To determine whether eating a high-carbohydrate meal between initial and delayed postexercise thallium-201 (Tl-201) imaging affects detection of Tl-201 redistribution during exercise stress testing, 16 patients with stable angina performed 2 Tl-201 treadmill exercise stress tests within a 14-day interval. Immediately after initial postexercise imaging, patients either drank a commercially available instant breakfast preparation for the intervention test or drank an equivalent volume of water for the control test. Comparable exercise workloads were achieved by exercising patients to the same heart rate for both tests. The order of the 2 (intervention and control) tests were randomized. All patients had at least 1 region of Tl-201 myocardial redistribution on either their eating or control test scans, although only 7 of the 16 had positive treadmill exercise test responses. Forty-six regions showing Tl-201 myocardial redistribution were identified in all 144 regions examined. Significantly more of these regions were identified on control test scans than on eating test scans: 11 of 46 on both test scans, 6 of 46 only on eating test scans and 29 of 46 only on control scans (p less than 0.001). Consistent with results of the quantitative regional analysis, the percentage of Tl-201 clearance over 4 hours in the 46 Tl-201 myocardial redistribution regions was 39 +/- 8% for the eating tests and 29 +/- 8% for control tests (mean +/- standard deviation, p less than 0.003). In 4 patients diagnosis of transient ischemia would have been missed because their 14 Tl-201 myocardial redistribution regions were detected only on the control test scans.(ABSTRACT TRUNCATED AT 250 WORDS)