Underestimation of the peak flow variability in asthmatic children: evaluation of a new formula

Asthma guidelines suggest evaluation of peak expiratory flow (PEF) variability, but timing for the two PEF measurements is not mentioned. Usual formula calculates amplitude as percentage of mean day-night PEF values. Since PEF circadian changes follow a sinusoidal function, we reasoned that variability might be calculated by measuring PEF at 4 pm (PEF(1)) and either at 10 am or 10 pm (PEF(2)) with the formula %variability = 200 . |PEF(1)-PEF(2)|/PEF(2). Children with stable, mild intermittent asthma were recruited from an Asthma Clinic and asked to perform PEF measurements at even hours during a week, until 12 measurements covering a 24-h period were accomplished. From these measurements we calculated PEF variability through several methods. Accuracy of such methods to predict actual PEF variability was assessed through the concordance correlation coefficient (r(c)). Thirty five asthmatic children were recruited. Actual PEF variability, calculated with the usual formula using the highest and lowest PEF obtained at any time of the 24 h cycle had a median of 37.3% (range, 0-88.5%). Variability calculated through other methods was: usual formula with highest and lowest PEF obtained from the sinusoidal curve, 21.4% (r(c) = 0.79); usual formula with PEF measured at 4 pm and 4 am, 17.8% (r(c) = 0.67); proposed formula using PEF measured at 4 pm and either 10 pm, 15.9% (r(c) = 0.68), or 10 am, 17.4% (r(c) = 0.69). Some examples with PEF measured in the morning (8 am or 10 am) and at night (8 pm or 10 pm) yielded median PEF variability from 4.0% (r(c) = 0.18) to 8.7% (r(c) = 0.38). Current methods for calculating PEF variability seemed not to be accurate enough as to be confident, suggesting that an in-deep reevaluation of the usefulness of PEF variability or, conversely, of the methods to assess it, should be done.