1. The rates of change in mean peak expiratory flow and in diurnal variation in peak flow were compared in 14 patients recovering from acute severe asthma.
2. Peak expiratory flow was measured on hospital admission, and at 6-hourly intervals for the next 3 weeks.
3. Diurnal variation in peak flow was assessed by measuring the following: amplitude (the highest minus the lowest peak expiratory flow during any given 24 h period), amplitude % mean (the highest minus the lowest peak expiratory flow during any given 24 h period divided by the mean peak expiratory flow over that period) and residual amplitude (the maximum variation about the mean peak expiratory flow during any given 24 h period).
4. Plots of diurnal variation in peak flow and peak expiratory flow against time were constructed for each patient. To enable comparison of changes in peak expiratory flow and diurnal variation in peak flow the data were transformed.
5. The rate of change for mean peak expiratory flow and for the three measures of diurnal variation in peak flow was assessed by fitting an exponential function to each set of data, and calculating the slope of the exponential curve halfway through the period of observation (10.5 days).
6. Median (range) slope for peak expiratory flow was 0.055 (0-2.57). The comparable value for amplitude was −3.15 (−1.27 to −4.22) (absolute median values compared, P = 0.0029), for amplitude % mean was −1.87 (−0.18 to −5.95) (P = 0.012) and for residual amplitude was −1.43 (−0.62 to −3.09) (P = 0.033).
7. Diurnal variation in peak flow therefore takes longer to reach a stable value than does mean peak expiratory flow. We conclude that the magnitude of diurnal variation in peak flow during recovery from an acute attack of asthma is not governed exclusively by mean airway calibre.