|Home | About | Journals | Submit | Contact Us | Français|
Studies exploring the relationship between prenatal vitamin D exposure and childhood asthma have yielded conflicting results. Higher vitamin D intake during pregnancy has been shown to lower the risk of childhood wheeze, yet a study of maternal late-pregnancy serum 25-hydroxyvitamin D suggested higher serum concentrations may be associated with increased childhood asthma.
To assess the relationship between mothers’ serum 25-hydroxyvitamin D status and asthma and wheeze phenotypes in their children at age 6 years. Secondly, to explore the relationship between maternal 25-hydroxyvitamin D status and objective measures of childhood atopy and lung function.
Serum 25-hydroxyvitamin D was measured at 34 weeks’ gestation in the mothers of 860 children born at term. Wheeze was classified as either transient or persistent/late using questionnaire data collated from 6, 12, 24 and 36 months and 6 years. At 6 years spirometry was performed and atopic status was determined by skin prick testing, exhaled nitric oxide was measured in 451 and bronchial hyperresponsiveness in 216 children.
There were no significant associations between maternal late-pregnancy 25-hydroxyvitamin D status and either asthma or wheeze at age 6 years. Maternal vitamin D status was not associated with transient or persistent/late wheeze; no significant association was found between persistent/late wheeze when subdivided according to atopic status. No associations were found with skin sensitisation or lung function.
This study provides no evidence that exposure to higher concentrations of 25-hydroxyvitamin D in maternal serum during late pregnancy increases the risk of childhood asthma, wheeze or atopy.
Vitamin D has multiple effects beyond those upon calcium metabolism and skeletal integrity. A role in asthma and atopic disease has been suggested as many immune cells possess vitamin D receptors and genetic association has been demonstrated between variants in this receptor and asthma. However, the relationship between vitamin D and asthma has proven controversial. Whilst Wjst and Dold proposed rising asthma prevalence to be a consequence of increased consumption of vitamin D-fortified foods, Litonjua and Weiss argued, in contrast, that deficiency of vitamin D may be responsible. Epidemiological support can be found for either viewpoint. A Southampton study found increased infantile eczema and childhood asthma in the children of mothers with higher late-pregnancy serum 25-hydroxyvitamin D. However, this small study was not specifically designed to look at atopic outcomes. Conversely, several cohort studies have found lower maternal vitamin D intake during pregnancy to be associated with increased childhood wheeze risk.[6-9]
Serum 25-hydroxyvitamin D reflects total body vitamin D. Most vitamin D is derived from photosynthesis in the skin. Serum measures may more accurately characterise total exposure than estimated intake records which do not account for sun exposure. Maternal 25-hydroxyvitamin D status was measured in the Dutch KOALA study but was not found to be associated with lung function at age 6 years; however, asthma prevalence, wheezing phenotypes and atopy were not assessed.
This is the first study to prospectively assess the relationship between maternal 25-hydroxyvitamin D status during pregnancy and childhood asthma, wheeze and atopy and the first to consider wheeze separately in atopic and non-atopic children. This is important as persistent wheeze with atopy differs in clinical presentation from non-atopic persistent wheeze and is likely to be of separate aetiology. Data from a large population-based cohort were used to investigate whether higher maternal 25-hydroxyvitamin D status at 34 weeks’ gestation is associated with increased childhood asthma or wheeze risk. Objective measures of lung function and skin sensitisation were used to test the secondary hypotheses that maternal 25-hydroxyvitamin D status is associated with evidence of altered immune or respiratory development.
Participants were mother-child pairs from the Southampton Women’s Survey. Women aged 20-34 years were recruited during 1998-2002; those who became pregnant were followed through pregnancy and their children visited at 6, 12, 24 and 36 months and 6 years. To exclude effects of prematurity, whilst maximising power, infants born < 35 weeks’ gestation were excluded. Children aged six years during 2006-2010 were invited for follow-up; 1485 mother-child pairs were eligible, maternal vitamin status and 6-year follow-up data were available for 860 pairs (Figure 1). Questionnaire data were collected during a home-visit, during this visit spirometry was also attempted. Due to resource limitations clinic-based bronchial provocation testing and exhaled nitric oxide measurements were limited to unselected subsets of participants. Parental consent was obtained and ethical approval was granted by the Southampton and South West Hampshire Local Research Ethics Committee (LREC Number 276/97, 307/97, 089/99 and 06/Q1702/104).
Maternal venous blood was sampled at 34 weeks’ gestation. To ensure stability samples were centrifuged, separated and stored at −80°C within 24 hours of sampling. 25-hydroxyvitamin D concentrations were measured by a Vitamin D External Quality Assessment Scheme compliant laboratory, with quality control samples in each batch. The radioimmunoassay used (DiaSorin, Minnesota, USA) had a coefficient of variability < 10%.
At 11 and 34 weeks’ gestation average frequencies of consumption over the preceding 3 months were recorded using a 100-item food frequency questionnaire (FFQ). The FFQ has been validated against 4-day food diaries and maternal micronutrient concentrations. Nutrient intakes were calculated by multiplying frequency of consumption by nutrient content for each food or supplement. Early and late pregnancy intakes were averaged to provide an estimate of pregnancy intake. Energy adjustment can be used to correct for over or underestimation of intake in the FFQ, however, intake data were not automatically adjusted for total energy intake as supplementary intake vitamin D from supplements do not increase in line with energy intake.
Skin prick testing was conducted at 1 and 3 years using cat, dog, house dust mite (Dermatophagoides pteronyssinus), grass pollens, egg and milk allergens (Hollister-Stier, Spokane, WA); at age 6, tree pollens (ALK Abelló Hørsholm, Denmark) were also tested. For validity a ≥ 3 mm positive and a 0 mm negative control response were required, atopy was defined as any allergen response ≥ 3 mm.
Exhaled nitric oxide (eNO) was measured online by trained research nurses according to ERS/ATS recommendations. Measurements were recorded during controlled expiration at 50 ml/sec using a NIOX® chemiluminescence analyser (Aerocrine, Sweden). A mean value was calculated from three readings where possible. eNO data were normalised using an inverse square root transformation then standardised as a z-score. The sign of the values was reversed so that high untransformed eNO values gave rise to high standardised scores.
Respiratory assessment was conducted by research nurses using questions from the ISAAC core questionnaire wheezing module. Specifically, mothers were asked at each visit whether their child had experienced ‘episodes of chestiness associated with wheezing or whistling in his/her chest since they were last seen’ and at 6 years whether their child had ‘ever been diagnosed with asthma by a doctor’. The asthma outcome was refined to current asthma by including only 6-year-old children diagnosed with asthma who had experienced asthma symptoms or received asthma medication within the last year. The wheeze phenotypes were based upon those of the Tuscon Children’s Respiratory Study; questionnaire data from 6, 12, 24 and 36 months and six years were combined to define:
The persistent and late-onset groups were combined because the late-onset group contained few children (Figure1). The persistent/late wheeze group was sub-classified according to atopic status determined by skin prick testing.
Spirometry was performed according to ATS/ERS guidelines, although without noseclips to avoid discomfort. Flow-volume loops were measured for all children by experienced research nurses using a portable Koko spirometer with incentive software (KoKo version 4; PDS Instrumentation; Louisville, USA). Absolute values of FEV1 and FVC were recorded with and without standardisation for height to explore whether any effect of maternal 25-hydroxyvitamin D status upon wheeze risk was mediated by an effect upon child’s height.
Bronchial hyperresponsiveness (BHR) was measured by bronchial provocation challenge. Methacholine was administered using a dosimeter (Koko; PDS Instrumentation; Louisville, USA) and a compressed air-driven nebuliser (Sidestream®; Respironics, UK). Challenges were conducted according to ATS/ERS guidelines using incremental methacholine concentrations (0.06 mg/ml to 16 mg/ml). Challenges were terminated following the 16 mg/ml dose or a 20% fall in FEV1. BHR was expressed as the inverse of the slope of the regression line through FEV1 drop and logged methacholine concentration: Log.slope=100/[regression slope of FEV1 drop and log10(cumulative methacholine dose) + 10] The constant removes negative values and the inverse transformation produces a normally distributed variable. Lower inverse log.slope values indicate increased BHR.
Poisson regression with robust variance was used to model relative risk for binary outcomes. This is more appropriate than logistic regression for common outcomes where odds ratios cannot be interpreted as relative risks and so are hard to interpret. Transient and persistent/late wheeze phenotypes were mutually exclusive; children suffering one of these types of wheeze were not at risk of the other so relative risks were calculated by comparing children with transient or persistent/late wheeze to those who had never wheezed. Relative risks for persistent/late wheeze with atopy and persistent/late wheeze without atopy were calculated using non-atopic children who had never wheezed as the comparator group. Relationships between maternal 25-hydroxyvitamin D status and continuous outcome variables were explored using linear regression.
Potential confounders were identified a priori and tested for association with each respiratory outcome. Models were developed comprising all the confounding variables listed in Table 1 which were significantly associated with each outcome (p < 0.1). Birthweight and gestation were initially excluded from the multivariable models as they may lie on the causal pathway. Similarly, season and year of blood sampling were initially excluded from the multivariable models to preserve variation in the exposure variable which may drive an effect upon outcome. The analyses were repeated including these variables if they were significantly associated with the outcomes. 25-hydroxyvitamin D was analysed as a continuous variable to maximise power, however, the relationship between this exposure and each outcome was checked for linearity by fitting a quadratic term.
As the analyses were designed a priori to test a limited number of hypotheses and not all the tests were independent, use of a Bonferroni correction was considered over-conservative. As a compromise we focused on results with P-values ≤ 0.025 and considered consistency of the findings in our interpretation.
Participant mothers were broadly similar in terms of asthma, atopy and allergic disorders to those mothers for whom maternal 25-hydroxyvitamin D status or follow-up data were incomplete. Participant mothers were slightly older, taller, less likely to smoke in pregnancy, more likely to be primiparous and of higher educational attainment and social class than those with incomplete data. Participant children were less likely to be exposed to environmental tobacco smoke and more likely to have been breastfed than those with incomplete data (Table 1). Similarly, those children contributing skin prick, spirometry, eNO or BHR data were broadly similar to those that did not.
The median, (IQR) maternal 25-hydroxyvitamin D concentration was 59.0 nmol/l, (40.5-84.9 nmol/l). The highest serum 25-hydroxyvitamin D value was 203 nmol/l, 29% of women had values greater than 80 nmol/l and 10% of women had serum levels greater than 110 nmol/l. Serum 25-hydroxyvitamin D concentrations were slightly lower in those women lost to follow-up (53.0, (38.5-79.2) nmol/l), probably reflecting socioeconomic and associated lifestyle factors. Eighty seven children (10.1%) had current doctor-diagnosed asthma, whilst 504 (58.9%) had experienced wheeze at or before six years of age. A total of 137 (16.0 %) children were assigned to the persistent/late wheeze phenotype; of these 51.1 % were atopic and 48.9 % non-atopic (Table 2). Technically acceptable measures of FEV1, BHR and eNO were available from 739, 216 and 451 children respectively (Figure 1).
There was no significant association between maternal 25-hydroxyvitamin D status at 34 weeks’ gestation and current asthma; neither was 25-hydroxyvitamin D status associated with wheeze at or before 6 years of age nor current wheeze in the 12 months preceding the 6-year assessment (Table 3). There were no significant associations with the transient or persistent/late wheeze phenotypes and subdividing the persistent/late phenotype by atopic status did not reveal any significant associations (Table 3). There was no evidence for a non-linear relationship between maternal 25-hydroxyvitamin D and any wheeze phenotype (data not shown).
Maternal 25-hydroxyvitamin D status at 34 weeks’ gestation was not associated with skin sensitisation at 1, 3 or 6 years or with exhaled nitric oxide at age 6 years (Table 4). There was no evidence for a non-linear relationship between maternal 25-hydroxyvitamin D and either skin sensitisation or eNO (data not shown).
Maternal 25-hydroxyvitamin D at 34 weeks’ gestation was not associated with either absolute or standardised values of FEV1 or FVC at 6 years. Maternal 25-hydroxyvitamin D status was not associated with BHR (Table 5). There was no evidence for a non-linear relationship between maternal 25-hydroxyvitamin D and any measure of lung function (data not shown).
Birthweight was not associated with any outcome and was therefore not considered a confounder. Gestation, however, was associated with all wheeze and skin sensitisation outcomes except current asthma and persistent/late wheeze with atopy; the absence of association between maternal 25-hydroxyvitamin D and these variables remained when gestation was included in the multivariable models. Birthweight was associated with absolute measures of FEV1 and FVC, but including birthweight in the multivariable models did not reveal any associations between maternal 25-hydroxyvitamin D and these measures (Tables E1-3). When adjustments were made for season and year of blood sampling the absence of any association between maternal 25-hydroxyvitamin D and childhood wheeze, atopy or lung function variables was unchanged (Tables E4-6).
The median (IQR) average total daily maternal vitamin D intake was 4.2 mcg/day (3.0 - 6.7 mcg/day). During early pregnancy 39% of women took supplements containing vitamin D, whilst during late pregnancy, 22% supplemented with vitamin D. For women taking supplements, median supplementary-intake (IQR) was 4.1 mcg/day (1.5-8.3 mcg/day) in early pregnancy and 5.8 mcg (2.5-12.2 mcg/day) vitamin D /day, in late pregnancy. Only 11.5% of women achieved an average intake of 10 mcg/day, currently recommended by the department of health. Correlation coefficients for early, late and average intake with status at 34 weeks’ gestation were 0.25, 0.33 and 0.33, respectively. Neither total nor food-derived maternal intake was associated with asthma, or any wheeze outcome (Tables E7 & E10). No significant associations were found between intake and skin sensitisation, (Tables E8 & E11) or any measure of lung function (Tables E9 & E12). A statistically significant positive association was found between food-derived vitamin D and eNO (Table E5) but the lack of a similar association with total intake (Table E2) suggests that this is not a clinically robust association. There was no evidence for a significant association between total maternal vitamin D intake and any outcome when, for consistency with previous birth cohort analyses, the results were energy-adjusted and analysed as a categorical variable according to quartile of vitamin D intake (Tables E13-15).
This study found no evidence that higher maternal serum 25-hydroxyvitamin D in late pregnancy is associated with an increased risk of asthma or atopy. There were no significant associations between maternal 25-hydroxyvitamin D status at 34 weeks’ gestation and asthma, transient or persistent/late wheeze, skin sensitisation at 1, 3 or 6 years or eNO, FEV1, FVC or BHR at age 6 years.
It has been suggested that serum 25-hydroxyvitamin D levels in pregnancy should be above 80 nmol/l, 29% of the women in this study had serum 25-hydroxyvitamin D concentrations above this level. Both supplementation and 25-hydroxyvitamin D levels were higher in the current study than in our previous cohort where a positive association was found between maternal 25-hydroxyvitamin D status and childhood eczema and asthma. Whilst a high prevalence of supplementation might reduce the ability of the present study to detect any effect of dietary insufficiency, failure to confirm a harmful effect of high 25-hydroxyvitamin D status cannot be attributed to lower exposure; it is more likely that the earlier study was underpowered to assess specific clinical outcomes.
Vitamin D supplementation during pregnancy is known to benefit calcium metabolism and bone health and many believe may protect against cardiovascular, autoimmune, and malignant disease via ‘fetal imprinting’. The National Institute for Health and Clinical Excellence has suggested that pregnant women may wish to consider vitamin D supplementation. Recently a randomised controlled-trial has demonstrated that daily supplementation with 4000 IU vitamin D can increase maternal serum 25-hydroxyvitamin D concentration without adverse events. However, few adequately powered studies have considered the effects of increased maternal 25-hydroxyvitamin D upon relevant clinical outcomes.
This study found no evidence that higher late-pregnancy maternal serum 25-hydroxyvitamin D is associated with increased asthma. Although an inverse relationship between energy-adjusted maternal vitamin D intake and asthma at age 5 was found in a Finnish cohort, this was significant only for food-derived not total intake. Associations with food derived intake only may be vulnerable to confounding by other nutrients present in vitamin D-rich food and socioeconomic factor or they may arise as a result of multiple comparisons. The majority of studies reporting inverse associations between early vitamin D exposure and adverse respiratory outcomes, reported associations with wheeze but not asthma. Many of these studies relied upon maternal intake data[6-9] and relatively short follow-up.[7-8]
Vitamin D intake studies are vulnerable to confounding by socio-economic and lifestyle factors and by the effects of other nutrients found in vitamin D-containing foods. Such confounding has been suggested to explain the absence of an association between 25-hydroxyvitamin D status and lung function in adults with chronic obstructive pulmonary disease, despite an association with vitamin D intake; absence of association between vitamin D receptor genotype and lung function strengthened this argument. Studies with short follow-up are obliged to use childhood wheeze as an outcome in the absence of a reliable asthma diagnosis in young children. Follow-up to 5 years was conducted in a New Zealand cohort and an inverse association found between cord blood 25-hydroxyvitamin D and wheeze, but again no association was found with asthma. An inverse association was also found between cord blood 25-hydroxyvitamin D status and respiratory infections in early infancy, leading these authors to conclude that beneficial effects of vitamin D upon innate immunity may indirectly reduce wheeze risk in early childhood. The present study did not address the issue of childhood infection.
Whilst it remains possible that vitamin D exerts an affect upon immune predisposition to wheeze with infection, rather than altered atopic immunity, an alternative explanation for the lack of demonstrable association between early vitamin D exposure and asthma could be the existence of different asthma phenotypes. Subdividing the persistent/late wheeze phenotype by atopic status did not reveal any association with maternal vitamin D status, although it remains possible that the phenotypes used in this study were excessively heterogeneous and that early vitamin D exposure may have differential effects upon late-onset compared with persistent wheeze, for example.
Animal studies suggest vitamin D may promote a proallergic Th2 phenotype. Similarly, epidemiological evidence suggests high early life exposure to vitamin D supplementation or high maternal 25-hydroxyvitamin D serum levels might predispose to allergic disorders. No previous prospective epidemiological study, however, has investigated the relationship between maternal 25-hydroxyvitamin D status and objective measures of atopy. In this respect, the null findings in this study are reassuring: higher serum 25-hydroxyvitamin D concentrations in late pregnancy do not appear to increase skin sensitisation at 1, 3 or 6 years or eosinophilic airways inflammation at 6 years.
Early vitamin D exposure has been shown to alter the volume dependence of lung mechanics in an animal model, suggestive of altered tissue structure. Altered development affecting lung structure and airway calibre would also be consistent with the results of maternal intake studies. However, this study confirms, in a larger cohort with more extensive characterisation of lung function including BHR, the findings of the KOALA study; there was no evidence of a clinically significant alteration of lung function according to maternal late-pregnancy 25-hydroxyvitamin D status.
As the repeatability of the serum 25-hydroxyvitamin D levels is relatively low, there is a significant chance that a single serum measurement will lead to misclassification of exposure. As this misclassification is random, this may bias studies, such as this, based upon single serum samples toward the null or no effect. Furthermore, epidemiological studies are limited in their ability to discriminate causal from closely linked factors; this study cannot exclude the existence of a relationship between vitamin D exposure and wheeze or atopy which is hidden by an opposing relationship between incompletely controlled for seasonal and other factors upon these outcomes. Another feature of the study design which may have limited the likelihood of identifying an association between maternal 25-hydroxyvitamin D status and childhood wheeze outcomes is the use of frequent prospective questionnaires; this may have set too low a threshold to reflect significant pathology.
Whilst the present study did not have complete follow-up and those followed up differed from those who were not in terms of several socioeconomic variables, this should not alter the main conclusions unless the nature of any relationship between maternal 25-hydroxyvitamin D status and wheeze or atopic outcomes differed according to socioeconomic status or if the relationship were non-linear. We have no evidence to support either assertion. The null results in this study may have arisen as a consequence of measuring 25-hydroxyvitamin D status in late pregnancy only, however, whilst much respiratory development, particularly that of the airways, occurs early in pregnancy, significant maturation of the immune system is believed to occur in late pregnancy. Furthermore, the null result was supported by analyses based upon intake data which covered both the first and second trimesters.
In summary, neither higher late-pregnancy maternal 25-hydroxyvitamin D status nor high vitamin D intake during pregnancy was significantly associated with asthma or any wheeze phenotype. Moreover there was no evidence that early exposure to higher concentrations of 25-hydroxyvitamin D had a deleterious effect upon either lung function or atopic sensitisation. Together, these findings suggest the risk posed by vitamin D supplementation in terms of asthma and atopic diseases may not be a concern.
The research was supported by infrastructure provided by the Southampton NIHR Respiratory Biomedical Research Unit and Nutrition, Diet and Lifestyle Biomedical Research Centre. Clinical investigations were conducted in Southampton Wellcome Trust Clinical Research Facility. We would also like to gratefully acknowledge the expertise of Professor Ramasamyiyer Swaminathan and Dr Sasala Wickramasinghe for providing 25-hydroxyvitamin D radioimmunoassays in the Chemical Pathology department at St. Thomas’ Hospital, London.
Funding: This work within the Southampton Women’s Survey has been funded by the Medical Research Council, University of Southampton, British Heart Foundation, and the Food Standards Agency (contract no N05071). The research is supported by infrastructure provided by the Southampton NIHR Respiratory Biomedical Research Units and Nutrition, Diet and Lifestyle Biomedical Research Centre. Dr KC Pike was supported by a grant from the British Lung Foundation.
The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive licence on a worldwide basis to the BMJ Group and co-owners or contracting owning societies (where published by the BMJ Group on their behalf), and its Licensees to permit this article (if accepted) to be published in Thorax and any other BMJ Group products and to exploit all subsidiary rights, as set out in our licence.
Competing interests: All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare that (1) none of the authors have support from any company for the submitted work; (2) KCP, HMI, SMR, JSAL, CC, NCH, KMG & GR have no relationships with companies that might have an interest in the submitted work in the previous 3 years; (3) their spouses, partners, or children have no financial relationships that may be relevant to the submitted work; and (4) KCP, HMI, SMR, JSAL, CC, NCH, KMG & GR have no non-financial interests that may be relevant to the submitted work.