Asthma and wheezing commonly occur during childhood, with a prevalence of approximately 10 % in Japan(Reference Ebisawa, Nishima and Ohnishi1,Reference Morikawa, Sasaki and Yoshida2) . There is no exact method for predicting or preventing childhood wheezing. While environmental factors have been reported as causal factors for asthma and wheezing during childhood(Reference Yu, Dai and Chen3–Reference Owora and Zhang5), pregnancy-associated factors have also been reported to influence the incidence of childhood asthma and wheezing(Reference Takata, Tanaka and Nagata6–Reference Murata, Kyozuka and Yasuda9). According to our previous study, maternal ritodrine hydrochloride administration for the prolongation of pregnancy was associated with an increased incidence of childhood wheezing in the offspring(Reference Murata, Kyozuka and Yasuda9). There is no clear explanation for this association; however, the foetuses of mothers whose gestational period was prolonged by long-term maternal ritodrine hydrochloride administration may have experienced stress due to ‘excessive’ uterine contractions, and there might have been secretion of foetal pituitary adrenocorticotropic hormone, leading to asthma development in the offspring in future(Reference Murata, Kyozuka and Yasuda9,Reference Matsubara, Takahashi and Ohkuchi10) . The association between prenatal factors and incidence of childhood wheezing in the offspring should be clarified considering underlying mechanisms.
Mg is a common factor involved in the function of enzymes and acts as a cofactor in over 600 enzymatic reactions(Reference Dalton, Ní Fhloinn and Gaydadzhieva11). Mg is important during pregnancy because it also affects uterine smooth muscle activity(Reference Zhang, Xun and Chen12), which could affect the pregnancy and offspring; additionally, it exerts various benefits, including anti-inflammatory and antioxidant effects(Reference Mousavi, Alizadeh and Asghari Jafarabadi13). Several studies have reported that adequate maternal magnesium intake (MMI) reduces preterm births, foetal growth restriction, small-for-gestational-age births and pre-eclampsia(Reference Tsakiridis, Kasapidou and Dagklis14). However, the association between MMI and offspring childhood health, including childhood wheezing, remains unclarified. A previous study did not report any association between MMI, which was estimated using a FFQ, and childhood asthma(Reference Nwaru, Erkkola and Ahonen15); however, this study showed an association between MMI and the decreased incidence of childhood eczema in offspring(Reference Nwaru, Erkkola and Ahonen15). Non-pregnant adults who received oral Mg supplements showed improved objective and subjective measures of bronchial reactivity, asthma control, and quality of life, respectively(Reference Kazaks, Uriu-Adams and Albertson16).
Therefore, we analysed the association between MMI and incidence of childhood wheezing in the offspring using data from a nationwide birth cohort study. Based on the above phenomenon(Reference Kazaks, Uriu-Adams and Albertson16), we hypothesised that higher MMI imparts anti-inflammatory and antioxidant effects that decrease childhood wheezing incidence in offspring.
Experimental methods
Study design
We analysed data from the Japan Environment and Children’s Study (JECS), a nationwide, government-funded, prospective birth cohort study initiated in January 2011 – participants were enrolled during 2011–2014 – aimed to investigate the effects of environmental factors on children’s health(Reference Kawamoto, Nitta and Murata17,Reference Michikawa, Nitta and Nakayama18) . Expectant mothers who (1) resided within the study area at the time of recruitment, with plans to keep residing in Japan in the foreseeable future; (2) had their expected delivery date between August 2011 and mid-2014; and (3) could easily participate in the study (i.e. understood the Japanese language and completed a self-administered questionnaire) were considered eligible for inclusion in the JECS.
The two modes of recruitment were as follows: (1) at the time of first prepartum examination by cooperating healthcare providers and (2) at local government offices issuing a pregnancy journal (i.e. the Maternal and Child Health Handbook) to all expectant mothers before receiving municipal services. Pregnant women were contacted through healthcare providers, and those who were willing to participate were registered. Self-administered questionnaires were used to collect information on demographic factors, medical history, physical and mental health, lifestyle, occupation, environmental exposure at home and in the workplace, housing condition, and socio-economic status during the first and second/third trimesters(Reference Kawamoto, Nitta and Murata17,Reference Michikawa, Nitta and Nakayama18) .
Data collection
The current analysis used data released in October 2019 (dataset: jecs-ta-20190930). Women with singleton pregnancies were included in the present study, while those who experienced abortion or stillbirth or had missing information were excluded.
Exposure variables
We used the FFQ administered in the early stage of pregnancy and in the second/third trimester (median: 27·0 weeks of gestation) in the JECS. Participants were asked to answer questions about their 1-year dietary intake at the early stage of pregnancy and between awareness of pregnancy and the second or third trimester. The FFQ used in this study was validated in a large-scale epidemiological study in Japan, and its details have been reported(Reference Yokoyama, Takachi and Ishihara19). Briefly, the long-type FFQ asked about the respondents’ habitual consumption of the listed food items using nine frequency and three portion sizes categories(Reference Yokoyama, Takachi and Ishihara19–Reference Eshak, Okada and Baba21). Frequencies ranged from less than once per month to more than seven times per d, and portions ranged from small, 0·5; medium, 1·0; and large, 1·5, compared with standard units(Reference Yokoyama, Takachi and Ishihara19–Reference Eshak, Okada and Baba21). Eating habits, including breakfast frequency, eating out and eating speed, were also assessed(Reference Yokoyama, Takachi and Ishihara19–Reference Eshak, Okada and Baba21). The validity of nutrient intake assessment using a FFQ in Japanese pregnant women has been previously reported(Reference Miura, Takamori and Hamazaki20–Reference Ogawa, Jwa and Kobayashi22).
The nutrient contents of each food were determined based on the Japanese food composition tables, 5th revised revision, while the daily intakes of nutrients were calculated by summing the contents from the food items after multiplying by the frequency of consumption(Reference Eshak, Okada and Baba21); MMI was also computed using FFQ data. We performed two analyses to adjust for the effects of maternal daily energy intake in assessing the association between MMI and the incidence of childhood wheezing in offspring: a standard model adjustment and nutrient density model adjustment(Reference Tomova, Arnold and Gilthorpe23). For the standard model adjustment, we focused on maternal daily energy intake as a confounding factor, after confirming the multicollinearity. For nutrient density model adjustment, we calculated adjusted MMI (aMMI) by rescaling the MMI as a proportion of total energy (MMI (mg/d)/maternal daily energy intake (kcal/d)). We plotted a histogram and performed the Shapiro–Wilk test, which showed that MMI and aMMI were not normally distributed. Participants were categorised based on (1) their MMI levels quintiles (< 148·00, 148·00–187·99, 188·00–228·99, 229·00–289·99 and ≥ 290·00 mg/d), (2) their aMMI levels quintiles (< 0·107, 0·107–0·119, 0·120–0·132, 0·133–0·149 and ≥ 0·150 mg/kcal), and (3) whether they had below- or above-ideal MMI levels (< 310·00 or ≥ 310·00 mg/d) according to the Reference Intakes for the Japanese population.
Main outcome measures and confounding factors
The main outcome measure was ‘wheezing ever’ at 3 years of age, which was assessed based on the International Study of Asthma and Allergies in Childhood (ISAAC) questionnaire(Reference Asher, Keil and Anderson24–Reference Ellwood, Asher and Beasley26). Information on ‘wheezing ever’ was obtained using self-reported questions from caregivers, including ‘Has your child ever had wheezing or whistling in the chest at any time in the past?’
Maternal age, BMI before pregnancy, parity, smoking status, educational status, annual household income, maternal daily energy intake, mode of delivery, preterm birth (before 37 weeks), low birth weight (< 2500 g), severe hypertensive disorders of pregnancy, maternal asthma, intra-uterine infection, MMI estimated at the early stage of pregnancy and maternal intake of other minerals during pregnancy, including Ca and Zn, that affect maternal Mg absorption via the digestive system were analysed as potential confounding factors. These factors were selected based on clinical importance(Reference Murata, Kyozuka and Yasuda9,Reference Nwaru, Erkkola and Ahonen15,Reference Kotani, Kim and Ishizaki27–Reference Litonjua, Rifas-Shiman and Ly30) .
Participants were divided into four age groups: < 20, 20–29, 30–39 and ≥ 40 years. Maternal BMI before pregnancy was categorised as follows(Reference Murata, Kyozuka and Yamaguchi31): < 18·5, 18·5–19·9, 20·0–22·9, 23·0–24·9 and ≥ 25·0 kg/m2. Participant parity was categorised as nulliparous and multiparous. Participants were requested to provide information on their smoking status by choosing one of the following: ‘Currently smoking’, ‘Never’, ‘Previously did, but quit before realising current pregnancy’ and ‘Previously did, but quit after realising current pregnancy’. Participants who chose ‘Currently smoking’ were included in the ‘smoking’ group; otherwise, participants were assigned to the ‘non-smoking’ group(Reference Murata, Kyozuka and Yasuda9). Maternal educational status was categorised into four groups as follows: junior high school, < 10 years; high school, 10–12 years; technical junior college, technical/vocational college, associate degree, or bachelor’s degree, 13–16 years; and graduate degree (master’s/doctorate), ≥ 17 years(Reference Murata, Kyozuka and Yasuda9). Annual household income was categorised into four levels as follows: < 2 000 000; 2 000 000–5 999 999; 6 000 000–9 999 999 and ≥ 10 000 000 Japanese yen(Reference Murata, Kyozuka and Yasuda9). Maternal daily energy intake was evaluated using a FFQ that was validated in previous cohort studies(Reference Yokoyama, Takachi and Ishihara19,Reference Eshak, Okada and Baba21) and calculated based on the Japan Standard Tables of Food Composition, 5th Revised Edition. Mode of delivery was divided into vaginal delivery and caesarean section based on medical record transcripts. Severe hypertensive disorders of pregnancy were defined as a persistently elevated blood pressure (≥ 160/110 mmHg) after 20 gestational weeks in an otherwise normotensive woman(Reference Brown, Magee and Kenny32). Maternal asthma was diagnosed based on the information in the questionnaire issued during the first trimester. Paternal asthma was not considered owing to a large amount of missing data. Intra-uterine infection was clinically diagnosed by the attending physician, and the corresponding data were retrieved from medical record transcripts. MMI via daily diet at the early stage of pregnancy was determined using the FFQ(Reference Yokoyama, Takachi and Ishihara19,Reference Eshak, Okada and Baba21) . After combining the second and third groups to form the moderate group, the participants were categorised into low-MMI (< 167·00 mg/d), moderate-MMI (167·00–287·99 mg/d) and high-MMI (≥ 288·00 mg/d) groups according to the quartiles of the MMI distribution in the early stage of pregnancy. Maternal intake of other minerals during pregnancy via the daily diet was determined using the FFQ(Reference Yokoyama, Takachi and Ishihara19,Reference Eshak, Okada and Baba21) . The participants were categorised into groups according to the quartiles of the distribution of maternal Ca intake in the second/third trimester. After combining the second and third quartiles to form the moderate group, we analysed the following three groups: low maternal Ca intake (< 320·00 mg/d), moderate maternal Ca intake (320·00–652·99 mg/d) and high maternal Ca intake (≥ 653·00 mg/d). Participants were likewise categorised into low, moderate and high maternal Zn intake groups according to the quartiles of the distribution of maternal Zn intake in the second/third trimester (< 5·40, 5·40–8·59 and ≥ 8·60 mg/d, respectively). For each confounder, ‘no answer’ was analysed as a single item.
Statistical analysis
Participant characteristics were summarised based on their daily MMI status. Univariable and multivariable logistic regression models were used to calculate the crude OR (cOR), adjusted OR (aOR) and 95 % CI for offspring childhood wheezing among participants (1) in each MMI, (2) aMMI and (3) above-ideal MMI categories, with the lowest MMI group used as the reference group for each classification criteria. Multivariable logistic regression analyses were adjusted for maternal age, BMI before pregnancy, parity, maternal smoking status, educational status, annual household income, maternal daily energy intake, mode of delivery, preterm birth, low birth weight, severe hypertensive disorders of pregnancy, maternal asthma, intra-uterine infection, MMI estimated at the early stage of pregnancy, and maternal intake of Ca and Zn during pregnancy. However, in analysis (2), multivariable logistic regression did not adjust maternal daily energy intake.
All statistical analyses were performed using SPSS version 26 (IBM Corp.). The model fitness of the multivariable logistic regression was confirmed using Hosmer and Lemeshow analysis. There was no multicollinearity, which was deemed present should there be an association between independent variables with a correlation coefficient of r > 0·8 and/or a variance inflation factor > 10.
Results
The total number of foetal records in the JECS was 104 062, and 79 907 individuals met the inclusion criteria (Fig. 1).
Enrolment flow chart.
Table 1 summarises participant characteristics stratified by MMI status. Differences were noted between the groups with respect to demographic, socio-economic, medical and other nutrient intake data. The incidence of preterm birth in the present study was relatively higher in the high-MMI group than in the remaining groups. Specifically, maternal daily energy intake, MMI estimated at the early stage of pregnancy, and maternal intake of Ca and Zn during pregnancy were the highest in the highest-MMI group. The ratio of wheezing ever was also the highest in this group. The median values of MMI in the early stage of pregnancy, maternal Ca intake during pregnancy and maternal Zn intake during pregnancy were 221·00 mg/d, 465·00 mg/d and 6·80 mg/d, respectively.
Characteristics and outcomes of participants based on maternal magnesium intake in the second/third trimester status
MMI, maternal magnesium intake; JPY, Japanese yen; HDP, hypertensive disorders of pregnancy.
Table 2 summarises the cOR, aOR and 95 % CI for childhood wheezing in the offspring of participants in each MMI group. The aOR for childhood wheezing in the offspring of participants in the highest MMI group was 1·09 (95 % CI, 1·00, 1·20), and no significant difference was found in the aOR between other groups.
Crude OR and adjusted OR for the incidence of ‘wheezing ever’ in women subdivided according to quintiles of maternal magnesium intake
cOR, crude OR; Ref, reference; aOR, adjusted OR.
Multivariable logistic regression analyses were adjusted for maternal age, BMI before pregnancy, parity, maternal smoking status, educational status, annual household income, maternal daily energy intake, mode of delivery, preterm birth, low birth weight, severe hypertensive disorders of pregnancy, maternal asthma, intra-uterine infection, maternal magnesium intake estimated at the early stage of pregnancy and maternal intake of Ca and Zn during pregnancy.
Table 3 summarises the cOR, aOR and 95 % CI for childhood wheezing in the offspring of participants in each aMMI group. The aOR for childhood wheezing in the offspring of participants in the higher aMMI group showed a decreased trend; however, no significant difference was found in the aOR.
Crude OR and adjusted OR for the incidence of ‘wheezing ever’ in women subdivided according to quintiles of adjusted maternal magnesium intake
cOR, crude OR; Ref, reference; aOR, adjusted OR.
Multivariable logistic regression analyses were adjusted for maternal age, BMI before pregnancy, parity, maternal smoking status, educational status, annual household income, mode of delivery, preterm birth, low birth weight, severe hypertensive disorders of pregnancy, maternal asthma, intra-uterine infection, maternal magnesium intake estimated at the early stage of pregnancy, and maternal intake of Ca and Zn during pregnancy.
Table 4 summarises the cOR, aOR and 95 % CI for childhood wheezing in the offspring of participants in the above-ideal MMI group. The aOR for childhood wheezing in the offspring of participants in the above-ideal MMI groups was 1·04 (95 % CI, 0·98, 1·10), and there was no significant difference in aOR between the reference group and above-ideal MMI group.
Crude OR and adjusted OR for the incidence of ‘wheezing ever’ in women in the above-ideal maternal magnesium intake group
cOR, crude OR; Ref, reference; aOR, adjusted OR.
Multivariable logistic regression analyses were adjusted for maternal age, BMI before pregnancy, parity, maternal smoking status, educational status, annual household income, maternal daily energy intake, mode of delivery, preterm birth, low birth weight, severe hypertensive disorders of pregnancy, maternal asthma, intra-uterine infection, maternal magnesium intake estimated at the early stage of pregnancy, and maternal intake of Ca and Zn during pregnancy.
Discussion
This study revealed a statistically significant but slight association between the highest levels of MMI during pregnancy and an increased incidence of childhood wheezing in offspring. However, there was no significant association between aMMI, above-ideal MMI levels and the incidence of childhood wheezing in the offspring. Contrary to our hypothesis, we conclude that higher MMI during pregnancy would not be useful for decreasing the incidence of childhood wheezing in offspring; a higher MMI during pregnancy, which reflects the real-world distribution of MMI across the nationwide population, is rather associated with a slightly elevated incidence of childhood wheezing in offspring.
The mechanism of a slightly but statistically significant increased incidence of childhood wheezing in offspring from mothers with the highest MMI is unknown. It is speculated that other maternal demographic, socio-economic, medical and other nutrient intake backgrounds would affect this association rather than MMI itself, despite adjusting these factors in multivariable logistic regression analyses. Meanwhile, there was a significant association between higher Mg levels in maternal urine samples collected before delivery and an increased incidence of offspring childhood asthma, although the underlying mechanism was unclear(Reference Karakis, Landau and Gat33). Another previous study reported that Mg intake in pre-school children might have contributed to wheezing or asthma development in childhood, although the mechanism also remains to be elucidated(Reference Emmanouil, Manios and Grammatikaki34). To the best of our knowledge, this study is the first to clarify the association of MMI with the incidence of childhood wheezing in offspring using a nationwide cohort study with appropriate confounding factors, which might strengthen these previous results(Reference Karakis, Landau and Gat33,Reference Emmanouil, Manios and Grammatikaki34) .
Analyses based on aMMI and ideal MMI showed no significant association between MMI and the incidence of childhood wheezing in offspring. According to our hypothesis, higher MMI may decrease inflammation and oxidative stress(Reference Mousavi, Alizadeh and Asghari Jafarabadi13) and lead to hyperactivity of pulmonary functions in the offspring, stabilisation of T cells and inhibition of mast cell degranulation to reduce inflammatory mediators(Reference Davalos Bichara and Goldman35). This was also supported by the study of non-pregnant adults who received oral Mg supplements(Reference Kazaks, Uriu-Adams and Albertson16). Moreover, higher MMI would maintain pregnancy and prolong gestation because higher maternal serum Mg levels were associated with the activity of uterine smooth muscle (preterm birth group: 0·93 mmol/L v. control group: 1·12 mmol/L, p < 0·01)(Reference Dalton, Ní Fhloinn and Gaydadzhieva11,Reference Zhang, Xun and Chen12) and may be associated with a decreased incidence of childhood wheezing in offspring. These may have protected the offspring from childhood asthma and wheezing. The contrary results could have been caused by differences in the mechanisms affecting childhood health via inflammatory and oxidative pathways between MMI, maternal serum Mg levels and Mg supplementation. This study’s incidence of preterm birth was relatively higher in the highest MMI group than in the other groups. The present study did not account for the effects of maternal serum Mg levels and maternal Mg supplementation; therefore, future prospective studies clarifying the impacts of MMI on childhood wheezing in offspring are warranted, considering the inflammatory and oxidative biomarkers and differences in MMI, maternal serum Mg levels, and Mg supplementation.
Overall, this study implies that MMI during pregnancy would not have significant clinical impact on the incidence of childhood wheezing in offspring. A previous study showed no association between MMI using FFQ collected during the 8th month of pregnancy and childhood asthma in offspring(Reference Nwaru, Erkkola and Ahonen15). This previous study was a population-based prospective birth cohort study that analysed 2441 5-year-old offspring; the mean MMI from food in this population was 474 mg/d, which was significantly higher than that in this study. However, the present results were partly consistent with this previous one because no significant change was found in the OR for childhood wheezing in offspring of mothers with above-ideal MMI levels. Moreover, modifying MMI could not significantly improve the incidence of childhood wheezing in offspring, as well as in the previous study(Reference Nwaru, Erkkola and Ahonen15).
This study has some limitations. First, the incidence of offspring childhood wheezing might have been over- or underreported; thus, the results should be interpreted with caution. However, the information on offspring childhood wheezing was based on the ISAAC questionnaire, and the parents’ reports are expected to have good agreement with the physicians’ diagnosis(Reference Roy, Kocak and Hartman36). Second, there may be unmeasured potential confounding factors to be considered. Detailed data on the daily maternal use of supplementation and drugs, including magnesium oxidate, antiasthmatic medications, antioxidative supplements and antipyretic analgesics, were not included. Moreover, data on childhood exposure to Mg in the offspring are lacking. Finally, there is a concern regarding potential selection bias because several participants who had missing data and satisfied the exclusion criteria were excluded.
In conclusion, contrary to our hypothesis, the highest MMI group showed a slight association with increased childhood wheezing incidence in offspring. Regarding research implications of this study, further studies are needed to investigate the effects of MMI during pregnancy on childhood wheezing incidence in offspring considering the impact of other maternal diets and inflammatory and oxidative biomarkers and differences in MMI, maternal serum Mg levels, and Mg supplementation. However, this association was insignificant based on aMMI and the ideal MMI level. As a clinical implication, this study would imply that MMI during pregnancy would not have significant clinical impact on childhood wheezing incidence in offspring, and modifying MMI to the ideal amount for pregnant women may not significantly decrease the incidence. Therefore, further studies to clarify the association between other prenatal factors and childhood wheezing incidence in offspring are needed to decrease the incidence of childhood wheezing and asthma.
Acknowledgements
The authors are grateful to all participants in this study. The members of the JECS Group as of 2021 are as follows: Michihiro Kamijima (principal investigator, Nagoya City University, Nagoya, Japan), Shin Yamazaki (National Institute for Environmental Studies, Tsukuba, Japan), Yukihiro Ohya (National Center for Child Health and Development, Tokyo, Japan), Reiko Kishi (Hokkaido University, Sapporo, Japan), Nobuo Yaegashi (Tohoku University, Sendai, Japan), Koichi Hashimoto (Fukushima Medical University, Fukushima, Japan), Chisato Mori (Chiba University, Chiba, Japan), Shuichi Ito (Yokohama City University, Yokohama, Japan), Zentaro Yamagata (University of Yamanashi, Chuo, Japan), Hidekuni Inadera (University of Toyama, Toyama, Japan), Takeo Nakayama (Kyoto University, Kyoto, Japan), Hiroyasu Iso (Osaka University, Suita, Japan), Masayuki Shima (Hyogo College of Medicine, Nishinomiya, Japan), Hiroshige Nakamura (Tottori University, Yonago, Japan), Narufumi Suganuma (Kochi University, Nankoku, Japan), Koichi Kusuhara (University of Occupational and Environmental Health, Kitakyushu, Japan) and Takahiko Katoh (Kumamoto University, Kumamoto, Japan).
This work was funded by the Ministry of the Environment, Japan. The findings and conclusions of this article are solely the authors’ responsibility and do not represent the official views of the Ministry of the Environment, Japan.
T. M., H. K., T. F., K. I., H. I., S. Y., A. Y., K. H., H. N., K. F., K. S., A. S. and Y. O.: substantial contributions to study conception and design, acquisition of data, and analysis and interpretation of data. T. M., M. H., S. Y., K. H., K. S., A. S., Y. O., H. N., K. F. and the JECS Group: drafting the article and revising it critically for important intellectual content.
There are no conflicts of interest.
This study was conducted according to the guidelines of the Declaration of Helsinki, and all procedures involving human participants were approved by the Ministry of the Environment’s Institutional Review Board on Epidemiological Studies (approval number: 100910001) and by the ethics committees of all participating institutions. Written informed consent was obtained from all participants.
