点击显示 收起
1 From the GSF–National Research Center for Environment and Health, Institute of Epidemiology, Neuherberg, Germany (SS, HEW, and JH); Dr von Hauner Children's Hospital (SK) and the Institute of Medical Data Management, Biometrics and Epidemiology (HEW), Ludwig-Maximilians–University of Munich, Munich, Germany; the Department of Pediatrics, Marien-Hospital Wesel, Wesel, Germany (BS and AvB); the Department of Environmental Immunology, UFZ–Center for Environmental Research, Leipzig, Germany (IL); the Departments of Pediatrics (MB) and Faculty of Medicine, Environmental Hygiene and Environmental Medicine (OH), University of Leipzig, Leipzig, Germany; the Municipal Hospital "St Georg" Leipzig, Children's Hospital, Leipzig, Germany (MB); and the Department of Human Exposure Research and Epidemiology, UFZ–Center for Environmental Research, Leipzig, Germany (OH)
2 Supported by grants 01 EG 9732 and 01 EG 9705/2 from the Federal Ministry for Education, Science, Research, and Technology. 3 Address reprint requests to J Heinrich, GSF–Institute of Epidemiology, Ingolstaedter Landstrasse 1, 85764 Neuherberg, Germany. E-mail: joachim.heinrich{at}gsf.de.
ABSTRACT
Background:Maternal diet during pregnancy might be one of the factors that influences fetal immune responses associated with childhood allergy.
Objective:We analyzed the association between maternal diet during the last 4 wk of pregnancy and allergic sensitization and eczema in the offspring at 2 y of age.
Design:Data from 2641 children at 2 y of age were analyzed within a German prospective birth cohort study (LISA). Maternal diet during the last 4 wk of pregnancy was assessed with a semiquantitative food-frequency questionnaire, which was administered shortly after childbirth.
Results:High maternal intake of margarine [adjusted odds ratio (aOR): 1. 49; 95% CI: 1.08, 2.04] and vegetable oils (aOR: 1.48; 95% CI: 1.14, 1.91) during the last 4 wk of pregnancy was positively associated and high maternal fish intake (aOR: 0.75; 95% CI: 0.57, 0.98) was inversely associated with eczema during the first 2 y in the offspring. High celery (aOR: 1.85; 95% CI: 1.18, 2.89) and citrus fruit (aOR: 1.73; 95% CI: 1.18, 2.53) intakes increased the risk of sensitization against food allergens. In turn, sensitization against inhalant allergens was positively related to a high maternal intake of deep-frying vegetable fat (aOR: 1.61; 95% CI: 1.02, 2.54), raw sweet pepper (aOR: 2.16; 95% CI: 1.20, 3.90), and citrus fruit (aOR: 1.72; 95% CI: 1.02, 2.92).
Conclusions:We suggest that the intake of allergenic foods and foods rich in n–6 polyunsaturated fatty acids during pregnancy may increase and foods rich in n–3 polyunsaturated fatty acids may decrease the risk of allergic diseases in the offspring.
Key Words: Maternal diet • pregnancy • allergic diseases • children
INTRODUCTION
Because prenatal life is a critical period for the development of the immune system, the role of intrauterine exposure in the etiology of allergic diseases has gained interest (1). In this context, maternal diet during pregnancy has been proposed to influence fetal immune responses that might predispose to childhood allergy (2).
During gestation, essential nutrients are transferred from the maternal to the fetal circulation across the placenta (3). Therefore, it might be possible that dietary factors associated with allergic diseases already exert their influence in utero. Complex transport mechanisms have been identified for antioxidants (4) and long-chain polyunsaturated fatty acids (5). Furthermore, it has been suggested that food and inhalant allergens ingested or inhaled by the mother cross the placental barrier (6).
Whether maternally derived allergens or immunogenic nutrients lead to an early specific sensitization in the fetus or induce allergen-specific tolerance is controversial (2, 7). Postnatally, a high dietary intake of n–6 polyunsaturated fatty acids (PUFAs) has been shown to be associated with an increased risk of allergic diseases because of their proinflammatory properties, whereas n–3 PUFAs and dietary antioxidants are supposed to have a protective effect on asthma and allergies (8).
Although there is a biological basis for a prenatal effect of dietary factors on the development of allergic diseases, the number of studies investigating this subject is very limited. The aim of the present study was to analyze prospectively in a large cohort whether maternal diet during the last 4 wk of pregnancy is associated with allergic sensitization or eczema in the offspring at 2 y of age.
SUBJECTS AND METHODS
Subjects
We analyzed data from the LISA (Influences of Lifestyle-related Factors on the Immune System and the Development of Allergies in Childhood) Study. The design and objective of this prospective birth cohort study were described in detail elsewhere (9). In brief, 3097 newborns were initially recruited between November 1997 and January 1999 from 4 German cities: Munich, Leipzig, Wesel, and Bad Honnef. Questionnaire data on family history of atopy, parental education, smoking during pregnancy, and maternal diet during the last 4 wk of pregnancy were obtained shortly after delivery (median: 3 d). We collected data on the children's health and on lifestyle factors by using repeated parental-completed questionnaires at regular time intervals during the first 2 y of life (ie, at 6, 12, 18, and 24 mo of age). At 2 y of age, blood samples were drawn for total and specific immunoglobulin E (IgE) analysis.
For the present analysis, we selected subjects who had participated in the follow-up at 2 y, excluding children with chronic diseases (eg, celiac disease and metabolic disorders; n = 9) and children for whom no information on their mothers' diet during the last 4 wk of pregnancy was available (n = 14). Thus, the final study population consisted of 2641 children. The number of participants in each follow up period and the procedure used to select the final study population are shown in Figure 1. The local ethics committees approved the study protocol, and informed consent was obtained from the parents.
FIGURE 1.. Flow chart showing the number of participants throughout the study. FFQ, food-frequency questionnaire.
Dietary assessment
We assessed maternal food intake during the last 4 wk of pregnancy by using a semiquantitative food-frequency questionnaire (FFQ) administered shortly after childbirth. For each food item, the mothers reported their average consumption frequency over the past 4 wk according to 5 categories ranging from "< 2 times/mo or never" to "4 times/wk." Quantitative information was gathered on milk and yogurt intake only according to the following 4 categories: "never," "sometimes," "up to 0.5 L/d," and ">0.5 L/d" for milk intake and "never," "sometimes," "up to 200 g/d," and ">200 g/d" for yogurt intake. The original wording of the 26 food items selected for the present analysis is shown in Appendix A.
View this table:
APPENDIX A. Categories derived from the original food-frequency questionnaire (FFQ)
To estimate the effect of high compared with low maternal food intake during pregnancy on allergic outcomes, the children were categorized into 2 groups according to maternal food consumption frequency. Therefore, each food-frequency variable was dichotomized at a cutoff close to the 66th percentile. In this way, we contrasted the upper tertile (tertile 3) with the combination of the lowest and the middle tertiles (tertile 1 + tertile 2) for most of the food variables (cheese, cream, margarine, vegetable oils, seeds, deep-frying vegetable fat, nuts, fish, cabbage, celery, raw tomatoes, salad, vegetable juice, citrus fruit, apples, and strawberries). For some foods, considerably >33% of all subjects were in the highest consumption category; therefore, this procedure was not feasible. The cutoff was set at the 33rd percentile, and the combined middle and upper tertiles (tertile 2 + tertile 3) was contrasted with the lowest tertile (tertile 1) (milk, yogurt, eggs, butter, raw carrots, spinach, raw sweet pepper, exotic fruit, pineapple, bananas, and fruit juice). Thus, we used different reference categories for the logistic regression models.
Depending on the intake distribution of each food item, high maternal intake referred to different consumption frequencies. The high intake category included subjects who consumed foods "2–3 times/mo" (deep-frying vegetable fat, spinach, celery, raw sweet pepper, and vegetable juice), "1–2 times/wk" (eggs, seeds, nuts, fish, raw carrots, cabbage, and strawberries), "3–4 times/wk" (cream, butter, vegetable oils, raw tomatoes, salad, citrus fruit, exotic fruit, and fruit juice), "4 times/wk" (cheese, margarine, apples, and bananas), or more than "sometimes" (milk and yogurt). A similar categorization strategy was recently described (10).
Outcome definition
Doctor-diagnosed eczema was based on a positive answer to the question, "Has a doctor diagnosed your child with allergic or atopic eczema in the past 6 mo? " Lifetime prevalence of doctor-diagnosed eczema was assumed if eczema has ever been diagnosed during the first 2 y of life.
Total and specific serum immunoglobulin E (IgE) concentrations were assayed by using the CAP-RAST FEIA system (Pharmacia Diagnostics, Freiburg, Germany) according to the manufacturer's instructions. The limit of detection for specific IgE was 0.35 kU/L, and values 0.35 kU/L were considered positive. Allergic sensitization against food allergens was defined as a specific serum IgE concentration 0.35 kU/L against pediatric food allergens (fx5) (egg, cow milk, wheat, peanut, soybean, and codfish). A positive result in this screening test was followed by measurement of the single allergens egg, cow milk, and peanut. Allergic sensitization against inhalant allergens was defined as a specific serum IgE concentration 0.35 kU/L against at least one of the following allergen mixes: house dust allergens (hx2) (Dermatophagoides pteronyssinus, Dermatophagoides farinae, German cockroach, and house dust), cat dander (e1), mixed molds (mx1) (Penicillium notatum, Cladosporium herbarum, Aspergillus fumigatus, and Alternaria alternata), or seasonal allergens (rx1) (timothy grass, mugwort, English plantain, ribwort, wall pellitory, and birch pollen).
Statistical methods
Differences in the prevalence of eczema and allergic sensitization between the 2 categories of maternal food intake were tested by chi-square test. We further applied multiple logistic regression analysis to estimate the association between maternal food intake during the last 4 wk of pregnancy and eczema and allergic sensitization in their offspring. Odds ratios (ORs) and the corresponding 95% CIs were computed. First, we examined the crude association of maternal food intake with eczema and allergic sensitization. Then, we calculated an adjusted model that included study area (Munich, Leipzig, Wesel, and Bad Honnef), sex, maternal age at delivery (31 or >31 y), maternal smoking during second or third trimester of pregnancy, level of parental education (very high, high, medium, and low), exclusive breastfeeding for 4 mo, parental history of atopic diseases (asthma, hay fever, or eczema; no parents atopic, one parent atopic, and both parents atopic), season of birth, and all dietary variables. Calculation of the variance inflation factor (11) did not indicate strong collinearity between dietary variables in the multiple logistic regression models.
We also tested for interactions between maternal food intake and maternal history of atopic diseases and duration of exclusive breastfeeding. All computations were performed by using the statistical analysis package SAS for WINDOWS (version 8.2; SAS Institute, Cary, NC). Two-sided P values <0.05 were considered statistically significant for all analyses.
RESULTS
The frequency of basic characteristics and possible confounding factors in our study population is shown in Table 1. Subjects were excluded when their mothers had completed the FFQ on maternal diet during pregnancy but did not return the 2-y questionnaire (n = 427). Compared with those, children of respondent mothers more likely lived in Munich (50.0% compared with 31.9%; P < 0.001) and were born to older mothers (31.3 y compared with 28.9 y; P < 0.001). They were less likely to have a smoking mother during pregnancy (10.1% compared with 24.2%; P < 0.001) and 2 older siblings (9.8% compared with 13.4%; P = 0.033), but were more likely to have parents with a very high level of education (54.6% compared with 31.8%; P < 0.001), at least one parent with atopic disease (52.7% compared with 43.0%; P < 0.001), and to have been breastfed for 4 mo (57.6% compared with 43.5%; P < 0.001). No statistically significant differences were observed between the responders and the excluded children with respect to sex (P = 0.375), birth weight (P = 0.713), or season of birth (P = 0.174).
View this table:
TABLE 1. Basic characteristics of the LISA study population
The prevalences of eczema and allergic sensitization in the study population are shown in Table 2. At 2 y of age, 17.7% of all children had had doctor-diagnosed eczema, 9.3% were sensitized against food allergens, and 4.8% were sensitized against inhalant allergens. Of those sensitized against food allergens, milk (5.1%) and egg sensitization (5.4%) was most common. Sensitization to inhalant allergens was mainly ascribed to house dust allergens (2.8%).
View this table:
TABLE 2. Prevalence of eczema and allergic sensitization in the LISA study population at 2 y of age
Maternal diet during the last 4 wk of pregnancy turned out to be related to study area, maternal age at delivery, maternal smoking during pregnancy, level of parental education, and duration of exclusive breastfeeding. Only some foods, such as milk, cream, eggs, spinach, cabbage, vegetable juice, bananas, and strawberries, seemed to be largely independent of these factors.
The prevalences of eczema and allergic sensitization according to the intake categories for each food are shown in Table 3. Children of mothers with a high intake of margarine, vegetable oils, and seeds during pregnancy had a statistically significant higher lifetime prevalence of doctor-diagnosed eczema at 2 y of age than did children whose mothers had a low intake of these foods. Allergic sensitization against food allergens was more prevalent in children whose mothers had a high intake of celery and citrus fruit during pregnancy, whereas children sensitized against inhalant allergens were more frequently born to mothers with a high intake of raw sweet peppers.
View this table:
TABLE 3. Prevalence of eczema and allergic sensitization in the children at 2 y of age according to intake categories of maternal food intake during pregnancy
In the logistic regression analysis, the results did not differ substantially between the crude and adjusted models. After adjustment for all potential confounders and dietary variables, the positive association between high maternal seed intake during the last 4 wk of pregnancy and eczema observed in the crude model attenuated to nonsignificance, whereas high fish intake showed a statistically significant protective effect on eczema for the first time. All other ORs became stronger, but the significance level did not change. The results of the adjusted model are presented in Table 4. A high maternal intake of margarine and vegetable oils during pregnancy was positively associated with doctor-diagnosed eczema in their offspring, and a high maternal fish intake had a protective effect. The risk of allergic sensitization against food allergens in children increased when the maternal intake of celery and citrus fruit was high. In turn, sensitization against inhalant allergens was positively related to a high maternal intake of deep-frying vegetable fat, raw sweet peppers, and citrus fruit.
View this table:
TABLE 4. Logistic regression results for the association between maternal food intake during pregnancy and eczema and allergic sensitization in the offspring at the age of 2 y1
We further analyzed the relation between maternal intake of dairy products during pregnancy and sensitization against cow milk allergen in the offspring (data not shown). Children of mothers with a high cream intake had an increased risk of cow milk sensitization at 2 y of age (aOR: 1.62; 95% CI: 1.07, 2.45). No association between maternal egg intake during pregnancy and sensitization against egg allergen in the offspring could be observed.
Formal testing for interaction between maternal diet during pregnancy and maternal history of atopic diseases and time of exclusive breastfeeding was not statistically significant (data not shown).
DISCUSSION
The results of this prospective study suggest that maternal diet during the last 4 wk of pregnancy has an effect on the development of allergic diseases in the offspring. In particular, high maternal intakes of margarine, vegetable oils, deep-frying vegetable fat, celery, citrus fruit, and raw sweet peppers were positively associated with either eczema or allergic sensitization in children at 2 y of age. In turn, high fish consumption during the last 4 wk of pregnancy seemed to decrease the risk of eczema in childhood.
These findings are consistent with the hypothesis that the development of allergic diseases in childhood can be affected by intrauterine exposure to maternally derived allergens or proinflammatory factors modulating fetal immune responses (7). Margarine and vegetable oils with a high content of n–6 PUFAs might be responsible for the observed positive association between maternal intake of these fats and eczema in their offspring (12). This assumption is supported by the results of several studies reporting that margarine intake has an adverse effect on allergic diseases in children and adults (13-16). In contrast, there is evidence for a protective effect of fish intake on asthma and other allergic diseases (17-19), possibly because of the antiinflammatory properties of n–3 PUFAs contained in oily fish. An association between celery, citrus fruit, or raw sweet pepper intake and allergic sensitization has not yet been described in the literature. Because potential allergens in celery (20), citrus fruit—in particular oranges (21, 22)—and bell peppers (23) have been characterized, the observed associations might be linked to the allergenicity of these foods. However, the high antioxidant content of these foods should also be considered, although it seems unlikely that they play a causal role. We further cannot completely rule out that the findings either occurred by chance or are confounded by lifestyle factors, which were not assessed in this study. Therefore, prospective interventional trials are necessary to confirm these new observations.
Only a small number of studies have investigated the effect of maternal diet during pregnancy on the development of allergic diseases in the offspring. The role of early dietary modification on immune responses in children at high risk of atopy has been studied in a randomized placebo-controlled clinical trial including 83 atopic pregnant women. Subjects in the intervention group received fish-oil capsules containing n–3 PUFAs from 20 wk gestation. Neonates whose mothers took fish-oil supplements during pregnancy had significantly lower interleukin 13 concentrations in their cord blood than did the control group (24). Follow-up of the study further showed that infants in the intervention group were less likely to be sensitized to egg allergen and had significantly less severe atopic dermatitis at 1 y of age (25). In a nested case-control study including 691 schoolchildren, maternal fish consumption during pregnancy was assessed by retrospective telephone interviews. Oily a fish intake of at least monthly during pregnancy was significantly associated with a reduced asthma risk in childhood, but only in children born to asthmatic mothers (26). Although the recall of diet over a long period may be inaccurate, these results agree with our observation that a high maternal fish intake during pregnancy reduces the risk of atopic diseases in childhood.
Studies of the effects of maternal dietary allergen avoidance during pregnancy on atopic diseases and sensitization in the children provide conflicting findings (27-31). Although experimental interventions were mainly due to the dietary exclusion of cow milk, eggs, and peanuts, no study aimed to assess a potential reduction in atopic disease through the elimination of celery, citrus fruit, or raw sweet pepper intakes.
Martindale et al (32) prospectively investigated whether maternal antioxidant intake during pregnancy is related to respiratory symptoms and atopic diseases in early childhood. They reported a negative association between maternal vitamin E intake during pregnancy and wheeze in children at 2 y of age. Children, whose mothers were atopic also had a reduced risk of eczema. However, vitamin C intake was positively associated with wheeze and eczema. Unfortunately, we were not able to analyze our data with respect to antioxidant intake because of the limited number of food items included in our FFQ. However, there was no indication that the consumption of foods rich in antioxidants protects against allergic diseases.
Several possible limitations of the study need to be considered. Misclassification of dietary exposure is always a major issue in epidemiologic studies. We applied a semiquantitative FFQ, which provided qualitative information on maternal consumption frequencies during pregnancy, but no details on usual serving sizes were collected. However, a body of evidence supports the idea that the increment of additional quantitative information on diet is extremely small (33). This has been attributed to the fact that frequency explains most of the variation in total food intake and that intraindividual variability in serving sizes is generally greater than interindividual variability. Anyway, random misclassification of exposure would tend to bias effect estimates toward the null value. Therefore, misclassification bias is unlikely to affect significant associations. Furthermore, we cannot completely rule out that the reported associations were modified by the diet of the children during the first 2 y of life. We estimated the correlation between maternal consumption frequencies of fish, celery, and citrus fruit during pregnancy and the time of introduction of these foods to the infant's diet during the first year of life. The correlation coefficients did not indicate a statistically significant association, but it might be possible that maternal food consumption frequencies are a surrogate marker of infant food consumption frequencies during the first 2 y of life. This would indicate that our findings might be at least partly a consequence of postnatal dietary influences. However, interventional trials are necessary to disentangle maternal diet from the diet of the child. Because many food items were grouped together in one category, it was not possible to evaluate the health effect of each single food, such as different kinds of nuts. Potential antagonistic effects of these foods might have masked some associations that might be especially relevant for allergenic foods. Furthermore, simultaneous adjustment for all dietary variables in the multiple logistic regression models might raise concern of overadjustment. We did not observe any substantial differences between the crude and adjusted models. Thus, it is unlikely that overadjustment affected our findings. One might also speculate that atopic mothers alter their diet during pregnancy and avoid known food allergens to prevent the onset of allergic diseases in their offspring. In this case, reverse causation might have been responsible for the observed associations. However, testing for interactions between food intake and maternal history of atopic diseases did not confirm the assumption that food intake during pregnancy is related to a higher risk of allergic diseases in the children of atopic mothers only. Because of a lack of statistical power, this study could not provide valid data to identify more specific associations between single foods and specific food sensitization. The possibility that significant associations occurred by chance as a result of multiple testing should also be kept in mind; however, because the observed significant effects mainly related to the initially proposed hypothesis, it is unlikely that the results might be explained by chance alone.
One of the major strengths of the present study was the large sample size, which enabled us to detect statistically significant associations, which are biologically plausible but have not been described previously. A prospective study design, as used in the present study, is certainly superior to a cross-sectional design for studying cause-effect relations.
In conclusion, these prospective data from a large cohort study indicate an association between maternal diet during pregnancy, particular foods rich in PUFAs and allergenic foods, and the risk of allergic diseases in the offspring. Currently, no recommendations are being made to mothers to modify their diets to prevent allergies in their children because of insufficient evidence of a beneficial effect. However, before any recommendations can be made, randomized clinical intervention trials should be performed to confirm the cause-effect relation observed in the present study.
ACKNOWLEDGMENTS
We thank all of the families for their participation. The LISA Study Group comprised the following organizations (all located in Germany): GSF–National Research Center for Environment and Health, Institute of Epidemiology, Neuherberg (HE Wichmann, J Heinrich J, G Bolte, P Belcredi, B Jacob, A Schoetzau, M Mosetter, J Schindler, and A Höhnke); Department of Pediatrics, University of Leipzig (M Borte, R Schulz, G Sierig, K Mirow, C Gebauer, B Schulze, and J Hainich); Institute of Clinical Immunology and Transfusion Medicine (U Sack and F Emmrich); Department of Pediatrics, Marien-Hospital Wesel (A von Berg, B Schaaf, C Scholten, and C Bollrath); Department of Human Exposure Research and Epidemiology, Centre for Environmental Research Leipzig-Halle Ltd (O Herbarth, U Diez, I Lehmann, M Rehwagen, and U Schlink); Division of Pediatric Infectious Diseases and Immunology, Ludwig-Maximilians-University Munich (M Weiss and M Albert); Friedrich-Schiller–University Jena, Institute of Clinical Immunology (B Fahlbusch); and the Institute of Occupational, Social and Environmental Medicine (W Bischof and A Koch).
SS was responsible for the data analysis, interpretation of data, and manuscript preparation. SK and IL assisted with the interpretation of results and critical revision of the manuscript. MB, OH, BS, AvB, HEW, and JH were responsible for the data collection and assisted with the interpretation of results. SS and JH developed the design and analysis plan. None of the authors had a conflict of interest.
REFERENCES