Newborn bathing is practised in many contexts on the day of birth as is thought to eliminate pollutants from the skin and prevent infection. However, this practice is based on cultural beliefs rather than evidence . There are various methods of bathing neonates like tub bathing, sponge bathing, swaddled bathing, and under running-water bathing . The two commonly used methods in hospitals include washing the neonate with a wet cloth (sponge-bath) or immersion of the body in tubs (tub-bath). Immersion bath has been shown to decrease heat loss in neonates, thereby decreasing the predisposition to hypothermia .
Bathing is a stressful procedure for a neonate and an early first bath has been shown to destabilize the vitals in apparently healthy neonates, especially temperature, glucose levels, and respiratory status [4,5]. Depending on severity, hypothermia can lead to poor activity, lethargy, and difficulty feeding which further predisposes the neonate to hypoglycaemia. Other consequences of the first bath may include tachypnoea, an increase in pulmonary vascular resistance, hypoxia, and pulmonary haemorrhage, secondary to hypothermia. Moreover, the timing of the first bath may be crucial for establishing breastfeeding and improving exclusive breastfeeding (EBF) rates. In many hospitals, the first bath is given soon after birth, separating the neonate from the mother, depriving her/him of the opportunity of skin-to-skin contact (SSC) with the mother, which can potentially hamper breastfeeding rates .
The World Health Organization (WHO) recommends delaying bathing until 24 hours (h) after birth, and when not possible, to be delayed for at least 6 h . However, this recommendation was based on expert consensus. Delaying the first bath may allow time for a neonate’s vitals to stabilize after birth. A pilot study showed that delaying the first bath until 24 h of life was associated with benefits from vernix caseosa on the skin and adequate time for SSC with the mother’s participation in her child’s bathing . Improved SSC with the mother may improve breastfeeding rates, body temperatures, and blood glucose levels in neonates .
There is currently no evidence-based recommendation for the timing of the first bath in healthy neonates. The objective of this review was to determine the impact of a first bath delayed for at least 24 h, compared to an early bath within the first 24 h, on neonatal mortality, hypothermia, hypoglycaemia, and exclusive breastfeeding rates in term healthy newborns.
Randomized controlled trials (RCTs) including cluster randomized trials or quasi-randomized trials in human neonates were eligible for this review. If the number of RCTs was found to be inadequate (<3) or the optimal information size was not met, we included observational studies (before-after/cohort/case-control/cross-sectional study analysed like case-control). The study population were term neonates (up to 28 completed days of life). We excluded the studies if most of the participants (at least 50%) were either low birth weight or preterm neonates. Studies were included if delayed first bath (after 24 h of age) was compared to early first bath (within 24 h of age) in neonates. The outcomes of interest were neonatal mortality (all-cause death in the first 28 days of life), systemic infections (sepsis, pneumonia, or possible serious bacterial infection), any respiratory morbidity (respiratory distress, pulmonary hemorrhage, or pulmonary hypertension), hypothermia (recorded temperature less than 36.5°C or 97.7°F), hypoglycaemia (recorded glucose level less than 2.5 mmol/L or 45 mg/dL), timing of breastfeeding initiation and exclusive breastfeeding rate at discharge and at 6 months of age.
The databases were searched independently by two authors (MP and BB). The search was conducted in the following databases: MEDLINE (1966 onwards) via PubMed, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library), Embase (1947 onwards), and CINAHL (1981 onwards). We conducted the first search till March 31, 2020, which was later updated till November 30, 2021. Searches were limited to human studies. There were no language or publication date restrictions. We also searched related conference proceedings (eg, Pediatric Academic Societies abstracts), clinical trial registries (eg, clinicaltrials.gov), and the reference list of all trials/studies identified. We contacted experts or researchers in the field, if necessary, for information on unpublished and ongoing trials. The search strategy is provided in Appendix S1 in Online Supplementary Document.
Data extraction and management
Two authors (MP and BB) extracted data independently using a pilot-tested data collection form to collect information on study setting, design, methods, participants, intervention (eg, timing and type of bath, duration, and the person implementing intervention), co-interventions (eg, use of warmer after bath), outcomes, and treatment effects from each included study. If data from study reports were insufficient, we contacted study authors to request further information or any clarifications, if required. Additional data was obtained by personal communication from the author for one study .
Assessment of risk of bias in included studies
Two authors (MP and BB) independently assessed the methodological quality of the selected trials/ studies. Quality assessment was undertaken using the Cochrane Risk of bias (RoB 2.0) tool for randomized trials, and using the ROBINS-I tool for observational studies [10,11]. Any disagreements between the authors were resolved by discussion.
Meta-analysis was performed using Stata 15.1 (StataCorp, College Station, TX, USA). Pooled estimates for categorical outcomes were calculated from the relative risk (RR) or odds ratios (OR) and their 95% confidence intervals (CIs) by the generic inverse variance method. We used the adjusted RR and OR from the studies where available. We used intention-to-treat analysis to calculate effect sizes for studies that provided the number of participants non-adherent to intervention with separate outcome data for these participants. We examined heterogeneity between study results by inspecting the forest plots and quantifying the impact of heterogeneity using the I2 statistic. We pooled the results of individual studies using the fixed-effect model if I2 was ≤60%. If we detected significant heterogeneity (I2>60%), we explored the possible causes by examining the population, intervention, outcome, and settings. If there was significant clinical heterogeneity, we used subgroup analyses or deferred from pooling the study results. If there was no clinical heterogeneity, we used the random-effects model for meta-analysis. We used the GRADEpro software for assigning the quality of evidence .
We included 16 studies in the review, 12 of which were included for meta-analysis (Figure 1, Table 1). The reasons for the exclusion of four studies from meta-analysis and their results have been summarized in Appendix S2 in Online Supplementary Document.
Figure 1. PRISMA flowchart depicting the selection of studies included in the review.
Table 1. Characteristics of the studies included in the review
HIC – high-income country, LIC – low-income country, LMIC – lower middle-income country, RCT – randomized controlled trial, UMIC – upper middle-income country, wk – week, h – hour, mo – month, BW – birth weight, gest – gestation, kg – kilograms, NICU – neonatal intensive care unit, OR – odds ratio, AOR – adjusted odds ratio, CI – confidence interval, mg/dL – milligram per decilitre, EBF – exclusive breastfeeding
Due to variation across studies in the definitions of early and delayed bath, we performed the analysis as follows: a pre-specified analysis was done for studies which defined early and delayed bath as in our review protocol, ie, delayed bath after 24 h of age compared to early bath before 24 h of age (ie, first bath at >24 h vs ≤24 h of age); and a post-hoc analysis for the most commonly used timing of the first bath in included studies, ie, delayed bath after 6 h of age compared to early bath before 6 h of age (ie, first bath at >6 h vs ≤6 h of age).
The designs of the included studies were before-after (n = 11), case-control (n = 2), randomized trial (n = 1), non-randomized trial (n = 1) and cross-sectional (n = 1). Before-after studies compared neonatal outcomes pre- and post-implementation of a new policy for delaying first bath in the hospital. Mullany et al.  performed a case-control study based on data from large community-based RCTs in Nepal, looking at the association of risk factors (including timing of the first bath) and hypothermia. Another case-control study was conducted in Ethiopia to look at the maternal and child health-related predictors of under-five and infant mortality . A 3-arm randomized trial studied the effects of 2-, 6- and 24-hour baths on newborn stabilization and behaviour . In another 3-arm but non-randomized trial, the authors compared the effects of 3-, 6- and 9-hour baths on axillary temperature of the newborns . Using data from the Demographic and Health Surveys, Mallick et al  studied neonatal thermal care and umbilical cord care practices in South Asia along with their associated mortality data.
The included studies were conducted in high-income (Canada, UK, and USA, n = 12), upper middle-income (Lebanon; n = 1), lower middle-income (Nepal and Pakistan, n = 2) and low-income (Ethiopia, n = 1) countries. The facility-based studies (11 before-after studies and 2 trials) enrolled neonates admitted in the community hospitals or mother-baby units. Three studies were based on data from the community in low- and middle-income countries (LMICs).
The review included 16 studies involving 39 020 neonates. 12 studies involving 14 421 neonates contributed to the meta-analysis. The participants were term and near-term healthy neonates, without any major comorbidities (sickness, ineligibility for breastfeeding, anomalies, etc.). Population characteristics were not clearly mentioned in six studies [5,14,17,19,20,22]; however, there was no indication to suggest that the newborns were not healthy.
Details of intervention
In facility-based studies, the timing of first bath was noted either prospectively or collected retrospectively from patient records in the hospital. These details in case-control studies were available from survey data collected in interviews from the caretakers conducted after a significant time interval from the event (first bath).
Timing of delayed bathing
Delayed bathing was defined variably by the included studies (Table S1 in Online Supplementary Document). Six studies defined delayed bathing as first bath at least 24 h after birth [8,13,15,17,22,24]. Among these, the mean bathing timing was mentioned as 30 h in one study .
Five studies defined delayed bath as first bath at least 12 h after birth. The mean timing of first bath was 17.9 , 14 , and 13.5 h  in three studies and not specified in two studies [3,5]. The cut-off of 9 h was used for delayed bathing by one study (mean time 13 h) . In three-arm trials, one study compared 3-, 6- and 9-hour baths , while another study compared 2-, 6- and 24-hour baths . In these trials, we considered 9- and 24-hour baths as delayed, respectively, and baths at ≤6 h (2-, 3- or 6-hour baths) as “early” to uniformly align the definition of early bath (≤6 h) in meta-analysis. One study classified first bath timing into six categories: >24, 12-23.9, 6-11.9, 3-5.9, 1-2.9 and <1 hour . All groups were compared against “>24-hour group” for the outcome of hypothermia in this study. One study did not define the exact timing of baths but mentioned mean timings (16 h for delayed and 4.6 h for early bath) .
Adherence to intervention
Two studies mentioned the number of neonates who did not adhere to the intervention and were bathed earlier [3,23]. There was no mention of the adherence rate to delayed bathing policy in other studies.
Two studies included skin-to-skin contact after bath in the new delayed bathing policy [21,24]. Two studies introduced a policy change which involved delaying the first bath as well as changing the type of bath (immersion bath instead of sponge bath) [3,13]. Three studies mentioned the possibility of potential impact of baby friendly hospital initiative (BFHI) program on breastfeeding rates [18,23,24].
The timing of “early” bath also varied across the studies, with different cut-off definitions (Table S1 in Online Supplementary Document). The most common definition of early first bath was bath within the first 6 h of life (12 studies). In one study, first bath within 12 h after birth was considered as early bath . Neonates were bathed within 2-4 h after birth (early bath) in five studies, which was a routine care practice in the pre-implementation phase [5,13,17,18,21]. Some studies reported the mean timing of early bath rather than a cut-off as 1.9 h , 6.9 h  and 3.5 h . One study used six different time-periods for comparison . In two studies, we defined first bath at or within 6 h of life, ie, at 2, 3 or 6 h after birth as early [8,9].
In two studies, 25%-27% of the newborns did not adhere to the timing of early bathing (control) in pre-implementation phase and received their first baths later [3,23]. There was insufficient information in the rest of the studies to clearly rule out contamination in the pre-implementation (control) group.
The critical outcomes reported by the included studies were neonatal and infant mortality, hypothermia, hypoglycemia, and exclusive breastfeeding (EBF) rates at discharge. None of the included studies reported sepsis or possible serious bacterial infection, or exclusive breastfeeding rates at 6 months.
Neonatal and infant mortality was reported by single studies [19,22]. One study reported the effect of delayed bath on neonatal mortality by analysing the health survey data from 4115 participants in Bangladesh and Nepal . The authors included neonatal deaths after day one of life till 28 days (excluding deaths on day of birth because these were unlikely to be related to bath practices).
Hypothermia and hypoglycaemia were reported by eight and four studies, respectively, six and three of which could be included in meta-analysis, respectively. A neonate who had at least one episode of hypothermia or hypoglycaemia after bath during hospital stay was considered to have an event. One study defined hypoglycaemia as blood glucose <49 mg/dL ; however, we pooled the study results with other studies (which used WHO definition of blood glucose <45 mg/dL) due to similarity in cut-off values.
Eight studies reported EBF rates at discharge, seven of which could be included in meta-analysis. Most studies used same definition for EBF (having received only breastmilk and no formula, water, or glucose water during the birth hospitalization); however, some studies did not provide a clear definition of EBF [15–17].
Risk of bias in included studies
A summary of the risk of bias assessment is provided in Appendix S3 in Online Supplementary Document. Six studies were judged to be at critical risk, and 10 studies at serious risk of bias mostly due to confounding effect.
Effects of interventions
The results are summarized separately for pre-specified analysis (first bath at >24 h vs ≤24 h of age) and post-hoc analysis (first bath at >6 h vs ≤6 h of age) (Table 2).
Table 2. Summary of findings: Delayed bath vs early bath in term healthy newborns
CI – confidence interval, OR – odds ratio, h – hour
*Most of the pooled effect provided by studies at “critical risk of bias”.
†Wide confidence interval crossing the line of no effect.
‡Most of the pooled effect provided by studies at “serious risk of bias”.
§Significant heterogeneity (I2 statistics ≥60%).
¶The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
Pre-specified analysis: Delayed first bath (>24 h after birth) vs early bath (≤24h after birth)
Infant mortality was reported by one study . 23% of infants died in the delayed bathing (>24h) compared to 40% infant deaths in the early bathing group (≤24h) (adjusted OR = 0.46, 95% CI = 0.28-0.77; low certainty evidence).
Hypothermia during hospital stay was reported by one study . 7% of the newborns developed hypothermia in the delayed bathing (>24h) compared to 13% newborns in the early bathing group (≤24h) (OR = 0.50; 95% CI = 0.28-0.88; low certainty evidence).
EBF rate at discharge was reported by one study . 57% of newborns were breastfeeding exclusively at discharge in the delayed compared to 62% the early bathing group (OR = 0.81, 95% CI = 0.58-1.12; very low certainty evidence).
Post-hoc analysis: Delayed first bath (>6 h, ie, at or after 9, 12 or 24 h after birth) vs early bath (≤6 after birth)
We could pool the results of ten studies under this comparison because the mean timings of early first bath in all these studies were within the first six h of life. The mean timing of early bath was 6.9 h in one study but was included in meta-analysis due to proximity to the cut-off timing (6 h) .
Neonatal mortality was reported by one study . The unadjusted OR was 0.71 (95% CI = 0.30-1.67; very low certainty evidence) for delayed first bath (>6 h after birth) compared to early bath (within 6 h after birth).
Ten percent of the newborns, who bathed more than 6 h after birth, developed hypothermia compared to 17% of those bathed within 6 h after birth (OR = 0.23, 95% CI = 0.10-0.54; five studies, 3582 newborns; low certainty evidence; Figure 2).
Figure 2. Forest plot for post-hoc analysis: Delayed first bath (>6 hours, ie, at or after 9, 12 or 24 hours after birth) vs early bath (≤6 hours after birth) in term, healthy newborns. Outcome: Incidence of hypothermia.
Three studies including data on 2775 neonates reported hypoglycaemia. There was 61% decrease in the odds of hypoglycaemia in neonates who underwent delayed bath (>6 h) compared to early bath (≤6 h) (OR = 0.39, 95%CI = 0.23-0.66; low certainty evidence; Figure 3).
Figure 3. Forest plot for post-hoc analysis: Delayed first bath (>6 hours, ie, at or after 12 or 24 hours after birth) vs early bath (≤6 hours after birth) in term, healthy newborns. Outcome: Incidence of hypoglycaemia.
Six studies including 6768 neonates reported EBF at discharge. The pooled OR was 1.20, favouring delayed bath (95% CI = 1.08-1.34; moderate certainty evidence; Figure 4).
Figure 4. Forest plot for post-hoc analysis: Delayed first bath (>6 hours, ie, at or after 9, 12 or 24 hours after birth) vs early bath (≤6 hours after birth) in term, healthy newborns. Outcome: Incidence of exclusive breastfeeding rates at hospital discharge.
Three studies reported proportion of neonates initiated on breastfeeding post-implementation of a delayed bath policy in 2221 neonates. There was no difference in the odds of breastfeeding initiation based on timing of the bath (OR = 1.35; 95% CI = 0.86-2.13; Figure S3 in Online Supplementary Document).
The results of the current review suggest that delaying first newborn bath for at least 24 h after birth may reduce infant mortality and neonatal hypothermia compared to early bath within first 24 h. The effect of delaying bathing for at least 24 h on EBF at discharge is very uncertain. Delaying bathing for at least 6 h after birth probably improves EBF at discharge and may reduce the risk of hypothermia and hypothermia during hospital stay. The effect of delayed bath for at least 6 h on neonatal mortality is very uncertain.
No prior reviews have evaluated the effect of delaying the first newborn bath on neonatal outcomes. Our findings in the review support the existing WHO recommendations to delay newborn bathing for at least 24 h after birth and if not possible, for at least 6 h after birth . This recommendation, however, was based on expert consensus, but the evidence generated by this review is supportive of the same.
The scarcity of evidence on newborn bath practices was apparent in a recent survey on newborn skincare policies across United States maternity hospitals . 87% of the surveyed 109 US hospitals practised delaying the first newborn bath by at least 6 h, but the evidence for these policies cited by hospitals was unclear. This was also evident in our review, as there were no large randomized trials evaluating the effects of delaying first bath systematically.
The findings in the review are supported by plausible biological mechanisms. Mardini et al  showed that delayed bath beyond 24 h was associated with vernix caseosa retention on the skin and adequate time for skin-to-skin contact with mother. Vernix caseosa may act an important role in preventing evaporative water loss, thermoregulation, and innate immunity . This, along with better STS contact opportunities, can explain the lesser incidence of hypothermia and hypoglycaemia and better breastfeeding rates.
This review tried to answer an important research question on the effect of delayed first newborn bath on mortality and morbidities in term healthy neonates. Rigorous methodology was followed to conduct this review, with an all-inclusive literature search and no language filters. Though this review included 16 studies, there were no RCTs in the meta-analysis. One pilot RCT involved a small number of neonates and did not contribute to the meta-analysis . Moreover, there is no uniformity in definition of delayed and early baths in literature. For example, two studies were excluded due to comparison of baths over a narrow time period (1- and 4-hour baths) or too wide a time period (24- and 48-hour baths) [27,28]. Thus, there is a scarcity of good quality trials that have assessed the impact of delayed first bath on important health outcomes in term healthy newborns. There is also a paucity of literature on the process of bathing (type of bath, water temperature, environment temperature, bathing products, any risk of adverse events like slippage or drowning, need for counselling etc.), which remains to be addressed by research.
None of the included studies reported outcomes of any serious morbidities (sepsis or possible serious bacterial infection), timing of breastfeeding initiation, exclusive breastfeeding rates at 6 months or any adverse events related to intervention. There was a wide variation in the definition of delayed and early first bath in the included studies. This led to difficulty in comparison of interventions across the studies. However, we could meta-analyse ten studies together after realizing the timing in early bath group to be within the first 6 h of life in all these studies. The evidence for comparison of delayed bath >24 h vs early bath <24 h was based on single study results. The evidence was rated as low- and very low-quality, in part, affected by very serious risk of bias. A few late preterm neonates (two trials, 1646 newborns) without major comorbidities were among the studied population. We did not consider this to be serious indirectness because they were otherwise healthy newborns which did not downgrade the evidence. Certain cointerventions could have independently affected the outcomes in the studies. For example, immersion (tub) bath is known to decrease heat loss during bath, and hence, it could have affected the temperatures of the neonates independently.
Delaying the first newborn bath for at least 24 h after birth may reduce infant mortality and hypothermia. Additionally, delayed first bath for at least 6 h after birth may prevent hypothermia and hypoglycemia, and likely improves EBF rates at discharge in healthy newborns. The available evidence supports delaying the first newborn bath by 24 h and, if not possible, by at least six h to improve thermoregulation and breastfeeding rates in term healthy newborns. However, most of these conclusions are based on low certainty evidence and needs further evaluation in well-designed randomized trials.
Disclaimer: The authors alone are responsible for the views expressed in this publication and they do not necessarily represent the views, decisions, or policies of the World Health Organization.
Acknowledgment: We are grateful to Dr Rajiv Bahl, WHO, Geneva for technical guidance and support.