ABSTRACT
Asthma affects over 850,000 Canadian children under 14, ranking among the top chronic childhood diseases. Prenatal dietary factors are hypothesised to influence its development. We examined associations between maternal prenatal dietary patterns and childhood asthma/wheeze in the CHILD cohort. Principal component analysis derived three dietary patterns (plant‐based, Western, balanced) from food frequency questionnaires. Logistic regression and generalised estimating equations (GEE) models assessed associations with physician‐diagnosed asthma (n = 1574) and recurrent wheeze (n = 1799–2374) at ages 1–3 years, adjusting for covariates. Among 2968 mother–child pairs, maternal asthma (OR = 2.16, 95% CI: 1.10–4.18), paternal asthma/wheeze (OR = 2.14, 1.26–3.56), and gestational age (OR = 0.79, 0.67–0.93) predicted asthma risk. Maternal wheeze (OR = 1.59, 95% CI: 1.05–2.38) and prior wheeze (OR = 4.30, 2.93–6.25) predicted recurrent wheeze. No associations were found between dietary patterns and asthma (plant‐based OR = 1.40/1‐point DS increase, 0.80–2.43; Western OR = 1.20, 0.86–1.65; balanced OR = 0.98, 0.74–1.26) or wheeze. While familial and perinatal factors strongly predict asthma risk, maternal prenatal diet patterns showed no association with childhood respiratory outcomes.
Keywords: CHILD, children's asthma, children's wheeze, cohort study, maternal diet, nutritional epidemiology
1. Introduction
Asthma, a chronic inflammatory airway disorder characterised by bronchoconstriction and airway hyperresponsiveness, represents a significant global public health challenge (Allan and Devereux 2011). With over 3.8 million affected Canadians, asthma is the most prevalent chronic disease among children under 14 years, accounting for 11% of cases and serving as a leading cause of school absenteeism and paediatric hospitalisations (Polisena et al. 2007). Globally, its prevalence continues to rise, particularly in Westernised nations, imposing substantial burdens on healthcare systems (Lawson et al. 2017; Masoli et al. 2004).
Emerging evidence underscores the developmental origins of asthma, wherein foetal exposure to environmental and nutritional factors during critical windows of gestation may programme immune and respiratory outcomes (Martino and Prescott 2011). The intrauterine environment, shaped by maternal diet, plays a pivotal role in regulating foetal immune development and susceptibility to allergic diseases. For instance, maternal deficiencies in vitamin D, vitamin E and zinc during pregnancy have been associated with elevated risks of childhood wheezing and asthma (Devereux et al. 2006). Conversely, higher prenatal fish consumption, rich in anti‐inflammatory omega‐3 fatty acids, correlates with reduced allergic sensitization in offspring (Calvani et al. 2006). These findings highlight the potential for maternal dietary interventions to modulate developmental trajectories of asthma.
Historically, nutritional research adopted a reductionist approach, focusing on single nutrients. However, this strategy has proven inadequate for assessing the role of diet in chronic disease prevention because human diets comprise complex interactions of foods and nutrients (Fardet and Rock 2014). Synergistic or antagonistic nutrient interactions—such as vitamin C enhancing nonheme iron absorption or iron inhibiting zinc uptake—illustrate the limitations of isolating individual dietary components (Sandström 2001). Moreover, individuals consume foods, not nutrients, necessitating a shift towards studying holistic dietary patterns. Such patterns, which reflect combinations of foods and eating behaviours, offer greater translational relevance for public health recommendations (Cespedes and Hu 2015).
Recent epidemiological studies have investigated prenatal dietary patterns and childhood asthma, yielding mixed results. Adherence to a Mediterranean diet—rich in fruits, vegetables, whole grains and fish—during pregnancy was linked to reduced risks of persistent wheezing and atopy in children (Chatzi et al. 2008). Similarly, a cohort study in Ireland associated higher maternal intake of fruits, vegetables and oily fish with lower asthma incidence by age 3 years (Fitzsimon et al. 2007). However, a meta‐analysis of six cohorts (n = 2516) found no consistent association between Mediterranean diets and asthma outcomes, underscoring heterogeneity in existing evidence (Lv et al. 2014). Discrepancies may arise from variations in dietary assessment methods, population characteristics or exposure and outcome definitions, necessitating further investigation in diverse cohorts.
The CHILD cohort study provides a unique opportunity to examine these relationships in a multi‐ethnic, population‐based cohort. This study evaluates the association between maternal dietary patterns during pregnancy and recurrent wheeze or asthma diagnosis in offspring.
2. Methods
2.1. Study Population
The CHILD cohort study is a multi‐centre, population‐based birth cohort that enrolled 3621 pregnant women between 2008 and 2012 across four major Canadian cities (Vancouver, Edmonton, Winnipeg and Toronto) and a rural region (Morden and Winkler, Manitoba) (Miliku et al. 2024). All the participants provided informed consent at the prior to data collection. For this analysis, data from 3454 mother–child pairs were extracted from derived datasets and questionnaire files. After excluding 486 records due to missing dietary score (exposure) information, a total of 2968 observations were included in the descriptive statistics.
At the 1‐year follow‐up, complete data for recurrent wheeze outcomes were available for 2434 mother–child pairs, with 534 observations missing primarily due to loss to follow‐up or incomplete questionnaire responses. At the 3‐year follow‐up, complete data for recurrent wheeze were available for 2590 pairs (378 missing), while physician‐diagnosed asthma data were available for 2312 pairs (656 missing), reflecting incomplete clinical follow‐up rather than selective outcome non‐response (Figure 1).
FIGURE 1.
CONSORT diagram for the sample size.
All analyses were conducted using all available data for each outcome, such that participants were included in outcome‐specific models whenever both the outcome and relevant covariates were observed.
2.2. Data Collection
The study data were collected and managed using REDCap (Research Electronic Data Capture) hosted at the British Columbia Children's Hospital Research Institute (BCCHR). REDCap is a secure, web‐based software platform designed to support data capture for research studies (Harris et al. 2019; Harris et al. 2009).
2.3. Dietary Pattern Classification
The plant‐based dietary pattern (score range: −2.6 to +4.9) was characterised by high positive loadings (> 0.50) for fruits, vegetables, whole grains, legumes and plant proteins (e.g., legumes: 0.62, whole grains: 0.71), with negative loadings for meats (−0.36). The Western pattern (score range: −3.9 to +5.0) featured high loadings for processed meats (0.55), refined carbohydrates (pasta: 0.55, pizza: 0.32) and sweet drinks (0.55). The balanced pattern (score range: −2.6 to +7.1) showed diverse loadings across proteins (fish: 0.50, poultry: 0.36), vegetables (cruciferous: 0.55) and fruits (0.53). Collectively, these three patterns explained 29% of the total variance in maternal dietary habits during pregnancy, as indicated by the cumulative variance from principal component analysis (PCA).
2.4. Assessment of Diet, Harmonisation and Dietary Pattern Analysis
Dietary intake during pregnancy was assessed using a validated Food Frequency Questionnaire (FFQ) adapted from the Fred Hutchinson Cancer Research Center. The FFQ captured detailed information on consumption of 152 items (de Souza et al. 2016), which were reduced into 36 predefined food groups based on nutrient profiles and food types. These groups were established to facilitate consistent dietary pattern analysis across different cohorts (de Souza et al. 2016).
PCA was employed to identify dietary patterns, utilising the ‘psych’ package (v.1.5.6) in R (v.3.1.2). An orthogonal ‘varimax’ rotation was applied to maximise interpretability of the derived patterns, which represent combinations of foods frequently consumed together. The number of dietary patterns retained was determined through visual inspection of scree plots, evaluation of eigenvalues (> 1), and the interpretability of the components. Three distinct dietary patterns emerged from the analysis, collectively explaining 29% of the total dietary variability within the cohort (Zulyniak et al. 2017). These patterns were labelled as: (1) plant‐based (adherence range: −2.6 to +4.9), (2) Western (adherence range: −3.9 to +5.0) and (3) balanced (adherence range: −2.6 to +7.1) (Appendix S1).
Each participant received an adherence score for each dietary pattern, with higher scores indicating stronger alignment with the respective pattern. To account for variations in total energy intake, adherence scores were energy‐adjusted using the residual method (Zulyniak et al. 2017). This approach ensures that the observed associations reflect dietary composition rather than differences in energy consumption.
2.5. Assessment of Asthma and Recurrent Wheeze
The primary outcome of this analysis was physician‐diagnosed asthma at age 3 years, as reported by healthcare providers through a standardised child health questionnaire. Secondary outcomes included parent‐reported recurrent wheeze at ages 1 and 3 years. Recurrent wheeze was defined as two or more episodes of wheezing within the past 12 months, as reported by the child's caregiver (Subbarao et al. 2015).
To assess the robustness of our findings, a sensitivity analysis was conducted using parent‐reported asthma at age 3 years as an alternative outcome measure. This approach allowed for the evaluation of potential discrepancies between physician‐diagnosed and caregiver‐reported asthma, providing additional insight into the reliability of the primary outcome.
2.6. Measurement of Other Variables
Data on potential covariates, including birthweight, maternal age, maternal history of wheezing, mother's education, family income, presence of furry pets in the household, paternal asthma, breastfeeding status and maternal ethnicity, were collected through questionnaires completed by the mothers. These known and suspected confounders were selected based on their potential influence on childhood asthma and wheeze outcomes, as supported by existing literature.
2.7. Statistical Analysis
Dietary patterns during pregnancy were derived using PCA with orthogonal varimax rotation, identifying three distinct patterns: plant‐based, Western and balanced (Zulyniak et al. 2017). Logistic regression was used to assess associations between maternal dietary patterns and childhood wheeze (at ages 1 and 3 years) and physician‐diagnosed asthma (at age 3 years). Additionally, generalised estimating equations (GEE) were employed to account for repeated measures of wheeze outcomes over time. Multicollinearity was assessed using variance inflation factors (VIF) and model fit was evaluated using the Hosmer–Lemeshow test. Missing data were handled using pairwise deletion, whereby all available observations were included in each outcome‐specific model. This approach allowed participants with partial follow‐up to contribute information to analyses of outcomes for which data were available, thereby maximising statistical efficiency while avoiding unnecessary listwise exclusion.
To assess the robustness of results to missing outcome data, sensitivity analyses were conducted using parent‐reported asthma at age 3 years as an alternative outcome. Agreement between physician‐diagnosed and parent‐reported asthma was evaluated using Cohen's κ, providing an empirical check on potential bias due to incomplete physician assessments. The psych package in R was used to compute κ with 95% confidence intervals (CIs). Agreement strength was interpreted using Landis and Koch's criteria (Landis and Koch 1977) (κ: 0.21–0.40 = fair; 0.41–0.60 = moderate). All regression models were specified to estimate the independent association between maternal dietary patterns during pregnancy and childhood asthma and wheeze, conditional on established parental, environmental and child‐level risk factors. Adjusted regression models controlled for covariates including maternal age, ethnicity, maternal smoking during pregnancy, gestational age, pre‐pregnancy BMI, breastfeeding and family history of asthma. Selection of covariates was guided by epidemiologic relevance and implemented using a forward stepwise selection approach. Candidate variables identified a priori based on existing literature and biologic plausibility were entered sequentially into the model, beginning with the exposure of interest. At each step, variables were retained if they were associated with the outcome at p ≤ 0.10 or if their inclusion improved overall model fit. This approach allowed the identification of a parsimonious multivariable model while accounting for potential confounding factors. Odds ratios (ORs) with 95% CIs were calculated to quantify associations, represent adjusted associations rather than the relative importance of variables in a predictive model. To assess the robustness of our findings to outcome definition, we conducted a sensitivity analysis using a composite outcome defined as the presence of either physician‐diagnosed asthma or recurrent wheeze at age 3. A multivariable logistic regression model was fitted using the same modelling strategy and covariate selection approach as the primary analyses. Estimates are presented as ORs with 95% CIs. All analyses were performed in R (version 4.3.3), with statistical significance set at p < 0.05.
2.8. Patient and Public Involvement
The CHILD cohort study emphasises participant and public involvement in all stages of its research. Research questions are informed and generated based on relevance to participants and their families to study those affected by childhood health conditions. Efforts are made to ensure participants' involvement is meaningful, comfortable and aligned with their preferences. Study results will be shared with participants through various channels, including newsletters and online platforms. Academic and public health publications will also communicate findings. We acknowledge and thank all participant advisers and study participants for their valuable contributions.
2.9. Ethics Approval
Ethical approval for the CHILD cohort study, including the oversight of the CHILD biological samples and the CHILD database (CHILDdb), was obtained from the local Research Ethics Board of each study site: the University of British Columbia, the University of Alberta, the University of Manitoba, the Hospital for Sick Children and McMaster University. The research described herein was approved by Hamilton Health Sciences Research Ethics Board (HiREB #11191).
3. Results
3.1. Demographic and Clinical Parameters
The study included 2434 mother–child pairs for recurrent wheeze at Year 1, 2590 for recurrent wheeze at Year 3 and 2312 for physician‐diagnosed asthma at Year 3. Maternal mean ages were ranging from 32.9–33.5 years for three different outcome groups, with 75.2% identifying as White Caucasian. Most mothers were married (94.4%–95%) and reported no gestational diabetes (95.6%–95.9%) or hypertension (97.9%–98.1%). Maternal wheeze (29.1%–29.6%) and asthma (22%–22.4%) were prevalent, with 37.4%–37.7% reporting allergies. Mean gestational age was 38.7–39.1 weeks, mean pre‐pregnancy BMI was 23.93–24.73 kg/m2 and 96.2%–96.3% of infants were breastfed. Most of the mothers' residential location were in the urban area (93.5%–94%). Mean birthweight ranged from 3396.9 to 3469.0 g and birth length from 51.0 to 51.6 cm. Households predominantly had two adults (84.4%–84.9%) and 43.4%–44.3% reported having furry pets at Year 1. Table 1 shows the details of demographic information of mother and child by outcome. Non‐white mothers had lower rates of wheeze (23% vs. 31.6%) and asthma (16.9% vs. 22.4%) but higher gestational diabetes (6.6% vs. 3.7%). Non‐white infants had lower birthweights (3348.5 g vs. 3485.9 g) and fewer furry pets (26.7% vs. 49.2%) Table 2.
TABLE 1.
Demographic information of mothers and offsprings by result sets.
| Variables | Children with data for recurrent wheeze (Year 1) (n, %, N) | Children with data for Recurrent Wheeze (Year 3) (n, %, N) | Children with data for physician's diagnosed Asthma (Year 3) (n, %, N) |
|---|---|---|---|
| Mothers | N = 2434 | N = 2590 | N = 2312 |
| Age mean (SD) | 32.94 (4.29) | 33.35 (4.14) | 33.47 (3.49) |
| Race | N = 2432 | N = 2586 | N = 2311 |
| Caucasian White | 1828 (75.2%) | 1917 (74.1%) | 1717 (74.3%) |
| Non‐white | 604 (24.8%) | 669 (25.9%) | 594 (25.7%) |
| Marital status of Mother | N = 2423 | N = 2577 | N = 2307 |
| Married/common law partner | 2302 (95%) | 2433 (94.4%) | 2191 (95%) |
| Divorced | 10 (0.4%) | 14 (0.5%) | 13 (0.6%) |
| Single/never married | 111 (4.6%) | 130 (5%) | 103 (4.5%) |
| Residential location | N = 2315 | N = 2469 | N = 2194 |
| Rural | 143 (6.2%) | 160 (6.5%) | 132 (6.0%) |
| Urban | 2172 (93.8%) | 2309 (93.5%) | 2062 (94.0%) |
| Mother's education | N = 2406 | N = 2560 | N = 2287 |
| Less than high school | 9 (0.4%) | 12 (0.5%) | 8 (0.03%) |
| Some high school | 51 (2.1%) | 56 (2.2%) | 39 (1.7%) |
| Completed high school | 138 (5.7%) | 146 (5.7%) | 115 (5.0%) |
| Some college | 145 (6.0%) | 151 (5.9%) | 138 (6.0%) |
| Completed college | 333 (13.8%) | 353 (13.8%) | 317 (13.9%) |
| Some university | 186 (7.7%) | 206 (8.0%) | 167 (7.3%) |
| Completed university | 1088 (45.2%) | 1141 (44.6%) | 1050 (45.9%) |
| Master's degree | 456 (19.0%) | 495 (19.3%) | 453 (19.8%) |
| Family income | N = 2406 | N = 2561 | N = 2287 |
| $0–$9999 | 18 (0.7%) | 18 (0.7%) | 15 (0.7%) |
| $10 000–$19 999 | 38 (1.6%) | 39 (1.5%) | 31 (1.4%) |
| $20 000–$29 999 | 44 (1.8%) | 46 (1.8%) | 37 (1.6%) |
| $30 000–$39 999 | 82 (3.4%) | 89 (3.5%) | 76 (3.3%) |
| $40 000–$49 999 | 106 (4.4%) | 112 (4.4%) | 96 (4.2%) |
| $50 000–$59 999 | 128 (5.3%) | 133 (5.2%) | 116 (5.1%) |
| $60 000–$79 999 | 296 (12.3%) | 310 (12.1%) | 274 (12.0%) |
| $80 000–$99 999 | 330 (13.7%) | 354 (13.8%) | 319 (13.9%) |
| $100 000–$149 999 | 636 (26.4%) | 658 (25.7%) | 604 (26.4%) |
| $150 000 or over | 513 (21.3%) | 565 (22.1%) | 517 (22.6%) |
| Prefer not to say | 215 (8.9%) | 237 (9.3%) | 202 (8.8%) |
| Wheeze of mother (yes) | 719 (29.6%) (2428) | 753 (29.2%) (2582) | 671 (29.1%) (2308) |
| Asthma | 543 (22.4%) (2426) | 573 (22.2%) (2580) | 507 (22%) (2307) |
| Asthma (diagnosed) | 518 (21.4%) (2422) | 544 (21.1%) (2576) | 481 (20.9%) (2303) |
| Allergy (pollen, tree, grass) | 915 (37.7%) (2426) | 964 (37.4%) (2580) | 868 (37.6%) (2307) |
| Food allergy | 548 (22.6%) (2427) | 574 (22.7%) (2526) | 515 (22.8%) (2257) |
| GDM | 107 (4.4%) (2434) | 109 (4.2%) (2590) | 95 (4.1%) (2312) |
| HTN | 52 (2.1%) (2434) | 51 (2%) (2590) | 44 (1.9%) (2312) |
| Preeclampsia | 89 (3.7%) (2434) | 94 (3.6%) (2590) | 84 (3.6%) (2312) |
| MultiVit during Preg | 315 (13%) (2414) | 340 (13.2%) (2569) | 296 (12.9%) (2299) |
| Gestational age (week) mean (SD) | 39.06 (1.36) | 38.98 (1.44) | 38.72 (1.69) |
| Gestational weight gain (Kg) mean (SD) | 15.34 (6.34) | 15.95 (6.71) | 16.44 (6.51) |
| Pre‐pregnancy BMI a mean (SD) | 24.73 (4.20) | 23.93 (3.53) | 24.14 (4.22) |
| Smoking during Preg | 229 (9.4%) (2427) | 239 (9.3%) (2581) | 190 (8.2%) (2308) |
| Years in Canada mean (SD) | 27.76 (9.37) | 25.46 (10.79) | 24.03 (10.45) |
| Infant | |||
| Birthweight mean, gm (SD) | 3469.02 (475.27) | 3423.49 (487.29) | 3396.89 (496.79) |
| Birth length mean, cm (SD) | 51.61 (2.37) | 51.41 (2.28) | 51.01 (2.49) |
| Ever breastfed (Year 1) | 2304 (96.2%) (2396) | 2125 (96.2%) (2209) | 1949 (96.3%) (2023) |
| Household | |||
| Number of adults (Year 1) | N = 2404 | N = 2556 | N = 2286 |
| 1 | 54 (2.2%) | 64 (2.5%) | 50 (2.2%) |
| 2 | 2040 (84.9%) | 2158 (84.4%) | 1936 (84.7%) |
| ≥ 3 | 310 (12.9%) | 334 (13.1%) | 300 (13.1%) |
| Number of adults (Year 3) | N = 1723 | N = 1844 | N = 1865 |
| 1 | 52 (3%) | 56 (3%) | 57 (3.1%) |
| 2 | 1439 (83.5%) | 1542 (83.6%) | 1553 (83.3%) |
| ≥ 3 | 232 (13.5%) | 246 (13.3%) | 255 (13.7%) |
| Furry pet in house (Year 1) | 1029 (43.9%) (2342) | 978 (44.3%) (2206) | 878 (43.4%) (2021) |
| Furry pet in house (Year 3) | 785 (40.3%) (1946) | 819 (38.3%) (2140) | 820 (38.2%) (2148) |
| Furry pet in house (prenatal) | 1201 (49.9%) (2406) | 1246 (48.7%) (2557) | 1088 (47.6%) (2285) |
| Dogs (of furry pets) | 772 (32.1%) (2407) | 791 (30.9%) (2557) | 693 (30.3%) (2285) |
| Cats (of furry pets) | 608 (50.7%) (1200) | 642 (51.6%) (1244) | 550 (50.6%) (1087) |
Note: N represents the number of participants responded to the specific variables, n represents the number of events and the % represents the percentage of events.
Abbreviations: HTN, hypertension; GDM, gestational diabetes, SD, standard diviation.
Pre‐pregnancy weight data were available for approximately 73% of women in the CHILD cohort. To address missing values, a multivariable prediction model was developed using data from White European participants enrolled in the CHILD and FAMILY cohorts. The predicted pre‐pregnancy weights generated from this model showed a moderate correlation (r = 0.42) with the observed pre‐pregnancy weight values. Then the BMI value was calculated using Mother's pre‐pregnancy weight in kg/Mother's height in m2 formula.
TABLE 2.
Demographic information of mothers and offsprings by ethnicity.
| Variables | White, n, (%), (N) | Non‐white, n, (%), (N) |
|---|---|---|
| Mother | ||
| Age (years) mean (SD) | 33.4 (4.3) | 33.4 (4.4) |
| Marital status of mother | ||
| Married/common law partner | 2086 (95%) | 697 (91%) |
| Divorced | 10 (1%) | 9 (1%) |
| Single/never married | 86 (4%) | 58 (8%) |
| Residential location | ||
| Rural | 151 (7.2%) | 21 (2.9%) |
| Urban | 1943 (92.8%) | 708 (97.1%) |
| Wheeze | 690 (31.6%) (2186) | 176 (23%) (766) |
| Asthma (diagnosed) | 490 (22.4%) (2183) | 129 (16.9%) (763) |
| Asthma | 510 (23.3%) (2185) | 141 (18.4%) (765) |
| Allergy (pollen, tree, grass) | 843 (38.6%) (2184) | 268 (35%) (765) |
| Food allergy | 484 (22.1%) (2186) | 167 (21.8%) (766) |
| GDM | 82 (3.7%) (2194) | 51 (6.6%) (768) |
| HTN of Mother | 36 (1.6%) (2194) | 19 (2.5%) (768) |
| Preeclampsia | 74 (3.4%) (2194) | 30 (3.9%) (768) |
| MultiVit during pregnancy | 281 (12.9%) (2172) | 120 (15.8%) (760) |
| Smoking during pregnancy | 225 (10.3%) (2185) | 80 (10.4) (766) |
| Gestational age (week) mean (SD) | 39.2 (1.3) | 38.9 (1.4) |
| Gestational weight gain (Kg) mean (SD) | 15.41 (6.19) | 14.62 (6.51) |
| Pre‐pregnancy BMI mean (SD) | 24.69 (4.44) | 24.19 (4.09) |
| Years in Canada mean (SD) | 28.9 (8.6) | 20.9 (12.4) |
| Infant | ||
| Birthweight mean, gm (SD) | 3485.9 (472.1) | 3348.5 (491.9) |
| Birth length mean, cm (SD) | 51.6 (2.6) | 51.2 (2.4) |
| Weight (Year 1, kg) (SD) | 9.9 (1.3) | 9.7 (1.3) |
| Ever breastfed (Year 1) | 1755 (95.8%) (1831) | 572 (96.6%) (592) |
| Breastfeed duration mean (SD) | ||
| Wheeze at Year 1 | 215 (11.8%) (1828) | 65 (10.8%) (604) |
| Wheeze at Year 3 | 175 (9.1%) (1917) | 85 (12.7%) (669) |
| Asthma of child at Year 3 | 62 (3.6%) (1717) | 31 (5.2%) (594) |
| Household | ||
| Number of adults (Year 1) | ||
| 1 | 42 (2%) | 28 (3.7%) |
| 2 | 1910 (88.3%) | 554 (73.9%) |
| ≥ 3 | 210 (9.7%) | 168 (22.4%) |
| Number of adults (Year 3) | ||
| 1 | 44 (3%) | 14 (3%) |
| 2 | 1250 (86.4%) | 342 (74.4%) |
| ≥ 3 | 154 (10.6%) | 104 (22.6%) |
| Furry pet in house (Year 1) | 900 (49.2%) (1830) | 159 (26.7%) (596) |
| Furry pet in house (Year 3) | 729 (43.7%) (1669) | 117 (21.5%) (543) |
| Furry pet in house (Prenatal) | 1180 (54.6%) (2162) | 228 (30.4%) (751) |
| Dogs (of furry pets) | 744 (34.4%) (2162) | 147 (19.5%) (752) |
| Cats (of furry pets) | 623 (52.8%) (1179) | 97 (42.7%) (227) |
Note: Sample sizes vary across variables due to missing data.
Abbreviations: HTN, hypertension; GDM, gestational diabetes, SD, standard deviation.
3.2. Dietary Patterns and Asthma
The association between maternal dietary patterns during pregnancy and asthma outcomes in children was examined using multiple logistic regression (Table 3). For physician‐diagnosed asthma at age 3 years, none of the dietary patterns—plant‐based (OR = 1.40, 95% CI: 0.80–2.43, p = 0.23), Western (OR = 1.20, 95% CI: 0.86–1.65, p = 0.27) or balanced (OR = 0.98, 95% CI: 0.74–1.26, p = 0.87)—showed significant associations. However, maternal diagnosed asthma (OR = 2.16, 95% CI: 1.10–4.18, p = 0.02) and paternal asthma or wheeze (OR = 2.14, 95% CI: 1.26–3.56, p < 0.001) were significantly associated with increased asthma risk. Additionally, gestational age (OR = 0.79, 95% CI: 0.67–0.93, p < 0.001) and mother's years in Canada (OR = 0.97, 95% CI: 0.85–0.99) were inversely associated with asthma. The model demonstrated good fit (Hosmer–Lemeshow χ 2 = 5.6, p = 0.69) and multicollinearity was minimal (VIF range: 1.02–1.68).
TABLE 3.
Multiple logistic regression models of diet patterns with asthma (n = 93 cases of asthma in 1704 children) and recurrent wheeze (n = 260 cases of recurrent wheeze in 1799 children) at age 3 years, Recurrent wheeze (n = 280 cases of recurrent wheeze at year 1 in 2374 children) at year 1, and Self‐reported Asthma (n = 334 SR Asthma in 1422 children) at year 3.
| Exposure | Recurrent Wheeze at Year 1 | Recurrent Wheeze at Year 3 | Physician's diagnosed Asthma at Year 3 | Self reported Asthma at Year 3 | ||||
|---|---|---|---|---|---|---|---|---|
| Odd ratio (95% CI) | p | Odd ratio (95% CI) | p | Odd ratio (95% CI) | p | Odd ratio (95% CI) | p | |
| Plant based diet | 0.89 (0.65, 1.23) | 0.49 | 0.80 (0.51, 1.25) | 0.33 | 1.40 (0.80, 2.43) | 0.23 | 0.69 (0.44, 1.05) | 0.08 |
| Western diet | 1.05 (0.88, 1.23) | 0.60 | 0.98 (0.77, 1.25) | 0.92 | 1.20 (0.86, 1.65) | 0.27 | 0.95 (0.75, 1.18) | 0.64 |
| Health‐Conscious diet | 1.06 (0.89, 1.25) | 0.47 | 1.07 (0.85, 1.34) | 0.53 | 0.98 (0.74, 1.26) | 0.87 | 1.18 (0.95, 1.46) | 0.12 |
| Mother's Wheeze (Y) | 1.59 (1.17, 2.16) | 0.003** | 1.59 (1.05, 2.38) | 0.03* | 0.67 (0.34, 1.28) | 0.24 | 0.87 (0.57, 1.31) | 0.50 |
| Mother's diagnosed Asthma (Y) | 1.24 (0.89, 1.73) | 0.19 | 1.26 (0.81, 1.97) | 0.29 | 2.16 (1.10, 4.18) | 0.02* | 1.30 (0.84, 2.02) | 0.24 |
| Mother's Allergy | 1.36 (1.03, 1.78) | 0.03* | 1.25 (0.89, 1.78) | 0.23 | 1.58 (0.93, 2.69) | 0.09 | 1.56 (1.11, 2.19) | < 0.01* |
| Gestational age (W) | 0.94 (0.86, 1.03) | 0.17 | 0.90 (0.80, 1.01) | 0.09 | 0.79 (0.67, 0.93) | < 0.001*** | 0.92 (0.81, 1.05) | 0.22 |
| Pre‐Pregnancy BMI | 0.94 (0.89, 0.98) | 0.004** | 0.93 (0.88, 0.97) | < 0.001*** | ||||
| Mother's Smoking during Pregnancy | 1.22 (0.80, 1.81) | 0.33 | 1.20 (0.47, 2.69) | 0.67 | ||||
| Mother's HTN (Yes) | 0.24 (0.01, 1.15) | 0.16 | ||||||
| Mother's Ethnicity (Non white) | 0.99 (0.65, 1.48) | 0.97 | 1.60 (0.89, 2.81) | 0.11 | 0.96 (0.64, 1.43) | 0.85 | ||
| Father's Asthma or Wheeze (Yes) | 1.29 (0.87, 1.88) | 0.19 | 2.14 (1.26, 3.56) | < 0.001*** | 1.93 (1.35, 2.75) | < 0.001*** | ||
| Mother's Years in Canada | 0.98 (0.96, 0.99) | 0.007** | 0.97 (0.95, 0.99) | 0.01** | 0.99 (0.97, 1.00) | 0.15 | ||
| Ever Breastfed during Year 1 (Yes) | 0.81 (0.35, 2.25) | 0.66 | 0.94 (0.30, 4.30) | 0.92 | ||||
| Child's Wheeze at Year 1 | 4.30 (2.93, 6.25) | < 0.001*** | 5.31 (3.10, 8.95) | < 0.001*** | 2.96 (1.97, 4.41) | < 0.001*** | ||
| Preeclampsia (Yes) | 2.26 (1.10, 4.36) | 0.02* | ||||||
| Birthweight (gm) | 1.00 (1.00, 1.00) | 0.22 | ||||||
| Birthlength (cm) | 0.99 (0.91, 1.06) | 0.71 | ||||||
| Variance Inflation Factor (VIF) range | 1.01–1.62 | 1.01–1.71 | 1.02–1.68 | 1.02–1.91 | ||||
| Hosmer Lemeshow Goodness of Fit (GOF) | χ2 = 12.89, p = 0.12 | χ2 = 8.95, p = 0.35 | χ2 = 5.6, p = 0.69 | χ2 = 5.71, p = 0.68 | ||||
Note: Covariables were determined based on forward‐stepwise regression and entered in the model if p < 0.1. Model tested all diet patterns and confounders. The grey blank shaded cell means that specific variable is not included in that specific model.
Abbreviation: (Y), year, (W), week.
p < 0.05.
p < 0.01.
p < 0.001.
For parent‐reported asthma at age 3 years, the plant‐based dietary pattern showed a borderline association (OR = 0.69, 95% CI: 0.44–1.05, p = 0.08), while neither the balanced diet (OR = 1.18, 95% CI: 0.95–1.46, p = 0.12) nor Western pattern (OR = 0.95, 95% CI: 0.75–1.18, p = 0.64) was significantly associated with asthma. Maternal allergy (OR = 1.56, 95% CI: 1.11–2.19, p < 0.01), paternal asthma or wheeze (OR = 1.93, 95% CI: 1.35–2.75, p < 0.001), preeclampsia (OR = 2.26, 95% CI: 1.10–4.36) and child wheeze at Year 1 (OR = 2.96, 95% CI: 1.97–4.41, p < 0.001) were independently associated with parent‐reported asthma. Mother's pre‐pregnancy BMI was inversely associated (OR = 0.93, 95% CI: 0.88–0.97) with parent‐reported asthma at age 3 years. The model fit was acceptable (Hosmer–Lemeshow χ 2 = 5.71, p = 0.68), with no evidence of multicollinearity (VIF range: 1.02–1.91). While higher maternal BMI has often been linked with increased asthma risk in offspring, several studies have reported heterogeneous or null associations depending on asthma phenotype and outcome definition. The inverse association observed in this analysis may therefore reflect differences in asthma classification, residual confounding or population characteristics.
The Cohen's kappa (κ) statistic was calculated to assess agreement between physician‐diagnosed asthma and parent‐reported asthma in children at age 3 years. The unweighted kappa value was κ = 0.37 (95% CI: 0.30–0.43) (Appendix S2), indicating fair agreement between the two measures. According to Landis and Koch's benchmark criteria (Landis and Koch 1977), this level of agreement is considered ‘fair’ (κ = 0.21–0.40). The narrow CI, which does not include zero, confirms statistically significant agreement beyond chance. However, the moderate strength of concordance suggests discrepancies in how asthma is identified by clinicians versus reported by caregivers. These differences may arise from variability in symptom interpretation, diagnostic thresholds or parental awareness of formal diagnoses. The analysis included 2232 participants, ensuring robust statistical power. While the results validate the use of parent‐reported outcomes as a supplementary measure, they underscore the importance of prioritising physician diagnoses for clinical and research accuracy.
3.3. Dietary Patterns and Recurrent Wheeze
For recurrent wheeze at age 1 year, none of the dietary patterns—plant‐based (OR = 0.89, 95% CI: 0.65–1.23, p = 0.49), Western (OR = 1.05, 95% CI: 0.88–1.23, p = 0.60) or balanced (OR = 1.06, 95% CI: 0.89–1.25, p = 0.47)—were significantly associated with the outcome. Maternal wheeze (OR = 1.59, 95% CI: 1.17–2.16, p = 0.003) and maternal allergy (OR = 1.36, 95% CI: 1.03–1.78, p = 0.03) were independently associated with recurrent wheeze. At age 3 years, recurrent wheeze was also not associated with dietary patterns (plant‐based: OR = 0.80, 95% CI: 0.51–1.25, p = 0.33; Western: OR = 0.98, 95% CI: 0.77–1.25, p = 0.92; balanced: OR = 1.07, 95% CI: 0.85–1.34, p = 0.53). However, mother's wheeze (OR = 1.59, 95% CI: 1.05–2.38), child wheeze at Year 1 (OR = 4.19, 95% CI: 2.89–6.02, p < 0.001) and maternal years in Canada (OR = 0.98, 95% CI: 0.96–0.99, p < 0.001) were significant contributors. Pre‐pregnancy BMI was inversely associated (OR = 0.94, 95% CI: 0.89–0.98) with child's wheeze at Year 3. Both models showed good fit (Hosmer–Lemeshow χ 2 = 12.89, p = 0.12 for Year 1; χ 2 = 8.95, p = 0.35 for Year 3), with no evidence of multicollinearity (VIF range: 1.01–1.71).
However, participants included in the regression analyses differed slightly from those excluded due to missing data. Excluded participants had higher mean pre‐pregnancy BMI (excluded vs. included means 25.04 vs. 24.35 respectively, t‐test p = 0.0001) and fewer years living in Canada (excluded vs. included means 26.24 vs. 27.26, t‐test p = 0.01). These differences suggest that missing data may have introduced some selection bias into the complete‐case analyses.
3.4. Composite Outcome (Sensitivity Analysis)
As a sensitivity analysis, a composite outcome combining physician‐diagnosed asthma and recurrent wheeze at age 3 years was examined (Appendix S3). The overall pattern of results was consistent with the primary analyses. Maternal dietary patterns (plant‐based, Western and balanced) were not significant predictors of the composite outcome. In contrast, prior child wheeze at Year 1 remained the strongest predictor (OR = 5.19, 95% CI: 3.57–7.52, p < 0.001), while gestational age (OR = 0.85, 95% CI: 0.76–0.96, p = 0.005), pre‐pregnancy BMI (OR = 0.95, 95% CI: 0.91–0.99, p = 0.025), and years in Canada (OR = 0.98, 95% CI: 0.96–0.99, p = 0.001) were inversely associated with the outcome. These findings are consistent with the direction and magnitude of effects observed in the individual outcome models, supporting the robustness of the primary results.
3.5. GEE Model for Recurrent Wheeze
The GEE model (Table 4), accounting for repeated measures of recurrent wheeze at Years 1 and 3, included 1574 clusters with a maximum cluster size of 2. Maternal dietary patterns—plant‐based (OR = 0.90, 95% CI: 0.63–1.27, p = 0.54), Western (OR = 1.02, 95% CI: 0.87–1.20, p = 0.81) and balanced (OR = 1.07, 95% CI: 0.90–1.26, p = 0.45)—were not significantly associated with childhood wheeze. However, maternal hypertension (OR = 0.24, 95% CI: 0.06–0.95, p = 0.04), wheeze (OR = 1.41, 95% CI: 1.01–1.97, p = 0.04) and allergy (OR = 1.53, 95% CI: 1.16–2.02, p < 0.001) were independently associated with wheeze across follow‐up. Father's asthma or wheeze also showed a marginal association (OR = 1.32, 95% CI: 0.97–1.79, p = 0.05). The model demonstrated good fit (Hosmer–Lemeshow χ 2 = 7.3, p = 0.5), with no evidence of multicollinearity (VIF range: 1.03–1.79).
TABLE 4.
GEE model results for Children's Wheeze and Asthma. We converted the data set to long format for repeated measures of Recurrent Wheeze of children at year 1 and year 3 and fit the GEE model for correlated data. Number of clusters: 1574; Maximum cluster size: 2.
| Exposure | Wheeze of Children | |
|---|---|---|
| OR (95% CI) | p | |
| Plant based diet | 0.90 (0.63, 1.27) | 0.54 |
| Western diet | 1.02 (0.87, 1.20) | 0.81 |
| Health‐conscious diet | 1.07 (0.90, 1.26) | 0.45 |
| Mother's HTN (Yes) | 0.24 (0.06, 0.95) | 0.04* |
| Mother's ethnicity (Non‐White) | 0.95 (0.68, 1.32) | 0.75 |
| Living area (Urban) | 0.71 (0.44, 1.15) | 0.17 |
| Mother's wheeze (Yes) | 1.41 (1.01, 1.97) | 0.04* |
| Mother's diagnosed asthma (Yes) | 1.30 (0.92, 1.84) | 0.14 |
| Mother's allergy (Yes) | 1.53 (1.16, 2.02) | < 0.001** |
| Father's Asthma or Wheeze (Yes) | 1.32 (0.97, 1.79) | 0.08 |
| Gestational age (W) | 0.94 (0.85, 1.03) | 0.17 |
| Pre pregnancy BMI | 0.98 (0.95, 1.01) | 0.11 |
| Mother's smoking during pregnancy (Yes) | 1.03 (0.64, 1.67) | 0.90 |
| Mother's Years In Canada | 0.99 (0.98, 1.01) | 0.27 |
| Ever breastfed during Year 1 (Yes) | 0.86 (0.40, 1.86) | 0.71 |
| Time | 0.94 (0.84, 1.04) | 0.23 |
| Variance inflation factor (VIF) range | 1.03–1.79 | |
| Hosmer and Lemeshow Goodness of Fit (GOF) test | χ2 = 7.3, p = 0.5 | |
p < 0.05.
p < 0.01.
4. Discussion
This study investigated the association between maternal dietary patterns during pregnancy and childhood asthma and recurrent wheeze in the Canadian CHILD cohort study. The findings revealed no significant associations between plant‐based, Western or balanced dietary patterns and physician‐diagnosed asthma or recurrent wheeze at ages 1 or 3 years. However, maternal and paternal asthma/allergy history, gestational age and maternal hypertension emerged as significant predictors. These results contribute to the growing body of literature on developmental origins of asthma and highlight the complex interplay of genetic, environmental and perinatal factors in shaping respiratory outcomes.
4.1. Dietary Patterns
The absence of statistically significant associations between maternal dietary patterns and childhood asthma or wheeze in this study is consistent with the heterogeneity observed across prior epidemiologic evidence. While some cohort studies have reported protective associations between adherence to a Mediterranean‐style diet during pregnancy and reduced risk of childhood wheeze or asthma (Chatzi et al. 2008), other studies and meta‐analyses have found inconsistent or null associations across populations (Lv et al. 2014). These discrepancies likely reflect differences in dietary assessment methods, population characteristics and the underlying conceptualisation of dietary exposures (Fitzsimon et al. 2007) (Cespedes and Hu 2015).
A key consideration in interpreting our findings is that the dietary patterns derived in this study are not directly equivalent to Mediterranean diet constructs. The plant‐based pattern identified through PCA was characterised by higher intakes of legumes, whole grains, fruits, vegetables and plant‐based dishes, while the balanced pattern included a combination of fish, fruits, vegetables, and mixed protein sources (Appendix S1). Although these features partially overlap with components of the Mediterranean diet, important defining elements—such as a high monounsaturated‐to‐saturated fat ratio (driven largely by olive oil), moderate alcohol intake and specific dietary scoring thresholds—were not explicitly captured in our PCA‐derived patterns (Trichopoulou et al. 2003). Moreover, the balanced pattern also included heterogeneous elements such as refined grains and condiments, while the plant‐based pattern did not reflect key lipid‐related exposures central to Mediterranean dietary mechanisms.
This distinction is critical because PCA‐derived dietary patterns are inherently data‐driven and population‐specific, reflecting prevailing consumption behaviours rather than predefined nutritional constructs (Hu 2002). As such, they may represent different combinations of foods and nutrients—and consequently, different biological pathways—compared to hypothesis‐driven dietary indices like the Mediterranean diet score. For example, anti‐inflammatory and anti‐oxidant effects often attributed to Mediterranean diets may be mediated through specific fatty acid profiles and micronutrient interactions that are not fully represented in our identified patterns (Fardet and Rock 2014). Therefore, the lack of association observed in our study should not be interpreted as directly confirming or refuting findings from Mediterranean diet research, but rather as highlighting the importance of dietary pattern definition in shaping observed associations.
The plant‐based pattern in our study emphasised fruits, vegetables and whole grains—components often linked to anti‐inflammatory effects (Julia et al. 2015). However, its lack of association with asthma may reflect residual confounding, or exposure misclassification of mother's education, family income or postnatal dietary influences, which were not fully accounted for. Similarly, the Western pattern, characterised by processed foods, showed no adverse effects, contrasting with studies linking processed meat intake to childhood wheeze (Calvani et al. 2006). This discrepancy may arise from the holistic nature of dietary patterns, where synergistic or antagonistic nutrient interactions could dilute individual food effects (Fardet and Rock 2014).
Overall, these findings underscore the methodological and conceptual challenges inherent in dietary pattern analysis and suggest that differences in how dietary exposures are defined and operationalised may substantially influence observed associations with childhood respiratory health outcomes.
4.2. Role of Familial and Perinatal Factors
The strong association between maternal diagnosed asthma (OR = 2.16) and childhood asthma underscores the heritable nature of respiratory diseases. This aligns with twin studies estimating that 35%–95% of asthma risk is attributable to genetic factors (Devereux et al. 2006). Paternal asthma/wheeze also significantly increased asthma risk (OR = 2.14), reinforcing the importance of familial atopy in developmental programming (Martino and Prescott 2011). Maternal allergy, particularly, was a robust predictor of both asthma (OR = 1.70) and recurrent wheeze (OR = 1.49), suggesting shared immunological pathways, such as transplacental transfer of IgE or epigenetic modifications (Prescott 2003).
Gestational age emerged as a protective factor (OR = 0.79 for asthma), consistent with evidence linking prematurity to impaired lung development and heightened airway inflammation (Lawson et al. 2017). Maternal hypertension, inversely associated with wheeze in GEE models (OR = 0.32), may reflect understudied placental adaptations or antihypertensive medication use, though further mechanistic studies are needed.
4.3. Implications for Public Health and Clinical Practice
While maternal data‐driven dietary patterns diet did not directly predict recurrent wheeze (Lange et al. 2010) or asthma, the prominence of familial atopy and perinatal factors highlights actionable targets for risk factors. Clinicians should prioritise identifying high‐risk families and advocating for early interventions, such as immunomodulatory therapies. The borderline association of the balanced diet with reduced self‐reported asthma (OR = 0.77, p = 0.20) warrants further exploration, because diverse diets may mitigate oxidative stress or modulate the gut microbiome—a known influencer of immune maturation (Zulyniak et al. 2017).
4.4. Strengths and Limitations
This study's strengths include its large, population‐based cohort, longitudinal design and adjustment for numerous confounders. The use of PCA‐derived dietary patterns aligns with contemporary nutritional epidemiology, which emphasises holistic dietary approaches over single‐nutrient analyses (Hu 2002). However, several limitations should be considered. First, FFQs are prone to recall bias and may not capture precise dietary exposures. Second, a considerable proportion of observations were excluded from the regression analyses due to missing data. To assess potential selection bias, characteristics of participants included in the final models were compared with those excluded, and only modest differences were observed in a few demographic variables, suggesting that the impact of missing data on the overall findings was limited. Third, residual confounding from unmeasured factors (e.g., paternal diet, postnatal feeding practices) could obscure associations. Additionally, the FFQ did not capture information on food sourcing, such as geographic origin, agricultural practices or transit time, which may influence nutrient composition and exposure to environmental factors, potentially contributing to measurement error. Fourth, the observational design precludes causal inferences. Finally, the predominance of Caucasian participants may limit generalisability to more diverse populations.
4.5. Future Directions
Future research should explore longitudinal dietary patterns beyond pregnancy, including postnatal maternal and child diets, to assess cumulative effects. Investigating gene‐diet interactions, particularly in high‐risk families, could uncover subgroups benefiting from dietary interventions. Mechanistic studies on maternal‐foetal immune crosstalk, such as placental cytokine profiles or epigenetic markers, may elucidate pathways linking maternal asthma/allergy to childhood outcomes. Additionally, incorporating objective biomarkers (e.g., vitamin D, omega‐3 levels) could refine dietary assessments and validate FFQ‐derived patterns (Allan and Devereux 2011).
In conclusion, this study found no evidence that maternal dietary patterns during pregnancy influence childhood asthma or recurrent wheeze. Instead, familial atopy, gestational age and maternal health conditions emerged as predominant predictors. These findings underscore the multifactorial aetiology of asthma and the need for integrated prevention strategies targeting genetic, environmental and perinatal risk factors. While diet remains a cornerstone of prenatal care, its role in asthma prevention may be secondary to addressing familial and immunological determinants.
Author Contributions
Fazle Rabbi: conceptualization, data curation, formal analysis, investigation, methodology, writing – original draft. Russell J. de Souza: conceptualization, data curation, formal analysis, investigation, methodology, supervision, writing – original draft, writing – review and editing. Pinaz Gulacha: formal analysis, writing – original draft. Lehana Thabane: formal analysis, investigation, methodology. Sonia S. Anand, Kozeta Miliku, Theo Moraes, Elinor Simons, Padmaja Subbarao: writing – review and editing.
Funding
Padmaja Subbarao holds a Canada Research Chair, Tier 1 in Pediatric Asthma and Lung Health. The CHILD study received funding from the Schroeder Foundation, Don & Debbie Morrison, the Canadian Institutes of Health Research, and Allergy, Genes and Enviroment Network.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Appendix S1: The principal component analysis (PCA) for CHILD FFQ.
Appendix S2: The Cohen's kappa (κ) statistic to assess agreement between physician‐diagnosed asthma and parent‐reported asthma in children at age 3 years.
Appendix S3: Sensitivity analysis using composite outcome for Wheeze at Year 3 and Asthma at Year 3.
Acknowledgements
We gratefully acknowledge the participants and research team whose contributions made this study possible.
Data Availability Statement
Data described in the manuscript are available by registration to the CHILD database https://childstudy.ca/childdb/ and the submission of a formal request. More information about data access for the CHILD cohort study can be found at https://childstudy.ca/for‐researchers/data access/. Researchers interested in accessing CHILD cohort study data for their research should contact child@mcmaster.ca.
References
- Allan, K. , and Devereux G.. 2011. “Diet and Asthma: Nutrition Implications From Prevention to Treatment.” Journal of the American Dietetic Association 111: 258–268. 10.1016/j.jada.2010.10.048. [DOI] [PubMed] [Google Scholar]
- Calvani, M. , Alessandri C., Sopo S. M., et al. 2006. “Consumption of Fish, Butter and Margarine During Pregnancy and Development of Allergic Sensitizations in the Offspring: Role of Maternal Atopy.” Pediatric Allergy and Immunology 17: 94–102. 10.1111/j.1399-3038.2005.00367.x. [DOI] [PubMed] [Google Scholar]
- Cespedes, E. M. , and Hu F. B.. 2015. “Dietary Patterns: From Nutritional Epidemiologic Analysis to National guidelines2.” American Journal of Clinical Nutrition 101: 899–900. 10.3945/ajcn.115.110213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chatzi, L. , Torrent M., Romieu I., et al. 2008. “Mediterranean Diet in Pregnancy Is Protective for Wheeze and Atopy in Childhood.” Thorax 63: 507–513. 10.1136/thx.2007.081745. [DOI] [PubMed] [Google Scholar]
- , de Souza, R. J. , Zulyniak M. A., Desai D., et al. 2016. “Harmonization of Food‐Frequency Questionnaires and Dietary Pattern Analysis in 4 Ethnically Diverse Birth Cohorts123.” Journal of Nutrition 146: 2343–2350. 10.3945/jn.116.236729. [DOI] [PubMed] [Google Scholar]
- Devereux, G. , Turner S. W., Craig L. C. A., et al. 2006. “Low Maternal Vitamin E Intake During Pregnancy Is Associated With Asthma in 5‐Year‐Old Children.” American Journal of Respiratory and Critical Care Medicine 174: 499–507. 10.1164/rccm.200512-1946OC. [DOI] [PubMed] [Google Scholar]
- Fardet, A. , and Rock E.. 2014. “Toward a New Philosophy of Preventive Nutrition: From a Reductionist to a Holistic Paradigm to Improve Nutritional Recommendations.” Advances in Nutrition 5: 430–446. 10.3945/an.114.006122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fitzsimon, N. , Fallon U., O'Mahony D., et al. 2007. “Mothers' Dietary Patterns During Pregnancy and Risk of Asthma Symptoms in Children at 3 Years.” Irish Medical Journal 100: 27–32. [PubMed] [Google Scholar]
- Harris, P. A. , Taylor R., Minor B. L., et al. 2019. “The REDCap Consortium: Building an International Community of Software Platform Partners.” Journal of Biomedical Informatics 95: 103208. 10.1016/j.jbi.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harris, P. A. , Taylor R., Thielke R., Payne J., Gonzalez N., and Conde J. G.. 2009. “Research Electronic Data Capture (REDCap)—a Metadata‐Driven Methodology and Workflow Process for Providing Translational Research Informatics Support.” Journal of Biomedical Informatics 42: 377–381. 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hu, F. B. 2002. “Dietary Pattern Analysis: A New Direction in Nutritional Epidemiology.” Current Opinion in Lipidology 13: 3–9. [DOI] [PubMed] [Google Scholar]
- Julia, V. , Macia L., and Dombrowicz D.. 2015. “The Impact of Diet on Asthma and Allergic Diseases.” Nature Reviews. Immunology 15: 308–322. 10.1038/nri3830. [DOI] [PubMed] [Google Scholar]
- Landis, J. R. , and Koch G. G.. 1977. “The Measurement of Observer Agreement for Categorical Data.” Biometrics 33: 159–174. 10.2307/2529310. [DOI] [PubMed] [Google Scholar]
- Lange, N. E. , Rifas‐Shiman S. L., Camargo C. A., Gold D. R., Gillman M. W., and Litonjua A. A.. 2010. “Maternal Dietary Pattern During Pregnancy Is Not Associated With Recurrent Wheeze in Children.” Journal of Allergy and Clinical Immunology 126: 250–255.e4. 10.1016/J.JACI.2010.05.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lawson, J. A. , Brozek G., Shpakou A., et al. 2017. “An International Comparison of Asthma, Wheeze, and Breathing Medication Use Among Children.” Respiratory Medicine 133: 22–28. 10.1016/j.rmed.2017.11.001. [DOI] [PubMed] [Google Scholar]
- Lv, N. , Xiao L., and Ma J.. 2014. “Dietary Pattern and Asthma: A Systematic Review and Meta‐Analysis.” Journal of Asthma and Allergy 7: 105–121. 10.2147/JAA.S49960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martino, D. , and Prescott S.. 2011. “Epigenetics and Prenatal Influences on Asthma and Allergic Airways Disease.” Chest 139: 640–647. 10.1378/chest.10-1800. [DOI] [PubMed] [Google Scholar]
- Masoli, M. , Fabian D., Holt S., Beasley R., and Program, Global Initiative for Asthma (GINA) . 2004. “The Global Burden of Asthma: Executive Summary of the GINA Dissemination Committee Report.” Allergy 59: 469–478. 10.1111/j.1398-9995.2004.00526.x. [DOI] [PubMed] [Google Scholar]
- Miliku, K. , Reyna M. E., Medeleanu M., et al. 2024. “From Fetus to Eight: The CHILD Cohort Study.” American Journal of Epidemiology 194: 1984–1998. 10.1093/aje/kwae397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Polisena, J. , Tam S., Lodha A., Laporte A., Coyte P. C., and Ungar W. J.. 2007. “An Economic Evaluation of Asthma Action Plans for Children With Asthma.” Journal of Asthma 44: 501–508. 10.1080/02770900701495728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Prescott, S. L. 2003. “Early Origins of Allergic Disease: A Review of Processes and Influences During Early Immune Development.” Current Opinion in Allergy and Clinical Immunology 3: 125–132. 10.1097/00130832-200304000-00006. [DOI] [PubMed] [Google Scholar]
- Sandström, B. 2001. “Micronutrient Interactions: Effects on Absorption and Bioavailability.” British Journal of Nutrition 85, no. 2: S181–S185. [PubMed] [Google Scholar]
- Subbarao, P. , Anand S. S., Becker A. B., et al. 2015. “The Canadian Healthy Infant Longitudinal Development (CHILD) Study: Examining Developmental Origins of Allergy and Asthma.” Thorax 70: 998–1000. 10.1136/thoraxjnl-2015-207246. [DOI] [PubMed] [Google Scholar]
- Trichopoulou, A. , Costacou T., Bamia C., and Trichopoulos D.. 2003. “Adherence to a Mediterranean Diet and Survival in a Greek Population.” New England Journal of Medicine 348: 2599–2608. 10.1056/NEJMoa025039. [DOI] [PubMed] [Google Scholar]
- Zulyniak, M. , de Souza R., Shaikh M., et al. 2017. “Does the Impact of a Plant‐Based Diet During Pregnancy on Birth Weight Differ by Ethnicity? A Dietary Pattern Analysis From a Prospective Canadian Birth Cohort Alliance.” BMJ Open 7: e017753. 10.1136/bmjopen-2017-017753. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1: The principal component analysis (PCA) for CHILD FFQ.
Appendix S2: The Cohen's kappa (κ) statistic to assess agreement between physician‐diagnosed asthma and parent‐reported asthma in children at age 3 years.
Appendix S3: Sensitivity analysis using composite outcome for Wheeze at Year 3 and Asthma at Year 3.
Data Availability Statement
Data described in the manuscript are available by registration to the CHILD database https://childstudy.ca/childdb/ and the submission of a formal request. More information about data access for the CHILD cohort study can be found at https://childstudy.ca/for‐researchers/data access/. Researchers interested in accessing CHILD cohort study data for their research should contact child@mcmaster.ca.