Abstract

Epidemiological problems arising from the absence of an agreed definition of asthma have led to the use of bronchial reactivity tests in community surveys of asthma prevalence. Since only a minority of the general population will develop bronchoconstriction in response to the dose of histamine considered acceptable for use in the community it is important to make maximum use of the data available. Several methods for summarising the information in the dose-response curve obtained from a histamine challenge test have been compared. A standardised histamine challenge test was administered to 797 subjects selected from two communities, and a repeat test to 106 subjects. The test was well accepted. For most subjects FEV1 rose initially after administration of histamine (median rise 100 ml), so maximum FEV1 was used as the baseline from which the 20% fall to achieve a PD20 was calculated. In order to use all the data rather than just two points on the FEV1-log dose graph, PD20 was estimated by means of curve fitting, and the values were compared with PD20 from linear interpolation. An exponential curve was found to fit the data well. Extrapolation from the maximum dose of 4 mumol up to 8 mumol was allowed in the estimation of PD20 by both methods. The curve fitting method gave slightly more reproducible PD20 values than did linear interpolation, and also gave more estimates in the range 0.03-8 mumol. The repeatability of PD20 compared well with that of asthmatic subjects tested in a clinical environment. Curve fitting has an advantage over linear interpolation in large community studies, for which analysis of data by computer is essential.