Forty-seven consecutive patients (mean age 61 +/- 8 years) referred for cardiac catheterization due to moderate to severe aortic (n = 30) or mitral (n = 17) valvular heart disease were examined by technetium-99m isonitrile tomography together with a high-dose dipyridamole infusion (0.7 mg/kg) and handgrip stress. Tomography did not identify coronary artery disease (CAD) in 3 of the 21 patients with angiographically proven disease (sensitivity 86%) and suggested false positive results in 5 of the 26 without the disease (specificity 81% and negative predictive accuracy 88%). No patient without angina pectoris and with negative scintigraphy (n = 14) had angiographically significant (greater than or equal to 50% diameter stenosis) CAD. Overall vessel sensitivity was 63%, and specificity was 92%. The frequency of side effects during the dipyridamole-handgrip test was only 7%. No serious complications occurred during stress tests. Thus, technetium-99m isonitrile tomographic imaging, together with high-dose dipyridamole and handgrip exercise, is a useful noninvasive method in excluding significant CAD in patients with valvular heart disease.