The purpose of this study was to assess whether expiratory flow limitation (FL), as measured by applying negative pressure at the mouth during tidal expiration, is a better predictor of dyspnea than routine spirometry measurements. The study population consisted of 117 ambulatory patients with COPD. Dyspnea was assessed according to the ATS-DLD respiratory Questionnaire. Expiratory flow limitation was measured in supine and sitting positions, and expressed as a percentage of the expired control tidal volume affected by flow limitation (FL, % VT). Using Spearman's rank correlation (rs), we found that the correlation of dyspnea scale with FL was stronger (rs > 0.5) than with FVC (rs < -0.3) or FEV1 (rs < -0.4) in both positions. In a multiple regression analysis FL remained the best predictor of dyspnea scale even after adjustment for FEV1 (% pred). Finally, FL was almost as sensitive as FEV1 (% pred) but much more specific in assessing the severity of dyspnea scale. These findings suggest that expiratory flow limitation as measured by the negative expiratory pressure technique may be more useful in the evaluation of dyspnea in patients with COPD than spirometry measurements.