MGSO4 TOCOLYSIS & RISK OF CEREBRAL PALSY IN VLBW INFANTS

点击翻译按钮获取中文摘要

The proposed study is designed to evaluate possible reduction in risk of congenital cerebral palsy (CP) in very low birthweight (VLBW) infants associated with tocolytic treatment with magnesium sulfate (MgSO4), a common antenatal treatment for preterm labor. Previous findings suggest that MgSO4 may offer the hope of a substantial reduction in risk of CPP among children who are at especially high risk because of unknown factors associated with their immaturity and VLBW. Serious limitations in these existing studies necessitate replication in a larger study with sufficient sample size to adequately test a hypothesis of association and to evaluate confounding. Using a retrospective case-control design and a two-stage approach to exclude women who would not have been eligible, under usual obstetric practice, to receive in-patient tocolytic treatment, medical record data will be obtained on mother-baby pairs from seven birth years (1988-1994) and 17 large hospitals in California which offer specialized care for newborns. Service agency and neonatal follow-up data will be reviewed to ascertain children with CP and to verify absence of CP in controls. Sample size will be sufficient to test the primary hypothesis of a reduction in risk associated with MgSO4 exposure in utero with power in the range of 90%. The primary purpose of this study is to replicate findings from previous, small observational studies and to provide a more precise estimate of the potential reduction in risk. The study will also examine confounding and effect modification by common maternal medical conditions, effective range of gestational age and birthweight, clinical characteristics of exposed neonates, and dose-response. The case-control design will also permit evaluation of other suggested risk factors for CP. The study population includes singleton survivors to two years of age with birthweight 500-1499 grams, or with birthweight 1500-1999 grams and gestational age lessor than 33 weeks or unknown. After ascertainment of cases, controls will be randomly sampled in a 2:1 ratio within 250 gram birthweight strata to mirror the birthweight distribution of cases. Univariate analyses will utilize Mantel-Haenszel odds rations adjusted for birthweight strata. Multivariate models using logistic regression will be constructed. Results will be available within two years from initiation of project.