Peak expiratory flow rate variability in population surveys. Does the number of assessments matter?

We investigated the effect of reducing the number of daily peak expiratory flow (PEF) measurements on (1) the amplitude of PEF variability and (2) the relationships of this amplitude to bronchial reactivity to methacholine. One hundred seventeen workers (mean age = 38.7 years +/- 9.5; men = 86.3%) recorded their highest of three PEF measurements, every 3 waking hours, ie, 5 times a day, for 7 days, each using a newly purchased peak flowmeter (Vitalograph), and underwent methacholine challenge tests. The variability of PEF of each subject was expressed using the three sets of indices: amp%mean, ie, highest of the daily measurements considered minus the lowest/mean x 100, averaged over 6 days from the second to the seventh, amp%highest (same as amp%mean, but with the highest daily measurements as denominators) and SD%mean (calculated initially as single measures using the data of the 6 days considered, with standard deviation (SD) of each subject's PEF measurements). For each set, we used the indices constructed with the five daily measurements of each day (gold standard), with the first, third, fourth, and fifth, the first, third, and fourth, the first and third, and the first and fourth. The PEF variability was significantly reduced when reducing the number of daily measurements, only when the amp%mean and the amp%highest sets were used. No decrease was observed with the SD%mean set of indices, and SD%mean constructed with the first, third, fourth, and fifth daily measurement was satisfactory. Whatever the sort of index used, three daily measurements were sufficient to identify the group of subjects with excessive variability in relation to methacholine reactivity.