The age of eczema onset and multiple food allergies

Abstract

Background:

Atopic dermatitis (AD) is the most common pediatric atopic disease and a risk factor for food allergy (FA) development. Children with AD are prone to developing allergic comorbidities. Its relationship with FA phenotype and timing of onset remains unclear.

Objective:

To evaluate associations between AD presence and onset timing with FA-related outcomes and atopic comorbidities.

Methods:

FORWARD is a longitudinal, multicenter study enrolling children younger than 12 years with physician-diagnosed FA between 2017 and 2024. Written informed consent was obtained. Data were collected from baseline survey and electronic medical records. Logistic regression models adjusted for demographic and socioeconomic factors assessed associations between AD with FA number, type of allergens, and atopic comorbidities.

Results:

Among 1309 children with physician-diagnosed FA, 77% reported history of AD, with 30.3% between 1 and 3 months, 50.2% at 4 to 12 months, and 19.5% after 12 months. Multiple FAs, asthma, and allergic rhinitis were each associated with higher odds of AD. Late-onset AD was associated to lower odds of multiple FAs compared with early onset AD (odds ratio 0.57, 95% CI 0.38-0.85), including lower odds of milk and egg allergies. Conversely, late-onset AD was associated with higher odds of asthma and allergic rhinitis.

Conclusion:

Among children with FAs, 77% report a history of AD, with more than 80% of children with FA developed AD within the first year of life. Children with multiple FAs and those with allergies to milk and egg are more likely to develop AD on earlier age onset, whereas asthma and allergic rhinitis were associated with late-onset AD.

Introduction

Atopic dermatitis (AD) is a chronic inflammatory skin condition that affects approximately 10% of children in the United States.¹ AD often starts early in life, with 50% to 60% of cases developing within the first year.² In addition to its significant impact on a child’s wellness and quality of life, AD is strongly associated with other allergic conditions, including asthma, allergic rhinitis, and food allergies (FAs).³ This relationship between AD and FA is a key component of the atopic march.⁴

Although studies have consistently linked AD to FA development, a critical question remains: How is the timing of AD onset associated with FA phenotype? Current literature suggests that early onset AD (within the first year of life) is a stronger risk factor for FAs than late-onset AD,⁵ but few studies have evaluated this relationship in detail. One study suggests that infants who develop AD by 3 months of age are at particularly high risk for persistent AD and FA,⁵ but the implications for FA phenotype, such as specific FAs, other atopic comorbidities, or multiple FAs, remain unclear.

Disruptions of the skin barrier in AD may facilitate sensitization to food allergens through microbial dysbiosis, which drive immune dyssregulation and promote FA development.⁶ Similarly, loss-of-function mutations in the filaggrin gene, which impair epidermal barrier integrity, are strongly associated with AD and subsequently contribute to the development and persistence of FA.⁷

However, the extent to which the history of AD or its onset timing (early vs late-onset AD) affects this pathway remains poorly understood. This information is important to plan early screening and develop intervention strategies, potentially leading to new approaches for preventing FA in at-risk children. To address this knowledge gap, our study investigated the relationship between the presence of AD and onset timing of AD with FA-related factors in a large cohort of children with FA. We aimed to determine whether early onset of AD is more strongly associated with multiple FAs, allergies to specific food allergens, and associated comorbidities compared with later onset AD.

Methods

The study was conducted as part of a large multicenter, longitudinal, observational cohort assessing the natural history and management of FA—FORWARD (Food Allergy Management and Outcomes Related to Racial/Ethnic Differences from Infancy through Adolescence).^8,9 Participants were identified and recruited consecutively from allergy clinics at 4 urban tertiary medical centers across the United States, including Ann & Robert H. Lurie Children’s Hospital of Chicago (Lurie), Rush University Medical Center, Children’s National Hospital, and Cincinnati Children’s Hospital Medical Center between 2017 and 2024. Children aged 2 years or younger with a physician-diagnosed FA were eligible for enrollment. On enrollment, parents or guardians of children with FA provided written consent and completed electronic surveys. Institutional review board approvals were obtained at all sites. The cohort initially enrolled non-Hispanic Black and White participants. Later, recruitment was expanded to include children of Hispanic and non-Hispanic Asian ethnicity, resulting in small sample sizes for these groups at the time of analysis.

Definitions of outcomes are as follows:

• FA: Physician-confirmed IgE-mediated FA, defined by a history of clinical reaction consistent with IgE-mediated symptoms and evidence of sensitization (positive skin prick test result or elevated food-specific IgE level). Oral food challenges were performed when clinically indicated .The number and type of FAs were reported by parents through the intake survey.

• AD: Parent-reported history of physician-diagnosed AD, collected through intake survey. Parents indicated whether the child had ever been diagnosed with AD (a condition in which the skin has a dry, itchy rash; the skin may be cracked or maybe oozing) and provided age at diagnosis.

• Asthma and allergic rhinitis: Physician-diagnosed conditions extracted from electronic health record based on documented clinical criteria.

The primary outcome measured in our study was the presence or absence of AD, including timing of AD onset. Outcome data were obtained from parent surveys. All patients’ AD status and FA data were determined using a parent-reported history of AD at the intake survey during their baseline study visit. Furthermore, we stratified the age of AD onset into 3 distinct age groups—early onset (1-3 months), 4 to 12 months onset, and late-onset (>12 months)—for further comparisons. We collected data on the number and type of food allergens, the 9 most common allergens in the United States (peanuts, tree nuts, milk, egg, wheat, soy, finfish, shellfish, and sesame), with an additional “other” category for less common allergens grouped together. This study also captured the history of physician-diagnosed allergic rhinitis and asthma which was confirmed by electronic medical record review. Incomplete surveys (11%) were excluded from statistical analysis, with missing/incomplete demographic or clinical data from intake surveys.

Data from each respective FORWARD site were exported from Research Electronic Data Capture (REDCap) databases and pooled for the current analyses. Race and ethnicity categories followed the National Institutes of Health guidelines (NOT-OD-15-089). Based on ethnicity, children whose parents identified them as Hispanic were grouped as Hispanic for analysis regardless of reported race. Non-Hispanic children with different racial backgrounds were grouped based on the racial group which they primarily identified. The gender defined as a socially constructed identity reported in the source data information on biological sex was not collected.

All analyses were conducted using IBM SPSS Statistics version 29.0.0.0 (241), and a two-sided P less than .05 was used to indicate statistical significance. Descriptive statistics were obtained to summarize cohort characteristics, including race, gender, age of AD onset, annual household income, and atopic comorbidities. The χ² tests were used to assess the associations between categorical variables. One-way analysis of variance was used to compare means across groups, such as AD vs no AD groups. Logistic regression models were constructed to evaluate associations between AD presence and its age of onset with independent variables including demographic factors (gender, race, and enrollment site) and socioeconomic status (household income). Next, we assessed the association between AD and its age of onset with clinical variables including FA-related outcomes and atopic comorbidities using logistic regression models, in which these clinical variables were treated as dependent outcomes. All models were adjusted for demographic variables (gender, race, and enrollment site), recruitment site, and household income.

Results

Cohort Demographics

A total of 1309 children with physician-diagnosed FA were included in the analysis. Participant race and ethnicity were distributed as follows: Hispanic (21.9%), non-Hispanic White (37.6%), non-Hispanic Black (35%), Asian (5%), and Multiracial (0.5%) (eTable 1). Demographic and clinical characteristics and comparisons between the 4 sites are detailed in eTable 1.

Among our cohort, 1008 (77%) reported a history of physician diagnosed AD (Table 1). Furthermore, 489 (37.4%) had asthma and 492 (37.6%) had allergic rhinitis. Prevalence of AD, asthma, and allergic rhinitis was similarly distributed across the enrollment centers.

Table 1.

Logistic Regression of Demographic Characteristics of Children With FA With and Without AD, Aged 0 to 12 Years Enrolled in FORWARD Between 2017 and 2024

Characteristics	No AD (Ref) n = 301 (23.0%)	AD
		n = 1008 (77.0%)	OR	(95% CI)	Adjusted P value
Race
Non-Hispanic White (ref)a	110 (36.5)	382 (37.9)	Ref	-	-
Non-Hispanic Black	99 (32.9)	359 (35.6)	0.92	0.64-1.33	.67
Hispanic	79 (26.2)	208 (20.6)	1.15	0.78-1.69	.49
Asian	13 (4.3)	53 (5.3)	0.58	0.29-1.14	.11
Multiracialb	0 (0)	6 (0.6)	-	-	-
Gender
Male (ref)	167 (55.5)	622 (61.7)
Female	134 (44.5)	386 (38.3)	1.30	0.99-1.69	.055
Annual household income
≥$100k (ref)	114 (37.9)	419 (41.6)		-	-
$50k-$99k	41 (13.6)	185 (18.4)	0.96	0.64-1.44	0.84
<$50k	76 (25.2)	292 (29.0)	0.80	0.52-1.22	.29
Nonresponderc	70 (23.3)	112 (11.1)	-	-

Abbreviations: AD, atopic dermatitis; FA, food allergy; n, sample size of the cohort; OR, odds ratio; ref, reference group.

NOTE. Data are presented as number (percentage), unless otherwise specified.

^aORs reveal the likelihood of having AD compared with not having AD (ref). The White race was the reference group for comparisons across race, as it had the largest total number of children recruited in the FORWARD study.

^bThe marked analysis was not possible due to a small sample size for this group.

^cNonresponders were not included in the analysis.

The most common types of food allergens among children in our cohort were peanut (65.2%), tree nut (52.1%), egg (48.7%), milk (27.8%), shellfish (17.3%), sesame (15.4%), finfish (13.9%), soy (10.8%) wheat (10.4%), and others (19.2%).

Atopic Dermatitis and Phenotype of Food Allergies

The logistic regression analyses revealed that AD was more frequently observed among children with multiple FAs (odds ratio [OR] 2.75, 95% CI 1.99-3.67). When evaluating individual allergens, AD was more common among children with peanut (OR 1.85, 95% CI 1.37-2.50), tree nut (OR 2.10, 95% CI 1.57-2.85), milk (OR 1.79, 95% CI 1.25-2.55), egg (OR 1.49, 95% CI 1.11-2.00), wheat (OR 2.24, 95% CI 1.31-3.82), soy (OR 1.48, 95% CI 1.07-2.36), and shellfish (OR 1.60, 95% CI 1.03-2.50). Similarly, children with asthma (OR 1.78, 95% CI 1.29-2.45) and allergic rhinitis (OR 1.76, 95% CI 1.27-2.52) had significantly higher odds of reporting AD compared with children without these conditions (Table 2).

Table 2.

Logistic Regression of Clinical Characteristics of Children With FA With and Without AD, Aged 0 to 12 Years Enrolled in FORWARD Between 2017 and 2024^a

Characteristics	No AD (ref) n = 301 (23.0%)	AD
		n = 1008 (77.0%)	OR	(95% CI)	Adjusted P value
Atopic comorbidities
Asthma	72 (23.9)	417 (41.4)	1.78	1.29-2.45	<.001
AR	86 (28.6)	406 (40.3)	1.76	1.27-2.5	<.001
Multiple FAs	166 (55.1)	770 (76.4)	2.7	1.99-3.67	<.001
Specific food allergens
Peanut	166 (55.1)	688 (68.3)	1.85	1.37-2.5	<.001
Tree nut	122 (40.5)	560 (55.6)	2.10	1.57-2.85	<.001
Milk	62 (20.6)	302 (30.0)	1.79	1.25-2.55	.001
Egg	120 (39.9)	518 (51.4)	1.49	1.10-2.00	.008
Wheat	17 (5.6)	119 (11.8)	2.66	1.4-5.04	.003
Sesame	40 (13.3)	162 (16.1)	1.37	0.90-2.13	.151
Soy	25 (8.3)	116 (11.5)	1.87	1.07-3.29	.03
Finfish	36 (12.0)	146 (14.5)	1.42	0.90-2.23	.14
Shellfish	43 (14.3)	183 (18.2)	1.60	1.032-2.5	.03
Other	48 (15.9)	203 (20.1)	1.30	0.83-1.77	.33

Abbreviations: AD, atopic dermatitis; FA, food allergy; n, sample size of the cohort; OR, odds ratio; ref, reference group.

NOTE. Each outcome was evaluated in a separate logistic regression model. Models were adjusted for demographic variables (gender, race, and enrollment site) and socioeconomic factors (household income). The dependent variables included asthma, allergic rhinitis, multiple FAs, and individual food allergens. Data are presented as number (percentage), unless otherwise specified. Results are significant at P is less than .05.

^aORs reveal the likelihood of having AD compared with not having AD (ref).

The Onset of Atopic Dermatitis

Most children with FA experienced AD within the first year of life, with 30.3% being diagnosed with AD between 1 and 3 months (called early onset in this report), 50.2% between the age of 4 and 12 months, and 19.5% after 12 months (called late onset or >12 months) (Table 3).

Table 3.

Multinominal Regression Analysis Was Used to Compare Demographic, Socioeconomic, and Clinical Characteristics of Children Aged 0 to 12 Years With Food Allergy Enrolled in the FORWARD Study: A Comprehensive Analysis by Age of Eczema Onset^a

Characteristics	Early onset (1-3 mo) (Ref) n = 305 (30.3%)	4-12 mo				Late onset (12+ mo)				Total n = 1006b
		n = 505 (50.2%)	OR	(95% CI)	Adjusted P value	n = 196 (19.5%)	OR	(95% CI)	Adjusted P value
Race
Non-Hispanic, White (Ref)c	119 (39.0)	212 (42.0)	Ref	-	-	51 (26.0)	-	-	-	382 (38.0)
Non-Hispanic, Black	102 (33.4)	170 (33.7)	1.09	0.73-1.63	.68	87 (44.4)	1.85	1.09-3.13	.022	359 (35.7)
Hispanic	68 (22.3)	93 (18.4)	0.82	0.52-1.27	.37	45 (23.0)	1.19	0.67-2.13	.55	206 (20.5)
Asian	16 (5.2)	24(4.8)	0.88	0.44-1.78	.73	13(6.6)	1.65	0.71-3.81	.24	53 (5.3)
Multiple raced	0 (0)	6 (1.2)	-	-	-	0 (0)	-	-	-	6 (0.6)
Gender
Male (Ref)	183(60)	306 (60.6)	Ref	-	-	131 (66.8)	-	-	-	620 (61.6)
Female	122 (40.0)	199 (39.4)	0.97	0.72-1.30	.84	65 (33.2)	0.74	0.51-1.09	.13	386 (38.4)
Annual household incomee
≥$100k (Ref)	125 (46.5)	228 (50.1)	Ref	-	-	64 (37.6)	-	-	-	417 (46.6)
<$50k	90 (33.5)	125 (27.5)	0.83	0.54-1.29	.41	77 (45.3)	1.42	0.83-2.47	.20	292 (32.7)
$50k-$99k	54 (20.1)	102 (22.4)	1.14	0.74-1.75	.55	29 (17.1)	1.03	0.57-1.86	.91	185 (20.7)
Institution
Northwestern/Lurie	85 (27.9)	177 (35.0)	Ref	-	-	80 (40.8)	-	-	-	342 (34.0)
Rush	74 (24.3)	96 (19.0)	0.63	0.42-0.95	.026	48 (24.5)	0.62	0.38-1.00	.051	218 (21.7)
CCHMC	82 (26.9)	123 (24.4)	0.67	0.45-1.00	.050	36 (18.4)	0.44	0.26-0.75	.002	241 (24.0)
CNHS	64 (21)	109 (21.6)	0.75	0.50-1.13	.17	32 (16.3)	0.51	0.30-0.87	.013	205 (20.4)

Abbreviations: CCHMC, Cincinnati Children’s Hospital Medical Center; CNHS, Children’s National Health Systems Washington, DC; n, sample size of the cohort; OR, odds ratio; Ref, reference group.

NOTE. Data are presented as number (percentage), unless otherwise specified.

^aORs reveal the likelihood of each characteristic across age of eczema onset groups relative to the reference age of onset group (1-3 months).

^bTwo participants were missing the timing of eczema diagnosis.

^cWhite race was the reference group for the comparisons across race due to having the largest total number of children recruited in the FORWARD study.

^dThe marked analysis was not possible due to a small sample size for this group.

^eNon-responders were not included in analysis.

We did not observe any significant differences between the age of AD onset across institutions or household income categories (Table 3). Race was associated with the age of AD onset. In comparison to non-Hispanic White children, non-Hispanic Black children were more likely to report late AD onset (>12 months at diagnosis) (OR 1.85, 95% CI 1.09-3.13) (Table 3).

The Impact of Age of Atopic Dermatitis Onset on Multiple Food Allergy, Type of Food Allergies, and Atopic Comorbidities

We then evaluated associations of age of AD onset with multiple FAs, individual food allergens, asthma, and allergic rhinitis as separate dependent variables, adjusting for gender, race, enrollment site, and household income. Children with AD onset at above or equal to 12 months had lower odds of multiple FAs compared with children with onset at 1 to 3 months (OR 0.57, 95% CI 0.38-0.85). Similarly, older onset was associated with reduced odds of specific FAs, including milk (OR 0.40, 95% CI 0.26-0.62) and egg (OR 0.41, 95% CI 0.28-0.59).

In contrast, when respiratory outcomes were analyzed, late-onset AD was associated with significantly higher odds of both asthma (OR 2.50, 95% CI 1.72-3.62) and allergic rhinitis (OR 1.79, 95% CI 1.23-2.62) compared with children whose AD began at 1 to 3 months (Table 4).

Table 4.

Logistic Regression Analysis Was Performed to Evaluate the Association Between Age of Onset and the Presence of MFA, Adjusting for Race, Gender, Institution, and Household Income

Characteristics	Early onset (1-3 mo) n = 305 (30.3%) (ref)	4-12 mo n = 505 (50.2%)				Late onset (12+ mo) n = 196 (19.5%)				Total n = 1006b
		n (%)	OR	(95% CI)	Adjusted P value	n (%)	ORa	(95% CI)	Adjusted P value
Atopic comorbidity
Asthma	112 (36.7)	187 (37.0)	1	0.74-1.34	.99	117 (59.7)	2.50	1.72-3.63	<.001	416 (41.4)
Allergic rhinitis	111 (36.4)	199 (39.4)	1.10	0.82-1.49	.52	96 (49.0)	1.8	1.23-2.6	.002	406 (40.4)
Number of food allergens
Single	67 (22.0)	104 (20.6)	Ref	-	-	66 (33.7)	-	-	-	237 (23.6)
MFAs	238 (78.0)	401 (79.4)	1.00	0.76-1.5	.68	130 (66.3)	0.57	0.36-0.85	.006	769 (76.4)
Specific food allergens
Peanut	202 (66.2)	366 (72.5)	1.34	0.98-1.84	.065	119 (60.7)	0.85	0.58-1.23	.377	687 (68.3)
Tree nut	163 (53.4)	289 (57.2)	1.Q5	0.86-1.54	.344	106 (54.1)	1.080	0.75-1.3	.68	558 (55.5)
Milk	112 (36.7)	152 (30.1)	0.72	0.54-0.98	.036	38 (19.4)	0.40	0.26-0.62	<.001	302 (30.0)
Egg	174 (57.0)	275 (54.5)	0.90	0.67-1.19	.45	68 (34.7)	0.41	0.28-0.60	<.001	517 (51.4)
Wheat	42 (13.8)	59 (11.7)	0.82	0.54-1.26	.37	18 (9.2)	0.60	0.33-1.08	.088	119 (11.8)
Sesame	58 (19.0)	80 (15.8)	0.77	0.53-1.10	.18	23 (11.7)	0.61	0.36-1.04	.060	161 (16.0)
Soy	35 (11.5)	60 (11.9)	1.03	0.66-1.63	.91	21 (10.7)	0.92	0.55-1.65	.76	116 (11.5)
Finfish	41 (13.4)	75 (14.9)	1.15	0.75-1.6	.58	29 (14.8)	1.1	0.65-1.85	.72	145 (14.4)
Shellfish	48 (15.7)	89 (17.6)	1.16	0.79-1.7	.44	45 (23.0)	1.55	0.99-2.46	.058	182 (18.1)
Other	66 (21.6)	97 (19.2)	0.86	0.86-0.61	.40	40 (20.4)	0.92	0.58-1.43	.68	203 (20.2)

Abbreviations: MFA, multiple food allergies; n, sample size of the cohort; OR, odds ratio; ref, reference group.

NOTE. Data are presented as number (percentage), unless otherwise specified.

Each outcome was evaluated in a separate logistic regression model. The dependent variables included asthma, allergic rhinitis, MFA, and individual food allergens.

The White race was the reference group for comparisons across races, as it had the largest total number of children recruited in the FORWARD study.

^aORs reveal the likelihood of each characteristic across age of eczema onset groups relative to the reference age of onset group (1-3 months). Models were variables (gender, race, and enrollment site) and socioeconomic factors (household income). Results are significant at P less than .05.

^bTwo participants were missing the timing of eczema diagnosis.

Discussion

In this cohort of children with physician-diagnosed FA, 77% had AD, and in more than 30% of those cases, AD began within the first 3 months of life. Interestingly, children with multiple FAs and allergies to milk and eggs were more likely to have early onset AD.

The relationship between FA and AD is well recognized but not uniform. Some children with AD develop sensitization without symptoms, and oral food challenges are needed to confirm true food triggers¹⁰ Others have a more complex pattern, in which sensitization and clinical reactions do not always align.¹¹ Recent prospective data add further clarity by revealing a bidirectional association; early AD increases later FA risk and early FA predicts later AD. This variation across children highlights the need to interpret FA-AD links within their broader clinical context.¹²

The association of AD and the timing of AD onset with multiple FA is supported by studies indicating that infants with AD are significantly more likely to develop FAs compared with those without AD.⁵

Several factors have been proposed to explain why AD increases the risk of FA. Disruption of the skin barrier in the context of AD, such as in individuals with filaggrin mutation, increases the susceptibility to cutaneous exposure to allergens.^13,14 In addition, environmental exposures, including allergens and certain skin care practices, have been investigated for their potential role in FA risk, although evidence remains mixed regarding early moisturizer use.^15-17 It is noteworthy that genetic predispositions that affect epidermal differentiation of the stratum corneum of the skin and compromise skin barrier function are risk factors for early onset AD and increase the risk of allergen sensitization, hence development of multiple FAs.^6,13,17 In the same line, as mentioned previously, early onset AD and dysfunctional skin barrier early in life create an increased opportunity for cutaneous immune exposure to food allergens before dietary introduction of these foods can induce oral tolerance.¹⁸ This is supported by our finding that children with cow’s milk, eggs, and wheat allergies were more inclined to develop early onset of AD. Those are the common foods in most households, and the chances of cutaneous exposure to their particles are very high. A study in Norway found milk and egg allergens in 39% and 22% of the mattress dust samples, respectively.¹⁹ In alignment with our findings, Kahveci et al²⁰ identified severe and early onset AD as a predictor of both the presence and severity of FAs in infancy, particularly increasing the risk of multiple FAs, including egg and cow’s milk. Similarly, Martin et al,⁵ in the HealthNuts cohort study, demonstrated that infants with AD have a higher likelihood of developing FAs, with those experiencing AD by 3 months at particularly high risk. Their findings revealed that infants with AD were 6 times more likely to develop egg allergy by 12 months compared with those without AD.⁵

Meta analyses and large reviews have revealed that the prevalence of AD among children with FA ranges from approximately 45% to 71% in different studies.^21,22 Similar proportion to our large FA cohort, 77% of children reported AD.

We also clearly revealed that AD increases the odds of asthma and allergic rhinitis among children with FA. Consistent with literature, 41.4% and 40.3% of children with a physician-diagnosed FA and AD had asthma and allergic rhinitis, respectively.^23-28

The relationship between AD and other atopic conditions is the key concept of the atopic march, which typically begins with AD in infancy, followed by the development of FA, allergic rhinitis, and asthma in later childhood.²⁹ Prior work indicates that environmental factors, including diet and microbial exposures, play a crucial role in the development of these allergic conditions in children with AD.⁸

Interestingly, we found that late-onset AD was associated with a higher risk of asthma and allergic rhinitis compared with those with early onset AD. It is possible that, in our study, children who were diagnosed with AD at a very early age received more frequent follow-up care, experienced changes in caregiver behavior, and had modifications made to their home environments to reduce exposure to environmental allergens, such as dust mites. These interventions may have contributed to a lower rate of sensitization to environmental allergens and, consequently, a reduced risk of developing asthma. This hypothesis is supported by the observed difference in our data revealing lower rates of allergic rhinitis and environmental allergen sensitization in this group.

Another possibility is that late-onset AD represents a different clinical pattern, one more closely linked with inhalant allergen sensitization and respiratory atopy, without the same food-sensitization profile we find in early onset disease. Although many children with early AD improve by school age, natural history studies demonstrate that a large proportion still go on to develop asthma or rhinoconjunctivitis later in childhood.³⁰ The phenotype data also reveal that late-onset AD is more strongly associated with allergic rhinitis and that any form of AD confers a higher asthma risk than having no AD at all.³¹

Adding to this, Taki et al³² found that persistent and late-onset AD are the phenotypes most strongly linked to respiratory allergic outcomes: In these groups, rates of asthma were higher, as well as allergic rhinitis, elevated total IgE level, greater aeroallergen sensitization, and higher exhaled nitric oxide, suggesting a more pronounced T_H2-airway inflammatory profile, even though their immune cell cytokine responses looked similar to other AD phenotypes.

The combination of these results suggests that the later the onset of AD or the longer its duration, the more the pathway may be driven by inhalant allergens, which would explain the higher rates of asthma and allergic rhinitis in children with late-onset AD found in our cohort. Further research is needed to explore this potential protective effect in more detail. Moreover, at this stage, this remains a hypothesis that requires further investigation through well-designed studies.

Race and household income were found to influence the onset and severity of AD in children. In our study, 81% of non-Hispanic White children developed AD before their first birthday, which was higher than the 67.1% rate found in non-Hispanic Black children. However, previous studies have indicated that non-Hispanic Black children are more likely to develop AD in early childhood compared with non-Hispanic White children.³³ Our observed lower reported rate of early onset AD among non-Hispanic Black children may reflect diagnostic challenges rather than true differences in prevalence.^34,35 AD in skin color can present features that vary from traditional descriptions. These variations can complicate diagnosis and reduce the accuracy of standard assessment tools, particularly those relying on erythema. This underscores the need for greater awareness and improved diagnostic approaches tailored to diverse skin types.

Our study may be subject to recall bias because information on AD and FA was obtained from parent-reported surveys. This approach was necessary to capture detailed onset timing and allergen profiles, which are often incomplete in medical records. Furthermore, asthma and allergic rhinitis diagnoses were extracted from electronic health record data, in which these conditions are reliably documented, ensuring accuracy for comorbidities. Including both sources ensured comprehensive data capture and consistency across sites. Moreover, incomplete surveys (11%) were excluded from the analysis to maintain data integrity.

A key strength of our study is that all participants had allergist-confirmed FAs, enabling us to specifically assess the impact of AD on FA-related factors. In addition, we evaluated phenotypic differences between children with and without AD and the influence of AD on the type and number of implicated food allergens.

This study highlights the importance of identifying early onset AD as a potential risk factor for multiple FA and sensitivities to specific foods. Recognizing this link not only helps inform and guide parents but may also serve as a valuable tool as our understanding of FA prevention strategies evolves.

Future research should explore whether the severity and early treatment of AD influence the progression of FA in early childhood. Investigating whether timely and effective management of AD can reduce the risk of developing multiple FAs is a critical next step.

Supplementary Material

Supplementary data related to this article can be found at doi.org/10.1016/j.anai.2025.12.001.