Background Exercise testing in women is associated with a high incidence of false-positive ECG changes and should be combined with an imaging study. The QT dispersion (QTD), recorded as the difference between maximum and minimum QT intervals on a 12-lead ECG, is sensitive to myocardial ischemia and may improve the accuracy of exercise testing in women.
Methods and Results Exercise ECGs were analyzed in 64 women who had undergone exercise ECG and coronary angiography for clinical indications: 20 patients with normal exercise stress test and nonsignificant (≤50% diameter narrowing of a major epicardial coronary artery) coronary artery disease (CAD) on angiography (true-negative; TN group), 20 patients with positive exercise stress tests (≥1 mm ST-segment depression or reversible perfusion defects) and significant CAD (true-positive; TP group), and 24 patients with positive exercise stress tests but no significant CAD (false-positive; FP group). The exercise QTD was 45±15 ms in TN, 80±23 ms in TP (P<.0001 versus TP), and 41±14 ms in FP (P=NS versus TN and <.0001 versus TP) groups. A stress QTD of >60 ms had a sensitivity of 70% and specificity of 95% for the diagnosis of significant CAD compared with 55% (P<.05) and 63% (P<.01), respectively, for ≥1 mm ST-segment depression during stress. When QTD of >60 ms was added to ST-segment depression as a condition for positive test, the specificity increased to 100%.
Conclusions Exercise QTD is an easily measurable ECG variable that significantly increases the accuracy of exercise testing in women.
The diagnosis of CAD in women with chest pain is a difficult problem in clinical medicine. Exercise testing is considered to be of limited value in the detection of CAD in the female population due to excessive false-positive results.1 The false-positive rate in women depends on disease prevalence and ranges from 25% in patients with typical angina to 50% in patients with atypical chest pain.2 3 4 5 Cummings et al4 reported ST-segment changes in 25% to 66% of asymptomatic women (depending on age) confounding the interpretation of routine treadmill exercise testing. Several mechanisms have been suggested to explain the high false-positive rate of ST depression in women, including estrogen and resting ST-T abnormalities.6 Because estrogen has a chemical structure similar to digitalis, it has been suggested that it may be partially responsible for the high prevalence of false-positive test results in women.
The QT interval is a measure of the duration of ventricular repolarization and is sensitive to myocardial ischemia. The interlead variation in the 12-lead ECG referred to as QTD reflects heterogeneity in the duration of myocardial repolarization and is increased in patients with ischemia and myocardial infarction.7 It has been reported that successful thrombolysis is associated with less QTD in post–acute myocardial infarction patients.8 9 The QTcD has also been used to improve the diagnostic accuracy of the treadmill stress test in men.10 We hypothesized that exercise-induced myocardial ischemia could change QT interval regionally in the area of ischemia and give rise to an increase in QTD in the 12-lead ECG, and this can improve the diagnostic accuracy of exercise stress testing in women.
We studied 64 women who underwent exercise stress testing and coronary angiography for clinical indications and were not receiving any antiarrhythmic agents. Thirty-one patients had sestamibi perfusion scanning. The patient characteristics are summarized in Table 1⇓ and in “Methods.”
Exercise Stress Test Protocol
All patients underwent symptom-limited treadmill stress tests with Ramp I or II, Bruce, or modified Bruce protocols. Patients receiving digoxin and those with resting ST changes also had sestamibi perfusion scans. All the medications that patients were receiving were continued through the treadmill testing. Patient symptoms, peak heart rate, blood pressure, and any ECG changes were noted. The exercise test was considered positive if there was >1-mm ST horizontal segment depression or 1-mm ST depression at 80 ms after the J point on the exercise ECG or a reversible perfusion defect on the sestamibi scan.
All patients underwent coronary angiography for clinical reasons. The coronary angiograms were reviewed, and significant CAD was defined as ≥50% luminal diameter narrowing of a major epicardial artery in any projection.
Based on the results of exercise testing and angiography, the patients were divided into the following three groups:
(1) The TN group consisted of 20 women who had a negative exercise stress test and nonsignificant CAD based on coronary angiography. The mean±SD age of the group was 54±12 years. Of the 20 subjects, 4 were taking a β-blocker, 6 were taking a calcium channel blocker, 2 were taking a nitrate, and 1 was taking digoxin (Table 1⇑).
(2) The TP group consisted of 20 patients with a mean±SD age of 66±10 years. These patients had positive exercise ECG or exercise-induced reversible perfusion defect on the sestamibi scan and an abnormal coronary angiogram. Fourteen patients in this group had a planar sestamibi scan, and all had reversible perfusion defects. Of 20 patients in the TP group, 2 were taking a β-blocker, 8 were taking a calcium channel blocker, 5 were taking a nitrate, and 2 were taking digoxin. Six of these patients had prior myocardial infarctions.
(3) The FP group consisted of 24 patients who had a positive treadmill test with nonsignificant epicardial CAD. Twelve of 14 in this group who had a sestamibi scan had reversible perfusion defects. The mean±SD age was 56±11 years. Of the 24 subjects in the FP group, 6 were taking a β-blocker, 6 were taking a calcium channel blocker, 5 were taking a nitrate, and none were taking digoxin.
QT Interval and QTD Measurements
The QT interval was measured in as many limb and precordial leads as possible (minimum of 9 and mean of 10.4 leads) in the 12-lead ECG (from the onset of Q wave to the end of T wave) at baseline and peak exercise with 2.5-fold magnification of the ECG through the use of a calibrated magnifying glass. When the T wave was fused with the U or P wave, a straight line was drawn tangential to the downstroke of the T wave, and the intersection of the latter with the baseline was taken as the end of the T wave. The PR segment was taken as the baseline to obviate the difficulty in identifying the end of the T wave in the presence of ST-segment depression (Fig 1⇓). The Q peak T interval was measured in a similar fashion from the onset of Q to the peak of T (if T wave was biphasic, the second peak was used) in a similar number of leads at baseline and peak exercise. The QT interval was corrected for heart rate using Bazett’s equation (QTc=QT/square root of RR interval in seconds).11 QTcD and Q peak Tc dispersion were measured as the difference between the maximum and the minimum QTc or Q peak Tc dispersion, respectively, for a given heart rate. The ΔQTD was defined as the difference between peak exercise and resting QTD, and in a similar fashion ΔQTcD was calculated. The QT interval measurements were performed while blinded to clinical, scintigraphic, and angiographic data.
Data are given as mean±SD values. Differences among the three groups (NL versus FP, TP versus FP, and NL versus TP) were analyzed by one-way ANOVA and/or χ2 test (SPSS for Windows version 6.1). A value of P<.05 was considered significant. Multiple regression analysis was used to arrive at the independent determinants of QTD. The reproducibility of the QTD measurements was tested in 20 subjects at both rest and peak exercise by two independent blinded observers. Agreement between two observers was verified using the Bland-Altman method.12
Reproducibility of QTD Measurements
The mean±SD difference between two independent observers in the measurements of resting QT dispersion (n=20) without ignoring the signs was −3.5±8.8 ms with a correlation coefficient between these independent measurements of .94. During exercise, the mean±SD difference between observations was 3.5±10.4 ms with a correlation coefficient between observations of .90. The mean of the absolute differences of measurements by the two observers was 8 ms for the resting ECGs and 9.5 ms for the exercise ECGs (Fig 2⇓).
Exercise Stress Test Data
As shown in Table 2⇓, the TN and FP groups achieved a higher heart rate than the TP group. The peak systolic blood pressure and duration of exertion, however, were similar. The TP and FP groups had a higher incidence of exercise-induced chest pain than the TN group. The TN group by definition did not have any ST-segment depression with exercise. The TP group had 1.02±0.88-mm exercise-induced ST-segment depression in 2.8±2.6 leads. The FP group had 0.98±0.76-mm ST-segment depression in 2.7±2.3 leads.
QTD and QTcD at Rest
QTD at rest was 72±27 ms in the TP group, 50±22 ms in the FP group, and 48±23 ms in the TN group, with a significant difference between TP and the other two groups (Table 3⇓ and Fig 3A⇓). QTcD at rest ranged from 23 to 174 ms (mean, 83±32 ms) in the TP group, from 13 to 126 ms (mean, 56±25) in the FP group, and from 24 to 134 ms (mean, 51±21) in the TN group, with a significant difference between TP and the other two groups (Fig 3B⇓).
Q Peak T Dispersion and Q Peak Tc Dispersion at Rest
Q peak T dispersion ranged from 20 to 120 ms (mean, 54±27) in the TP group, from 20 to 100 ms (mean, 47±19) in the FP group, and from 20 to 80 ms (mean, 43±17) in the TN group, with no significant difference between the three groups. The Q peak T dispersion corrected for heart rate was 62±31, 51±21, and 46±18 ms, respectively, for the TP, FP, and TN groups, with a significant difference (P<.008) between the TN and TP groups.
QTD and QTcD at Peak Exercise
QTD at peak exercise was 80±23 ms in the TP group, 41±14 ms in the FP group, and 45±15 ms in the TN group, with a significant difference between TP and the other two groups (P<.0001). No significant difference was found between the TN and FP groups (Table 2⇑ and Fig 4A⇓). QTcD at peak exercise ranged from 48 to 159 ms (mean, 109±33 ms) in the TP group, from 28 to 95 ms (mean, 61±18 ms) in the FP group (P<.0001), and from 30 to 120 ms (mean, 67±21 ms) in the TN group. There also was a significant difference between the TP and TN groups (P<.0001). There were no significant differences between FP and NL groups (Fig 4B⇓). Table 3⇑ summarizes the relation of resting and exercise QTD to ST-segment depression, perfusion defects, and angiographic findings in individual patients.
Q Peak T Dispersion and Q Peak Tc Dispersion at Peak Exercise
Q peak T dispersion at peak exercise ranged from 30 to 120 ms (mean, 64±26 ms) in the TP group, from 20 to 70 ms (mean, 43±11 ms) in the FP group, and from 10 to 70 ms (mean, 40±16 ms) in the TN group, with a significant difference between the TP group and the other two groups. The Q peak T dispersion corrected for heart rate at peak exercise was 86±35, 63±15, and 61±27 ms, respectively, for the TP, FP, and TN groups, with a significant difference (P<.014) between the TN and TP groups and a significant difference between TP and FP (P=.01). There was no significant difference between TN and FP groups.
ΔQTD and ΔQTcD
ΔQTD was 7±26 ms in the TP group, −8±20 ms in the FP group, and −1±20 ms in the TN group, with a significant difference only between the FP and TP groups. ΔQTcD ranged from −30 to 74 ms (mean, 8±31 ms) in the TN group, from −55 to 89 ms (mean, 25±35 ms) in the TP group, and from −62 to 60 ms (mean, 5±26 ms) in the FP group. There was no significant difference between the FP and TN groups and the TP and TN groups; there was a significant difference between the FP and TP groups (P=.05).
Diagnostic Value of QTD and QTcD
The diagnostic value of exercise QTD and QTcD alone and in combination with ST-segment deviation is summarized in Table 4⇓. An exercise QTD of >60 ms had a sensitivity of 70% and specificity of 95% for the diagnosis of CAD compared with 55% (P<.05) and 63% (P<.01), respectively, for ≥1-mm ST-segment depression. When a QTD of >60 ms was added to ST-segment depression as a condition for a positive test, the specificity increased to 100%. It is interesting to note that in this study population, reversible sestamibi defect was extremely sensitive but not specific for the diagnosis of CAD. An exercise QTcD of >70 ms significantly increased the sensitivity of exercise testing for CAD to 85%, and its combination with ST-segment depression had a specificity of 100%.
QTD and QTcD During Exercise
The results of this study indicate that QTD and QTcD are increased during exercise-induced myocardial ischemia in women with significant CAD. The TP group had a wider baseline QTD, which increased further with exercise. There has been no previous report of this observation in women. In addition, the Q peak T dispersion was wider in the TP group than in the TN and FP groups, although the difference was not significant between FP and TP when corrected for heart rate. QTD has been reported to be increased during acute myocardial infarction.7 8 9 Dibs et al13 showed QT peak dispersion to be a useful adjunct in the diagnosis of CAD. Although in our study Q peak T was greater for the TP group than for the other two groups, the separation was not as striking as for QTD or QTcD. Patients with an acute anterior and inferior myocardial infarction have a prolongation of QTD (70±30 and 73±32 ms, respectively) compared with control subjects (46±18 ms).3 These findings provide support for the hypothesis that QTD may reflect regional differences in ventricular recovery time or repolarization, which is very sensitive to ischemia.
Diagnostic Value of Exercise QTD and QTcD
The values of QTD and QTcD were assessed in terms of their potential to improve the accuracy of interpretation of stress ECG in women. The diagnostic accuracy of ST-segment depression of ≥1 mm alone has a sensitivity of 55% compared with 85% for QTcD of >70 ms and 70% for QTD of >60 ms. If the test was called positive when either ST depression or QTD of >60 ms was found, abnormal sensitivity rose to 85% with a drop of specificity to 59%. When both ST-segment depression and QTD prolongation were needed for the test to be called positive, the specificity increased to 100%. Exercise-induced QTD >60 ms had a positive predictive value of 87% for the prediction of significant CAD. Exercise-induced QTD of ≤60 ms identified 23 of the 24 patients with FP treadmill or sestamibi results. Again, QTcD of >70 ms had a specificity of 74% for the detection of CAD with a sensitivity of 85%. Based on these results, it may be stated that if the goal of exercise testing in a study population is to detect the presence of any CAD, then one should use any abnormality in ST-segment deviation or an increase in QTD or QTcD as a potential marker of CAD. If the goal is to reduce false-positive results, then QTD or QTcD should be used as an additional condition for a positive test.
Reproducibility of QTD Measurement
The QTD measurement was highly reproducible at both baseline and exercise in this study. Apparently better reproducibility of QTD measurement in this study compared with the results of Kautzner et al14 is probably due to wider QTD encountered in this population. Certainly, for a given level of error, error as a percentage of the actual measurement is smaller in subjects with greater QTD compared with normal subjects who have a narrower QTD. Kautzner et al tested the long-term reproducibility of QTD using ECGs obtained at two different time periods in 28 healthy volunteers14 and found a variability of 25% to 30% in its measurement, which translates to ≈10 ms in normal individuals who have a narrow QTD. The mean QTD in their group was 37 ms for men and 28 ms for women.15 Despite a similar absolute variability, the percent variability is less with wider QT dispersion, as in our study group. In addition, Kautzner et al14 tested the reproducibility in tracings obtained a mean of 8 days apart; a change in physiological state in the interim might have changed the actual QTD. Acceptable reproducibility of QTD during exercise makes this a useful marker of exercise-induced ischemia in women, potentially paralleling the value of echocardiography and perfusion scintigraphy.
The baseline QTD in patients with CAD was longer than that in the other groups, although not as dramatic during exercise. This may be due to the effect of coronary disease on regional repolarization, which may be more sensitive to ischemia than other myocardial functions. Abnormalities of regional diastolic functions have been observed in patients with one-vessel CAD and normal resting systolic wall motion, and this is normalized by coronary angioplasty.16 Therefore, the finding of prolongation of resting QTD in patients is interesting but not totally surprising, and presence of prior myocardial infarction might have contributed to some of this abnormality.
The specificity of sestamibi scan in this study was lower than that given in the literature for men. There are scant data on the specificity of exercise sestamibi in women. Amanullah et al17 reported a specificity of 78% for adenosine sestamibi scan in women. The high incidence of false-positive sestamibi scan in our patients may be due to the possibility that many of these patients might have had syndrome X—most of them were undergoing evaluation for chest pain. The presence of normal epicardial arteries, especially in this population, cannot rule out syndrome X or small vessel disease.
One of the limitations of this study was the limited size of the study population. It is also possible that the measured QT interval in a given lead may depend on the T-wave vector, especially the terminal forces. However, we did not measure QT intervals in leads in which the T wave was flat or its termination was unclear. In addition, the same errors in QT interval measurements would have affected QTD calculations in all three groups. Although the QTD measurement was made with the observer blinded to the sestamibi and angiographic data, it was not possible to blind the observer to ST-segment depression, and this factor can potentially introduce a bias in its measurement. At very fast heart rates, T and P waves may be fused; the QT interval may sometimes be difficult to measure, and the T wave needs to be extrapolated to the baseline. Also, in such a situation, QT peak T dispersion may be helpful.13
This preliminary study indicates that the measurement of QTD or QTcD in women is a useful adjunct in the interpretation of exercise ECG and may markedly improve the accuracy of exercise ECG in women.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|QTc||=||corrected QT interval|
|Tc||=||corrected T interval|
- Received October 24, 1996.
- Revision received February 6, 1997.
- Accepted February 20, 1997.
- Copyright © 1997 by American Heart Association
Ellestad MH, Savitz S, Bergdall D, Teske J. The false-positive stress test: multivariate analysis of 215 subjects with hemodynamic, angiographic and clinical data. Am J Cardiol. 1977;40:681-685.
Detry JMR, Kapita BM, Cosyns J, Sottiaux B, Brasseur LA, Rousseau MF. Diagnostic value of history and maximal exercise electrocardiography in men and women suspected of coronary artery disease. Circulation. 1977;56:756-761.
Weiner DA, Ryan TJ, McCabe CH, Kenedy JW, Schloss M, Tristani F, Chaitman BR, Fisher LD. Exercise stress testing correlation among history of angina, ST-segment response and prevalence of coronary artery disease in the Coronary Artery Surgery Study (CASS). N Engl J Med. 1979;301:230-235.
Cummings GR, Dufresne C, Kich L, Sann J. Exercise electrocardiogram patterns in normal women. Br Heart J. 1973;35:1055-1061.
Guiteras P, Chaitman BR, Waters DD, Bourassa MG, Scholl JM, Ferguson RJ, Wagniart P. Diagnostic accuracy of exercise ECG lead systems in clinical subsets of women. Circulation. 1982;65:1465-1474.
Clark PI, Glasser SP, Lyman GH, Krug-fite J, Root A. Relation of results of exercise stress test in young women to phases of menstrual cycle. Am J Cardiol. 1988;61:197-199.
Higham PD, Furniss SS, Campbell RW. QT dispersion and components of the QT interval in ischaemia and infarction. Br Heart J. 1995;73:32-36.
van de Loo A, Arendts W, Hohnloser SH. Variability of QT dispersion measurements in the surface electrocardiogram in patients with acute myocardial infarction and in normal subjects. Am J Cardiol. 1994;74:1113-1118.
Moreno FL, Villanueva T, Karagounis LA, Anderson JL. Reduction in QT interval dispersion by successful thrombolytic therapy in acute myocardial infarction. Circulation. 1994;90:94-100.
Stoletniy LN, Pai RG. Usefulness of QTc dispersion in interpreting exercise electrocardiograms. Am Heart J. 1995;130:918-921.
Bazett HC. An analysis of the time-relations of electrocardiograms. Heart. 1920;7:353-370.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307-310.
Dibs SR, Childs DB, Tenaglia AN. QT peak dispersion: a potential adjunct in noninvasive diagnosis of coronary artery disease. Circulation. 1995;92(suppl I):I-540. Abstract.
Kautzner J, Yi G, Camm AJ, Malik M. Short and long term reproducibility of QT, QTc and QT dispersion measurement in healthy subjects. J Am Coll Cardiol. 1994;23:178A. Abstract.
Fei L, Statters DJ, Camm AJ. QT-interval dispersion on 12-lead electrocardiogram in normal subjects: its reproducibility and relation to the T wave. Am Heart J. 1994;127:1654-1655.
Bonow RO, Vitale DF, Bacharach SL, Frederick TM, Kent KM, Green MV. Asynchronous left ventricular regional function and impaired global diastolic filling in patients with coronary artery disease: reversal after coronary angioplasty. Circulation. 1985;71:297-307.
Amanullah AM, Kiat H, Friedman JD, Berman DS. Adenosine technetium-99m sestamibi myocardial perfusion SPECT in women: diagnostic efficacy in detection of coronary artery disease. J Am Coll Cardiol. 1996;27:803-809.