Supported by an unrestricted educational grant from Wyeth-Lederle Vaccines and Pediatrics

---

Volume III, Issue 2 - March 2000

Although perinatal disease caused by GBS has been significantly reduced, much more progress is essential. It is estimated that there were 2,200 cases of early-onset and 1,400 of late-onset (>=7 days of age) infant GBS infection in 1998, resulting in 140 deaths.¹ Furthermore, because IAP strategies demand intravenous antibiotic administration to large numbers of pregnant women, concerns about the risk of serious adverse events and the development of antimicrobial resistant neonatal infections are appropriate. In part because there now are efficacious methods for the prevention of most early-onset GBS infections, the number of questions from physicians and parents actually appears to have increased. This review will provide possible responses to the questions most frequently posed by physicians and parents.

GBS colonization of the genital and/or gastrointestinal tract is frequent in women and men. The organism is a component of the "normal flora" at these sites and produces no symptoms or signs. Approximately 1 in 4 pregnant women are GBS carriers. An estimated 0.5% to 1% of infants exposed to GBS through maternal colonization develop invasive early-onset disease if their mothers do not receive appropriate IAP. Maternal factors dictating health or disease in the neonate exposed through maternal colonization are defined. They include fever during labor, chorioamnionitis, delivery before 37 weeks' gestation, rupture of membranes for more than 18 hours, and low concentrations of IgG directed against the capsular polysaccharide of the colonizing GBS strain (Fig. 2). "Heavily" colonized women also have a greater risk of giving birth to a neonate who will have invasive infection, and strain-to-strain variation in GBS virulence is presumed. While most physicians are familiar with these risk factors, it is not recognized that up to 50% of early-onset infections occur among neonates born at term to GBS colonized women without a clinically identifiable risk factor.

Properly obtained culture specimens at 35 to 37 weeks of gestation are highly predictive of the GBS colonization status at delivery (up to 100% sensitivity and specificity). The appropriate sites for culture are the lower vagina and the anorectum. Collection of specimens from a single site or from an alternate site (i.e., cervix) results in substantially lower detection rates or "false negatives." Besides inappropriate specimen collection, improper processing of specimens remains another potential explanation for the development of early-onset disease in an infant born to a "GBS-screen-negative" mother. Once separate or combined swabs specimens are sent in transport media to the laboratory, optimal processing includes placement of the swabs into commercially available, antibiotic-containing Todd-Hewitt broth medium (Lim^TM or selective broth medium); incubation of broth overnight; and subculture onto solid blood agar medium. Education about this process for all laboratories must be achieved in order to fully implement effective IAP.

It also is important to emphasize the necessity of ascertaining maternal GBS status by culture rather than by one of the rapid GBS antigen tests. None of the immunoassays for detection of GBS in vaginal swab specimens is sufficiently sensitive to exclude colonization and they are not approved by the Food and Drug Administration (FDA) for the anorectal site. The best of these assays detects "heavy" GBS vaginal colonization with 100% sensitivity but has an accuracy of only 40% for "lightly" colonized women.³ No rapid vaginal antigen test is recommended for use in screening for GBS colonization. If used at all, rapid antigen tests should be considered for women with no GBS culture screening who are admitted for rupture of membranes or early labor without risk factors. In this latter circumstance, a positive antigen test is reliable because the specificity of these tests is high (98% to 99%) but a negative test does not exclude colonization.

This has become an increasingly common question. Patient brochures describing GBS colonization, disease, and prevention are available from the CDC and ACOG and their use should be encouraged for all who provide prenatal care. The consensus guidelines state that there are two "equally acceptable methods" for the prevention of early-onset GBS disease. When the risk-based method is used, a discussion of GBS prevention by IAP usually is delayed until active labor or rupture of membranes. This timing often results in omission of a discussion. Using the risk factor-based strategy of prevention, women with one or more factors known to increase the risk for early-onset neonatal infection are offered IAP but identification of these women may occur too late to assure effective IAP. Using the culture-based approach, all women identified as carriers between 35 and 37 weeks gestation are offered IAP during labor. This encourages administration of IAP at admission for labor or rupture of membranes, a circumstance that promotes a sufficient number of penicillin G dosages before delivery to assure prevention. Although the risk-based method, unlike the culture-based method, had not been evaluated in clinical trials, it was predicted in 1996 that it would be less effective than the culture-based strategy. It now has been demonstrated that substantially more cases are prevented when the culture-based method is chosen (Table 1).

Parents begin to ask this almost as soon as the potential for death has passed. The current mortality rate from early-onset disease is 5% overall and <3% in infants born at 37 or more weeks gestation. Late-onset infection has a fatal outcome in 3% of cases.² There is great social expectation pertaining to outcomes in our current healthcare climate. It is tempting to want to reassure parents that "all will be well". Survivors of GBS disease without meningitis usually have no long-term residua of infection directly attributable to GBS. The exception is that of periventricular leukomalacia and its attendant neurologic impairment that may follow septic shock in the premature infant. There are no contemporary data regarding the long-term outcome of infants recovering from GBS meningitis. Approximately 30% of these infants evaluated in the 1970s and early 1980s had substantial neurologic sequelae.⁴ With improvements in the intensive care of sick neonates over the past 2 decades, this probably represents the worst-case estimate for outcome in the year 2000.

For infants with uncomplicated meningitis, 14 days of treatment usually is sufficient. For those with a more complicated course, 21 days is recommended. However, the optimal duration of therapy has not been determined by controlled clinical trials. It is my practice to obtain a computed tomography (CT) scan of the head with enhancement toward the end of the second week of treatment. If there are vascular complications (e.g., infarcts or cerebritis), persistent meningeal enhancement or enhancing subdural fluid collections, these indicate an ongoing inflammatory process. In this circumstance, I extend antimicrobial therapy a total of 21 days and repeat the CT scan toward the end of the third week. By then, enhancement usually has resolved. Although abscess formation is rare as a consequence of GBS meningitis, it has been described and is one of the only indications for a more prolonged course of antibiotic therapy.

Despite lack of supporting controlled clinical trial data, I recommend lumbar puncture near the 14th day of therapy. Adequate resolution of the inflammatory process is suggested by a cerebrospinal fluid (CSF) white blood cell count (WBC) of <100 cells/mm³, by a neutrophil proportion of <30%, and by return of the protein towards normal range (values as high as 100 to 150 mg/dL are acceptable). The CSF glucose frequently remains low in many infants and shouldn't provoke longer treatment. The recommendation for repeating the lumbar puncture certainly should apply to the infant with a complicated clinical course (such as prolonged fever and focal neurologic findings). If the CSF parameters suggest incomplete resolution, a third week of empiric antibiotic treatment is recommended.

It is true that an estimated 0.4% to 3% of infants experience a recurrence of invasive GBS infection. Persistent GBS colonization at mucous membrane sites, despite parenteral penicillin therapy, is one proposed source for recurrence. Recently, we treated 18 infants who had recovered from invasive GBS disease and their 17 mothers with oral rifampin for 4 days beginning 1 week after hospital discharge.⁵ Seven infants and 13 mothers were colonized before treatment and 4 infants and 7 mothers remained colonized afterwards. Although their GBS strains had uniformly low minimal inhibitory concentrations to rifampin, this agent was not bactericidal for most strains. Thus, either the lack of bactericidal activity or a suboptimal treatment schedule may explain the failure of rifampin to reliably eradicate GBS colonization.

In vitro testing of 229 isolates of GBS from a variety of patients with invasive GBS infection during an interval spanning the introduction of IAP demonstrated uniform penicillin G susceptibility (Table 2).⁶ These results support the continued use of penicillin G or ampicillin as IAP for the prevention of early-onset GBS infection. However, 7% of the isolates were resistant to erythromycin and 3% to clindamycin. Other investigators have reported resistance in up to 15% of colonizing isolates. The clinical significance of the apparent increase in resistance of GBS to erythromycin and clindamycin is unknown but it suggests caution in relying on use of these agents as alternatives for IAP in the penicillin-allergic woman. Unless the penicillin-allergy history is documented to be serious (e.g., anaphylaxis), I recommend cefazolin as the first-line alternative to penicillin G. The empiric initial dose is 2 grams intravenously followed by 1 gram every 4 hours until delivery. GBS isolates have been shown to be uniformly susceptible to first generation cephalosporins.

Few data exist to dictate the optimal duration or number of dosages that assure IAP efficacy. However, available data indicate that 1) amnionitic fluid concentrations of penicillin G sufficient to kill GBS are reached about 3 hours after a maternal intravenous dose is given; 2) transmission of GBS from a colonized mother to neonates is interrupted in 98% of cases between 2 and 4 hours after an intravenous dose of ampicillin; and 3) in one observational trial of IAP, no neonates developed early-onset disease when women received ampicillin four or more hours before delivery. Limited evaluation and observation for at least 48 hours is indicated for healthy infants of <35 weeks gestation and for those whose mothers received IAP for <4 hours (<2 dosages) before delivery (Table 3). Thus, the infant queried above has received adequate prophylaxis and only requires observation for 48 hours with no laboratory evaluation. Observation alone is indicated for infants of at least 35 weeks gestation whose mothers received IAP at least 4 hours before delivery. Limited or full evaluation should occur if an infant's clinical status suggests infection. Similarly, the second 24 hours of observation can be carried out in a home setting if it is established that this is optimal.

In some infants, late-onset infection does follow birth canal GBS exposure and it might have been predicted that interrupting vertical transmission of colonization might have reduced the incidence of late-onset disease. However, approximately 50% of late-onset infections have a nonmaternal origin. Although there is a small downward trend in the incidence of late-onset disease since the institution of IAP, this modest decline is not statistically significant.² Thus, prevention strategies other than IAP are needed. The most promising is provision of protective concentrations of IgG directed against the capsular polysaccharides of the 5 prevalent serotypes of GBS (Ia, Ib, II, III, and V) in serum at delivery through active immunization.

Progress toward development of a safe and immunogenic multivalent GBS polysaccharide-protein vaccine continues. These conjugate vaccines have completed phase 1 and 2 trials in healthy nonpregnant adults and results are very promising.⁷ The next hurdles are a corporate partner to develop GBS conjugate vaccines despite concerns about their potential use in third-trimester pregnant women and the difficulty of planning efficacy trials in the era of routine use of IAP in the US and Canada.

Morven S. Edwards, MD
Department of Pediatrics
Baylor College of Medicine
Houston, Texas

References

• Back to CLINICAL UDATES IN PEDIATRIC INFECTIOUS DISEASES
• Back to CLINICAL UPDATES