Abstract

INTRODUCTION

Recent epidemiologic studies have shown an inverse association of maternal intake of vitamin D during pregnancy with childhood wheezing^{1, 2} and possibly childhood asthma.^{3, 4} Similarly, cross-sectional studies have found associations of low vitamin D status with both persistent wheezing in African American children⁵ and worse asthma severity among asthmatic children.^{6, 7}These studies and others have led some investigators to propose that insufficient exposure to sunlight, with subsequent reduction of vitamin D, may be a cause of the asthma epidemic.^{8, 9}

Complicating these studies, however, is the fact that most early wheezing is transient and not associated with classic atopic asthma,¹⁰ raising the question of whether vitamin D status influences risk for infectious wheezing or the allergic components of asthma. Indeed, recently published data from a New Zealand birth cohort showed that cord blood 25(OH)D levels were inversely associated with risk of respiratory infection and wheezing but had no association with incident asthma at age 5 years.¹¹ Moreover, vitamin D supplementation was recently shown to reduce influenza A infection in a randomized controlled trial of Japanese school children.¹²

The relation of vitamin D to asthma pathogenesis is further complicated by studies suggesting that vitamin D levels may be associated with increased markers of atopy and increased risk ofallergic conditions.^13–16 It has even been proposed that increased (not decreased) intake of vitamin D is responsible for the global epidemic of allergic disease.¹⁷

Using an established birth cohort in Tucson, Arizona, a desert city at 32°N latitude, we examined associations between cord blood levels of 25(OH)D with both atopy outcomes (total IgE, detectable inhalant IgE, and skin-prick test positivity) and childhood allergic diseases (allergic rhinitis and asthma) through age 5 years. Our objective was to study these associations in a population with high year-round exposure to direct sunlight and, consequently, higher levels of circulating 25(OH)D.

METHODS

RESULTS

Children with both cord blood 25(OH)D data and at least one outcome measure are included in this analysis (n=219). Included children did not differ significantly from other children enrolled in the IIS (n=263) with respect to gender, maternal or paternal atopy, maternal or paternal asthma, household income or household smoking (all P>0.05). However, compared to excluded children, included children were more likely to have been born between May and October, and to have mothers who were older, non-Hispanic white, or had more years of education. A complete breakdown of these traits for included and excluded subjects is provided in a supplementary section (S-Table I).

Cord Blood 25(OH)D Levels

Cord blood levels of 25(OH)D ranged from 13 to 173 nmol/L, with 58 children (26.5%) having levels less than 50 nmol/L; 90 children (41.1%) with levels between 50 to 74.9 nmol/L; 47 children (21.5%) with levels between 75 to 99.9 nmol/L; and 24 children (11.0%) with levels greater than or equal to 100 nmol/L. The median cord level of 25(OH)D was 64 nmol/L (interquartile range, 49 to 81 nmol/L) (Figure 1).

Open in a separate window

Figure 1

Histogram of cord blood 25(OH)D levels.

Demographic traits and prenatal exposures related to cord blood 25(OH)D levels included maternal ethnicity, maternal education, household income, birth season, and household smoking (Table I). When all of these factors were entered into a multivariable model of 25(OH)D level, only maternal ethnicity, birth season and household smoking remained statistically significant (data not shown). Both forward and backward stepwise regression modeling confirmed these same 3 factors as predictors of 25(OH)D levels, which were therefore considered as potential confounders in outcome analyses.

Table I

Demographic traits and prenatal exposures, by cord blood 25(OH)D levels.

				Cord 25(OH)D level (nmol/L)
			<50	50–74.9	75–99.9	≥100.0
		N	%(n)	%(n)	%(n)	%(n)	P-value*
Total		219	26.5 (58)	41.1 (90)	21.5 (47)	11.0 (24)	-
Gender
	Male	107	30.8 (33)	43.0 (46)	16.8 (18)	9.3 (10)
	Female	112	22.3 (25)	39.4 (44)	25.9 (29)	12.5 (14)	0.24
Maternal Ethnicity
	Non-Hispanic White	168	20.2 (34)	42.3 (71)	23.8 (40)	13.7 (23)
	Hispanic	39	48.7 (19)	38.5 (15)	10.3 (4)	2.6 (1)
	Other/Unknown	12	41.7 (5)	33.3 (4)	25.0 (3)	0.0 (0)	0.004
Household Income
	< $25,000	28	50.0 (14)	35.7 (10)	10.7 (3)	3.4 (1)
	$25,000–45,000	50	18.0 (9)	44.0 (22)	22.0 (11)	16.0 (8)
	> $45,000	127	22.8 (29)	43.3 (55)	24.4 (31)	9.4 (12)	0.06
Maternal Atopy
	No	70	32.9 (23)	37.1 (26)	20.0 (14)	10.0 (7)
	Yes	143	23.8 (34)	44.1 (63)	23.1 (33)	9.1 (13)	0.52
Maternal Asthma
	No	169	26.6 (45)	43.8 (74)	19.5 (33)	10.1 (17)
	Yes	43	25.6 (11)	30.2 (13)	27.9 (12)	16.3 (7)	0.25
Birth Season
	Nov – Apr	93	31.2 (29)	46.2 (43)	18.3 (17)	4.3 (4)
	May – Oct	126	23.0 (29)	37.3 (47)	23.8 (30)	15.9 (20)	0.02
Househod smoking
	No	176	23.9 (42)	41.5 (73)	21.6 (38)	13.1 (23)
	Yes	43	37.2 (16)	39.5 (17)	20.9 (9)	2.3 (1)	0.10
Maternal age (years)
	mean (sd)	219	28.7 (6.2)	29.5 (6.4)	30.7 (5.8)	29.8 (6.5)	0.42
Maternal education (years)
	mean (sd)	219	14.8 (2.7)	15.7 (2.7)	16.0 (2.5)	15.5 (1.4)	0.08

Open in a separate window

^*P-values from Fisher's exact tests and one-way ANOVA (for maternal age and education).

DISCUSSION

This birth cohort study shows a nonlinear relation between vitamin D status at birth and immune system markers of atopy through age 5 years. Both low and high levels of cord blood 25(OH)D were associated with higher levels of total IgE and inhalant specific IgE. In addition, we showed that high, though not low, levels of 25(OH)D at birth were significantly associated with increased SPT positivity in children at age 5. A similar pattern was seen for allergic rhinitis at age 5, although this relation was not statistically significant. In contrast, no relation was evident between cord blood 25(OH)D and physician diagnosed active asthma at age 5 years.

Several recent studies suggest that low vitamin D status at birth is associated with increased risk for childhood wheezing illness. Camargo et al.¹¹ found an inverse relation of cord blood 25(OH)D to wheezing by age 15 months, 3 years and 5 years in a New Zealand birth cohort. Earlier studies are consistent with these 25(OH)D findings by showing that maternal vitamin D intake in pregnancy, through either foods or supplementation, is inversely related to wheeze^{2, 4} or recurrent wheeze¹ in young children. In addition, recent studies have shown associations between vitamin D deficiency and asthma severity.^{6, 23} Given this body of research and the strong association between wheezing illness and IgE levels,^{24, 25} we were not surprised to find that low levels of 25(OH)D at birth were associated with elevated total and inhalant specific IgE.

We also found, however, that high levels of 25(OH)D at birth appear to convey risk for elevated markers of atopy, specifically increased total IgE, inhalant specific IgE, and positive skin test response prevalence at age 5 years. Other investigations have similarly suggested that vitamin D supplementation may be associated with increased risk for atopy, variously defined, and actual allergic diseases. In a historical cohort study, Hypponen et al. found that supplementation with high daily doses (≥2000 IU/d) of vitamin D in the first year of life was associated with increased SPT response and allergic rhinitis in more than 5000 subjects at age 31.¹⁵ Two small prospective studies showed 1) that vitamin D intake during infancy was positively associated with atopic dermatitis with and without asthma or rhinitis symptoms at age six;¹³ and 2) that maternal serum levels of 25(OH)D were positively associated with childhood eczema and asthma.¹⁴ A much larger prospective study found that supplementation during infancy with vitamins A and D was associated with increased risk of asthma and allergic sensitization at 4 years of age.¹⁶ Finally, a large cross-sectional study in over 7000 adults showed that both low and high serum 25(OH)D levels were associated with increased total IgE,²⁶ a finding similar to that which we found in this longitudinal study of children. This evidence of a nonlinear relation with atopy suggests that it will be important to disentangle relations with immune system markers of atopy from those associated with actual allergic diseases to explicate the complex health effects of vitamin D.

The association between high levels of 25(OH)D at birth and subsequent markers of atopy was not consistently reflected in clinical outcomes in this population. Although statistical power was limited, there was a non-significant trend toward higher rates of physician-diagnosed allergic rhinitis at age 5 in association with elevated levels of 25(OH)D that mirrored our findings pertaining to IgE and skin test response. By contrast, we found no evidence of an association between cord blood levels of 25(OH)D and risk of childhood asthma. The literature regarding relations of vitamin D and allergic diseases is also inconsistent. Ekkola et al. showed an inverse relation of maternal dietary intake of vitamin D with both asthma and allergic rhinitis in their children to age 4.³ Other studies have reported a positive association between adult allergic rhinitis and concurrent serum 25(OH)D levels²⁷ or vitamin D supplementation in the first year of life.¹⁵ Since both total IgE and inhalant allergen specific IgE are strongly associated with subsequent risk of allergic rhinitis and asthma,²⁸ it will be important to continue to assess these associations as our population ages, and to address this complex issue in larger cohorts with greater statistical power.

The apparent nonlinear relation between vitamin D status and allergic/asthmatic outcomes observed both in this and previous studies may result from two opposing mechanisms operating simultaneously. On one hand, there is evidence from epidemiologic studies that vitamin D deficiency is a risk factor for respiratory infection. For example, Camargo et al. showed an inverse association of cord vitamin D levels and respiratory infections in almost 1000 New Zealand infants.¹¹ Similarly, a randomized clinical trial among Japanese school children showed that supplementation with 1200 units of vitamin D was associated with significantly reduced risk of incidence of influenza A infection, and, among asthmatics, a reduced frequency of asthma attacks.¹² On the other hand, the positive association between high levels of 25(OH)D levels at birth with total and inhalant specific IgE may be attributable to a different mechanism. Vitamin D is an immune system regulator, involved in both cell-mediated and humoral antibody responses, and supports a Th2-mediated anti-inflammatory cytokine profile.²⁹ In a murine model of experimental allergic asthma, mice that lack a critical vitamin D receptor fail to develop airway inflammation, eosinophilia, or airway hyperresponsiveness, despite high IgE concentrations and elevated Th2 cytokines.³⁰ Similarly, mice challenged with ovalbumin who received injections of 25(OH)D showed elevations in production of Th2 cytokines (IL-4 and 13) and total IgE, while the inflammatory response in the lungs was concurrently reduced, as indicated by eosinophils and IL-5 in BAL fluid and lung tissue.³¹ These studies suggest that increased levels of 25(OH)D might lead to sustained Th2 skewing in immune system markers in circulation, while simultaneously creating opposing, anti-inflammatory effects in the lung. Additional research into these complex relationships is clearly warranted.

This study has several strengths. First, it is one of the few birth cohort studies to have measured 25(OH)D levels in cord blood as opposed to estimating infant levels based on 25(OH)D levels in pregnant women ¹⁴ or maternal intake of vitamin D during pregnancy.^1–4 Secondly, we collected outcome data at multiple ages, providing more robust measures of potential health effects than single time-point outcomes could provide. Finally, our study population was unique with respect to high annual sun exposure (i.e. 350 days of sunshine annually), providing a natural setting for assessing effects of higher 25(OH)D levels. Indeed, cord blood 25(OH)D levels in our population were considerably higher than those recently reported in a New Zealand study,³² and were higher than levels in women during late pregnancy in the United Kingdom.¹⁴ The large range of 25(OH)D levels in our population is a substantial strength given that our goal was to assess health effects of vitamin D in a population with high sun exposure.

This study also had several potential limitations. First, we cannot be sure that birth levels of 25(OH)D are causally linked with the outcomes studied. High levels of 25(OH)D at birth may be indicative of a lifestyle of increased sun exposure or increased supplementation, and thus children with high levels of 25(OH)D at birth may have high levels throughout childhood, which may be the true driver of the observed associations. Unfortunately, as this vitamin D analysis was not planned when this study was begun in 1997, our questionnaires did not specifically ask about vitamin D supplementation. Second, the number of subjects with allergic rhinitis (n=38) or asthma (n=29) in this analysis was relatively small, which limits our power to see relations with these outcomes, either positive or negative, with cord blood levels of 25(OH)D. Third, subjects were followed only to age 5 years, at which age asthma is still evolving from being infection-related to being predominantly related to allergy. For example, in an older cohort we have shown similar null effects between frequent wheeze at this age and day care, which was positively and negatively associated with frequent wheeze before and after that age, respectively.³³ Finally, our sample included a disproportionate number of children born between May and October, months of higher sun exposure. However, our goal was to assess associations with higher levels of cord 25(OH)D which were evident in these children, and we adjusted for birth season in all analyses.

Clearly, vitamin D deficiency is undesirable from multiple standpoints, particularly bone health, and thus improving the vitamin D status of those who are deficient could have substantial health benefits. Indeed, most studies suggesting a protective effect of supplementation on childhood respiratory health^{1–3, 33} have been conducted in populations at northern latitudes with limited sun exposure, conditions associated with a greater prevalence of vitamin D deficiency. However, our findings of elevated atopy outcomes in a population of children with relatively high cord levels of 25(OH)D raise concerns about “one size fits all” interventions involving vitamin D supplementation of pregnant women. If such trials are conducted in a population with high baseline cord levels of 25(OH)D, vitamin D supplementation may actually contribute to increased atopy in the child. The findings from the present study suggest that it may be important to monitor prenatal levels of 25(OH)D in order to provide supplementation in a dose-appropriate manner with the goal of preventing vitamin D deficiency, rather than universally supplementing pregnant women, regardless of baseline level.

In summary, this study in a population with high levels of annual sun exposure showed that both low and high levels of cord blood 25(OH)D were associated with increased total IgE and aeroallergen sensitization. Although statistical power was limited for clinical outcomes, vitamin D status at birth was not statistically associated with risk of allergic rhinitis or asthma at age 5 years. This nonlinear relation with immune system markers of atopy may help to bridge seemingly contradictory findings from earlier studies involving relations of vitamin D with atopy and allergic disease.

Supplementary Material

ACKNOWLEDGMENTS

Abbreviations

Footnotes

REFERENCES