This study evaluated a research method to examine an exacerbation of asthma induced by corticosteroid withdrawal. Ten non-smoking adult asthmatics who were stable on treatment with inhaled steroid underwent a graded reduction of the daily dose by 200 μg at weekly intervals until an exacerbation of symptoms occurred. A daily symptom, peak expiratory flow rate (PEF) and medication diary was kept. Weekly clinic visits were used to assess symptoms, spirometry, methacholine airway responsiveness (expressed as the provocative concentration to cause a fall in FEV1 of 20%, PC20), circulating eosinophils, basophils and their progenitors (Eo/B-CFU), and sputum inflammatory cells. The laboratory tests were performed blind to the clinical details. Each subject developed an exacerbation of symptoms, on average at 16 (70–26) days after the onset of steroid reduction. This was accompanied by a deterioration in each of the objective measures. There was a fall in FEV1 by 320 ml (s.e.m. 9.5) and in PC20 from 0.8 to 0.43 mg/ml. Circulating eosinophils rose from 114 (24) × 103/ml to 227 (50) × 103/ml and Eo/B-CFU rose from 31 (5.6) to 44(11.3)/106 cells. Sputum developed in five subjects and contained 36 (5.2)% eosinophils and 1.98 (0.21)% metachromatic cells (mast cells or basophils). The symptom diary and weekly questionaire were demonstrated to be valid and responsive to change. A deterioration indicated by the daily symptom score preceded changes in PEF. Treatment by an increase in steroid was followed by reversal of each of the changes. We conclude that this research method can safely produce and examine a mild exacerbation of asthma, and that an increase in airway inflammation is an early feature.