We studied respiratory system resistance by the interrupter technique (Rint) in healthy and asthmatic children, paying special attention to the effect of cheek compliance and the effects of supporting the cheeks, the influence of lung volume at which interruption was performed, the effect of direction of flow before occlusion (inspiration vs. expiration), and short-term reproducibility of this method. One hundred and thirty-two children (36 controls and 96 asthmatics) were included in the study (mean, 9.0 +/- SD 3.6 years). Rint was calculated from the ratio of the alveolar pressure (estimated from moth pressure during occlusion) to the flow prior to interruption. We observed that 1) underestimation of airway resistance due to upper airway compliance can be minimized by supporting the cheeks; Rint was significantly lower when measured without supporting the cheeks than with support of the cheeks in controls (3.9 +/- 0.9 vs. 4.7 +/- 1.2 cmH2O.L-1.s, respectively) and asthmatics (5.2 +/- 1.6 vs 6.9 +/- 2.0 cmH2O.L-1.s); the quantitative differences of Rint with and without cheek support was larger in small children and in the most obstructed children; 2) performing occlusion at mid-tidal volume accurately reflected the respiratory system resistance of the whole respiratory cycle since we observed no difference in Rint when performing occlusion at different volumes during quiet respiration or at the middle of tidal volume; 3) Rint measured during expiration was higher than Rint obtained during inspiration in controls and in asthmatics; moreover, the effect of direction of flow before occlusion was greater in the small children; 4) Rint was closely correlated to height in controls (r = -0.82; P < 0.001); and 5) short-term reproducibility (at 15 min intervals) was satisfactory in controls and asthmatics (coefficients of variation were 9% and 7%, respectively).