Allergic diseases are among the most common chronic diseases, especially in developed countries(Reference Asher, Montefort and Björkstén1). One of the earliest manifestations of allergic diseases is cows’ milk allergy (CMA), which affects 2–6 % of children in Finland(Reference Saarinen, Juntunen-Backman and Järvenpää2,Reference Pyrhönen, Näyhä and Kaila3) .
In developed countries, the dietary intake of n-6 PUFA has increased, and simultaneously, the intake of n-3 PUFA has decreased(Reference Blasbalg, Hibbeln and Ramsden4). This changed ratio of intake of n-6/n-3 PUFA has led to a more pro-inflammatory environment, as the n-6 PUFA are observed to promote allergic inflammation by releasing allergy promoting eicosanoids from the cells, whereas n-3 PUFA reduce the allergic inflammation by damping the release of pro-inflammatory factors from the cells(Reference Miles, Calder and Miles5). Because the fatty acid status of the fetus and the newborn infant may be modulated by maternal fatty acids, and that the development of the immune system starts already in utero, it has been implicated that maternal fatty acid intake may affect the risk of development of allergic diseases in children(Reference Prescott and Dunstan6).
A systematic review of epidemiological studies and meta-analysis of randomised controlled trials (RCT) has suggested a protective role of higher intake of n-3 PUFA and fish during pregnancy for allergic diseases, especially for eczema, wheeze and asthma(Reference Best, Gold and Kennedy7). However, some null results have also been reported(Reference Miyake, Sasaki and Tanaka8–Reference Nwaru, Erkkola and Lumia11), and one of the studies had used sensitisation for cows’ milk protein as an outcome(Reference Nwaru, Ahonen and Kaila10). The RCT of the role of maternal supplementation of n-3 PUFA in pregnancy have reported both inverse(Reference Furuhjelm, Warstedt and Fagerås12) and null results(Reference Dunstan, Mori and Barden13,Reference Palmer, Sullivan and Gold14) in relation to the risk of food allergy in the offspring. However, none of these RCT used CMA as an outcome. Thus, the role of maternal fatty acid intake during pregnancy for the development of CMA in the offspring remains unclear.
The aim of our study was to examine the maternal intake of different fatty acids during pregnancy and the development of CMA in the offspring. To our knowledge, this is the first study to assess this association. We hypothesised that maternal higher intake of n-3 PUFA decreases and higher intake of n-6 PUFA increases the risk of CMA in the offspring.
Methods
Study population
We obtained the data from the Finnish Type 1 Diabetes Prediction and Prevention (DIPP) nutrition study, which is a multidisciplinary prospective population-based birth cohort study done in a framework of the ongoing DIPP study. The DIPP study carried out in Finland in the area of Turku, Tampere and Oulu University Hospitals aimed at generating novel insights into the pathogenesis of type 1 diabetes(Reference Kupila, Muona and Simell15). All newborn infants in these areas have been invited for screening of their human leucocyte antigen -conferred susceptibility for type 1 diabetes. Children carrying genotypes conferring high or moderate disease risk (14 % of the infants) are invited to enrol in the follow-up study. Children with severe congenital abnormalities or diseases, whose parents were of non-Caucasian origin, or did not understand Finnish, English or Swedish, were excluded from the DIPP study.
The DIPP nutrition study is conducted among children born in the area of Oulu and Tampere University Hospitals. The present study comprises children born between August 1997 and September 2004 (n 6288), with available personal identity code. Altogether, 4921 children had information on maternal diet during pregnancy and child’s CMA by the age of 3 years (Fig. 1, Tables 1–3). In addition to the dietary data, we collected information on parental history of allergic rhinitis or asthma from the Asthma and Allergy sub-study, where all children still in follow-up at the age of 5 years (n 4075) were invited to participate (n 3781). Altogether, 2327 children had information on maternal diet during pregnancy, basic background factors, child’s CMA by the age of 3 years and parental history of allergic rhinitis or asthma (Table 3:2. Adjustment column).
Fig. 1. Flow chart of the participants. CMA, cows’ milk allergy.
Table 1. Distribution of background characteristics among the whole study population (n 4921) and cases with cows’ milk allergy
(Numbers and percentages, n 448)
* Comparison done with χ 2 test, comparing the distribution of the cows’ milk allergy across the categories.
† Information collected only from the children participating in the Asthma and Allergy sub-study at the age of 5 years.
Table 2. Maternal daily intake of fatty acids during pregnancy from diet and supplements together
(Mean values and standard deviations, n 4921)
Table 3. Associations between maternal daily intake of energy-adjusted fatty acids during pregnancy with the risk of cows’ milk allergy in the offspring by the age of 3 years. The OR are presented per 1 standard deviation increment of the particular fatty acid. Both the dietary intake and total intake (dietary + supplement) of each fatty acid were analysed if the supplemental intake was meaningful
(Odd ratios and 95 % confidence intervals)
* Number of children with cows’ milk allergy.
† Number of children in the analysis.
‡ Adjusted for study centre, sex, birth weight, maternal age and education, maternal smoking during pregnancy, duration of gestation, mode of delivery, season of birth, number of mother’s previous deliveries, length of breast-feeding and urbanity of the living environment.
§ Adjusted additionally for the parental history of allergic rhinitis and asthma, pets inside home during the child’s first year of life and visits to the stable during the child’s first year of life. This additional information was collected when the child was 5 years old.
The present study was conducted according to the guidelines of the Declaration of Helsinki, and all procedures involving patients were approved by the respective ethical committees of the hospital districts of Oulu and Tampere. Written informed consent was obtained from all parents, and they were informed that they can withdraw from the study at any time without further explanations.
Dietary data
The mothers completed a validated 181-item semi-quantitative FFQ , which was designed to reflect the diet during the 8th month of pregnancy (1 month preceding the start of the Finnish maternity leave). The FFQ was validated against 10-d food records(Reference Erkkola, Karppinen and Javanainen16). The FFQ was mailed to the mothers after delivery and returned at the child’s 3-month study visit. The use of food ingredients and dishes was reported as common serving sizes and the frequency of use as no use, daily, weekly or monthly use. The individual variation of used fats in cooking, baking and salad dressings was also queried. Further, the information about the vitamin and mineral supplements used during the whole pregnancy, including supplement’s brand name and the amount of used supplements (tablets, drops, spoonfuls or millilitres), was collected. The FFQ that were filled in inadequately or contained more than ten missing values were excluded (n 53, 1·1 %). The data processing of the FFQ has been described earlier(Reference Erkkola, Karppinen and Javanainen16). Briefly, we double entered the dietary data into the database. We used Finnish Food Composition Database ‘Fineli’(17) and in-house software of the Finnish Institute of Health and Welfare to calculate the estimate of the daily average of the studied fatty acids for each mother.
The fatty acids investigated in the present study were SFA, MUFA, PUFA, n-3 PUFA, n-6 PUFA and trans fatty acids. In addition, we investigated the ratios of n-3 PUFA to n-6 PUFA and linoleic acid to α-linolenic acid. In the validation of the FFQ against two 5-d food records, the Pearson correlations were SFA (0·55), MUFA (0·34), PUFA (0·47), n-3 PUFA (0·39) and n-6 PUFA (0·49)(Reference Erkkola, Karppinen and Javanainen16).
Confounding factors
At the time of enrolment, we asked from the parents their age, occupation, education level and the place of residence. We collected the information on the child’s sex, delivery type, gestational age, pregnancy and delivery complications, birth weight and height, mother’s earlier deliveries and maternal smoking during pregnancy from the Medical Birth Registers of Tampere and Oulu University Hospitals. We asked about breast-feeding from parents at study visits at the child’s age of 6 months, 1, 2 and 3 years. Among children who participated in the Asthma and Allergy sub-study, we collected the information of maternal and paternal history of allergic rhinitis or asthma, pet keeping and contacts to the farm animals during the child’s first year of life by the parental questionnaire completed at the 5-year study visit.
Endpoints
Our definition of CMA was based on information obtained from the special reimbursement register maintained by the Social Insurance Institution of Finland and linked using personal identity codes(Reference Gissler and Haukka18). We complemented the register data with a parental report on CMA queried by the validated questionnaire, which was filled in at study visits, when the child was 6 and 12 months old and annually thereafter, until the age of 9 years(Reference Tuokkola, Kaila and Pietinen19,Reference Tuokkola, Luukkainen and Kaila20) .
From the register, we obtained the information about a valid special reimbursement for the costs of special infant formula needed in the treatment of diagnosed CMA (ICD-10 codes L27.2 or K52.2). In Finland, every child with CMA, diagnosed by a paediatrician, is entitled for this reimbursement up to the age of 2 years, irrespective of the family’s socio-economic status or place of residence. At the time of the present study, the CMA diagnosis in Finland was usually based on an open oral food challenge(Reference Kaila, Vanto and Valovirta21).
Statistical methods
We analysed the difference in background factors between children with and without CMA by the χ² test. We selected the variables used in the adjusted models based on previous knowledge(Reference Vassallo, Banerji and Rudders22–Reference Metsala, Lundqvist and Kaila24) and their association with CMA in the present study. We analysed the association between maternal fatty acid intake and the risk of CMA in the offspring by logistic regression. The possible reliance among siblings was accounted for using the generalised estimating equations with the sandwich estimator of variance to estimate regression coefficients in logistic regression.
After the logarithmic transformation, we adjusted the nutrients for energy intake by the residual method(Reference Willet25) and we used standardised scores in the analyses. We also divided the energy-adjusted dietary intake into quartiles, and the first and last quartiles were compared with the combined mid quartiles in our analysis. We analysed fatty acid intakes from food and from food and supplements, if the supplemental intake was meaningful. Our first adjusted model included study centre, sex, birth weight, maternal age and education, maternal smoking during pregnancy, duration of gestation, mode of delivery, season of birth, number of older siblings, length of breast-feeding and urbanity of the living environment as covariates. We made the unadjusted and the first adjusted model among children with information on maternal diet during pregnancy, child’s CMA status and basic background characteristics (n 4921). We made the second adjusted analysis among children participating in the Asthma and Allergy sub-study (n 2327). The variables used in the second adjusted model were all those in the first adjusted model in addition to maternal and paternal history of allergic rhinitis or asthma, and both visits to a stable and pets inside the home during the study participant’s first year of life. We tested the interaction of maternal history of allergic rhinitis or asthma and fatty acids among the population participating in the Asthma and Allergy sub-study. If the interaction was significant (P-value < 0·05), we studied the association separately among mothers with and without a history of allergic rhinitis or asthma using the fully adjusted model. Missing data were addressed using complete case analysis. SAS version 9.3 (SAS Institute Inc.) and IBM SPSS Statistics for Windows, version 27 (IBM corp.) were used in the analysis.
Results
We identified 448 children with CMA (9·1 %). Children with CMA were more often male, had less often pets inside the home during the first year of life, and their mothers were more often non-smokers during pregnancy and had higher vocational education. Further, their mothers and fathers were more often affected by asthma or allergic rhinitis (Table 1). Children who participated in the Asthma and Allergy sub-study had older, more educated and less frequently smoking mothers when compared with non-participants(Reference Tuokkola, Luukkainen and Kaila26).
The maternal intake of studied fatty acids is shown in Table 2. Maternal total intake of SFA, MUFA, PUFA, n-3 PUFA, n-6 PUFA, trans fatty acids, ratio of n-3 PUFA to n-6 PUFA or ratio of linoleic acid to α-linolenic acid was not associated with the risk of CMA in the offspring (Table 3). When fatty acid intakes and ratios were analysed as quartiles, we did not observe significant associations between either the lower or the higher quartile as compared with the mid-half and CMA (data not shown).
We observed an interaction between maternal history of allergic rhinitis or asthma and the maternal intake of α-linolenic acid (P-value = 0·012). For mothers without a history of allergic rhinitis or asthma (n 1265), α-linolenic acid was associated with a decreased risk (OR 0·72; 95 % CI 0·56, 0·93) of CMA in the offspring (n 237), but not in mothers with a history of allergic rhinitis or asthma (n 1062, OR 1·05; 95 % CI 0·88, 1·26).
Discussion
We did not observe an association between maternal intake of fatty acids and the development of CMA in the offspring. When we took the maternal history of allergic rhinitis or asthma into account, the maternal intake of α-linolenic acid was inversely associated with the risk of CMA in the offspring of the mothers without a history of allergic rhinitis and asthma, but not in mothers with such a history.
To our knowledge, this is the first study to report the association between maternal intake of fatty acids during pregnancy and the development of CMA in the offspring. The associations between maternal intake of fatty acids during pregnancy and development of allergic rhinitis, asthma, eczema and wheeze in the offspring have been previously reported from the same cohort as in the present study(Reference Nwaru, Erkkola and Lumia11,Reference Lumia, Luukkainen and Tapanainen27) . Higher maternal intake of α-linolenic acid during pregnancy was associated with a decreased risk of asthma(Reference Lumia, Luukkainen and Tapanainen27) and allergic rhinitis in the offspring(Reference Nwaru, Erkkola and Lumia11), but for other fatty acids the findings were more inconsistent.
Our results are supported by the RCT where no associations were observed between maternal supplementation of n-3 PUFA during pregnancy and the risk of food allergy in the offspring(Reference Dunstan, Mori and Barden13,Reference Palmer, Sullivan and Gold14) . However, a protective effect has been observed in one RCT(Reference Furuhjelm, Warstedt and Fagerås12). The inconsistent results might be explained by the methodological differences between the studies. The dosage of n-3 PUFA supplementation has varied, and it is possible that there exists a dose-dependent association. In addition, the variable timing of the supplementation may have resulted in inconsistencies, as the crucial timing of the immune development in infants is yet to be determined.
Our results are in apparent contrast to the recent systematic review concluding that maternal fish oil supplementation during pregnancy may reduce the risk of sensitisation for egg and peanut in the offspring(Reference Garcia-Larsen, Ierodiakonou and Jarrold28). However, in epidemiological studies null results have been observed when studying the association between maternal dietary intake of fatty acids and food sensitisation in the offspring(Reference Notenboom, Mommers and Jansen29,Reference Yu, Chan and Calder30) , one study reporting the sensitisation to cows’ milk proteins(Reference Nwaru, Ahonen and Kaila10). As egg and peanuts are consumed only in small quantities, and maybe not as early in life as cows’ milk products, the maternal intake of fatty acids may have a greater role in the prevention of egg and peanut allergy. Further, the association between sensitisation and food allergy varies, and it is possible to have sensitisation without clinical food allergy, as well as food allergy without sensitisation(Reference Sicherer, Wood and Abramson31).
Our result that α-linolenic acid was inversely associated with the development of CMA only in mothers without a history of allergic rhinitis or asthma may be explained by the possibility that mothers transfer the risk of allergy to their offspring(Reference Cook-Mills32). Thus, it is possible that the maternal history of allergy overpowers the protective effect of α-linolenic acid.
The major strength of our study is prospectively collected data from a relatively large sample size, which minimise the selection bias. Our endpoint was based on register-based information and complemented with a parental questionnaire which is validated to represent exceptionally well the physician diagnosed CMA(Reference Tuokkola, Kaila and Pietinen19,Reference Tuokkola, Luukkainen and Kaila20) . Further, our food consumption data had good coverage and were collected by validated FFQ, specifically designed for the present study. In addition, the FFQ took into account the individual habit of used fats, which increases the accuracy of the intake of specific fatty acids.
The major limitation of our study is the restriction of the participants to those with human leucocyte antigen conferred susceptibility to type 1 diabetes. As previously reported, these infants may have increased intestinal permeability(Reference Vaarala, Atkinson and Neu33); also, the knowledge of the risk of type 1 diabetes may alter the behaviour of the family, and the parents may seek medical advice more eagerly leading to receiving the diagnosis of CMA more often. These factors may explain the higher incidence of CMA in our study population compared with what is previously reported in Finland(Reference Saarinen, Juntunen-Backman and Järvenpää2,Reference Pyrhönen, Näyhä and Kaila3) . Therefore, the generalisability of our results to the general paediatric populations may be limited. In addition, we did not have data on the duration between the onset of symptoms of CMA and the date of diagnosis or start of the elimination diet, even though these should be coincidental. The FFQ was designed to represent the total diet during the 8th month of pregnancy and thus presents an estimate of the diet during the whole pregnancy. In the validation study, the FFQ was observed to slightly overestimate the nutrient intake; this should be ameliorated by the usage of energy-adjusted nutrient intakes. As the FFQ was mailed to the participants after the delivery, it is open for recall bias. However, the FFQ was validated in the same design as the present study was performed and was found to be suitable to measure the maternal diet during pregnancy and has shown acceptable reproducibility and validity(Reference Erkkola, Karppinen and Javanainen16); thus, the risk of recall bias should be minor. The use of food supplements was not taken into account in the validation study. As our results were substantially similar for the fatty acid intakes from food and from food and supplements together, this should not have major relevance in our study. The fact that we did not have data on the child’s dietary intake may result in some residual confounding.
In conclusion, the present study provides novel information about the association between maternal intake of fatty acids during pregnancy and the development of CMA in the offspring. We did not observe an association between the maternal intake of fatty acids and the development of CMA in the offspring, except the maternal intake of α-linolenic acid which was associated with decreased risk of CMA in the offspring of mothers without a history of allergic rhinitis or asthma. Thus, additional benefits may not be expected for the prevention of CMA in the offspring by advising the pregnant women to use supplements containing fatty acids in addition to a healthy and balanced diet.
Acknowledgments
We are extremely grateful to all the families taking part of the study. We would like to acknowledge every research nurse, doctor, nutritionist and laboratory staff in the DIPP study for their excellent collaboration over the years.
The study was supported by the Academy of Finland (grants 63672,79685, 79686, 80846, 201988, 210632, 129492, 126813, 276475), the Finnish Paediatric Research Foundation, the Juho Vainio Foundation, the Yrjö Jahnsson Foundation, the Competitive Research Funding of the Tampere University Hospital (grants 9L035, 9M029, 9P017, 9P057, 9R012, 9R055, 9S015, 9S074, 9T072, 9U016, 9U065, 9V012, 9V072, 9X062), Medical Research Funds of Turku and Oulu University Hospitals, the European Foundation for the Study of Diabetes supported by EFSD/JDRF/Lilly, the Juvenile Diabetes Research Foundation (grants 197032, 4-1998-274, 4-1999-731, 4-2001-435), the Novo Nordisk Foundation and EU Biomed 2 (BMH4-CT98-3314), Doctoral Programs for Public Health, Foundation for Allergy Research, Research Foundation of Orion Corporation, Tampere Tuberculosis Foundation, Päivikki and Sakari Sohlberg Foundation and the Jalmari and Rauha Ahokas Foundation. None of the funders had any role in the design, analysis or writing of this article.
S. M. V., M. K., A. L. and J. M. were responsible for formulating the research questions. J. I., M. K., J. T. and R. V. are members of the steering committee of the DIPP study. S. M. V. designed the DIPP nutrition study and within the DIPP nutrition study the allergy study was designed by S. M. V. and M. K. S. M. V. and M. K. were responsible for carrying out the study. R. V. was responsible for the clinical work in Oulu, M. K. was responsible for the clinical work in Tampere. A. L. and H.-M. T. were responsible for analysing the data. A. L. wrote the first version of the article. J. M., M. K., and S. M. V. participated in the writing process. S. M. V. and M. K. had the primary responsibility for the final work with equal contribution. All the authors participated in the critical revision of the manuscript and have accepted the final version.
All authors declare that there are no conflicts of interest.
