ABSTRACT
BACKGROUND: Group B streptococcal (GBS) infection contributes significantly to neonatal morbidity and mortality. In 1996, the guidelines put forward by the American College of Obstetricians and Gynecologists, the American Academy of Pediatrics, and the Centers for Disease Control and Prevention (CDC) recommended that prenatal care providers adopt either a “screening” or “risk-based” approach to guide intrapartum antibiotic prophylaxis for neonatal early-onset GBS infection. The efficacy of these 2 alternative approaches in clinical practice was compared in this population-based study.
POPULATION STUDIED: A total of 5144 birth records (including 312 cases of early-onset GBS disease) were randomly selected from 629,912 live birth records in 8 geographic areas monitored by the CDC during 1998 through 1999. The screened and risk-based groups were similar, but the risk-based group had a statistically significant greater number of Hispanics, women with inadequate prenatal care, and preterm deliveries. The screened group had a higher proportion of women with GBS bacteriuria and women with a history of previously delivering a GBS-infected neonate.
STUDY DESIGN AND VALIDITY: In this retrospective cohort study, the authors compared screening and risk-based approaches for prevention of neonatal GBS infection. At least 500 records from each CDC surveillance area were included. All GBS-infected neonates were included in the sample. The sample was stratified by surveillance area, year, and hospital, and each record was given a constant statistical weight based on the inverse of its probability of selection. This weight was adjusted to account for records without charts and for preterm births, ensuring that the number of preterm births was representative of the number in the general population. Abstractors blinded to infant GBS status gathered record information. Women who had a documented GBS culture at least 2 days prior to delivery were included in the screened group, and the remaining women were placed in the risk-based group. Univariate and multivariate analyses were used in comparing the groups, with the infant disease status as the outcome variable.
OUTCOMES MEASURED: The primary outcome measured was early-onset invasive GBS infection in neonates. Abstracters also recorded information on maternal demographics, prenatal care, GBS screening, risk factors for GBS infection, intrapartum antibiotic administration, and gestational age at birth.
RESULTS: The adjusted relative risk of neonatal infection in the screened group compared with the risk-based group was 0.46 (95% confidence interval [CI], 0.36–0.60). Intrapartum fever and a history of a previous infant with early GBS disease were the strongest predictors of GBS disease in both univariate and multivariate analyses. Women in the screened group who had positive culture results were much more likely to receive antibiotic prophylaxis than women in the risk-based group who had risk factors (89% vs 61%, P < .001). However, even with projected perfect administration in the risk-based group, the incidence of GBS disease would still be less in the screened group (0.32 vs 0.44 per 1000 live births). In addition, with perfect implementation of either preventive strategy, the anticipated overall rate of antibiotic use would be similar (31% in the screened group vs 29% in the risk-based group).
Screening is clearly more effective than risk stratification in preventing early-onset neonatal GBS infection. Universal screening for GBS should be adopted. The CDC recently issued updated guidelines that advocate universal screening.1