Methods

Study population

The Osaka Maternal and Child Health Study (OMCHS) is a prospective cohort study. Details of the OMCHS have been described elsewhere.8 We recruited pregnant women at baseline. Neyagawa City is one of the 43 municipalities in Osaka Prefecture, a metropolis in Japan. Of the 3639 eligible subjects in Neyagawa City, 627 women (17.2%) participated in the OMCHS between November 2001 and March 2003. In order to increase the sample size, pregnant women living in municipalities other than Neyagawa City were also recruited. Eight pregnant women who did not live in Neyagawa City but who had become aware of the OMCHS at an obstetric clinic before August 2002 decided by themselves to participate in the OMCHS. In addition, there were 77 participants who received details of the OMCHS from public health nurses in six other municipalities from August 2002 to March 2003. From October 2002 to March 2003, 290 participants were recruited from a university hospital and three obstetric hospitals in three other municipalities.

A total of 1002 pregnant women at any stage of pregnancy gave their fully informed consent in writing and completed the baseline survey. Of the 1002 subjects, 867 mother-child pairs participated in the second survey from 2 to 9 months postpartum. Of the 867 mother-child pairs, 763 participated in the third survey from 16 to 24 months postpartum. The population analysed consisted of these 763 mother-child pairs.

Measurements

At baseline, each participant filled out a set of two self-administered questionnaires. Also, a self-administered questionnaire was used in both the second and third survey. Participants mailed these materials to the data management centre at the time of each survey. Research technicians completed missing or illogical data by telephone interview.

In the baseline survey, dietary habits during the preceding month were assessed using a self-administered, semi-quantitative, comprehensive diet history questionnaire (DHQ). Details of the structure of the DHQ, calculation of dietary intake and validity of commonly studied nutritional factors have been published elsewhere.9 10 Estimates of daily intake for foods (150 items in total), energy and selected nutrients were calculated using an ad hoc computer algorithm for the DHQ9 10 based on the Standard Tables of Food Composition in Japan.11 12 Information on dietary supplements was not used in the calculation of dietary intake. The correlation coefficients for nutrient intake between those estimated from the DHQ and those observed by a 3-day dietary record were 0.75, 0.50, 0.37 and 0.49 for saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids and cholesterol, respectively, in women.9 A highly positive correlation was also observed between intake of n-3 polyunsaturated fatty acids of marine origin estimated by the DHQ and the corresponding concentration in the serum phospholipid fraction in women (r = 0.69).10 Energy-adjusted intake by the residual method was used for the analyses.13

A second questionnaire at baseline asked about maternal age, gestation, family income, maternal and paternal education, maternal and paternal history of asthma, atopic eczema and allergic rhinitis and changes in diet in the previous month. A paternal or maternal history of asthma, atopic eczema and allergic rhinitis was defined as positive if the respective parent had been treated with medications for any of these allergic disorders at some time before the start of the survey.

A self-administered questionnaire in the second survey elicited information on the sex, birth weight and date of birth of the baby, the number of older siblings, maternal smoking during pregnancy and smoking in the household. A self-administered questionnaire in the third survey included questions on breastfeeding duration in months and symptoms of wheeze and eczema based on the ISAAC phase I questionnaire.14 15 Wheeze was defined as present if the mother answered “yes” to the written question: “Has your child had wheezing or whistling in the chest in the last 12 months?”. For eczema, affirmative answers to the following three questions were required: “Has your child ever had an itchy rash which was coming and going for at least 6 months?”, “Has your child had this itchy rash at any time in the last 12 months?” and “Has this itchy rash at any time affected any of the following places: the folds of the elbow, behind the knees, in front of the ankles, under the buttocks, or around the neck, ears, or eyes?”.

Statistical analysis

The intake of dietary factors under investigation was categorised at quartile points based on the distribution in 763 subjects. The following variables were a priori selected as potential confounding factors: maternal age, gestation at baseline, residential municipality, family income, maternal and paternal education, maternal and paternal history of asthma, atopic eczema and allergic rhinitis, maternal intake of vitamins D and E during pregnancy, changes in maternal diet in the previous 1 month, season when data at baseline were collected, maternal smoking during pregnancy, baby’s older siblings, baby’s sex, baby’s birth weight, household smoking in the same room as the infant, breastfeeding duration and time of delivery before the third survey.

Logistic regression analysis was used to estimate crude odds ratios (ORs) and 95% confidence intervals (CIs) of wheeze and eczema for each category of dietary intake under study in comparison with the lowest intake category. Multiple logistic regression analysis was employed to control for potential confounding factors. Trend of association was assessed by a logistic regression model assigning consecutive integers (1–4) to the quartiles of the exposure variables. All statistical analyses were conducted using the SAS software package Version 9.1 (SAS Institute, Cary, North Carolina, USA).

Results

Of the 763 infants, the third survey revealed that 169 (22.1%) and 142 (18.6%) had symptoms of wheeze and eczema based on the ISAAC criteria, respectively. About 75% of infants had been breastfed for 6 months or longer (table 1). Slight or substantial changes in diet in the previous month were reported by 551 mothers due to nausea gravidarum (n = 453), maternal and fetal health (n = 89) and other reasons (n = 9). Maternal mean daily total energy intake and energy-adjusted consumption of n-3 and n-6 polyunsaturated fatty acids during pregnancy were 7629 kJ, 2.4 g and 11.5 g, respectively (table 2).

Table 3 presents ORs and 95% CIs for the risk of wheeze and eczema in the offspring according to maternal dietary intake of specific types of fatty acids and cholesterol during pregnancy. After adjustment for the confounding factors under study, the maternal intake of total fat, saturated fatty acids, monounsaturated fatty acids and n-3 polyunsaturated fatty acids during pregnancy was not measurably related to the risk of wheeze or eczema in the offspring. Compared with maternal α-linolenic acid intake during pregnancy in the first quartile, such intake in the fourth quartile was independently associated with a reduced risk of wheeze in the offspring (adjusted OR 0.52, 95% CI 0.28 to 0.97), although the inverse exposure-response relationship was not statistically significant. There was no association between maternal α-linolenic acid intake during pregnancy and the risk of infantile eczema. Maternal eicosapentaenoic acid intake during pregnancy was not related to the risk of childhood wheeze or eczema. Maternal consumption of docosahexaenoic acid in the second and fourth quartiles was independently related to a decreased risk of wheeze in the infants, although the inverse linear trend was not statistically significant (adjusted OR between extreme quartiles 0.37, 95% CI 0.15 to 0.91). No such relationship was detected for infantile eczema. Maternal intake of n-6 polyunsaturated fatty acids and linoleic acid during pregnancy in the highest quartile was independently associated with an increased risk of infantile eczema, but not wheeze, showing clear inverse exposure-response relationships (adjusted ORs 2.25 (95% CI 1.13 to 4.54) and 2.11 (95% CI 1.06 to 4.26), p for trend  = 0.01 and 0.03, respectively). No statistically significant exposure-response relationships were observed between maternal intake of arachidonic acid and cholesterol and the ratio of n-3 to n-6 polyunsaturated fatty acid consumption during pregnancy and the risk of wheeze or eczema in the offspring after multivariate adjustment.

After adjustment for potential confounders excluding maternal intake of vitamins D and E during pregnancy, inverse exposure-response associations between maternal intake of total fat, saturated fatty acids, monounsaturated fatty acids, n-3 polyunsaturated fatty acids and α-linolenic acid during pregnancy and the risk of wheeze in infants were statistically significant whereas maternal intake of n-6 polyunsaturated fatty acids and linoleic acid during pregnancy was not related to the risk of infantile eczema.

There were no evident relationships between maternal intake of meat and fish during pregnancy and the risk of wheeze or eczema in the offspring (table 4).

Discussion

During immune maturation, naive Th0 cells are able to modulate the predominant expression of a Th1- or Th2-like cytokine pattern according to numerous eicosanoids and cytokines synthesised by antigen-presenting cells.16 Polyunsaturated fatty acids and their derivatives can influence the Th1:Th2 ratio. An intervention study among atopic pregnant Australian women showed that fish oil supplementation achieved a significant increase in n-3 polyunsaturated fatty acids in neonatal erythrocyte membranes compared with the control group and that the neonatal in vitro interleukin 10 (IL10) response to cat allergen was significantly lower in the fish oil-supplemented group than in the control group.17 Other reports of that intervention study showed that IL13 levels in cord plasma and neonatal leukotriene B₄ production by cord blood-derived neutrophils were significantly lower in infants from the intervention group than from the placebo group.18 19 In a randomised clinical trial of 528 pregnant Danish women, the risk of allergic asthma, atopic dermatitis or allergic rhinitis during the first 15–16 years of life was significantly lower in children whose mothers had received supplements of fish oil (2.7 g marine n-3 polyunsaturated fatty acids per day) compared with those in the placebo group (olive oil), although the risk in children of mothers in the no-oil group was remarkably similar to the risk in children whose mothers had received fish oil.20 These findings agree with our results in relation to maternal intake of α-linolenic acid and docosahexaenoic acid during pregnancy and the risk of infantile wheeze. The lack of association of wheeze in offspring with maternal eicosapentaenoic acid intake in the present study might to some extent be ascribed to the fact that docosahexaenoic acid is prefentially transferred to the fetus from the maternal circulation.21 A UK prospective study showed that the levels of 11 individual n-3 and n-6 fatty acids in both cord blood red cells and maternal blood red cells obtained after 20 weeks of pregnancy were not related to the risk of infantile wheezing and eczema.22 Japanese people consume a large quantity of fish compared with people in Western countries, but the median value of daily marine origin n-3 polyunsaturated fatty acid intake in the present study was 0.48 g, which was much lower than the supplemented quantity in the Danish intervention study cited above.20 A clear protective relationship between maternal intake of marine origin n-3 polyunsaturated fatty acids and the risk of infantile atopic eczema might be substantiated if consumption was much higher than in this study. The non-significant association of wheeze and eczema in the offspring with maternal fish intake in the current study could be due to unrecognised active agents in fish that interfered with the expected benefit to the offspring of maternal fish intake during pregnancy. For example, methylmercury and dioxins are accumulated in fish and shellfish through the marine food web. This hypothetical adverse effect might not be detected in Western populations where fish intake is low.2 3 5 6

Linoleic acid is converted into arachidonic acid which enhances inflammatory processes via its derivatives prostaglandin E₂ and leukotriene B₄. Our results regarding maternal intake of n-6 polyunsaturated fatty acids and linoleic acid and infantile eczema might be reasonable. On the other hand, lipoxins derived from arachidonic acid decrease the synthesis of tumour necrosis factor α and leucocyte chemotaxis.16 Arachidonic acid, as well as docosahexaenoic acid, is essential to the placenta and the fetus.16 A US birth cohort study found that higher levels of cord blood eicosapentaenoic acid and arachidonic acid were associated with attenuation of cord blood lymphocyte proliferation and decreased secretion of interferon γ in response to allergen stimulation, whereas higher levels of linoleic acid were correlated with increased allergen-stimulated IL13 levels.23 Given that the beneficial effect of arachidonic acid on fetal immune development counteracts the detrimental effect, the present findings on null associations between maternal intake of n-6 polyunsaturated fatty acids, linoleic acid and arachidonic acid during pregnancy and infantile wheeze also might be reasonable.

This study had methodological strengths. Study subjects were homogeneous with the same residential background. Extensive data on potential confounding factors were controlled for. In particular, maternal intake of vitamins D and E during pregnancy was the most important confounder. Such intake was inversely related to the risk of infantile wheeze and eczema. The prospective design was likely to reduce the possibility of recall bias. The definition of wheeze and eczema was based on the ISAAC questions. However, the outcomes under investigation were measured at varying ages between 16 and 24 months and, in addition, wheeze in infancy is known not to be a good predictor of asthma over the longer term. The consequence would have been an underestimation of values in our results.

Other weaknesses of the current study should be borne in mind. Our DHQ could only approximate consumption. The consequence would bias the estimates of the observed association toward the null. The DHQ was designed to assess recent dietary intake (ie, for 1 month before completing the questionnaire). This disadvantage is likely to be alleviated after adjustment for the season when data were collected, however. Changes in diet in the past month were controlled for because pregnant women are likely to change their diet.

The follow-up rate was not good. Of the 1002 participants at baseline, 239 mother-child pairs did not take part in the third survey. There were no material differences between the 239 non-participants and the 763 participants in the third survey with regard to distribution of maternal age, maternal and paternal history of asthma, atopic eczema and allergic rhinitis, and maternal intake of total energy, monounsaturated fatty acids, n-3 polyunsaturated fatty acids, n-6 polyunsaturated fatty acids other than arachidonic acid, meat and fish. Compared with non-participants in the third survey, participants were less likely to report low family income and a low maternal and paternal educational level and were more likely to have a high intake of total fat, saturated fatty acids, cholesterol, arachidonic acid and vitamins D and E. Moreover, at baseline the participation rate in Neyagawa City was only 17.2% and in other areas the participation rate could not be calculated because the exact number of eligible subjects among the sources from which participants were recruited was not available. Thus, the mother-child pairs in this study were probably not representative of Japanese mother-child pairs in the general population. In fact, educational levels were higher in the mothers in our study than in the general population. According to the 2000 population census of Japan, the proportions of women aged 30–34 years in Osaka Prefecture with years of education of <13, 13–14, 15+ and unknown were 49.2, 32.3, 13.6 and 4.9%, respectively.24 The present population might have had a greater awareness about health than the general population.

The current finding regarding the null association between maternal meat intake during pregnancy and infantile eczema was at variance with our previous results, showing a positive association between maternal meat intake during pregnancy and suspected atopic eczema in infants aged 3–4 months. The definition of suspected atopic eczema might have been crude. Of 58 infants identified with suspected atopic eczema, 25 (43.1%) were classified as having eczema based on the ISAAC criteria in the third survey. In view of the multiple analyses and the p values, it is possible that some or all of the findings in the current study and in our previous study occurred by chance.

Our results suggest that maternal intake of α-linolenic acid and docosahexaenoic acid during pregnancy may be preventive against infantile wheeze whereas maternal intake of n-6 polyunsaturated fatty acids, especially linoleic acid, during pregnancy may increase the risk of childhood eczema. Further epidemiological investigations are required to ascertain whether the relationships observed in the current study are replicated in other populations.