Maternal magnesium sulfate and the development of neonatal periventricularleucomalacia and intraventricular hemorrhage

Objective: Neonatal periventricular leucomalacia and intraventricular hemor rhage are strong correlates of cerebral palsy. Our objective was to evaluat e the effect of maternal magnesium sulfate exposure on the incidence and se verity of periventricular leucomalacia and intraventricular hemorrhage in p reterm neonates. Methods: Nine hundred eighteen consecutive inborn neonates with birth weigh ts from 500 to 1750 g were divided primarily into two groups on the basis o f maternal exposure to magnesium sulfate. The groups were divided secondari ly into two clinical groups, a physician-initiated group, which consisted o f neonates delivered for maternal or fetal indications, and a preterm deliv ery group, which included neonates delivered as a result of preterm labor o r preterm premature rupture of membranes. These clinical groups were strati fied further into magnesium sulfate-exposed and -unexposed subgroups. Neona tal neurosonograms were performed on days 3 and 7 of life and described as normal or abnormal. Abnormal sonograms included any periventricular leucoma lacia or intraventricular hemorrhage. Severe lesions included periventricul ar leucomalacia, periventricular leucomalacia with intraventricular hemorrh age, or grades 3 or 4 intraventricular hemorrhage. The magnesium sulfate gr oups and the clinical groups with their magnesium sulfate strata were compa red for the incidence and severity of abnormal sonograms. They also were co mpared for maternal and neonatal characteristics. Results: Maternal magnesium sulfate exposure was not associated with reduct ion in the incidence of abnormal sonograms when compared with the unexposed group (27% compared with 33%, P = .06). However, fewer severe lesions were observed in the exposed group (14% compared with 21%, P = .004). When clin ical groups were examined, magnesium sulfate was not associated with a decr ease in abnormal sonograms (adjusted odds ratio [OR] 1.09, 95% confidence i nterval [CI] 0.78, 1.52, P = .40) or severe lesions (adjusted OR 1.11, 95% CI 0.73, 1.68, P = .42). Logistic regression analyses of magnesium sulfate exposure within clinical groups controlling for the confounding effects of maternal and neonatal characteristics revealed no protective effect of magn esium sulfate exposure on the incidence of abnormal sonograms (adjusted OR 1.01, 95% CI 0.70, 1.44, P = .97) or severe lesions (adjusted OR 1.01, 95% CI 0.70, 1.74, P = .69). Within clinical groups, the preterm delivery group exhibited an increased risk for abnormal sonograms (adjusted OR 1.63, 95% CI 1.01, 2.67, P = .05) and severe lesions (adjusted OR 9.79, 95% CI 3.27, 29.29, P = .001) when compared with the physician-initiated delivery group, independent of maternal magnesium sulfate exposure. Conclusions: Maternal magnesium sulfate exposure had no protective effect o n the incidence or severity of periventricular leucomalacia and intraventri cular hemorrhage in preterm neonates. The prevalence of these lesions was c orrelated better with the clinical group of origin and indication for its u se. (C) 1999 by The American College of Obstetricians and Gynecologists.