Adherence to Mediterranean Diet During Pregnancy, Breastfeeding, and Development of Food Allergy in the Offspring: Results From the MEDALLION Cohort Study

ABSTRACT

Background

Prevention is a key strategy for reducing the burden of food allergies (FA). The maternal diet during pregnancy and breastfeeding is increasingly recognized as a factor contributing to FA risk.

Methods

We analyzed data from a sub‐cohort of the MEDALLION study to evaluate the association between maternal adherence to the Mediterranean Diet (MedDiet) and the development of FA in offsprings. Maternal dietary intake during pregnancy and lactation was assessed using a validated scale, i.e., the MedDiet Score, along with individual food group consumption. Multiple logistic regression models were performed.

Results

A total of 430 mother‐offspring dyads were included in the analysis. Higher maternal adherence to the MedDiet was associated with reduced odds of FA in offspring during both pregnancy (adjusted OR 0.94, 95% CI 0.89–1.00) and lactation (adjusted OR 0.94, 95% CI 0.88–1.00). Increased weekly consumption of fruits and full‐fat dairy products during pregnancy, and vegetables during breastfeeding, was associated with reduced likelihood of FA in the offspring. Conversely, higher intake of poultry and red meat (> 3 servings/week) during both periods, and fish consumption during pregnancy (> 1 serving/week), were associated with increased odds of FA.

Conclusions

This study points out that the MedDiet and, especially, some specific components, such as fruits, vegetables, and full‐fat dairy products, may contribute to lowering the risk of FA in infants, whereas a higher intake of poultry, red meat, and fish may increase it. These findings support the potential of maternal nutrition to reduce childhood FA.

This study analyzed data from a sub‐cohort of the MEDALLION study to examine the association between maternal adherence to the Mediterranean diet and the development of food allergy in offspring. Maternal dietary intake during pregnancy and lactation was assessed using the validated MedDiet Score, together with individual food group consumption; multiple logistic regression models were applied. Integrating Mediterranean diet principles into guidelines, with an emphasis on increased consumption of fruits, full‐fat dairy, and vegetables, may help lower the risk of food allergy. FA, food allergy; MEDALLION, Mediterranean Allergy Prevention Study; OR, odds ratio.

Abbreviations

FA

food allergy

MEDALLION

Mediterranean Allergy Prevention Study

OR

odds ratio

1. Introduction

Over the last decades, food allergies (FA) have become an exponentially increasing public health concern, with prevention recognized as a promising strategy to reduce their burden. FA are referred to as the “second wave” of the allergy epidemic –following asthma, which constituted the “first wave” [1]. The most common form of FA is IgE‐mediated [2], and its prevalence in Western countries ranges from 8% to 10%, primarily affecting infants and young children.

In contrast, FA prevalence in developing countries shows marked variability: urbanized areas such as certain regions in China display rates comparable to those in the West, while rural parts of Asia and Africa tend to report significantly lower prevalence [3]. These differences suggest that FA prevalence may be influenced by factors such as age, diet, and environment, highlighting complex interactions between the genome, microbiome, and external exposures [4, 5]. Although approximately half of individuals with atopy report a family history of allergic disease, environmental factors appear to play a more substantial role than genetics in the development of FA [6].

The role of environmental modulators has also received increasing attention [7]. Among these factors, exposure to tobacco smoke, whether active or passive, has been consistently linked to an increased risk of allergic conditions in children [8, 9, 10, 11, 12], thereby heightening the risk of atopic dermatitis and FA in infancy [8]. Similarly, alterations in neonatal gut microbiota (e.g., due to cesarean delivery), which can lead to reduced microbial diversity and colonization by beneficial bacteria [13], might impair immune tolerance and increase the risk of allergic diseases, including FA [14, 15, 16, 17]. Furthermore, current recommendations endorse the early introduction of allergenic foods to promote immune tolerance and prevent FA [18]. However, the impact of early‐life exposures remains controversial in many cases [17, 18, 19, 20, 21].

Another area of growing interest is the role of maternal diet during pregnancy and lactation, given its influence on fetal immune system development and programming [22, 23]. In this context, traditional dietary patterns, especially the Mediterranean diet (MedDiet), have been proposed to exert immunoprotective effects. The MedDiet is rich in monounsaturated fats, omega‐3 polyunsaturated fatty acids (PUFAs), fruits, vegetables, legumes, and whole grains, which support the establishment of a healthy gut microbiome and immune tolerance. In contrast, the Western diet, characterized by high intake of saturated fats, refined sugars, and processed foods, promotes systemic inflammation, disrupts gut barrier function, and reduces intake of nutrients essential for immune regulation [24].

Despite the potential benefits of maternal diet, many women avoid common allergenic foods during pregnancy or breastfeeding, particularly if their infants exhibit signs of eczema or gastrointestinal discomfort. However, this approach may be counterproductive. Dietary proteins are present in breast milk in very small concentrations, up to 1000 times lower than endogenous human milk proteins, yet they may play a beneficial role by promoting oral tolerance to allergens [25, 26]. Allergic reactions to endogenous breast milk proteins are extremely rare and support the concept that gradual low‐dose exposure, as seen in human milk, could help prevent FA [25].

Based on this evidence, our large retrospective MEDiterranean ALLergy PreventION Study (MEDALLION) cohort study aimed to evaluate the potential effect of the MedDiet and investigate how consumption of different food groups during pregnancy or breastfeeding may influence the development of FA in early life.

2. Methods

The MEDALLION study is a large cohort study in Greece that followed 2306 mothers from six different regions (Athens, Thessaloniki, Ioannina, Alexandroupoli, Peloponnese, Crete). This study aimed to investigate the role of early life exposures, i.e., the period from pregnancy up to 3 years of age [27], and the risk of developing FA during infancy. For the MEDALLION study, pediatric allergy specialists from these regions were asked to refer infants diagnosed with FA during the first 6 months of life to the first author (E.V.). The diagnosis of FA was established in the following cases:

For IgE‐mediated responses, diagnosis was based on:

1. A combination of a clinical history suggestive of FA and positive serum food allergen‐specific IgE (sIgE) and skin prick tests (SPTs), or

2. A positive oral food challenge in cases where diagnostic certainty could not be achieved with (a) [28].

For non‐IgE‐mediated responses, diagnosis was determined through food elimination followed by reintroduction of the suspected allergen [29].

Immediately upon diagnosis, potentially eligible subjects were invited to participate in an in‐person meeting to explain in detail the study procedures, obtain the written informed consent, and perform an interview.

Healthy controls matched by age, sex, and living area were also enrolled. These subjects were referred by primary care pediatricians and followed the same procedures as the cases.

Exclusion criteria for both cases and controls were mothers with pre‐existing chronic diseases affecting diet or breastfeeding that were not accounted for in the study, such as diabetes mellitus, and the ongoing participation in other dietary intervention studies. We aimed to include approximately two healthy infants without a diagnosis of FA as a control for every case of FA.

A total of 791 mothers of infants with physician‐diagnosed FA diagnosis and 1515 mothers of healthy controls were enrolled in the MEDALLION study. During the interview conducted immediately after the enrollment, the following information was collected: demographic data such as birthplace, nationality, and parental education; parental history of allergies; and any atopic history in older siblings, as well as neonatal data like gestational age, birth weight, mode of delivery, and maternal antibiotic use during pregnancy [27]. Additional data gathered were parental smoking habits, maternal diet during pregnancy and breastfeeding (collected through MedDiet Score questionnaire), any dietary restriction during the same periods, complementary feeding strategies, family history of allergies, the presence of pets in the household, and antibiotic use during the first 3 years of life (the full study's questionnaire is provided in Table S6).

The study was approved by the Ethics Committee of the 6th Health Region of Peloponnese, Ionian Islands, Epirus and Western Greece, University General Hospital Panayia Voithia (Ref No. 9220/6.4.22), the “Panagiotis and Aglaia Kyriakou” Children's Hospital, School of Medicine, University of Athens (Ref No. 10156/080622), and written informed consent was obtained from each participating mother, adhering to the principles of the Declaration of Helsinki.

2.1. Participants in the Sub‐Cohort Study

From the original MEDALLION study (n = 2306), a total of 521 mother–child pairs met the inclusion criteria for this sub‐cohort: (i) exclusive breastfeeding for ≥ 2 months; (ii) completion of the MedDiet Score questionnaire at enrollment in the MEDALLION study, and (iii) origin from three regions in Greece (Athens, Thessaloniki, and Ioannina). The exclusion criteria for our sub‐cohort were: (i) absence of breastfeeding, (ii) incomplete MedDiet Score questionnaire responses at enrollment. When children reached 24–36 months, eligible mothers were re‐contacted and invited to complete a follow‐up interview including a second MedDiet questionnaire (focused on the breastfeeding period) and additional questions on feeding practices, allergic conditions, antibiotic exposure, and parental smoking. Of the 521 invited, 469 (90%) participated, forming the final sub‐cohort, of whom 430 provided complete responses.

2.2. MedDiet Score Questionnaire

Maternal adherence to the MedDiet was assessed using a modified MedDiet Score questionnaire (MedDiet Score), based on the scoring system by Panagiotakos et al. [30] and modified by Demetriou et al. [31]. This score evaluates the weekly consumption of 11 food groups: unrefined cereals, potatoes, fruits, vegetables, legumes, fish, red meat products, poultry, full‐fat dairy products, and olive oil usage in cooking, along with alcoholic beverages. Each food group is scored from 0 to 5. The resulting total score ranges from 0 to 55, with calculated tertiles indicating low (0–13), moderate (14–27), good (28–41), and high (42–55) adherence to the MedDiet.

2.3. Statistical Analysis

For the analyses of this study, only participants without missing data on MedDiet were considered. Descriptive statistics included frequencies and percentages for categorical variables and means and standard deviations or medians and interquartile ranges (IQR) for continuous variables, depending on the normality of the data as assessed graphically. Initially, the similarities in MedDiet patterns (considering both MedDiet Score and food groups) between two groups of mothers—those with infants diagnosed with IgE‐mediated FA and those with infants diagnosed with non‐IgE‐mediated FA—were examined using t‐tests or Wilcoxon test as appropriate. As no significant differences were found between the two groups, they were combined into a single group of mothers with infants diagnosed with FA (Table S1).

Single food groups were examined through an ordinal scale (0–5) according to the portions/week (Tables S2 and S3). Then, the weekly consumption levels were categorized into high and low based on cumulative percentages that best distinguished infants with FA from healthy infants. For most food groups, these cut‐offs aligned with the Mediterranean dietary pattern as depicted in the pyramid by Panagiotakos et al. [32].

Univariate analyses were conducted to evaluate: (i) the effect of the MedDiet Score on FA using the t‐test, (ii) the effect of each individual food group on FA between the two groups using the Mann–Whitney U test, and (iii) the identification of confounders. A variable was considered a confounder if it was significantly associated with both the outcome (FA) and the MedDiet. Potential confounders, including maternal and paternal history of allergies, smoking during pregnancy and breastfeeding, antibiotic use during breastfeeding, duration of breastfeeding, and food exclusion, were analyzed in relation to FA using the Chi‐square test and in relation to the MedDiet Score using the t‐test.

Multivariate analyses were conducted to assess the association between the MedDiet (both the overall score and individual food groups) and FA, adjusting for identified confounders. This involved two distinct logistic regression models, each using the binary classification of FA as the outcome variable. The independent variables were: (i) the continuous MedDiet Score and (ii) the individual food groups that showed a significant association with FA in the univariate analysis. Both models were adjusted for identified confounders and were conducted separately for variables related to pregnancy and breastfeeding. Additionally, further analysis was performed using the binary classification of weekly food group consumption to evaluate the effectiveness of the cutoffs.

3. Results

3.1. Descriptive Characteristics

A total of 430 mothers with complete data were included in the final analysis (Table 1). This sub‐cohort of the MEDALLION study comprised 94 mothers of healthy infants and 336 mothers of infants diagnosed with FA. Among the FA cases, 159 were IgE‐mediated and 177 were non‐IgE‐mediated. The non‐IgE‐mediated cases included 121 infants with allergic proctocolitis, 13 with gastroesophageal reflux, 15 with eosinophilic gastritis, 34 with food protein‐induced enterocolitis syndrome (FPIES), and 5 with eosinophilic esophagitis (EoE). Some infants were diagnosed with more than one non‐IgE‐mediated condition. Regarding the mode of birth, 170 (40%) were delivered via normal birth and 260 (60%) via cesarean section.

TABLE 1.

Characteristics of study participants.

	N (%)
Sex
Male	240 (56)
Female	190 (44)
Age at the first diagnosis, months (median [IQR])	5 [1–11]
Mode of birth
Normal	170 (40)
Caesarian	260 (60)
Wheezing	17 (4)
Atopic dermatitis	125 (31)
Anaphylaxis emergency kit	82 (19)
Duration of breastfeeding, months (median [IQR])	6 [1–12]
Type of formula
Cow's milk based	97 (23)
Partially hydrolyzed	81 (19)
Extensively hydrolyzed	61 (14)
Amino acid	24 (5)
Other/missing or don't remember	167 (39)
Food exclusion during pregnancy	55 (13)
Maternal smoking during pregnancy	56 (14)
Antibiotics during pregnancy	91 (21)
Food exclusion during breastfeeding	68 (16)
Maternal smoking during breastfeeding	66 (15)
Maternal FA	59 (14)
Maternal atopic history	129 (30)
Paternal FA	49 (11)
Paternal atopic history	123 (29)

Note: Data are presented as frequencies and percentages, unless otherwise indicated.

Abbreviations: FA, food allergy; IQR, interquartile range.

Wheezing was reported in 17 (4%) infants, while 125 (31%) had a history of atopic dermatitis. Anaphylaxis emergency kits were prescribed to 82 (19%) of the infants.

During pregnancy, 56 mothers (14%) smoked, 91 mothers (21%) used antibiotics, and 55 mothers (13%) reported excluding certain foods from their diet, primarily sugar (N = 16), coffee (N = 18), and milk (N = 9). The most commonly reported reason for exclusion was experiencing combined discomfort upon consumption (N = 20).

The median duration of breastfeeding was 6 months (IQR = 2–12). Regarding formula feeding, 97 infants received cow's milk‐based formula, 81 partially hydrolyzed formula, 61 extensively hydrolyzed formula, and 24 were given amino acid‐based formula. During breastfeeding, 66 mothers (15%) reported smoking, and 68 (16%) mothers reported excluding certain foods from their diet, with the most avoided items being milk (N = 42) and nuts (N = 16). The main reasons for exclusion were maternal discomfort upon consumption (N = 12), fear that the baby would develop colic (N = 14), and observed disturbances in the infant (N = 17).

A total of 129 mothers (30%) had a history of atopy, including 59 (14%) with FA. A history of atopy was present in 123 fathers (29%), with 49 (11%) reporting FA.

3.2. Mediterranean Diet

Adherence to the MedDiet was high during both pregnancy (32.4 ± 4.4) and breastfeeding (33.0 ± 4.3) in the participants of this sub‐cohort. Mothers of healthy infants had significantly higher MedDiet Scores compared to mothers of infants with FA during pregnancy (p = 0.02) and breastfeeding (p = 0.02), as shown in Tables 2 and 3.

TABLE 2.

Univariate analysis of the Mediterranean diet during pregnancy (using MedDiet Score and food groups) and infant food allergy (FA). Data are presented as mean ± SD for continuous variables (MedDiet Score) and median with IQR for ordinal data (food groups). Differences between means and medians were assessed using the t‐test and Mann–Whitney U test, respectively.

	Total sample	Control	FA	p
MedDiet score	32.4 ± 4.4	33.3 ± 4.3	32.2 ± 4.4	0.02
Unrefined cereals	2 [1–2]	2 [1–2.75]	2 [1–2]	0.26
Potatoes	2 [2–2]	2 [1–2]	2 [2–2]	0.19
Fruits	2 [2–3]	3 [2–4]	2 [1–3]	< 0.001
Vegetables	2 [1–3]	2 [1–3]	2 [1–2]	0.07
Legumes	2 [1–2]	2 [1–2]	2 [1.75–2]	0.42
Fish	2 [1–2]	1 [1–2]	2 [1–2]	< 0.001
Red meat and products	4 [3–4]	4 [4–5]	4 [3–4]	< 0.001
Poultry	4 [3–5]	5 [5–5]	4 [3–5]	< 0.001
Full fat dairy	4 [3–5]	3 [2–4]	4 [4–5]	< 0.001
Olive oil	5 [5–5]	5 [5–5]	5 [5–5]	0.003

TABLE 3.

Univariate analysis of the Mediterranean diet during breastfeeding (using MedDiet Score and food groups) and infant food allergy (FA). Data are presented as mean ± SD for continuous variables (MedDiet Score) and median with IQR for ordinal data (food groups). Differences between means and medians were assessed using the t‐test and Mann–Whitney U test, respectively.

	Total sample	Control	FA	p
MedDiet score	33.0 ± 4.3	33.9 ± 4.6	32.7 ± 4.1	0.02
Unrefined cereals	1 [1–2]	2 [0–2]	1 [1–2]	0.95
Potatoes	2 [1–2]	2 [1–3]	2 [1–2]	0.23
Fruits	2 [2–3]	3 [2–3]	2 [2–3]	0.20
Vegetables	2 [1–2]	3 [1.25–3]	2 [2–3]	0.002
Legumes	2 [1–2]	2 [1–2]	2 [1–2]	0.34
Fish	2 [1–2]	2 [1–2]	2 [1–2]	0.08
Red meat and products	4 [3–4]	4 [4–5]	4 [3–4]	< 0.001
Poultry	5 [4–5]	5 [5–5]	4 [3.75–5]	< 0.001
Full fat dairy	4 [3–5]	4 [3–4.75]	4 [3–5]	0.20
Olive oil	5 [5–5]	5 [5–5]	5 [5–5]	0.11

Univariate logistic regression analysis revealed that higher adherence to the MedDiet by mothers during pregnancy (OR 0.94, 95% CI 0.89–0.99) and breastfeeding (OR 0.93, 95% CI 0.87–0.98) was significantly associated with a lower likelihood of the child developing FA (Tables 4 and 5).

TABLE 4.

Univariate analysis for the identification of confounders. Data are presented as absolute frequencies and percentages to assess associations between potential confounders and infant food allergy (FA, Chi‐square test), and as mean ± standard deviation for MedDiet score (Student's t‐test).

	Total	Control	FA	p	MedDiet pregnancy	MedDiet breastfeeding
	N = 430	N = 94	N = 336		Mean ± SD	Mean ± SD
Pregnancy information
Food exclusion during pregnancy
Yes	55 (13)	2 (2)	53 (16)	0.001	30.2 ± 4.6	—
No	355 (87)	92 (98)	283 (84)		32.7 ± 4.3
					< 0.001
Maternal smoking during pregnancy
Yes	56 (14)	8 (8)	48 (15)	0.13	34.3 ± 4.3	—
No	355 (86)	86 (92)	269 (85)		32.1 ± 4.3
					< 0.001
Antibiotics during pregnancy
Yes	91 (21)	13 (14)	78 (23)	0.06	32.1 ± 4.9	—
No	339 (79)	81 (86)	258 (77)		32.5 ± 4.2
					0.53
Breastfeeding information
Food exclusion during breastfeeding
Yes	68 (16)	2 (2)	66 (20)	< 0.001	—	31.5 ± 4.2
No	362 (84)	92 (98)	270 (80)			33.2 ± 5.0
						0.005
Maternal smoking during breastfeeding
Yes	66 (15)	11 (12)	55 (16)	0.19	—	34.5 ± 3.5
No	364 (85)	83 (88)	281 (84)			32.7 ± 4.3
						< 0.001
Family characteristics
Maternal FA
Yes	59 (14)	15 (16)	44 (13)	0.61	33.4 ± 4.8	33.6 ± 4.6
No	371 (86)	79 (84)	292 (87)		32.2 ± 4.3	32.9 ± 4.2
					0.08	0.23
Maternal atopic history
Yes	129 (30)	16 (17)	113 (34)	0.002	33.1 ± 4.4	33.0 ± 4.3
No	301 (70)	78 (83)	223 (66)		32.1 ± 4.4	32.9 ± 4.3
					0.03	0.81
Paternal FA
Yes	49 (11)	15 (16)	34 (10)	0.18	34.4 ± 3.0	34.3 ± 3.2
No	381 (89)	79 (84)	302 (90)		31.1 ± 4.9	32.8 ± 4.4
					< 0.001	0.003
Paternal atopic history
Yes	123 (29)	15 (16)	108 (32)	0.003	33.2 ± 3.6	33.4 ± 4.5
No	307 (71)	79 (84)	228 (68)		32.1 ± 4.6	32.8 ± 4.6
					0.005	0.14

TABLE 5.

Multivariate analysis of the Mediterranean diet during pregnancy (using MedDiet Score and food groups with ordinal scale) and infant food allergy (FA). Model 1: Logistic regression of MedDiet score on FA. Model 2: Logistic regression of food groups on FA.

	Unadjusted	Adjusted a
	OR (95% CI)	OR (95% CI)
Model 1
MedDiet score	0.94 (0.89, 0.99)	0.94 (0.89, 1.00)
Model 2
Fruits	0.42 (0.30, 0.57)	0.42 (0.30, 0.58)
Fish	1.61 (1.09, 2.39)	1.57 (1.04, 2.39)
Red meat products	0.73 (0.56, 0.93)	0.67 (0.51, 0.87)
Poultry	0.67 (0.50, 0.89)	0.69 (0.51, 0.92)
Full fat dairy products	2.38 (1.84, 3.17)	2.44 (1.85, 3.29)
Olive oil	0.43 (0.03, 1.38)	0.54 (0.04, 1.73)

^a Adjusted for maternal food exclusion during pregnancy, and maternal and paternal atopy.

3.3. Food Group Consumption During Pregnancy and Breastfeeding

Univariate analyses revealed significant differences in maternal food consumption during pregnancy between mothers of healthy infants and those of infants with FA in the following components: fruits (p < 0.001), fish (p < 0.001), red meat products (p < 0.001), poultry (p < 0.001), full‐fat dairy products (p < 0.001), and olive oil (p = 0.003). During pregnancy, mothers of healthy infants had higher consumption of fruits, vegetables, and olive oil, while they consumed less fish, red meat, and poultry (Table 2). During breastfeeding, significant differences were observed between the two groups for vegetables (p = 0.002), red meat products (p < 0.001), and poultry (p < 0.001). Mothers of healthy infants consumed more vegetables, while those of infants with FA had higher consumption of red meat products and poultry (Table 3).

3.4. Multivariate Analysis

The following confounders were identified and included in the multiple adjusted analyses: maternal and paternal history of atopy and food exclusion during pregnancy and breastfeeding. These variables were significantly associated with both FA and the MedDiet Score (Table 4). The adjusted regression models indicated that higher adherence to the MedDiet by mothers during pregnancy remained marginally inversely associated with the likelihood of their child developing FA (OR 0.94, 95% CI 0.89–1.00). A similar association was observed during breastfeeding (OR 0.94, 95% CI 0.88–1.00).

The multiple logistic regression analyses on individual food groups confirmed that a high consumption of fruits, full‐fat dairy products, and a low consumption of red meat, fish, and poultry during pregnancy was a protective factor against FA in offspring (Table 5). Additionally, higher vegetable intake and lower consumption of red meat and poultry during breastfeeding also reduced the odds of FA (Table 6).

TABLE 6.

Multivariate analysis of the Mediterranean diet during breastfeeding (using MedDiet Score and food groups with ordinal scale) and infant food allergy (FA). Model 1: Logistic regression of MedDiet score on FA. Model 2: Logistic regression of food groups on FA.

	Unadjusted	Adjusted a
	OR (95% CI)	OR (95% CI)
Model 1
MedDiet score	0.93 (0.87, 0.98)	0.94 (0.88, 1.00)
Model 2
Vegetables	0.62 (0.46, 0.83)	0.67 (0.50, 0.90)
Red meat products	0.65 (0.49, 0.84)	0.60 (0.44, 0.80)
Poultry	0.73 (0.55, 0.94)	0.75 (0.57, 0.98)

^a Adjusted for maternal food exclusion during breastfeeding, and maternal and paternal atopy.

When food consumption was categorized into binary variables, the results were consistent (Tables S4 and S5). In particular, higher maternal consumption of full‐fat dairy products (> 15 servings/week), fruits (> 8 servings/week) during pregnancy, and vegetables (> 12 servings/week) during breastfeeding was associated with reduced odds of FA in infants. Conversely, intake of > 3 servings/week of poultry and red meat during both periods was associated with a higher prevalence of FA. The consumption of fish (> 1 serving/week) during pregnancy was linked to a potential increase in the odds of FA in offspring.

4. Discussion

The results of this study showed that maternal adherence to the MedDiet during pregnancy and breastfeeding was associated with a reduced likelihood of infant FA. In addition to overall adherence to the MedDiet, the analysis of individual food groups, such as full‐fat dairy products, fruits, vegetables, poultry, red meat, and fish, revealed both protective and risk‐enhancing associations. The discussion will focus on these associations.

The potential of maternal MedDiet on FA in offspring is debated. A systematic review published in 2020 found no association between the MedDiet Score and the prognosis of FA in offspring [30]. However, a cohort study by Venter et al. reported a strong association between a 1‐unit increase in the MedDiet adherence and a decrease in the odds of developing atopic dermatitis or any other allergic disease, except for wheezing [33]. Our data provide additional evidence on the positive effects of maternal MedDiet on FA in offspring.

The MedDiet pattern during pregnancy is considered to reduce the risk of atopy in infants, due to the high content of biodiverse and low‐processed foods [34, 35]. Nonetheless, not all studies support a protective role: for example, in the Avon Longitudinal Study of Parents and Children, no associations were found between adherence to the MedDiet and the development of atopy in infants [36]. Future studies should investigate the underlying mechanisms beyond the role of maternal MedDiet both during pregnancy and lactation and the development of FA in the first years of a child's life.

Our findings showed an inverse association between maternal intake of full‐fat dairy products and FA in offspring, in line with previous studies [37]. Importantly, we also observed no increase in FA risk related to maternal milk elimination during breastfeeding, contrary to findings by Järvinen et al. [38], which suggested that avoidance of cow's milk could increase allergy risk. Our data indicate that mothers of allergic infants were more likely to eliminate foods, especially dairy, likely in response to early symptoms such as eczema or gastrointestinal disturbances, consistent with prior observations [25]. However, dietary proteins in human milk may help promote tolerance [39, 40]. The results of our study suggest that emphasizing specific food groups in the MedDiet may enhance its protective effects against allergies. Increasing maternal weekly consumption of fruits and full‐fat dairy products during pregnancy was associated with a reduced prevalence of FA in offspring. Similarly, higher dietary intake of vegetables during lactation was linked to a lower risk of FA in offspring. In a prior study conducted by our team, higher maternal fruit consumption was associated with a decreased risk of allergic proctocolitis. Castro‐Rodriguez et al. [22] found that low maternal fruit consumption negatively impacts the development of atopy in the offspring [41]. A study by Venter et al. found that vegetables in pregnancy were associated with FA prevention in infants, whereas 100% pure fruit juices had the opposite effect [42]. Specifying fruit and vegetable portions has also been underscored due to inconsistencies and contradictions in previous observational studies regarding their impact on FA prevention [23]. A study by Kim et al. observed that during pregnancy, a diet rich in baked and sugary products, along with a prolonged lactation period, may contribute positively to the prevalence of FA in the offspring [43]. A recent study revealed that higher maternal consumption of white bread and thiamine‐fortified bread flour during pregnancy increased the risk for FA and eczema development in infants [44].

Our results also suggest a connection between higher maternal intake of red meat and poultry during pregnancy and breastfeeding, with an increased prevalence of FA in infants, while a rather protective role against FA was observed when red meat products and poultry were consumed sparingly. These results are consistent with previous reports on the link between high red meat consumption [22, 23, 42, 45, 46] and poultry consumption [47] with increased odds of offspring allergy.

Our sub‐cohort analysis revealed an increased risk of FA associated with maternal fish intake of more than twice per week. These findings are in line with a previous study that found an inverse association between shellfish consumption during pregnancy and risks of wheezing and eczema in the offspring [48] On the other hand, these findings contrast with previous studies suggesting no role [49] or even a protective effect of increased fish consumption against FA and other allergic diseases [50]. These conflicting findings may, in part, be explained by the dual effect of fish consumption observed in other studies, which have shown that harmful substances (e.g., chemical pollutants) present in fish may exert effects that counteract the potentially protective properties of the fish's nutritional components against various diseases [51].

5. Limitations

This study has several limitations that should be considered when interpreting the findings. Firstly, the retrospective design and reliance on maternal recall, particularly regarding diet during pregnancy and breastfeeding, introduce potential recall bias. Data were collected at enrollment (within the first 6 months of life) and again via phone follow‐up when children were 24–36 months old. The time gap may affect recall accuracy and exposure assessment. Differential recall between cases and controls may also contribute to this potential bias. Although we used the validated MedDiet Score questionnaire to mitigate this, some misreporting cannot be ruled out. Second, the unbalanced ratio of FA cases to controls may reflect differential follow‐up or participation rates. Third, although FA diagnoses were established by physicians, the observational nature of the study precludes causal inference. We adjusted for several known confounders, including parental atopy and maternal food avoidance behaviors, but residual confounding remains possible. Fourth, the high baseline consumption of certain MedDiet components (e.g., olive oil) in this Greek population may have influenced the overall MedDiet Score, limiting generalizability into non‐Mediterranean settings.

Finally, differences in study design may help explain inconsistencies between our findings and those reported in prospective cohort studies [42, 52]. Prospective designs offer the advantage of real‐time dietary reporting, reducing recall bias and improving temporal accuracy. Nonetheless, our results provide valuable insights into maternal dietary patterns and their potential influence on FA development within a Mediterranean context.

6. Conclusion

This study indicates that greater maternal adherence to the MedDiet, particularly with high intake of full‐fat dairy products, fruits, and vegetables, was associated with reduced odds of FA in infants. Conversely, higher consumption of poultry, red meat, and fish during pregnancy and breastfeeding may increase this risk. These findings from the MEDALLION sub‐cohort underscore the need for future dietary intervention studies to explore how promoting traditional dietary patterns may contribute to FA prevention in early life.

Author Contributions

E.V., G.P.M., and C.V.: conceptualization; EV: data interpretation, data analysis, supervision, writing first draft; E.V., C.K., A.C., and C.A.: statistical analysis, data interpretation, critical review and approved manuscript; M.‐K.M.: data collection and writing first draft; N.G.P., M.P., M.L., S.T., and G.P.M.: data collection and data interpretation. All authors critically reviewed the manuscript.

Conflicts of Interest

C.V. has received research funding from Reckitt and has provided consultancy or lectures for Nestle Nutrition Institute, Danone, Reckitt, Abbott, Else Nutrition, Ausnutria, and HAL Allergy. G.P.M. received a grant from Angelini S.P.A. and acted as advisory for scientific projects. Furthermore, he received a grant from Reckitt Benckiser Healthcare S.P.A. All other authors have no conflicts of interest to declare.

Supporting information

Appendix S1: all70054‐sup‐0001‐AppendixS1.docx.

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.