Discussion
In this RCT in high-risk infants, we found that daily emollient use initiated in the first week of life until 2 months is associated with a significant reduction in the cumulative incidence of AD at 12 months. Daily emollient use was associated with a 50% and 29% reduction in the risk of the cumulative incidence of AD at 6 and 12 months, respectively. Similar risk reductions were observed in the per-protocol analyses where only those in the intervention and control groups were included if they used emollients at least once daily and <4 days a week, respectively. However, these were not significant for the primary outcome which may be due to the conservative adherence criteria applied and thus, lower numbers included in the analysis and therefore lower power to detect differences between the groups.
While some AD cases diagnosed before 6 months had resolved by 12 months, there was no difference in transient cases between the groups. As we did not collect longer term data, we cannot exclude the possibility that the intervention may have only delayed the onset of AD beyond 12 months. A recent meta-analysis reported a protective effect of emollients but only when there was no interval between the emollient treatment and AD assessment 14. However, there was significant heterogeneity between the four studies included in that analysis. In our study, a 29% reduction in the risk of cumulative AD at 12 months, was maintained 10 months after the intervention.
Our findings are at variance with recent findings from two large RCTs, where no evidence of a protective effect of emollient use in the first year against AD was found 10,11. Among the most notable differences between these RCTs and ours was the timing of the intervention. The treatment in STOP AD began within days of birth during a dynamic period of skin maturation and adaption to the dramatic environmental changes of life ex utero. In STOP AD, infants were randomised within 4 days of birth with the IG advised to begin the emollient treatment immediately. In BEEP, the median (IQR) age that emollient use began was 11 days (7, 17) days, with only 89% starting emollient application before 3 weeks. In PreventADALL, the intervention began from 2 weeks of age.
The emollients used in BEEP and PreventADALL were basic petroleum and paraffin-based formulations, respectively. The emollient used in this study consists of a formulation with added ceramides developed specifically for very dry itchy skin. Two small studies that also used more complex ceramide-rich emollients reported non-significant trends towards a protective effect against AD 15,24. Following one of these 15, a larger scale RCT, the PEBBLES study, involving twice-daily application of the same ceramide-based emollient from 0-6 months is ongoing25. Here we showed a reduced risk of AD at 12 months with a short 2-month intervention period, which may represent a more feasible and family friendly strategy for AD prevention.
Our high adherence rates demonstrate the feasibility of implementing a regimen of daily emollient use during the first 2 months of life. Adherence rates using the diaries were lower than reported on the questionnaires but 82.4% still reported using emollients on ≥75% of days equating to over 5 days a week. While infants in this study were followed closely during the intervention period, similar rates of adherence were observed in BEEP which involved limited contact, but used a less strict definition for adherence (emollient use ≥3 days/week)10. Only 27% of the IG fully adhered to the protocol in PreventADALL which may have influenced the absence of a protective effect 11.
While we did assess food allergy outcomes, this study was not powered to detect a reduction in food allergy risk. Unlike BEEP, where a non-significant increase in food allergy in the IG has been prominently reported (15), we found no difference in the prevalence of food allergy between the groups. While we did not use SPTs to screen for food allergy, almost all infants had tried the most common food allergens - milk, egg and peanut - by 12 months, so the rate of food sensitization and allergy reported is likely reflective of the true rate in our groups. BEEP reported a higher rate of skin infections in the IG, with suggestions of the possibility of greater pathogen exposure with emollient application 10. We did not find evidence of an increased risk of skin infections with short-term emollient use.
Despite the reduction of AD risk in the IG, there was no difference in TEWL throughout the first year between the groups. Other studies on emollient use during infancy reported a similar absence of an effect of the intervention on TEWL 15,24. TEWL measurements are influenced by environmental factors and more crucially for infants, subject-specific parameters including stress and crying26. This may have affected our ability to detect differences between the groups.
The major strength of this study is the initiation of emollient use within days of birth in the IG. Other strengths include the close follow-up of infants, a high rate of adherence in the IG and a low rate of contamination in the CG.
A limitation to this study is that in response to the COVID-19 pandemic, many AD diagnoses were made remotely. To mitigate this, detailed information and photographs were collected when making a diagnosis. SCORAD assessments were also completed remotely, which may have affected assessments of AD severity. Validated diagnostic criteria could not be applied when diagnosing earlier onset AD (<6 months), where cases were diagnosed based on presence of AD lesions. However, of the 73 infants diagnosed with AD ≤6 months, 71 (97.3%) met the UKWPDC at 6 months. The prevalence of cumulative AD in this group was higher than expected based on the rates among infants with parental history of atopy in an Irish birth cohort 23. A possible explanation for this is the a priori recruitment of high-risk infants and the close monitoring of skin health in this study. Only a third (32.1%) of those eligible for this study were recruited. One of the main reasons for the refusal to participate was the demanding follow-up schedule involved, particularly during the intervention period that started before going home with their newborn baby, suggesting that more motivated individuals were recruited. We also had a higher rate of withdrawals in the first two weeks of life, particularly in the intervention group, mainly due to withdrawal of consent and not due to early onset of AD by this time. This is a consideration in assessing the feasibility of advising daily emollient use in the early postnatal period to a more general population.
We have demonstrated that early initiation of daily specialized emollient use until 2 months reduces the incidence of AD in the first year of life in high-risk infants. The mechanisms behind this are unclear but analysis of microbiome diversity and inflammatory biomarkers in a subgroup of this study is ongoing and may provide further information. While several recent studies do not support a protective effect of emollient use in infancy, future studies should examine the use of more complex emollients directed at enhancing the skin barrier, while identifying a treatment window that is both effective and acceptable to parents.