Bronchodilator testing: confidence intervals derived from placebo inhalations.

We developed criteria for a statistically significant response to an inhaled bronchodilator by measuring changes in spirometry during blinded placebo inhalation in 40 patients referred to our pulmonary function laboratory for routine testing. Ninety-five percent confidence intervals for the absolute difference and the percent change after placebo inhalation were developed for PEFR, FVC, FEV1, FEV1%, and FEF25-75% using two-tailed t-distributions. Baseline pulmonary functions and clinical factors had no apparent effect on the placebo responses. Using these confidence intervals, we evaluated 72 patients referred to our pulmonary function laboratory for bronchodilator testing. Forty-three patients had a significant change in at least one of the above parameters. We could identify 38 (88%) of these responders using the confidence interval for the percent change in FEV1, percent change in FEV25-75%, and the absolute difference in FEV1, which were 12.3%, 45.1%, and 0.18 L, respectively. In contrast, ACCP criteria would have identified only 20 of these responders. We conclude that this method provides a simple and accurate approach for evaluating changes in routine spirometry.