Antiischemic effectiveness of long-term urokinase therapy and isovolemic hemodilution therapy has been reported in patients with symptomatic coronary artery disease, but both interventions have never been compared. In patients with refractory angina pectoris and end-stage coronary artery disease (clinical functional class III), isovolemic hemodilution (n = 9) (hydroxyethyl starch solution 6%, 1-2 times/week), and urokinase therapy (n = 11) (500,000 U urokinase per i.v. injection, 3 times a week) were performed over a period of 12 weeks, each additionally to maximal conventional treatment. Apart from the assessment of clinical symptoms and rheologic parameters, invasive hemodynamic measurements were carried out at rest and during exercise testing before and after treatment. After treatment with urokinase, patients showed a significant reduction of clinical symptoms (from 19.8 +/- 6.5 to 5.0 +/- 4.3 anginal events/week, p < 0.001), fibrinogen (from 410 +/- 88 to 238 +/- 40 mg/dl, p < 0.001), plasma viscosity (from 1.45 +/- 0.10 to 1.33 +/- 0.03 mPa x s-1, p < 0.01), and no changes of hematocrit (from 0.45 +/- 0.02 to 0.45 +/- 0.02) and whole blood viscosity (from 4.7 +/- 0.5 to 4.4 +/- 0.7 mPa x s-1); however, hemodilution resulted in a decrease of hematocrit (from 0.46 +/- 0.01 to 0.39 +/- 0.01, p < 0.001) and whole blood viscosity (from 4.7 +/- 0.5 to 4.0 +/- 0.3 mPa x s-1, p < 0.001) and no changes of initially comparable levels of clinical symptoms, fibrinogen, and plasma viscosity. Hemodynamic parameters at rest improved after urokinase therapy with a reduction of pulmonary capillary wedge pressure (from 9.1 +/- 5.1 to 5.5 +/- 2.8 mmHg, p < 0.05) at comparable levels of systemic vascular resistance (from 1510 +/- 340 to 1420 +/- 510 dyn x s x cm-5). Hemodilution did not result in any significant hemodynamic changes. Apart from clinical symptoms, long-term intermittent urokinase therapy reduces pulmonary capillary wedge pressure at rest. This may reflect an improved diastolic function due to a rheological enhancement of myocardial perfusion at the level of the coronary microcirculation. Isovolemic hemodilution seems to be of no benefit.