In an effort to decrease the intrapartum transmission of the group B streptococcus (GBS), the CDC issued guidelines in May 1996 for the “Prevention of Perinatal Group B Streptococcal Disease,”1 which were reviewed and accepted by The American Academy of Pediatrics and The American College of Obstetricians and Gynecologists. Many clinicians adopted the “Screening-Based Approach” which recommended maternal GBS rectal and vaginal cultures at 35 to 37 weeks gestation and the use of intrapartum antibiotic prophylaxis for colonized mothers and those with identified GBS risk factors. A recent large multicenter review suggested greater efficacy of the screening-based approach 2 compared with the alternative strategy of using intrapartum antibiotics in all situations where risk factors exist, without screening. Recently the CDC has revised these guidelines to recommend a strategy of using intrapartum antibiotic prophylaxis based on universal prenatal screening cultures. 3
One consistent outcome of both the 1996 and 2002 guidelines is that pediatricians are frequently faced with the challenge of evaluating and managing infants born of mothers who have received incomplete antibiotic prophylaxis as part of those guidelines. These infants are identified as being “at risk” newborns, defined as those born to mothers who had positive vaginal/anal cultures for GBS and/or other risk factors and did not receive antibiotics at least 4 h before birth. In both the 1996 and 2002 guidelines, it is recommended that otherwise asymptomatic newborns ≥35 weeks gestation receive a limited evaluation to include at least a CBC and blood culture 1, 2; routine antimicrobial therapy is not recommended. Both sets of guidelines suggest a full diagnostic evaluation and empiric therapy for infants with signs and symptoms of sepsis. The 2002 guidelines argue that this neonatal management algorithm should actually reduce costs by avoiding the unnecessary evaluation of asymptomatic infants who received appropriate intrapartum antibiotics, 3 although both guidelines avoid discussion of the utility and cost effectiveness of performing hundreds of laboratory tests on the asymptomatic at risk newborn population. The aims of this study were to determine whether use of CBC and blood cultures obtained in asymptomatic at risk newborn infants, according to the CDC guidelines, added to the ability of clinicians to detect early onset neonatal sepsis before clinical signs and symptoms were present.
The study was conducted at Holy Cross Hospital, a community hospital in suburban Washington, DC, with ∼7000 deliveries annually. The study was approved by the Institution Review Board at the hospital. This was a review of the CBC and blood culture results from all initially asymptomatic at risk infants born between May 1, 1996 and July 31, 1999 and an extensive medical record review of all newborns ≥ 35 weeks gestation with a discharge diagnosis of clinical sepsis and/or positive blood cultures during the first week of life. Newborns were classified as at risk if gestational age was ≥35 weeks, without other complications and had one or more of the CDC-defined criteria; if they were born to a mother with: proven or unknown GBS colonization, premature rupture of membranes >18 h, fever ≥38.0°C or with a previous infant with invasive GBS disease and who did not receive at least one dose of appropriate antibiotic 4 h before delivery. All infants in this group were initially asymptomatic for at least 4 h after birth. All of the infants classified as being at risk during the study period had a CBC with manual differential count and an aerobic blood culture obtained and sent from the well-baby nursery as soon after birth as possible (within 4 h) based on a rigorously enforced nursery protocol concordant with the 1996 CDC guidelines. The maternal GBS colonization rate was obtained from a log of all newborns admitted to the well-baby nursery, which included maternal GBS status. Laboratory records of infants who had a blood culture and CBC obtained from the term nursery during the study period were reviewed and analyzed. An abnormal CBC was defined as either a total white blood cell (WBC) count of ≤5000 or ≥30 000/mm3, an absolute neutrophil count of <1500/mm3 or an immature-mature neutrophil ratio of >0.2. Blood cultures were obtained in a sterile manner and were incubated and read with a nonradiometric continuous monitoring system (Bactec 9000 series; BD Biosciences, Franklin Lakes, NJ).
The study hospital has a Level 3 neonatal intensive care unit (NICU) staffed by on site neonatologists 24 h per day, and infants with signs and symptoms consistent with early onset sepsis are evaluated by a neonatologist and transferred to the NICU as soon as possible. A group of neonatologists and staff hospital pediatricians who reviewed the newborn records determined that the following were findings suggestive of early onset sepsis: fever; hypothermia; lethargy; tachypnea; apnea/bradycardia; cyanosis; and hypoglycemia (not explained by other diagnosis). The following additional data were abstracted from medical records of all infants transferred to the NICU with the diagnosis of sepsis: additional final diagnoses; CBC and culture results; gestational age of infant; maternal GBS status; the presence of maternal risk factors; the number of doses of maternal intrapartum antibiotics received; and the specific clinical signs of sepsis. The time of transfer from the newborn nursery to the NICU was recorded and compared with the time of birth as well as the time of the first dose of intrapartum antibiotics. In addition to the chart review, laboratory records provided by the microbiology department were reviewed to identify all infants from whom blood cultures were drawn at ≤7 days of age, data that were then cross-checked with clinical records.
During the study period there were 20 554 deliveries; 19 320 (94%) were at ≥35 weeks gestation. The mothers identified themselves (through self-reported ethnicity) as: African-American, 58%; Caucasian, 30%; Hispanic, 6%; Asian, 3%’ and other categories, 3%. The prevalence of maternal GBS colonization was 17%.
Of the 19 320 newborns, 1665 (8.6%) were initially asymptomatic yet met the CDC/AAP definition of at risk newborn infants. A total of 300 term newborns were given a discharge diagnosis of sepsis, although only 17 came from the initially asymptomatic at risk newborn subset (Fig. 1). All infants with a diagnosis of early onset neonatal sepsis were symptomatic within the first 48 h of life, with or without a positive blood culture. The frequency of sepsis signs and symptoms was: tachypnea, 58%; cyanosis, 25%; lethargy, 20%; apnea/bradycardia, 20%; hypoglycemia (without other etiology), 8.4%; fever, 7%; hypothermia, 3.5%. Nearly all (91%) of the neonates with the diagnosis of sepsis had more than 1 sign/symptom, and 77% had at least 3. Maternal risk factors for sepsis were present in 39%, with fever being the most common (56%). Maternal colonization for GBS was found in 89 of 300 (30%), and results were unknown in 30 (10%). Only 4% of these mothers had greater than one risk factor. Maternal intrapartum antibiotic prophylaxis was given to 153 of these mothers (51%), with 71 (46%) receiving 1 dose, 52 (34%) receiving 2 doses and 30 (20%) receiving >2 doses. Ninety-one (59%) received 1 dose or more at least 4 h before delivery. Immediate transfer to the NICU from the delivery room because of the severity of symptoms occurred in 134 (39%). The remaining newborns with symptoms of sepsis were subsequently transferred to the NICU from the term nursery.
None of the 17 initially asymptomatic at risk newborns who developed clinical sepsis ever had positive blood cultures. Of the remaining 283 infants with sepsis, 14 had positive blood cultures. All were symptomatic with 1 symptom or more at the time the blood culture was known to be positive. Blood cultures were positive for the following pathogens:Escherichia coli, 3; group B Streptococcus, 2;Klebsiella sp., 1; and Streptococcus mitis, 2. Six blood cultures (43%) were considered to be contaminants, including the following species:Staphylococcus epidermidis, 4;Streptococcus capitus, 1; and Staphylococcus warneri, 1. The two newborns with blood cultures positive for GBS were born to mothers who were culture-negative for GBS.
All 8 newborns with blood cultures positive for pathogens had abnormal white blood cell indices; 2 had a total WBC of <5000/mm3, whereas only 1 had a total WBC of ≥30 000/mm3. All had a band form-polymorphonuclear cell ratio of >0.2. Only 1 had an absolute neutrophil count of <1500/mm3. This infant also had a blood culture positive for GBS. The total number of initially asymptomatic at risk newborns with abnormal WBC indices was 454 of 1665 (27%). Only 7 (1.5%) of those with abnormal WBC developed clinical sepsis. The sensitivity and specificity of an abnormal WBC in predicting sepsis in the asymptomatic at risk newborns were 41 and 73%, respectively. The positive predictive value of an abnormal CBC in identifying sepsis in the asymptomatic at risk group was 1.5%, whereas the negative predictive value was 99%. Likelihood ratios (LR) were calculated as: positive LR, 1.52; negative LR, 0.81; with an odds ratio (positive LR-negative LR, 1.88).
Before the institution of the CDC guidelines, there was greater variability in the obstetricians’ use of intrapartum antibiotics at our hospital, and the annual rates of culture-proved early onset GBS among full term births was on average 1 per 1000. Although the number of term births has remained stable at ∼6500 annually throughout the 1990s, after institution of the guidelines in 1996 there were only 2 cases of culture-proved early onset GBS sepsis in term newborns, a reduction in rates from 1.2 per 1000 to 0.3 per 1000 births (P < 0.0001).
In the years since the publication of the 1996 guidelines for the prevention of early onset group B streptococcal sepsis, widespread institution of these guidelines (or similar practices) led to a reduction in the incidence of this infection by ∼70%, as summarized in the 2002 revised guidelines. 3 During this study we noted a 4-fold reduction in rates at our institution, when comparing rates from pre- and post-institution of the guidelines. In addition to the benefit, one result of institutionalizing guidelines to prevent early onset GBS has been the identification of a large number of infants who in the past may not have been particularly scrutinized, those otherwise asymptomatic “at risk newborn” infants. In many ways it is more straightforward to manage clinically symptomatic infants, given that all those identified at our institution during the study period were ill within 48 h of birth (39% immediately), the majority having multiple symptoms and many with abnormal laboratory results. Implementation of the 1996 GBS guidelines (and adherence to the 2002 management algorithm for newborns exposed to intrapartum prophylaxis) has created a significant population of asymptomatic term or near term at risk infants who are undergoing screening with CBCs and blood culture. We determined that only 1% of the initially asymptomatic at risk newborns identified in this study eventually developed clinical sepsis, a rate identical with that determined by Escobar et al. 4 in the same subset of their population-based study of 18 299 ≥2000-g newborn infants.
For a generation pediatricians have sought to develop tools to help screen newborn infants for findings consistent with early sepsis. The results of many of these studies have been influenced significantly by the definition of the at risk neonatal population subject to the screening method chosen. The identification of a blood culture that becomes positive for a known neonatal pathogen has long been considered a standard for defining sepsis in the newborn infant. The sensitivity of a blood culture as a screening tool is hard to evaluate as older studies of GBS sepsis, such as the 7-year experience of Lannering et al. 5 with group B streptococcal disease published in 1983, and a similar study published in 1992 by Weisman et al. 6 used positive blood culture as the criterion for inclusion into the study population. In recent years few studies have addressed the utility of obtaining a blood culture in a largely asymptomatic population identified by the guidelines. Safier et al. 7 recently estimated that based on the rate of GBS bacteremia found in their newborn population; and by counting all blood cultures performed for both specific clinical indications as well as screening tests in asymptomatic newborns, it would take ∼10 000 blood cultures to identify 1 case of GBS sepsis. Even in the high risk subpopulation of 300 clinically unstable patients, we identified only 8 episodes of bacteremia with pathogenic organisms and nearly an equal number (6) of probable contaminants.
Much attention has also been placed on aspects of the CBC in identifying early onset sepsis. Many clinicians look for either the extremes of neutrophil counts (particularly neutropenia) and the immature-total neutrophil ratio, as proposed by Manroe et al., 8, 9 although the specificity of the immature-total neutrophil ratio has been questioned more recently by Schelonka et al. 10 In this study we determined a particular lack of sensitivity, as well as poor positive predictive value of either of these criteria in identifying the small number of initially asymptomatic at risk newborns who progressed to clinical sepsis, similar to the results reported by Escobar et al. 4
Perhaps the most reassuring result of this study was that essentially all of the infants who were diagnosed with clinical sepsis became ill early, within the usual newborn nursery admission time postdelivery. This finding is similar to that of Ascher et al., 11 who in their study of 30 000 births at 5 military medical centers reported that when intrapartum antibiotics failed, most infants were ill immediately after birth. A more recent study of 319 infants with group B streptococcal disease by Bromberger et al. 12 indicated that exposure to intrapartum antibiotics had no effect on the timing of illness presentation, with >95% of newborns demonstrating clinical signs of infection in the first 24 h of life. Like Ascher et al. 11 we found no newborns with asymptomatic GBS bacteremia, whereas Bromberger et al. 12 reported a small incidence of asymptomatic bacteremia (6 of 83 episodes in term infants), and both Lannering et al. 5 and Weisman et al. 6 reported rates of asymptomatic GBS bacteremia of >20%. It is difficult to interpret these differences because all these studies used different criteria to identify GBS infected infants, and none reported maternal GBS colonization rates. The recent reductions of the incidence of asymptomatic bacteremia is most likely related to an increased use of intrapartum antibiotic prophylaxis.
A limitation of this study is that we did not have the ability to detect newborns who developed symptoms after discharge at >48 h of age. Although we would have expected symptomatic newborns to have returned for evaluation and readmission to the pediatric unit or NICU at the study hospital, it is possible that some infants who became ill after discharge from the nursery could have returned to another area hospital for evaluation and treatment. Thus the results of this study are applicable only for early onset disease that presents <48 h after birth. However, the previous cohort study by Escobar et al. 4 demonstrated that presentation of early onset sepsis after 48 h is unlikely.
We conclude that although the implementation of the CDC guidelines for prevention of perinatal GBS disease has significantly reduced the rate of early onset GBS sepsis at our institution, the recommendation for a limited evaluation of asymptomatic at risk newborns including CBC and differential and blood culture add little to clinical observation in detecting and managing early onset sepsis in term or near term asymptomatic newborns.
1. Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease: a public health perspective. MMWR 1996; 45 (RR-7):1–24.
2. Schrag SJ, Zell ER, Lynfield R, et al. A population-based comparison of strategies to prevent early-onset group B streptococcal disease in neonates. N Engl J Med 2002; 347: 233–9.
3. Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease: revised guidelines from CDC. MMWR 2002; 51 (RR-11):1–22.
4. Escobar GJ, Li D, Armstrong MA, et al. Neonatal sepsis workups in infants >2000 grams at birth: a population-based study. Pediatrics 2000; 106: 256–63.
5. Lannering B, Larsson LE, Rojas J, Stahlman MT. Early onset group B streptococcal disease: seven year experience and clinical scoring system. Acta Pediatr Scand 1983; 72: 597–602.
6. Weisman LE, Stoll BJ, Cruess DF, et al. Early-onset group B streptococcal sepsis
: a current assessment. J Pediatr 1992; 121: 428–33.
7. Safier RA, Robins SH, Picone CM, Tafari N. Blood culture
screening of newborns at risk for early onset neonatal group B streptococcal disease, [Abstract 1768]. Presented at the Pediatric Academic Societies’ Annual Meeting, Baltimore, May 4 to 7, 2002. Pediatr Res 2002; 51 (Suppl): 304A.
8. Manroe BL, Rosenfeld CR, Browne R, Weinberg AG. The differential leukocyte count in the assessment and outcome of early-onset neonatal group B streptococcal disease. J Pediatr 1977; 91: 632–7.
9. Manroe BL, Weinberg AG, Rosenfeld CR, Browne R. The neonatal blood count in health and disease: I. Reference values for neutrophilic cells. J Pediatr 1979; 95: 89–98.
10. Schelonka RL, Yoder BA, desJardins SE, Hall RB, Butler J. Peripheral leukocyte count and leukocyte indexes in healthy newborn
term infants. J Pediatr 1994; 125: 603–6.
11. Ascher DP, Becker JA, Yoder BA, et al. Failure of intrapartum antibiotics to prevent culture-proven neonatal group b streptococcal sepsis
. J Perinatol 1993; 13: 212–16.
12. 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: 244–50.