Seven RCTs with a total of 2,149 women carrying 2,176 unborn infants were identified; 862 remained undelivered more than 7 days after treatment allocation, and so were included in the review.
There was no statistically significant difference in the risk of stillbirths between those exposed to a course of prenatal corticosteroids and the control group, (3 trials; RR 1.67, 95% CI: 0.86, 3.25, P=0.13).
The risk of neonatal death was three times greater in the group that had been exposed to corticosteroids than in the control group (1 trial; RR 3.24, 95% CI: 1.32, 7.96, P=0.01). The higher number of neonatal deaths in the treatment group was not explained by one cause of death alone, but there was a higher number of congenital anomalies in the treatment group. After removing infants with congenital anomalies from the analysis, there was still a greater number of deaths in the treatment group than the control group, but this difference was no longer statistically significant (RR 2.53, 95% CI: 0.91, 7.05, P=0.11). The risk of perinatal mortality was more than twice as high in the corticosteroid-exposed group than the control (3 trials; RR 2.13, 95% CI: 1.27, 3.57, P<0.01). When infants with congenital anomalies were excluded from the analysis, the difference between groups remained statistically significant (RR 1.86, 95% CI: 1.08, 3.22, P<0.03).
There was no statistically significant difference in the risk of respiratory distress syndrome between infants exposed to a course of prenatal corticosteroids and infants in the control group, (7 trials; RR 0.72, 95% CI: 0.49, 1.07, P=0.11). Other infant outcomes were only reported in one trial. Infants exposed to corticosteroids had a significantly lower gestational age than infants in the control group; the mean difference was -5.00 days (95% CI: -9.15, -0.85, P=0.02). No statistically significant differences in birth weight, low Apgar scores, neonatal infection, or cerebroventricular haemorrhage were found between the two groups.
Mothers who received corticosteroids more than 7 days before giving birth were at almost three times greater risk of chorioamnionitis than mothers in the control group, (1 trial; RR 2.91, 95% CI: 1.25, 6.74, P=0.01). No significant differences were found in pyrexia or type of delivery between the two groups.
The removal of trials with inadequate or unreported methods of treatment allocation from the analysis had little effect on the results. There was no evidence of statistical heterogeneity in the pooled analyses. There was some evidence of imbalance in gestational age, race and the use of alcohol as a tocolytic across treatment groups in some studies.