Obesity is associated with sustained symptomatology and unique inflammatory features in children with asthma

Abstract

Background.

Obesity complicates the clinical manifestations of asthma in children. However, few studies have examined longitudinal outcomes or markers of systemic inflammation in obese asthmatic children.

Objective.

We hypothesized that obese children with asthma would have: 1) poorer clinical outcomes over 12 months, 2) decreased responsiveness to systemic corticosteroid administration, 3) greater markers of systemic inflammation and 4) unique amino acid metabolites associated with oxidative stress.

Methods.

Children 6-17 years (lean, N=257; overweight, N=99; obese, N=138) completed a baseline visit and follow-up visit at 12 months. Outcome measures included asthma control, quality of life, lung function, and exacerbations. A subset received intramuscular triamcinolone and were re-evaluated at 7(+7) days. Leptin, adiponectin, C-reactive protein, total cholesterol, IL-1β, IL-6, IL-17, IFNγ, TNFα, MCP-1, and amino acid metabolites were also quantified in plasma as potential biomarkers of outcomes in obese children.

Results.

Obesity was associated with more symptoms, poorer quality life, and more exacerbations that persisted over one year despite greater medication requirements. Obese children also had minimal clinical improvement in asthma control and lung function after intramuscular triamcinolone. Leptin, C-reactive protein, and amino acid metabolites associated with glutathione synthesis and oxidative stress differed in obese children. Within the obese group, lower concentrations of arginine-related metabolites also distinguished uncontrolled from controlled asthma at 12 months.

Conclusion.

Obesity is associated with poorer asthma outcomes and unique systemic inflammatory features that may not be adequately modified with conventional asthma therapies. Novel approaches may be needed given increased symptoms and unique inflammation and oxidative stress in obese children with asthma.

Background

In the United States, the prevalence of obesity is currently 18.4% among 6-11 year old children and 20.6% among adolescents age 12-19 years.¹ In children with current asthma, the prevalence of obesity may be substantially higher.^{2, 3} Obesity is not only a potential risk factor for asthma onset, but is also a risk factor for greater asthma morbidity in affected children. For example, several epidemiologic studies have shown that obesity often precedes incident asthma^4-6 and that this risk may begin quite early, during prenatal or early postnatal life.^{7, 8} There is also mounting evidence that obesity may modify the clinical manifestations of asthma in children with an established asthma diagnosis, since obese children tend to have poorer symptom control and greater asthma severity.^9-11

Despite increased awareness of the impact of obesity on childhood asthma, studies in children remain quite limited. In existing literature, there is wide variability in the populations of children studied and the outcomes reported. The majority of pediatric studies are also cross-sectional in nature and do not address the temporal stability of the obesity/asthma associations. Furthermore, in contrast to adult populations,^{10, 12, 13} there have been few attempts to characterize systemic inflammation in obese asthmatic children.

Given the public health impact of obesity in children, we questioned whether obesity in children with asthma is associated with poorer outcomes over 12 months of follow-up and after a supervised intramuscular corticosteroid intervention. We also questioned whether children with obesity and asthma have differential expression of biomarkers (i.e., inflammatory cytokines and amino acid metabolites) compared to overweight and lean children with asthma. We hypothesized that obese children with asthma would have: 1) poorer asthma control and quality of life, more airflow limitation, and more exacerbations over 12 months, 2) decreased clinical responsiveness to the systemic corticosteroid intervention, 3) greater markers of systemic inflammation, and 4) unique amino acid metabolites associated with oxidative stress.

Methods

Children 6 through 17 years of age with physician-diagnosed asthma at an academic specialty center for asthma in Atlanta, Georgia, were eligible for the study if they: 1) were receiving asthma controller therapy, 2) had asthma that was not well controlled per asthma treatment guidelines, defined as daytime symptoms more than twice weekly, nighttime symptoms twice or more per month, or two or more exacerbations in the previous year,¹⁴ and 3) had current or historical evidence of ≥ 12% reversibility in their forced expiratory volume in one second (FEV₁) relative to baseline after bronchodilator administration, or alternatively, airway hyperresponsiveness to methacholine, with a provocative concentration of methacholine causing a 20% drop in FEV₁ (PC₂₀) of ≤16 mg/mL. Exclusion criteria included premature birth before 35 weeks of gestation and other chronic airway disorders that could mimic asthma such as pulmonary aspiration or vocal cord dysfunction. Permission to proceed with this study was granted by the Emory University and Children’s Healthcare of Atlanta Institutional Review Boards. Informed written consent was obtained from legal guardians. Verbal assent was obtained from children 6-10 years and written assent was obtained from children and adolescents 11 to 17 years.

Study design and outcome measures.

Participants completed a baseline outpatient research visit and a follow-up visit at 12 months. These visits were postponed if an asthma exacerbation treated with systemic corticosteroids was reported within the preceding two weeks. After the baseline visit, a subset of children also underwent systemic corticosteroid responsiveness testing with intramuscular triamcinolone acetonide (1 mg/kg, 60 mg maximum dose) administered in the gluteal muscle. Corticosteroid responsiveness was assessed at 7 days (window +7 days) after triamcinolone receipt. Participants who received triamcinolone recorded daily symptoms and inhalations of rescue bronchodilators used on diary cards between visits.

Participants were classified as either lean, overweight or obese according to age- and sex-specific body mass index (BMI) percentiles with the BMI-for-age percentile growth charts from the Centers for Disease Control and Prevention. Overweight was defined as a BMI at the 85^th to less than the 95^th percentile, while obesity was defined as a BMI at or above the 95^th percentile. Outcome measures included asthma control, asthma-related quality of life, lung function, and asthma exacerbations treated with systemic corticosteroids.¹⁵ Asthma exacerbations were verified by a review of medical records.

Clinical characterization procedures.

Participants and their caregivers completed questionnaires pertaining to symptoms, medical history and demographics. Asthma control was assessed with the 6-item Asthma Control Questionnaire (ACQ)¹⁶ administered by a trained interviewer.¹⁷ Asthma-related quality of life over the preceding two weeks was assessed with the Pediatric Asthma Quality of Life Questionnaire (PAQLQ)¹⁸ with technical assistance recommended for younger children.¹⁹ Aeroallergen sensitization was assessed by specific IgE testing (Children’s Healthcare of Atlanta, Atlanta, GA) or skin prick testing with eight extracts: tree mix, grass mix, weed mix, mold mix (Alternaria alternata, Aspergillus fumagatis, Cladosporium herbarum), dog dander, cat dander, Blatella germanica, and dust mite mix (Dermatophagoides farinae and Dermatophagoides pteronyssinus) (Greer® Laboratories, Lenoir, North Carolina). Venipuncture was also performed for plasma collection and quantification of blood eosinophils (Children’s Healthcare of Atlanta). Spirometry (KoKo® PDS, Ferraris, Louisville, Colorado) was performed according to technical standards²⁰ with 360 mcg albuterol sulfate for bronchodilator reversibility testing. The best of three forced vital capacity (FVC) maneuvers was interpreted according to Global Lung Function Initiative prediction equations.²¹ A subset of participants without evidence of FEV₁ bronchodilator reversibiliy also underwent bronchoprovocation testing with methacholine concentrations of 0 to 16 mg/mL (Provocholine; Methapharm Inc, Coral Springs, Florida) delivered by a Rosenthal dosimeter (Pulmonary Data Service Instrumentation, Louisville, Colorado). Bronchoprovocation was limited to participants with baseline FEV₁ >70% predicted.

Laboratory methods for biomarker quantification.

At the baseline visit, whole blood was collected into EDTA tubes and immediately centrifuged for isolation of plasma. Plasma samples were stored at −80°C and analyzed in a single batch. Leptin (EZHL-80SK, Millipore, Burlington), adiponectin (EZHADP-61K, Millipore), C-reactive protein (ab9995, Abcam, Cambridge, Massachusetts) and total cholesterol (STA-384, Cell BioLabs, San Diego, California) were quantified in plasma with commercially available kits according to manufacturer instructions. Interleukin (IL)-1β, IL-6, IL-17, interferon gamma (IFNγ), tumor necrosis factor alpha (TNFα), and monocyte-chemoattractant protein-1 (MCP-1) (HSTCMAG-28SK and HCYTOMAG-60K, Millipore) were quantified in plasma with a Luminex MAGPIX system (Millipore) according to technical standards.

Amino acid metabolites (52 in total) were measured in plasma by solid phase extraction followed by derivatization and liquid/liquid extraction (EZ:faast Kit, Phenomenex, Torrance, California) according to the manufacturer’s instructions. Samples were mixed with internal standards (homoarginine, methionine-d3 and homophenylalanine), extracted, and derivatized with propyl chloroformate. The organic phase was evaporated at room temperature under a stream of nitrogen and re-dissolved in mobile phase. Samples were analyzed using a Thermo Vanquish ultra-high performance liquid chromatograph coupled to a Thermo TSQ Quantis triple quadrupole mass spectrometer (Thermo Scientific, Waltham, Massachusetts). Using an autosampler at 4°C, a volume of 1 μL was injected onto a 250 x 2.0 mm x 4 μ AAA-MS column (Phenomenex) at a flow rate of 0.25 mL/min. The column was held at 35°C. Mobile phase A was 10 mM ammonium formate in water, and mobile phase B was 10 mM ammonium formate in methanol. Samples were separated using an 18-minute gradient, from 68 to 83% of mobile phase B, with a 7-minute re-equilibration between samples. The ion transfer tube and vaporizer were maintained at 275°C and 225°C respectively. Positive electrospray ionization mode at 5000 V was used to monitor selected reaction transitions as outlined in the EZ:faast manual. Transitions were optimized for the mass spectrometer using derivatized standards. Quantitation of amino acids was performed using TraceFinder software (Thermo Scientific).

Statistical analyses.

Data were analyzed with SPSS® Statistics (Version 26, IBM, Armonk, NY). Group differences at baseline and 12 months were assessed with Chi-square tests or analysis of variance with Tukey’s Least Significant Difference post-hoc tests. Within groups paired t-tests were used to compare clinical features before and after triamcinolone administration. A p value <0.05 was used as the threshold for statistical significance without adjustment for multiple comparisons. Metabolites were first visualized with linear discriminant analysis, which was performed using the Fisher method after z-score normalization as described previously.²² This procedure yields a set of discriminant functions based on the linear combination of the variables that provide the best discrimination between groups. Discriminant functions were plotted to visualize the proximity of each participant to others in the same group and were also used to generate predictive models of obesity. Individual metabolites that differed significantly between groups after Bonferroni correction were further analyzed with generalized linear models and post-hoc tests utilizing a two-tailed α of 0.05. Metabolite pathways were visualized with the Kyoto Encyclopedia of Genes and Genomes database pathway maps.

Results

Four hundred ninety-four children with asthma (lean, N=257; overweight, N=99; obese, N=138) were enrolled and completed the 12-month visit. Features of the enrolled participants are shown in Table 1. Age, asthma duration, sex, ethnicity, race, family history of asthma, and household educational attainment and tobacco smoke exposure were not different between BMI groups. However, gastroesophageal reflux and eczema were more prevalent in obese children. Obese children were also more likely to report worsening of asthma symptoms with daily activity and respiratory infections despite significantly greater asthma controller medication use. Other features of allergic sensitization, including the percentage of positive aeroallergens, blood eosinophil counts, and total serum IgE concentrations, did not differ between the BMI groups (Table 1).

Table 1.

Features of the enrolled participants. Data represent the mean ± standard deviation, the median (25^th, 75^th percentile) or the number of participants (%) unless otherwise stated.

	Lean N=257	Overweight N=99	Obese N=138
BMI percentile	50.8 ± 24.5	90.6 ± 2.6	97.9 ± 1.5
Age (years)	11.9 ± 3.7	11.9 ± 3.5	11.6 ± 3.0
Asthma duration (years)	8.8 ± 4.5	8.9 ± 3.7	9.1 ± 3.8
Males	143 (55.6)	57 (57.6)	82 (59.4)
Hispanic ethnicity	19 (7.4)	5 (5.1)	10 (7.2)
Race
White	95 (37.0)	32 (32.3)	34 (24.6)
Black	139 (54.1)	58 (58.6)	87 (63.0)
Other	23 (9.0)	9 (9.1)	17 (12.3)
Parent with asthma	135 (55.3)	48 (50.0)	88 (64.7)
College degree in household	96 (37.4)	40 (40.4)	41 (29.7)
Tobacco smoke exposure	35 (14.2)	15 (15.5)	26 (19.0)
Comorbid conditions (any in past year)
Sinusitis	68 (26.5)	35 (35.4)	49 (35.5)
Pneumonia	120 (46.7)	53 (53.5)	73 (52.9)
Gastroesophageal reflux	56 (21.8)	21 (21.2)	44 (31.9)*
Asthma controller medications
Inhaled corticosteroid (any dose)	242 (81.5)	95 (79.8)	135 (88.2)
Inhaled corticosteroid (high dose)	139 (46.8)	68 (57.1)	95 (62.1)*
Long-acting beta agonist	158 (53.2)	74 (62.2)	103 (67.3)*
Montelukast	182 (61.3)	73 (61.3)	102 (66.7)
Omalizumab	6 (2.0)	3 (2.5)	8 (5.2)
Self-reported asthma triggers
Walking or daily activity	112 (43.6)	47 (47.5)	89 (64.5)*^
Exercise or sports	147 (57.2)	61 (61.6)	91 (65.9)
Respiratory infections	222 (86.4)	93 (93.9)	133 (96.4)*
Reduction in inhaled corticosteroid dose	103 (40.1)	44 (44.4)	65 (47.1)
Allergic sensitization features
Eczema	139 (54.1)	58 (58.6)	89 (64.5)*
Positive aeroallergens (of 8 tested)	3 (1, 7)	3 (1, 9)	5 (2, 7)
Blood eosinophil count (cells/microliter)	281 (156, 492)	245 (126, 481)	308 (145, 490)
Serum IgE (kU/L)	299 (92, 809)	275 (69, 934)	462 (146, 875)

^*p<0.05 vs. lean asthma

^{^}p<0.05 vs. overweight asthma

Other features of asthma in the participants are shown in Table 2. At baseline, obese children had significantly poorer asthma control, reflected by higher ACQ-6 scores, and poorer asthma-related quality of life, reflected by lower PAQLQ scores. Obese children also had greater exacerbation occurrence in the previous year and a marginal reduction in FEV₁/FVC, although this was due to a slightly higher FVC with no differences in FEV₁. Forced expiratory flow at 25-75% of vital capacity (FEF_25-75), FEV₁ bronchodilator reversibility, exhaled nitric oxide, and methacholine PC₂₀ concentrations were not significantly different between groups (Table 2).

Table 2.

Asthma features at baseline. Data represent the mean ± standard deviation, the median (25^th, 75^th percentile) or the number of participants (%).

	Lean N=257	Overweight N=99	Obese N=138
Asthma control
Daytime symptoms >twice weekly	95 (36.9)	35 (35.4)	70 (50.7)*^
Nocturnal symptoms ≥twice monthly	102 (39.7)	43 (43.4)	72 (52.2)*
ACQ-6 score	0.98 ± 0.91	1.21 ± 1.09	1.23 ± 1.03*
Asthma quality of life
PAQLQ total score	5.62 ± 1.25	5.27 ± 1.34*	5.17 ± 1.23*
PAQLQ symptom domain	5.54 ± 1.34	5.25 ± 1.38	5.10 ± 1.34*
PAQLQ activity domain	5.61 ± 1.30	5.26 ± 1.43*	5.31 ± 1.12*
PAQLQ emotional domain	5.77 ± 1.39	5.45 ± 1.60	5.31 ± 1.56*
Asthma exacerbation (previous year)
Any exacerbation	146 (58.6)	71 (73.2)	98 (71.0)*
≥3 exacerbations	59 (23.0)	29 (29.3)	52 (37.7)*
Emergency department visit	109 (42.4)	60 (60.6)	83 (60.1)*
Hospitalization	62 (24.2)	30 (30.3)	43 (31.2)
Lung function (% predicted value)
FVC	100 ± 12.7	103 ± 12.3*	104 ± 14.1*
FEV1	90.5 ± 14.9	90.0 ± 16.1	91.0 ± 16.5
FEV1/FVC	90.2 ± 11.1	86.3 ± 11.3*	86.9 ± 9.3*
FEF25-75	71.5 ± 26.7	67.0 ± 26.3	68.6 ± 24.6
FEV1 bronchodilator reversibility
Relative reversibility (%)1	12.8 ± 15.5	15.9 ± 16.9	13.8 ± 18.3
Absolute reversibility (%)2	10.2 ± 10.8	12.7 ± 11.0	11.6 ± 11.8
Exhaled nitric oxide (ppb)	37.1 ± 37.4	33.2 ± 32.7	32.4 ± 28.1
Methacholine PC20 (mg/mL)3	1.54 (0.43, 5.66)	1.95 (0.72, 9.22)	1.95 (0.82, 5.93)

¹Defined as: (FEV_{1 post-bronchodilator} – FEV_{1 pre-bronchodilator})/FEV_{1 pre-bronchodilator}

²Defined as: FEV_{1 post-bronchodilator %predicted} – FEV_{1 pre-bronchodilator %predicted}

³lean, n=119; overweight, n=35; obese, n=39

^*p<0.05 vs. lean asthma

^{^}p<0.05 vs. overweight asthma

Systemic corticosteroid responses.

A subset of 231 participants consented to systemic corticosteroid responsiveness testing with intramuscular triamcinolone at the completion of the baseline visit. Demographic features of the subset that received triamcinolone, including age, sex, race, ethnicity, household education and medication use, were not significantly different from the parent sample (data not shown). ACQ-6 scores statistically improved in obese children after triamcinolone administration (Table E1), particularly in the subset with “not well-controlled asthma” defined by an ACQ-6 score >0.75.²³ However, the percentage of patients achieving the minimal important difference of 0.5¹⁷ for the ACQ-6 instrument was not significantly different between groups (Figure 1A) and ACQ-6 scores after triamcinolone remained significantly higher in obese children (Figure 1B). Obese children also used more frequent short-acting bronchodilators for asthma symptoms in the seven days following the triamcinolone intervention (Figure 1C). Lung function percent predicted indices were largely unchanged (Table E1) and did not differ between groups after the triamcinolone intervention (lean vs. overweight vs. obese, FVC: 103 ± 16 vs. 105 ± 15 vs. 101 ± 13, p=0.152; FEV₁: 90 ± 18 vs. 94 ± 14 vs. 89 ± 16, p=0.216; FEV₁/FVC: 87 ± 12 vs. 89 ± 9 vs. 88 ± 10, p=0.427). Exhaled nitric oxide concentrations improved in both the obese and lean groups after triamcinolone administration (mean differences are shown in Table E1). After triamcinolone, exhaled nitric oxide concentrations did not differ between BMI groups (mean concentration in ppb, lean: 26 ± 24; overweight: 20 ± 18; obese: 24 ± 22 ppb, p=0.215).

Figure 1.

(A) Percentage of participants achieving the minimal important difference (MID) on the Asthma Control Questionnaire (ACQ) 6-item tool, (B) ACQ-6 raw scores, and (C) daily inhalations of albuterol for asthma symptoms in lean (red), overweight (yellow), and obese (blue) children after intramuscular triamcinolone injection. Boxplot whiskers depict the 5^th and 95^th percentiles.

Outcomes at 12 months.

At 12 months, 90% of lean children remained lean, while 85% of obese children remained obese (Figure E1). Obese children also remained symptomatic. Although there were no significant differences in the percentage of children reporting daytime symptoms more than twice weekly between BMI groups (lean: 27%; overweight, 28%; obese, 35%, p=0.414), obese children reported more nocturnal symptoms at least twice monthly (lean: 29%; overweight: 28%; obese: 46%; p=0.012) and had higher ACQ-6 scores (Figure 2A). By 12 months, obese children also had a higher occurrence of any exacerbation treated with systemic corticosteroids and any exacerbation necessitating an Emergency Department visit (Figure 2B). Obese children also had poorer asthma-related quality of life (Figure 2C) that was largely driven by asthma symptoms (Figure 2D, Table E2). Quality of life related to activities and emotions, lung function indices, and exhaled nitric oxide concentrations were not different between groups at 12 months (Table E2).

Figure 2.

(A) Asthma Control Questionnaire (ACQ) 6-item scores, (B) exacerbation occurrence, and (C) Pediatric Asthma Quality of Life Questionnaire (PAQLQ) total scores and (D) PAQLQ symptom domain scores in lean (red), overweight (yellow), and obese (blue) children at 12 months. Boxplot whiskers depict the 5^th and 95^th percentiles.

Biomarkers of obesity and outcomes.

Obese children, compared to lean and overweight children, had significantly higher plasma concentrations of C-reactive protein and leptin (Figure 3A, 3B). Adiponectin was also somewhat higher in obese children (Figure 3C), but the leptin-to-adiponectin ratio remained substantially elevated in this group (Figure 3D). There were no differences in total cholesterol (lean: 91.5 ± 47.2; overweight: 87.7 ± 43.5; obese: 88.8 ± 43.2 mg/L, p=0.870) and the inflammatory markers IL-1β, IL-6, TNFα, and MCP-1 between BMI groups (Table E3). However, obese children had significantly lower concentrations of IFNγ (mean ± standard error of the mean, lean: 6.0 ± 0.3; overweight: 5.2 ± 0.3; obese: 5.1 ± 0.3 pg/mL, p=0.035) and IL-17 (lean: 5.3 ± 0.3; overweight: 4.9 ± 0.4; obese: 4.4 ± 0.3 pg/mL, p=0.049).

Figure 3.

(A) C-reactive protein, (B) leptin, and (C) adiponectin concentrations and (D) the ratio of leptin to adiponectin in the plasma of lean (red), overweight (yellow), and obese (blue) children. Boxplot whiskers depict the 5^th and 95^th percentiles.

Linear discriminant analysis of the combination of all 52 plasma metabolites demonstrated significant differences between lean, overweight and obese children (Figure 4A). The resulting model yielded the greatest separation between obese versus lean children (discriminant function 1 Eigenvalue=0.524, 64.9% of variance explained, Wilks’ lambda=0.511, p<0.001), whereas overlap was noted between overweight children and the other groups (discriminant function 2 Eigenvalue=0.283, 35.1% of variance explained, Wilks’ lambda=0.779, p=0.061). Individual metabolites that were significantly different in obese children are shown in Table 3. Obese children had higher concentrations of 4-aminobenzoic acid, an intermediate in the synthesis of folate, and higher concentrations of the branched chain amino acid, isoleucine. Obese children also had higher concentrations of the tyrosine precursor, phenylalanine, and the tyrosine derivative, 3-hydroxytyrosine, and lower concentrations of the collagen-derived oligopeptide, prolyl-hydroxyproline. Metabolites associated with glutathione synthesis and metabolism, including glutamine, glycine and pyroglutamic acid, were significantly lower in obese children. Cystine, the oxidized form of the amino acid cysteine, was significantly higher in obese children compared to the other groups. The glutathione pathway is illustrated in Figure 4B.

Figure 4.

(A) Plot of canonical discriminant functions generated from the linear combination of all amino acid metabolites that best classifies lean (red), overweight (yellow) and obese (blue) children with asthma, and (B) metabolites associated with glutathione and arginine synthesis that are altered in children with obese asthma.

Table 3.

Concentrations of plasma metabolites in lean, overweight and obese children. Data are shown as the mean ± standard error of the mean.

Metabolite	Lean N=163	Overweight N=61	Obese N=75
Branched chain amino acids
Isoleucine	29.2 ± 2.13	40.2 ± 3.62*	43.0 ± 2.87*
Tyrosine synthesis and metabolism
Phenylalanine	45.8 ± 1.92	51.4 ± 2.35	55.9 ± 3.24*
3-hydroxytyrosine (L-DOPA)	2.84 ± 0.32	3.65 ± 0.83	5.62 ± 0.85*^
Benzoic acids
4-aminobenzoic acid	36.9 ± 0.93	39.4 ± 1.26*	44.3 ± 1.36*^
Collagen-derived oligopeptides
Prolyl-hydroxyproline	1.97 ± 0.12	1.76 ± 0.15	1.33 ± 0.10*
Glutathione synthesis/ metabolism
Cystine	2.99 ± 0.34	3.80 ± 0.84	5.80 ± 0.90*^
Glutamine	1123 ± 79.5	925.2 ± 101.2	788.4 ± 42.0*
Glycine	3683 ± 273.6	2623 ± 340.7	2419 ± 271.6*
Pyroglutamic acid (oxoproline)	2406 ± 239.4	1576 ± 266.7	1404 ± 164.2*

^*Bonferroni-corrected p<0.05 vs. lean

^{^}p<0.05 vs. overweight

Exploratory analyses were further performed within the obese group only to determine associations between biomarkers and uncontrolled asthma at 12 months, defined as an ACQ-6 score >1.5, daytime symptoms more than twice weekly, or nighttime symptoms twice or more per month.¹⁴ Within the obese group, uncontrolled asthma at 12 months was associated with lower adiponectin, lower glycine, higher phenylalanine, and higher 4-aminobenzoic acid concentrations (Figure 5A-D). Obese children with uncontrolled asthma also had significantly lower concentrations of glutamate, ornithine, and arginine (Figure 5E-G). The arginine pathway is also shown in relation to the glutathione pathway in Figure 4B.

Figure 5.

(A) Adiponectin, (B) ratio of adiponectin to leptin, (C) phenylalanine, (D) 4-aminobenzoic acid, (E) glycine, (F) arginine, (G) ornithine, and (H) glutamate concentrations in plasma of obese children with controlled (red) and uncontrolled (yellow) asthma at 12 months. Boxplot whiskers depict the 5^th and 95^th percentiles.

Discussion

This study addressed associations between obesity and longitudinal clinical outcomes in school-age children and adolescents with confirmed asthma. In this cohort, obesity was associated with greater symptom burden, poorer asthma control, more exacerbations, and poorer quality life that persisted over one year of follow-up, despite more intensive asthma controller medications. Although asthma control scores statistically improved in obese children after intramuscular triamcinolone administration, clinical features of poor asthma control such as symptoms and short-acting bronchodilator use persisted in this group. Alternations in leptin, C-reactive protein, and amino acid metabolites were also observed in obese children with asthma and may contribute to the poorer asthma outcomes observed in these patients.

The results of our study are similar to others that have shown greater symptom burden and reduced quality of life in obese versus lean children with asthma.^{24, 25} Our observation of greater exacerbation occurrence in obese children with asthma is also similar to other reports,^{26, 27} but this observation has not been consistently replicated.^{28, 29} For example, a recent meta-analysis observed a significant association between obesity and asthma exacerbations, but this effect was rather small (odds ratio of 1.17) and was generated from a very large (i.e., N>45,000) pooled sample of both cross-sectional and prospective studies.³⁰ In contrast, a separate analysis of five large asthma clinical trials with prospective evaluation of respiratory tract infections, unscheduled visits for asthma, and asthma exacerbations treated with systemic corticosteroids found no associations between BMI classification and the rate of visits from upper respiratory tract infections, all-cause asthma events (including exacerbations), or respiratory tract-associated asthma events.³¹

In the present study, we also found no consistent differences in lung function measures between obese and lean children with asthma. Although studies in adults have shown clear associations between increases in BMI and reductions in FVC and FEV₁,^{32, 33} a recent meta-analysis also revealed that unlike adults, there were no differences in FEV₁ or FVC percent predicted values in obese versus lean asthmatic children.³⁴ However, that same study and a recent large, multi-center study also reported a lower ratio of FEV₁ to FVC of approximately 1 to 1.5% in children with obese asthma,^{34, 35} which may be due to airway dysanapsis, whereby obese children have incongruence in the growth of the lungs (evidenced by lung volumes) versus the growth of the airways.³⁶ It is also important to note that these prior studies did not express FEV₁/FVC as a percent predicted value and utilized data obtained with a variety of different lung function interpretation equations.^{34, 35} In the present study, we expressed FEV₁/FVC as a percentage of predicted using the Global Initiative for Lung Function equations,²¹ since FEV₁/FVC is not a fixed ratio and follows a curvilinear pattern from childhood to adulthood.³⁷ The Global Initiative for Lung Function reference equations also have limited agreement with other reference equations that were previously used in children.³⁸

To our knowledge, this is the first study to characterize systemic inflammatory markers and amino acid metabolites in a longitudinal study of obese children with asthma. In animal and in vitro studies, increased adiposity promotes the secretion and release of inflammatory mediators such as leptin, IL-1β, IL-6, TNFα, and MCP-1, which inhibit adiponectin and promote a pro-inflammatory state.³⁹ Whereas studies in adults have likewise noted increased IL-1β, IL-6, TNFα and concentrations in obese patients with asthma,^40-43 we found no such associations between these biomarkers and obesity in children with asthma. However, many of these prior studies included adults with metabolic syndrome, which is associated with excess systemic inflammation and insulin resistance.⁴⁴ In children, the construct of metabolic syndrome remains controversial, as it is poorly defined and has does not discriminate clinical outcomes better than obesity alone.⁴⁵ Our observation of increased leptin and C-reactive protein concentrations in obese asthmatic children is similar to other reports^46-48 and does highlight increased systemic inflammation in these children, which could be a predictor of increased cardiometabolic risk.⁴⁹ Although we also observed lower IFNγ and IL-17 concentrations in obese children in the present study, the clinical relevance of this finding is unclear, as systemic cytokines have also not been well characterized in pediatric obese asthma populations. For example, one study found lower systemic Th1 cytokines in obese asthmatic children,⁵⁰ whereas a separate study concluded that obese versus asthma in children is associated with a Th1-skewed immune response.⁵¹ Lower IFNγ and IL-17 concentrations in our obese patients could reflect more Th2-mediated inflammation given the markedly elevated blood eosinophil counts in each of our study groups. Alternatively, higher doses of inhaled corticosteroids and more frequent systemic corticosteroid courses in the obese group may have influenced systemic cytokine and chemokine measurements in the present study. Indeed, obese children had more prior exacerbations at baseline and therefore greater cumulative corticosteroid exposure at study entry when biomarkers were analyzed.

The amino acid metabolites that differed in obese children in the present study also have biologic plausibility. While higher concentrations of 4-aminobenzoic acid, isoleucine, phenylalanine, and 3-hydroxytyrosine in obese children could be due to excessive dietary intake, other studies have shown that these markers are potentially inflammatory and may increase cardiovascular risk. For example, higher concentrations of aromatic amino acids such as phenylalanine and branched-chain amino acids such as isoleucine have been observed in adults with obesity^{52, 53} and are associated with risk of cardiovascular disease in large population studies.^{52, 54} 3-hydroxytyrosine is a precursor to catecholamine, but is also a marker of oxidative and nitrosative stress from the reactions of tyrosine with oxidizing free radicals.⁵⁵ Heightened oxidative stress in obese patients may also account for the higher 4-aminobenzoic acid levels observed in these patients, since derivatives of this metabolite have free radical-scavenging properties⁵⁶ and inhibit insulin resistance.⁵⁷ Lower concentrations of prolyl-hydroxyproline may also be a compensatory response to inhibit insulin resistance in these children, since prolyl-hydroxyproline stimulates synthesis of hyaluronic acid, which promotes insulin resistance in the setting of a high-fat diet.⁵⁸ Other features of oxidative stress observed in obese children, including lower concentrations of glutamine, glycine and pyroglutamic acid and higher concentrations of cystine, are also consistent with other reports in obese patients.^59-64 These reports have also observed lower concentrations of the antioxidant, glutathione,^{65, 66} which were restored with weight loss intervention.⁶² Likewise, low concentrations of arginine and related metabolites have also been observed in obese patients and are associated with systemic inflammation, insulin resistance, and other risk factors of cardiovascular disease.^67-69 Supplementation with L-arginine in obese patients is associated with a reduction in waist circumference⁷⁰ and improvement in cardiovascular risk markers.⁷¹ In obese asthmatics, supplementation with L-citrulline, a precursor of L-arginine, also improved asthma control.⁷²

Strengths of the present study are the large and highly characterized sample of children and adolescents with asthma confirmed by bronchodilator reversibility or airway hyperresponsiveness testing, since asthma can be easily misdiagnosed in this population.⁷³ Our inclusion criteria, which required that enrolled children have inadequately controlled asthma, also ensured a study population with sufficient disease burden at greatest risk for poorer outcomes. To our knowledge, this is also the first study to examine clinical responses to systemic corticosteroids in children with obese asthma. Our observation of poorer asthma control in obese children after triamcinolone injection is similar to other studies which have shown a poorer clinical response in these patients after receipt of inhaled corticosteroids.^74-77 For example, in a post-hoc analysis of the Childhood Asthma Management Research Program, obese children had lesser improvement in pulmonary function measures and no reduction in the risk of emergency department visits or hospitalizations after inhaled corticosteroid initiation.⁷⁷ A more recent analysis also found an earlier time-to-management failure, defined as any step-up in therapy, systemic corticosteroid prescription, acute care visit or hospitalization, in obese children treated with Step-2 therapy.⁷⁴

Nonetheless, this study does have limitations. First, the outcomes selected are multifactorial and may not be directly attributed to adiposity. We also utilized BMI percentiles as a measure of obesity and this measure does not fully characterize adiposity. However, a recent study found that, unlike in adults, BMI exhibited good to moderate correlations with adiposity measured by dual-energy x-ray absorptiometry in children.⁷⁸ Increasing BMI has also been associated with greater amounts of adipose tissue within the airway wall.⁷⁹ Furthermore, we are also unable to comment on the temporal stability of the measured biomarkers since samples were not available for analysis at the 12-month follow-up visit. It is also recognized that systemic measurement of inflammation is limited and may not reflect inflammation in the airways. It is also recognized that our selected outcomes are multifactorial and could be influenced by unmeasured sources of confounding including adherence to asthma medications, which was not adequately assessed. We also acknowledge that obese patients may have higher medication requirements due to differing symptom perceptions irrespective of lung function.⁸⁰ Indeed, we have previously shown that obese children with asthma have decreased functional residual capacity, which may result in overestimation of asthma symptoms due to enhanced perception of dyspnea resulting from altered mechanical properties of the chest wall.⁸¹ Unfortunately, lung volumes were not measured in the present study, so physiological underpinnings of asthma symptoms cannot be adequately addressed.

In summary, we demonstrate that poor clinical outcomes in obese children with asthma are sustained over time and are not readily responsive to systemic corticosteroid treatment. Heighted systemic inflammation and metabolites of oxidative stress are present in obese children with asthma, but are not clearly related to distinct patterns of T cell priming. Although independent replication is warranted, these findings question whether obesity-related morbidity in children and adolescents can be adequately modified with conventional asthma therapies such as inhaled corticosteroids. Our findings further suggest that obesity-driven inflammation may be distinct in obese children with asthma and may warrant novel approaches for evaluation and management, including further evaluation of the role of transdiagnostic symptoms, given the significant burden of asthma experienced by these patients.

Highlights.

What is already known about this topic?

Obesity complicates the clinical manifestations of asthma in children. However, unlike studies in adults, few studies have examined longitudinal outcomes or markers of systemic inflammation in obese asthmatic children.

What does this article add to our knowledge?

Poor clinical outcomes in obese children with asthma are sustained over time. Systemic corticosteroid responsiveness is also impaired in obese children with asthma and is accompanied by unique patterns of systemic inflammation and oxidant stress.

How does this study impact current management guidelines?

Obesity-driven inflammation may be distinct in children with asthma and may not be adequately modified with conventional asthma therapies. Novel therapeutic approaches may be warranted given the significant burden of symptoms in these children.

Abbreviations

ACQ

Asthma Control Questionnaire

BMI

Body mass index

FEF_25-75

Forced expiratory flow at 25-75 percent of vital capacity

FEV₁

Forced expiratory volume in one second

FVC

Forced vital capacity

IFNγ

Interferon gamma

IL

Interleukin

MCP-1

Monocyte-chemoattractant protein-1

PAQLQ

Pediatric Asthma Quality of Life Questionnaire

PC₂₀

Provocative concentration causing a 20% drop in FEV₁

TNFα

Tumor necrosis factor alpha