Maternal dietary inflammatory index during pregnancy and the risk of offspring allergic disease

Abstract

Background

Maternal diet during pregnancy is considered a potential modifiable risk factor for allergic diseases in offspring. The dietary inflammatory index (DII) is a tool to assess the inflammatory potential of the diet and has been suggested to be associated with offspring allergy development. Its association with food allergy and immunoglobulin E (IgE) sensitization in children remains understudied.

Methods

This study analyzed 4709 mother‐partner‐child triads from the NorthPop Birth Cohort in Sweden. Maternal DII scores were calculated from a food frequency questionnaire administered at gestational week 34. Allergy outcomes at 18 months included parent‐reported physician‐diagnosed food allergy, parent‐reported eczema and atopic eczema according to UK Working Party criteria, parent‐reported ever wheeze, parent‐reported physician‐diagnosed asthma, and IgE sensitization to food and airborne allergens. Associations between maternal DII scores (continuous and quartiles) and allergic outcomes were assessed using logistic regression, adjusting for maternal age, allergic heredity, farm living, region of birth, siblings, and education.

Results

At age 18 months, 4.9% of children had physician‐diagnosed food allergy, 30.6% had eczema, 11.4% had atopic eczema, 15.9% reported ever wheeze, 4.1% had physician‐diagnosed asthma, and 19% were IgE sensitized. No significant associations were found between maternal DII scores and the allergic outcomes of interest.

Conclusion

This large birth cohort study found no association between maternal DII during pregnancy and allergic diseases or IgE sensitization in 18‐month‐old children, suggesting that a proinflammatory diet during pregnancy does not influence early allergic outcomes. Further research is needed to clarify the role of maternal diet in offspring immune development.

Key message.

We found no association between maternal dietary inflammatory index during pregnancy and the risk of offspring allergic disease including IgE sensitization at 18 months in a large longitudinal birth cohort including 4709 mother‐infant‐partner triads.

1. INTRODUCTION

Over the last decades, there has been a worldwide increase in the prevalence of allergic diseases such as asthma and eczema, followed by rising rates of food allergy. ¹ The etiology is considered multifactorial, and early modifiable dietary habits have been of great interest. Maternal diet during pregnancy represents a potential window of opportunity to modulate the developing fetal immune system, potentially mediated by placental transfer of nutrients and immune factors, as well as by epigenetic mechanisms. ² Studies on specific nutrients such as vitamin D and Omega‐3 during pregnancy have shown a clearer association with offspring allergies than studies on dietary patterns. ³ , ⁴ Studying dietary patterns is a more complex task but often desirable since the diet encompasses more than just specific nutrients and foods. ⁴

In 2014, Shivappa et al. developed the dietary inflammatory index (DII), a literature‐derived population‐based index for assessing the inflammatory potential of an individual diet. ⁵ The DII is calculated based on 45 “food parameters” comprising whole foods, nutrients, and other bioactive compounds, all associated with inflammatory markers. A negative DII score represents an anti‐inflammatory diet. In contrast to commonly used dietary indices in research, such as the Healthy Eating Index and Mediterranean Diet Score, the DII is neither based on adherence to dietary recommendations nor derived directly from the present study population. Instead, the DII is based on study participant food intake in relation to a suggested “global mean” food intake. ⁶ The DII has been validated in both pregnant and non‐pregnant populations worldwide. ⁷ , ⁸ Maternal DII during pregnancy has been associated with different offspring outcomes including congenital heart defects, ⁹ emotional and behavioral symptoms ¹⁰ and bone health. ¹¹ Given that exposure to chronic low‐grade inflammation in utero may affect both airway development, lung function, and immune response ² , ⁴ , ¹² , ¹³ the DII is a possible tool for quantifying this exposure.

Studies examining the association between maternal DII during pregnancy and offspring allergy risk have mainly focused on respiratory outcomes, with conflicting results. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ However, a recent meta‐analysis reported no association with respiratory disease in childhood. ¹⁸ Notably, studies examining the association between maternal DII and offspring immunoglobulin E (IgE) sensitization, a key biomarker of allergic disease, ¹⁹ are lacking. ⁴ To our knowledge, only one single study has examined maternal DII during pregnancy and the offspring's risk of IgE‐mediated food allergy ²⁰ highlighting another major knowledge gap. Large prospective birth cohorts are particularly suited to clarify how maternal diet‐induced inflammation during pregnancy may influence the risk of IgE sensitization and allergic diseases in offspring.

The overall aim of this study was to examine the association between maternal DII and the risk of allergic outcomes in offspring. We hypothesized that a more proinflammatory diet during pregnancy, according to the DII, increases the child's risk of allergic outcomes, including both Immunoglobulin E (IgE) sensitization to common allergens and physician‐diagnosed food allergy in young children.

2. METHODS

2.1. Study population

We used data from the large longitudinal population‐based NorthPop Birth Cohort Study in Västerbotten County, Sweden. ²¹ Eligible families were recruited at the time of the routine ultrasound examination. The inclusion criteria for the NorthPop Birth Cohort Study were a viable pregnancy at gestational weeks 14–24, pregnant woman ≥18 years of age, comprehending the Swedish language, and intention to give birth and reside in the catchment area in the next few years.

Participants for the current study were recruited between May 2016 and April 2021. A sample size estimation (Table S1) was conducted in advance for the binary outcomes of interest based on preliminary prevalence estimates from the NorthPop Birth Cohort Study. Exclusion criteria for the current study were applied sequentially in the following order: intrauterine deaths and terminations, multiple pregnancy, preterm birth defined as gestational age <37 weeks, and implausible daily total energy intake in pregnancy (<500 and >3500 kcal/day). Figure 1 depicts the enrollment of the study population. The study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology – Nutritional Epidemiology (STROBE‐nut) guidelines. ²²

FIGURE 1.

Flowchart of the inclusion of the final study population from the NorthPop Birth Cohort Study.

2.2. Data Collection

Data on parental medical history, allergic heredity, country of birth, family situation, and maternal diet were collected using web‐based questionnaires (Appendix A) provided to the pregnant women and their partners at gestational ages 14–24, 26, and 34. Data on parental age at gestation, gestational age, infant sex, maternal body mass index (BMI), neonatal care, delivery mode, and parental educational level were obtained from the NorthPop Birth Cohort Study database, ²¹ Swedish Pregnancy Register, ²³ Swedish Neonatal Quality Register, ²⁴ and the Swedish Longitudinal Integrated Database for Health Insurance and Labour Market Studies. ²⁵

When the child was 4, 9, and 18 months old, the families responded to questions regarding breastfeeding, family situation, and maternal and child health, including early allergic manifestations (Appendix A). Eczema was parent‐reported and categorized as atopic eczema if UK Working Party criteria ²⁶ were fulfilled based on data from questionnaires administered between pregnancy and child age 18 months. A Patient‐Oriented Eczema Measure (POEM) score ²⁷ , ²⁸ was obtained from participants with parent‐reported eczema at 18 months of age. Parents also reported on physician diagnosis of food allergy and asthma at that age. Heredity for allergic disease was defined as parent‐reported eczema, food allergy, or asthma among first‐degree relatives.

All children were invited at 18 months of age for a blood sample to assess IgE sensitization to common food allergens (Food mix fx5; cow's milk, egg white, wheat, cod, peanut and soybean) and aeroallergens (Phadiatop) using ImmunoCAP, Thermo Fisher Scientific/Phadia, Uppsala, Sweden, according to the manufacturer's instructions. An IgE level ≥0.35 kU/l for Food mix fx5 and an IgE level ≥0.35 PAU/l for Phadiatop were considered positive.

2.3. Exposure

Dietary intake was assessed by a self‐administered food frequency questionnaire (FFQ) at around gestational week 34, as previously described. ²⁹ The FFQ included 125 food items and portion‐size estimations of vegetable, carbohydrate, and protein sources. Solid foods were reported in five frequencies: never or rarely, 1–3 times/week, 4–6 times/week, 1–2 times/day, or ≥3 times/day, except bread and sandwich toppings, for which frequency was reported as slices/day, together with information on how often different bread types were consumed, from 1 (most often consumed) to 6 (least often consumed). Beverages were reported in six frequencies: never or rarely, 1–3 glasses/week, 4–6 glasses/week, 1–2 glasses/day, 3–4 glasses/day, and ≥5 glasses/day. Intake frequencies were converted into daily intakes in quantity per day using reported portion sizes combined with standard portion sizes. ³⁰ Nutrient intake was calculated using the Swedish Food Composition Database provided by the Swedish Food Agency. ³¹ Dietary supplements were not included in the calculation of the DII.

The construction of the DII has been described in detail elsewhere. ⁵ In summary, it was developed through a comprehensive review of the literature, identifying 45 food parameters linked to inflammatory markers such as interleukin (IL)‐1β, IL‐4, IL‐6, IL‐10, Tumor Necrosis factor (TNF)‐α, and C‐reactive protein (CRP). Each parameter was assigned an inflammatory effect score based on the strength and quality of its association with these markers. Using dietary data from eleven populations, global means and standard deviations (SD) were calculated. A Z‐score was computed by subtracting the global mean from the reported intake and dividing by the global SD. To reduce skewness, Z‐scores were converted to centered percentiles. Finally, the pregnant woman's total DII score was calculated by multiplying the centered percentile of each available food parameter by its inflammatory effect score and summing these values.

In this study, 30 of the 45 food parameters were available from the FFQ to calculate the DII score for each study participant. The available 30 food parameters were total energy, total fat, saturated fat, trans fatty acids, cholesterol, monounsaturated fat, polyunsaturated fat, omega‐6 fatty acids, omega‐3 fatty acids, carbohydrates, protein, fiber, alcohol, caffeine, tea, folic acid, selenium, niacin, iron, zinc, thiamin, vitamin B2, vitamin B6, vitamin B12, vitamin A, vitamin E, vitamin C, vitamin D, magnesium, and β‐carotene. We lacked intake data on eugenol, garlic, ginger, onion, saffron, turmeric, flavan‐3‐ol, flavones, flavonols, flavanones, anthocyanidins, isoflavones, pepper, thyme/oregano, and rosemary, and these food parameters were omitted in the calculation of the total DII score.

2.4. Outcomes

The outcomes were the prevalence of IgE sensitization and the following five individual allergic disease outcomes at 18 months of age: parent‐reported physician‐diagnosed food allergy, parent‐reported eczema, atopic eczema according to UK Working Party criteria, ²⁶ parent‐reported ever wheeze, and parent‐reported physician‐diagnosed asthma. IgE sensitization at 18 months of age was defined as a positive Food mix fx5 and/or Phadiatop.

2.5. Statistical Analysis

The total study population characteristics are described with continuous data reported as means ± SD for symmetrically distributed data and medians with interquartile range (IQR) for skewed distributions. Categorical data are reported as frequencies with observed proportions.

We used logistic regression models to estimate the association between maternal DII score and the outcomes of interest as binary outcomes. Separate models were fitted for continuous DII score and DII scores categorized into quartiles, using first quartile (Q1) as reference and representing the participants with the lowest DII scores. The models were adjusted for potential confounders and reported as crude and adjusted odds ratios (OR and aOR).

We used Directed Acyclic Graphs (DAGs) ³² to select relevant confounders a priori, inspired by the DAG suggested for epidemiological studies in childhood food allergy by Peng et al. ³³ for each outcome. The adjusted models for all outcomes included the following covariates and factors: maternal age (years), farm living (yes/no), heredity for allergic disease (yes/no), maternal region of birth (Nordic country, European country outside the Nordic region, non‐European country), siblings/other children in the household (yes/no) and maternal education level (elementary or lower, high school, university) (Figures [Link], [Link], [Link], [Link]).

Missing values were handled by applying multiple imputation using the Multivariate Imputation by Chained Equations (MICE) using the mice package version 3.16.0 ³⁴ with 170 imputations and 30 iterations. The number of imputations needed was calculated using the method suggested by von Hippel. ³⁵ Multiple imputations were applied to both the outcome variables and the set of independent variables.

Statistical analysis was conducted using IBM SPSS Statistics, version 28.0 (IBM Corp, 2021, Armonk NY) and R 4.2.2 (R Core Team, 2022, Vienna, Austria). A p‐value <.05 was considered to indicate statistical significance.

2.6. Ethics

Ethical permission was granted by the regional Ethical Committee in Umeå, Sweden (2016/349‐31 and 2018/504‐32). All parents were informed both verbally and in writing, and informed consent was collected from both parents.

3. RESULTS

3.1. Study population

A total of 4709 mother‐partner‐child triads were included in the study, of whom 2429 (51.6%) of the children were boys. The majority of mothers were born in a Nordic country (92.0%) and had a university‐level education (69.7%). The mean DII score was −0.98 ± 1.84 with a range of −5.64 to 4.00. Detailed demographics for the study population characteristics are presented in Table 1.

TABLE 1.

Study population characteristics.

Characteristic	Study population	Missing, no. (%)
Participants, no. (%)	4709 (100)	0 (0)
Boys, no. (%)	2429 (51.6)	0 (0)
Gestational wk., median (IQR)	40 (2)	0 (0)
Caesarean section, no. (%)	738 (15.7)	0 (0)
Maternal BMI (kg/m2), median (IQR)	23.8 (5.3)	109 (2.3)
Energy intake (kcal/d), mean (±SD)	2131 (562)	340 (7.2)
Dietary inflammatory index, mean (±SD)	−0.98 (1.84)	340 (7.2)
Birthweight (g), median (IQR)	3570 (620)	0 (0)
Neonatal care a , no. (%)	420 (8.9)	0 (0)
Maternal age (y), median (IQR)	31 (6)	0 (0)
Mother's partner age (y), median (IQR)	32 (7)	0 (0)
Presence of other children in the household, no. (%)	2161 (50.7)	449 (9.5)
Farm living	88 (2.0)	282 (6.0)
Maternal country of birth, no. (%)		284 (6.0)
Nordic country	4073 (92.0)
European country outside the Nordic region	122 (2.8)
Non‐European country	230 (5.2)
Highest educational level of mother, no. (%)		11 (0.2)
University	3273 (69.7)
High school	1256 (26.7)
≤Elementary school	169 (3.6)
Highest educational level of mother's partner, no. (%)		115 (2.4)
University	2329 (50.7)
High school	2063 (44.9)
≤Elementary school	202 (4.4)
Heredity for allergic diseases b , no. (%)		1001 (21.3)
Family history of eczema	958 (25.8)
Family history of food allergy	1041 (28.1)
Family history of asthma	1271 (34.3)
Breastfeeding, no. (%)
Exclusively at 4 mo	2614 (68.6)	897 (19.0)
Partially at 4 mo	3198 (83.4)	897 (19.0)
Partially at 9 mo	1460 (40.7)	1122 (23.8)

Note: Valid percentages calculated from participants responding to each specific question. Missingness calculated based on the whole study population.

Abbreviations: IQR, interquartile range; Mo, month; No, number; SD, standard deviation; Wk, week.

^a Admitted to the neonatal unit.

^b Among first‐degree relatives.

Physician‐diagnosed food allergy was reported in 144 (4.9%) children (Table 2). Milk allergy, followed by egg allergy, was the most common specified food allergy (data not shown). At that age, eczema was parent‐reported among 1050 (30.6%) of the children, whereas 391 (11.4%) were categorized as having atopic eczema according to UK Working Party criteria. Among participants with parent‐reported eczema, the POEM scores were in general very low, with the majority of parents (55.1%) reporting clear or almost clear skin. The mean POEM scores were significantly higher (p < .001) among participants with atopic eczema according to UK Working Party criteria (5.0 ± SD 4.2) compared to participants with eczema only (1.8 ± SD 2.2). The prevalence of ever having wheezed was 15.9%, and the prevalence of asthma was 4.1% at 18 months of age. Data on IgE levels were available for 2436 (51.7%) children at 18 months, of whom 462 (19.0%) had a positive Food Mix fx5 and/or Phadiatop (Table 2).

TABLE 2.

Prevalence of allergic diseases and IgE sensitization at 4, 9 and 18 months of age in the NorthPop Birth Cohort Study.

Allergic diseases of child	Study population	Missing no. (%)
Parent‐reported eczema, no. (%)
At age 4 mo	501 (13.1)	891 (18.9)
At age 9 mo	856 (23.8)	1111 (23.8)
At age 18 mo	1050 (30.6)	1273 (27.0)
POEM‐group 18 mo		38 (3.6)
Clear/almost clear	558 (55.1)
Mild	340 (33.6)
Moderate	107 (10.6)
Severe	7 (0.7)
Very severe	0 (0)
Parent‐reported atopic eczema a
At age 9 mo	269 (7.5)	1114 (23.7)
At age 18 mo	391 (11.4)	1278 (27.1)
POEM‐group 18 mo		14 (3.6)
Clear/almost clear	121 (32.1)
Mild	159 (42.2)
Moderate	90 (23.9)
Severe	7 (1.9)
Very severe	0 (0)
Physician‐diagnosed FA, no. (%)
Any FA at age 9 mo	125 (3.5)	1118 (23.7)
Any FA at age 18 mo	144 (4.9)	1782 (37.8)
Parent‐reported ever wheeze
At age 9 mo	419 (11.6)	1111 (23.6)
At age 18 mo	547 (15.9)	1270 (27.0)
Physician‐diagnosed asthma, no. (%)
At age 9 mo	26 (0.7)	1111 (23.6)
At age 18 mo	141 (4.1)	1271 (27.0)
Serology at age 18 mo, no. (%)		2273 (48.3)
Phadiatop ≥0.35 PAU/l	111 (4.6)
Food Mix fx5 ≥0.35 kUA/l	433 (17.8)
Phadiatop and/or Food Mix fx5 positive	462 (19.0)

Note: Valid percentages calculated from participants responding to each specific question. Missingness calculated based on the whole study population. Food Mix fx5; Mix of egg white, milk, fish, wheat, peanut and soybean. Phadiatop; Mix of representative inhalant allergens.

Abbreviations: FA, food allergy; Mo, month; No, number; POEM, patient‐oriented eczema measure.

^a According to UK Working Party criteria. ²⁶

3.2. DII and allergy outcomes at 18 months of age

There were no associations between maternal DII score during pregnancy as a continuous variable and the six outcomes of interest at 18 months of age; parent‐reported physician‐diagnosed food allergy (aOR 1.053; 95% CI: 0.962–1.151), parent‐reported eczema (aOR 0.977; 95% CI: 0.938–1.018), atopic eczema according to UK Working Party criteria (aOR 1.009; 95% CI: 0.952–1.070), parent‐reported ever wheeze (aOR 1.037; 95% CI: 0.986–1.090), parent‐reported physician‐diagnosed asthma (aOR 1.051; 95% CI: 0.960–1.150) and IgE sensitization (aOR 0.992; 95% CI: 0.938–1.049). The results were similar in both unadjusted and adjusted analyses (Table 3). Further comparisons of maternal DII scores by quartiles showed no association with the outcomes of interest in adjusted and unadjusted analyses (Table 4).

TABLE 3.

Associations between one unit increase of Dietary Inflammatory Index score and offspring allergic outcomes at age 18 months in the NorthPop Birth Cohort Study.

Allergy outcome at 18 months	Unadjusted			Adjusted
	OR	95% CI	p‐Value	aOR	95% CI	p‐Value
Physician‐diagnosed FA	1.053	0.983–1.150	.256	1.053	0.962–1.151	.262
Parent‐reported eczema	0.982	0.944–1.022	.365	0.977	0.938–1.018	.268
Parent‐reported atopic eczema a	1.006	0.950–1.065	.841	1.009	0.952–1.070	.756
Parent‐reported ever wheeze	1.042	0.992–1.094	.100	1.037	0.986–1.090	.157
Physician‐diagnosed asthma	1.076	0.985–1.176	.102	1.051	0.960–1.150	.281
Phadiatop and/or Food Mix fx5 positive	0.989	0.936–1.044	.680	0.992	0.938–1.049	.775

Note: Multiple imputation of missing values in both the adjusted and unadjusted multivariate logistic regression models. Adjustment for maternal age, farm living, heredity for allergic disease, maternal region of birth, siblings/other children in the household and maternal education level. N = 4709 for both unadjusted and adjusted analyses. Food Mix fx5; Mix of egg white, milk, fish, wheat, peanut and soybean. Phadiatop; Mix of representative inhalant allergens.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; FA, food allergy; OR, odds ratio.

^a According to UK Working Party criteria. ²⁶

TABLE 4.

Associations between Dietary Inflammatory Index score as quartiles and offspring allergic outcomes at age 18 months in the NorthPop Birth Cohort Study.

Outcome	Exposure (quartiles)	OR (95% CI)	p‐Value	aOR (95% CI)	p‐Value
	DII
Physician‐diagnosed FA	Q1 vs. Q2	1.221 (0.756–1.972)	.415	1.220 (0.755–1.973)	.417
	Q1 vs. Q3	1.172 (0.720–1.909)	.523	1.172 (0.718–1.913)	.525
	Q1 vs. Q4	1.326 (0.843–2.087)	.222	1.324 (0.839–2.089)	.228
Parent‐reported eczema	Q1 vs. Q2	1.033 (0.845–1.263)	.750	1.025 (0.838–1.255)	.807
	Q1 vs. Q3	0.979 (0.799–1.200)	.839	0.973 (0.792–1.195)	.794
	Q1 vs. Q4	0.943 (0.767–1.161)	.581	0.927 (0.751–1.145)	.484
Parent‐reported atopic eczema a	Q1 vs. Q2	1.061 (0.789–1.428)	.694	1.067 (0.792–1.438)	.668
	Q1 vs. Q3	1.005 (0.740–1.364)	.974	1.027 (0.754–1.397)	.867
	Q1 vs. Q4	1.116 (0.827–1.506)	.474	1.138 (0.839–1.545)	.406
Parent‐reported ever wheeze	Q1 vs. Q2	1.129 (0.874–1.459)	.353	1.118 (0.864–1.447)	.397
	Q1 vs. Q3	1.186 (0.910–1.544)	.206	1.173 (0.899–1.532)	.240
	Q1 vs. Q4	1.134 (0.873–1.472)	.346	1.107 (0.848–1.445)	.454
Physician‐diagnosed asthma	Q1 vs. Q2	1.112 (0.677–1.828)	.675	1.075 (0.653–1.768)	.776
	Q1 vs. Q3	1.577 (0.980–2.535)	.060	1.480 (0.914–2.395)	.111
	Q1 vs. Q4	1.358 (0.830–2.223)	.223	1.218 (0.738–2.008)	.440
Phadiatop and/or Food Mix fx5 positive	Q1 vs. Q2	0.944 (0.721–1.238)	.678	0.944 (0.717–1.242)	.678
	Q1 vs. Q3	0.851 (0.645–1.123)	.254	0.867 (0.653–1.151)	.324
	Q1 vs. Q4	0.987 (0.749–1.301)	.927	1.010 (0.762–1.340)	.943

Note: First quartile as reference. Adjustment for maternal age, farm living, heredity for allergic disease, maternal region of birth, siblings/other children in the household and maternal education level. Food Mix fx5; Mix of egg white, milk, fish, wheat, peanut and soybean. Phadiatop; Mix of representative inhalant allergens.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; FA, food allergy; OR, odds ratio; Q, quartile.

^a According to UK Working Party criteria. ²⁶

4. DISCUSSION

To our knowledge, this is the first study examining the association between maternal DII score and the risk of IgE sensitization in offspring using IgE serology. We also contribute to existing research on the associations between maternal DII during pregnancy and allergic diseases. In this large, population‐based birth cohort study, we found no association between maternal DII scores reported in the last trimester and the risk of food allergy, eczema, atopic eczema, wheeze, asthma, or IgE sensitization in offspring at 18 months of age.

Our results on sensitization and food allergy are consistent with a recent study from Australia involving a smaller cohort of infants with at least one first‐degree relative with allergic disease. ²⁰ In comparison, our population‐based cohort had a lower prevalence of physician‐diagnosed food allergy (4.9% vs. 14.1%) and sensitization (19.0% vs. 25.8%) although assessed later in the Australian study (18 vs. 12 months) and by skin prick test. Similarly, we found no association between maternal DII and offspring eczema, aligning with the two previously conducted studies as well. ¹⁶ , ²⁰ The similar findings in both high‐risk and low‐risk populations strengthen the conclusion that a higher maternal DII during pregnancy is not associated with an increased risk of eczema, food allergy, or sensitization in the offspring.

In line with a previous meta‐analysis of seven European birth cohorts, ¹⁸ we found no association between maternal DII and offspring ever wheeze or asthma at this very early age. Diagnosing asthma before 3 years of age is challenging and primarily based on recurring symptoms combined with a probability assessment considering clinical findings, the presence of atopic disease, and family history. ³⁶ Other studies focusing on DII and asthma have assessed the prevalence of asthma at later ages, ranging from 3 to 9 years of age ¹⁴ , ¹⁵ , ¹⁷ , ¹⁸ which is an important difference when comparing the results. Wheeze is a common asthma symptom, and three studies, ¹⁵ , ¹⁶ , ¹⁸ including the above‐mentioned meta‐analysis, have investigated the association between DII and early wheeze, with conflicting results. Only the study by Hanson et al. ¹⁵ reports any association between maternal DII and early wheeze. A recent study from the ELFE birth cohort in France reported a weak association between the energy‐adjusted DII and belonging to the “cluster asthma‐only” up to five and a half years of age, but no association to clusters including food allergy, eczema, allergic conjunctivitis, and rhinitis. ³⁷

Studies examining maternal DII during pregnancy and its association to inflammatory cytokine levels have also reported conflicting results. Higher maternal DII scores have been linked to higher maternal CRP levels, ⁷ IL‐6 levels ³⁸ and TNF‐α levels ³⁹ while others report no association to these or other inflammatory markers. ⁴⁰ , ⁴¹ , ⁴² Adding to this complexity, Cui et al. reported a non‐linear, U‐shaped relationship between the DII score and IL‐1β and monocyte chemoattractant MCP‐1 levels in the third trimester ⁴³ which should be accounted for in future studies. To date, no meta‐analyses have been conducted to clarify the relationship between maternal DII during pregnancy and inflammatory markers, and such efforts would be a valuable contribution to the field. The association between maternal DII and inflammatory markers in maternal blood has not been validated in our study population, which is a potential limitation.

Immunological adaptations during pregnancy are complex and dynamic, involving trimester‐specific changes in immune activity. The first and third trimesters are typically characterized by a proinflammatory state, whereas the second trimester is associated with an anti‐inflammatory environment. ⁴⁴ In this study, we assessed maternal diet during the third trimester, the same time point as the studies by both Sivula ¹⁶ and Pretorius ²⁰ with colleagues, which also did not find any association between maternal DII and offspring allergic disease. However, the two original studies reporting an association between maternal DII and offspring respiratory outcomes, including asthma ¹⁴ , ¹⁵ assessed maternal diet in the first trimester. The meta‐analysis used FFQ data from both pre‐, early, and late pregnancy. ¹⁸ The time point of data collection may be of importance since one postulated mechanism of action, i.e., the epigenetic modifications, mainly occur during early embryogenesis and gametogenesis, well within the first trimester. ⁴⁵

Furthermore, studies also report significant changes in the individual diet between the first and later trimesters with increasing intake of micronutrients ⁴⁶ , ⁴⁷ and conflicting results regarding changes in dietary patterns. ⁴⁸ , ⁴⁹ Gete et al. reported an increased score in the Health Eating index during pregnancy compared to preconception and that the increase was further influenced by a higher maternal educational level. ⁴⁸ In Sweden, the National Food Agency provides advice on nutritional and safe eating for pregnant women. ⁵⁰ This could lead to a more uniform diet among the study participants in the last trimester, potentially masking important differences in early pregnancy. A study from Norway, however, highlighted low adherence to the Nordic Nutrition Recommendations regarding total carbohydrates, saturated fat, iron, folate, calcium, iodine, and vitamin D. ⁵¹ The timing of dietary exposure assessment is therefore critical and may significantly influence the observed associations with both maternal inflammation and offspring allergy outcomes.

The major strengths of this study are its prospective design and the large, population‐based cohort (n = 4709) including data on IgE sensitization among 2436 participants at 18 months of age. We also identified potential confounders a priori using a DAG, made possible by the extensive data collection in the NorthPop Birth Cohort Study, which included both questionnaire data and national registers. Statistical analyses were conducted using both the DII score as a continuous variable and as quartiles with no association or trend with the outcomes of interest. The study had a relatively high proportion of missing values for certain variables, which were assumed to be missing at random. Therefore, we applied multiple imputation instead of complete‐case analysis to minimize bias ⁵² and preserve the sample size, thereby retaining statistical power.

Limitations are the use of self‐reported dietary data. ⁵³ The FFQ used included 30 of the 45 food parameters used in the DII. All the 15 missing food parameters had a negative inflammatory effect score on the total DII, which represents anti‐inflammatory properties that can skew the total DII score towards a more proinflammatory diet. However, a validation study showed no reduction in the ability to predict interval changes in high‐sensitivity CRP values using questionnaires containing 28 out of 45 parameters compared to more comprehensive methods. ⁵⁴ Notably, none of the previous studies on maternal DII in pregnancy and offspring allergic disease were able to include all 45 food parameters. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ²⁰ The mothers in the study population were mostly of Swedish descent and had a high educational level in general, which may limit the generalizability of our findings to more diverse populations. A previous analysis from a study conducted within the NorthPop Birth Cohort Study further highlights potential selection bias, as non‐responders tended to have younger maternal age and lower educational level. ⁵⁵ Participants who attended the blood sampling at 18 months were also more likely to have a mother with food allergy and a higher educational level. ⁵⁵

It has been demonstrated that self‐reported food hypersensitivity and food allergy will overestimate the prevalence as compared with a diagnosis confirmed with objective diagnostic methods. ⁵⁶ Here, we included parent‐reported physician diagnosis of food allergy only, which may lead to an underestimation of the prevalence, but on the other hand limits detection bias. Eczema is an umbrella term, including both atopic and non‐atopic eczema ⁵⁷ which is why we applied the UK Working Party criteria for the diagnosis of atopic eczema. The prevalence of parent‐reported eczema was high (30.6%) at 18 months of age. Using the UK Working Party criteria provided good alignment with global data on the prevalence of atopic eczema. ⁵⁸ Additionally, POEM scores indicated that most cases of parent‐reported eczema were mild or nearly absent, whereas the POEM scores were significantly higher in the group of parent‐reported atopic eczema.

In conclusion, we found no association between maternal DII during pregnancy and early allergic outcomes in offspring. Importantly, our study found no association with IgE sensitization at 18 months, which is a novel finding. This suggests that the inflammatory potential of maternal diet in pregnancy may not significantly influence early sensitization processes in children. Further research is needed to explore possible effects of other aspects of maternal diet during pregnancy on offspring immune development and risk of allergic disease.

FUNDING INFORMATION

This study was funded by the Swedish Research Council grant number 2018‐02642 and 2021‐01637 (CEW) and 2023‐01784 (MD); the Heart‐Lung Foundation grant number 20180641 (CEW); the Ekhaga Foundation grant number 2018‐40 (CEW) and the Västerbotten County Council (ALF) grant numbers RV 832 441, RV 840 681, and RV 967 569 (CEW). The NorthPop Birth Cohort study infrastructure receives funding from Västerbotten County Council and Umeå University (CEW and MD). Thermo Fisher/Scientific Phadia provided reagents for the Immunoglobulin E tests. The funding bodies had no role in study design, data collection and analysis, nor in the preparation of the manuscript.

CONFLICT OF INTEREST STATEMENT

CEW has received research funding, which was paid directly to the institution, from Thermo Fisher Scientific/Phadia. The other authors have no conflicts of interest to declare.

Supporting information

Figure S1

Figure S2

Figure S3

Figure S4

Table S1

APPENDIX A. QUESTIONNAIRE ITEMS FROM THE NORTHPOP BIRTH COHORT STUDY TRANSLATED FROM SWEDISH TO ENGLISH

A.1. Mother questionnaire gestational age 14–24 weeks

• Where were you born?

  Sweden/other country

• If other country: Which country?

• Do you or your biological parents, siblings, or children have any of the following allergic diseases (asthma, food allergy, allergy to furry pets, hay fever, eczema)?

  Yes/No

• What type of housing do you mainly live in?

  Single‐family detached home, apartment building, agricultural property/farm, other type of housing

• Who else lives in your household?

  The expected child's biological father, another partner/cohabiting person who is not the expected child's biological father, relatives, friends, or other adults, children

A.2. Partner questionnaire gestational age 14–24 weeks

• Where were you born?

  Sweden/other country

• If other country: Which country?

• Do you or your biological parents, siblings, or children have any of the following allergic diseases (asthma, food allergy, allergy to furry pets, hay fever, eczema)?

  Yes/No

A.3. Mother questionnaire gestational age 26 weeks

• Have you smoked cigarettes during this pregnancy?

  No, Yes, I smoked during the last month before I became pregnant, Yes, I have smoked during this pregnancy

• Does anyone else in your household smoke?

  No, Yes, but the person only smokes outdoors, Yes, the person smokes indoors but only under the kitchen fan, Yes, the person smokes indoors and not only under the kitchen fan

A.4. Mother questionnaire gestational age 34 weeks

• A food frequency questionnaire including 125 items and portion‐size estimations for vegetables, carbohydrates, and protein sources was used. The intake frequency for solid foods was reported as: never or rarely, 1–3 times/week, 4–6 times/week, 1–2 times/day, or ≥3 times/day. Bread and sandwich toppings were reported as slices per day, along with information on how often different types of bread were consumed, ranked from 1 (most frequently) to 6 (least frequently). Beverages were reported using six frequency categories: never or rarely, 1–3 glasses/week, 4–6 glasses/week, 1–2 glasses/day, 3–4 glasses/day, and ≥5 glasses/day.

A.5. Child 4‐month questionnaire

• Does the child share a home with siblings or other children?

  Yes/no

• Which type of milk is the child currently receiving?

  Breastmilk, Only formula, Both breastmilk and formula

• Does your child have eczema?

  Yes/No

  If yes, the following questions from the POEM score were asked.

  • ◦
    Over the last week, on how many days has your child's skin been itchy because of the eczema?
  • ◦
    Over the last week, on how many nights has your child's sleep been disturbed because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been bleeding because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been weeping or oozing clear fluid because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been cracked because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been flaking off because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin felt dry or rough because of the eczema?

A.6. Child 9‐month questionnaire

• Does your child have eczema?

  Yes/No

• Has the child been diagnosed with food allergy by a physician?

  Yes/No

• If physician‐diagnosed food allergy: Which food(s) does the child have a physician‐diagnosed allergy/intolerance to?

• Has the child ever wheezed?

  Yes/No

• Has the child been diagnosed with asthma by a physician?

  Yes/No

• Do you breastfeed your child now?

  Yes/No

A.7. Child 18‐month questionnaire

• Has the child ever wheezed?

  Yes/No

• Has the child been diagnosed with asthma by a physician?

  Yes/No

• Does your child have eczema? Yes/No. If yes, the following questions from the POEM score were asked

  • ◦
    Over the last week, on how many days has your child's skin been itchy because of the eczema?
  • ◦
    Over the last week, on how many nights has your child's sleep been disturbed because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been bleeding because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been weeping or oozing clear fluid because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been cracked because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin been flaking off because of the eczema?
  • ◦
    Over the last week, on how many days has your child's skin felt dry or rough because of the eczema?

• Has the child been diagnosed with food allergy by a physician?

  Yes/No

• If physician‐diagnosed food allergy: Which food(s) does the child have a physician‐diagnosed allergy/intolerance to?

Österlund J, Bodén S, Granåsen G, et al. Maternal dietary inflammatory index during pregnancy and the risk of offspring allergic disease. Pediatr Allergy Immunol. 2025;36:e70148. doi: 10.1111/pai.70148