Seven randomised controlled trials (1,214 patients, range 31 to 325) were included in the review. The overall bias assessment indicated that three trials were unlikely to be at risk of bias, two trials had a low risk of bias and two trials had a moderate risk of bias.
Statistically significant benefits were observed in newborns treated with therapeutic hypothermia for the composite outcome of death or major neurodevelopmental disability at 18 months (RR 0.76, 95% CI 0.69 to 0.84 and NNT=7, 95% CI 5 to 10; Ι²=0%; seven trials). Significant reductions were observed in risk of death at 18 months (RR 0.75, 95% CI 0.63 to 0.88 and NNT=11, 95% CI 7 to 26; Ι²=0%).
Among newborns who survived to 18 months, therapeutic hypothermia was associated with statistically significant reductions in risk of major disability (RR 0.68, 95% CI 0.56 to 0.83 and NNT=8, 95% CI 5 to 16; Ι²=12%), cerebral palsy (RR 0.62, 95% CI 0.49 to 0.78 and NNT=8, 95% CI 5 to 16; Ι²=33%), developmental delay (RR 0.66, 95% CI 0.52 to 0.82 and NNT=8, 95% CI 5 to 18 patients; Ι²=25%) and blindness (RR 0.56, 95% CI 0.33 to 0.94 and NNT=23, 95% CI 12 to 207; Ι²=0%). No significant differences were observed for risk of deafness between newborns treated with hypothermia and newborns who maintained normothermia. Survival with normal neurological function was increased in the therapeutic hypothermia group (RR 1.63, 95% CI 1.36 to 1.95 and NNT=7, 95% CI 5 to 11; Ι²=0%; six trials).
The results of sensitivity and subgroup analyses were similar to the results for the primary analyses. No evidence of publication bias was identified by visual appraisals of funnel plots in the review