Waterbirth: do the benefits outweigh the harms?

A review of waterbirth in the Journal of Pediatrics and Child Health succinctly summarizes the current state of knowledge about waterbirth.

Water births and the research required to assess the benefits versus the harms by Mark W. Davies starts by making it clear that there is a big difference between laboring in water and giving birth in water:

There are two separate and distinct aspects to the use of water immersion in labour:

1 the use of immersion for women in labour (without birth into the water); and
2 immersion for women in the second stage of labour with birth into the water – water birth.

This separation must be re-emphasised whenever discussing the use of water immersion in labour.

Davies points out the polarized nature of the discussion on waterbirth:

On the one hand, there are those who cannot imagine why you would want to deliver a baby into water and put them in harm’s way; on the other hand, there are those who believe that immersion in the second stage of labour offers significant benefits to the mother and is safe. However, the questions that must be asked about any health-care intervention are: first, is it useful?; second, does it do any harm?; and third, do any benefits outweigh any harms?…

The threshold question is whether waterbirth has any benefits.

There is some level 2 evidence available. However, there is only one randomised controlled trial (RCT) that has studied women who were randomised to either no immersion (n = 60)
or immersion in the second stage of labour with birth into the water (n = 60). The results … have been included in the Cochrane systematic review by Cluett et al.: It has not been published in full in the peer reviewed literature… The only outcome that showed a statistically significant difference was the subjective outcome of whether the women were satisfied with pushing efforts: There were no significant differences in any objective assessments of benefit.

The other two RCTs allocated women to either no immersion or immersion:Women in the immersion group were able to use immersion in the first or second stage of labour, or both, with or without birth into the water. The study by Woodward and Kelly was undertaken as a feasibility exercise, and was too small and greatly underpowered to assess efficacy… The most recent RCT8 also used either no immersion (n = 53) or immersion in both the first and second stages of labour (n = 53). The results are difficult to interpret as many of the basic CONSORT reporting requirements are missing… [T]he authors report that second-stage duration was the same in both groups, but there was a significant difference in rates of ‘gave birth naturally’ (outcome not defined). All those in the immersion group ‘gave birth naturally’ compared with 79% in the no immersion group. More information than is currently available would be required to assess the validity of this trial…

How about the harms?

The only RCTs available were greatly underpowered to detect any significant differences for any harmful effects to the mother or infant, especially the uncommon outcomes such as
perinatal death…

What is the evidence from case reports?

There are deaths reported directly attributed to water birth and significant morbidity directly attributed to water birth. Morbidity includes near drowning and other respiratory
difficulties including stridor, hyponatraemia and seizures secondary to hyponatraemia, infection such as Legionella pneumonia, hypoxic-ischaemic encephalopathy and avulsion of the umbilical cord.

That’s not surprising given what we know about fetal and newborn physiology.

It should be remembered that unrecognised asphyxia can occur during any delivery, that asphyxiated babies gasp (pre-, intra- and postpartum), and that if gasping occurs in infants born into the water, they will gasp under water and aspirate bath water, further compromising gas exchange and delaying resuscitation. This mechanism is almost certainly the cause for many of the morbidities described above.

Do the benefits outweigh the harms. There’s no evidence that they do.

First, there’s no evidence that there is a benefit to delivering under water (as opposed to laboring in water). Second, although the existing RCTs are underpowered to detect difference in perinatal death rates, there is a large and growing body of case studies of waterbirths that results in deaths and serious injuries to babies. That’s not surprising since, contrary to the claims of waterbirth advocates, it is incontrovertible that born and partially born infants gasp and can and do aspirate the fecally contaminated bath water.

Davies recommends that there should be no waterbirths except as part of randomized controlled trials with informed consent. What would such trials involve?

• Treatment allocation should be randomised – to reduce selection bias
• Treatment allocation should be concealed by a central mechanism such as a central telephone/Internet-based service – to reduce selection bias
• Treatment allocation should occur at the start of second stage – to reduce selection bias and performance bias (especially co-intervention)
• No crossover should be allowed – to reduce performance bias (especially contamination)
• Outcome assessment should be complete – to prevent attrition bias
• Outcome assessors should be blinded to treatment allocation – to prevent detection bias
• Adequate numbers should be enrolled to give the study adequate power to detect important differences in neonatal mortality and morbidity (such a RCT will require at least 3500 women in each group to detect a 100% increase in perinatal death rate (i.e. from 2 to 4 per 1000) with 80% power and an a of 0.05)
• Long-term follow-up of infants to assess long-term neurodevelopmental outcome.

To be adequately generalisable, the trial should only recruit women who are of low risk for complications related to labour and birth, and only recruit women who want a water birth.

As is typical in natural childbirth and homebirth advocacy, a procedure has been put into practice with no evidence that it is either safe or effective and a growing body of evidence that it is neither. That’s not surprising since there is nothing natural about waterbirth.