ABC | Volume 114, Nº2, February 2020

Original Article Vieira et al. Evaluation myocardial ischemia with iFR Arq Bras Cardiol. 2020; 114(2):256-264 Figure 2 – Classification tree for the logistic regression model 1; stent placement was observed in 69.2% of patients with iFR (instantaneous wave-free ratio) ≤ 0.87; and in 17.4% of patients with iFR > 0.87. iFR p < 0.001 1 Node 2 (n = 26) Node 3 (n = 69) > 0.87 ≤ 0.87 69.2% 30.8% 17.4% 82.6% 0.8 0.6 0.4 0.2 0 0 1 0 1 0.8 0.6 0.4 0.2 0 Table 2 – Logistic regression models Variable Estimative Standard error Odds ration (95%CI) p Model 1 (Intercept) 7.8161 3.0611 0.0107 Diabetes mellitus 0.4511 0.6360 1.570 (0.451; 5.461) 0.4782 Dyslipidemia 0.9722 0.7391 2.644 (0.621; 11.256) 0.1884 Moderate stenosis -1.5000 0.5819 0.223 (0.071; 0.698) 0.0099 iFR -9.7182 3.4198 0.000 (0.000; 0.049) 0.0045 Model 2 (Intercept) 9.7209 2.8715 0.0007 Moderate stenosis -1.2414 0.5389 0.289 (0.100; 0.831) 0.0212 iFR -10.9861 3.2441 0.000 (0.000; 0.010) 0.0007 CI: confidence interval. a considerable increase in flow, similarly to a coronary flow reserve (which is a directly measured parameter) greater than 2.0. In this context, the presence of an iFR > 0.90 and an FFR < 0.80 has been associated with a coronary flow reserve not limited by flow. 28 In the present study, an iFR > 0.70 was found in the moderate lesions, and a higher variability was observed in severe lesions (0.61-1.00), mostly treated with stent placement. Such variability may be due flow changes associated with collateral supplied by microcirculation, more commonly seen in chronic lesions and in vessels that the irrigated area is not significant. In addition, there were 23 lesions in diagonal, marginal, posterior descending and posterior ventricular branches, corroborating previous hypothesis. Recently, the iFR/FRR-guided assessment has been suggested in complete revascularization in coronary three-vessel disease, venous grafts, and grafts in the circumflex system. 29 The logistic regression models and the classification tress enabled the identification of the variables more frequently related with the coronary flow reserve. Diabetes mellitus, dyslipidemia, the presence of moderate stenosis and an iFR lower than 0.87 were predictors of stent implantation in moderate and severe lesions of CAD patients, in which results obtained from non‑invasive tests and those of coronary angiography were discordant. However, when the model was constructed with significant variables only, only iFR < 0.87 and the presence of moderate stenosis remained in the model, indicating the importance of a functional analysis in this group of patients. The main limitation of this study is the lack of both short‑term and long-term follow-ups, which would allow us to evaluate whether there was an improvement in the clinical outcomes of the patients. Although a mere visual estimation of the lesion is a known limitation because of interobserver variation, it in fact reflects real-world clinical practice. The primary objective of the 260