IJCS | Volume 33, Nº3, May / June 2020

267 the stress QTd, the sensitivity was 44.4% and the specificity was 81.6% (AUC 0.585;CI 95% 0.465-0.699) with a cutoff of 46 ms. For stress QTc, the sensitivity was 58.3% and the specificity was 63.2% (AUC 0.593;CI 95% 0.472-0.706) with a cutoff of 57 ms. In relation to the sensitivity and specificity values of the traditional treadmill test, and considering the presence of ST segment depression or typical chest pain on exertion, we found a sensitivity of 72% and a specificity of 32% in our sample. Since the confidence intervals found for the QTd and QTc cutoff values included the 0.5 value, we decided not to aggregate the QT dispersion values into the traditional stress test, because any improvements in the sensitivity and specificity that we could possibly find would not have been reliable. When we analyzed the three groups formed (TN, TP and FP), the following was found: the mean values of QT dispersion at rest did not show statistically significant difference between the three groups. Respectively, 58 ± 30 ms, 47 ± 22 ms and 43 ± 19 ms, for the TN, TP and FP groups (p = 0.172). In addition, we did not observe significant difference between the mean values of the QTc dispersion at rest: 67 ± 40 ms, 55 ± 26 ms and 49 ± 21 ms, respectively, for the TN, TP and FP groups (p = 0,163). Thus, we moved on to analyze the mean QT dispersion values found between the three groups during effort. Similarly to what happened in relation to QT dispersion at rest, we found close mean values with no statistical difference between the three groups: 32 ± 11ms, 48 ± 28 ms and 42 ± 22 ms, respectively, for the TN, TP and FP groups (p = 0.124). However, when we analyzed the data of QTc dispersion of effort, we verified that the values of QTc between the three groups were different: TN (47 ± 17 ms), TP (72 ± 42 ms) and FP (61 ± 31 ms), with p = 0.003. When we compared TN and TP, we found p < 0.05; when comparing TN and FP, we also found p < 0.05; however, when VP and FP were compared, we found p > 0.05 (Table 2). In order to better assess the changes in stress induced coronary depolarization, we created a delta QT dispersion value (ΔQTD) which was obtained by the following equation: ΔdQT= QTd stress – QTd rest. Likewise, we obtained a delta value of QTc dipersion by a similar equation: ΔdQTc = QTc stress – QTc rest. The ΔdQT was −25 ± 33 ms in the TN group, 1 ± 27 ms in the TP and −2 ± 23 ms in the FP group, with statistical difference between the three groups, with p = 0.013. Comparing TN and TP, we found p < 0.05;the same was observed when TN was compared with FP (p < 0.05);in Table 2 - Clinical and electrocardiographic variables during stress test TP FP TN p-value Number of patients 26 23 14 Rest HR (bpm) 77 ± 13 79 ± 11 79 ± 14 p = 0.922 Peak HR (bpm) 134 ± 15 131 ± 21 128 ± 22 p = 0.637 Peak SP (mmHg) 171 ± 21 173 ± 39 193 ± 29 p = 0.071 Chest pain on exertion (%) 53% 45% 0 ST depression (%) 88,40% 79,20% 0 Pain with ST depression (%) 42,3% 25% 0 Rest QTd (ms) 47 ± 22 43 ± 19 58 ± 30 p = 0.172 Rest QTc (ms) 55 ± 26 49 ± 21 67 ± 40 p = 0.163 Peak QTd (ms) 48 ± 28 42 ± 22 32 ± 11 p = 0.124 Peak QTd (ms) 72 ± 42 61 ± 31 47 ± 17 p = 0.003 ΔQTD (ms) 1 ± 27 −2 ± 23 −25 ± 33 p = 0.013 ΔQTDc (ms) 17 ± 40 11 ± 30 −20 ± 45 p = 0.013 QTd: QT dispersion; QTc: QT dispersion corrected for heart rate; ΔQTD: Stress peak QT dispersion minus rest QT dispersion; ΔQTDc: QT dispersion corrected for stress peak HF minus QT dispersion corrected for HF at rest. contrast, the comparison between TP and FP showed p > 0.05. The mean ΔQTc dispersion was −20 ± 45 ms in the TN group, 17 ± 40 ms in the TP group and 11 ± 30 ms in the FP group. Again, the same “p” value of 0.013 was found between the three groups, as well as the same values of “p” for the other comparisons: TN vs VP (p < 0.05), VN vs FP (p < 0.05) and VP vs FP (p > 0.05). We did not find any statistical difference between TP vs FP. Due to the statistical difference found between the three groups, in relation to the mean values of ΔQTc, we decided to illustrate the behavior of QTc dispersion from rest to stress peak in the three groups. Figures 3, 4 and 5 show the behavior of the three groups. Discussion The aim of this study was to assess the relationship between QT interval dispersion and chronic CAD. The focus of our investigation was to evaluate the feasibility Barcelos et al. QT interval dispersion in exercise test Int J Cardiovasc Sci. 2020; 33(3):263-271 Original Article