To identify and quantify the amount of viable hibernating myocardium in patients with chronic coronary artery disease, resting ²⁰¹Tl single photon emission computed tomography (SPECT) was compared with ^99mTc-methoxyisobutyl isonitrile (MIBI) SPECT after nitrate infusion (nitrate-^99mTc-MIBI) and ²⁰¹Tl SPECT after ²⁰¹Tl with glucose-insulin-potassium infusion (²⁰¹Tl-GIK) in 25 patients. Twenty-one patients also underwent completely left ventriculography beforehand and 5±4 months afterwards. SPECT images were divided into 9 segments and scored visually from 0 (normal uptake) to 3 (absent). The defect score was calculated as the summation of the total scores (TDS) in each patient. The TDS of nitrate-^99mTc-MIBI images (6.3±4.3) and ²⁰¹Tl-GIK images (5.8±4.2) were significantly lower than the 7.4±4.3 of resting ²⁰¹Tl images (p<0.01). Based on the improvement of wall motion after coronary revascularization, the sensitivity of ²⁰¹Tl-GIK imaging (85%) was significantly higher (p<0.05), and that of nitrate-^99mTc-MIBI imaging (79%) also tended to be higher (p=0.08), than that of ²⁰¹Tl imaging (62%) in detecting viable myocardium. The specificity of the 3 methods was almost the same. The nitrate-^99mTc-MIBI and ²⁰¹Tl-GIK methods were more useful than the resting ²⁰¹Tl method for evaluating viable hibernating myocardium. Furthermore, the ²⁰¹Tl-GIK method may provide a more accurate estimate of the amount of viable myocardium than the nitrate-^99mTc-MIBI method.