Red X iconGreen tick iconYellow tick icon


Although waterbirth is widely practised in New Zealand in all intrapartum settings, there is not universal agreement amongst health care professionals who provide maternity and neonatal care about its safety. Previous studies do not report an associated increase in poorer neonatal or maternal outcomes. However, case reports and case series report rare adverse neonatal outcomes.


This retrospective observational cohort study, used a New Zealand national midwifery dataset that included low risk births between 1st January 2012 and 31st December 2014. The study's aim was to determine whether there were poorer maternal and newborn outcomes for women who gave birth underwater in comparison with women giving birth conventionally. Logistic regression was used determine adjusted odds ratio (aOR) for selected maternal and neonatal outcomes. In addition non inferiority analysis was to determine whether waterbirth outcomes were clinically non inferior to conventional birth for these low risk women.


There were 47 651 women in the overall cohort, with 7265 (15.3%) in the waterbirth group and 40 386 (84.7%) in the conventional birth group. Women having waterbirths were more likely to be of European ethnicity and less likely to be Māori, Pacific, Asian or Indian. They were more likely to give birth at home or in a primary maternity unit, give birth at a later gestational age and register for care earlier in pregnancy, and less likely to be obese or recorded as smokers, than the conventional birth group.

Babies in the waterbirth group had greater aOR for Apgar score <7 at 5 minutes (aOR 1.17 (0.89 to 1.53)) and Apgar <7 at 1 minute (aOR 1.18 (1.04 to 1.32) p=0.008)). Neither outcome was found to be clinically non-inferior, with results inconclusive for both of these findings. The cumulative incidence of the primary neonatal outcome (Apgar score <7 at 5 minutes) was low (<1%) and the difference between groups minimal (0.3%), with the number needed to harm (NNH) high (3333).

Despite a higher adjusted odds ratio (aOR) of Apgar scores <7 at 1 and 5 minutes, babies in the waterbirth group had a significantly lower aOR of being resuscitated (aOR 0.85 (0.77 to 0.93) p=0.001). Further Mantel-Haenszel analysis, stratifying results by waterbirth and conventional birth groups, found having an Apgar <7 at 1 minute was not associated with an increased risk of an Apgar <7 at 5 minutes for the waterbirth group (compared to the conventional birth group).

The aORs for all maternal outcomes were significantly lower for the waterbirth group, blood loss ≥500mls (aOR 0.81 (0.81 to 0.8, p≤0.001)), third or fourth degree tear (aOR 0.60 (0.49 to 0.75, p≤0.001)) and episiotomy, second, third or fourth degree tear (aOR 0.71 (0.67 to 0.75, p≤0.001)). Results from non-inferiority analysis demonstrated waterbirth outcomes were non-inferior (or superior) in comparison with the conventional birth group.


There are methodological limitations of observational waterbirth research. In spite of these, when considered in the context of previously published findings, this study demonstrates that there is no clear evidence for either increased harm or benefit to babies born by waterbirth compared with conventional birth, and there is a clear benefit associated with waterbirth for women. This study builds on the body of observational waterbirth evidence and supports the view that waterbirth is a safe option for low risk women and babies.

Supervisors: Phil Hider & Jonathan Williman

Back to top