MAARIE Template for Evaluating Evidence Based Health Research

Journal of Midwifery &Women’s Health www.jmwh.org Original Research

Maternal and Newborn Outcomes Following Waterbirth: The Midwives Alliance of North America Statistics Project, 2004 to 2009 Cohort Marit L. Bovbjerg, PhD, MS, Melissa Cheyney, PhD, CPM, LDM, Courtney Everson, MA, PhD

Introduction: Data on the safety of waterbirth in the United States are lacking.

Methods: We used data from the Midwives Alliance of North America Statistics Project, birth years 2004 to 2009. We compared outcomes of neonates born underwater waterbirth (n = 6534), neonates not born underwater nonwaterbirth (n = 10,290), and neonates whose mothers intended a waterbirth but did not have one intended waterbirth (n = 1573). Neonatal outcomes included a 5-minute Apgar score of less than 7, neonatal hospital transfer, and hospitalization or neonatal intensive care unit (NICU) admission in the first 6 weeks. Maternal outcomes included genital tract trauma, postpartum hospital transfer, and hospitalization or infection (uterine, endometrial, perineal) in the first 6 weeks. We used logistic regression for all analyses, controlling for primiparity.

Results:Waterbirth neonates experienced fewer negative outcomes than nonwaterbirth neonates: the adjusted odds ratio (aOR) for hospital trans- fer was 0.46 (95% confidence interval [CI], 0.32-0.68; P � .001); the aOR for infant hospitalization in the first 6 weeks was 0.75 (95%CI, 0.63-0.88; P � .001); and the aOR for NICU admission was 0.59 (95% CI, 0.46-0.76; P � .001). By comparison, neonates in the intended waterbirth group experienced more negative outcomes than the nonwaterbirth group, although only 5-minute Apgar score was significant (aOR, 2.02; 95% CI, 1.40-2.93; P � 0001). For women, waterbirth (compared to nonwaterbirth) was associated with fewer postpartum transfers (aOR, 0.65; 95% CI, 0.50-0.84; P = .001) and hospitalizations in the first 6 weeks (aOR, 0.72; 95% CI, 0.59-0.87; P � 0.001) but with an increased odds of genital tract trauma (aOR, 1.11; 95%CI, 1.04-1.18; P = .002).Waterbirth was not associated withmaternal infection.Women in the intended waterbirth group had increased odds for all maternal outcomes compared to women in the nonwaterbirth group, although only genital tract trauma was significant (aOR, 1.67; 95% CI, 1.49-1.87; P � .001).

Discussion:Waterbirth confers no additional risk to neonates; however, waterbirth may be associated with increased risk of genital tract trauma for women. J Midwifery Womens Health 2016;61:11–20 c© 2016 by the American College of Nurse-Midwives.

Keywords: childbirth, complications, perineal trauma, safety, waterbirth

INTRODUCTION

Waterbirth is highly controversial in the United States,1–8 despite being an accepted practice in other high-resource nations.9–12 Proponents of waterbirth cite anthropological ev- idence from Odent and Tjarkovsky regarding childbearing traditions that include immersion13,14; the maternal benefits of laboring in water, such as pain relief and reduced stress on tissues secondary to buoyancy15,16; the potential benefits to a neonate of being born into a warm, liquid environment similar to the amniotic fluid17; and a series of studies, mostly small and observational, suggesting no adverse effects for ei- ther the laboring woman or the neonate.18,19 By contrast, in the spring of 2014, the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pedi- atrics (AAP) jointly issued clinical recommendations advising strongly against allowing women to labor in water after the first stage of labor is complete.3

Waterbirth is generally defined as a neonate being in- tentionally born underwater. Provided that the neonate is

Address correspondence to Marit Bovbjerg, PhD, MS, College of Public Health and Human Sciences, Oregon State University, Milam Hall 103, Corvallis, OR 97331. E-mail: [email protected]

promptly brought to the surface, it is thought that the diving reflex, which mechanically blocks the airway of submerged infants (although not older children or adults), will pre- vent the newborn from aspirating the water.20 The category waterbirth does not include women who labor in water but give birth to their newborn into air. Laboring in water is con- sidered safe; the current question in the literature is whether waterbirth is safe.3

Published reports of outcomes following waterbirth in the United States currently consist solely of case series1,2,4,5,21–24 rather than studies with robust designs and adequate power. However, there are several cohort studies from Europe describing waterbirth outcomes,25–35 nicely summarized by Nutter et al in a recent review.19 The results of these studies collectively suggest that waterbirth is not associated with an increased risk of morbidity for the newborn (eg, low Apgar score, neonatal intensive care unit [NICU] admission, neonatal injury, or death), although small sample sizes hinder comparisons for all but themost common events.18,19 Regard- ing maternal outcomes, previous literature suggests women do not experience an increase in perineal trauma, infection, or hemorrhage.18,19 Nonetheless, it can be argued that both the US population and US health care system are unique, and thus results from Europe might not be generalizable to the

1526-9523/09/$36.00 doi:10.1111/jmwh.12394 c© 2016 by the American College of Nurse-Midwives 11

✦ Using data collected from the Midwives Alliance of North America Statistics Project (MANA Stats 2004-2009), this study reports waterbirth outcomes for a large sample of midwife-attended births occurring at home and in birth centers in the United States (N = 18,343 women); 35% of the women (n = 6521 women; 13 sets of twins) had a waterbirth.

✦ Neonates of womenwho had awaterbirth were less likely to experience a low 5-minute Apgar score, neonatal transfer to the hospital, and hospitalization or neonatal intensive care unit admission in the first 6 weeks when compared to nonwaterbirth neonates.

✦ For women, waterbirth was associated with decreased odds of hospitalization, either immediately postpartum or within the first 6 weeks, but increased odds of genital tract trauma.

✦ Waterbirth was not associated with increased risk of maternal infection.

United States. The purpose of this study, therefore, was to re- port waterbirth outcomes from a large sample of midwife- led births occurring at home and in birth centers in the United States.

METHODS

Data Source and Sample Description

The data for this study come from the Midwives Alliance of North America Statistics Project, commonly referred to as MANA Stats.36 MANA Stats is an ongoing, Web-based data collection effort designed to capture complete courses of care from the medical records of women who have had midwife-led pregnancies and births. Any midwife, regardless of birth setting, is eligible to contribute data. However, in practice, most MANA Stats records are for planned home or planned birth center births (97.6% for the years 2004-2009) attended by certified professional midwives (CPMs) in the United States. Of the births in the 2004 to 2009 MANA Stats dataset, 73% of the births were attended by CPMs.36

A midwife who is a MANA Stats contributor enters data on all women in her care from the first prenatal visit through the final visit, which is usually at 6 weeks postpartum. Midwives are required to preregister or log patients into the MANA Stats system early in care, before the outcome of the pregnancy is known. This prospective logging helps to ensure that all births from participating midwives are captured, regardless of outcome, thus reducing selection bias in the sample.

Women give informed consent allowing their deidentified data to be included in MANA Stats, and the consent includes explicit permission for the data to be used for research. Should a woman decline consent, her data are not included, but this decision to decline consent would occur early in pregnancy and thus could not be affected by pregnancy outcome. In prac- tice, very few women decline this consent; based on practice data reported bymidwives, we estimate thatMANA Stats cap- tures 97% of births attended by midwife contributors.36 The high rate of maternal participation in this population is not unique to the MANA Stats dataset; it has also been reported in other studies enrolling women planning home and birth center births.37,38

The institutional review board at Oregon State University approved this analysis, which usesMANA Stats data for births

that occurred between 2004 and 2009. Evidence of reliabil- ity and validity of the MANA Stats 2004 to 2009 dataset, as well as detailed data collection protocols, is presented elsewhere.36

We limited our sample to births that were planned home births or planned birth center births at the onset of labor wherein the neonate was actually born in the intended set- ting (ie, no intrapartum transfer to a hospital occurred). Thus, excluded from the sample were the data from women who planned a hospital birth and from those for whom a hospital birth was not planned but occurred following an intrapartum transfer (Figure 1).

These cases were excluded for 2 reasons. First, during the research years (2004-2009), very few hospitals in the United States offered the option of giving birth underwater.39 Second, hospital births are almost always the most complicated preg- nancies and labors in the MANA Stats database; the major- ity of contributors to the project specialize in home or birth center birth and transfer care to hospitals only when com- plications arise.40 Thus, including women with more compli- cated pregnancies or labors resulting in the transfer of care to a hospital-based provider, either before or during labor, com- bined with the reduced likelihood of encountering the expo- sure, would have introduced bias in the direction of making waterbirth appear safer relative to nonwaterbirth. Mother– newborn dyads who transferred to the hospital during the postpartum period were retained in the sample because, if in fact waterbirth itself introduces risk (eg, infection, respi- ratory distress), we would expect to see more women and newborns with immediate postpartum newborn or maternal complications requiring transfer and possible hospitalization following waterbirth. We also excluded 12 singleton pregnan- cies for which the waterbirth variable was missing. Apply- ing these inclusion criteria resulted in a sample size of 18,397 neonates (N = 18,343 pregnancies), as shown in Figure 1.

Variables

The main exposure, waterbirth, was collected as a 3-level cat- egorical variable. To the question “Baby born underwater?” the midwife had 3 answer options: “no”; “yes”; or “intended, but not born underwater.” Throughout this article, we refer to these categories as nonwaterbirth, waterbirth, and intended waterbirth, respectively.

12 Volume 61, No. 1, January/February 2016

n=24,848 pregnancies (n=24,969 newborns; 119 sets of twins and 1 set of triplets)

21,414 pregnancies 21,495 newborns (81 sets of twins)

18,363 pregnancies 18,417 newborns (54 sets of twins)

3433 (3473 newborns) plus 1 record that had missing data = 3 434

Stopped receiving care from this midwife prior to the onset of labor

yes

no

Location of birth: home or birth center yes

no 2949 (2976 newborns) hospital 100 ‘other’ 2 missing

18,355 pregnancies 18,409 newborns (54 sets of twins)

Neonate born alive yes

no 8 IUFDs (intrauterine fetal demise) after the onset of labor, but prior to birth

10,252 pregnancies 10,290 newborns (38 sets of twins)

6521 pregnancies 6534 newborns (13 sets of twins)

1570 pregnancies 1573 newborns (3 sets of twins)

non -waterbirth

waterbirth

intended waterbirth

12 singleton pregnancies were missing data on waterbirth

Figure 1. Sample Size Delimitation Begins with all records entered into MANA Stats for birth years 2004 to 2009. Women who changed providers (ie, stopped receiving care from the midwife filling out the data form) prior to the onset of labor are excluded. Some of these women may be included as separate records if they changed providers to another MANA Stats contributor; however, many developed a complication requiring maternity specialty care and therefore would not be included. Also excluded are women who did not give birth at home or at a birth center, and mother—fetus dyads if the fetus died prior to birth. All numbers are counts of singleton pregnancies, unless otherwise indicated.

Abbreviation: MANA Stats, Midwives Alliance of North America Statistics.

Neonatal outcomes included a 5-minute Apgar score of less than 7 (yes/no), postpartum transfer to the hospital for a newborn indication (referred to hereafter as neonatal trans- fer, yes/no), any infant admission to the hospital during the first 6 weeks of life (yes/no), and any NICU admission during the first 6 weeks of life (yes/no). Although we knew the num- ber of events would be small based on our previous work with this dataset,33 we also included early (prior to 7 completed days of life) and late (at least 7 completed days of life, but not yet 28) neonatal deaths as end points.

Maternal outcomes included postpartum reproductive tract infection at any time during the first 6 weeks postpar- tum (presence of: uterine infection, urinary tract infection, or delayed perineal healing/infection), postpartum transfer for a maternal indication (referred to hereafter as postpartum transfer; yes/no), anymaternal admission to the hospital dur- ing the first 6 weeks postpartum (yes/no), and genital tract trauma. The degree of genital tract trauma was evaluated first as a simple dichotomous variable and then further as a mul- tilevel nominal variable with the following categories: none, episiotomy only, first- or second-degree perineal only, third-

or fourth-degree perineal only, mild labial only, more severe labial only (defined on the data collection formas required re- pair), other trauma requiring repair, trauma at multiple sites, and trauma not otherwise specified. The latter category con- sisted of women for whom the midwife indicated that, yes, there was trauma, but then did not answer the follow-up ques- tions regarding location and severity.

For the neonatal and postpartum transfers,midwiveswere able to indicate multiple indications for transfer. It is there- fore possible that one mother–newborn dyad could have ex- perienced both a postpartum transfer and a neonatal transfer if, for example, the midwife indicated both “extensive lacera- tion repair requiring anesthesia” and “evaluation of congenital anomalies” as reasons for transfer.

Analysis

We used logistic regression to analyze all dichotomous out- comes (5-minute Apgar score of � 7, neonatal transfer, NICU admission, infant hospitalization in the first 6 weeks, post- partum transfer, postpartum reproductive tract infection,

Journal of Midwifery &Women’s Health � www.jmwh.org 13

maternal hospitalization in the first 6 weeks, dichotomized trauma). Directed Acyclic Graph (DAG) methodology41,42 was used to determine potential confounders.Directed acyclic graphs are a type of causal model that allow the researcher to determine a complete set of potential confounders that maxi- mizes use of available data while simultaneously reducing bias that would result from adjusting for highly collinear covari- ables. The DAG/causal model that we drew for this analysis is available from the authors on request; as a result of the DAG analysis, we controlled for primiparity in all models.

The 3-level waterbirth exposure variable was entered into the models as a nominal variable, with nonwaterbirth as the reference category. Our results are thus presented as adjusted odds of a given outcome for waterbirth compared to nonwa- terbirth and, separately, adjusted odds of a given outcome for intended waterbirth compared to nonwaterbirth.

As expected, cell counts were very low for the neona- tal death outcomes. Thus, for these outcomes we report the raw data but did not calculate adjusted odds ratios and corre- sponding 95% confidence limits.

For the multilevel nominal outcome (genital tract trauma), we used multinomial (ie, not ordered) logistic regression and controlled for primiparity. No trauma was the reference category. Data were analyzed using SPSS 19.0.0.1 (IBM Corp, Armonk, NY) and S-Plus Version 8.1 (Tibco Spotfire, Seattle, WA).

RESULTS

We report results from 18,343 births, which included 18,397 neonates (see Figure 1). Of these, 10,252 women (10,290 neonates) were in the nonwaterbirth group, 6521 women (6534 neonates) were in the waterbirth group, and 1570 women (1573 neonates) were in the intended waterbirth group. Demographics of the women in this sample are shown in Table 1. Briefly, the majority were white, married, and col- lege educated. The mean age at conception was 29.9 years (standard deviation 5.3). Our sample does include larger proportions of both Amish/Mennonite (6.0%) and grand- multiparous women (8.5%), relative to the US population as a whole; as expected, there is a large overlap between these 2 groups. Additionally of note, 968 women had vaginal births after cesarean (VBACs), 134 gave birth to a neonate in breech presentation, and 54 women had twins. Because our sample was limited to those women who gave birth at home or in a birth center, all of these were vaginal births.

Neonatal Outcomes

Neonates born underwater (waterbirths) fared better than their nonwaterbirth counterparts on all neonatal outcome measures, when controlling for primiparity (Table 2). The ad- justed odds ratio (aOR) for neonatal transfer to the hospital, for waterbirth neonates compared to nonwaterbirth neonates, was 0.46 (95% confidence interval [CI], 0.32-0.68; P � .001). The aOR for NICU admission during the first 6 weeks was 0.59 (95% CI, 0.46-0.76; P � .001); the aOR for any hospital admission during the first 6 weeks was 0.75 (95% CI, 0.63- 0.88; P � .001) (Table 2). There was no evidence of 5-minute

Apgar scores below 7 being more common in the waterbirth group (aOR, 0.88; 95% CI, 0.65-1.19; P = .42).

By contrast, neonates whose mothers intended a water- birth but did not have one (intended waterbirths) fared worse than neonates whosemothers had not planned to give birth in the water. Neonates in the intended waterbirth category had a 102% increase in the odds of a 5-minute Apgar score of less than 7 (aOR, 2.02; 95% CI,1.40-2.93; P � 0.001).

Although the number of events was too small for firm conclusions, we found no evidence of increased neonatal deaths (early or late) among neonates in the waterbirth group. There were 9 deaths (6 early, 3 late) in the nonwaterbirth comparison group; these were attributed to hypoxia, congen- ital anomalies, chorioamnionitis, cord accidents (3), and un- known causes (3). Among the waterbirth group, there were 3 neonatal deaths (2 early, one late), attributed to hypoxic ischemic encephalopathy, congestive heart failure, and un- known causes. In the intended waterbirth group, there were also 3 neonatal deaths (2 early, one late), attributed to placen- tal abruption, shoulder dystocia with compressed cord, and hypoxia (Table 2).

Maternal Outcomes

Women who completed the second stage while immersed in water (waterbirths) had a 35% reduction in odds of postpar- tum transfer (aOR, 0.65; 95% CI, 0.50-0.84; P = .001) and a 28% reduction in odds of maternal hospitalization in the first 6 weeks (aOR, 0.72; 95% CI, 0.59-0.87; P � .001) (Table 3). By contrast, women who had a waterbirth also had an 11% in- crease in odds of experiencing any genital tract trauma (aOR, 1.11; 95% CI, 1.04-1.18; P = .002). When categories based on trauma location and severity were assessed, however, there was no discernible pattern between the women in the 3 co- horts (Table 4). Women who planned a waterbirth but did not have one (intended waterbirths) experienced substantially in- creased odds of any genital tract trauma (aOR, 1.67; 95% CI, 1.49-1.87; P � .001) (Table 3).

DISCUSSION

Main Findings and Interpretation

This retrospective cohort study is the largest study on this topic to date and one of the first to focus on a US population. We found thatwaterbirthwas not associatedwith an increased risk of 5-minute Apgar score of less than 7, immediate neona- tal transfer of care to a hospital, any neonatal hospitalization in the first 6 weeks, or NICU admission in the first 6 weeks. For women, we found that waterbirth was associated with re- duced risk of both immediate postpartum transfer of care to a hospital and any maternal hospitalization in the first 6 weeks. However, waterbirth was associated in our data with an in- creased risk of genital tract trauma. We found no evidence of an association between waterbirth and uterine, endometrial, or perineal infection.

A recent Cochrane review on waterbirth stated that: “Im- mersion during the 2nd stage of labour needs further investi- gation, but at present there is no clear evidence to support or not to support a woman’s decision to give birth in water.”18 This review, as is the case for all Cochrane Collaboration

14 Volume 61, No. 1, January/February 2016

Table 1. Sample Demographics and Pregnancy Characteristics for 18,343Women (18,397 neonates) WhoGave Birth at Home or in a Birth Center with a Midwifea

Total Sample Nonwaterbirth Waterbirth IntendedWaterbirth

N () n () n () n () Race/ethnicity (categories are not mutually exclusive)b

African or Caribbean 386 (2.1) 220 (2.1) 122 (1.9) 44 (2.8)

Asian 820 (4.5) 471 (4.6) 274 (4.2) 75 (4.8)

White 16,785 (91.8) 9336 (91.1) 6027 (92.5) 1422 (90.7)

Hispanic 786 (4.3) 448 (4.4) 262 (4.0) 76 (4.8)

Native American 192 (1.0) 117 (1.1) 64 (1.0) 11 (0.7)

Other 373 (2.0) 206 (2.0) 129 (2.0) 38 (2.4)

Special groups (categories are not mutually exclusive)

Amish/Mennonite/other Plain Church community 1081 (5.9) 963 (9.4) 121 (1.9) 23 (1.5)

Immigrant 504 (2.7) 298 (2.9) 156 (2.4) 50 (3.2)

Educationc

Less than high school 1338 (7.5) 1094 (10.9) 200 (3.1) 44 (2.9)

Completed high school 3126 (17.5) 1756 (17.5) 1134 (17.8) 236 (15.4)

Completed 1-3 years of college 4552 (25.4) 2425 (24.2) 1726 (27.1) 401 (26.2)

Completed at least 4 years of college 8885 (49.6) 4739 (47.3) 3297 (51.9) 849 (55.5)

Marital statusd

Married 16,139 (88.0) 9031 (88.1) 5750 (88.2) 1358 (86.5)

Partnered (although not married) 1717 (9.4) 950 (9.3) 589 (9.0) 178 (11.3)

Single (includes separated/divorced) 420 (2.3) 238 (2.3) 153 (2.4) 29 (1.9)

Pregnancy characteristics

Primiparouse 4001 (21.8) 2316 (22.5) 1246 (19.1) 439 (28.0)

Multiparousf 13,184 (71.8) 7121 (69.5) 4984 (76.4) 1079 (68.7)

Grand multiparousg 1158 (6.3) 815 (7.9) 291 (4.5) 52 (3.3)

History of cesareanh 967 (5.3) 535 (5.2) 327 (5.0) 105 (6.7)

Breech birthi (denominator is newborns) 134 (0.7) 85 (0.8) 29 (0.4) 20 (1.3)

Multiple birth 54 (0.3) 38 (0.4) 13 (0.2) 3 (0.2)

aData come from the Midwives Alliance of North America Statistics Project, birth years 2004 to 2009. bFourteen women were missing data on race. cThree hundred ninety-six women were missing data on education. dEight women were missing data on partner status. eOne woman was missing data on parity. fIncludes women with 2-4 previous births only; primiparous, grand multiparous, and multiparous are mutually exclusive groups. gGrand multiparous is defined as at least 5 previous births (live births or stillbirths after 20 weeks’ gestation). hEight women were missing data on history of cesarean. iSixty-one newborns were missing data on presentation.

publications, focused exclusively on results of randomized controlled trials (RCTs), of which there are only a few compar- ing waterbirth to nonwaterbirth43,44—and these were small and underpowered. Another problem inherent in these stud- ies, as well as in any large, future, hypothetical RCTs, is that women in the waterbirth group whose labors become com- plicated are often directed to discontinue immersion,45,46 but women randomized to nonwaterbirth would not suddenly be asked to get in the water. This one-way noncompliance with the assigned intervention group could introduce differential misclassification bias, making interpretation of intention-to- treat results problematic. This bias is also apparent in obser- vational studies: womenwho develop complications get out of the tub, but rarely do womenwho never intended a waterbirth suddenly decide to get in. To our knowledge, we are the first to address this issue by reporting separately on the intended waterbirth group.

These methodologic issues, taken together, led the Cochrane review’s authors to state unprecedentedly (the Cochrane Group is well known for their randomized trial preference) that: “Large audits and cohort studies should be undertaken in units which provide a pool facility to provide evidence for practice.”18 In this article, we have presented re- sults from just such a large cohort study: 18,397 neonates, 35% (n= 6534) of whowere born underwater. Furthermore, this is the first waterbirth study to report results for a US population.

We found that neonates born underwater were not at in- creased risk of adverse outcomes; on the contrary, they fared better than their nonwaterbirth counterparts on all outcomes. Rather than being a true benefit of waterbirth, however, this finding is likely secondary to the misclassification bias de- scribed above: outcomes in the intended waterbirth group were uniformly worse than those observed in the (planned) nonwaterbirth group. Taken together, these findings suggest

Journal of Midwifery &Women’s Health � www.jmwh.org 15

Table 2. Comparison of Newborn Outcomes in Newborns Born Underwater (Waterbirth), Newborns Not Born Underwater (Nonwaterbirth), and NewbornsWhoseMothers Intended a Waterbirth But Did Not Have One (IntendedWaterbirth)a

Frequencies Adjusted Results

Nonwaterbirth Waterbirth

Intended

Waterbirth Waterbirth IntendedWaterbirth

n = , n =  n =  Nonwaterbirth aOR ( CI) P Value aOR ( CI) P Value 5-minute Apgar score<7,b n (%)

No 10,125 (98.8) 6451 (99.0) 1526 (97.6) –

Yes 120 (1.2) 66 (1.0) 38 (2.4) 1.0 0.88 (0.65, 1.19) .42 2.02 (1.40, 2.93) � .001

Transfer to the hospital during the postpartumperiod for a neonatal indication,c n (%)

No 10,094 (98.8) 6463 (99.5) 1532 (98.4) –

Yes 118 (1.2) 34 (0.5) 25 (1.6) 1.0 0.46 (0.32, 0.68) � .001 1.33 (0.86, 2.06) .20

Infant hospitalization, first 6 weeks,d n (%)

No 9809 (95.5) 6301 (96.6) 1484 (94.5) –

Yes 465 (4.5) 220 (3.4) 87 (5.5) 1.0 0.75 (0.63, 0.88) � .001 1.21 (0.95, 1.53) .12

NICU admission, first 6 weeks,e n (%)

No 10,013 (97.6) 6426 (98.6) 1521 (97.1) –

Yes 242 (2.4) 90 (1.4) 45 (2.9) 1.0 0.59 (0.46, 0.76) � .001 1.18 (0.85, 1.63) .32

n (rate/1000) n (rate/1000) N (rate/1000)

Early neonatal death (7 completed days)f

No, n 10,280 6532 1569

Yes, n

(rate/1000)

6g (0.58/1000) 2h (0.31/1000) 2i (1.27/1000) k k k

Late neonatal death (at least 7, but not yet 28, completed days)m

No, n 10,276 6529 1568

Yes, n

(rate/1000)

3n (0.29/1000) 1p (0.15/1000) 1q (0.64/1000) k k k

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; NICU, neonatal intensive care unit. aRaw data as well as adjusted odds ratios (aOR) and 95% confidence intervals (95% CI) are presented; the nonwaterbirth group serves as the reference category. Analyses adjust for primiparity. bSeventy-one neonates were missing data on 5-minute Apgar score. cAcross all 3 groups, the most common reasons for transfer were respiratory distress and/or low Apgar score. One hundred thirty-one neonates were missing data on neonatal transfer. dThirty-one neonates were missing data on hospitalizations in the first 6 weeks. eSixty neonates were missing data on NICU admissions in the first 6 weeks. fThere are no missing data for the death variables. The 4 nonwaterbirth and 2 intended waterbirth neonates not accounted for here died of autopsy-confirmed congenital anomalies that are incompatible with life and were removed from analysis, as is customary when examining fetal/neonatal death as an outcome. gCauses of death: 1) hypoxia (neonate never attempted to breathe; life support removed after 2 days; no autopsy); 2) patent ductus arteriosis was official cause but no autopsy; 3) prelabor brain damage (tight nuchal cord; neonate never attempted to breathe; life support removed after 2 days; no autopsy but brain damage was receiving obstetrician’s diagnosis); 4) acute chorioamnionitis (autopsy-confirmed); 5) cord rupture/hemorrhage (autopsy-confirmed); and 6) unknown (no autopsy; official cause was cardiac failure). hCauses of death: 1) hypoxic ischemic encephalopathy (confirmed with magnetic resonance imaging on day 4; life support subsequently removed); and 2) congestive heart failure (autopsy-confirmed). iCauses of death: 1) placental abruption (placenta came out with neonate; no autopsy); and 2) shoulder dystocia with cord compression (autopsy-confirmed). kCell counts too low to produce reliable effect estimates. mThere are no missing data for the death variables. The one nonwaterbirth and 2 waterbirth neonates not accounted for here died of autopsy-confirmed congenital anomalies that are incompatible with life and were removed from analysis, as is customary when examining fetal/neonatal death as an outcome. Also not included here are neonates who died during the early neonatal period; the denominator for late neonatal death rates has been limited to those infants still at risk of the outcome. nCauses of death: 1) cord prolapse (no autopsy); 2) unknown (neonate never attempted to breathe and did not respond to resuscitation; no autopsy; this woman had a subsequent neonate with the same issue); and 3) unknown (no autopsy). pCause of death: unknown (no autopsy). qCause of death: official cause was hypoxia, but there was no autopsy.

that, when potential complications arose during labor, the at- tending midwife recommended that the woman discontinue water immersion, perhaps to allow for closer monitoring of fetal heart tones or to visualize fetal scalp color, etc. Thus, the intended waterbirth category reflects a cohort of women who were experiencing more complicated labors or births. It is not surprising that the outcomes forwomen and neonates in the intended waterbirth category were worse. Following, it is also not surprising that the neonates of women who remained in the tub or pool (the waterbirth group) had better out-

comes: all of the nonstraightforward (ie, potentially elevated risk) mother–newborn dyads had been asked to get out of the tub. The substantially improved outcomes observed among neonates who were born underwater is almost certainly an artifact of this unavoidable but appropriate clinical manage- ment. The nonwaterbirth comparison (reference) group con- sists of women who never wanted (or who did not have access to)waterbirth. It seems likely that some of thesewomenwould have developed complications during labor that would have led to discontinuing immersion, had they been laboring in

16 Volume 61, No. 1, January/February 2016

Table 3. Maternal Outcomes ofWomenWhoHad a Waterbirth, Nonwaterbirth, or IntendedWaterbirth But Gave Birth Outside of the Planned Water Immersiona

Frequencies Adjusted Results

Nonwaterbirth

n = ,

Waterbirth

n = 

Intended

Waterbirth

n =  Waterbirth IntendedWaterbirth

n () n () n () Nonwaterbirth aOR ( CI) P Value aOR ( CI) P Value Maternal genital tract trauma (any)b

No 5162 (50.7) 3186 (49.3) 581 (37.4) –

Yes 5018 (49.3) 3272 (50.7) 973 (62.6) 1.0 1.11 (1.04-1.18) .002 1.67 (1.49-1.87) � .001

Postpartum transferc

No 9990 (98.0) 6415 (98.8) 1514 (97.2) –

Yes 199 (2.0) 80 (1.2) 43 (2.8) 1.0 0.65 (0.50, 0.84) .001 1.36 (0.97, 1.90) .07

Maternal hospitalization, first 6 weeksd

No 9890 (96.6) 6354 (97.6) 1502 (95.9) –

Yes 343 (3.4) 154 (2.4) 65 (4.1) 1.0 0.72 (0.59, 0.87) � .001 1.17 (0.89, 1.54) .21

Maternal postpartum reproductive tract infectione

No 10,002 (97.8) 6389 (98.2) 1525 (97.6) –

Yes 221 (2.2) 119 (1.8) 38 (2.4) 1.0 0.87 (0.69, 1.09) .23 1.08 (0.78, 1.52) .69

Abbreviations: aOR, adjusted odds ratios; CI, confidence interval. aRaw data as well as aOR and 95% CIs (95% CI) are presented; all analyses adjust for primiparity. bOne hundred fifty-one women were missing data on genital tract trauma. cOne hundred and two women were missing data on postpartum transfer. dThirty-five women were missing data on hospitalization in the first 6 weeks postpartum. eForty-nine women were missing data on postpartum reproductive tract infection (includes endometrial, perineal, or uterine infection in the first 6 weeks postpartum).

water. Therefore, the nonwaterbirth reference group contains bothwomenwho could have hadwaterbirths and thosewhose labors were more complicated, and who therefore would have been asked to get out of the tub; for women who plan water- births, these 2 exposure groups have been separated into wa- terbirth and intended waterbirth.

Nonetheless, the neonates in the waterbirth group were born underwater, and we found no evidence of any sub- sequent increased risk for any adverse neonatal outcome. Safety of waterbirth (as opposed to merely laboring in wa- ter) is currently the issue under heaviest scrutiny in the recent ACOG/AAP guidelines,3 as well as the evidence gap identi- fied by the authors of the Cochrane Review.18 Based on the strength of evidence presented here, we cannot conclude that waterbirth is beneficial; however, based on our results, wa- terbirth certainly is not harmful to the neonate. Our findings also suggest that US midwives attending home and birth cen- ter births on the whole are appropriately managing women as they labor and birth in water.

Our newborn morbidity results are consistent not only with those in the Cochrane review18 but also with those of most other published observational studies, which al- most universally report no adverse outcomes for waterbirth neonates.26–35,46,47 By contrast, Carpenter et al48 reported an increase in respiratory distress symptoms among waterbirth neonates; however, the sample size was small (N = 38), and the authors reflect that ascertainment bias may have played a prominent role. Hawkins,49 in a small (N = 32), non- randomized cohort study, reported more neonatal infections (and therefore greater morbidity) in the waterbirth group, al- though during the study period a lack of adherence to tub-

cleaning protocols required intervention by the hospital’s in- fection control team, likely affecting the study’s results. All other published reports of increased neonatal morbidity fol- lowing waterbirth are case reports or case series without con- trol groups.1,14,21–24,45,50–57

We lacked sufficient power to calculate reliable adjusted odds ratios for early and late neonatal deaths. Nonetheless, there was no evidence of elevated mortality among the wa- terbirth group, a finding that is consistent with our morbidity results as well as with those published by others.26,58 Further- more, of the 3 neonatal deaths in the waterbirth group, none were attributed to causes that might stem from the waterbirth (ie, no drowning, no cord avulsion, no respiratory distress).

Women in our sample who gave birth while immersed in water had no increased risk of adverse outcomes except for genital tract trauma: they experienced an 11% increase in odds of genital tract trauma, although without a discernible pattern as to trauma location and severity. Cortes et al59 also reported an increased risk of perineal trauma for women who gave birth while immersed in water. The Cortes study, however, contradicts the bulk of the literature, which has reported re- duced risk of genital tract trauma, or no change in risk, among women who had a waterbirth.26–28,32–34,44,46,47,60 Because ours is the first large cohort study to report results in a US popu- lation, and the largest cohort study to date, our finding of in- creased risk of maternal genital tract trauma requires further investigation. In the meantime, clinicians should discuss the possible risk of genital tract trauma as part of shared decision making and the informed consent process around waterbirth.

Other studies have also reported no increased risk of infection among women who have waterbirths,26,29,32,61

Journal of Midwifery &Women’s Health � www.jmwh.org 17

Table 4. Comparison of Location and Severity ofMaternal Birth Canal Trauma byWhether or Not the BirthWas a Waterbirth, a Nonwaterbirth, or an IntendedWaterbirtha

Frequencies Adjusted Results

Nonwaterbirth

n = ,

Waterbirth

n = 

Intended

Waterbirth

n =  Waterbirth IntendedWaterbirth

n () n () n ()

Non-wat-

erbirth aOR ( CI) P Value aOR ( CI) P Value None 5162 (50.6) 3186 (48.9) 581 (37.3) –

Episiotomy only 21 (0.2) 1 (0.02) 11 (0.7) b b b

1st- or 2nd-degree

perineal only

3,140 (30.8) 1980 (30.4) 547 (35.1) 1.0 1.06 (0.98, 1.14) .13 1.52 (1.34, 1.73) � .001

3rd- or 4th-degree

perineal only

63 (0.6) 27 (0.4) 13 (0.8) 1.0 0.79 (0.50, 1.24) .30 1.73 (0.94, 3.18) .07

Mild labial only 551 (5.4) 487 (7.5) 101 (6.5) 1.0 1.50 (1.32, 1.71) � .001 1.74 (1.39, 2.17) � .001

More severe: labial

only

142 (1.4) 87 (1.3) 27 (1.7) 1.0 1.10 (0.84, 1.45) .50 1.61 (1.06, 2.46) .028

Other trauma

requiring repair

21 (0.2) 6 (0.09) 2 (0.1) b b b

Trauma at multiple

sites

1080 (10.6) 684 (10.5) 263 (16.9) 1.0 1.10 (0.99, 1.23) .085 2.09 (1.77, 2.47) � .001

Trauma NOS 63 (0.6) 61 (0.9) 14 (0.9) 1.0 1.60 (1.12, 2.28) .01 1.96 (1.09, 3.52) .025

Abbreviations: aOR, adjusted odds ratio; CI, confident interval; NOS, not otherwise specified. aRaw data as well as aOR and 95% CIs (95% CI) are presented. The multinomial logistic regression model controlled for primiparity. bCell counts too low to produce reliable effect estimates.

although one study found an elevated risk.35 In our study, we found no association between waterbirth and maternal per- ineal, uterine, or endometrial infection.

Women in the intended waterbirth category had in- creased odds of all adverse outcomes, although only genital tract trauma was statistically significant. As discussed above, this finding could be an artifact of the misclassification bias wherein women experiencing complicated labors were asked to discontinue water immersion. Alternatively, perhaps midwives attempted to speed the second stage in cases of fetal distress by cutting an episiotomy or allowing/encouraging tearing by verbally coaching a woman to continue to push, or push harder, in the final moments of birth. Alternatively, some underlying condition or risk factor (eg, smoking or poor nutrition) that predisposes to fetal complications could also place the woman at risk for reduced tissue integrity.62

Strengths and Limitations

This research uses data from the MANA Stats Project, birth years 2004 to 2009. Although this dataset has many strengths, including evidence of reliability and validity, data collection procedures that preclude a midwife from entering data only from births with good outcomes, a large sample size, ex- tremely high participation by women, and large numbers of covariables,36 it does have limitations. Chief among these is the fact that data are collected by a voluntary sample of midwives. We previously estimated that approximately 30% of midwives attending home or birth center births in the United States contribute data,36 and it is certainly possible that contributing midwives are not representative of all midwives

attending US home and birth center births.We cannot predict how this would affect our results on waterbirth.

Additionally, our results are based on women who gave birth at home or in a birth center. These women likely are not representative of the US childbearing population as a whole. Additionally, they by definition gave birth withminimalmed- ical intervention, which also is not the norm for US women, for whom the majority of births occur in the hospital under biomedical management.63,64 However, because the few US hospitals that do offer birth tubs typically prohibit their use by higher risk women and by women who utilize epidural anal- gesia or other ongoing interventions,28 and because even in hospitals the vast majority of laboring women are low-risk at the onset of labor, it seems reasonable that our results could be applied in US hospital settings where waterbirth is offered.

CONCLUSION

This study is the largest cohort study to date on waterbirth, the first large study from the United States, and the first to re- port separately on outcomes for mother–newborn dyads who did not complete a planned waterbirth because of risk factors that arose in the intrapartum period. Our results indicate that waterbirth does not confer an increased risk of morbidity or mortality for the newborn, but women completing the second stage immersed in water may experience more genital tract trauma. Our results are congruent with findings from studies in other settings, and contrary to the recently published ACOG/AAP clinical guidelines, suggest that waterbirth is a reasonably safe option for use in low-risk, low-intervention

18 Volume 61, No. 1, January/February 2016

births—especially when the risks associated with other forms of pharmacologic pain management are considered.

AUTHORS

Marit L. Bovbjerg, PhD, MS, is Instructor in the Epidemiol- ogy Program, College of Public Health and Human Sciences at Oregon State University in Corvallis, Oregon. She is also Director of Data Quality for the Midwives Alliance of North America Division of Research.

Melissa Cheyney, PhD, CPM, LDM, is Associate Professor of medical anthropology and reproductive biology in the De- partment of Anthropology at Oregon State University in Cor- vallis, Oregon. She is also a certified professional midwife, licensed in the State of Oregon, and Chair of the Division of Research for the Midwives Alliance of North America.

Courtney Everson,MA, PhD, is aMedical Anthropologist and the Graduate Dean at the Midwives College of Utah in Salt LakeCity, Utah. She is also theDirector of Research Education for the MANA Division of Research.

CONFLICT OF INTEREST

The authors have no conflicts of interest to disclose.

ACKNOWLEDGMENTS

Wewould like to thank themany volunteers who have worked to make the MANA Stats project a success, particularly Bruce Ackerman, Wendy Gordon, Ellen Harris-Braun, Saraswathi Vedam, and Trinlie Wood. We also acknowledge the hard work of theMANADivision of Research Coordinating Coun- cil, as well as that of the MANA Board of Directors, without whose support this project would not be possible. Ongoing data collection efforts for the MANA Stats Project are funded by MANA and the Foundation for the Advancement of Mid- wifery.

REFERENCES

1.Bowden K, Kessler D, Pinette M,Wilson E. Underwater birth: Missing the evidence or missing the point? Pediatrics. 2003;112(4):972-973.

2.PinetteMG,Wax J,Wilson E. The risks of underwater birth.Am J Ob- stet Gynecol. 2004;190(5):1211-1215. doi:10.1016/j.ajog.2003.12.007.

3.Committee on Obstetric Practice, American Academy of Pedi- atrics. Committee Opinion: Immersion in Water During Labor and Delivery.; 2014. Available at: http://www.acog.org/Resources And Publications/Committee Opinions/Committee on Obstetric Practice/Immersion in Water During Labor and Delivery. Accessed March 29, 2014.

4.Schafer R. Umbilical cord avulsion in waterbirth. J Midwifery Wom- ens Health. 2014;59(1):91-94. doi:10.1111/jmwh.12157.

5.Fritschel E, Sanyal K, Threadgill H, Cervantes D. Fatal legionellosis after water birth, Texas, USA, 2014. Emerg Infect Dis. 2015;21(1):130- 132. doi:10.3201/eid2101.140846.

6.Water birth? Don’t do it. Adv Neonatal Care Off J Natl Assoc Neona- tal Nurses. 2014;14(3):139. doi:10.1097/ANC.0000000000000098.

7.Harper B. Birth, bath, and beyond: The science and safety of water im- mersion during labor and birth. J Perinat Educ. 2014;23(3):124-134. doi:10.1891/1058-1243.23.3.124.

8.WeaverMH.Water birth in the hospital setting.NursWomens Health. 2014;18(5):365-369. doi:10.1111/1751-486X.12144.

9.UK Department of Health, UK National Health Service. Maternity Matters: Choice, Access and Continuity of Care in a Safe Service. Available at: http://www.dh.gov.uk/prod consum dh/groups/dh digitalassets/@dh/@en/documents/digitalasset/dh 074199.pdf. Accessed January 25, 2012.

10.Zanetti-Dällenbach R, Lapaire O, Maertens A, Holzgreve W, Hösli I. Water birth, more than a trendy alternative: A prospective, observational study. Arch Gynecol Obstet. 2006;274(6):355-365. doi:10.1007/s00404-006-0208-1.

11.Royal College of Obstetricians and Gynaecologists, Royal College of Midwives. Immersion in Water during Labour and Birth; 2012. http://www.waterbirth.org/assets/documents/rcog rcm birth in water.pdf. Accessed September 29, 2015.

12.Henderson J, Burns EE, Regalia AL, Casarico G, Boulton MG, Smith LA. Labouring women who used a birthing pool in obstetric units in Italy: Prospective observational study. BMC Pregnancy Childbirth. 2014;14:17. doi:10.1186/1471-2393-14-17.

13.Burns E, Greenish K. Water birth. Pooling information. Nurs Times. 1993;89(8):47-49.

14.Odent M. Birth under water. Lancet. 1983;2(8365-8366):1476-1477. 15.Mackey MM. Use of water in labor and birth. Clin Obstet Gynecol.

2001;44(4):733-749. 16.Torkamani S, Kangani F, Janani F. The effects of delivery in water on

duration of delivery and pain compared with normal delivery. Pak J Med Sci. 2010;26(3):551-555.

17.Daniels K. Water birth: The newest form of safe, gentle, joyous birth. J Nurse Midwifery. 1989;34(4):198-205.

18.Cluett ER, Burns E. Immersion in water in labour and birth. Cochrane Database Syst Rev. 2009;(2):CD000111. doi:10.1002/ 14651858.CD000111.pub3.

19.Nutter E, Meyer S, Shaw-Battista J, Marowitz A. Waterbirth: An in- tegrative analysis of peer-reviewed literature. J Midwifery Womens Health. 2014;59(3):286-319. doi:10.1111/jmwh.12194.

20.Johnson P. Birth under water: To breathe or not to breathe. Br J Obstet Gynaecol. 1996;103(3):202-208.

21.Hagadorn JI, Guthrie E, Atkins VA, DeVine CW, Hamilton L. Neonatal aspiration pneumonitis and endotracheal colonization with Burkholderia pickettii following home water birth. Pediatrics. 1997;100(Suppl):506.

22.Parker PC, Boles RG. Pseudomonas otitis media and bacteremia fol- lowing a water birth. Pediatrics. 1997;99(4):653.

23.Schroeter K. Water births: A naked emperor. Pediatrics. 2004;114(3):855-858. doi:10.1542/peds.2004-0145.

24.Church LK. Water birth: One birthing center’s observations. J Nurse Midwifery. 1989;34(4):165-170. doi:10.1016/0091-2182(89)90076-1.

25.Geissbuehler V, Eberhard J, Lebrecht A. Waterbirth: Water tem- perature and bathing time: Mother knows best! J Perinat Med. 2002;30(5):371-378. doi:10.1515/JPM.2002.058.

26.Geissbuehler V, Stein S, Eberhard J. Waterbirths compared with landbirths: An observational study of nine years. J Perinat Med. 2004;32(4):308-314. doi:10.1515/JPM.2004.057.

27.Thoeni A, Zech N, Moroder L, Ploner F. Review of 1600 wa- ter births. Does water birth increase the risk of neonatal in- fection? J Matern-Fetal Neonatal Med. 2005;17(5):357-361. doi:10.1080/14767050500140388.

28.Thöni A, Moroder L. Waterbirth: A safe and natural delivery method. Experience after 1355 waterbirths in Italy.Midwifery Today Int Mid- wife. 2004;(70):44-48.

29.Waldenström U, Nilsson CA. Warm tub bath after spontaneous rup- ture of the membranes. Birth Berkeley Calif. 1992;19(2):57-63.

30.Zanetti-Daellenbach RA, Tschudin S, Zhong XY, Holzgreve W, La- paire O, Hösli I. Maternal and neonatal infections and obstetri- cal outcome in water birth. Eur J Obstet Gynecol Reprod Biol. 2007;134(1):37-43. doi:10.1016/j.ejogrb.2006.09.012.

31.SchornMN,McAllister JL, Blanco JD.Water immersion and the effect on labor. J Nurse Midwifery. 1993;38(6):336-342.

32.Lenstrup C, Schantz A, Berget A, Feder E, Rosenø H, Hertel J. Warm tub bath during delivery.ActaObstet Gynecol Scand. 1987;66(8):709- 712.

Journal of Midwifery &Women’s Health � www.jmwh.org 19

33.Burke E, Kilfolyle A. Retrospective comparative study of water birth and bedbirth.Midwives J. 1995;(January):3-7.

34.Otigbah CM, Dhanjal MK, Harmsworth G, Chard T. A retrospective comparison of water births and conventional vaginal deliveries. Eur J Obstet Gynecol Reprod Biol. 2000;91(1):15-20.

35.Eriksson M, Ladfors L, Mattsson LA, Fall O. Warm tub bath during labor. A study of 1385 women with prelabor rupture of the mem- branes after 34 weeks of gestation. Acta Obstet Gynecol Scand. 1996;75(7):642-644.

36.Cheyney MJ, Bovbjerg ML, Everson C, Gordon W, Hannibal D, Vedam S. Development and validation of a national data registry for planned home and birth center births: TheMidwives Alliance ofNorth America Statistics Project 2.0 dataset. J Midwifery Womens Health. 2014;59(1):8-16. doi: 10.1111/jmwh.12165.

37.Stapleton SR, Osborne C, Illuzzi J. Outcomes of care in birth cen- ters: Demonstration of a durablemodel. J MidwiferyWomens Health. 2013;58(1):3-14. doi:10.1111/jmwh.12003.

38.JohnsonKC,Daviss B-A.Outcomes of planned home births with certi- fied professional midwives: Large prospective study inNorth America. BMJ. 2005;330(7505):1416. doi:10.1136/bmj.330.7505.1416.

39.Harper B. Waterbirth International. Available at: http://www. waterbirth.org/. Accessed August 21, 2013.

40.Cheyney M, Bovbjerg M, Everson C, Gordon W, Hannibal D, Vedam S. Outcomes of care for 16,924 planned home births in the United States: The Midwives Alliance of North America Statistics Project, 2004 to 2009. J Midwifery Womens Health. 2014;59(1):17- 27. doi:10.1111/jmwh.12172.

41.Shrier I, Platt RW. Reducing bias through directed acyclic graphs. BMC Med Res Methodol. 2008;8:70. doi:10.1186/1471-2288- 8-70.

42.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiol Camb Mass. 1999;10(1):37-48.

43.Chaichian S, Akhlaghi A, Rousta F, SafaviM. Experience of water birth delivery in Iran. Arch Iran Med. 2009;12(5):468-471.

44.Woodward J, Kelly SM. A pilot study for a randomised controlled trial of waterbirth versus land birth. BJOG Int J Obstet Gynaecol. 2004;111(6):537-545. doi:10.1111/j.1471-0528.2004.00132.x.

45.Deans AC, Steer PJ. Labour and birth in water. Temperature of pool is important. BMJ. 1995;311(7001):390-391.

46.Baxter L. What a difference a pool makes: Making choice a reality. Br J Midwifery. 2006;14(6):368-372.

47.Geissbühler V, Eberhard J. Waterbirths: A comparative study. A prospective study on more than 2,000 waterbirths. Fetal Diagn Ther. 2000;15(5):291-300. doi:21024.

48.Carpenter L, Weston P. Neonatal respiratory consequences from water birth. J Paediatr Child Health. 2012;48(5):419-423. doi:10.1111/j.1440-1754.2011.02241.x.

49.Hawkins S. Water vs conventional births: Infection rates compared. Nurs Times. 1995;91(11):38-40.

50.Cro S, Preston J. Cord snapping at waterbirth delivery.Br J Midwifery. 2002;10(8):494-497.

51.Franzin L, Cabodi D, Scolfaro C, Gioannini P. Microbiological inves- tigations on a nosocomial case of Legionella pneumophila pneumonia associated with water birth and review of neonatal cases. Infez Med. 2004;12(1):69-75.

52.Mammas IN, Thiagarajan P. Water aspiration syndrome at birth - re- port of two cases. J Matern-Fetal Neonatal Med. 2009;22(4):365-367. doi:10.1080/14767050802556067.

53.Nguyen S, Kuschel C, Teele R, Spooner C. Water birth: A near- drowning experience. Pediatrics. 2002;110(2 Pt 1):411-413.

54.Rawal J, Shah A, Stirk F, Mehtar S.Water birth and infection in babies. BMJ. 1994;309(6953):511.

55.Sotiridou E, Mukhopadhyay S, Clarke P. Neonatal aspiration syn- drome complicating a water birth. J Obstet Gynaecol. 2010;30(6):631- 633. doi:10.3109/01443615.2010.491569.

56.Alderdice F, Renfrew M, Marchant S, et al. Labour and birth in water in England and Wales. BMJ. 1995;310(6983):837.

57.Barry CN. Water births. Could saline in the pool reduce the potential hazards? BMJ. 1995;310(6994):1602.

58.Gilbert RE, Tookey PA. Perinatal mortality and morbidity among ba- bies delivered in water: Surveillance study and postal survey. BMJ. 1999;319(7208):483-487.

59.Cortes E, Basra R, Kelleher CJ. Waterbirth and pelvic floor injury: A retrospective study and postal survey using ICIQ modular long form questionnaires. Eur J Obstet Gynecol Reprod Biol. 2011;155(1):27- 30. doi:10.1016/j.ejogrb.2010.11.012.

60.Pagano E, De Rota B, Ferrando A, Petrinco M, Merletti F, Gregori D. An economic evaluation of water birth: The cost- effectiveness of mother well-being. J Eval Clin Pract. 2010;16(5):916- 919. doi:10.1111/j.1365-2753.2009.01220.x.

61.Robertson PA, Huang LJ, Croughan-Minihane MS, Kilpatrick SJ. Is there an association between water baths during labor and the devel- opment of chorioamnionitis or endometritis? Am J Obstet Gynecol. 1998;178(6):1215-1221.

62.Smith LA, Price N, Simonite V, Burns EE. Incidence of and risk factors for perineal trauma: A prospective observational study. BMC Preg- nancy Childbirth. 2013;13(1):59. doi:10.1186/1471-2393-13-59.

63.Declercq E, Sakala C, CorryM,ApplebaumS.Listening to Mothers II: Report of the Second National US Survey of Women’s Childbearing Experiences. Available at: www.childbirthconnection.com. Accessed April 6, 2009.

64.MacDorman MF, Menacker F, Declercq E. Trends and characteristics of home and other out-of-hospital births in the United States, 1990- 2006. Natl Vital Stat Rep. 2010;58(11):1-14, 16.

20 Volume 61, No. 1, January/February 2016