Fourteen RCTs involving 1,533 women were included.
Women receiving amnioinfusion had a significantly lower rate of: Caesarean section rate, foetal heart rate abnormalities, overall Caesarean section rate, foetal acidaemia at birth, foetal heart rate abnormalities during labour, and Apgar scores under 7 at 5 minutes. Postpartum endometritis rates were similar among treatment groups. Apart from acidaemia at birth, studies were homogeneous for outcomes. For most outcomes, funnel plots showed no evidence of publication bias. The included studies were of high quality.
Caesarean section rate for foetal heart rate abnormalities (12 RCTs with 1,240 women): OR 0.23 (95% CI: 0.15, 0.35); heterogeneity p-value 0.44; NNT 8 (95% CI: 6, 10).
Overall Caesarean section rate (13 RCTs with 1,487 women): OR 0.52 (95% CI: 0.40, 0.68); heterogeneity p-value 0.34; NNT 11 (95% CI: 7, 18).
Foetal heart rate abnormalities during labour (9 RCTs with 1,054 women): OR 0.24 (95% CI: 0.17, 0.34); heterogeneity p-value 0.22; NNT 5 (95% CI: 4,7).
Apgar scores under 7 at 5 minutes (11 RCTs with 1,242 women): OR 0.52 (95% CI: 0.29, 0.91); heterogeneity p-value 0.72; NNT 42 (95% CI: 21, 500).
Postpartum endometritis (7 RCTs with 1,044 women): OR 0.60 (95% CI: 0.36, 1.01); heterogeneity p-value 0.11; NNT 32 (95% CI: 16, 1000).
Foetal acidaemia at birth (11 RCTs with 1,173 women): OR 0.40 (95% CI: 0.30, 0.55); heterogeneity p-value 0.002; NNT 8 (95% CI: 6, 11). After removing two studies with higher rates of acidaemia in the control groups, subgroup analysis showed significant difference favouring amnioinfusion.
There were differences among studies in inclusion criteria and interventions. Subgroup analysis after the exclusion of 3 RCTs that did not quantify the degree of polyhydramnios, and 2 RCTs that used transabdominal amnioinfusion, did not alter the overall results.