The correlation between angiographic assessment of coronary stenoses and fractional flow reserve (FFR) is weak. Whether and how risk factors impact the diagnostic accuracy of angiography is unknown. We sought to evaluate the diagnostic accuracy of angiography by visual estimate and by quantitative coronary angiography when compared with FFR and evaluate the influence of risk factors (RF) on this accuracy.Methods and Results—
In 1382 coronary stenoses (1104 patients), percent diameter stenosis by visual estimation (DSVE) and by quantitative coronary angiography (DSQCA) was compared with FFR. Patients were divided into 4 subgroups, according to the presence of RFs, and the relationship between DSVE, DSQCA, and FFR was analyzed. Overall, DSVE was significantly higher than DSQCA (P<0.0001); nonetheless, when examined by strata of DS, DSVE was significantly smaller than DSQCA in mild stenoses, although the reverse held true for severe stenoses. Compared with FFR, a large scatter was observed for both DSVE and DSQCA. When using a dichotomous FFR value of 0.80, C statistic was significantly higher for DSVE than for DSQCA (0.712 versus 0.640, respectively; P<0.001). C statistics for DSVE decreased progressively as RFs accumulated (0.776 for ≤1 RF, 0.750 for 2 RFs, 0.713 for 3 RFs and 0.627 for ≥4 RFs; P=0.0053). In addition, in diabetics, the relationship between FFR and angiographic indices was particularly weak (C statistics: 0.524 for DSVE and 0.511 for DSQCA).Conclusions—
Overall, DSVE has a better diagnostic accuracy than DSQCA to predict the functional significance of coronary stenosis. The predictive accuracy of angiography is moderate in patients with ≤1 RFs, but weakens as RFs accumulate, especially in diabetics.