Epidemiology of neonatal group B streptococcal disease in the Netherlands before and after introduction of guidelines for prevention

OBJECTIVES: (1) To describe the epidemiology of neonatal group B streptococcal (GBS) disease over five years (1997-2001) in the Netherlands, stratified for proven and probable sepsis and for very early (<12 h), late early (12 h - <7 days) and late (7-90 days) onset sepsis. (2) To evaluate the effect of the introduction in January 1999 of guidelines for prevention of early onset GBS disease based on risk factors. METHODS: Data on cases were collected in collaboration with the Dutch Paediatric Surveillance Unit and corrected for under-reporting by the capture-recapture technique. RESULTS: Total incidence of proven very early onset, late early onset and late onset GBS sepsis was 0.32, 0.11 and 0.14 per 1000 live births, respectively, and of probable very early onset, late early onset and late onset GBS sepsis was 1.10, 0.18 and 0.02 per 1000 live births, respectively. Maternal risk factors were absent in 46% of the proven early onset cases. Considerably more infants with proven GBS sepsis were boys. 64% of the infants with proven very early onset GBS sepsis were first born compared with 47% in the general population. After the introduction of guidelines the incidence of proven early onset sepsis decreased considerably from 0.54 per 1000 live births in 1997-8 to 0.36 per 1000 live births in 1999-2001. However, there was no decrease in the incidence of meningitis and the case fatality rate in the first week of life. The incidence of late onset sepsis also remained unchanged. CONCLUSION: After the introduction prevention guidelines based on risk factors there has been a limited decrease in the incidence of proven early onset GBS sepsis in the Netherlands. This study therefore recommends changing the Dutch GBS prevention guidelines.     

Fatal late onset group B streptococcal meningitis following maternal postpartum sepsis

Although maternal screening and the administration of prophylactic intrapartum antibiotics have decreased the incidence of early onset group B streptococcal (GBS) disease in neonates, there is still significant morbidity and mortality as a result of neonatal GBS disease.Maternal GBS infections are not uncommon, but with appropriate therapy there is almost a uniformly good outcome. Little is written about the appropriate management of well infants born to mothers with postpartum GBS sepsis.The question of whether well infants born to mothers with GBS puerperal sepsis should be treated empirically with antibiotics and the lack of literature concerning this issue became apparent when an untreated term infant died of late onset GBS meningitis following maternal puerperal GBS sepsis. We describe this event in the following case presentation.With the current paucity of literature regarding the management of well infants born to mothers with postpartum GBS sepsis, it seems prudent to treat such infants empirically with antibiotics (following a full septic work-up) until this matter has been investigated further.     

Changing etiology and outcome of neonatal septicemia in Riyadh, Saudi Arabia

To study the etiology of neonatal septicemia and factors associated with outcome, all charts of neonates with bacteremia and clinical sepsis admitted to a neonatal unit in Saudi Arabia, from 1 November 1980 to 31 October 1984 were reviewed. The results were compared to a previous study period in the unit (1 November 1976-31 October 1980). Septicemia was diagnosed on 50 occasions in 49 neonates. The incidence of neonatal sepsis among patients born in the hospital was 2.5/1,000 live births. Mortality from sepsis was 33% and was associated with neutropenia in 63%. The most commonly isolated bacteria were E. coli, Klebsiella and Staphylococcus aureus. Salmonella enteritidis serotypes were isolated in 4% of the cases. Group B streptococci (GBS) were isolated, for the first time, from blood of 3 neonates. Salmonella species were less frequently and GBS more often isolated than previously. GBS have now appeared as etiologic organisms in neonatal sepsis also in Saudi Arabia. Salmonella septicemia remains more common in Saudi Arabia than in the West.     

Neonatal bacterial sepsis in a neonatal intensive care unit: A 5 year analysis

Objective : To study the pattern of neonatal sepsis in a neonatal intensive care unit (NICU) during a 5 year period and assess the relationship between maternal risk factors and early onset sepsis (EOS).
Methodology : The study reported here was a retrospective analysis of 209 episodes of septicaemia and 5 episodes of bacterial meningitis in 198 newborn infants, 22 of whom died. Eighty-one infants had EOS (≤72h) and 117 infants had late onset sepsis (LOS >72 h). All infants had clinical evidence of sepsis, a computerized haematological score for sepsis of 4 or greater, and either treatment with antibiotics for 7 days or more or had earlier death due to sepsis. The organisms causing neonatal sepsis were analyzed according to the day of onset, gestational age, birthweight and year of infection.
Results : Sepsis occurred in 5.6 per 1000 live births and 3.8% of NICU admissions. There were 81 episodes of EOS and 128 of LOS. Coagulase negative staphylococci (CONS) 38.8%, group B Streptococcus (GBS) 20.1% and Gram-negative bacilli (GNB) 20.1% were the common causes of sepsis; and GBS (50.6%) and CONS (60.9%) were the most common organisms in EOS and LOS, respectively. The mean gestational age and birthweight were heigher in babies with EOS than compared with LOS. The higher likelihood of probable rather than definite infection in infants with EOS was related to more mothers in the EOS group receiving intrapartum antibiotics. GNB infection was more common in their babies.
Conclusions : GBS and CONS were the most common causes of EOS and LOS, respectively. The use of maternal intrapartum antibiotics interferes with neonatal blood culture results. Because blood cultures are not always positive in neonatal septicaemia, a combination of clinical, haematological and other microbiological evidence should be used when diagnosing neonatal septicaemia.     

Bacterial profile of sepsis in a neonatal unit in south India

OBJECTIVE: To study the pattern of sepsis in a neonatal unit in south India and assess the influence of maternal factors on early onset sepsis (EOS). DESIGN: Prospective survey from 1995-1996. SETTING: Medical College Hospital. SUBJECTS: All inborn babies who had clinical signs of sepsis or were born to mothers with potential risk factors for infection were screened for sepsis. Neonatal septicemia was defined as a disease of infants who were younger than 1 month of age, were clinically ill, and had positive blood cultures. RESULTS: Among 13,367 live births in the study period, there were 131 episodes of neonatal septicemia among 125 newborn infants, 18 (14.4%) of whom died. Thirty (24%) had EOS (< or = 48 hours) and 95 (76%) had late onset sepsis (LOS) (> or = 48 hours). Sepsis occurred in 9.8 per 1000 livebirths and 4.4% of all nursery admissions. E. coli and E. fecalis were the predominant organisms causing EOS, while Klebsiella and E. fecalis were the predominant organisms in LOS. The mean gestational age (GA) and birth weight (BW) of babies with EOS was significantly higher than those with LOS. Maternal factors significantly associated with EOS were meconium staining of liquor and multiple vaginal examinations. CONCLUSIONS: The incidence of neonatal bacterial sepsis is 9.8 per 1000 livebirths. E. coli and Klebsiella were the most common organisms causing EOS and LOS, respectively. E. fecalis was also a major pathogen, both in EOS and LOS.     

Significance of a positive urine group B streptococcal latex agglutination test in neonates

We assessed the clinical significance of a reactive urine latex agglutination (LA) test in neonates without bacteriologically confirmed group B streptococcal (GBS) infection. In a retrospective review of a 3 1/2-month period, during which 367 urine specimens from newborn infants evaluated for suspected sepsis were tested by LA, 25 infants (6.9%) with sterile blood cultures but positive urine LA test results were compared with a control group of 112 infants with both blood cultures and urine LA test results negative for GBS. When the data were studied with stepwise discriminant analysis, the only variables significantly associated with a positive urine LA test result were immature to total neutrophil ratios greater than or equal to 0.16 at 0 and 12 hours. The influence of mucosal GBS colonization on urine LA test results was then investigated prospectively in 98 healthy infants (83 born to mothers colonized with GBS and 15 born to mothers with negative GBS cultures). Eight (8.2%) of the infants studied, or 8 of 52 (15.4%) infants colonized with GBS, had a positive urine LA test result. GBS was isolated from urine cultures of all infants with a positive urine LA test result. A positive urine LA test result was associated with positive GBS rectal and vaginal cultures and with increased density of colonization at those sites. We conclude that contamination of bag specimens of urine with GBS from perineal and rectal colonization may produce a positive urine LA test result in an infant with no systemic sign of infection.     

Neonatal early onset Escherichia coli sepsis: trends in incidence and antimicrobial resistance in the era of intrapartum antimicrobial prophylaxis

BACKGROUND: Although intrapartum antimicrobial prophylaxis has lowered the incidence of early onset group B Streptococcus (GBS) sepsis, there are concerns that the increased use of antibiotics may raise the incidence of non-GBS antimicrobial-resistant infections. The objective of this study was to determine trends in the incidence and antimicrobial resistance of early onset sepsis caused by Escherichia coli in the era of antimicrobial prophylaxis. METHODS: All neonates with early onset E. coli infection who were born at La Paz Hospital, Madrid, from January 1, 1992, through December 31, 2002, were identified from a microbiologic register of all neonatal infections. To evaluate the effect of the guidelines for GBS prevention, data were pooled and compared for: 1992 through 1995 (Period 1); 1996 through 1998 (Period 2); and 1999 through 2002 (Period 3). RESULTS: Early onset E. coli infection was diagnosed in 41 of 84 612 live births. The infection rate did not change significantly during the 3 time periods (0.56, 0.24 and 0.55 per 1000 during Periods 1, 2 and 3, respectively; P = 0.936, linear-by-linear association). The proportion of E. coli infections that were resistant to ampicillin increased significantly among preterm infants, from 25% (1 of 4) in Period 1, to 100% (2 of 2) in Period 2 and to 91% (10 of 11) in Period 3 (P = 0.017, linear-by-linear association), but not among term infants, with 67% (8 of 12) in Period 1, 50% (1 of 2) in Period 2 and 44% (4 of 5) in Period 3 (P = 0.317, linear-by-linear association). CONCLUSIONS: Although the incidence of early onset sepsis caused by E. coli remained stable during the study period, antibiotic-resistant E. coli infections increased among preterm infants. On the whole these trends are reassuring with respect to GBS prophylaxis. However, the increase in the proportion of ampicillin-resistant infections in preterm infants suggests that continuing evaluation of the risks and benefits of prophylaxis in this group is critical.     

Changing Etiology and Outcome of Neonatal Septicemia in Riyadh, Saudi Arabia

ABSTRACT. To study the etiology of neonatal septicemia and factors associated with outcome, all charts of neonates with bacteremia and clinical sepsis admitted to a neonatal unit in Saudi Arabia, from 1 November 1980 to 31 October 1984 were reviewed. The results were compared to a previous study period in the unit (1 November 1976-31 October 1980). Septicemia was diagnosed on 50 occasions in 49 neonates. The incidence of neonatal sepsis among patients bom in the hospital was 2.5/1000 live births. Mortality from sepsis was 33% and was associated with neutropenia in 63%. The most commonly isolated bacteria were E. coli , Klebsiella and Staphylococcus aureus. Salmonella enteritidis serotypes were isolated in 4% of the cases. Group B streptococci (GBS) were isolated, for the first time, from blood of 3 neonates. Salmonella species were less frequently and GBS more often isolated than previously. GBS have now appeared as etiologic organisms in neonatal sepsis also in Saudi Arabia. Salmonella septicemia remains more common in Saudi Arabia than in the West.     

Ten-year study on the effect of intrapartum antibiotic prophylaxis on early onset group B streptococcal and Escherichia coli neonatal sepsis in Australasia

BACKGROUND: Intrapartum antibiotics have reduced the incidence of neonatal early onset (EO) group B streptococcal (GBS) disease. Some surveillance data suggest that this success may be at the cost of increasing rates of non-GBS infection, especially in premature neonates. OBJECTIVE: To examine rates of EOGBS infection and EO Escherichia coli neonatal sepsis in Australasia. METHODOLOGY: Analysis of trends in EO (<48 h age) GBS and E. coli sepsis from longitudinal prospective surveillance data collected from representative tertiary obstetric hospitals in each state of Australia and selected centers in New Zealand during a 10-year period from 1992 through 2001. Statistical analysis used Poisson regression. RESULTS: 206 GBS and 96 E. coli cases occurred in 298,319 live births during the study period. The EOGBS sepsis rate fell from a peak of 1.43/1000 live births in 1993 to 0.25/1000 in 2001 (P < 0.001). The overall EO E. coli sepsis rate was 0.32/1000. In babies with birth weight <1500 g, it was 6.20/1000. There was an overall trend to decreasing EO E. coli sepsis (P = 0.07), and there was no significant change in E. coli sepsis in babies <1500 g (P = 0.60). Sixty-nine percent of E. coli cases occurred in the <1500 g cohort; the case fatality rate in this group was 50%. The overall case fatality rate from E. coli sepsis was 36%, and this rate remained stable during the study period (P = 0.47). CONCLUSIONS: The increasing use of intrapartum antibiotics produced a steady decline in EOGBS disease in Australasia. There was also a trend to decreasing EO E. coli sepsis in all babies, and the rate in very low birth weight infants remained stable.     

Early-onset group B streptococcal sepsis: a current assessment.

Group B streptococcus (GBS) is a common cause of early-onset sepsis in neonates. The most recent reviews describing incidence, diagnosis, treatment, and outcome evaluated data on patients from the early 1980s. To obtain current information about this disease, we retrospectively evaluated data on neonates with GBS early-onset sepsis from nine hospitals in the United States between Jan. 1, 1987, and Dec. 31, 1989. There were 245 infants with GBS bacteremia identified among 61,809 live births, resulting in an incidence of 0.32%. Ninety-six infants (39%) were preterm (less than 38 weeks of gestational age). Maternal risk factors for infected preterm and term infants were similar. Antibiotics were administered during parturition in 10% of infants with bacteremia. Mothers of preterm infants received antibiotics up to 48 hours before delivery; mothers of term infants received antibiotics less than 4 hours before delivery. All preterm infants with bacteremia had symptoms; 22% of term infants with bacteremia had no symptoms. Group B streptococcal meningitis was confirmed in 6.3% of infants. Although 86% survived, GBS sepsis increased the birth weight-specific mortality rate up to eightfold in preterm infants and more than 40-fold in term infants. Although the incidence of GBS early-onset sepsis is not changing, we speculate that the improved birth weight-specific survival rate and the changing clinical presentation are due to improved intrapartum and neonatal management.     

Placental pathology compared with clinical outcome: a retrospective blind review

The usefulness of the microscopic examination of the placenta, associated membranes, and umbilical cord was tested in a retrospective clinical review. Fifty-nine patients with inflammation were matched by sex, race, and gestation with 59 patients without inflammation. Blind review of the clinical course of these infants revealed five cases of culture-positive septicemia, 28 cases of probable sepsis, 39 cases of possible sepsis, and 46 normal infants. The clinical categorization was significantly correlated with the microscopic appearance of the placenta, membranes, and cord. Triple vessel vasculitis in the umbilical cord vessels and chorionic microabscesses were significantly related to the incidence of proven, probable, and possible clinical sepsis. The microscopic examination of the umbilical cord and placenta provides a useful, but not infallible, tool in the evaluation of sepsis in the newly born infant.     

Neonatal Group B Streptococcal bacteraemia in India: ten years'experience

Group B Streptococcus (GBS) is an infrequent cause of neonatal septicaemia in many developing countries. In a perinatal centre in India with 60,119 live births between 1988 and 1997, GBS was isolated from blood cultures of 10 babies. Thus the incidence of GBS bacteraemia was 0.17 per 1000 live births. Lethargy, respiratory distress and poor perfusion were the presenting features in eight symptomatic babies. Two babies had meningitis, three required ventilatory support and one died. There were no cases of late onset disease. The low incidence could be due to the low rate of colonisation and high prevalence of protective antibody in the mothers.     

Infection and neonatal meningitis: epidemiology, pathogen spectrum, therapy

Data from 196 infants were analyzed who had been treated for neonatal septicemia and/or meningitis between 1962-1974 (n = 88) and 1975-1985 (n = 108). In addition to an increase in the incidence of septicemia (1962-1974: 0.88 cases/1000 live births/year; 1975-1985: 1.8 cases/1000 live births/year) there was also a change in the pattern of infection. Group B streptococcal infections were first observed in 1975. Infections with Escherichia coli increased (1962-1974: 0.25 cases/1000 live births/year; 1975-1985: 0.65 cases/1000 live births/year). Although the incidence of meningitis was similar in both periods (1962-1985: 0.45 cases/1000 live births/year) the relative number of cases declined (1962-1974: 51 of 88 patients; 1975-1985 25 of 108). Mortality also decreased during the second period (1962-1974: 53%; 1975-1985: 29%). All infants were primarily treated with a combination of ampicillin and gentamicin. The decision to discontinue this therapy was based on the clinical course of the patient and the results of culture and susceptibility studies. Ampicillin and/or gentamicin were effective in vitro against all microorganisms which caused septicemia and/or meningitis within the first four days of life. In contrast this antimicrobial combination was only active in vitro against 77% of the pathogens isolated after this time period.     

新生儿B族链球菌感染现状的多中心前瞻性研究

目的 了解孕妇B族链球菌（group B Streptococcus,GBS）定植及新生儿早发型GBS疾病（early-onset GBS disease,GBS-EOD）的发生情况,并探讨GBS定植孕妇的子代发生GBS-EOD的影响因素。 方法 前瞻性纳入2019年5月1日至2020年4月30日在厦门市妇幼保健院、首都医科大学附属北京妇产医院和漳州正兴医院建档的16 384例孕妇及其分娩的16 634例新生儿作为研究对象。各研究中心采用统一的GBS筛查时间、培养方法和产时抗生素预防（intrapartum antibiotic prophylaxis,IAP）适应证,调查孕妇GBS定植率和新生儿GBS-EOD发生率,并应用多因素logistic回归分析评估GBS定植孕妇的子代发生GBS-EOD的影响因素。 结果 3所医院妊娠晚期孕妇GBS培养阳性率为11.29%（1 850/16 384）,新生儿GBS-EOD发生率为0.96‰（16/16 634）。GBS阳性组孕妇分娩活产儿收住院率高于GBS阴性组（P<0.05）。GBS阳性组孕妇分娩活产儿的GBS-EOD发生率［6.38‰（12/1 881）］高于GBS阴性组［0.27‰（4/14 725）］（P<0.05）。多因素logistic回归分析显示,胎盘拭子培养GBS阳性和新生儿出生时胃液培养GBS阳性是GBS定植孕妇子代发生GBS-EOD的独立预测因素（P<0.05）,而充分IAP为保护因素（P<0.05）。 结论 孕妇妊娠晚期GBS定植对其子代有不良影响。明确孕妇产前GBS定植状态,并依据IAP适应证给予充分IAP是减少其子代发生GBS-EOD的重要措施。 引用格式:     

Early-onset neonatal group B streptococcal infections in New Zealand 1998-1999

OBJECTIVE: To determine in New Zealand infants the attack rates, risk factors, preventive policies, strain serotype and antibiotic susceptibilities of early-onset neonatal group B streptococcus (GBS) infection. METHOD: A 2-year prospective active surveillance study was conducted in New Zealand's 19 neonatal units. Cases had to present within 48 h of delivery, be unwell, possess abnormal haematological indices and have GBS isolated from sterile sites. RESULTS: Of the 112 402 infants born in New Zealand during 1998-1999, 56 had early-onset GBS infection, an attack rate of 0.5 per 1000 live births (95% confidence interval [CI] 0.38, 0.65). Seven had meningitis and there was one death (case fatality rate of 1.8%; upper 95% CI 9.5%). Univariate analysis identified young maternal age, parity, preterm labour, prolonged membrane rupture, maternal fever and assisted delivery as risk factors. Preventive policies for GBS were reported by 14 (74%) obstetric centres associated with neonatal units. Of the 56 cases, five (9%) were born to mothers receiving intrapartum antibiotics, 32 (57%) had mothers with risk factors but were not treated with antibiotics, and 19 (34%) were born to mothers without identifiable risk factors for GBS prevention. Serotypes Ia and III predominated, while two isolates were resistant to erythromycin and/or clindamycin. CONCLUSIONS: Rates of early-onset GBS infection are similar to other countries following the introduction of prevention policies. Further reductions are possible with full implementation of these guidelines. Meanwhile, emergence of antibiotic resistance complicates the management of women with penicillin allergy. Vaccine development therefore remains a priority.     

Group B streptococcal carriage and disease: a 6-year prospective study

A prospective study of group B streptococcal (GBS) carriage and disease was conducted over 6 years. Carriage rates at delivery for mothers and infants were 20% and 12%, respectively. Forty-five cases of GBS disease occurred in infants, 24 "early-onset" disease and 21 "late-onset" disease. The combined attack rate for early and late disease was 3.3 per 1000 live births over the 6 years. The rate of early-onset disease was highest in infants found to be heavily colonized at birth: 50 per 1000 live births. Twenty-three of 24 had evidence of intrauterine-acquired infection. All GBS serotypes were represented. Preterm delivery, prolonged labor, premature rupture of membranes, and maternal infection enhanced the risk of early disease. Septicemia was the predominant form of late-onset disease (15 of 21 cases); GBS type III accounted for 19 of 21 cases. Ten of 21 infants with late infections were colonized at birth with the GBS type that subsequently caused disease. Thus a maternal source of infection was identified in 34 of the 45 infants. These data reveal consistent year-to-year carriage and disease rates in the study population.     

Systemic group B streptococcal disease in neonates and young infants in Norway 1985-94

In the period 1985-94, 237 out of 575 248 (0.41 per 1000) live born infants in Norway were reported to suffer culture-confirmed systemic group B streptococcal disease before their 90th day of life. The annual incidence increased from 0.20 per 1000 live births in 1985 to 0.64 in 1994, due solely to an increase in cases with an onset before the seventh day of life. Future studies should address the possible causes of this increase.     

Sepsis in neonatal intensive care in the late 1990s

AIM: To determine the incidence and clinical characteristics of sepsis in ventilated infants from an Australian neonatal intensive care unit (NICU) in the late 1990s. METHODS: Demographic data was collected from babies requiring assisted ventilation (AV) over the 6-month period from 1 July to 31 December 1998. Sepsis was divided into early onset sepsis (EOS; " 72 h of age) and late onset sepsis (LOS; >72 h of age), including both definite (culture-proven + abnormal markers) and probable (culture negative + abnormal markers) episodes. RESULTS: Two hundred and eleven babies required AV over this period. Of these, 64 (30.3%) had at least one infection, with 85 episodes of sepsis (40.3 episodes per 100 admissions requiring AV). There were 22 babies with 22 episodes of EOS, and 45 with 63 episodes of LOS. Three babies had both EOS and LOS. The rate of EOS was 10.4 infected infants (10.4 infections per 100 admissions requiring AV). The rate of LOS was 21.3 infected infants (29.9 infections per 100 admissions requiring AV). The rates of both EOS and LOS were higher than previously reported by Australian studies in the early 1990s. In both EOS and LOS, risk factors for infection were common. Group B streptococcus was the commonest cause of definite EOS. The mortality rate from sepsis in the EOS group was 14% (3/22). Coagulase-negative staphylococci were the commonest cause of LOS. The mortality rate from sepsis in the LOS group was 11% (5/45). CONCLUSIONS: EOS and LOS are significant problems in ventilated NICU infants in the late 1990s.     

孕妇B族链球菌定植及其早产儿感染状况的研究

目的探讨孕妇B族溶血性链球菌（GBS）定植及其分娩早产儿的GBS感染状况,评估早产儿GBS定植的危险因素。方法采用前瞻性队列研究方法,纳入2017年1月至2018年1月分娩的859例早产孕妇作为研究对象。入院时采集孕妇阴道下段1/3和直肠拭子行GBS培养,其中515例行实时PCR GBS DNA检测。采集所纳入孕妇分娩的早产儿的口咽分泌物、胃液或血液进行GBS培养。取孕妇外周血及其分娩的早产儿脐血测定抗GBS荚膜多糖抗体水平。调查早产儿GBS感染情况和影响定植的围产因素。结果 859例孕妇阴道、直肠GBS培养阳性率为14.8%（127/859）。515例GBS DNA检测的阳性率为15.1%（78/515）。859例孕妇共分娩活产早产儿976例,其中43例（4.4%）GBS培养阳性;4例发生早发型GBS疾病,其中2例肺炎,2例早发型GBS败血症。127例GBS阳性孕妇分娩的127例早产儿中,34~<37周早产儿组GBS阳性率明显低于<34周早产儿组（P=0.013）,抗GBS荚膜多糖抗体水平明显高于<34周早产儿组（P=0.001）。多因素logistic回归分析显示胎膜早破>18 h和绒毛膜羊膜炎是早产儿GBS定植的独立危险因素（分别OR=6.556、6.160,均P < 0.05）。结论早产儿GBS阳性率及抗GBS荚膜多糖抗体水平与胎龄相关。胎膜早破>18 h和绒毛膜羊膜炎可增加早产儿GBS定植的风险。     

Failure of rifampin to eradicate group B streptococcal colonization in infants

BACKGROUND: Mucous membrane colonization with group B streptococci (GBS) frequently persists in infants after treatment of invasive infection and may be associated with recurrent disease. OBJECTIVE: To determine the frequency with which GBS colonization persists at mucous membrane sites after treatment of invasive early or late onset infection and to determine the efficacy of oral rifampin in eradicating colonization in these infants and their mothers. METHODS: Cultures for isolation of GBS were obtained from infants and their mothers after completion of the infant's parenteral therapy, 1 week later when rifampin therapy was initiated and at approximately 1 and 4 weeks after completion of rifampin therapy. Rifampin was administered (10-mg/kg dose; maximum, 600 mg) twice daily for 4 days. RESULTS: Ten of 21 infants (48%) and 13 (65%) of their 20 mothers were colonized with GBS at throat or rectal (infant) or vaginal, rectal or breast milk (mother) sites before rifampin was initiated. One week or less after rifampin treatment, 7 (70%) infants and 4 (31%) mothers remained colonized with GBS. At study completion 6 infants and 7 mothers had GBS colonization. Persistent colonization was not related to GBS serotype, to initial rifampin minimal inhibitory concentration or to the development of rifampin resistant strains. CONCLUSIONS: Rifampin treatment for four days utilized as a single agent after completion of parenteral therapy failed to reliably eradicate GBS colonization in infants.