Asthma is one of the most common chronic conditions among children [1
]. It is a complex heterogeneous disease that affects both the respiratory and the immune system [2
]. It manifests by many phenotypes that vary by sex, age at onset, presence of obesity, as well as the severity of atopy, allergic sensitization, bronchial obstruction, and hyperresponsiveness. Although genetic factors and childhood exposure to environmental triggers, such as tobacco smoke, air pollution, viral infections or aeroallergens play a major role in the development of childhood asthma [3
], it has been suggested that environmental exposures during gestation may also be important [4
]. Asthma is more prevalent among boys until puberty when a shift towards higher prevalence among girls and women is observed [5
]. In children, this is thought to be due to sex differences in lung development and inflammatory profile [7
Vitamin D is a fat-soluble vitamin and a secosteroid hormone playing an important role in both skeletal and non-skeletal functions [8
]. It readily crosses the placenta and the fetal supply is totally dependent on the supplies of the mother [9
]. Results from animal and human studies have shown that prenatal vitamin D insufficiency may influence both the intrauterine immune system [10
] and lung development [11
], causing permanent changes to these systems. Such changes constitute a plausible biological basis to suggest that vitamin D insufficiency during gestation may have a programming effect contributing to the risk of childhood asthma. This is especially relevant given the high prevalence of vitamin D insufficiency in pregnant women [13
]. The research on the association between prenatal vitamin D status and childhood asthma risk in the offspring is quite extensive. Results from several systematic reviews and meta-analyses of observational studies, however, have been inconsistent [14
]. Two randomized clinical trials (RCT) have recently been conducted; both found non-significant 20% reduced risk of asthma/recurrent wheeze among 0–3 year old children whose mothers were supplemented with vitamin D during pregnancy [19
], while a combined analysis of the two trials showed a significant reduction in offspring’s asthma/recurrent wheeze risk after vitamin D supplementation in pregnancy [21
]. Despite sex differences in asthma prevalence and possible sex difference in the effect of vitamin D on asthma development, few observational studies and none of the RCTs have run analyses differentiated by sex. Furthermore, most of the previous studies had limited numbers of participants and short follow-up periods [14
In Denmark, between 1937 and 1985, it was mandatory to fortify margarine with 1.25 µg vitamin D per 100 g [22
]. The fortification accounted for on average 13% (3–29%) of total vitamin D intake from food in the Danish population [22
]. Despite that, the mandatory margarine fortification policy was canceled in June 1985. This study utilized the design of this societal experiment grounded on the distinct in time termination of the Danish margarine vitamin D fortification policy. The objective of the present study was to examine if children born to women exposed to the margarine fortification policy with a small dose of extra vitamin D during pregnancy had a reduced risk of developing asthma until age 9 years compared to children whose mothers were unexposed to the fortification policy during pregnancy. Furthermore, this study examined whether the association between exposure and asthma risk varied by age, sex, and month of birth.
Overall, we did not see a difference in the risk of asthma between children of both sexes age 0–9 years born to mothers exposed to the margarine fortification policy with a small extra dose of vitamin D during their entire pregnancy compared to those who were born to mothers who were unexposed. However, our analysis indicated that young children (0–3 years old), and in particular boys born to exposed mothers, had more than 20% lower hazards for developing asthma compared to boys born to unexposed mothers.
Several observational studies have examined the association between prenatal vitamin D and the development of asthma in the offspring, however, the results have been mixed. There are several potential reasons for this inconsistency. The observational studies have differed in study design, number of participants, adjustment for covariates, the time at exposure assessment, the source material for biomarker analysis, the concentration of maternal or cord blood 25(OH)D, and the assay method. Early observational studies looking at vitamin D intake during pregnancy, usually assessed by food frequency questionnaire (FFQ), and later development of asthma and wheezing, tended to find inverse associations [30
]. Vitamin D may have served as a marker of a healthier diet in general and other dietary-related factors such as differences in the maternal microbiome [33
]. The cohort studies examining maternal or cord blood 25(OH)D tended to find no association between 25(OH)D level in pregnancy and offspring asthma [34
]. This was true across different timing of exposure and outcome assessment as well as geographical settings. Total 25(OH)D level may not be a sufficient indicator of the biologically available vitamin D, as other factors such as level of vitamin D binding protein could influence the metabolism and bioavailability [37
]. These studies also did not look at the difference in asthma risk between different sex nor stratify according to pre-pregnancy vitamin D status which may be more important for reducing asthma risk than vitamin D status in pregnancy. In two recent randomized trials where mothers were supplemented with daily dose of 60 and 100 µg vitamin D respectively during pregnancy [19
], inverse associations between 25(OH)D level and asthma/recurrent wheeze were strongest among women with a high 25(OH)D status at baseline (≥75 nmol/L). This could indicate that the programming effect already takes place in early pregnancy or even prior to pregnancy, and therefore, having optimal vitamin D status already at conception might be important for reducing the risk of childhood asthma. This may be supported by the results from observational studies assessing vitamin D intake as women with higher vitamin D intake during pregnancy are likely to have had the same dietary pattern prior to pregnancy and thus have higher vitamin D status at conception. In our study, the mothers were exposed to extra vitamin D from fortification both during a long period prior to pregnancy and during the entire pregnancy. Thus, it is expected that the exposed mothers had a higher vitamin D status pre-pregnancy than unexposed mothers. It is important to stress that the vitamin D dose administered to the subjects in the two trials was substantially higher than what women were exposed to from fortification in our study. We have previously calculated, based on the fortification dose (1.25 µg/100g) and margarine intake statistics, that an average additional 0.4–0.6 µg of vitamin D per day could be provided by the fortified margarine [39
]. Compared to vitamin D supplementation in the two trials (i.e. 60 and 100 µg), and in light of the recommended total intake of 10 µg vitamin D per day for skeletal actions [40
], the Danish margarine fortification policy provided a very low extra vitamin D dose. However, the optimal vitamin D dose for the prenatal development of the immune system and/or lung development is unknown. Much higher supplementation doses provided in the two previously mentioned trials seemed to be inadequate to reach a level of sufficiency among women during their pregnancy to prevent asthma development in the offspring. In our study, a constant intake of a very low extra dose of vitamin D via fortified foods consumption over entire pregnancy (and before pregnancy), might have been sufficient to reduce the risk in the offspring with the highest risk of developing asthma, the youngest boys [7
]. Methodological advantages and disadvantages of our societal experiment design study, if compared to clinical trials, are discussed below under strengths and limitations.
Asthma phenotypes are defined based on age at debut, the presence of atopy, allergic sanitization, lung function, responsiveness to steroids, obesity, sex, and inflammatory profile. Many of these different phenotypes vary according to age, making age an important parameter in identifying asthma phenotypes [42
]. Therefore, to try to isolate specific phenotypes, we stratified the analysis by age at onset. Our results of a slight reduction in the risk of inpatient asthma admission among 0–3 years old boys born to mothers who were exposed to the margarine fortification policy with extra vitamin D during pregnancy are in accordance with the results from the combined analysis of the two recently published RCTs showing a 26% reduction in the risk of asthma/recurrent wheeze at age 0–3 years among those born to mothers who took vitamin D supplementation during pregnancy [21
]. This could indicate that vitamin D has an effect on asthma phenotypes that are prevalent in the youngest children, especially boys, or that the programming effect is relatively weaker than the effect of the risk factors accumulating during childhood. On the other hand, among children younger than 6 years, and particularly those younger than 3 years, the asthma diagnosis is based on symptoms (wheezing, cough, and breathlessness), clinical history of these symptoms, or family history of asthma or atopy, whereas among older children and adults asthma is diagnosed based on symptoms and confirmed with objective measures (spirometry, beta2 agonist reversibility test, bronchial provocation test, and peak flow measurements) [3
]. Wheezing is the most common clinical manifestation of asthma onset. As very young children who experience severe and persistent wheezing are more likely to also have asthma in childhood or adulthood [43
], persistent wheezing in very young children usually is considered to be an asthma indication. At the same time, many young children often experience wheezing due to viral respiratory tract infections (RTI) without having or later developing asthma [44
]. Notably, a recent review on the association between maternal vitamin D status and RTI in offspring showed an inverse association between exposure to vitamin D and RTIs (highest vs. lowest 25(OH)D level: OR 0.64, 95% CI: 0.47–0.87), but no association with asthma or wheezing [18
]. Hence, it cannot be ruled out that the protective effect observed among the 0–3-year-olds in our study was driven by a proportion of children misdiagnosed with asthma because of wheezing due to viral RTI.
In regard to the sex differences, there are anatomic differences in lung size, maturity, and function between the sexes with boys having larger lungs than girls, however, girls have higher forced expiratory flow rates [7
]. The lungs of newborn boys are also less mature than the lungs of newborn girls [45
] and thus are more vulnerable to respiratory infection, and consequently, asthma. Moreover, severe allergic asthma, or asthma with multiple sensitizations to allergens, characterized by early onset, high eosinophil count, and a low response to corticosteroid, is more prevalent among boys [29
]. Children included in our study were children with asthma diagnosed during hospitalization, thus our outcome measure captured the more severe cases of asthma. It is therefore possible that the protective association observed among the youngest boys in our study was driven by the effects on an allergic type of asthma. Interestingly, despite sex differences in the prevalence of asthma, and different asthma phenotypes in boys and girls [44
], very few studies have examined whether the effect of prenatal vitamin D exposure on the development of asthma is sex-specific. We identified only one other study that conducted sex-stratified analyses, and similarly to the results of the present study, the authors found an inverse association between maternal vitamin D and asthma risk at 6 years among boys only [46
]. Consequently, vitamin D status in fetal life may contribute differently to the regulation of the immunologic responses in boys and girls.
Based on the present study, we cannot conclude that exposure to the margarine fortification policy with extra vitamin D during the prenatal period influences the risk of offspring asthma. However, our results and those of prior studies may indicate that vitamin D intake, even in small amounts, early in pregnancy or even before conception could influence the risk of asthma development, especially among more vulnerable groups such as very young male children. Future studies of the association between prenatal vitamin D and offspring asthma risk should assess vitamin D status or intervene with vitamin D supplements periconceptually or prior to pregnancy. Furthermore, they should focus on different phenotypes of asthma, using biomarkers shown to have a good prognostic value, and assessment of possible sex differences.
Strengths and Limitations
In this study we utilized a societal experiment of an abrupt abolishment of an obligatory vitamin D fortification, that exposed all individuals living in Denmark to extra vitamin D from fortified margarine during a distinct period of time when the fortification policy was implemented, and not thereafter. A semi-ecological study design has both strengths and limitations.
The strength is that we could include all individuals born in Denmark from entire birth cohorts from the Danish population, capturing all inpatient asthma admissions, which makes our results generalizable for the entire Danish population. This was possible due to the excellent Danish administrative and health registers, extensively used in Danish epidemiological research [47
]. Additionally, asthma diagnoses in DNPR have been previously validated [27
]. Another strength of the design was that potential confounding (i.e., from differences in socioeconomic status, obesity, maternal lifestyle and diet during pregnancy) can be considered to be equally distributed in both the exposed and the unexposed groups as they included all individuals born in adjacent birth cohorts around the time of the policy change. Nevertheless, secular trends in potential confounders during this short period of time between 1983–1988 could have introduced residual confounding. Intake of margarine was remarkably stable during the years around the policy change, and we were not able to identify changes in national recommendations for intake and/or supplementation of vitamin D or in fortification practice. However, a study from Denmark shows that smoking among women, a risk factor for offspring asthma, decreased slightly during this period [49
]. This would have attenuated our findings and therefore speaks in favor of a true association between exposure to the margarine fortification policy with extra vitamin D and the development of asthma. Furthermore, we know that overweight and obesity, a risk factor for asthma, has been increasing among women [50
], and from 1987–1988 there were changes in fiscal policy in Denmark that caused economic challenges for many households [51
]. It can be speculated that this economic crisis influenced overall diet quality, such as decreased the intake of fish or fish oil, which is also shown to have a protective effect on asthma development among offspring.
The limitation of this semi-ecological study design is that information on our exposure represents ecological data; we do not have any information on actual vitamin D intake, neither from the fortified margarine nor from other dietary sources and hence cannot make actual intake recommendations. A further limitation of the study is that we only included in-patient discharge diagnoses and not outpatients diagnoses and diagnoses by general practitioners, as the DNPR does not collect this information. Consequently, we have a subgroup of asthmatic children, most likely those with severe or uncontrolled asthma, and can generalize our results only to this group. Another study limitation related to the nature of register data, which does not specify different phenotypes of asthma, nor clinical parameters. To approximate this information, we conducted analyses stratified by age and sex; however, even though the phenotypes differ by age and sex, there are other parameters that are equally important, such as sensitization to allergens, responsiveness to steroids, and inflammatory profile that would also be interesting to examine.