Asthma Phenotypes, Risk Factors, and Measures of Severity in a National Sample of US Children
http://www.100md.com
《小儿科》
Rollins School of Public Health, Emory University, Atlanta, Georgia
Air Pollution and Respiratory Health Branch, Division of Environmental Hazards and Health Effects, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia
Division of Pulmonary and Critical Care Medicine, University of Kentucky Medical Center, Lexington, Kentucky
ABSTRACT
Objective. To examine a nationally representative sample of US children aged 6 to 16 years old and determine whether there are differences in risk factors and measures of severity between children with different asthma phenotypes.
Methods. We analyzed data from the Third National Health and Nutrition Examination Survey. We used questionnaire and skin-prick testing data to separate children into the following mutually exclusive categories: atopic asthma, nonatopic asthma, resolved asthma, frequent respiratory symptoms with no asthma diagnosis, and normal. We used multivariate regression to determine whether demographic or potential risk factors varied between phenotypes and whether measures of severity varied by phenotype.
Results. We found that 4.8% of children had atopic asthma, 1.9% had nonatopic asthma, 3.4% had resolved asthma, and 4.3% had frequent respiratory symptoms. Risk factors varied by phenotype, for example, the mean BMI was higher among children with nonatopic asthma, prenatal maternal smoking was a risk factor for resolved asthma, and child care attendance was a risk factor for frequent respiratory symptoms with no asthma diagnosis. Patients with atopic and nonatopic asthma were similar for most measures of asthma severity (medication use, health status, and lung function impairment). In contrast, patients with resolved asthma had fewer symptoms but a similar level of lung function impairment to that seen in patients with current asthma, whereas children with frequent respiratory symptoms but no asthma diagnosis had normal lung function.
Conclusions. Asthma risk factors and measures of severity vary between children with different asthma phenotypes.
Key Words: asthma atopy allergy children lung function
Abbreviations: NHANES III, Third National Health and Nutrition Examination Survey SES, socioeconomic status PIR, pov-erty/income ratio FEV1, forced expiratory volume in 1 second FVC, forced vital capacity ED, emergency department
Asthma is a chronic inflammatory disorder of the airways that causes recurrent episodes of wheezing and other respiratory symptoms in an estimated 5 million children in the United States.1,2 Asthma prevalence, morbidity, and mortality has increased in the United States since 1980 for reasons that are not clear.3,4
For decades, clinical differences in asthma presentation have been recognized and led to its description as a heterogeneous disorder.5 Subtypes originally identified were referred to as intrinsic and extrinsic asthma, but this terminology has since been abandoned and relabeled as atopic and nonatopic asthma, often delineated on the basis of positive skin tests to common allergens or the presence of antibodies in the blood.5–9 In addition to these 2 clinical phenotypes, another phenotype describing children whose asthma resolves as they age has been reported and linked in the literature to premature birth and prenatal maternal smoking.8,10 Finally, many studies have addressed the prevalence and risk factors of "undiagnosed asthma" defined by frequent respiratory symptoms or wheezing.11–14
Although many studies have described the similarities and differences between asthma phenotypes for immunologic markers and airway lability,6,15–17 fewer have described the epidemiologic and clinical characteristics of asthma across different phenotypes in a generalizable sample.8,18 We analyzed data among children who were aged 6 through 16 years from the Third National Health and Nutrition Examination Survey (NHANES III) and classified them into 5 respiratory phenotypes: current physician-diagnosed atopic asthma, current physician-diagnosed nonatopic asthma, resolved physician diagnosed asthma, frequent respiratory symptoms with no asthma diagnosis, and normal. Potential asthma risk factors and measures of severity were assessed across these phenotypes.
METHODS
Study Population
The National Center for Health Statistics of the Centers for Disease Control and Prevention (Atlanta, GA) conducted NHANES III from 1988 through 1994.19 NHANES III was approved by the Institutional Review Board of the National Center for Health Statistics. Survey participants completed extensive questionnaires in the household and underwent comprehensive physical examination, including pulmonary function testing, at a specially equipped mobile examination center. A knowledgeable proxy, usually a parent or a guardian, completed questionnaires for participants who were younger than 17 years.
Participants and Demographics
We limited the analysis to children who were aged 6 to 16 years and had data on age, gender, race, and BMI. NHANES III participants underwent a physical examination that included allergy skin testing for children aged 6 years and pulmonary function testing for children aged 8 years. Allergens tested included 3 indoor allergens (house mite, cat, and cockroach), 6 outdoor allergens (ragweed, perennial rye, Alternaria, Bermuda grass, Russian thistle, and white oak), and 1 food allergen (peanut). Laboratory analysis, including serum cotinine levels, was completed on children who were 4 years of age.19
Phenotype Definition
To delineate children who had asthma from those who did not have asthma, we used a positive answer to the question, "Has a doctor ever told you that your child has asthma" Children then were classified as currently having asthma when the respondent answered positively when asked, "Does the child still have asthma" Among this group, the children were stratified further by their allergen skin test status. Those with a positive reaction (wheal diameter 3 mm than saline control) for 1 or more allergens were classified as having atopic asthma. Those who reported a history of positive skin testing or eczema but did not undergo skin testing (n = 5) were also included in this group. Those with negative skin testing were classified as having nonatopic asthma. Patients with current asthma but no skin testing were excluded from the study (n = 48). Those who had a previous diagnosis but denied having current asthma were classified as having resolved asthma. Children without a diagnosis of asthma but who reported moderate or severe respiratory symptoms (see below) were classified as having frequent respiratory symptoms. Finally, all children who did not meet the previous criteria were classified as normal.
Variable Definitions
We classified the race/ethnicity of the participating children as white, black, Mexican-American, or other. We used education of the responding adult and poverty/income ratios (PIRs) as proxies for socioeconomic status (SES). The PIR is determined on the basis of the family income and number of people in the household.19 Age- and gender-specific percentiles were calculated for BMI.20 Children were classified as having been exposed to prenatal maternal smoke when the respondent answered positively to the question, "Did mother smoke while pregnant with child" (asked only of children 11 years and younger). Finally, children were classified as having attended child care before age 4.
Serum cotinine levels were determined using high-performance liquid chromatography atmospheric-pressure chemical ionization tandem mass spectrometry, as described elsewhere.21 Those with cotinine levels >15 ng/mL were considered active smokers.21 For PIR, BMI, education, and smoke exposure, we used continuous data in the regression models.
Measures of Severity
Severity of respiratory symptoms was classified as no symptoms when the respondent reported no episodes of coughing, wheezing, or upper respiratory tract infection in the past 12 months. Mild symptoms were defined as 1 to 11 episodes, moderate symptoms as 12 to 300 episodes, and severe symptoms as >300 episodes (typically reported as "daily") in the past 12 months. Symptom aggravation was determined by positive answers to questions that asked whether respiratory symptoms were brought on by pollen, house dust, animals, and exercise or cold air. Skin-test positivity was stratified into indoor allergen, outdoor allergen, and food allergen groups as defined above. Respondents were asked to list any prescription medications that the children were using and the diagnosed condition for these medications. We searched for medication that is typically used for asthma (eg, inhaled bronchodilators, inhaled steroids) and classified these as inhaled steroids, inhaled bronchodilators, or other medication used for asthma. Indicators of health status were determined by the occurrence of at least 1 hospitalization for wheezing in the past 12 months, at least 1 emergency department (ED) or unscheduled doctor’s office visit for wheezing in the past 12 months, and >5 school absences in the past 12 months. Also, the respondent classified the child’s health status according to his or her impression. This is reported as excellent/very good or as good/fair/poor. Finally, the examining doctor also classified the child’s health status according to his or her opinion. This is also reported as excellent/very good or as good/fair/poor.
Spirometry was conducted on children who were 8 years or older using a dry rolling seal spirometer in the mobile examination center. Procedures for testing were based on the 1987 American Thoracic Society Recommendations.22 To obtain spirometry acceptable according to the protocol, we performed 5 to 8 forced expirations. Several measures of lung function were used: the forced expiratory volume in 1 second (FEV1), the forced vital capacity (FVC), and the FEV1/FVC ratio. Published prediction equations based on NHANES III data23 were used to determine which participants had a low FEV1, defined as <80% of the predicted value. The FEV1/FVC ratio was also dichotomized with 0.80 as the cutoff.
Statistical Analysis
We calculated all estimates using the appropriate sampling weight to represent US children who were 6 to 16 years of age. For analyses, we used both SAS and SUDAAN.24,25 Weighted percentages are reported stratified by phenotype for variables and outcome measures except where geometric mean is indicated. We developed logistic and linear regression models to predict asthma risk factors by comparing children with each phenotype, separately, with normal children. We developed similar models to examine the outcomes of asthma severity, but, in addition, compared children in the asthma phenotype groups with children with atopic asthma. Variables included in these models were age, gender, race/ethnicity, PIR, education, BMI, cotinine level, prenatal maternal smoking, and child care attendance. Categorical outcome variables were dichotomized as follows: ethnicity, white versus all others; respiratory symptoms, severe or moderate versus mild or no symptoms; adult respondent’s impression of health status, excellent or very good versus less than very good; and physician’s impression of health status, excellent or very good versus less than very good. P < .05 for the regression coefficient was considered significant.
RESULTS
Of the 13944 children who participated in NHANES III, 8261 were younger than 6 years and excluded from the analysis. We also excluded 390 for missing BMI, 48 children with current asthma but no skin testing, and 1 for missing cotinine levels. This resulted in an analytic sample of 5244 children who represented 39.6 million US children. The 439 children who were 6 years and older and excluded from the study sample were similar to the children who were included with regard to age, gender, race/ethnicity, reported prenatal smoking, and child care attendance (P < .05 for all).
Demographic characteristics for the analytic sample are depicted in Table 1. We classified 4.8% of children as having atopic asthma, 1.9% as having nonatopic asthma, 3.4% as having resolved asthma, 4.3% as having frequent respiratory symptoms with no asthma diagnosis, and 85.6% as normal (Tables 1 and 2). Children with the 4 asthma phenotypes were similar to normal children for most demographic or potential risk factors. Exceptions were that in multivariate models that compared the 4 groups with normal children, children with atopic asthma had a lower mean PIR and a higher mean parental education level, children with nonatopic asthma had a higher mean BMI and a higher mean parental education level, children with resolved asthma had a higher prevalence of prenatal maternal smoking, and children with frequent respiratory symptoms had a higher prevalence of child care attendance and a higher mean parental education level (Table 2; the confidence intervals for these associations are available from the authors).
Among children with atopic asthma, almost all (96.5%) reacted to at least 1 indoor allergen (Table 2). This finding can be contrasted with 41.1% of normal children who reacted to at least 1 indoor allergen (Table 2).
Children with resolved asthma were, in general, more similar to normal children than those with atopic asthma (Table 3). They did, however, have significantly lower lung function, as determined by the FEV1/FVC ratio, than did normal children (Table 3).
Children who reported frequent respiratory symptoms had more symptom aggravation by pollen, dust, or animals and had a higher proportion reporting hospitalization for wheezing or an ED visit for wheezing in the previous year when compared with normal children, consistent with a definition requiring frequent respiratory symptoms for inclusion (Table 3). Their lung function was similar to that of normal children and significantly better than that in children with atopic asthma (Table 3).
DISCUSSION
In this nationally representative sample of US children, we classified 4.8% of children as having atopic asthma, 1.9% as having nonatopic asthma, 3.4% as having resolved asthma, and 4.3% as having frequent respiratory symptoms with no asthma diagnosis. We found important similarities and difference for both potential risk factors and measures of severity among our 4 asthma phenotypes. Children with nonatopic asthma had higher mean BMIs. Prenatal maternal smoking and child care attendance were associated with resolved asthma phenotype and frequent respiratory symptom phenotype, respectively. Children with atopic and nonatopic asthma share the highest burden of hospitalization, ED visits, and school absences. A substantial proportion of all 3 phenotypes of physician-diagnosed asthma demonstrated lung function impairment. Of note, the frequent respiratory symptoms phenotype had a slightly higher prevalence of certain adverse events (symptom aggravation, ED visits, and school absences) than normal children but far less than those with a current physician diagnosis of asthma.
The association of increased BMI and asthma has been shown internationally in children and adults.26–28 Our study adds to this knowledge by examining the effect of BMI across asthma phenotype. Children with nonatopic asthma had a significantly higher mean BMI (68.5 vs 57.7; P < .05) than normal children. These results support the existence of a complex relationship between asthma and BMI that may vary by asthma phenotype; however, the true relationship needs to be examined in longitudinal studies.
In our analysis, we expected serum cotinine levels to emerge as a phenotypic indicator because involuntary smoking by children has been linked to respiratory infections, middle ear disease, and asthma.29–31 Current exposure to environmental tobacco smoke, however, did not predict any of the respiratory phenotypes in our cross-sectional analysis. Our data did concur with previous reports that children who are exposed in utero to tobacco smoke are at increased risk for wheezing that later resolves with age.10,32,33 This group of children with "resolved" asthma had evidence that their respiratory disease had not truly resolved, but, perhaps, were less symptomatic at the time of the evaluation. Ominously, these children had lung function impairment similar to that seen in children with "current" asthma, suggesting that they may have ongoing lung inflammation, airway remodeling, or scarring related to the previous asthma diagnosis, tobacco smoke exposure, or other factors.34
Another important finding in this analysis was that atopic and nonatopic asthma do not differ substantially with respect to most asthma severity measures, with the notable exception of symptom aggravation by dust or animals. This finding was surprising in that the literature documents differences in airway lability in children and risk factors in adults between atopic and nonatopic asthma.9,15 However, there is evidence that immunologic findings (eosinophilia and elevated serum immunoglobulin E) are similar among the 2 groups.6,17 Our data support these findings in a new arena and suggest that atopic and nonatopic asthma may be more similar than different clinically as well as epidemiologically.
Those who experience frequent respiratory symptoms in the absence of an asthma diagnosis are often classified in the literature as having probable or undiagnosed asthma.11–14 In our study, this phenotype was distinctly different from those with a physician diagnosis of asthma, particularly with regard to asthma severity measures and lung function. Therefore, it is possible that many children with frequent respiratory symptoms are not undiagnosed or have probable asthma but represent a process different from that of asthma. A small proportion of these children (<5%) were on "asthma" medications but for diagnoses other than asthma (Table 3).
This study is potentially limited by its cross-sectional design as it can identify only associations and cannot establish causation. In addition, the possibility of misclassification exists as much of the data were self-reported by an adult proxy. Finally, the phenotypic definitions may not represent true differences in asthma presentation. Asthma is a complex disorder that may display various characteristics from each of the defined phenotypes.
In conclusion, asthma, which is an important cause of morbidity in US children, probably represents several different clinical entities with different risk factors and outcomes. Better subclassifications of both children and adults with asthma may ultimately lead to better interventions and treatments. A particularly worrisome finding in this analysis is the high proportion of children with "resolved" asthma who also have lung function impairment, suggesting that these children may be at risk for lower lung function or the development of chronic obstructive pulmonary disease as adults and may merit close clinical monitoring.
FOOTNOTES
Accepted Jul 28, 2004.
No conflict of interest declared.
REFERENCES
Maddox L, Schwartz DA. The pathophysiology of asthma. Annu Rev Med. 2002;53 :477 –498
Akinbami LJ, Schoendorf KC. Trends in childhood asthma: prevalence, health care utilization, and mortality. Pediatrics. 2002;110(suppl) :315 –322
Mannino DM, Homa DM, Pertowski CA, et al. Surveillance for asthma—United States, 1960–1995. CDC Surveill Summ. 1998;47 :1 –27
Mannino DM, Homa DM, Akinbami LJ, Moorman JE, Gwynn C, Redd SC. Surveillance for asthma—United States, 1980–1999. CDC Surveill Summ. 2002;51 :1 –13
Rackeman FM. A working classification of asthma. Am J Med. 1947;3 :601 –606
Bettiol J, Bartsch P, Louis R, et al. Cytokine production from peripheral whole blood in atopic and nonatopic asthmatics: relationship with blood and sputum eosinophilia and serum IgE levels. Allergy. 2000;55 :1134 –1141
Ulrik CS, Backer V, Dirksen A, Pedersen M, Koch C. Extrinsic and intrinsic asthma from childhood to adult age: a 10-yr follow-up. Respir Med. 1995;89 :547 –554
Martinez FD. What have we learned from the Tucson Children’s Respiratory Study Paediatr Respir Rev. 2002;3 :193 –197
Romanet-Manent S, Charpin D, Magnan A, Lanteaume A, Vervloet D. Allergic vs nonallergic asthma: what makes the difference Allergy. 2002;57 :607 –613
Taussig LM, Wright AL, Holberg CJ, Halonen M, Morgan WJ, Martinez FD. Tucson Children’s Respiratory Study: 1980 to present. J Allergy Clin Immunol. 2003;111 :661 –675
Yeatts K, Davis KJ, Sotir M, Herget C, Shy C. Who gets diagnosed with asthma Frequent wheeze among adolescents with and without a diagnosis of asthma. Pediatrics. 2003;111(suppl) :1046 –1054
Nathell L, Larsson K, Jensen I. Determinants of undiagnosed asthma. Allergy. 2002;57 :687 –693
Silverman EK, Chapman HA, Drazen JM, et al. Genetic epidemiology of severe, early-onset chronic obstructive pulmonary disease. Risk to relatives for airflow obstruction and chronic bronchitis. Am J Respir Crit Care Med. 1998;157(suppl) :1770 –1778
Clark NM, Brown R, Joseph CL, et al. Issues in identifying asthma and estimating prevalence in an urban school population. J Clin Epidemiol. 2002;55 :870 –881
Stein RT, Holberg CJ, Morgan WJ, et al. Peak flow variability, methacholine responsiveness and atopy as markers for detecting different wheezing phenotypes in childhood. Thorax. 1997;52 :946 –952
Walker C, Bode E, Boer L, Hansel TT, Blaser K, Virchow JC Jr. Allergic and nonallergic asthmatics have distinct patterns of T-cell activation and cytokine production in peripheral blood and bronchoalveolar lavage. Am Rev Respir Dis. 1992;146 :109 –115
Humbert M, Menz G, Ying S, et al. The immunopathology of extrinsic (atopic) and intrinsic (non-atopic) asthma: more similarities than differences. Immunol Today. 1999;20 :528 –533
Kurukulaaratchy RJ, Fenn MH, Waterhouse LM, Matthews SM, Holgate ST, Arshad SH. Characterization of wheezing phenotypes in the first 10 years of life. Clin Exp Allergy. 2003;33 :573 –578
Plan and operation of the Third National Health and Nutrition Examination Survey, 1988–94. Series 1: programs and collection procedures. Vital Health Stat 1. 1994;(32) :1 –407
Executive summary of the clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Arch Intern Med 1998;158 :1855 –1867
Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Maurer KR. Exposure of the US population to environmental tobacco smoke: the Third National Health and Nutrition Examination Survey, 1988 to 1991. JAMA. 1996;275 :1233 –1240
Standardization of spirometry—1987 update. Statement of the American Thoracic Society. Am Rev Respir Dis. 1987;136 :1285 –1298
Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med. 1999;159 :179 –187
SAS Institute. SAS Language: Reference, Version 6. Cary, NC: SAS Institute; 1990
Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.0. Research Triangle Park, NC: Research Triangle Institute; 1996
von Mutius E, Schwartz J, Neas LM, Dockery D, Weiss ST. Relation of body mass index to asthma and atopy in children: the National Health and Nutrition Examination Study III. Thorax. 2001;56 :835 –838
Young SY, Gunzenhauser JD, Malone KE, McTiernan A. Body mass index and asthma in the military population of the northwestern United States. Arch Intern Med. 2001;161 :1605 –1611
Soyseth V, Kongerud J, Boe J. Postnatal maternal smoking increases the prevalence of asthma but not of bronchial hyperresponsiveness or atopy in their children. Chest. 1995;107 :389 –394
Infante-Rivard C. Childhood asthma and indoor environmental risk factors. Am J Epidemiol. 1993;137 :834 –844
Cook DG, Strachan DP. Health effects of passive smoking-10: summary of effects of parental smoking on the respiratory health of children and implications for research. Thorax. 1999;54 :357 –366
Stein RT, Holberg CJ, Sherrill D, et al. Influence of parental smoking on respiratory symptoms during the first decade of life: the Tucson Children’s Respiratory Study. Am J Epidemiol. 1999;149 :1030 –1037
Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M, Morgan WJ. Asthma and wheezing in the first six years of life. The Group Health Medical Associates. N Engl J Med. 1995;332 :133 –138
Elliot J, Vullermin P, Robinson P. Maternal cigarette smoking is associated with increased inner airway wall thickness in children who die from sudden infant death syndrome. Am J Respir Crit Care Med. 1998;158 :802 –806(Colleen F. Kelley, MD, Da)
Air Pollution and Respiratory Health Branch, Division of Environmental Hazards and Health Effects, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia
Division of Pulmonary and Critical Care Medicine, University of Kentucky Medical Center, Lexington, Kentucky
ABSTRACT
Objective. To examine a nationally representative sample of US children aged 6 to 16 years old and determine whether there are differences in risk factors and measures of severity between children with different asthma phenotypes.
Methods. We analyzed data from the Third National Health and Nutrition Examination Survey. We used questionnaire and skin-prick testing data to separate children into the following mutually exclusive categories: atopic asthma, nonatopic asthma, resolved asthma, frequent respiratory symptoms with no asthma diagnosis, and normal. We used multivariate regression to determine whether demographic or potential risk factors varied between phenotypes and whether measures of severity varied by phenotype.
Results. We found that 4.8% of children had atopic asthma, 1.9% had nonatopic asthma, 3.4% had resolved asthma, and 4.3% had frequent respiratory symptoms. Risk factors varied by phenotype, for example, the mean BMI was higher among children with nonatopic asthma, prenatal maternal smoking was a risk factor for resolved asthma, and child care attendance was a risk factor for frequent respiratory symptoms with no asthma diagnosis. Patients with atopic and nonatopic asthma were similar for most measures of asthma severity (medication use, health status, and lung function impairment). In contrast, patients with resolved asthma had fewer symptoms but a similar level of lung function impairment to that seen in patients with current asthma, whereas children with frequent respiratory symptoms but no asthma diagnosis had normal lung function.
Conclusions. Asthma risk factors and measures of severity vary between children with different asthma phenotypes.
Key Words: asthma atopy allergy children lung function
Abbreviations: NHANES III, Third National Health and Nutrition Examination Survey SES, socioeconomic status PIR, pov-erty/income ratio FEV1, forced expiratory volume in 1 second FVC, forced vital capacity ED, emergency department
Asthma is a chronic inflammatory disorder of the airways that causes recurrent episodes of wheezing and other respiratory symptoms in an estimated 5 million children in the United States.1,2 Asthma prevalence, morbidity, and mortality has increased in the United States since 1980 for reasons that are not clear.3,4
For decades, clinical differences in asthma presentation have been recognized and led to its description as a heterogeneous disorder.5 Subtypes originally identified were referred to as intrinsic and extrinsic asthma, but this terminology has since been abandoned and relabeled as atopic and nonatopic asthma, often delineated on the basis of positive skin tests to common allergens or the presence of antibodies in the blood.5–9 In addition to these 2 clinical phenotypes, another phenotype describing children whose asthma resolves as they age has been reported and linked in the literature to premature birth and prenatal maternal smoking.8,10 Finally, many studies have addressed the prevalence and risk factors of "undiagnosed asthma" defined by frequent respiratory symptoms or wheezing.11–14
Although many studies have described the similarities and differences between asthma phenotypes for immunologic markers and airway lability,6,15–17 fewer have described the epidemiologic and clinical characteristics of asthma across different phenotypes in a generalizable sample.8,18 We analyzed data among children who were aged 6 through 16 years from the Third National Health and Nutrition Examination Survey (NHANES III) and classified them into 5 respiratory phenotypes: current physician-diagnosed atopic asthma, current physician-diagnosed nonatopic asthma, resolved physician diagnosed asthma, frequent respiratory symptoms with no asthma diagnosis, and normal. Potential asthma risk factors and measures of severity were assessed across these phenotypes.
METHODS
Study Population
The National Center for Health Statistics of the Centers for Disease Control and Prevention (Atlanta, GA) conducted NHANES III from 1988 through 1994.19 NHANES III was approved by the Institutional Review Board of the National Center for Health Statistics. Survey participants completed extensive questionnaires in the household and underwent comprehensive physical examination, including pulmonary function testing, at a specially equipped mobile examination center. A knowledgeable proxy, usually a parent or a guardian, completed questionnaires for participants who were younger than 17 years.
Participants and Demographics
We limited the analysis to children who were aged 6 to 16 years and had data on age, gender, race, and BMI. NHANES III participants underwent a physical examination that included allergy skin testing for children aged 6 years and pulmonary function testing for children aged 8 years. Allergens tested included 3 indoor allergens (house mite, cat, and cockroach), 6 outdoor allergens (ragweed, perennial rye, Alternaria, Bermuda grass, Russian thistle, and white oak), and 1 food allergen (peanut). Laboratory analysis, including serum cotinine levels, was completed on children who were 4 years of age.19
Phenotype Definition
To delineate children who had asthma from those who did not have asthma, we used a positive answer to the question, "Has a doctor ever told you that your child has asthma" Children then were classified as currently having asthma when the respondent answered positively when asked, "Does the child still have asthma" Among this group, the children were stratified further by their allergen skin test status. Those with a positive reaction (wheal diameter 3 mm than saline control) for 1 or more allergens were classified as having atopic asthma. Those who reported a history of positive skin testing or eczema but did not undergo skin testing (n = 5) were also included in this group. Those with negative skin testing were classified as having nonatopic asthma. Patients with current asthma but no skin testing were excluded from the study (n = 48). Those who had a previous diagnosis but denied having current asthma were classified as having resolved asthma. Children without a diagnosis of asthma but who reported moderate or severe respiratory symptoms (see below) were classified as having frequent respiratory symptoms. Finally, all children who did not meet the previous criteria were classified as normal.
Variable Definitions
We classified the race/ethnicity of the participating children as white, black, Mexican-American, or other. We used education of the responding adult and poverty/income ratios (PIRs) as proxies for socioeconomic status (SES). The PIR is determined on the basis of the family income and number of people in the household.19 Age- and gender-specific percentiles were calculated for BMI.20 Children were classified as having been exposed to prenatal maternal smoke when the respondent answered positively to the question, "Did mother smoke while pregnant with child" (asked only of children 11 years and younger). Finally, children were classified as having attended child care before age 4.
Serum cotinine levels were determined using high-performance liquid chromatography atmospheric-pressure chemical ionization tandem mass spectrometry, as described elsewhere.21 Those with cotinine levels >15 ng/mL were considered active smokers.21 For PIR, BMI, education, and smoke exposure, we used continuous data in the regression models.
Measures of Severity
Severity of respiratory symptoms was classified as no symptoms when the respondent reported no episodes of coughing, wheezing, or upper respiratory tract infection in the past 12 months. Mild symptoms were defined as 1 to 11 episodes, moderate symptoms as 12 to 300 episodes, and severe symptoms as >300 episodes (typically reported as "daily") in the past 12 months. Symptom aggravation was determined by positive answers to questions that asked whether respiratory symptoms were brought on by pollen, house dust, animals, and exercise or cold air. Skin-test positivity was stratified into indoor allergen, outdoor allergen, and food allergen groups as defined above. Respondents were asked to list any prescription medications that the children were using and the diagnosed condition for these medications. We searched for medication that is typically used for asthma (eg, inhaled bronchodilators, inhaled steroids) and classified these as inhaled steroids, inhaled bronchodilators, or other medication used for asthma. Indicators of health status were determined by the occurrence of at least 1 hospitalization for wheezing in the past 12 months, at least 1 emergency department (ED) or unscheduled doctor’s office visit for wheezing in the past 12 months, and >5 school absences in the past 12 months. Also, the respondent classified the child’s health status according to his or her impression. This is reported as excellent/very good or as good/fair/poor. Finally, the examining doctor also classified the child’s health status according to his or her opinion. This is also reported as excellent/very good or as good/fair/poor.
Spirometry was conducted on children who were 8 years or older using a dry rolling seal spirometer in the mobile examination center. Procedures for testing were based on the 1987 American Thoracic Society Recommendations.22 To obtain spirometry acceptable according to the protocol, we performed 5 to 8 forced expirations. Several measures of lung function were used: the forced expiratory volume in 1 second (FEV1), the forced vital capacity (FVC), and the FEV1/FVC ratio. Published prediction equations based on NHANES III data23 were used to determine which participants had a low FEV1, defined as <80% of the predicted value. The FEV1/FVC ratio was also dichotomized with 0.80 as the cutoff.
Statistical Analysis
We calculated all estimates using the appropriate sampling weight to represent US children who were 6 to 16 years of age. For analyses, we used both SAS and SUDAAN.24,25 Weighted percentages are reported stratified by phenotype for variables and outcome measures except where geometric mean is indicated. We developed logistic and linear regression models to predict asthma risk factors by comparing children with each phenotype, separately, with normal children. We developed similar models to examine the outcomes of asthma severity, but, in addition, compared children in the asthma phenotype groups with children with atopic asthma. Variables included in these models were age, gender, race/ethnicity, PIR, education, BMI, cotinine level, prenatal maternal smoking, and child care attendance. Categorical outcome variables were dichotomized as follows: ethnicity, white versus all others; respiratory symptoms, severe or moderate versus mild or no symptoms; adult respondent’s impression of health status, excellent or very good versus less than very good; and physician’s impression of health status, excellent or very good versus less than very good. P < .05 for the regression coefficient was considered significant.
RESULTS
Of the 13944 children who participated in NHANES III, 8261 were younger than 6 years and excluded from the analysis. We also excluded 390 for missing BMI, 48 children with current asthma but no skin testing, and 1 for missing cotinine levels. This resulted in an analytic sample of 5244 children who represented 39.6 million US children. The 439 children who were 6 years and older and excluded from the study sample were similar to the children who were included with regard to age, gender, race/ethnicity, reported prenatal smoking, and child care attendance (P < .05 for all).
Demographic characteristics for the analytic sample are depicted in Table 1. We classified 4.8% of children as having atopic asthma, 1.9% as having nonatopic asthma, 3.4% as having resolved asthma, 4.3% as having frequent respiratory symptoms with no asthma diagnosis, and 85.6% as normal (Tables 1 and 2). Children with the 4 asthma phenotypes were similar to normal children for most demographic or potential risk factors. Exceptions were that in multivariate models that compared the 4 groups with normal children, children with atopic asthma had a lower mean PIR and a higher mean parental education level, children with nonatopic asthma had a higher mean BMI and a higher mean parental education level, children with resolved asthma had a higher prevalence of prenatal maternal smoking, and children with frequent respiratory symptoms had a higher prevalence of child care attendance and a higher mean parental education level (Table 2; the confidence intervals for these associations are available from the authors).
Among children with atopic asthma, almost all (96.5%) reacted to at least 1 indoor allergen (Table 2). This finding can be contrasted with 41.1% of normal children who reacted to at least 1 indoor allergen (Table 2).
Children with resolved asthma were, in general, more similar to normal children than those with atopic asthma (Table 3). They did, however, have significantly lower lung function, as determined by the FEV1/FVC ratio, than did normal children (Table 3).
Children who reported frequent respiratory symptoms had more symptom aggravation by pollen, dust, or animals and had a higher proportion reporting hospitalization for wheezing or an ED visit for wheezing in the previous year when compared with normal children, consistent with a definition requiring frequent respiratory symptoms for inclusion (Table 3). Their lung function was similar to that of normal children and significantly better than that in children with atopic asthma (Table 3).
DISCUSSION
In this nationally representative sample of US children, we classified 4.8% of children as having atopic asthma, 1.9% as having nonatopic asthma, 3.4% as having resolved asthma, and 4.3% as having frequent respiratory symptoms with no asthma diagnosis. We found important similarities and difference for both potential risk factors and measures of severity among our 4 asthma phenotypes. Children with nonatopic asthma had higher mean BMIs. Prenatal maternal smoking and child care attendance were associated with resolved asthma phenotype and frequent respiratory symptom phenotype, respectively. Children with atopic and nonatopic asthma share the highest burden of hospitalization, ED visits, and school absences. A substantial proportion of all 3 phenotypes of physician-diagnosed asthma demonstrated lung function impairment. Of note, the frequent respiratory symptoms phenotype had a slightly higher prevalence of certain adverse events (symptom aggravation, ED visits, and school absences) than normal children but far less than those with a current physician diagnosis of asthma.
The association of increased BMI and asthma has been shown internationally in children and adults.26–28 Our study adds to this knowledge by examining the effect of BMI across asthma phenotype. Children with nonatopic asthma had a significantly higher mean BMI (68.5 vs 57.7; P < .05) than normal children. These results support the existence of a complex relationship between asthma and BMI that may vary by asthma phenotype; however, the true relationship needs to be examined in longitudinal studies.
In our analysis, we expected serum cotinine levels to emerge as a phenotypic indicator because involuntary smoking by children has been linked to respiratory infections, middle ear disease, and asthma.29–31 Current exposure to environmental tobacco smoke, however, did not predict any of the respiratory phenotypes in our cross-sectional analysis. Our data did concur with previous reports that children who are exposed in utero to tobacco smoke are at increased risk for wheezing that later resolves with age.10,32,33 This group of children with "resolved" asthma had evidence that their respiratory disease had not truly resolved, but, perhaps, were less symptomatic at the time of the evaluation. Ominously, these children had lung function impairment similar to that seen in children with "current" asthma, suggesting that they may have ongoing lung inflammation, airway remodeling, or scarring related to the previous asthma diagnosis, tobacco smoke exposure, or other factors.34
Another important finding in this analysis was that atopic and nonatopic asthma do not differ substantially with respect to most asthma severity measures, with the notable exception of symptom aggravation by dust or animals. This finding was surprising in that the literature documents differences in airway lability in children and risk factors in adults between atopic and nonatopic asthma.9,15 However, there is evidence that immunologic findings (eosinophilia and elevated serum immunoglobulin E) are similar among the 2 groups.6,17 Our data support these findings in a new arena and suggest that atopic and nonatopic asthma may be more similar than different clinically as well as epidemiologically.
Those who experience frequent respiratory symptoms in the absence of an asthma diagnosis are often classified in the literature as having probable or undiagnosed asthma.11–14 In our study, this phenotype was distinctly different from those with a physician diagnosis of asthma, particularly with regard to asthma severity measures and lung function. Therefore, it is possible that many children with frequent respiratory symptoms are not undiagnosed or have probable asthma but represent a process different from that of asthma. A small proportion of these children (<5%) were on "asthma" medications but for diagnoses other than asthma (Table 3).
This study is potentially limited by its cross-sectional design as it can identify only associations and cannot establish causation. In addition, the possibility of misclassification exists as much of the data were self-reported by an adult proxy. Finally, the phenotypic definitions may not represent true differences in asthma presentation. Asthma is a complex disorder that may display various characteristics from each of the defined phenotypes.
In conclusion, asthma, which is an important cause of morbidity in US children, probably represents several different clinical entities with different risk factors and outcomes. Better subclassifications of both children and adults with asthma may ultimately lead to better interventions and treatments. A particularly worrisome finding in this analysis is the high proportion of children with "resolved" asthma who also have lung function impairment, suggesting that these children may be at risk for lower lung function or the development of chronic obstructive pulmonary disease as adults and may merit close clinical monitoring.
FOOTNOTES
Accepted Jul 28, 2004.
No conflict of interest declared.
REFERENCES
Maddox L, Schwartz DA. The pathophysiology of asthma. Annu Rev Med. 2002;53 :477 –498
Akinbami LJ, Schoendorf KC. Trends in childhood asthma: prevalence, health care utilization, and mortality. Pediatrics. 2002;110(suppl) :315 –322
Mannino DM, Homa DM, Pertowski CA, et al. Surveillance for asthma—United States, 1960–1995. CDC Surveill Summ. 1998;47 :1 –27
Mannino DM, Homa DM, Akinbami LJ, Moorman JE, Gwynn C, Redd SC. Surveillance for asthma—United States, 1980–1999. CDC Surveill Summ. 2002;51 :1 –13
Rackeman FM. A working classification of asthma. Am J Med. 1947;3 :601 –606
Bettiol J, Bartsch P, Louis R, et al. Cytokine production from peripheral whole blood in atopic and nonatopic asthmatics: relationship with blood and sputum eosinophilia and serum IgE levels. Allergy. 2000;55 :1134 –1141
Ulrik CS, Backer V, Dirksen A, Pedersen M, Koch C. Extrinsic and intrinsic asthma from childhood to adult age: a 10-yr follow-up. Respir Med. 1995;89 :547 –554
Martinez FD. What have we learned from the Tucson Children’s Respiratory Study Paediatr Respir Rev. 2002;3 :193 –197
Romanet-Manent S, Charpin D, Magnan A, Lanteaume A, Vervloet D. Allergic vs nonallergic asthma: what makes the difference Allergy. 2002;57 :607 –613
Taussig LM, Wright AL, Holberg CJ, Halonen M, Morgan WJ, Martinez FD. Tucson Children’s Respiratory Study: 1980 to present. J Allergy Clin Immunol. 2003;111 :661 –675
Yeatts K, Davis KJ, Sotir M, Herget C, Shy C. Who gets diagnosed with asthma Frequent wheeze among adolescents with and without a diagnosis of asthma. Pediatrics. 2003;111(suppl) :1046 –1054
Nathell L, Larsson K, Jensen I. Determinants of undiagnosed asthma. Allergy. 2002;57 :687 –693
Silverman EK, Chapman HA, Drazen JM, et al. Genetic epidemiology of severe, early-onset chronic obstructive pulmonary disease. Risk to relatives for airflow obstruction and chronic bronchitis. Am J Respir Crit Care Med. 1998;157(suppl) :1770 –1778
Clark NM, Brown R, Joseph CL, et al. Issues in identifying asthma and estimating prevalence in an urban school population. J Clin Epidemiol. 2002;55 :870 –881
Stein RT, Holberg CJ, Morgan WJ, et al. Peak flow variability, methacholine responsiveness and atopy as markers for detecting different wheezing phenotypes in childhood. Thorax. 1997;52 :946 –952
Walker C, Bode E, Boer L, Hansel TT, Blaser K, Virchow JC Jr. Allergic and nonallergic asthmatics have distinct patterns of T-cell activation and cytokine production in peripheral blood and bronchoalveolar lavage. Am Rev Respir Dis. 1992;146 :109 –115
Humbert M, Menz G, Ying S, et al. The immunopathology of extrinsic (atopic) and intrinsic (non-atopic) asthma: more similarities than differences. Immunol Today. 1999;20 :528 –533
Kurukulaaratchy RJ, Fenn MH, Waterhouse LM, Matthews SM, Holgate ST, Arshad SH. Characterization of wheezing phenotypes in the first 10 years of life. Clin Exp Allergy. 2003;33 :573 –578
Plan and operation of the Third National Health and Nutrition Examination Survey, 1988–94. Series 1: programs and collection procedures. Vital Health Stat 1. 1994;(32) :1 –407
Executive summary of the clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Arch Intern Med 1998;158 :1855 –1867
Pirkle JL, Flegal KM, Bernert JT, Brody DJ, Etzel RA, Maurer KR. Exposure of the US population to environmental tobacco smoke: the Third National Health and Nutrition Examination Survey, 1988 to 1991. JAMA. 1996;275 :1233 –1240
Standardization of spirometry—1987 update. Statement of the American Thoracic Society. Am Rev Respir Dis. 1987;136 :1285 –1298
Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med. 1999;159 :179 –187
SAS Institute. SAS Language: Reference, Version 6. Cary, NC: SAS Institute; 1990
Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.0. Research Triangle Park, NC: Research Triangle Institute; 1996
von Mutius E, Schwartz J, Neas LM, Dockery D, Weiss ST. Relation of body mass index to asthma and atopy in children: the National Health and Nutrition Examination Study III. Thorax. 2001;56 :835 –838
Young SY, Gunzenhauser JD, Malone KE, McTiernan A. Body mass index and asthma in the military population of the northwestern United States. Arch Intern Med. 2001;161 :1605 –1611
Soyseth V, Kongerud J, Boe J. Postnatal maternal smoking increases the prevalence of asthma but not of bronchial hyperresponsiveness or atopy in their children. Chest. 1995;107 :389 –394
Infante-Rivard C. Childhood asthma and indoor environmental risk factors. Am J Epidemiol. 1993;137 :834 –844
Cook DG, Strachan DP. Health effects of passive smoking-10: summary of effects of parental smoking on the respiratory health of children and implications for research. Thorax. 1999;54 :357 –366
Stein RT, Holberg CJ, Sherrill D, et al. Influence of parental smoking on respiratory symptoms during the first decade of life: the Tucson Children’s Respiratory Study. Am J Epidemiol. 1999;149 :1030 –1037
Martinez FD, Wright AL, Taussig LM, Holberg CJ, Halonen M, Morgan WJ. Asthma and wheezing in the first six years of life. The Group Health Medical Associates. N Engl J Med. 1995;332 :133 –138
Elliot J, Vullermin P, Robinson P. Maternal cigarette smoking is associated with increased inner airway wall thickness in children who die from sudden infant death syndrome. Am J Respir Crit Care Med. 1998;158 :802 –806(Colleen F. Kelley, MD, Da)