Predictive treadmill score to identify coronary artery disease

DONG Hao-jian; YU Dan-qing; ZHANG Guo-lin; LIU Zhi; ZHOU Ying-ling

摘要：Background Exercise treadmill testing(ETT)is widely used for the diagnosis of coronary artery disease(CAD).The high false-positive and false-negative rates hamper its clinical application.Hypothesis:We examined a hypothesis that combined ETT and risk-factors of CAD to develop an easily applied predictive treadmill score(PTS)and the superiority of the novel score over the conventional ETT criteria was tested.This score can improve the diagnostic accuracy of treadmill testing for Chinese patients with suspected CAD and plan management strategies for such patients.Methods The present study comprised a retrospective analysis of 300 ETTs of Chinese patients with chest pain referred to suspected CAD who also had coronary angiography within one month after ETT.Clinical characteristics and ETT results of those with and without angiographic CAD were compared.A logistic regression model was used to construct a PTS that could accurately predict clinically significant CAD,defined as the presence of at least one＞70% angiographic stenosis in a major epicardial coronary artery.The accuracy of the new PTS was compared to conventional criteria of ETT for CAD diagnosis.Results The study included 185 patients with clinically significant CAD(61.6%)and 115(38.4%)without CAD.Patients with significant CAD were older,more cigarettes smokers,and had higher prevalence of hypercholesterolemia,diabetes mellitus and serum CRP concentration.The PTS for CAD was derived from the logistic regression equation:Y(-1,1)=-6.94 + 0.07Age + 0.56Smoking + 0.7DM + 0.6TC/HDL + 0.6ST +3.5Symptom-0.01HR.According to the derived Y value,the study subjects were classified as low,intermediate and high risk CAD groups,if Y＜-1,-1≤Y≤1,and Y＞1 respectively.The area under the curve(AUC)of our PTS on the receiver operating curve(ROC)was 0.863.Using the PTS,the specificity,false-positive rate and the predictive accuracy for CAD detection were significantly improved compared to conventional ETT diagnostic criteria(73.6% vs.56.9%;26.4% vs.43.1%;77.6% vs.71.8%,P＜0.05).Conclusions A new predictive treadmill score for CAD diagnosis was validated and found superior to the conventional criteria of Etr for the diagnosis of CAD in Chinese patients presenting with chest pain.