Neonatal Sepsis Caused Simultaneously by Two Serotypes of Group B Streptococcus

Group B Streptococcus (GBS) has been a frequent cause of neonatal sepsis and meningitis for more than two decades. Recently there have been some changes in the epidemiology of GBS infections. The incidence of early onset disease has declined after publication of the recommendations regarding maternal intrapartum chemoprophylaxis,1-3 and GBS infections are less frequently fatal (estimated case-fatality ratio, 5%)4 than formerly. New serotypes of GBS have been identified, notably type V in the United States5-7 and serotypes VI and VIII as colonizing strains among pregnant women in Japan.8 Also a shift in the distribution of serotypes causing early onset infection has been reported.9, 10 The serotypes of GBS are based on distinct capsular polysaccharide antigens and, to date, nine are recognized (serotypes Ia, Ib, II to VIII). Among neonates with early onset GBS disease, serotypes Ia and III account for ∼65% of isolates, and type V is third in frequency at an estimated 15%.3, 9 Although colonization of the female genital tract with more than one serotype of GBS has been noted,11 neonatal sepsis caused by two GBS serotypes has not been described previously. We describe a term newborn infant with early onset sepsis and isolation of serotypes II and V GBS from the blood. Case report. A male infant weighing 2715 g was born in January, 1998, at 39 weeks of gestation, to a 23-year-old African-American woman who had had no prenatal complications. Maternal cultures 2 weeks before delivery did not grow GBS (sites and microbiologic methods not known). The membranes ruptured spontaneously ∼7 h before vaginal delivery. Ten minutes before delivery the maternal temperature was 102.4° F. Intravenous ampicillin and gentamicin therapy was started immediately postpartum, and within 24 h the mother was discharged home to complete a 7-day course of oral antimicrobial therapy. At birth the infant was tachypneic with a respiratory rate of 70 to 80/min and had intercostal retractions. After blood was obtained for culture and complete blood count, intravenous ampicillin (300 mg/kg/day) and gentamicin (2.5 mg/kg/dose) were initiated. The white blood cell count was 12 580/mm3 with 36% polymorphonuclear leukocytes, 52% lymphocytes and 12% monocytes. The blood culture grew beta-hemolytic streptococci that were serologically confirmed as GBS. On Day of Life 3 cerebrospinal fluid revealed 10 WBC/mm3 (12% polymorphonuclear leukocytes), a glucose of 57 mg/dl, protein of 51 mg/dl and a Gram stain that showed no organisms. The cerebrospinal fluid culture was sterile as was a blood culture obtained after 36 h of antimicrobial therapy. On the sixth day of treatment penicillin G alone (200 000 units/kg/day) was initiated to complete a total of 10 days of therapy, and recovery was uneventful. Two days after discontinuation of therapy, swabs from the throat and rectum of the infant and from the lower vagina and rectum of the mother were obtained for isolation of GBS. The same sites from the infant again were cultured 7 days later. Microbiologic studies. Swabs from the infant's throat and rectum and the mother's vagina and rectum were placed into transport media (Amies' medium; Difco, West Molesey, UK) and later into selective broth medium containing nalidixic acid (15 μg/ml) and gentamicin (8 μg/ml) (Becton Dickinson, Cockeysville, MD). The broth was incubated overnight at 35°C and subcultured onto 5% sheep blood agar medium (Becton Dickinson Microbiology Systems, Cockeysville, MD). Both sets of infant cultures grew GBS, but the maternal cultures, obtained 2 days after completion of oral antibiotic therapy, were negative. Minimal inhibitory concentration and minimal bactericidal concentration of penicillin for types II and V of invasive and colonizing isolates was identical at 0.015 μg/ml. The infants blood isolate was serotyped by the capillary precipitin technique12 with the hyperimmune rabbit antisera prepared with prototype GBS strains 090 (Ia), H36B (Ib), 18RS21 (II), M732 (III) and CNCTC 10/84 (V). Strong reactions with types II and V antisera were noted with initial and confirmatory testing. The original stock culture of the blood isolate that had been frozen in Todd-Hewitt broth with 10% glycerol was thawed, streaked onto 5% sheep blood agar medium and incubated overnight at 35°C. The colonial morphology was typical of GBS and was uniform in character. Ten individual colonies were selected for serotyping. Five of these reacted with type II, four with type V and one with both type II and V antisera. GBS isolated from the throat and rectum cultures also were serotyped. Extracts from both sets of throat cultures reacted with type II antisera, whereas both sets of rectal isolates reacted with type V. Discussion. Neonatal sepsis with isolation of GBS from blood cultures is common, but concomitant bloodstream invasion by two distinct serotypes has not been previously described. Because bacteremia in our patient was documented within 1 h of birth, the most plausible pathogenesis for this early onset infection was exposure to both GBS serotypes through aspiration of infected amniotic fluid with subsequent bacteremia. The mother had intrapartum fever and clinical amnionitis, but neither blood nor amniotic fluid cultures were obtained before initiation of broad spectrum antibiotic therapy. Later attempts to document maternal GBS colonization with dual GBS serotypes postpartum were unsuccessful, either because her antibiotic therapy eradicated GBS13 or because this therapy diminished the inoculum at the time of culture assessment. Additional evidence that the mother was the source for the infant's infection was isolation of distinct serotypes from different mucosal sites (throat, type II; rectum, type V). Vaginal colonization of women with more than one serotype of GBS has been observed,11 but infrequently (estimated at <1%). Both serotypes of GBS that were isolated from our patient can cause invasive neonatal sepsis.9 Type V has emerged in the past decade as a frequent cause of invasive GBS infections.6, 7, 9, 10 It is estimated to cause ∼15% of neonatal disease,7, 9 and an even greater proportion of invasive infections in pregnant women and nonpregnant adults.7, 10 Serotypes II and V are infrequently associated with focal GBS infections in infants, especially meningitis, suggesting limited virulence for the neonate.14 This observation and the prompt institution of antimicrobial therapy may have accounted for this infant's benign hospital course and complete recovery. The isolation of distinct GBS serotypes from the blood is rare. However, serotyping is not a routine laboratory procedure, and most research laboratories that perform serotyping use methods that are not designed for optimal detection of this phenomenon. Typically only one GBS colony on a blood agar plate is selected for serotyping.9, 12, 15 The observation that some women may be colonized with more than one GBS serotype derived from a procedure in which 10 GBS colonies from each vaginal specimen subculture were selected for serotyping.11 However, in our laboratory the same technician has serotyped ∼250 invasive GBS isolates since 1992 and dual serotype infection has not been previously observed. The development of more practical and precise laboratory methods would be required to determine the true incidence of colonization and infection by more than one GBS serotype. Marisol Fernandez, M.D. Marcia A. Rench, B.S.N., R.N. Carol J. Baker, M.D.