To assess whether satisfying American Thoracic Society (ATS) end-of-test spirometry criteria can be enhanced by modifying the patient's expiratory technique, we conducted a cross-over trial of two expiratory techniques in 48 patients with a range of pulmonary functions (Group 1, n = 12: FEV1/FVC < 0.45; Group 2, n = 11: FEV1/FVC, 0.45 to 0.60; Group 3, n = 16: FEV1/FVC, 0.61 to 0.74; Group 4, n = 9: FEV1/FVC > or = 0.75). After randomizing the order of testing, each patient performed three exhalations using a "standard" forced expiratory maneuver and a modified expiratory technique consisting of an initial maximal expiratory effort followed by a "relaxed expiration" for as long as possible. Patients initiated "relaxed expiration" when instructed by the supervising technician, who issued the instruction to relax when expiratory airflow fell to < or = 200 ml/s (as determined by flow-volume loop analysis). ATS end-of-test criteria were satisfied significantly more often using the modified expiratory technique (58.3% of testing sessions) than using the standard technique (18.7% of sessions, p = 0.001) because of prolongation of the forced expiratory time (FET) with the modified technique in all patient groups. In the 38 patients with FEV1/FVC < or = 0.75, the largest FVC and FET rose significantly using the modified expiratory technique, without compromising the largest FEV1 in any group. In patients with FEV1/FVC > or = 0.75, FET increased without concomitant changes in FVC or FEV1.(ABSTRACT TRUNCATED AT 250 WORDS)