Rijcken B, Schouten J P, Weiss S T, Rosner B, De Vries K, Van der Lende R
Department of Epidemiology, University Hospital, Faculty of Medicine, University of Groningen, The Netherlands.
Am Rev Respir Dis. 1993 Oct;148(4 Pt 1):944-9. doi: 10.1164/ajrccm/148.4_Pt_1.944.
Long-term variability of bronchial responsiveness has been studied in a random population sample of adults. During a follow-up period of 18 yr, 2,216 subjects contributed 5,012 observations to the analyses. Each subject could have as many as seven observations. Bronchial responsiveness was assessed with a histamine challenge test. The threshold value was defined as the concentration of histamine that caused a decrease in FEV1 of 10% of more (PC10). After a 3-yr interval, a change of the threshold value with two or more doubling concentrations occurred in 21% of the subjects. This proportion increased to 43% if the time between two measurements increased to 18 yr. Of all subjects with multiple observations, 41% were always nonresponders (PC10 > or = 32 mg/ml), 11% were always responders (PC10 < or = 16 mg/ml), and 48% changed responder status. Of the subjects with greater than three observations, 65% changed responder status. Regression analyses, stratified by symptom status, sex, and smoking habit, were used to estimate intraclass correlation coefficients (ICC) as a measure of variability of responsiveness. On average, ICC values were low (0.40), also indicating considerable variability. With adjustment for FEV1, age, area of residence, and eosinophil count, the ICC values decreased slightly. This indicated that with adjustment, the total variability decreased more than the within-subject variability. ICC values did not differ significantly between asymptomatic and symptomatic subjects. Analyses of responsiveness as a binary variable revealed similar results.
