The incremental prognostic value of thallium 201 imaging in patients with nonspecific ST-T abnormalities on the resting electrocardiogram (ECG) may be different from those with a normal resting ECG.

Nine hundred thirty-nine consecutive patients with nonspecific ST-T abnormalities on their resting ECG who had undergone exercise ²⁰¹Tl imaging were followed for a median duration of 7.0 y (94% complete). The Cox proportional hazards regression model was used in a stepwise fashion to generate (1) a clinical (Cl) model, (2) a clinical and exercise (Cl + Ex) model, (3) and a clinical, exercise, and thallium (Cl + Ex + Tl) model, for the prediction of cardiac death.

Age, sex, and diabetes composed the Cl model (χ ² = 63, P < .0001). The Duke treadmill score added to the Cl + Ex model (χ ² = 71, P < .0001). Increased lung uptake (P < .0001) added significantly and summed reversibility score (P = .03) added modestly to the Cl + Ex + Tl model (χ ² = 96, P < .0001). On the basis of the Cl + Ex + Tl model, the low-, intermediate-, and high-risk groups had a 7-y survival free of cardiac death of 99%, 88%, and 58%, respectively (P < .0001). Using the Cl + Ex + Tl model, only a small number of low-risk and high-risk patients by the Cl + Ex model were reclassified. However, 48% of the 230 patients in the intermediate-risk group by the Cl + Ex model were reclassified as low risk or high risk.

²⁰¹Tl imaging has incremental prognostic value in patients with nonspecific abnormalities on their resting ECG. However, patients classified as low risk or high risk by exercise testing using the Cl + Ex model do not require ²⁰¹Tl imaging. Intermediate-risk patients should be further risk-stratified by ²⁰¹Tl imaging.