach. The differences and similarities between the CDC guidelines published in 1996 and 200224 and our protocol implemented at PMH in 1995 are provided in Table 5. In essence, our protocol favors the intrapartum administration of ampicillin every 6 hours rather than penicillin every 4 hours and neonatal prophylaxis. In addition, the risk assessment did not include GBS bacteriuria because the result of routine urine cultures was not readily available at delivery.
Among the 32 cases of early-onset GBS infection that occurred after implementation of the chemoprophylaxis protocol, the most frequent maternal risk factor was intrapartum fever. In addition, maternal intrapartum fever was present in 80% of the cases in which the mothers received chemoprophylaxis but their infants developed early-onset GBS infection. Several studies have reported that use of intrapartum antibiotics in women with chorioamnionitis reduces the risk of neonatal sepsis as a result of GBS and other organisms.25,26 However, Lin et al21 found that the effectiveness of intrapartum antibiotics was lowest in infants born to mothers with intrapartum fever (72%; 95% CI: −8% to 92%). These investigators concluded that once infection was established, the preventive effect of chemoprophylaxis is diminished. Similar to our findings, Ascher et al27 reported that chorioamnionitis was present in 81% of cases of early-onset GBS disease that occurred despite the use of intrapartum antibiotics. The duration and number of doses of maternal antibiotic treatment before delivery also are important factors for reducing chemoprophylaxis failures in mothers with chorioamnionitis. Pylipow et al9 reported that asymptomatic infants of GBS-colonized mothers who received 2 or more doses of ampicillin before delivery had normal laboratory evaluation and clinical course without antibiotics, whereas among the 43 infants whose GBS-colonized mothers had only 1 dose, 2 had positive blood cultures for GBS. Lin et al21 reported that the overall effectiveness of intrapartum antibiotic prophylaxis was 86% (95% CI: 66%–94%) and 89% (95% CI: 70%–96%) when the first dose was given at least 2 hours before delivery and 71% (95% CI: −8% to 92%) when given within 2 hours. In our study, among the women who received chemoprophylaxis for intrapartum fever, 75% received only 1 dose. It may be that these infants already were infected in utero and that a single dose of ampicillin was insufficient to sterilize the blood culture. It also highlights the need for antibiotic therapy of newborns born to mothers with chorioamnionitis. The frequent association of intrapartum fever with early-onset GBS infection and the suboptimal antibiotic treatment of these mothers argue for prevention of chorioamnionitis, an advantage that has been seen with screening-based strategies.23
Other potential reasons for failure of our combined chemoprophylaxis protocol are absence of identifiable maternal risk factors because no maternal GBS screening was performed, late presentation to the labor suite so that there is not enough time for antibiotics to have an effect before delivery, and unforeseeable communication gaps or errors. Among the 19 mothers who had an identifiable risk factor, 4 (21%) did not receive chemoprophylaxis. Two of these 4 mothers presented in labor, developed fever, but delivered within 25 minutes and before chemoprophylaxis could be initiated. The reason for the lack of chemoprophylaxis in the other 2 mothers could not be ascertained from chart review. In addition, 41% of cases did not receive chemoprophylaxis because there was no identifiable maternal risk factor. This is clearly a disadvantage of all risk-based strategies because at least 25% to 40% of cases of early-onset GBS occur among infants who are born to women without risk factors.3,8,28,29 Because of this and on the basis of the studies of Siegel et al10,11 at our institution, we included in our chemoprophylaxis regimen a neonatal component that consisted of the administration of IM penicillin to all infants within 1 hour of birth. Four (31%) of 13 GBS-infected infants who were born to mothers without risk factors received IM penicillin. However, no case of early-onset GBS infection was documented when both the mother and the infant received chemoprophylaxis. Similarly, studies of Boyer and Gotoff6 and Garland and Fliegner30 that demonstrated a beneficial effect of intrapartum chemoprophylaxis in prevention of early-onset GBS infection also included a neonatal component. Boyer and Gotoff treated newborns who were born to mothers who were colonized with GBS with ampicillin every 12 hours for 48 hours, whereas Garland and Fliegner provided a single dose of IM penicillin to full-term infants. CDC, however, recommends that at-risk asymptomatic newborns have a complete blood cell count and blood culture but no prophylactic antibiotic therapy.3,24 Finally, the definition of early-onset GBS infection as those occurring within the first 72 hours of age rather than in the first week of life may have affected our results. Although almost all cases of GBS infection that occur in the first week of life have their onset in the first 72 hours of age, additional cases of GBS infection may have been detected with a longer surveillance period.
Our combined maternal and neonatal GBS chemoprophylaxis strategy significantly reduced both the incidence and the mortality of early-onset GBS disease. In addition, our findings support the addition of a neonatal component to current maternal chemoprophylaxis regimens. Using this combined protocol, we have been able to achieve rates of early-onset GBS infection comparable to those achieved using a screening-based strategy without the cost and time required to process >13 000 screening cultures per year.
This work was presented in part at the 38th Annual Meeting of the Infectious Disease Society of America, September 7–10, 2000, New Orleans, Louisiana.
We thank Fiker Zeray, RN, for assistance in data collection; the Microbiology Laboratory (R. M. Gander, PhD, and Linda Byrd, MT-ASCP-SM), Parkland Health and Hospital Systems, for assistance in surveillance of bacterial cultures; and Karen Kirby for secretarial expertise.
- ↵Baker CJ, Edwards MS. Group B streptococcal infections. In: Remington JS, Klein JO, eds. Infectious Disease of the Fetus and Newborn Infant. 5th ed. Philadelphia, PA: WB Saunders; 2001:1091–1156
- ↵Pylipow M, Gaddis M, Kinney JS. Selective intrapartum prophylaxis for group B streptococcus colonization: management and outcome of newborns. Pediatrics.1994;93 :631– 635
- ↵Siegel JD, McCracken GH Jr, Threlkeld N, DePasse BM, Rosenfeld CR. Single-dose penicillin prophylaxis of neonatal group-B streptococcal disease. Lancet.1982;1426– 1430
- ↵Schuchat A, Zywicki SS, Dinsmoor MJ, et al. Risk factors and opportunities for prevention of early-onset neonatal sepsis: a multicenter case-control study. Pediatrics.2000;105 :21– 26
- ↵Bromberger P, Lawrence JM, Braun D, Saunders B, Contreras R, Petitti DB. The influence of intrapartum antibiotics on the clinical spectrum of early-onset group B streptococcal infection in term infants. Pediatrics.2000;106 :244– 250
- ↵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– 1098
- ↵Gilson GJ, Christensen F, Romero H, Bekes K, Silva L, Qualls CR. Prevention of group B streptococcus early-onset neonatal sepsis: comparison of the Centers for Disease Control and Prevention screening-based protocol to a risk-based protocol in infants at greater than 37 weeks’ gestation. J Perinatol.2000;20 :491– 495
- ↵Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease. MMWR Morb Mortal Wkly Rep.2002;51(RR-11) :1– 23
- ↵Boyer KM, Gadzala CA, Burd LI, Fisher DE, Paton JB, Gotoff SP. Selective intrapartum chemoprophylaxis of neonatal group B streptococcal early-onset disease. I. Epidemiologic rationale. J Infect Dis.1983;148 :795– 801
- ↵Gotoff SP, Boyer KM. Prevention of early-onset neonatal group B streptococcal disease. Pediatrics.1997;99 :866– 869
- Copyright © 2003 by the American Academy of Pediatrics