Magnesium for seizure prophylaxis in patients with mild preeclampsia.

OBJECTIVE To estimate whether magnesium therapy is the optimal management for women with mild preeclampsia. METHODS A decision analytic model was designed for women with mild preeclampsia to estimate whether empiric magnesium therapy or no magnesium therapy is the optimal management strategy. The model considered relevant clinical events: seizure, placental abruption, and magnesium toxicity. Modeled outcomes were maternal mortality, neonatal mortality, and neurologic neonatal compromise. The two strategies were compared based on the probability of clinical events and outcomes and the utilities or values assigned to the outcomes by prior research. Probabilities and utilities were derived from the literature. RESULTS The base-case analysis showed that although the no-magnesium strategy results in a 15% reduction in neonatal mortality and avoids most maternal drug toxicity, it produces a twofold increase in maternal death and more neurologically compromised neonates compared with empiric magnesium. The two strategies are essentially equivalent with regard to aggregate maternal and neonatal outcomes (0.9792 compared with 0.9781 utilities). Multivariable sensitivity analysis using Monte Carlo simulation confirmed the decision to be a "toss-up," yielding a similar mean utility for no-magnesium and magnesium strategies (0.9789+/-0.1374 compared with 0.9784+/-0.1390, respectively). CONCLUSION Our decision model indicates that either strategy, using or not using empiric magnesium sulfate therapy, is acceptable. The clinical decision of whether to use magnesium in patients with mild preeclampsia for seizure prophylaxis should be determined by the physician or institution, considering patient values or preferences and the unique risk-benefit trade-off of each strategy.