Low diagnostic yield of invasive coronary angiography: is the noninvasive stress testing guilty? Data from a cohort of 2600 patients

Purpose: Many patients referred for invasive coronary angiography (ICA) following clinical assessment complemented with noninvasive stress testing (NIST) do not have obstructive coronary artery disease (OCAD). We aimed to determine the predictors of OCAD and the incremental value of NIST. Methods: We analyzed a cohort of patients referred for ICA for stable CAD diagnosis in a single tertiary-care center (2006-2011). Traditional and nontraditional CAD risk factors, modified Framingham risk score (FRS), symptoms, left ventricle function (LVF), NIST and ICA results were assessed. OCAD: luminal narrowing ≥70% (≥50% for left main artery). OCAD predictors were determined by uni and multivariate analysis. To assess the incremental predictive value of each factor, a stepwise logistic regression analysis was performed, starting with (1) FRS and progressively adding (2) nontraditional risk factors, (3) symptoms, (4) pretest (NIST) probability of CAD, (5) LVF and (6) NIST result. The discriminatory power at each step was evaluated by the area under the ROC curve (AUC). Results: 2600 patients included: 65.2±9.9 years, 59% male, mean 10-year Framingham risk 18%, 10% depressed LVF, 81% positive NIST (treadmill exercise ECG/SPECT). Only 49% had OCAD. Factors independently associated with OCAD: class 3 (CCS) angina (OR 5.3, 95% CI 3.1-8.9), typical angina (OR 3.2, 95% CI 2.6-4.1), male gender (OR 2.9, 95% CI 2.4-3.7), depressed LVF (OR 2.9, 95% CI 1.9-4.3) and high FRS (OR 1.2, 95% CI 1.1-1.2), (all p<0.05). Stepwise analysis shown in the Figure. ![Figure][1] Conclusions: The diagnostic yield of ICA was low. The single most important predictor of OCAD was severe angina, while NIST did not increase the discriminatory power over the clinical judgment. Better strategies are needed to avoid unnecessary referrals to ICA. [1]: pending:yes