Atopic outcomes at 5 years in the CORAL cohort born during the 2020 COVID‐19 lockdown

Abstract

Background

The 2020 SARS‐CoV‐2 lockdown in Ireland was among the most stringent in Europe. The CORAL (Impact of Coronavirus Pandemic on Allergic and Autoimmune Dysregulation in Infants Born during Lockdown) study was developed to investigate early‐life allergic outcomes in an altered environment. At 2 years we showed lower food allergy and higher atopic dermatitis rates. The current report reassesses this same cohort at 5 years to evaluate whether early pandemic‐related trends were transient or represented enduring alterations in childhood allergy.

Methods

This longitudinal observational study assessed outcomes at 5 years. Clinical assessments included evaluation of atopic dermatitis, skin prick testing for food and aeroallergens, and collection of stool and blood samples. Parent‐completed questionnaires collected data on environmental and household exposures, allergic and respiratory outcomes, and healthcare use.

Results

Aeroallergen sensitisation increased from 8.9% at 2 years to 24% at 5 years (p < .001). Sensitisation strongly predicted allergic rhinitis symptoms (OR 5.68, p < .001). Parent‐reported wheeze was unchanged between 2 years (23.8%) and 5 years (22.1%) and strongly associated with asthma diagnosis (χ ² = 33.0, p < .001). Wheeze occurred predominantly in children with a recent infection. Most atopic dermatitis had resolved and food allergy remained uncommon at 5 years of age.

Conclusion

This study provides the first longitudinal evaluation of pandemic‐associated early‐life exposures and atopic outcomes in children. Our findings suggest that pandemic‐associated early‐life environmental changes may have selectively modified food allergy risk, while the overall pattern of allergic disease remained broadly consistent with the accepted evolution of allergic conditions.

Trajectories of atopic conditions between 1y and 5y of age followed known patterns in infants born during COVID‐19 lockdown in 2020. Rates of other health outcomes except actual COVID‐19 infection all increased predictably.

Key message.

This study provides the first 5‐year longitudinal follow‐up of infants born during the COVID‐19 pandemic. In infancy, altered microbial and social exposures coincided with unexpectedly low rates of food allergy, potentially supported by increased breastfeeding, reduced antibiotic use and structured early allergen introduction, which positively impacted gut microbiota composition and immunoregulatory metabolic pathways. These benefits persisted to 5 years, with food allergy and eczema remaining uncommon, whereas aeroallergen sensitisation increased. Early pandemic‐associated exposures appear to have shifted early atopic manifestations without preventing the later emergence of allergic and respiratory disease.

1. INTRODUCTION

Early‐life environmental and microbial exposure plays a crucial role in shaping immune tolerance and allergy risk with ecological decreases in infection and smaller family size being the basis of the ‘hygiene hypothesis’ to explain increases in allergic conditions in the last 50 years. ¹ , ² The SARS‐CoV‐2 (COVID‐19) pandemic (2020–2023) profoundly altered the environmental and social exposures experienced by infants born during this period. Public health restrictions, reduced childcare attendance, changes in healthcare access and increased parental presence at home reshaped the early‐life environment during the critical window for immune and microbial development. The pandemic therefore created a unique real‐world climate to evaluate the hygiene hypothesis in the context of unusual, extended early‐life social isolation.

We established the CORAL (Impact of Coronavirus Pandemic on Allergic and Autoimmune Dysregulation in Infants Born during Lockdown) study to investigate early‐life allergic outcomes during the COVID‐19 pandemic. Earlier CORAL findings showed age‐related increases in mild to moderate atopic dermatitis prevalence within the cohort at 12 and 24 months, with overall prevalence comparable to expected pre‐pandemic levels, but lower rates of food allergy compared with pre‐pandemic infants. ³ , ⁴ We hypothesised that the lower prevalence of food allergy reflected pandemic‐related changes in early‐life exposures, including higher breastfeeding rates facilitated by lockdown‐related changes in maternal work patterns, reduced antibiotic exposure associated with decreased infection transmission and social mixing, and early dietary allergen introduction advice provided by the CORAL research group. While these early findings highlighted short‐term shifts in atopic outcomes, the present study offers the first longitudinal assessment of exclusively pandemic‐associated early‐life exposures in children. Our present study reassesses participants from ages 2 to 5 years using the same core outcome measures and definitions. By extending follow‐up to 5 years, the study clarifies whether the early pandemic‐associated patterns represent transient findings or sustained shifts in childhood allergy.

2. METHODS

2.1. Study design and population

The CORAL study is a longitudinal observational cohort examining the effect of the coronavirus pandemic on atopic outcomes and gut microbial diversity. Our initial cohort of 365 infants was recruited during the first, most restrictive COVID‐19 lockdown in Ireland (March–May 2020) and followed prospectively for 2 years. Outcomes were compared with a pre‐pandemic Irish cohort BASELINE. ³ , ⁴ , ⁵ For the present follow‐up study, all participants who completed the 2‐year assessment were invited for reassessment at 5 years of age.

2.2. Study sample

The primary study participants eligible for inclusion were term singletons resident in Ireland and born between March and May 2020 in two Dublin maternity hospitals. Exclusion criteria were major congenital abnormalities, neonatal antibiotic exposure, or a confirmed household contact with COVID‐19 during pregnancy. Of the 3065 eligible infants invited to participate, 365 were recruited into the original CORAL at birth, and 320 remained at the 2‐year study endpoint. ³ For the current follow‐up, 209 participants re‐consented to continue in the study; of these, 180 attended the in‐person visit and completed the questionnaire, 28 completed the questionnaire only and 1 attended the appointment only. Participant recruitment, retention and completion at 5 years are illustrated in Figure 1.

FIGURE 1.

CORAL Study recruitment, retention and completion at 5y.

2.3. Study measures

At 5 years, allergic and respiratory outcomes, medication use, healthcare interactions and key environmental exposures were collected via a parent‐completed questionnaire. Consenting participants attended face‐to‐face at 5 years of age (mean age 5.25 years), during which they underwent physical examination including assessment of atopic dermatitis, skin prick testing for food and aeroallergens, and collection of stool and blood samples. Families unable to attend in‐person visits were invited to complete equivalent online questionnaires to maximise standardisation of data capture.

2.4. Definitions

Food and aeroallergen sensitisation was defined as a positive wheal ≥ 3 mm. ⁶ Food allergy was defined as a clinical history of immediate reaction and positive SPT/IgE. Oral food challenge was considered if clinically indicated. Atopic dermatitis was defined according to the UK Working Party Diagnostic Criteria. ⁷ Atopy was defined as a positive skin prick test to any aeroallergen or a confirmed food allergy at 5 years.

SCORAD was employed to objectively assess atopic dermatitis severity. ⁸ Study children underwent skin prick test in line with local and international recommendations. ⁹ Three core foods (egg, milk and peanut) and five locally dominant aeroallergens (house dust mite [Dermatophagoides pteronyssinus], dog, cat, tree mix and timothy grass) were offered to all children. Additional food SPTs were only performed if clinically indicated by history. In cases where a child had a positive SPT in combination with a compatible clinical history, food‐specific IgE testing was undertaken. Reporting of any suspected allergic reactions to the study team was encouraged.

2.5. Ethics, consent, data handling

Ethical approval was granted by the Children's Health Ireland Research Ethics Committee (CHI REC‐513‐24). Parents gave informed consent for study measures and storage of online data. De‐identified datasets were transferred into Stata/SE 17.0 (StataCorp) for statistical analysis.

2.6. Statistical analysis

Statistical analyses were performed using Stata BE 19 (StataCorp LLC, College Station, TX, USA). Descriptive statistics are presented as frequencies and percentages for categorical variables and medians with interquartile ranges for continuous variables. Analyses were conducted on available data only (complete‐case analysis); missing data were not included, and denominators therefore vary between outcomes.

Categorical variables were compared using Pearson's chi‐square or Fisher's exact tests, as appropriate. Unadjusted associations between binary variables were summarised using crude odds ratios with 95% confidence intervals. Prevalence comparisons between 2 and 5 years were analysed as unpaired due to differing denominators. Food‐allergy analyses were descriptive and used Fisher's exact tests; no multivariable models were performed due to small sample size.

Diagnostic performance of aeroallergen SPT in predicting parent‐reported symptoms was evaluated using standard measures: sensitivity, specificity, positive and negative predictive values, and likelihood ratios, each with exact binomial 95% confidence intervals. Univariable logistic regression was used to explore associations between environmental exposures and the following outcomes: aeroallergen sensitisation, wheeze and physician‐diagnosed asthma at 5 years. Multivariable logistic regression models were then fitted for the same outcomes. Variables were selected a priori based on biological plausibility and previous literature. Collinearity was assessed using variance inflation factors, with none exceeding two. A separate logistic regression model assessed the association between sibling number and atopy (defined as aeroallergen sensitisation or confirmed food allergy), treating sibling count as a continuous predictor with children who had no siblings as the reference group. All tests were two‐sided, with p < .05 considered statistically significant.

3. RESULTS

3.1. Participant follow‐up and dataset completeness

A total of 209 children participated in the 5‐year follow‐up. 181 children attended the in‐person visit; 175 consented for SPT to food and aeroallergens. These denominators were used for questionnaire‐based outcomes (N = 208), clinical assessments (N = 181) and sensitisation analyses (N = 175), respectively. Participant flow is illustrated in Figure 1.

3.2. Health outcomes in the CORAL cohort between 2 and 5 years

Participant health measures were compared between the 2 and 5‐year using questionnaire‐derived data (Table 1). Significant changes were observed across all evaluated domains over time. SARS‐CoV‐2 infections predictably decreased between age 2 and 5 years. In contrast, reported infections, systemic antibiotic exposure, hospital admissions and out‐of‐home childcare or school attendance all increased.

TABLE 1.

Environment and health outcomes.

	CORAL at 2 years N = 319	CORAL at 5 years N = 208 a	p
COVID‐19 infections	157 (49.2%)	71 (34.1%)	<.001
All reported infections	253 (79.3%)	185 (88.9%)	<.01
Systemic antibiotics	168 (52.6%)	162 (77.8%)	<.001
Hospital admissions	28 (8.7%)	52 (25%)	<.001
Out of home childcare or education	219 (68.6%)	196 (94.2%)	<.001

^a Total N varies by outcome due to loss to follow‐up.

3.3. Atopic outcomes in the CORAL cohort

3.3.1. Atopy trajectories over time

Between 2 and 5 years, atopic dermatitis declined (21.3%–10.4%), food allergy remained uncommon and aeroallergen sensitisation increased (8.9%–24.0%). These trends are illustrated in Figure 2, summarised in Table 2 and shown longitudinally in Table S10.

FIGURE 2.

Trajectories of key atopic outcomes from 1 to 5 years.

TABLE 2.

Atopic outcomes at 2 and 5 years.

	CORAL 2 years N = 320	CORAL 5 years N = 175 a	Chi‐square (χ 2 )	p
Atopic dermatitis	68	19 (N = 181) b	8.58	p < .01
Peanut allergy	2	1	0.005	.94
Tree nut allergy	3	3	0.57	.45
Milk allergy	0	0	N/A	N/A d
Egg allergy	0	0	N/A	N/A d
Fish allergy	0	1	1.832	N/A d

	CORAL 2 years N = 292 c	CORAL 5 years N = 175 a	Chi‐square (χ 2 )	p
Aeroallergen sensitisation	26	42	18.85	<.001

	CORAL 2 years N = 320	CORAL 5 years N = 208	Chi‐square (χ 2 )	p
Parent reported wheeze	76	46	0.001	.98
Doctor diagnosed asthma	7	21	15.7	<.001

^a Total N for skin prick tests is 175 due to compliance at 5‐year follow‐up.

^b Total N reflects patient attendance.

^c Total N varies where outcomes were not available for all participants.

^d χ ² and p‐value not applicable because there were no cases in either group; therefore statistical comparison cannot be performed.

3.3.2. Atopic dermatitis

At 5 years, 31/181 (17.1%) parents reported atopic dermatitis. Clinical evaluation identified active atopic dermatitis in nineteen (10.4%; Table 2) children based on SCORAD assessment (median SCORAD 10.9; IQR 8.35–15.3). Eighteen cases were mild, one was moderate and none were severe. Among children with atopic dermatitis at 2 years, 49 of 68 had resolved by 5 years, 11 had persistent atopic dermatitis (predominantly mild) and eight new mild cases emerged. Children with atopic dermatitis had higher food allergy rates than those without atopic dermatitis, although this association was not statistically significant (Table S1).

3.3.3. Food allergy

Food allergy remained uncommon at 5 years (Table 2). No new cases of IgE milk or egg allergy were identified. Peanut allergy decreased from two cases to one, while tree nut allergy showed turnover, with two cases resolving and two new cases emerging. One new fish allergy was identified between 2 and 5 years. Given the low number of food allergy cases, no multivariable analyses were performed and no significant associations with environmental exposures were identified (Table S2).

3.3.4. Aeroallergens, asthma and parent reported wheeze

Aeroallergen sensitisation increased substantially between 2 and 5 years (Table 2). House dust mite was the predominant sensitiser (n = 33), followed by grass (n = 25), dog (n = 11) and tree (n = 3).

13% of children were polysensitised. At 5 years, aeroallergen sensitisation was strongly associated with allergic rhinitis symptoms, with polysensitised children at the highest risk of allergic rhinitis (Table S3).

Among individual allergens at 5 years, sensitisation to house dust mite and animal dander was associated with wheeze, whereas no allergen was associated with physician‐diagnosed asthma (Tables S4 and S5). In a multivariable model adjusting for environmental exposures, no factors were significantly associated with sensitisation (Table S6). In contrast, respiratory outcomes showed a different pattern; wheeze at age 5 was predominantly associated with recent infection, as outlined in the following section.

3.3.5. Respiratory outcomes

At 5 years, 46/208 (22.1%) children had parent‐reported wheeze, and 21/208 (10.1%) had doctor‐diagnosed asthma. Parent‐reported wheeze was strongly associated with doctor‐diagnosed asthma (χ ² = 33.0, p < .001). Wheeze prevalence remained stable between 2 and 5 years, whereas asthma prevalence increased (Table 2).

In multivariable analysis, urban residence (adjusted OR 2.86, 95% CI 1.00–8.15, p = .049) and household smoking exposure (adjusted OR 2.30, 95% CI 1.02–4.76, p = .044) were independently associated with parent‐reported wheeze at 5 years. No significant associations were observed with pets, farm animal exposure, daycare attendance, or antibiotic use. Recent infection was strongly associated with wheeze but could not be included in the multivariable model because of complete separation (Table S7). Doctor diagnosed asthma at 5 years showed no significant associations with any early‐life environmental exposures (Table S8).

3.3.6. Co‐occurrence of atopic outcomes at 5 years

The overlap of food allergy, aeroallergen sensitisation, eczema and physician‐diagnosed asthma at age two and five is shown in Figure 3.

FIGURE 3.

Venn diagram of CORAL atopic outcomes at 2 and 5 years. Asthma counts at 5 years are restricted to children attending the in‐person visit; prevalence analyses additionally include questionnaire data from non‐attenders.

3.3.7. Family structure

Sibling number was evaluated as a proxy for early‐life microbial exposure. Logistic regression modelling sibling number as a continuous variable showed no significant association with atopy. Although a trend towards lower odds of atopy with increasing sibling number was observed, this did not reach statistical significance (Figure 4; Table S9).

FIGURE 4.

Odds ratios (95% CI) for atopy according to number of siblings. Odds ratios were estimated using logistic regression modelling sibling number as a continuous variable, with children who had no siblings as a reference group (OR = 1.0). The solid blue line represents the model‐predicted odds ratio as a function of sibling number; the shaded area indicates the model‐based 95% confidence interval. Observed odds ratios with 95% CI are shown by orange points with vertical error bars. The model coefficient corresponds to an odds ratio of 0.75 per additional sibling (95% CI 0.48–1.18; p = .21).

4. DISCUSSION

This prospective 5‐year longitudinal study provides insight into the long‐term atopic outcomes of infants born during the COVID‐19 pandemic. By 5 years of age, atopic dermatitis had largely resolved and food allergy remained uncommon, whereas aeroallergen sensitisation increased substantially and was strongly associated with allergic rhinitis symptoms. Parent‐reported wheeze remained stable despite an increase in doctor‐diagnosed asthma and was more strongly associated with recent infection than aeroallergen sensitisation. Lockdown‐associated factors may have conferred protection against early food allergy, but by 5 years children were fully socially reintegrated. We propose that lockdown‐associated early‐life environmental and lifestyle changes including higher breastfeeding rates, lower antibiotic exposure and early allergenic food introduction, as previously reported, ⁴ may have altered early risk for food allergy, a benefit that persisted to 5 years, while the broader evolution of allergic disease remained consistent with recognised age‐related trajectories.

The marked reduction in atopic dermatitis between 2 and 5 years aligns with established natural history patterns and earlier CORAL findings. By 5 years, atopic dermatitis had fallen to 10.4%, with almost all cases mild and none severe. These findings align with major pre‐pandemic cohorts. ¹⁰ , ¹¹ Comparable observations were also reported in the MIS BAIR study, a pre‐pandemic birth cohort whose 5‐year follow‐up coincided with the COVID‐19 lockdowns. ¹² The convergence of our findings with those diverse populations, PreventADALL in particular, suggests that the early trajectory of atopic dermatitis is highly robust across settings, and that intrinsic factors such as genetic predisposition and skin‐barrier characteristics early in life may be more influential in determining long‐term outcomes than short‐term fluctuations in environmental exposures.

Food allergy remained uncommon at 5 years, with most early allergies resolving and very few new cases emerging. The overall prevalence at 5 years was substantially lower than that reported in major pre‐pandemic population‐based cohorts. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ The CORAL cohort's initial trajectory ran counter to the hygiene hypothesis, which would anticipate higher allergy rates following altered microbial exposures in early life. We have previously proposed that clinician‐supported early allergen introduction, increased family compliance, reduced antibiotic use and increased breastfeeding duration facilitated by the pandemic lockdown may have supported durable oral tolerance induction, potentially through early‐life vertical microbial transmission (e.g., bifidobacteria) and appropriate dietary advice. ¹⁷ , ¹⁸ The persistence of this finding to 5 years suggests that pandemic‐associated environmental changes influenced early food allergy risk beyond infancy, rather than simply delaying disease expression.

Aeroallergen sensitisation rose from 9% at age 2 to 24% at age 5, with the largest increase seen for house dust mite, alongside a substantial rise in grass pollen sensitisation. This is consistent with the typical emergence of inhalant allergens during the preschool years reported in other birth cohorts. ¹⁹ , ²⁰ By age 5, 173/208 children (81%) were attending formal childcare or school, representing a shift from predominantly home‐based environments in infancy to broader environmental exposures. This transition may be associated with a shift in aeroallergen sensitisation over this period, with seasonal outdoor allergens such as grass pollen becoming more prominent. The contrasting persistence of reduced food allergy alongside age‐appropriate increases in aeroallergen sensitisation suggests that different allergic phenotypes may differ in their susceptibility to early‐life environmental modification.

Sensitisation to house dust mite and animal dander was associated with parent‐reported wheeze. Notably, recent infection showed complete separation with wheeze, indicating a very strong association between infection and wheeze at age 5. These findings suggest that wheeze at age 5 largely reflects infection‐associated respiratory illness, with aeroallergen sensitisation acting as a secondary risk factor rather than a primary driver. In contrast, physician‐diagnosed asthma was uncommon and was not associated with either infection or aeroallergen sensitisation. Pandemic‐related shifts in early‐life exposures in this cohort, including higher breastfeeding rates and reduced antibiotic use, were associated with distinct gut microbiota profiles. These profiles were characterised by higher Bifidobacterium abundance at 6 months and increased butyrate‐producing taxa at 12 months and were inversely associated with atopic dermatitis and food allergy skin prick test positivity. ¹⁸ Such early immunoregulatory signals may plausibly modify or delay progression along the atopic pathway towards asthma, potentially contributing to the absence of observed associations with asthma at 5 years of age. However, this interpretation remains speculative given the low prevalence of asthma and the possibility that asthma phenotypes may not yet be fully established.

Hospital admissions more than doubled between ages 2 and 5, coinciding with increased out‐of‐home childcare and social mixing following the lifting of public health restrictions. This is consistent with recent models of altered population susceptibility following periods of reduced pathogen exposure. ²¹

This study has several strengths. CORAL is a globally unique longitudinal birth cohort established during the earliest months of the COVID‐19 pandemic restrictions. It provides a unique opportunity to examine whether the profound alterations in early‐life environmental exposures associated with the COVID‐19 pandemic influenced the natural history of allergic disease beyond infancy. Detailed allergy phenotyping was undertaken at both 2 and 5 years using the same standardised questionnaires, ensuring consistency across time points. The cohort represents a stable population with a retention rate of approximately 65%, supporting the validity of longitudinal analyses. Limitations include modest sample sizes for some outcomes and subgroup analyses, particularly food allergy, doctor diagnosed asthma and sibling‐related analyses, which limited statistical power. Asthma at age 5 was ascertained by parent report of a doctor's diagnosis and may represent a heterogeneous and evolving phenotype, such that associations with early‐life risk factors may become more apparent with longer follow‐up. The cohort was drawn predominantly from a socioeconomically advantaged population, which may limit generalisability. Some participants (33/208) declined skin prick testing, although this is unlikely to have materially biased prevalence estimates. As with all longitudinal studies, loss to follow‐up may introduce selection bias. Although some participants formally withdrew consent, most losses reflected non‐response to follow‐up contact. As we did not compare participants retained at 5 years with those lost to follow‐up; therefore the potential for attrition bias cannot be excluded.

Even after an unprecedented period of social isolation and altered microbial exposures in infancy, the order in which allergic conditions emerged in CORAL broadly followed the classical atopic march, although the prevalence of individual conditions, such as food allergy, differed from expectations. Thus providing evidence that although early environmental disruptions may influence the initial expression of atopy, they do not necessarily alter the later trajectory through which allergic disease unfolds. Continued longitudinal follow‐up into later childhood will be essential to determine whether early aeroallergen sensitisation translates into increased asthma risk and to clarify the longer‐term clinical implications for this cohort.

AUTHOR CONTRIBUTIONS

Anitha Sokay: Conceptualization; investigation; writing – original draft; methodology; validation; visualization; writing – review and editing; software; formal analysis; data curation; project administration. Sinead Burke: Investigation; project administration; data curation; resources. Susan Byrne: Conceptualization; investigation; methodology; supervision. Maria Iatan: Investigation; project administration; data curation; resources. Jordan Crupper: Investigation; project administration; data curation. Sarah Byrne: Investigation; data curation; project administration. Jonathan O'B Hourihane: Conceptualization; investigation; funding acquisition; methodology; validation; visualization; writing – review and editing; project administration; data curation; supervision; resources. Siobhan Verdon: Investigation; project administration; data curation. Liam O'Mahony: Writing – review and editing; supervision; validation.

FUNDING INFORMATION

Research Ireland Frontiers for the Future Programme (23/FFP‐A/12076) provided funding.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Supporting information

Table S1. Food allergy and AD severity at 2 and 5 years.

Table S2. Food allergy and exposure variables.

Table S3. Sensitivity and specificity calculated using reported allergic rhinitis symptoms as the reference standard.

Table S4. Association between sensitisation to individual aeroallergens and parent‐reported wheeze at age 5.

Table S5. Association between sensitisation to individual aeroallergens and doctor diagnosed asthma at age 5.

Table S6. Adjusted multivariable model for environmental exposures and aeroallergen sensitisation.

Table S7. Multivariable logistic regression model for predictors of wheeze at 5 years.

Table S8. Multivariable logistic regression model for predictors of asthma at 5 years.

Table S9. Prevalence of atopy by number of siblings.

Table S10. Prevalence of atopic outcomes from 1 to 5 years and statistical comparison of change.

Sokay A, Burke S, Iatan M, et al. Atopic outcomes at 5 years in the CORAL cohort born during the 2020 COVID‐19 lockdown. Pediatr Allergy Immunol. 2026;37:e70477. doi: 10.1111/pai.70477

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.