Prior research has shown a reduced risk for cerebral palsy and other neurologic defects among premature infants exposed antenatally to magnesium sulfate.
Methods and results
Over 16 months, 149 gravidas between 24 and 34 weeks’ gestation in active labor were randomized to magnesium tocolysis, “other” tocolysis, “neuroprotective” magnesium, or placebo. In the tocolytic groups, 37 singletons and 9 pairs of twins were born to mothers administered magnesium via a 4-g bolus followed by an hourly infusion of 2 to 3 g, and 41 singletons and 5 pairs of twins were born to women receiving ritodrine, terbutaline, indomethacin, or nifedipine, at the discretion of the attending physician. In the preventive groups, 28 singletons and 1 pair of twins were randomized to a 4-g bolus of magnesium (without further infusion), and 27 singletons and 1 pair of twins were exposed to placebo.
Umbilical cord blood was collected at delivery to determine serum ionized magnesium levels, and neonatal cranial ultrasound scans were performed. Of 165 infants examined, 37 experienced neonatal intraventricular hemorrhage, periventricular leukomalacia, cerebral palsy, and/or death; those infants had higher umbilical cord magnesium levels.
The researchers concluded that the use of magnesium was associated with a higher risk of adverse perinatal outcome.
Who may be affected by these findings?
Infants exposed to magnesium.
Mittendorf et al recommend “abandoning magnesium for use as routine tocolytic therapy,” but they fail to offer sufficient evidence to warrant such a conclusion.
For example, maternal and cord blood samples were obtained on only half of the population. And while placental cultures were obtained, the results were not reported.
Many potentially important differences between the groups were not explicated in this study. Instead, data are presented for the combined groups and divided into adverse versus no adverse outcomes. For instance, 101 of 142 patients for whom data are reported had preterm premature rupture of membranes, but we don’t know how many fell into either the tocolytic or preventive arms of the study.
The distribution of gestational ages in the 2 separate trials was not revealed—only that 42 of the 147 participants were at less than 28 weeks’ gestation. In addition, the use of betamethasone was roughly equivalent in both the adverse and no adverse outcome groups. Presumably, a higher proportion of women in the tocolytic arm were able to receive the full 24 hours of treatment than in the neuroprotective arm. Further, the rate of steroid use was available for only 135 of the 149 women. Since steroids have been shown to decrease the rate of intraventricular hemorrhage, this particular issue needs to be more clearly explored.
Finally, for both the tocolytic and preventive groups, results were reported as composite adverse outcomes. As such, an infant who died with an intraventricular hemorrhage had 2 adverse events. This method of counting may be misleading.
This study does not support the conclusion that tocolytic use of magnesium should be limited to controlled trials. Rather, we should await the results of 2 large, randomized magnesium trials before deciding to halt the use of magnesium. Until then, it is appropriate to continue judicious use of tocolytics to allow for steroid administration and effectiveness.