QT Hysteresis Index Improves the Power of Treadmill Exercise Test in the Screening of Coronary Artery Disease

Abstract

Background: QT hysteresis phenomenon exists in healthy subjects, and is more exaggerated in patients with coronary artery disease (CAD) and long QT syndrome. The purpose of this study was to establish an appropriate method to evaluate the magnitude of QT hysteresis, and assess the value of QT hysteresis index in the treadmill exercise test (TET) in predicting CAD.

Methods and Results: A total of 138 subjects with suspected CAD and referred for TET and selective coronary angiography (SCA) were divided into positive (n=77) and negative (n=61) SCA groups. Dynamic ECG were recorded during TET. QT/RR curves were constructed and QTp (Q-T_peak) and QTe (Q-T_end) hysteresis indices were calculated for each subject. SYNTAX score in the positive SCA group was determined. The QTp and QTe hysteresis indices in the positive SCA group were significantly higher than in the negative SCA group. The combination of QTe hysteresis index and conventional TET criteria had the highest sensitivity and negative predictive value according to receiver operating characteristic curve, and was an independent predictor on multivariate logistic regression. QT hysteresis indices significantly correlated with SYNTAX score in the positive SCA group.

Conclusions: QTe hysteresis index enhances the specificity of predicting CAD in TET. It improves the diagnostic value of TET for CAD significantly when combined with conventional criteria and is associated with the severity of CAD. (Circ J 2014; 78: 2942–2949)

Under steady-state conditions, the QT interval correlates with the RR interval, and responds to temporary deviations in instantaneous heart rate. Under dynamic conditions, however, the QT interval at a given heart rate during recovery is shorter than that during exercise.¹ This phenomenon is termed QT hysteresis, which exists in healthy subjects² and is more exaggerated in patients with coronary artery disease (CAD)^3,4 and long QT syndrome.⁵

Although invasive methods of diagnosing CAD such as selective coronary angiography (SCA) are widely accepted, investigations into other electrocardiogram (ECG) indicators of CAD are continuing.^6–8 Exercise testing is one of the most widely used non-invasive methods for assessing CAD, but the sensitivity and predictive value of conventional ST-segment deviation criteria are limited in comparison with other non-invasive modalities such as coronary computed tomography (CT) angiography and nuclear imaging.^9,10 For this reason, researchers explored other strategies to combine oxygen uptake kinetics with the conventional ST segment change to increase the sensitivity and specificity for predicting CAD of exercise testing.¹¹ Recently, we developed a new parameter from exercise ECG for quantitatively evaluating the phenomenon of QT hysteresis in the treadmill exercise test (TET), called the QT hysteresis index, and we tested the value of using QT hysteresis index in predicting CAD. We hypothesized that the QT hysteresis index may improve the power of TET in CAD screening.

Methods

Subjects

From September 2011 to May 2013, consecutive subjects with suspected CAD who were referred for TET and SCA at Renmin Hospital of Wuhan University were enrolled in this study. Exclusion criteria included known CAD, age <18 years or >70 years, cardiac pacemaker placement, cardiomyopathy, acute myocarditis, uncontrolled hypertension, digoxin or anti-arrhythmia drug treatment, atrial fibrillation, pre-excitation syndrome, repolarization abnormality, interventricular conduction defect, or history of myocardial infarction. Patients were eligible if they were willing to provide informed consent, and were able to walk on the treadmill.

The protocol of the study and informed consent were approved by the Ethics Review Board of Renmin Hospital of Wuhan University.

TET

Before TET, treatment with β-blocker, nitrates or calcium channel blocker was ceased for at least 5 half-lives. Each subject underwent symptom-limited TET (CASE 8000; GE Medical, Waukesha, USA) according to the modified Bruce protocol.^2,12–14 The grade and speed of the initial stage was 10° and 2.7 km/h, and increased to 12° and 4.0 km/h at stage 2, 14° and 5.4 km/h at stage 3, and 16° and 6.7 km/h at stage 4, each stage lasting for 3 min. The recovery phase involved 30 s of walking at 10° and 2.7 km/h followed by an additional 5-min sitting, or until any exercise abnormality resolved. The 12-lead ECG were recorded at baseline in the supine position and continuously during the entire exercise protocol. Blood pressure was monitored at baseline in both the supine and standing positions and at 1-min intervals during the exercise test.

TET was discontinued if any of the following criteria were met: (1) reaching a target heart rate of [(220–age)×0.9] beats/min; (2) occurrence of limiting angina pain; (3) ST-segment depression >3 mm in at least 1 lead; (4) unable to continue because of fatigue.¹⁵

The conventional TET criteria used in this study consisted of (1) typical symptoms of angina pectoris; (2) horizontal or down-sloping ST-segment depression ≥0.1 mV lasting ≥2 min in at least 2 contiguous leads; and (3) ST-segment elevation ≥0.1 mV in inferior and lateral leads or ST-segment elevation ≥0.2 mV in at least 2 contiguous precordial leads lasting ≥2 min. The deviation of ST segment was measured at 80 ms after the J point. Subjects who did not meet these criteria were defined as TET negative.^16,17

Duke Treadmill Score (DTS)

Using the following equation, DTS was calculated for each patient. DTS=exercise time (min)–(5×max ST deviation in mm)–(4×treadmill angina index). Treadmill angina index was defined as follows: no angina during exercise, 0; non-limiting angina, 1; stopped exercise due to angina, 2. The score typically ranges from –25 to +15. These values correspond to low-risk (DTS ≥5), moderate-risk (DTS –10 to +4), and high-risk (DTS ≤–11) categories.^18,19

SCA and SYNTAX Score

Within 2 weeks after TET, each patient underwent SCA in multiple views using the standard Judkins techniques,²⁰ and the results were analyzed by at least 2 interventional physicians. Any one or more main coronary arteries, including left circumflex branch, left anterior descending branch, and right coronary artery, showing reduction of lumen diameter ≥50% was considered a positive result.²¹ Such patients were assigned to the positive SCA group.

For subjects with positive SCA the SYNTAX score was determined, which evaluated the severity and complexity of CAD on the basis of coronary anatomic risk factors including the number of lesions and their functional impact, location, and morphologic features.^22,23 The overall SYNTAX score was derived from the summation of individual scores for each separate lesion (defined as stenosis ≥50% in vessel ≥1.5 mm) using the SYNTAX score calculator (measurement key available at http://www.syntaxscore.com). All angiograms in the positive SCA group were measured independently by 2 experienced cardiologists and 1 technician who were blinded to clinical information, and the SYNTAX score for each subject was defined as the average of those 3 scores.

Measurement of RR and QT Interval

At baseline the QTe interval was measured on 12-lead ECG for QT dispersion calculation, which was defined as the difference between the maximum and minimum QTe interval occurring in any of the 12 ECG leads.²⁴ During TET, 12-lead ambulatory ECG was done using a Holter recorder (Marquette-3000; GE Medical). For each TET recording, ECG traces of the lead with the highest T wave amplitude (usually V2) were used for RR and QT interval measurement. QTp (from the onset of QRS complex to the peak of T wave) and QTe (from the onset of QRS complex to the end of T wave) intervals were measured every 10 s in the ECG trace using commercial software (MARS 7.2; GE Medical). RR interval was measured in median beats of the 10-s segments. Each measurement was reviewed by 2 experienced technicians blinded to the clinical information, and corrections were made to the measurements as necessary.

Quantitative Evaluation of QT Hysteresis

The method used to quantify QT hysteresis is shown in Figure 1. QT/RR curves were plotted by connecting the points in sequence using RR intervals on the x-axis and QTp or QTe intervals on the y-axis.

Figure 1.

Calculation of QT hysteresis area and QT hysteresis index. Yellow, QT hysteresis area (ms²) delineated by C–E, and F. QT hysteresis index (ms)=QT hysteresis area (ms²)/distance between A and B.

The QT hysteresis area was calculated over a range of RR intervals from 10% to 90% of RR interval variation. To determine the variation range of RR interval, the maximum RR interval was recorded at the onset of exercise or the termination of recovery and the minimum RR interval was defined as at the peak of exercise. Two vertical lines were drawn at RR intervals representing 10% and 90% of the RR interval variation range, respectively. These 2 lines and QT/RR curves during exercise and recovery formed a loop, the area of which is the QT hysteresis area (ms²). The QT hysteresis index (ms) is defined as the QT hysteresis area divided by the 10–90% RR interval variation range (distance between A and B in Figure 1).

Statistical Analysis

Continuous variables that followed a normal distribution are presented as mean±SD. Categorical variables are presented as frequency. Continuous variables were compared using independent-samples t-test or Student Newman-Keuls multiple comparison test. Discrete variables were compared using chi-squared or Fisher exact tests. A receiver operating characteristic (ROC) curve was constructed to determine sensitivity and specificity for QT hysteresis index. The optimal cut-off values were identified using the Youden index method,²⁵ which defines the cut-off in terms of the maximum sum of sensitivity and specificity (Youden index=sensitivity+specificity–1). A prediction probability for the combination of each QT hysteresis index with conventional TET criteria was generated on logistic regression analysis to assess the combined diagnostic value on ROC curve. Multiple logistic regression analysis was carried out to evaluate the prediction value of various indices related to TET. Positive predictive and negative predictive values, and area under the ROC curve (AUC) were calculated. Kappa statistic was used to assess the consistency and reliability of QT hysteresis index, DTS and conventional TET criteria compared with SCA. The differences of AUC between ROC curves were compared using MedCalc (version 13.2). Linear regression analysis with Spearman coefficient was used to clarify the relationship between QT hysteresis index and SYNTAX score. All statistical analysis was performed using SPSS (version 17.0, SPSS, Chicago, IL, USA) except that specifically noted. P<0.05 was considered statistically significant.

Results

Baseline Characteristics

A total of 138 subjects were enrolled in this study. Baseline characteristics according to SCA result are listed in Table 1. The positive SCA group had a significantly higher frequency of risk factors including male gender and hyperlipidemia, and symptoms of chest pain (both atypical and typical). The QRS width and QT intervals were significantly longer in the positive than in the negative SCA group. There was no significant difference between the 2 groups in systolic and diastolic blood pressure before and at peak exercise and heart rate before exercise, but heart rate at peak exercise in the positive SCA group was significantly slower than in the negative group.

Table 1. Baseline Characteristics vs. SCA

	Negative SCA (n=61)	Positive SCA (n=77)	P-value
Male	43 (70.5)	66 (85.7)*	0.029
Age (years)	53.5±6.7	55.6±7.8	0.097
Hypertension	21 (34.4)	36 (46.8)	0.144
Diabetes	12 (19.7)	15 (19.5)	0.391
Hyperlipidemia	19 (31.1)	38 (49.4)*	0.031
Smoker	32 (52.5)	50 (64.9)	0.138
Family history of CAD	25 (37.9)	36 (46.8)	0.498
Asymptomatic	3 (1.6)	0 (0)*	0.049
Atypical chest pain	48 (78.7)	25 (32.5)**	<0.001
Typical chest pain	10 (16.4)	52 (67.5)**	<0.001
QRS width (ms)	83.3±9.4	89.4±15.3*	0.015
QT (ms)	371.0±31.8	387.6±23.3*	0.015
QTd (ms)	33.3±10.8	29.8±10.1	0.080
Before exercise
SBP (mmHg)	123±15	126±19	0.331
DBP (mmHg)	72±9	73±12	0.486
Heart rate (beats/min)	73±1	71±9	0.296
Peak exercise
SBP (mmHg)	178±27	175±25	0.477
DBP (mmHg)	76±12	77±13	0.861
Heart rate (beats/min)	152±10	142±17**	<0.001
QTp hysteresis index (ms)	7±6	19±11**	<0.001
QTe hysteresis index (ms)	8±6	24±13**	<0.001
QTp hysteresis area (ms2)†	2,224±1,208	6,352±2,644**	<0.001
QTe hysteresis area (ms2)†	2,357±1,335	7,656±3,236**	<0.001
10–90% RR interval variation range (ms)‡	335±67	323±59	0.145
DTS	3.62±5.06	−4.95±7.32**	<0.001

Data given as mean±SD or n (%). *P<0.05, **P<0.001. ^†Area surrounded by C–E, and F in Figure 1; ^‡distance between A and B in Figure 1. CAD, coronary artery disease; DBP, diastolic blood pressure; DTS, Duke treadmill score; QTd, QT dispersion; QTe, Q-T_end; QTp, Q-T_peak; SBP, systolic blood pressure; SCA, selective coronary angiography.

Higher QT Hysteresis Indices in CAD Patients

As shown in Table 1, QTp and QTe hysteresis indices were significantly higher (P<0.001) in the positive SCA group (19.3±10.7 ms and 24.0±12.8 ms, respectively) than in the negative group (6.5±6.4 ms and 7.6±6.3 ms, respectively). QTp and QTe hysteresis areas were significantly larger (P<0.001) in the positive SCA group (6,352±2,644 ms² and 7,656±3,236 ms², respectively) than in the negative group (2,224±1,208 ms² and 2,357±1,335 ms², respectively). There was no significant difference in the RR interval variation range between the 2 groups. The DTS was significantly lower in the positive SCA group than in the negative group (–4.95±7.32 vs. 3.62±5.06, P<0.001).

Typical QT/RR curves of subjects with different SCA or TET results are shown in Figure 2. In subjects who had positive SCA, regardless of whether TET was positive or negative, the QT/RR curves showed clear separation between exercise and recovery phases, which represented an obvious QT hysteresis phenomenon and indicated that myocardial ischemia may increase QT hysteresis indices (Figures 2A,B). Subjects who had negative SCA and TET positive or negative had QT/RR curves that were slightly higher or no different during exercise than during recovery, which indicated a weak QT hysteresis phenomenon (Figures 2C,D).

Figure 2.

Representative QT/RR curves vs. selective coronary angiography (SCA) and treadmill exercise test (TET) results. (A) Both TET and SCA positive; (B) TET negative and SCA positive; (C) TET positive and SCA negative; (D) both TET and SCA negative. The QT hysteresis area and index correlated well with the SCA results. For the positive SCA group, QT hysteresis is noticeable. For the negative SCA negative group, QT hysteresis is minimal. QTe, red; QTp, green.

Efficacy of QT Hysteresis Indices in Diagnosis of CAD

ROC curves using QT hysteresis indices, DTS, TET criteria and QT hysteresis indices in combination with TET criteria are shown in Figure 3. The diagnostic accuracy of using either QTe hysteresis index (AUC=0.905), QTp hysteresis index (AUC=0.856), combination of QTp hysteresis index with TET criteria (AUC=0.884) or the combination of QTe hysteresis index with TET criteria (AUC=0.904) was significantly higher compared to that using DTS (AUC=0.843) or conventional TET criteria (AUC=0.688; Table 2).

Figure 3.

Receiver operating characteristic curves. (A) QTp hysteresis index; (B) QTe hysteresis index; (C) Duke treadmill score (DTS); (D) combination of QTp hysteresis index and conventional treadmill exercise test (TET) criteria; (E) combination of QTe hysteresis index and conventional TET criteria; (F) conventional TET criteria. AUC, area under the curve; CI, confidence interval.

Table 2. Diagnostic Accuracy in CAD

	Cut-off	Sensitivity	Specificity	PPV	NPV	κ agreement	AUC (95% CI)	P value
QTp hysteresis index	11 ms	0.779	0.852	0.870	0.754	0.623	0.856**,† (0.793–0.919)	<0.001
QTe hysteresis index	18 ms	0.740	0.967	0.966	0.747	0.855	0.905**,†† (0.851–0.959)	<0.001
DTS	−3.25	0.673	0.956	0.700	0.708	0.609	0.843* (0.764–0.923)	<0.001
Conventional TET criteria	–	0.818	0.558	0.949	0.705	0.384	0.688 (0.596–0.799)	<0.001
QTp hysteresis index+TET	–	0.766	0.885	0.937	0.750	0.639	0.884**,† (0.828–0.941)	<0.001
QTe hysteresis index+TET	–	0.831	0.885	0.901	0.806	0.709	0.904**,†† (0.851–0.956)	<0.001

*P<0.05, **P<0.001 compared to conventional TET criteria; ^†P<0.05, ^††P<0.01 compared to DTS. AUC, area under ROC curve; CI, confidence interval; DTS, Duke treadmill score; NPV, negative predictive value; PPV, positive predictive value; TET, treadmill exercise test. Other abbreviations as in Table 1.

Youden index for each ROC curve in Figure 3 was calculated. According to the maximum Youden index point, the cut-offs for the QTp and QTe hysteresis indices were 11 ms and 18 ms, respectively, and the cut-off of DTS was –3.25. Accordingly, we calculated the sensitivity and specificity of diagnosing CAD using the QTp and QTe hysteresis indices, DTS, conventional TET criteria and 2 combined methods (Table 2). The combination of QTe hysteresis index and conventional TET criteria had the highest sensitivity (0.831) and the highest negative predictive value (0.806). The QTe hysteresis index had the highest specificity (0.967) and the highest positive predictive value (0.966; Table 2).

Using SCA as the gold standard for diagnostic criteria, the consistency of the QTp and QTe hysteresis indices, and of DTS, TET, and the 2 combination methods was evaluated on κ statistics (Table 2). The QTp hysteresis index (κ agreement=0.623) and QTe hysteresis index (κ agreement=0.855) each had a higher consistency than DTS (κ agreement=0.609) or TET (κ agreement=0.384).

On multivariate logistic regression analysis the combination of QTe hysteresis index and TET was an independent predictor of CAD (Table 3).

Table 3. Multivariate Predictors of CAD

Variables	β-value	P-value
TET	−0.959	0.501
RR interval variation range†	0.006	0.399
QTp hysteresis index	0.169	0.382
QTe hysteresis index	−0.099	0.521
QTp hysteresis area‡	0.000	0.617
QTe hysteresis area‡	0.000	0.946
QTp hysteresis index+TET	−0.798	0.866
QTe hysteresis index+TET*	8.852	0.038

*P<0.05. ^†Distance between A and B in Figure 1; ^‡area surrounded by C–E, and F in Figure 1. Abbreviations as in Tables 1,2.

QT Hysteresis Index and SYNTAX Score

QT hysteresis index had a modest correlation with severity of CAD, assessed using SYNTAX score. In the positive SCA group, the mean SYNTAX score was 13.4±7.7 (range, 2–36). There were significant differences in the QTp and QTe hysteresis indices according to SYNTAX score tertile (lowest tertile, <11; intermediate tertile, 11–22; and highest tertile, >22; Figure 4). Furthermore, the QTp and QTe hysteresis indices significantly correlated with SYNTAX score in the positive SCA group (r=0.450, P<0.001 and r=0.591, P<0.001, respectively; Figure 5).

Figure 4.

QT hysteresis indices correlated with severity of coronary artery disease according to SYNTAX score in the positive selective coronary angiography group (P<0.001).

Figure 5.

QTp (A) and QTe (B) hysteresis indices significantly correlated with SYNTAX score in the positive selective coronary angiography group (P<0.001).

Discussion

Main Findings

The phenomenon of QT hysteresis has been studied for many years since Sarma et al first demonstrated the dynamic relationship between the QT and RR intervals.¹ To the best of our knowledge, this is the first clinical study to assess the value of QT hysteresis during exercise and recovery in predicting CAD. Given that the methods for evaluating the magnitude of QT hysteresis have not been unified,^3,26,27 we have proposed a new parameter to reflect the QT hysteresis phenomenon as objectively as possible.

From Figure 1, it is not appropriate to characterize QT hysteresis by comparing the QT intervals at 1 or several RR intervals during exercise with those during recovery, or by simply calculating the area of the loop created by QT/RR curves during exercise and recovery because the area of the loop is largely dependent on the variation range of heart rate. In this study, we quantified QT hysteresis considering both the area of the QT/RR loop and the RR interval variation range. Moreover, given that the end of the T wave is difficult to identify at high heart rate (which may enlarge the deviation) and that at low heart rate the QT/RR curves have a tendency to overlap, we disregarded 10% of the RR interval at both extremes of the RR interval variation range. Using this algorithm, we found that CAD patients had higher QT hysteresis index, which is in consistent with previous studies.³

Given that SCA is widely accepted, TET is more often chosen as the primary screening for subjects with suspected CAD. Thus the sensitivity of TET needs to be improved to screen the patients with atypical clinical manifestations as far as possible, and the loss in specificity can be supplemented by subsequent SCA. For this reason, we chose a relatively loose and simple standard (any one or more main coronary arteries with reduction of lumen diameter ≥50%) for the diagnosis of CAD in this study.

According to the ROC curves, QT hysteresis index can improve the efficacy in predicting CAD, especially QTe hysteresis index and the combination of QTe hysteresis index and conventional TET criteria. QTe hysteresis index had the highest specificity, positive predictive value and κ agreement, and combining the QTe hysteresis index with conventional TET criteria may enhance the sensitivity and negative predictive value of CAD compared to a single criterion. We note that the sensitivity of conventional TET criteria was relatively high (0.818) while the specificity was relatively low (0.558). Meanwhile QT hysteresis indices had a high specificity (QTp hysteresis index, 0.852; QTe hysteresis index, 0.967), complementary to the conventional TET criteria, and in this way improves the power of TET in the screening of CAD. On multivariate logistic regression the combination of QTe hysteresis index and TET was a more prominent predictor of CAD than the other criteria. The reason why QTe hysteresis index is superior to the combination of this index with TET with regard to specificity, positive predictive value, κ agreement, and AUC may be due to the generation of the combined predictors using the logistic regression analysis algorithm. Taking AUC for example, the AUC of the QTe hysteresis index was the highest but the AUC of the conventional TET criteria was the lowest. The power of the former may be weakened by the latter, when combined.

The present data suggest that this method of describing QT hysteresis is reliable and that QT hysteresis index is an appropriate quantitative parameter that reflects the degree of QT hysteresis.

Probable Mechanisms Underlying QT Hysteresis

During TET, the body is characterized by wide-ranging fluctuations of heart rate in an acceleration-deceleration pattern, progressive myocardial ischemia and increased sympathetic tone. The overshoot in respiratory gas variables was also observed during exercise recovery, which has been confirmed as being related to cardiac disease.²⁸ Each of those physiological or pathophysiological alterations may facilitate the QT hysteresis phenomenon. First, the QT hysteresis phenomenon is considered to be tightly associated with the accelerating and decelerating heart rate changes, although QT interval variation may exhibit different characteristics in other conditions. Second, the stenosis of coronary arteries is accompanied by hypoxia and acidosis, namely myocardial ischemia in CAD patients. Continuous ischemia can cause shortening of action potential durations,^29–31 and in this way contributes to QT hysteresis. In this study, we found that the QT hysteresis index is significantly correlated with SYNTAX score, which is an angiographic grading system based on the severity and complexity of coronary artery lesions. A prior study showed that SYNTAX score correlates well with myocardial ischemia as assessed on ^99mTc-sestamibi single-photon emission CT.³² Thus, the significant correlation between QT hysteresis index and SYNTAX score in the positive SCA group suggests that myocardial ischemia may underlie the mechanisms of QT hysteresis. Moreover, the autonomic nervous system may play a key role in differences in repolarization during exercise and recovery.^1,25 A previous study found that rapid pacing in an acceleration-deceleration pattern did not yield the same level of QT hysteresis as the process of exercise and recovery.³³ Given that atrial pacing alone does not evoke an overall increase in cardiac sympathetic tone,³⁴ QT hysteresis is likely to be caused by a higher catecholamine concentration during the early recovery period, or by abnormal activation of ion channels by catecholamines.^35,36 In contrast, another study showed that the hysteresis effect is driven largely by differential parasympathetic inputs during exercise and recovery.³⁷ Thus, the mechanisms underlying QT hysteresis remain to be identified.

We note that the QTe hysteresis index was superior to the QTp hysteresis index in diagnosing CAD. This may be because the terminal component of T waves (the peak of T wave to the end of T wave) is influenced by disease as well as exercise.³⁸ Important abnormalities in the QT interval, including abnormalities in rate dependence, can be missed if the QTp interval is used rather than the QTe interval.

Clinical Rationale for Use of QT Hysteresis Index

The efficacy of TET when used for the diagnosis of CAD remains controversial. The relatively low specificity and κ agreement of the conventional TET ST-segment deviation method has been demonstrated in this study. As we know, the ST segment is a reflection of phase 2 of the myocardial action potential. The influence of myocardial ischemia on various ionic currents (such as the ATP-sensitive potassium current) will lead to heterogeneity of repolarization of the 3 layers of ventricular myocardium and may be reflected as ST-segment deviation.³⁹ In addition, myocardial ischemia will also alter the action potential duration and increase the regional differences in repolarization heterogeneity, which can be reflected in QT interval variations. Therefore, the electrophysiological effect due to myocardial ischemia will be inevitably reflected in the potential associated index (ST-segment deviation)⁴⁰ and interval-associated index (QT interval variation) at the same time.⁴¹

The value of the QTe hysteresis index in predicting CAD is far superior to that of TET, which depends primarily on the degree and duration of ST-segment deviation. The present results indicate that ST-segment deviation and QT interval variation, each of which is related to ventricular repolarization, are complementary to each other for enhancing the accuracy of diagnosing CAD.

Study Limitations

The number of subjects enrolled in the study was limited. The use of SCA results (which are limited to evaluation of the degree of coronary artery stenosis) to define CAD may have led to a conclusion distinct from that had myocardial perfusion been used to define CAD. The assessment of coronary artery hemodynamics, such as fractional flow reserve, was not carried out. We did not collect data on QT hysteresis index after a relief of myocardial ischemia using any intervention. The cellular and ionic mechanisms of QT hysteresis in patients with CAD have not been fully elucidated.

Conclusions

QTe hysteresis index can serve as a specific parameter for CAD screening, especially combined with conventional TET criteria, and can improve the efficacy of TET to diagnose CAD. QT hysteresis index correlated with severity of CAD, assessed using SYNTAX score.

Acknowledgments

Funding: Natural Science Foundation of Hubei Province, China (No. 2012FFB04332); Scientific Research Starting Foundation for Returned Overseas Chinese Scholars, Ministry of Education, China.

Disclosures

No authors have any conflicts of interest.

References

• 1.    Sarma  JS,  Venkataraman  SK,  Samant  DR,  Gadgil  U. Hysteresis in the human RR-QT relationship during exercise and recovery. Pacing Clin Electrophysiol 1987; 10: 485–491.
• 2.    Kannankeril  PJ,  Harris  PA,  Norris  KJ,  Warsy  I,  Smith  PD,  Roden  DM. Rate-independent QT shortening during exercise in healthy subjects: Terminal repolarization does not shorten with exercise. J Cardiovasc Electrophysiol 2008; 19: 1284–1288.
• 3.    Lauer  MS,  Pothier  CE,  Chernyak  YB,  Brunken  R,  Lieber  M,  Apperson-Hansen  C, et al. Exercise-induced QT/R-R-interval hysteresis as a predictor of myocardial ischemia. J Electrocardiol 2006; 39: 315–323.
• 4.    Starobin  JM,  Cascio  WE,  Goldfarb  AH,  Varadarajan  V,  Starobin  AJ,  Danford  CP, et al. Identifying coronary artery flow reduction and ischemia using quasistationary QT/RR-interval hysteresis measurements. J Electrocardiol 2007; 40: 91–96.
• 5.    Wong  JA,  Gula  LJ,  Klein  GJ,  Yee  R,  Skanes  AC,  Krahn  AD. Utility of treadmill testing in identification and genotype prediction in long-QT syndrome. Circ Arrhythm Electrophysiol 2010; 3: 120–125.
• 6.    Hisamatsu  T,  Ohkubo  T,  Miura  K,  Yamamoto  T,  Fujiyoshi  A,  Miyagawa  N, et al. Association between J-point elevation and death from coronary artery disease: 15-year follow up of the NIPPON DATA90. Circ J 2013; 77: 1260–1266.
• 7.    Hisamatsu  T,  Ohkubo  T,  Miura  K,  Horie  M,  Ueshima  H; Nippon Data90 Research Group. Is more aggressive prevention of coronary artery disease required for patients with early repolarization syndrome?: Reply. Circ J 2013; 77: 1643.
• 8.    Barra  SN,  Providência  R. Is more aggressive prevention of coronary artery disease required for patients with early repolarization syndrome? Circ J 2013; 77: 1642.
• 9.    Froelicher  VF,  Lehmann  KG,  Thomas  R,  Goldman  S,  Morrison  D,  Edson  R, et al. The electrocardiographic exercise test in a population with reduced workup bias: Diagnostic performance, computerized interpretation, and multivariable prediction: Veterans Affairs Cooperative Study in Health Services #016 (QUEXTA) Study Group: Quantitative Exercise Testing and Angiography. Ann Intern Med 1998; 128: 965–974.
• 10.    Jeong  HC,  Kim  I,  Park  KH,  Sim  DS,  Hong  YJ,  Kim  JH, et al. New strategy for detection of subclinical coronary atherosclerosis in asymptomatic patients with type 2 diabetes based on cardiac multi-detector computed tomography and treadmill test. Circ J 2014; 78: 671–678.
• 11.    Tajima  A,  Itoh  H,  Osada  N,  Omiya  K,  Maeda  T,  Ohkoshi  N, et al. Oxygen uptake kinetics during and after exercise are useful markers of coronary artery disease in patients with exercise electrocardiography suggesting myocardial ischemia. Circ J 2009; 73: 1864–1870.
• 12.    Doan  AE,  Peterson  DR,  Blaekmon  JR,  Bruce  RA. Myocardial ischemia after maximal exercise in healthy men. Am Heart J 1965; 69: 11–25.
• 13.    Ashley  EA,  Myers  J,  Froelicher  V. Exercise testing in clinical medicine. Lancet 2000; 356: 1592–1597.
• 14.    Fletcher  GF,  Froelicher  VF,  Hartley  LH,  Haskell  WL,  Pollock  ML. Exercise standards. A statement for health professionals from the American Heart Association. Circulation 1990; 82: 2286–2322.
• 15.    Marwick  TH,  Mehta  R,  Arheart  K,  Lauer  MS. Use of exercise echocardiography for prognostic evaluation of patients with known or suspected coronary artery disease. J Am Coll Cardiol 1997; 30: 83–90.
• 16.    Gibbons  RJ,  Balady  GJ,  Beasley  JW,  Bricker  JT,  Duvernoy  WF,  Froelicher  VF, et al. ACC/AHA guidelines for exercise testing: Executive summary: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Exercise Testing). Circulation 1997; 96: 345–354.
• 17.    Mark  DB,  Shaw  L,  Harrell  FE Jr,  Hlatky  MA,  Lee  KL,  Bengtson  JR, et al. Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Engl J Med 1991; 325: 849–853.
• 18.    Shaw  LJ,  Peterson  ED,  Shaw  LK,  Kesler  KL,  DeLong  ER,  Harrell  FE Jr, et al. Use of a prognostic treadmill score in identifying diagnostic coronary disease subgroups. Circulation 1998; 98: 1622–1630.
• 19.    Gulati  M,  Arnsdorf  MF,  Shaw  LJ,  Pandey  DK,  Thisted  RA,  Lauderdale  DS, et al. Prognostic value of the Duke treadmill score in asymptomatic women. Am J Cardiol 2005; 96: 369–375.
• 20.    Oldroyd  KG,  Phadke  KV,  Phillips  R,  Carson  PH,  Clarke  M,  Davis  JA. Cardiac catheterisation by the Judkins technique as an outpatient procedure. BMJ 1989; 298: 875–876.
• 21.    Rubinshtein  R,  Halon  DA,  Gaspar  T,  Schliamser  JE,  Yaniv  N,  Ammar  R, et al. Usefulness of 64-slice multidetector computed tomography in diagnostic triage of patients with chest pain and negative or nondiagnostic exercise treadmill test result. Am J Cardiol 2007; 99: 925–929.
• 22.    Sianos  G,  Morel  MA,  Kappetein  AP,  Morice  MC,  Colombo  A,  Dawkins  K, et al. The SYNTAX score: An angiographic tool grading the complexity of coronary artery disease. EuroIntervention 2005; 1: 219–227.
• 23.    Serruys  PW,  Onuma  Y,  Garg  S,  Sarno  G,  van den Brand  M,  Kappetein  AP, et al. Assessment of the SYNTAX score in the Syntax study. EuroIntervention 2009; 5: 50–56.
• 24.    Pye  M,  Quinn  AC,  Cobbe  SM. QT interval dispersion: A non-invasive marker of susceptibility to arrhythmia in patients with sustained ventricular arrhythmias? Br Heart J 1994; 71: 511–514.
• 25.    Hajian-Tilaki  K. Receiver operating characteristic (ROC) curve analysis for medical diagnostic test evaluation. Caspian J Intern Med 2013; 4: 627–635.
• 26.    Krahn  AD,  Klein  GJ,  Yee  R. Hysteresis of the RT interval with exercise: A new marker for the long-QT syndrome? Circulation 1997; 96: 1551–1556.
• 27.    Krahn  AD,  Yee  R,  Chauhan  V,  Skanes  AC,  Wang  J,  Hegele  RA, et al. Beta blockers normalize QT hysteresis in long QT syndrome. Am Heart J 2002; 143: 528–534.
• 28.    Suzuki  T,  Koike  A,  Nagayama  O,  Sakurada  K,  Tsuneoka  H,  Kato  J, et al. Overshoot phenomena of respiratory gas variables during exercise recovery in cardiac patients. Circ J 2012; 76: 876–883.
• 29.    Shigematsu  S,  Arita  M. Anoxia-induced activation of ATP-sensitive K⁺ channels in guinea pig ventricular cells and its modulation by glycolysis. Cardiovasc Res 1997; 35: 273–282.
• 30.    Friedrich  M,  Benndorf  K,  Schwalb  M,  Hirche  H. Effects of anoxia on K and Ca currents in isolated guinea pig cardiocytes. Pflugers Arch 1990; 416: 207–209.
• 31.    Kleber  AG. Resting membrane potential, extracellular potassium activity, and intracellular sodium activity during acute global ischemia in isolated perfused guinea pig hearts. Circ Res 1983; 52: 442–450.
• 32.    Tanaka  H,  Chikamori  T,  Hida  S,  Igarashi  Y,  Shiba  C,  Usui  Y, et al. Relationship of SYNTAX score to myocardial ischemia as assessed on myocardial perfusion imaging. Circ J 2013; 77: 2772–2777.
• 33.    Rickards  AF,  Norman  J. Relation between QT interval and heart rate: New design of physiologically adaptive cardiac pacemaker. Br Heart J 1981; 45: 56–61.
• 34.    Dominiak  P,  Delius  W,  Grobecker  H. Changes in plasma catecholamine concentrations in patients with coronary artery disease during pacing and physical exercise. Clin Cardiol 1985; 8: 77–81.
• 35.    Magnano  AR,  Holleran  S,  Ramakrishnan  R,  Reiffel  JA,  Bloomfield  DM. Autonomic nervous system influences on QT interval in normal subjects. J Am Coll Cardiol 2002; 39: 1820–1826.
• 36.    Dimsdale  JE,  Hartley  LH,  Guiney  T,  Ruskin  JN,  Greenblatt  D. Postexercise peril. Plasma catecholamines and exercise. JAMA 1984; 251: 630–632.
• 37.    Pelchovitz  DJ,  Ng  J,  Chicos  AB,  Bergner  DW,  Goldberger  JJ. QT-RR hysteresis is caused by differential autonomic states during exercise and recovery. Am J Physiol Heart Circ Physiol 2012; 302: 2567–2573.
• 38.    Davey  PP. QT interval measurement: Q to T_Apex or Q to T_End? J Intern Med 1999; 246: 145–149.
• 39.    Vincent  GM,  Abildskov  JA,  Burgess  MJ. Mechanisms of ischemic ST-segment displacement. Evaluation by direct current recordings. Circulation 1977; 56: 559–566.
• 40.    Kligfield  P,  Lauer  MS. Exercise electrocardiogram testing: Beyond the ST segment. Circulation 2006; 114: 2070–2082.
• 41.    Korantzopoulos  P,  Letsas  KP,  Christogiannis  Z,  Kalantzi  K,  Massis  I,  Milionis  HJ, et al. Exercise-induced repolarization changes in patients with stable coronary artery disease. Am J Cardiol 2011; 107: 37–40.