Department of Family Practice Journal Club

1
Department of Family Practice Journal Club

• Theresa A. Allison, M.D.
• September 4, 2002

2
A Population-Based Comparison of Strategies to
Prevent Early-Onset Group B Streptococcal Disease
in Neonates

• Schrag, S.J. Zell, E.R. Lynfield, R. Roome,
  A. Arnold, K.E., Craig, A.S. Harrison, L.H.
  Reingold, A. Stefonek, K. Smith, G. Gamble, M.
  and Schuchat, A. for the Active Bacterial Core
  Surveillance Team.New England Journal of
  Medicine, Volume 347/4 (July, 25, 2002) 233-239.

3
Background

• 1996 ACOG, AAP and CDC guidelines for intrapartum
  Group B Strep (GBS) prophylaxis recommend either
  
• Risk based treatment, or
• Routine screening at 35-37 weeks
• Note that, as of March, 2002, the CDC is revising
  their recommendations.

4
Risk-based treatment

• Clinical Group B Streptococcal risk factors
• Prior history of infant with GBS sepsis
• Intrapartum fever (gt38 degrees C)
• Prolonged (gt18 hours) rupture of membranes
• Preterm delivery (lt37 weeks GA)
• GBS bacteriuria

5
Routine screening

• Looks for GBS colonization in all women,
  regardless of risk factors
• Vaginal/Rectal swab cultures obtained between 35
  and 37 weeks

6
Study Design

• Multistate, retrospective, cohort study
• Comparison of screening approach to early-onset
  GBS sepsis prevention versus risk-based approach
• Univariate and multivariate models used to
  determine relative risk in screened group
  relative to risk-based group

7
Validity

• Users Guides to the Medical Literature, part
  IV, How to Use an Article about Harm. Levine, M.
  et al. JAMA 271/20 (May 25, 1994) 1615-1619

8
Population

• Stratified, random sample of 629,912 live births
  (1998-9)
• Eight geographical areas
• 5,144 births (sample size)
• 312 neonates with GBS sepsis (all cases)

9
Weighting of cases

• Statistical weight assigned to each birth
• Weight 1/probability of selection
• Early-onset invasive GBS infection, weight1
• Nonresponse (i.e. chart unavailable) assumed
  weight was representative of stratum
• Wieghting to adjust for birth hospital,
  surveillance area and year
• Further adjusted to reflect incidence of preterm
  births in overall population

10
Data Collection

• Abstraction from LD records
• Demographics of mother
• Screening for GBS
• Clinical risk factors
• Intrapartum antibiotic use
• Gestational age of infant (Birth registry data)
• Abstracters blinded to infection status of infants

11
Analysis

• Relative risk for the infant of acquiring
  early-onset invasive GBS infection.
• Comparison of screening approach versus
  risk-based approach.
• Dependent variable disease status of infant.
• Independent variables screening, risk factors,
  potential confounders.

12
Further analysis

• All women not screened were initially included in
  the risk-based group, RR 0.46 (0.36-0.60).
• In order to adjust for women whose providers may
  have had no GBS prevention strategy, the authors
  then excluded the 207 patients who had risk
  factors but did not receive antibiotics.
  Adjusted RR0.48 (0.37-0.61).

13
Results

Table 2. Factors Associated with Early-Onset
Group B Streptococcal Disease in the Univariate
Analysis.
14
Efficacy

• Efficacy (1-relative risk) assessed using
  antibiotic prophylaxis as independent variable.
• Among screened women with no risk factors, the
  efficacy of antibiotics in preventing infection
  was 88.6 (66.4-96.1).
• Projected comparison of perfect implementation of
  risk-based approach suggests an absolute risk
  reduction from 0.5 per 1000 to 0.44 per 1000 live
  births.
• Study showed absolute risk for screened patients
  of 0.32 per 1000.

15
Conclusions

• Routine screening for Group B streptococcus
  during pregnancy prevents more cases of
  early-onset disease than the risk-based approach
  (Schrag et al, 2002233).
• Screening reduced by 54 (40-64 within 95
  confidence interval) the incidence of early-onset
  GBS in the multivariate analysis.

16
Conclusions

• Protective effect stems from
• Identification of risk factor-free GBS carriers
  (18 of women in the screened group).
• The fact that this cohort was more likely to
  receive antibiotics (89 vs 50-79 for different
  clinical risk factors).

17
Implication for Our Practice

• SFGH currently uses a risk-based approach with
  Penicillin G intrapartum prophylaxis.
• Should we switch to a screening approach?

18
Changes in pathogens causing early-onset sepsis
in very-low-birth-weight infants. Stoll BJ,
Hansen N, Fanaroff AA, et al. New England Journal
of Medicine 2002347240-7.
Table 3. Rates of Early-Onset Sepsis and
Associated Pathogens in 1991-1993 and 1998-2000.
19
Demographics

Table 1. Characteristics of the Women in the
Screened and Risk-Based Groups.