During exercise, patients with chronic obstructive pulmonary disease (COPD) increase their pulmonary arterial wedge (Ppaw) and left ventricular (LV) end-diastolic pressures more than normal control subjects. The increase in pressure is commonly attributed to an increase in intrathoracic pressure (Pit). However, mean esophageal pressure (Pes) does not increase with supine exercise in patients with COPD. Because changes in Pes may not represent changes in Pit when recorded in the supine position, we measured Ppaw and Pes during upright exercise in 8 patients with severe air-flow limitation (mean +/- SD) FEV1, 0.88 +/- 0.27 L secondary to COPD and no history or electrocardiographic abnormalities suggesting a previous myocardial infarct, history of angina, evidence of systemic hypertension, or use of cardiac medications. In addition, all patients completed a progressive exercise test to exhaustion without angina or ST segment changes, and all had normal LV function at rest assessed by equilibrium radionuclide ventriculography. The Ppaw increased a mean of 7.2 +/- 4.3 mmHg, whereas Pes increased a mean of only 1.3 +/- 1.6 mmHg. By multiple linear regression analysis, Ppaw was significantly associated with the work level performed (p less than 0.01), but had no significant association with Pes (p greater than 0.1). The change in Ppaw could not be attributed to changes in Pes. If changes in Pes during upright exercise are representative of changes in Pit or juxtacardiac pressure, a rise in Pit does not explain the exercise-induced increase in Ppaw and LV end-diastolic pressure that occurs in patients with COPD.