A previous study of cold air challenge (CACh) in children with asthma produced indirect evidence for the existence of a reaction plateau. To examine directly this phenomenon, 17 children, mean age 13 1/2 years, were challenged by isocapnic hyperventilation with -10 degrees C air at 75% of maximal voluntary ventilation for 10 minutes (extended CACh [ECACh]). Each minute FEV1 and maximum expiratory flow at 25% remaining vital capacity (V25) were measured. During "recovery" these measurements were repeated every minute for 10 minutes. Analysis of the cumulative dose-response curves constructed from the measurements revealed that a reaction plateau was reached in the last minutes of the challenge by both functions in each child. Induced changes in absolute terms were similar for both functions (FEV1: -28 +/- 12% predicted; V25: -29 +/- 17% predicted), but V25 started at a lower baseline (FEV1: 81 +/- 15% predicted; V25: 48 +/- 27% predicted) and thus arrived at a higher degree of obstruction (FEV1: 53 +/- 19% predicted; V25: 19 +/- 13% predicted). Together with the reaction plateau developing faster for V25, this indicated a dose-response influenced by baseline small airway obstruction. Although the size of the reaction differed from one child to the other (FEV1: -5% to -50% predicted; V25: -4% to -77% predicted), interindividual variability in the development of the reaction decreased toward the end of the challenge and was minimal at the plateau and in the early minutes of "recovery."(ABSTRACT TRUNCATED AT 250 WORDS)