The effect of waterbirth on neonatal mortality and morbidity: a systematic review and meta-analysis

Women have been giving birth in water in many centers across the globe; however, the practice remains controversial. Qualitative studies highlight the benefits that waterbirth confers on the laboring woman, though due to the nature of the intervention, it is not surprising that there are few randomized controlled trials available to inform practice. Much of the criticism directed at waterbirth focuses on the potential impact on the neonate.

The objective of this review was to systematically synthesize the best available evidence regarding the effect of waterbirth, compared to landbirth, on the mortality and morbidity of neonates born to low risk women.

This review considered studies that included low risk, well, pregnant women who labor and birth spontaneously, at term (37-42 weeks), with a single baby in a cephalic presentation. Low risk pregnancies are defined as pregnancies with an absence of co-morbidity or obstetric complication, such as maternal diabetes, previous cesarean section, high blood pressure or other illness. Women may be experiencing their first or subsequent pregnancy. The fetus must also be well and without any co-morbidity or complication.The intervention of interest is waterbirth. The comparator is landbirth. Women and their babies must be cared for by qualified maternity healthcare providers throughout their labor and birth. The birth setting must be clearly described but can include home, hospital or birth center, either freestanding or attached to a hospital.This review considered randomized controlled trials, quasi-experimental studies and observational prospective and retrospective cohort studies.

A multi-step search strategy was utilized to find published and unpublished studies, in English between January 1999 and June 2014.

The first author assessed the quality of all eligible studies. The three secondary authors independently assessed six studies each, followed by group discussion using the appropriate Joanna Briggs Institute appraisal checklist.

Data were extracted using a standardized extraction tool from Joanna Briggs Institute.

Quantitative studies were pooled, where possible, for meta-analysis using software provided by Cochrane. Effect sizes were expressed as odds ratio or relative risk, according to study design, and the 95% confidence intervals were calculated. Heterogeneity was assessed statistically using the standard Chi-square test.

The meta-analyses of 12 studies showed that for the majority of outcomes measured in this review there is little difference between waterbirth and landbirth groups. Meta-analysis was not conducted for mortality within 24 days of birth. Heterogeneity was significant between studies for APGAR (Appearance, Pulse, Grimace, Activity, and Respiration). scores ≤7 at one minute and admission to Special Care nursery. Sensitivity analysis for case control studies describing infection found results that were not statistically significant (OR 0.74, 95% CI 0.05-11.06). Results of meta-analysis were also not significant for studies describing resuscitation with oxygen (OR 1.12, 95% CI 0.14-8.79) and Respiratory Distress Syndrome (OR 0.81, 95% CI 0.44-1.49). Results comparing APGAR scores ≤7 at five minutes for waterbirth and landbirth groups results for included RCTs demonstrated results that were not statistically significant (OR 6.4, 95% CI 0.63-64.71). However, results for included cohort studies describing APGAR scores ≤7 at 5 minutes indicate neonates are less likely to have scores ≤7 in the waterbirth group (OR 0.32, 95% 0.15-0.68). Data were not statistically significant for meta-analysis describing admission to NICU (OR 0.51, 95% CI 0.13-1.96) between water and landbirth groups. The differences in arterial (MD 0.02, 95% CI 0.01-0.02) and venous (MD 0.03, 95% CI 0.03-0.03) cord pH, while statistically significant, were clinically negligible.

Analyses of data reporting on a variety of neonatal clinical outcomes comparing land with waterbirth do not suggest that outcomes are worse for babies born following waterbirth. Meta-analysis of results for five-minute APGAR scores ≤7 should be treated with caution due to the different direction of results for meta-analysis of data from randomized controlled trials and cohort studies. Data measuring cord pH (an objective measure of neonatal wellbeing) were robust and showed no difference between groups. Overall this review was limited by heterogeneity between studies and meta-analysis could not be conducted on a number of outcomes. Waterbirth does not appear to be associated with adverse outcomes for the neonate in a population of low risk women.

There is no evidence to suggest that the practice of waterbirth in a low risk population is harmful to the neonate.

There is a paucity of high level evidence to guide practice in the area of waterbirth. It is unlikely that randomized controlled trials on waterbirth will be acceptable to childbearing women or maternity caregivers. Observational studies are a more appropriate choice for researchers in this field as they offer a more practical and ethical approach.