Background: a subclinical left ventricle diastolic dysfunction (LVDD) has been described in patients with chronic obstructive pulmonary disease (COPD). Objectives: to evaluate the prevalence of LVDD in stable severe COPD patients, to analyze its relationship with exercise capacity and to look for its possible causes (lung hyperinflation, ventricular interdependence or inflammatory mechanisms). Methods: we evaluated 106 consecutive outpatients with severe COPD (FEV1 between 30-50%). Thirty-three (31%) were excluded because of previous heart disease. A pulmonary function test, a 6-minute walking test (6MWT), a Doppler echocardiography test, including diastolic dysfunction parameters, and an analysis of arterial blood gases, NT-proBNP and serum inflammatory markers (CRP, leucocytes), were performed in all patients. Results: the prevalence of LVDD in severe stable COPD patients was 90% (80% type I, n=57, and 10% type II, n=7). A significant association between a lower E/A ratio (higher LVDD type I) and a lower exercise tolerance (6-minute walked distance (6MWD)) was found (r=0.29, p<0.05). The fully adjusted multivariable linear regression model demonstrated that a lower E/A ratio, a DLCO in the quartile 4th and a higher tobacco consumption were associated with a lower 6MWD (76, 57 and 0.7 metres, respectively, p<0.05). A significant correlation between E/A ratio and PaO2 was observed (r=0.26, p<0.05), but not with static lung hyperinflation, inflammation or right ventricle overload parameters. Conclusion: in stable severe COPD patients, the prevalence of LVDD is high and this condition might contribute in their lower exercise tolerance. Hypoxemia could have a concomitant role in their pathogenesis.