In patients with aortic valve disease and normal coronary angiograms coronary reserve was determined by the coronary sinus thermodilution technique. Three groups of patients were studied: 37 preoperative patients; 18 different patients 12-52 months after aortic valve replacement and seven control subjects with no cardiac disease. Coronary flow ratio (dipyridamole/rest) was diminished in preoperative compared with postoperative patients (1.66 +/- 0.44 vs 2.22 +/- 0.85; P less than 0.05) as well as with controls (2.80 +/- 0.84; P less than 0.01), and corresponding coronary resistance ratio (dipyridamole/rest) was higher in preoperative patients than in both other groups (0.61 +/- 0.17 vs 0.48 +/- 0.14; P less than 0.05 vs 0.37 +/- 0.10; P less than 0.01). Differences in the flow ratio, but not in the resistance ratio, were significant (P less than 0.05) in patients after aortic valve replacement compared with controls. Total coronary sinus blood flow at rest was elevated in preoperative compared with both postoperative patients and controls (252 +/- 99 vs 169 +/- 63; P less than 0.01; vs 170 +/- 35 ml.min-1, P less than 0.05), whereas flows after maximal vasodilation did not differ among the three groups (416 +/- 184 vs 361 +/- 150 vs 488 +/- 235 ml.min-1). Postoperative patients showed a distinct, though not total regression of left ventricular angiographic muscle mass index and wall thickness. Nine of the 18 postoperative patients showed a normal coronary flow reserve and nine showed subnormal response. These two subgroups did not differ with respect to preoperative macroscopic and microscopic measures of hypertrophy. Thus in aortic valve disease, the reduced coronary vasodilator capacity is mainly due to an elevated coronary flow at rest, while the maximal coronary blood flow achieved is identical to that of postoperative patients and controls. With regression of left ventricular hypertrophy, flow at rest decreases and this leads to a distinct improvement of coronary flow reserve.