Reproducibility of maximum flows in air and He-O2 and of delta Vmax50 in the assessment of the site of airflow limitation.

Maximum expiratory flows at 50% of vital capacity while breathing air and a He-O2 mixture and delta Vmax50 observed during three sittings (morning, noon and morning the day after) in reference subjects and in patients with bronchial hyperreactivity were submitted to statistical analysis by between sittings comparison, looking for correlation coefficients, paired "t" test and variation coefficient. Results of paired "t" test ruled out the influence of fatigue and of training in performing forced expiratory manoeuvres. Correlation coefficients relevant to absolute flows are highly significant in both groups, whereas coefficients of variation are satisfactory in reference subjects and slightly higher although still acceptable in patients with bronchial hyperreactivity. As far as delta Vmax50 is concerned, the results of statistical analysis are positively unacceptable, thus demonstrating that this arithmetic manipulation of absolute flows is the ground for a poor reproducibility of results.