OBJECTIVE:

To compare pregnancy outcome prospectively after phenytoin and carbamazepine monotherapy with outcome in matched mother-child pairs exposed to nonteratogens to evaluate the relative fetal safety of these drugs.

DESIGN:

A prospective, controlled, and blinded observational study.

PATIENTS:

Thirty-six mother-child pairs exposed to carbamazepine monotherapy and 34 pairs exposed to phenytoin monotherapy, all prospectively studied, were compared with mother-child pairs exposed to nonteratogens. The controls were matched for maternal age, time of consultation, obstetric history, and socioeconomic status.

MAIN OUTCOME MEASURE:

The primary end point of interest was the children's global IQ measured by either the Bayley or the McCarthy scale according to their ages.

SETTING:

A teratology consultation program and two neurology services in Toronto, Ontario.

RESULTS:

Children exposed to phenytoin in utero had a mean (+/- SD) global IQ 10 points lower (95% confidence interval, 4.9 to 15.8 points) than their matched controls (113.4 +/- 13.1 and 103.1 +/- 25.1; P = .038). The Reynell language development scores followed a similar trend, with children exposed to phenytoin scoring significantly lower than their controls. Phenytoin-exposed children had a global IQ of 84 or less significantly more often than the control group (P < .01). Children exposed in utero to carbamazepine did not differ from their controls on any of the neurobehavioral tests.

CONCLUSIONS:

Our study suggests a clinically important negative effect of phenytoin on neurobehavioral development, independent of maternal or environmental factors, causing a substantial number of children to achieve a lower score than expected on cognitive tests. No similar effects could be shown after gestational use of carbamazepine.