In this study we aimed to determine the relationship between exercise capacity, clinical ratings of dyspnoea and lung function parameters in patients with severe chronic obstructive pulmonary disease (COPD) by means of the statistical method of factor analysis. Sixty two patients (mean age +/- SD, 66 +/- 9 yrs) in stable clinical condition, with a forced expiratory volume in one second (FEV1) < 65% of predicted were investigated. Before the study, therapy was optimized, including inhaled bronchodilators, theophylline and steroids. Exercise capacity was determined from the best 6 min walking distance achieved in five self-paced treadmill walks performed on consecutive days. Lung function testing comprised spirometry and body plethysmography. Four different tools were chosen to rate dyspnoea and quality of life: the Baseline Dyspnoea Index (BDI), the Oxygen Cost Diagram (OCD), a modified Medical Research Council (MRC) Scale, and the Chronic Respiratory Disease Questionnaire (CRQ). Principal component factor analysis revealed that the data could be reduced to three hypothetical underlying variables (factors), which accounted for 79% of the total variance. BDI, MRC, OCD, CRQ and walking distance were attributed to the first factor, forced expiratory volume in one second and airway resistance to the second factor, and lung volumes to the third factor. Thus, our data suggest that the pathophysiological condition of severe COPD is characterized by three statistically independent entities: 1) exercise capacity, dyspnoea and quality of life ratings; 2) airway obstruction; and 3) pulmonary hyperinflation.(ABSTRACT TRUNCATED AT 250 WORDS)