VARIABILITY OF PEAK EXPIRATORY FLOW-RATE IN CHILDREN - SHORT AND LONG-TERM REPRODUCIBILITY

Background - Variability of peak expiratory flow (PEF) has been proposed as a surrogate for bronchial hyperresponsiveness. The normal range of variability of PEF for children has been reported and the test has been used to screen for asthma in population based studies. However, there is Little information on the reproducibility of the method in epidemiological settings. Methods - In a cohort study of primary school children the variability in PEF was recorded in two consecutive years for one week (first survey) and two weeks (second survey) using mini Wright peak flow meters. PEF was recorded twice daily (morning and evening) and average amplitude as a percentage of mean was calculated as a standard measure of PEF variability for each single week of PEF measurement. Children with PEF variability exceeding the 90% percentile of the distribution for the specific time period were regarded as having increased variability of PEF. Results - Of 66 children with increased PEF variability in the first year, 13 (19.7%) had an abnormal test in the first week of the second year. Of 543 children with normal PEF variability in the first year, 44 (8.1%) had an abnormal test in the second study year (odds ratio 2.8, confidence interval (CI) 1.4 to 5.4). Of 646 children in the second survey 61 (9.4%) were abnormal during the first week and 68 (10.5%) had an increased PEF variability during the second week, but only 24 (3.7%) children had an increased PEF variability in both weeks. The sensitivity (specificity) for doctor-diagnosed asthma (12 month period prevalence) was 36.4% (91.0%) in the first week of the second survey. When measurements of both weeks of the second survey were used to calculate PEF variability there was little improvement in the sensitivity (38.1%) and specificity (91.5%), mainly because of decreased compliance in the second measurement week. Conclusions - In young children assessment of PEF variability in order to screen for asthma is of limited value because of the low reproducibility of the method.