Atopic outcomes at 2 years in the CORAL cohort, born in COVID‐19 lockdown

The CORAL study is a cohort of infants born during the first weeks of the first SARS‐CoV‐2 (COVID‐19) lockdown. This cohort has had lower antibiotic exposure, higher breastfeeding rates and lower infection rates, especially in the first year of life. We hypothesized that the altered early‐life environment of infants born during lockdown would change the incidence of allergic conditions.

Results: AD was more common in CORAL group at both 12 (26.5% vs. 15.5%;p < .001)and 24 months (21.3% vs. 15.9%;p = .02)compared with pre-pandemic BASELINE cohort.Within the CORAL group, those with AD at both 12-and 24-month appointments had a more severe AD phenotype associated with a higher risk of allergic sensitization.There was less milk (0% vs. 1%; p = .09),peanut (0.6% vs. 1.8%; p = .3),and egg allergy (0% vs. 2.9%; p < .001) in the CORAL group at 24 months compared with the BASELINE cohort.Aeroallergen sensitization increased between 12 and 24 months in the CORAL cohort (1.5% vs. 8.9%; p < .001),as did parent-reported wheezing episodes (9% vs. 24%; p < .001).Conclusions: Despite higher AD incidence in the CORAL cohort, the incidence of food sensitization and allergy are lower than expected pre-pandemic rates possibly reflecting the early introduction and maintenance of dietary allergens enhanced by changes in infant infections, antibiotic use, and breastfeeding in the first 2 years of life in the group.These beneficial effects of the lockdown could be outweighing the expected risk of less early-life microbial encounters outlined by the hygiene hypothesis.

| INTRODUC TI ON
The Republic of Ireland had one of the longest and strictest lockdowns in Europe with 38 weeks (75%) of the first year following the declaration of the pandemic in the highest state of lockdown.
The SARS-CoV-2 (COVID-19) health-related lockdown measures adopted in Ireland, and globally, included physical distancing, the wearing of masks and changes in hand hygiene and home hygiene practices.We have reported differences in the social and communication skills of infants born during the pandemic when compared with pre-pandemic groups 1 and we and others noted significant changes in pediatric healthcare access in the first months on the pandemic. 2,3Health outcome data at 6 months 4 and health and atopic outcome data at 12 months 5 have been published previously.
Strachan first proposed the hygiene hypothesis as an explanation for the allergy epidemic more than 30 years ago. 6Mounting evidence supports the theory that lack of microbial biodiversity, associated with modern living, is implicated in the rapidly rising rates of allergy worldwide. 7The COVID-19 lockdown presented an exceptional opportunity to examine the hygiene hypothesis and its role in the development of atopy in a real-world context, with exaggerated and prolonged social isolation.

| Study sample
Infants eligible for inclusion were term singletons born between March and May 2020 in two Dublin city maternity hospitals who are resident in Ireland.Those excluded were infants with major congenital abnormalities, a history of neonatal antibiotic exposure or women with a confirmed household contact with COVID-19 during the pregnancy.Of the 3065 eligible term infants invited to participate in the study 365 infants were recruited, and 320 infants remained recruited at the 2-year study end point (Figure S1).

| Comparison cohort
Allergic outcomes in the CORAL cohort were compared with a national pre-pandemic birth cohort from the BASELINE study. 8The BASELINE study recruited babies born in Ireland between 2008 and 2011.Further details about recruitment, protocol, and prior results of the BASELINE study have been published and are available online (www.baselinest udy.net).The questionnaires used in the CORAL study were derived from those used in the BASELINE cohort and the definitions for AD, food sensitization and allergy were consistent between the two groups.A table comparing the main demographic details between the CORAL and BASELINE cohorts is included in the supplemental material (Table 1).

| Study measures
Infants attended appointments at 6-, 12-, and 24-months at which they had a physical examination, including atopic dermatitis assessment, SARS-CoV-2 antibody testing and at 12 and 24 months had skin prick testing (SPT) and stool and blood samples were taken.Detailed epidemiological information was collected at enrolment, and at each subsequent study appointment, using parent completed questionnaires.
In infants with visible atopic dermatitis at their research appointment, a ScorAD (Scoring Atopic Dermatitis) was calculated at 12 and 24 months and a total and objective ScorAD recorded. 9T procedure was conducted in line with local and international allergic outcomes, COVID-19, hygiene hypothesis

Key message
Infants born during the first months of the COVID-19 lockdown had a higher incidence of atopic dermatitis at 12 and 24 months when compared to a pre-pandemic Irish cohort.
Food allergy was lower in the pandemic group by 2 yearspossibly reflecting the impact of early introduction advice in the CORAL research group.Additionally, the lower antibiotic use and higher breastfeeding rate in the group may have had beneficial effects in promoting tolerance mediated by the gut microbiome.
recommendations and was performed to additional foods only as clinically indicated by history. 10A panel of three foods (cow's milk, hen's egg, and peanut) and the three locally dominant aeroallergens (house dust mite (dermatophagoides pteronyssinus), cat and grass (timothy) pollen were offered to all infants.Study participants were encouraged to contact the CORAL team at any time about any suspected hypersensitivity reaction.Oral Food Challenges (OFC) were conducted on all infants with positive SPT (>3 mm) who were not regularly tolerating a food at their 24-month appointment.OFCs were conducted using the PRACTALL guidelines.The egg OFC protocol used raw pasteurized egg. 11In line with national practice, infants with newly diagnosed cow's milk allergy or egg allergy were started on graded reintroduction, using ladder-based programs. 12,13l parents received national weaning advice at face-to-face appointments at 6 months regarding the early introduction of peanut and this was reviewed and repeated, if necessary, at 12 months.Data regarding COVID-19 contacts and infections were collected through parent-reported questionnaires and through lateral flow SARS-CoV-2 antibody testing at research appointments.

| Definitions
Atopic dermatitis (AD) was defined as infants who fulfilled the UK Working Party Diagnostic criteria 14 (Itch plus 3 minor features of: visible dermatitis, generalized dry skin, family history of atopy, flexural involvement, and onset under 2 years) using data collected from parent questionnaires and/or those with doctor diagnosed AD (general practitioner or at study appointment).
Sensitization was defined as a positive wheal >3 mm. 11Food allergy at 12 months was defined as sensitization plus a history of immediate reaction or positive OFC.Food allergy at 24 months was defined by positive OFC.

| Statistical analysis
Chi-square and Fisher's exact test were used to compare AD, food sensitization, and food allergy between the CORAL and BASELINE groups, and between the CORAL group at 12 and 24 months.Comparison data on aeroallergen sensitization, allergic airways disease and asthma was not available from the BASELINE cohort.Within CORAL, AD phenotypic groups were further analyzed for differences using one-way analysis of variance (ANOVA) for continuous variables, such as Sco-rAD, and Chi-square for nominal variables.To examine the impact of individual variables on allergic risk simple logistic regression (reported as odds ratio) was performed followed by multiple logistic regression (reported as adjusted odds ratio [aOR]) where possible.The "rule of thumb" of a minimum of 10 events per one independent variable was adopted to avoid overfitting of data.Therefore, multiple logistic regression was possible for AD and allergic sensitization but not for food allergy due to the small numbers of participants with this outcome.

Detailed breakdown of simple and multiple logistic regression analysis
for each atopic outcome are available in the supplemental data.As this was an exploratory observational study, significance was set at p < .05.

| Health outcomes in the CORAL cohort between 12 and 24 months
Between 12 and 24 months, there were significant changes in parentreported infections, systemic antibiotic use, and day care outside of the home in the CORAL cohort.Key metrics are summarized in Table 1.S3 and S4).

| Food allergy
The total point prevalence of OFC confirmed food allergy in the CORAL group reduced slightly between 12 and 24 months (4.7% vs. 1.7%; Fishers exact test: p = .07).Food allergic patient demographic and environmental features are outlined in Table S5.Given the small number of children in the cohort with food allergy at 24 months, multiple logistic regression was not possible.

Peanut
One child had persistent peanut allergy at both 12 and 24 months.
One child had peanut allergy which resolved between 12 and 24 months and one child who previously tolerated peanut developed new peanut allergy between 12 and 24 months.Peanut allergy was not parentally suspected in any individual that did not attend for SPT (declined or remote participant).

Egg
There was one new egg sensitization case at 24 months.7/8 (87.5%) of children with a history of egg allergy before 12 months were tolerating some form of baked egg but not eating it in all forms and one parent was not giving egg in any form.OFC was conducted in 7/9 (78%) with egg sensitization at 24 months and all challenges were negative.

Tree nuts
Seven tree nut OFCs were conducted in six children with four positive and three negative challenges bringing the point prevalence of tree nut allergy at 24 months in the cohort to 3/292 (1%).

| Allergic sensitization
SPT positivity in CORAL is outlined in Table 3.

| DISCUSS ION
The CORAL study is to our knowledge unique, being the only longitudinal prospective and allergy-focused study of children born in the most severe global lockdown, with rigorous follow up as societal restrictions were lifted.We hypothesized that being born during the COVID-19 lockdown would alter the infant environment and may impact the incidence of allergic disease in the cohort.
At 12 months, we reported very low antibiotic use with lower rates of parent-reported infections and higher breastfeeding rates. 5As society re-opened between this cohort's ages of 12 and 24 months, parent-reported infections and systemic antibiotic use Compared with the pre-pandemic BASELINE cohort, there are higher rates of AD at 12 and 24 months.One explanation for this is a gradual increase in the incidence of AD in the Irish population between 2011 and 2022.Although there is substantial, variation between countries the childhood European incidence of AD is approximately 20%. 15As both CORAL and BASELINE used the same definition for AD and infants were assessed at the same time points this does not explain any differences in incidence.Differences in the early-life environment of infants born during the COVID-19 lockdown may also have influenced the incidence of AD.We identified three distinct patterns of AD in the cohort with differences in severity and allergic sensitization in the persistence group.AD severity and atopic sensitization have previously been identified as important prognostic indicators of AD persistence into later childhood. 16,17spite higher levels of AD at 12 and 24 months, food sensitization and allergy in the CORAL group are lower than in the prepandemic BASELINE. 8Compared to BASELINE the prevalence of peanut sensitization was significantly lower at 24 months.Peanut allergy was also lower in the CORAL cohort, but this did not reach significance, possibly due to the very small number of children with this outcome at 24 months.
We have previously reported how the lower peanut sensitization and allergy prevalence found in CORAL may be secondary to secular changes in allergy practice between the BASELINE and CORAL cohorts with CORAL researchers encouraging both the introduction and maintenance of common allergens into infants' diets. 18g sensitization at 24 months was similar between the CORAL and BASELINE cohorts.However, egg allergy was statistically significantly lower.All children with milk and egg allergy at 12 months who remained in the study at 24 months had full resolution of their allergy.This likely reflects changes in allergy practice using baked egg and milk ladders since the BASELINE study.
In addition, to these changes in allergy practice, it is possible that the environmental differences in this locked down cohort may have allowed rapid progression along allergy ladders and early tolerance.
CORAL parents had increased time at home for food preparation and parental supervision during ladder up-dosing.CORAL infants had very low early-life antibiotic consumption and more prolonged breastfeeding which may have promoted a more tolerogenic host gut environment.
Tree nut allergy was more common than peanut allergy at 24 months although comparison case numbers are small (1% vs. 0.7% [95% CI: −0.0169, 0.0229] p = .65).This may reflect changes in allergy practice and dietary diversity in Irish infants' diets which have focused on peanut and may have not emphasized the introduction and maintenance of other allergens in children's diets.
It is known that early food introduction is allergen specific and early introduction of peanut does not reduce the risk of tree nut allergy. 19 expected, there was an association between all allergic conditions in the CORAL group.We identified more allergic sensitization and food allergy in children with AD.Allergic sensitization at 12 and 24 months was associated with AD at both time points and with asthma diagnosis at 24 months.We examined differences in the infant environment between the atopic and non-atopic cohorts.At 12 months, no single environmental factor reached significance for influencing the risk of AD; however, siblings and rural household showed a protective signal.At 24 months, systemic antibiotic use and childcare outside the home increased the odds of AD.Children minded at home received less antibiotics, avoiding the resultant negative impacts on the gut microbiota.Previous studies have also identified a link between early antibiotic use and increased risk of AD. [20][21][22] The infant environment appeared to exert a greater influence on the risk of allergic sensitization than on risk of AD at 12 months.At 12 months, presence of siblings was protective against allergic sensitization.This echoes the original hygiene hypothesis 6 and has been identified in other notable publications which have shown a dose-related protective effect between siblings and reduced risk of allergic sensitization. 23 24 months, despite there being more AD in children attending day care outside the home there was less allergic sensitization in this group.At 24 months, aeroallergen sensitization was more prominent than food sensitization and infants were more likely to be sensitized to an aeroallergen if exclusively cared for at home.
An explanation for this may be longer exposure to higher dust mite allergen levels in the home setting than in center-based childcare, such as nursery.Both community-acquired infections and antibiotic use increased substantially between 12 and 24 months.Many viral infections, including rhinovirus-C and respiratory syncytial virus, are associated with an increased risk of allergic airways disease. 24,25Antibiotic use in early life has also been associated with an increased risk of asthma. 20,26This is the first study examining the impact of SARS-2-CoV infection in early life on the risk of aeroallergen sensitization and asthma, with no additional adverse effect identified.The long-term impact of SARS-2-CoV infection on persistent wheeze and asthma risk has yet to be determined and this will be studied in due course in the CORAL group.

| Strengths and limitations
Strengths of the CORAL study include the novelty of the cohort during a time of limited environmental exposures and its prospec- on the generalizability of the data.The small sample size has meant that we may not have identified associations which would be apparent if the cohort was larger.Individual participant demographics for the BASELINE AD and food allergy groups were not available which meant it was not possible to adjust for differences between these specific groups.However, the BASELINE and CORAL cohorts were well matched overall.

| CON CLUS ION
The CORAL cohort has undergone major changes as children have entered society and their environmental and dietary exposures have diversified.Despite higher AD incidence in the CORAL cohort, the incidences of food sensitization and food allergy are lower than the comparative pre-pandemic BASELINE rates.These CORAL study is a longitudinal observational study examining the effect of the Coronavirus pandemic on allergic and autoimmune outcomes in infants born during the first, maximal COVID-19 lockdown in Ireland.The overall aims of the CORAL study included comparing the incidence of allergic disease in children born during the lockdown with pre-pandemic published Irish data from the BASELINE cohort, 8 determining the impact of lockdown measures on infant exposure and infection with SARS-CoV-2 virus and finally studying how the lockdown impacted stool microbial diversity.This paper will present comparisons between atopic outcomes in the CORAL and BASELINE cohorts, compare allergic outcomes within the CORAL cohort at 12 and 24 months and demonstrate participant factors associated with allergic outcomes in CORAL.
Ethical approval was granted by the National Research Ethics Committee for COVID-19-related Research (20-NREC-COV-067).At enrolment, families gave informed consent for the above study measures and for data to be stored on an online data collection and storage platform -Castor EDC.Anonymized data were exported from Castor to Stata/SE Version 17.0 (StataCorp) for analysis.
increased.COVID-19 infant infections increased 10-fold between 12 and 24 months, showing that lockdown measures are effective in protecting infants in the first year of life.The longitudinal follow-up of the cohort to 2 years has allowed us to further characterize atopic disease within the cohort by identifying patterns of AD onset and resolution, track the rapid resolution of infants with egg and milk allergies and diagnose aeroallergen sensitization and reactive airways disease that present after infancy.
tive follow-up, using validated assessment tools and defined clinical outcomes, such as food challenges and ScorAD assessments of AD.Limitations of the study are the small size of the cohort and the generalizability of the data.The cohort represented just 12% (364/3065) of the eligible children born in the two maternity hospitals during the recruitment period increasing the risk of selection bias.The maternal education attainment level, low parental smoking rates, and uniquely high breastfeeding rates for Ireland suggest that this parent cohort may not be representative of the general Irish population again suggesting the potential limitation of selection bias changes may reflect the effect of changing allergy practices regarding the early introduction of allergens and the importance of maintenance of allergens in the diet.Lockdown-related changes in infant infections, antibiotic use, and breastfeeding in the first 2 years of life may have enhanced the effects of allergy interventions and mediated the risk of AD through their impact on the microbiome.These beneficial effects of the lockdown could be outweighing the expected risk of less early-life microbial encounters outlined by the hygiene hypothesis.This study has added to the literature on the impact of early-life infant environment on the Atopic outcomes in the CORAL group and BASELINE groups at 12 and 24 months (Fisher's exact test used when comparing groups with <9 -no test statistic, p-value only).
Tree nut SPT only performed if clinical history of reaction; Fisher's exact test used when comparing groups with <9 -no test statistic, p-value only.