1. We tested the hypothesis that increasing myocardial cyclic GMP levels would reduce myocardial O2 consumption and areas of low O2 supply/consumption balance, using zaprinast, a selective cyclic GMP-phosphodiesterase inhibitor. 2. The study was conducted in three groups (vehicle, 10(-3) and 3 x 10(-3) mol/L zaprinast) of anaesthetized open-chest New Zealand white rabbits (n = 24). Coronary blood flow (radioactive microspheres), arterial and venous O2 saturation (microspectrophotometry), O2 consumption, cyclic GMP content (competitive binding) and cyclic GMP-phosphodiesterase activity (conversion of 3H-cyclic GMP to 3H-GMP) were determined. 3. Agents were applied to a patch on the myocardial surface and did not cause significant haemodynamic changes, except for bradycardia in the vehicle and low dose group. 4. The total myocardial cyclic GMP-phosphodiesterase activity was 148 +/- 14 while the zaprinast (10 mumol/L) inhibitable activity averaged 63 +/- 8 pmol/mg protein per min. Cyclic GMP content was increased with increasing doses of zaprinast (vehicle, 4.308 +/- 0.349 pmol/g; low dose zaprinast, 4.803 +/- 0.279 and high dose zaprinast, 7.938 +/- 1.304 pmol/g). 5. Coronary blood flow was not different after treatment (198 +/- 11, 209 +/- 10 and 153 +/- 9 mL/min per 100 g for the vehicle, low and high dose zaprinast, respectively). 6. Under control conditions, 48% of the small veins had O2 saturations below 50%. With zaprinast, this value was reduced to 19% for the low and 24% for the high dose. 7. Average venous O2 saturation increased with zaprinast (49 +/- 2%, 61 +/- 3% and 59 +/- 1%).(ABSTRACT TRUNCATED AT 250 WORDS)