Obesity and adiposity indicators, asthma, and atopy in Puerto Rican children

Abstract

Background:

Whether adiposity indicators other than body mass index should be used in studies of childhood asthma is largely unknown. The role of atopy in “obese asthma” is also unclear.

Objective:

To examine the relation among adiposity indicators, asthma, and atopy in Puerto Rican children, and to assess whether atopy mediates the obesity-asthma association.

Methods:

In a study of Puerto Rican children with (n=351) and without (n=327) asthma, we measured body mass index (BMI), percent body fat (PBF), waist circumference (WC), and waist-to-hip ratio (WHR). The outcomes studied included asthma, lung function, measures of atopy, and, among cases, indicators of asthma severity or control. We performed mediation analysis to assess the contribution of atopy to the relationship between adiposity and asthma.

Results:

BMI, PBF, and WC were associated with increased odds of asthma. Among cases, all three measures were generally associated with lung function, asthma severity/control, and atopy; however, there were differences depending on the adiposity indicator analyzed. Atopy considerably mediated the adiposity-asthma association in this population: allergic rhinitis accounted for 22-53% of the association with asthma, and sensitization to cockroach mediated 13-20% of the association with FVC and 29-42% of the association with emergency room visits for asthma.

Conclusions:

Adiposity indicators are associated with asthma, asthma severity/control, and atopy in Puerto Rican children. Atopy significantly mediates the effect of adiposity on asthma outcomes. Longitudinal studies are needed to further investigate the causal role, if any, of adiposity distribution and atopy on “obese asthma” in childhood.

Clinical Implications:

Assessment of adiposity rather than sole reliance on BMI may be important in studies of childhood asthma. Atopy is an important mediator of the relation between obesity and asthma in Puerto Rican children.

Capsule Summary:

In a cohort of Puerto Rican children, measures of adiposity were associated with asthma, asthma severity/control, and atopy; however, some differences existed depending on the adiposity indicator utilized. Atopy significantly mediated the association between adiposity indicators and asthma.

INTRODUCTION

Childhood asthma and obesity are both major public health concerns worldwide, and the prevalence of both diseases has risen markedly in the last several decades^(1-3). There is ample and growing evidence of an association between obesity and asthma, both in children and adults^(4-8). Compared to children of normal weight, those who are overweight or obese have a greater risk of incident asthma, more severe or frequent symptoms, and a decreased response to inhaled corticosteroids⁽⁹⁾. While there is growing evidence for an “obese asthmatic” phenotype^{(10, 11)}, little is known about its specific characteristics.

Body mass index (BMI) has been extensively used as a proxy for overweight or obesity in epidemiologic studies of asthma. Whether other adiposity measures (e.g. percent body fat [PBF] or waist-to-hip ratio [WHR]) provide phenotypic information that differs from or adds to that obtained by measuring BMI for studies of asthma is largely unknown. This is important, as BMI alone may not adequately characterize the relation between overweight or obesity and complex diseases such as asthma. For example, adults with “normal weight central obesity” (normal BMI but high WHR may have the highest risk for coronary artery disease⁽¹²⁾.

Several plausible mechanisms have been proposed to explain the observed association between obesity and asthma, including enhanced systemic inflammation⁽¹³⁾. Given conflicting findings from studies of overweight or obesity (largely assessed by BMI) and atopy or atopic diseases (e.g. allergic rhinitis)^(14-17), the role of atopy or allergic airway inflammation in the “obese asthmatic” phenotype is currently unclear.

Puerto Ricans share a disproportionate burden of asthma and overweight/obesity^(18-20). Very few studies have examined overweight or obesity and childhood asthma in Puerto Ricans^{(7, 21)}, and none has assessed adiposity indicators other than BMI in relation to asthma severity or control, lung function, or markers of allergic sensitization (e.g. allergy skin testing).

In this report, we examine the relation between indicators of adiposity/obesity, allergy markers, and measures of asthma severity or control (e.g. lung function) in Puerto Rican children with asthma living in San Juan, PR. We hypothesized that indicators of adiposity other than BMI would help characterize the “obese asthmatic” phenotype in Puerto Rican children, in whom an association between overweight or obesity and asthma severity or control could be mediated by atopy.

METHODS

Subject recruitment

A detailed description of study methods is provided in the Online Repository. From March of 2009 to June of 2010, children in San Juan (SJ) were chosen from randomly selected households. In brief, households in the Standard Metropolitan Area of SJ were selected using a multistage probability sample design⁽²²⁾. Primary sampling units (PSUs) were randomly selected neighborhood clusters based on the 2000 U.S. Census, and secondary sampling units were randomly selected households within PSU. A household was eligible if ≥1 resident was a child 6- 14 years old. A total of 6,401 households selected for inclusion were contacted. Of these, 1,111 households had ≥1 child who met inclusion criteria other than age (four Puerto Rican grandparents and residence in the same household for ≥1 year). Of these 1,111 households, 438 (39.4%) had ≥1 eligible child with asthma (a case, defined as having physician-diagnosed asthma and wheeze in the prior year). From these 438 households, one child with asthma was selected (at random if there was more than one such child). Similarly, only one child without asthma (a control subject, defined as having neither physician-diagnosed asthma nor wheeze in the prior year) was randomly selected from the remaining 673 households. In order to reach our target sample size (~700 children), we attempted to enroll 783 of the 1,111 eligible children selected for inclusion. Parents of 105 (13.4%) of these 783 children refused to participate or could not be reached, leaving 678 study participants (351 cases and 327 control subjects). There were no significant differences in age, gender, or area of residence between eligible children who did (n=678) and did not (n=105) agree to participate.

Study procedures

A detailed description of the study procedures is provided in the Online Repository. Study participants completed a protocol that included questionnaires on respiratory health and household characteristics, spirometry, allergy skin testing (AST), and collection of blood and house dust samples. Dust samples were obtained from three areas in the home: one in which the child slept (usually his/her bedroom), living room/television room, and kitchen. The dust was sifted through a 50-mesh metal sieve, and the fine dust was reweighed, extracted, and aliquoted for analysis of allergens from dust mite (Der p 1), cockroach (Blatella germanica [Bla g 2]), and mouse (mouse urinary protein [Mus m 1]) using monoclonal-antibody Multiplex array assays using the same reagents employed in the established ELISA⁽²³⁾. Allergen levels were analyzed as continuous (after log10-transformation), with non-detectable levels assigned a constant (half the lowest detectable value).

Measures of obesity and adiposity

BMI was calculated from weight in kilograms and height in meters. Percent body fat (PBF) was calculated from tricipital (TC) and subscapular (SS) skin folds⁽²⁴⁾, which were obtained by trained study personnel using calibrated calipers; the averages of three TC and SS measurements were used for PBF calculation. All measures were transformed to z-scores to obtain standardized/comparable coefficients and odds ratios, as follows: BMI z-scores were calculated using a program based on the 2000 CDC growth charts⁽²⁵⁾; PBF z-scores were calculated using a recent study on reference equations for U.S. children and adolescents⁽²⁶⁾; and waist circumference (WC) and waist-to-hip ratio (WHR) were standardized using the distribution of our study sample.

Ethics statement

Written parental consent and written assent were obtained for participating children. The study was approved by the Institutional Review Boards (IRBs) of the University of Puerto Rico (SJ [Protocol # 0160507]), Brigham and Women’s Hospital (Boston, MA [Protocol # 2007P- 001174]), and the University of Pittsburgh (Pittsburgh, PA [Protocol # PRO10030498]).

Statistical analysis

Our outcomes of interest included asthma (defined as above), lung function measures (FEV₁, FVC and FEV₁/FVC), allergic rhinitis (defined as current naso-ocular symptoms apart from colds and at least one positive skin test to allergens), allergy markers (skin test reactivity [STR] to allergens and serum total IgE), and other indicators of asthma severity or control, as follows: 1) number of days on oral or intravenous steroids in the prior year (categorized as 0, 1-8, 9-40, and over 40); 2) missed school days due to asthma in the prior year (categorized as 0, 1-2, 3-5 or at least 6); 3) exercise-induced symptoms in the prior year (categorized as never, occasionally, frequently, or always); and 4) number of visits to the emergency department (ED) for asthma, ever.

Bivariate analyses were conducted using Fisher’s exact tests for binary variables and two-tailed t tests for pairs of binary and continuous variables. Linear or logistic regression was used to examine the relation between each adiposity/obesity indicator and the outcomes of interest, while adjusting for potential confounders. All multivariate models included age, sex, household income (< vs. ≥ $15,000/year [the median household income for Puerto Rico in 2008-2009⁽²⁷⁾]), parental (maternal or paternal) history of asthma, and percentage of African racial ancestry (determined using genome-wide genotypic data⁽²⁸⁾, see Online Repository). All analyses of FEV₁ and FVC were additionally adjusted for height and height squared, and analyses of STR were additionally adjusted for levels of indoor allergens (see Online Repository for details).

We performed mediation analysis to assess whether part or all of the association between adiposity indicators (e.g. BMI) and outcomes of interest (e.g. FEV₁) is explained by atopy via a mediated or ‘indirect effect’ (see Online Repository for details). This analysis was performed via structural equation modeling for continuous and ordinal data, and by the Karlson-Holm- Breen decomposition method⁽²⁹⁾ for binary outcomes, which adjusts for the rescaling issues that arise from cross-model comparison of nonlinear models^{(30, 31)}. Mediation analysis was performed only on measures of atopy (i.e. allergic rhinitis, STR to cockroach) that were associated with both the adiposity indicators and the asthma outcomes. Other indicators of atopy (e.g. total IgE) did not meet this criterion and were thus not included in the mediation analysis. All statistical analyses were performed using SAS statistical software, version 9.3 (SAS Institute; Cary, NC), with the exception of the mediation analysis (which was conducted using Stata 12.1 [StataCorp, College Station, TX]).

RESULTS

The characteristics of the 678 study participants are shown in Table 1. BMI was significantly associated with increased odds of asthma after adjusting for covariates. PBF and WC were also associated with asthma, but these associations only approached significance (p=0.06 and p=0.08, respectively) (Table 2). As expected, all four obesity/adiposity measurements were significantly correlated with each other (P < 0.0001), although the degree of correlation and the slope of the regression coefficient varied (Figure 1).

Table 1.

Characteristics of study participants

Demographics	Cases N=351	Controls N=327
Male gender	201 (57.3%)**	159 (48.6%)
Age (years)	10.0 (2.59)**	10.5 (2.73)
Parental asthma history	68.1%**	33.1%
Household income ≥ $15,000/yr	34.6%	37.2%
African ancestry (%)	25.2 (11.7)	24.8 (12.5)
Household dust allergen levels1:
Dust mite (μg/g)	2.42 [4.5-9.6]	1.97 [4.4-9.3]
Cockroach (U/g)	0.73 [1.3-4.3]	0.71 [1.2-2.7]
Mouse (ng/g)	2.0 [7.0-31.0]	2.0 [7.0-28.0]
Obesity/adiposity measures2
Body mass index (BMI)	0.69 (1.19)**	0.51 (1.12)
Percent body fat (PBF)	0.29 (0.86)*	0.22 (0.77)
Waist circumference (WC)	0.03 (1.04)*	−0.03 (0.95)
Waist-to-hip ratio (WHR)	0.001 (0.85)	−0.001 (1.13)
Pulmonary function tests
FEV1 (L)	1.88 (0.67)**	2.05 (0.74)
FVC (L)	2.33 (0.82)**	2.47 (0.88)
FEV1/FVC (%)	80.9% (9.0%)**	83.5% (8.9%)
Asthma severity
No. of ED/UC visits for asthma1	10 [5-20]	n/a
Severity scores, past year1:
- Prednisone courses (0-4)	2 [1-2]	n/a
- Exercise symptoms (0-3)	1 [1-3]	n/a
- Missed school days(0-3)	1 [0-2]	n/a
Atopy measures
Total serum IgE (IU/mL)1	346 [116-881]**	158 [44-600]
Allergic rhinitis3	53.6%**	19.2%
Skin-test reactivity (STR) to:
- Dust mite	55.4%**	41.6%
- Cockroach	39.4%**	26.3%
- Alternaria	25.2%	24.8%
- Mold	12.2%	14.4%
- Mouse	26.2%	21.2%
- Any STR+	69.0%**	56.0%

Numbers for continuous variables represent mean (SD), except

¹median [IQR].

²All transformed to z-scores.

³Defined as history of allergic rhinitis plus current symptoms plus STR to at least one allergen.

^*P<0.10,

^**P<0.05.

ED/UC: Emergency department / urgent care.

Table 2.

Indicators of adiposity, allergy markers and indicators of asthma severity or control

	Body mass index (BMI)	Percent body fat (PBF)	Waist circumference (WC)	Waist-to-hip ratio (WHR)

Asthma status

1.27 (1.1-1.5)**

1.24 (.99-1.6)*

1.18 (.98-1.4)*

1.12 (.93-1.35)

Lung function in cases

FEV1 (mL)1	+68.8 (34.7 – 103.0)**	+37.5 (-10.5 – 84.5)	+20.1 (-18.3 – 60.2)	+27.3 (-19.4 – 74.0)
FVC (mL)1	+98.0 (59.0 – 137.1)**	+63.4 (8.9 – 117.8)**	+61.1 (15.9 – 106.2)**	+50.1 (−4.0 – 104.3)*
FEV1/FVC (%)	−0.6 (−1.5 – 0.5)	−0.6 (−1.9 – 0.7)	−1.0 (−2.1 – 0.0)*	−0.7 (−2.1 – 0.7)

Asthma severity

Urgent care visits, ever	+3.0 (0.01 – 6.1)**	+4.6 (0.50 – 8.79)**	+3.4 (−0.01 – 6.8)*	+2.4 (−2.0 – 6.9)
>1 urgent care visit, last yr	1.23 (0.96 – 1.59)*	1.40 (0.99 – 1.97)**	0.99 (0.75 – 1.31)	1.01 (0.70 – 1.45)
Severity scores:
- Prednisone courses	+0.08 (−0.01 – 0.17)*	+0.15 (0.03 – 0.28)**	+0.07 (−0.03 – 0.14)	+0.07 (−0.05 – 0.20)
- Missed school days	+0.13 (0.02 – 0.25)**	+0.17 (0.01 – 0.33)**	+0.13 (0.005 – 0.26)**	+0.10 (−0.05 – 0.26)
- Exercise symptoms	+0.09 (−0.02 – 0.11)	+0.15 (0.03 – 0.29)**	+0.18 (0.07 – 0.29)**	+0.18 (0.04 – 0.32)**

Atopy measures in cases

Allergic rhinitis	1.19 (0.96 – 1.49)	1.33 (0.97 – 1.81)*	1.36 (1.05 – 1.77)**	1.31 (0.95 – 1.90)*
Total IgE (IU/mL)2	0.99 (0.84 – 1.18)	1.04 (0.83 – 1.31)	1.2 (1.002 – 1.45)**	1.03 (0.83 – 1.29)
STR to3:
- Dust mite	0.97 (0.77 – 1.23)	1.17 (0.85 – 1.61)	1.07 (0.83 – 1.39)	0.95 (0.70 – 1.30)
- Cockroach	1.37 (1.04 – 1.79)**	1.57 (1.11 – 2.24)**	1.49 (1.11 – 2.01)**	1.41 (0.99 –2.01)*
- Alternaria	1.46 (1.08 – 1.97)**	1.80 (1.20 – 2.70)**	1.76 (1.26 – 2.45)**	1.44 (1.01 – 2.06)**
- Mold	2.54 (1.54 – 4.19)**	2.08 (1.23 – 3.52)**	1.80 (1.20 – 2.71)**	1.17 (0.73 – 1.88)
- Mouse	1.36 (1.02 – 1.83)**	1.70 (1.15 – 2.51)**	1.38 (1.01 – 1.88)**	1.21 (0.85 – 1.74)
- Any STR+	1.30 (1.02 – 1.64)**	1.54 (1.10 – 2.17)**	1.36 (1.03 – 1.78)**	1.35 (0.96 – 1.91)*

Results for adjusted regression analysis in asthmatic children. All models adjusted for gender, age, parental history of asthma, household income, and percent African ancestry. Numbers represent β coefficients for continuous/ordinal outcomes and OR for binary outcomes (with 95%CI in parentheses) per 1.0 z-score increment in each adiposity measure.

^*P<0.10,

^**P<0.05. P>0.10 dimmed in gray.

¹Analyzed as absolute values due to lack of predictive equations for Puerto Ricans; adjusted additionally for gender, age, height, and height squared.

²Analyzed as log₁₀.

³All STR models additionally adjusted for dust mite allergen level, except those for STR to cockroach or mouse, which were adjusted for levels of the respective allergens.

Figure 1.

Correlation between BMI and indicators of adiposity.

Table 2 shows the results of the multivariate analysis of each measure of obesity/adiposity and indicators of asthma severity or control in children with asthma (n=351). In this analysis, each 1.0 z-score increment in BMI was significantly associated with an ~69 ml higher FEV_1. All adiposity measures were positively associated with FVC, ranging from an ~50mL increment per each z-score in WHR to an ~98mL increment per z-score increment in BMI, with intermediate results obtained for PBF and WC. Of the four adiposity measures, only WC was significantly associated with a decrement in FEV₁/FVC. All adiposity measures except WHR were associated with increased lifetime ED / urgent care (UC) visits for asthma, ranging from ~3 additional ED/UC visits per each z-score increment in BMI to ~4.6 additional ED/UC visits per z-score increase in PBF. Similar results were obtained for the analysis of school absences due to asthma. BMI and PBF (but not WC or WHR) were also associated with an increased number of courses of systemic steroids for asthma in the previous year. In addition, PBF, WC, and WHR (but not BMI) were associated with increased exercise-induced asthma symptoms.

In the multivariate analysis of allergic rhinitis and allergy markers (Table 2), PBF and WC were each associated with increased odds of allergic rhinitis, and WC was associated with increased total IgE. In this analysis, all indicators of obesity/adiposity were associated with increased odds of STR to cockroach and Alternaria; and all measures except WHR were also associated with increased odds of STR to mold and mice. There was no significant association between any of the adiposity measures and STR to dust mite.

In order to assess whether the observed association between obesity/adiposity and asthma-related outcomes is mediated through atopy, we performed a mediation analysis (see Methods and Figure S1 in the Online Repository). On the basis of their high prevalence in our study population and their association both with indicators of adiposity and with asthma outcomes, we examined allergic rhinitis as a mediator for asthma and STR to cockroach as a mediator for FVC and ED/UC visits. Allergic rhinitis significantly mediated the associations between each of three indicators of adiposity or obesity (BMI, PBF and WC) and asthma (Table 3); the estimated mediation effect explained 22%, 53%, and 43%, respectively, of each association. Among children with asthma, STR to cockroach mediated ~20% and ~13% of the estimated effects of PBF and WC on FVC, respectively. STR to cockroach also mediated ~28-42% of the association between indicators of adiposity/obesity (BMI, PBF, and WC) and the number of ED/UC visits for asthma. While mediation does not imply causation, these results demonstrate a significant contribution of atopy to the obesity-asthma relationship.

Table 3.

Estimated mediation of the association between obesity/adiposity indicators and asthma outcomes by atopy

	Body mass index (BMI)	Percent body fat (PBF)	Waist circumference (WC)

Asthma status by allergic rhinitis

Total effect	1.32 (1.11–1.58) .002	1.26 (0.99–1.61) .06	1.20 (0.99–1.46) .067
Direct effect	1.24 (1.04–1.48) .015	1.12 (0.87–1.42) .38	1.11 (0.92–1.35) .29
Indirect effect	1.06 (1.01–1.13) .032	1.13 (1.04–1.23) .004	1.08 (1.01–1.16) .022
Percent mediated	22.3%	52.9%	42.6%
Total/direct effect	1.29	2.12	1.74

FVC by STR+ to cockroach in cases

Total effect (mL)	+97 (56 – 138) <.001	+60 (3 – 117) .04	+75 (26 – 125) .003
Direct effect	+90 (49 – 131) <.001	+48 (−8 – 105) .09	+65 (17 – 114) .008
Indirect effect	+7 (−2 – 16) .11	+12 (−1 – 26) .08	+10 (−1 – 22) .08
Percent mediated	7.5%	20.1%	13.3%
Total/direct effect	1.08	1.25	1.15

Number of ED/UC visits by STR+ to cockroach

Total effect	+2.8 (−.3 – 6) .08	+4.7 (0.5 – 9) .03	+2.7 (−0.9 – 6) .14
Direct effect	+2.0 (−1 – 5) .21	+3.4 (−1 – 8) .13	+1.6 (−2 – 5) .39
Indirect effect	+.8 (−0.08 – 2) .07	+1.4 (0.01 – 3) .048	+1.1 (0.01 – 2) .048
Percent mediated	28.7%	28.9%	41.9%
Total/direct effect	1.40	1.41	1.72

Results from decomposition (binary outcomes) or structural equation modeling (continuous outcomes). Table shows odds ratios (binary outcomes) or beta coefficients (continuous outcomes), 95% confidence intervals, and p-values. Indirect effect: Mediation of allergic rhinitis in the relationship between adiposity measure and outcome. Percent mediated: Percentage of the total effect explained by the mediation of atopy. ED/UC: Emergency department / urgent care.

DISCUSSION

In this study we show that BMI, PBF, and WC are each significantly associated with asthma and indicators of asthma severity or control in Puerto Rican children. We also report that atopy may be an important mediator of the relation between adiposity or obesity and asthma morbidity in these children.

Although BMI has been the most widely used proxy of adiposity, its usefulness and predictive value have been questioned in studies of cardiovascular disease and diabetes^{(32, 33)}. Our results for PBF or WC are consistent with those of several studies showing an association between BMI and asthma^(4-6) or worsened asthma control⁽⁷⁾ in children. In contrast to our findings for asthma, however, limiting our assessment to BMI would have led to non-detection of significant associations between obesity/adiposity (measured by PBF or WC) and exercise-induced asthma symptoms, allergic rhinitis, and total IgE. Our results thus suggest that sole reliance on BMI may partly explain inconsistent findings for asthma severity or atopy in prior studies. Of interest, recent studies have shown that WC predicts total body fat (measured by dual-energy X-ray absorptiometry) more accurately than BMI, which may require different height correction factors depending on age, gender, and race⁽³⁴⁾.

To our knowledge, the only previous study that examined PBF and asthma control⁽³⁵⁾ found an association in adolescent girls but not boys. WHR, which as been touted as an important marker of cardiovascular disease^{(12, 33)}, demonstrated the fewest associations with asthma outcomes in Puerto Rican children.

Adipose tissue may be related to inflammation and immune responses through production of cytokines/adipokines and macrophage activation⁽¹³⁾. However, distinct types of adiposity may differentially affect various diseases in children and adults. For example, visceral fat has been associated with cardiovascular disease in adults but not with insulin sensitivity in children⁽³⁶⁾. Similarly, large subcutaneous adipocytes may be more important for glucose/insulin regulation than visceral fat in obese women⁽³⁷⁾. Alternatively, one cannot rule out that measures of adiposity other than BMI may better reflect poor fitness rather than just asthma. Our results highlight the importance of determining which indicator(s) of obesity/adiposity should be used in future studies of obesity and asthma^{(38, 39)}, particularly as we characterize obese asthmatic phenotypes (e.g. non-atopic vs. atopic, etc.).

Obese asthmatic phenotypes may differ between children and adults. Whereas obese adults with asthma have been shown to have a restrictive ventilatory deficit (low FVC and FEV₁ but normal FEV₁/FVC )⁽⁴⁰⁾, obese children with asthma tend to have either an obstructive ventilatory deficit or dysanaptic lung growth, representing a mismatch between growth of the airway and that of the lung parenchyma (normal or high FVC and FEV₁ but decreased FEV₁/FVC)^(41-44). In general but not complete agreement with these findings, we show that BMI is associated with higher FEV₁, that all adiposity measures are associated with higher FVC, and that WC is associated (albeit non-significantly) with a lower FEV₁/FVC in Puerto Rican children.

As with asthma and obesity, there are clinical and experimental data linking obesity and atopy: adipose tissue contains high concentrations of aromatase and can increase circulating levels of estrogen in obese women, and estrogen has been shown to enhance eosinophil function and modulate IL-4 and IL-13 production by monocytes^{(45, 46)}. Obesity in mouse models of asthma has been shown to lower the threshold for allergic sensitization, measured by IgE, interleukin (IL)-5, and eosinophilia⁽⁴⁷⁾. Cluster analysis in adults has shown that certain “obese asthmatic” phenotypes have increased IgE⁽¹⁰⁾.

We have previously shown that atopy is common in Puerto Rican children with asthma^{(22, 48)}. In this study, we report that BMI, PBF, and WC are each consistently associated with increased odds of STR to cockroach, STR to mold, and STR to mouse in Puerto Rican children with asthma. In addition, we show that WC is associated with higher total IgE and allergic rhinitis in these children. Previous studies have reported conflicting results for BMI and atopy, with some studies showing a positive association with allergic sensitization in all children (independently of asthma status)^{(14, 49)} or in girls only ^{(15, 50)}, and others yielding negative results⁽¹⁷⁾. Discrepancies among reports, including ours, may be explained by differences in genetic and environmental/lifestyle factors across study populations.

Whether atopy mediates or modifies the association between obesity and asthma is unclear. To our knowledge, this is the first report utilizing mediation analysis to address this question. Findings from our analysis suggest that a significant proportion of the association between adiposity indicators and asthma-related outcomes in Puerto Rican children is mediated by atopy. Up to 22% of the increased asthma risk associated with BMI was explained by allergic rhinitis, with consistent results for WC and PBF (up to 42% and 53%, respectively). Among cases, atopy also mediated a significant proportion of the associations between obesity indicators, FVC, and ED/UC visits for asthma. Of interest, when structuring the models in the opposite direction (to answer the question: “is the association between atopy and asthma mediated by obesity?”), most indirect effects were non-significant (data not shown). Our results are in contrast with those of a prior study showing that, while BMI was associated with total IgE, BMI was only associated with asthma in non-atopic children⁽¹⁶⁾. Our results also contrast with those of a previous study that concluded that PBF provided no additional information to that given by BMI with regard to asthma severity⁽³⁵⁾. This suggests that adiposity or obesity may be more likely to influence asthma through atopy in Puerto Rican children than in children in other ethnic groups.

There are several potential limitations to our study. As with any cross-sectional study, the temporality of the observed associations cannot be ascertained, and we cannot thus exclude “reverse causation”. However, a link between obesity and asthma has been established in several longitudinal studies of children^{(5, 8)}. Similarly, a mediation analysis allows for ‘decomposition’ of an association into a ‘direct’ effect and a ‘mediated’ effect, but cannot determine causality. Of note, we did not assess puberty staging. However, we obtained very similar findings after additional adjustment for age as a proxy for puberty onset (set at ≥12 years for boys and at ≥11 years for girls) in our multivariate models (data not shown). Future studies should include Tanner staging, as hormonal differences before and after puberty may affect the relation among gender, obesity and asthma. Finally, we may have been underpowered to detect small effects of adiposity measures on certain asthma-related outcomes. However, such effects may not be clinically relevant.

In summary, we report that measures of obesity/adiposity are associated with asthma, asthma severity/control and atopy in Puerto Rican children. While our results were generally consistent, there were several differences according to the adiposity indicator analyzed. In this group of children, atopy was a significant mediator of the effect of adiposity on asthma and asthma- related outcomes. Future studies should aim to elucidate the roles of adiposity distribution and atopic sensitization on “obese asthma” in childhood.

Abbreviations

AST

Allergy skin testing

BMI

Body mass index

ED/UC

Emergency department or urgent care visit

FVC

Forced vital capacity

FEV1

Forced expiratory volume in one second

IgE

Immune globulin E

IL

Interleukin

PBF

Percent of body fat

PSU

Primary sampling unit

SD

Standard deviation

SS

Subscapular

STR

Skin test reactivity (positive AST)

TC

Tricipital

WC

Waist circumference

WHR

Waist-to-hip ratio