Asthma as a Risk Factor for COPD in a Longitudinal Study

Graciela E. Silva, MPH; Duane L. Sherrill, PhD; Stefano Guerra, MD, PhD, MPH; Robert A. Barbee, MD, FCCP


CHEST. 2004;126(1) 

In This Article

Materials and Methods

Data for the present study were collected as part of the Tucson Epidemiologic Study of Airway Obstructive diseases. Subjects were selected from a random, stratified cluster sample of white, non–Mexican-American households. Details of the study design have been published previously.[16] In brief, the sampling frame consisted of census blocks stratified on the basis of ethnic group, age of the head of household, and socioeconomic status based on 1970 census data. Subjects were enrolled between 1972 and 1973. There were 12 periodic follow-up surveys obtained approximately 1.5 to 2 years apart for a total of 20 years. During each survey, subjects completed standardized questionnaires assessing respiratory symptoms, pulmonary diseases, and smoking history, as well as other diseases. Spirometric and flow-volume data were obtained during each survey with a pneumotachograph using American Thoracic Society criteria,[17] with the exception of survey 4. The initial participants numbered 3,805 from 1,655 households, with new enrollees added by marriage and births. The total number of participants at the end of the study was 5,261. In the present study we include 3,099 subjects who fulfill the selection criteria and who were ≥ 20 years old at their initial survey.

Asthma status and potential confounders including age, sex, smoking, IgE, and skin test reactivity were assessed at the initial survey. Smoking status was determined from answers to the following questions: "Have you ever smoked cigarettes regularly? (1) Yes, and I still smoke (current smoker); (2) Yes, but I no longer smoke (ex-smoker); or (3) No, I never smoked (never-smoker)."

Total serum IgE concentrations were measured by the paper radioimmunosorbent method (Pharmacia Diagnostics; Piscataway, NJ). The reported values represent the mean of duplicate tests. Blood for serum IgE levels was obtained initially from subjects who were at least 6 years of age.[18] In the present analyses, we include levels obtained from 2,302 subjects who were ≥ 20 years old at the initial survey. A logarithm base-10 transformation of IgE values was used in the present analyses to normalize the distribution. Allergy skin tests were performed using the skin-prick test technique.[19] The skin test antigens were obtained from Hollister-Stier Laboratories (Spokane, WA), and included house dust (used at a dilution of 1:10), Dematiaceae mold mix (1:100), Bermuda grass (1:20), tree mix (1:20), and weed mix (1:20). A control solution of 50% glycerine was also applied. Subjects with one or more reactions of at least 2-mm greater than that produced by the glycerin control were placed in the skin test-positive group.

Single-breath diffusing capacity of the lung for carbon monoxide (Dlco) measurements were obtained from 580 subjects during survey 7 and survey 12. The Dlco measurement technique has been previously described.[20] The DS model automated system (W.E. Collins; Braintree, MA) was used during both surveys with no changes in the procedures. The washout volume was 1.0 L, except for participants with a vital capacity of < 2.0 L for whom a washout volume of 500 mL was used. The test gas was 0.3% carbon monoxide, 10% helium, and balance nitrogen. At least two Dlco maneuvers were performed. The mean Dlco value from the two that matched within 2 mL/min/mm Hg was reported. For the purpose of these analyses, Dlco measurements were normalized for age, sex, and height utilizing the prediction equations of Crapo and Morris[21] in accordance with American Thoracic Society recommendations.[22] The percentage of predicted value of Dlco was used in this study.

Asthma status was determined at initial survey from answers to the question: Have you ever had asthma? Subjects who answered "yes, and I still have it" were considered to have active asthma, subjects who answered "yes, but I no longer have it" were considered to have inactive asthma, and subjects who answered "no" were considered to have no asthma. Subjects with active asthma were included in the analyses only if they had seen a physician about their asthma. There were 220 subjects with active asthma; of these, 211 subjects had physician-confirmed asthma. Development of chronic bronchitis (CB), physician-diagnosed emphysema, or COPD were determined at each follow-up survey. Subjects were classified as having CB if they reported having cough and phlegm for most days for as much as 3 months of the year in at least 2 consecutive years (according to the American Thoracic Society standards),[8] and had FEV1 values < 80% predicted. Subjects were classified as having physician-diagnosed emphysema if they ever reported having a physician-confirmed diagnosis of emphysema. In addition, subjects had to have FEV1 or Dlco values < 80% predicted. COPD was defined as having either CB or emphysema or both, and having either FEV1 or Dlco values < 80% predicted. All subjects included in the present study had negative findings for CB and emphysema at the initial survey. Furthermore, subjects who underwent heart or lung surgery before or at any time during the study were excluded because this may have affected their lung function. There were a total of 185 subjects excluded who had reported heart or lung surgery, including 12 from the inactive asthma and 19 from the active asthma categories. In addition, to avoid having those subjects with emphysema mask the symptoms of CB, subjects who concurrently fulfilled the criteria for CB and emphysema at any given follow-up survey were excluded from the analyses for CB. The Tucson Epidemiologic Study of Airway Obstructive Diseases was approved by the Institutional Review Board for Human Studies, and informed written consent was obtained from all subjects at the time of their enrollment into the study.

The χ2 test was used to compare differences in proportions between each of the asthma categories. The Student t test and analyses of variance were used to compare differences in mean values, and the score χ2 test for trend of odds was used to test for linear trend among the asthma categories. Independent multivariate Cox proportional hazards models were used to determine the adjusted hazard ratios (HRs) associated with each asthma category and other potential confounders at the initial survey and subsequent development of CB, emphysema, or COPD. Potential confounders included age at initial survey, sex, smoking, IgE level, and skin test. Age was centered by subtracting the mean to each subject's age. Interaction terms between log IgE and skin test reactivity, and between log IgE and smoking were also evaluated in the model. Variables with multiple categories (≥ 2) ie, asthma and smoking, were entered using indicator variables. The Wald test was used to test for significant difference in linear contrast of coefficients between inactive and active asthma categories. Cox regression survival curves using time to first event of CB, emphysema, or COPD as the primary outcome and adjusting for potential confounders were compared across the asthma categories.

Some studies[4,23] have indicated that subjects with longer duration of asthma symptoms have an increased rate of decline in lung function. Therefore, we evaluated the effect that longer asthma duration had on development of CB, emphysema, or COPD for subjects in the active asthma group. For this purpose, asthma duration was defined as the number of years with asthma prior to initial survey. This was computed by subtracting the age at which subjects reported having their first asthma attack from the subject's age at initial survey. Asthma duration was dichotomized into subjects with ≤ 50th percentile of the distribution (≤ 22.08 years) who were considered to have short asthma duration, and those > 50th percentile (> 22.08 years) who were considered to have long asthma duration. In addition, we evaluated active asthmatics according to childhood or adulthood asthma onset. Active asthmatics were categorized on whether their asthma onset was at ≤ 14 years old (childhood asthma onset, n = 88), or at > 14 years old (adulthood asthma onset, n = 123). Statistical tests were performed using statistical software (Intercooled Stata, version 7.0 for Windows; Stata Corporation; College Station, TX). A significance level of 0.05 was used for all statistical tests.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: