Early emollient bathing is associated with subsequent atopic dermatitis in an unselected birth cohort study

Abstract Background Skin barrier dysfunction is a key component of the pathogenesis of atopic dermatitis (AD). Recent research on barrier optimization to prevent AD has shown mixed results. The aim of this study was to assess the relationship between emollient bathing at 2 months and the trajectory of AD in the first 2 years of life in a large unselected observational birth cohort study. Methods The Babies After SCOPE: Evaluating the Longitudinal Impact Using Neurological and Nutritional Endpoints Birth Cohort study enrolled 2183 infants. Variables extracted from the database related to early skincare, skin barrier function, parental history of atopy, and AD outcomes. Statistical analysis was performed to adjust for potential confounding variables. Results One thousand five hundred five children had data on AD status available at 6, 12, and 24 months. Prevalence of AD was 18.6% at 6 months, 15.2% at 12 months, and 16.5% at 24 months. Adjusted for potential confounding variables, the odds of AD at any point were higher among infants who had emollient baths at 2 months (OR (95% CI): 2.41 (1.56 to 3.72), p < .001). Following multivariable analysis, the odds of AD were higher among infants who had both emollient baths and frequent emollient application at 2 months, compared with infants who had neither (OR (95% CI) at 6 months 1.74 (1.18–2.58), p = .038), (OR (95% CI) at 12 months 2.59 (1.69–3.94), p < .001), (OR (95% CI) at 24 months 1.87 (1.21–2.90), p = .009). Conclusion Early emollient bathing was associated with greater development of AD by 2 years of age in this population‐based birth cohort study.

factors such as microbial exposure. 12 Loss-of-function mutations in FLG (encoding filaggrin, which contributes to skin barrier integrity), 13 represent the greatest genetic risk factor for AD. 14 Emollients have been targeted to repair skin barrier defects in AD. Two small pilot trials showed that early emollient application to infants at high risk of developing AD reduced the development of AD significantly. 15,16 However, two randomized controlled trials (RCT) showed no benefit in early emollient use, 17,18 although a more recent RCT showed a significant reduction in the development of AD in high-risk infants if emollients were applied between birth and 2 months and then stopped. 19 A Cochrane review evaluating skin care interventions concluded that, based on low-moderate certainty evidence, skin care interventions such as emollients during the first year of life in healthy infants are probably not effective for preventing eczema and probably increase the risk of skin infection. 20 We aimed to assess the impact of early emollient bathing on AD development over the first 2 years of life in a large observational birth cohort study.

| Study subjects
This study was a secondary analysis of the Cork Babies After SCOPE: Evaluating the Longitudinal Impact Using Neurological and Nutritional Endpoints (BASELINE) Birth Cohort study (http://www. basel inest udy.net/). The purpose of BASELINE was to examine the effects of environmental exposures during pregnancy and infancy on childhood health and development. BASELINE recruited healthy first-born term babies and performed assessments at birth, 2, 6, 12, and 24 months involving parental questionnaires and physical assessment, including serial skin barrier assessment using transepidermal water loss (TEWL) measurements. Questionnaires at 2 months included questions about bathing and moisturizing habits, including frequency of bathing and moisturizing, and products used; and a specific screening question for AD ("Has you baby had an itchy rash on the face or in folds of arms or legs?"). Emollient baths were defined as baths with oil or emulsifierbased additives. Emollient application was defined as the application of a "leave on" emollient directly onto the skin. Parental atopy was defined as a self-reported parental history of AD, asthma, or allergic rhinitis. Experienced healthcare professionals diagnosed AD at 6, 12, and 24 months using validated diagnostic criteria. 21

| Data extraction
Data extracted from the database included: sex, birthweight, gestational age, and parental atopy; frequency of moisturizing, frequency of bathing, use of emollient bath additives, presence of parent-reported "itchy rash," and TEWL at 2 months; and AD status at 6, 12, and 24 months.

| Statistical analysis
Categorical variables were described using frequency (percentage) and continuous variables using mean (standard deviation (SD)). To investigate whether changes in AD prevalence were influenced by emollient bathing at 2 months, a mixed effects logistic regression model was used. Time points (6,12, and 24 months), emollient bathing at 2 months (yes, no) and their interaction (time*emollient use) were categorical fixed effects in the model. If the interaction term was not significant, the model was rerun with the interaction term removed. To adjust for potential confounding variables, the model (without the interaction term) was rerun with the potential confounding variables included as fixed effects.
Univariable and multivariable logistic regression models were used to investigate relationships between use of emollient baths and frequency of emollient application at 2 months and the presence of AD at 6, 12, and 24 months of age separately. To investigate whether the relationship between emollient bathing/application and AD differed by itchy skin status at 2 months, the interaction term emollient use*itchy skin was included in a multivariable model.
Infants were grouped according to the combination of emollient bathing and emollient application practices and univariable and multivariable logistic regression was performed. Models were run with those with "itchy rash" at 2 months included and then excluded. Emollient bathing was divided into those who had emollient baths at 2 months versus those who did not. Emollient application was divided into those who had emollient applied more than once weekly ("frequent") versus those who had emollient applied once a week or less ("infrequent"). Variables were combined to create four categories: 1. Emollient bath, frequent emollient application.
2. Emollient bath, infrequent emollient application. The potential confounders included in all multivariable models were sex, birth weight, bathing frequency at 2 months, TEWL at 2 months, "itchy rash" at 2 months, and presence of maternal/paternal atopy (AD, asthma, rhinitis). For all independent variables, unadjusted and adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs) are presented. Prior to performing the multivariable analyses, multicollinearity among the independent variables was tested using the variance inflation factor. All tests were two-sided, and a p-value <.05 was considered statistically significant. Statistical analysis was performed using Stata (version 15.1, StataCorp LP).

| Effect of emollient bathing at 2 months on AD at 24 months
Values in bold represent statistically significant results (p < .05).
In the univariable analysis (

| Outcomes of combination emollient bathing and application at 2 months
Subjects were further grouped according to the combination of emollient bathing and application practices. At all time points, the prevalence of AD was highest in the group who had emollient baths and frequent emollient application, with similar trends seen if infants with "itchy rash" at 2 months were excluded (Table 2, Figure 2A,B).
Univariable and multivariable logistic regression analysis showed similar findings, with significantly higher odds ratios in infants who had emollient baths and frequent emollient application, compared with infants who had no emollient baths and infrequent emollient application ( Table 3). The findings were similar when infants with "itchy rash" at 2 months were excluded.

| Impact of emollient practices at 2 months on TEWL
A linear mixed model was used to investigate whether changes in TEWL between 2 days and 2 months differed by emollient bathing/ application. The fixed effects included in the model were group, time, and the interaction of group by time.
At 2 days, the median (IQR) TEWL was 7.00 g/m 2 /h (5.00-9.00) in the group that did not go on to have emollient baths (n = 1097) and 7.23 g/m 2 /h (5.00-9.25) in the group that did go on to have emollient baths (n = 422). At 2 months, the median (IQR) was 9.32 g/m 2 /h (7.00-12.01) in the group that did not have emollient baths (n = 1215) and 9.66 g/m 2 /h (7. 15-13.26) in the group that did use emollients (n = 457). In the linear mixed model, the interaction between group and time was not statistically significant (p = .797), indicating that the changes in TEWL over time did not differ by emollient bathing practice.
The groups were further divided according to frequency of emollient application (once weekly or less vs. greater than once weekly) and emollient bathing practice (yes vs. no; Table 4). TEWL was significantly higher in the group that had emollient baths and had emollient application greater than once weekly compared to the other three groups (p = .003 for group in linear mixed model excluding the interaction group*time).

TA B L E 2
Prevalence of atopic dermatitis (AD) at 6, 12, and 24 months with combinations of use of emollient baths (yes/no) and frequency of emollient application (frequent defined as more than once weekly, infrequent defined as once weekly or less) at 2 months, including and excluding infants with "itchy rash" at 2 months.  Note: Group 1 = emollient bath, frequent emollient application; Group 2 = emollient bath, infrequent emollient application; Group 3 = no emollient bath, frequent emollient application; and Group 4 = no emollient bath, infrequent emollient application.

| Effect of parental history of atopy on outcomes
Atopic dermatitis outcomes were also examined in relation to the use of emollient baths in those with a parental history of atopy (AD/asthma/rhinitis) versus those without a parental history of atopy. At 6 months, the relationship between emollient bathing and increased risk of AD did not differ by history of parental atopy

| DISCUSS ION
This secondary analysis of a large unselected first-born birth cohort showed that emollient bathing at 2 months was associated with the development of AD by 2 years of age. Results were consistent even after accounting for confounding factors such as TEWL at 2 months, parent-reported "itchy rash" at 2 months, and parental atopy. The relationship was similar at 6, 12, and 24 months, with similar trajectories for infants with an early "itchy rash" versus those without. Over a quarter of infants had emollient baths at 2 months, without being Confounding factors were controlled for as far as possible. Delivery by cesarean section was not shown to be associated with AD in this cohort. 25 Limitations include the lack of randomization, so early emollient use may have been a parental response to signals of impending AD. However, the statistical model controlled for the presence of an "itchy rash" at 2 months and upper quartile TEWL at 2 months, a functional marker of skin barrier integrity. In addition, subgroup exclusion of infants reported as having an "itchy rash" at 2 months may have helped to exclude infants who had emollient baths due to parental concerns about dry skin and impending AD. Early application of emollients may have masked subclinical AD at 2 months, which became apparent at subsequent time points due to progression in AD activity. The rate of dropout from the study may have differed between the groups who used emollients versus those who did not and between the groups who developed AD and those who did not. The duration of use of emollient baths or emollient application was also not captured, and effects may have been greater with longer use.
The type of emollient/moisturizer was not specified, and products may vary in their effects. The most frequently used emollient on young infants at the time of this cohort study was vegetable oil (30%), 26 and olive oil is known to impede the lamellar structure of the epidermal barrier. 27 The composition of the emollient/moisturizer may also be important as the local water quality in Cork is classified as hard to very hard. 28 Surfactants such as sodium laureth sulfate deposit more easily on the skin when used with hard water, with an associated increase in TEWL and potential for irritation in individuals with normal skin, and more severe symptoms in patients with AD associated with filaggrin mutations. 29 Thus, the findings may not be as relevant for regions were bathing water is less hard.
Parental history of AD, asthma, and rhinitis was self-reported and therefore may not be completely reliable. This was a monocentric study with a relatively homogenous population.
The findings support recent RCT that show that emollients as bath additives or "leave on" products did not prevent the development of AD. 17,18 The PreventADALL trial showed an increased risk of AD by 12 months in infants randomized to regular emollient bathing between 2 weeks and 8 months, with an odds ratio of 1.57 (95% CI: 1.10, 2.23). 18,20 Emollient bathing may interfere with natural lipids in the stratum corneum that reduce TEWL and protect against TA B L E 3 Odds ratios (OR) and 95% confidence intervals (95% CI) for atopic dermatitis (AD) at 6, 12, and 24 months for infants who had emollient baths and frequent emollient application, relative to infants who had no emollient baths and infrequent emollient application (frequent defined as more than once weekly, infrequent defined as once weekly or less) at 2 months, including and excluding infants with "itchy rash" at 2 months.

ACK N OWLED G M ENTS
We would like to thank the children and parents and our colleagues in the BASELINE study and colleagues in the INFANT research center.

CO N FLI C T O F I NTE R E S T S TATE M E NT
JO'BH receives research funding related to this field from the City

PE E R R E V I E W
The peer review history for this article is available at https://