The recent Cochrane Review and similar metaanalysis by Conde-Agudelo and Romero establish a fetal neuroprotective role for antenatal magnesium sulfate (MgSO ₄ ) given to women at risk of preterm birth, with significant reductions in the risk of cerebral palsy (CP) and gross motor dysfunction. However, the upper gestational age that warrants treatment is not clear. CP is 20 times more common in babies born at <28 weeks’ gestation (14.6% prevalence) than in those of 32-36 weeks (0.7% prevalence). Perinatal and neonatal factors are more prominent in the etiology of CP in less mature infants, suggesting that an intervention immediately predelivery is more likely to be effective the earlier the gestation. Thus the number needed to treat will increase significantly with advancing gestational age.

Both metaanalyses analyze all trials recruiting below certain gestational ages together, thus including all the most immature babies at highest risk of CP in every analysis. This approach does not inform how effective MgSO ₄ might be at higher gestational ages. Four of the individual trials recruited at ≤33, ≤32, ≤31, and ≤29 weeks’ gestation, with the fifth supplying data for women ≤36 weeks. If the 2 trials including pregnancies of 33 weeks’ gestation are combined, the relative risk (RR) for CP is 0.76 (95% confidence interval [CI], 0.21–2.80). The 3 trials recruiting women at 32 weeks also have no significant reduction in CP (RR, 0.71; 95% CI, 0.43–1.15). Even in these trials, most CP will be in the less mature infants.

The cited trial by Rouse et al reported gestational age subgroups. There was a significant reduction in moderate or severe CP in babies of women recruited at <28 weeks (RR, 0.45; 95% CI, 0.23–0.87), but not in those of 28-31 weeks’ gestation: 8/599 MgSO ₄ exposed vs 8/599 controls. Combining Rouse et al’s <28-week subgroup with Crowther et al’s data on ≤29-week pregnancies, there is a significant reduction in moderate or severe CP (RR, 0.55; 95% CI, 0.35–0.88). The 2 metaanalyses show no protective effect for mild CP. Thus, the evidence suggests that only the most immature infants are protected.

There is no evidence of a protective effect of MgSO ₄ at 32 or 33 weeks’ gestation. There is doubt of any effect at ≥28 weeks. Since the treatment has significant maternal side effects and appreciable resource implications, it seems reasonable to restrict MgSO ₄ treatment to women threatening to deliver at ≤29 weeks’ gestation.