- Email: [email protected]
Newer QT Correction Formulae to Correct QT for Heart Rate Changes During Exercise
CLINICAL INVESTIGATION
Newer QT Correction Formulae to Correct QT for Heart Rate Changes During Exercise Simon W. Rabkin, MD and XinBo (Justin) Cheng, BASc ABSTRACT Background: The QT interval is a marker for drug-induced cardiac toxicity, electrolyte abnormalities and genetic mutations with a high risk of sudden death. Aim: The objective was to determine the optimal QT-heart rate correction when heart rate is increased. Materials and Methods: A total of 40 persons had QT interval measured before at the end of each stage of a Bruce protocol. Currently used heart rate correction formulae (QTc) were compared to recently proposed QTc formulae derived from large population studies. Results: Comparing the data at each stage of exercise found that QTc using the Bazett formula (QTcBZT) increased with exercise while the QTc proposed by Fridericia (QTcFRD) and by Framingham (QTcFRM) decreased with exercise. In contrast QTc proposed by Dmitrienko (QTcDMT) and Rautaharju (QTcRTHa) were relatively constant despite the increase in heart rate during exercise, whereas QTc proposed by Hodges (QTcHDG) was more variable. With exercise, the differences between QTcBZT or QTcFRD and the other correction formulae became greater and highly significant. Next, the slope of QTc or RR regression was calculated for each individual during the exercise test. The rank order of the slopes (from the smallest to largest absolute value) was QTcRTHa, QTcDMT, QTcBZT, QTcHDG, QTcFRD and QTcFRM. Furthermore the slope of the QT/heart rate relationship was significantly (P o 0.0001) different between the older formulae proposed by Bazett or Fridericia compared to the newer formulae QTcDMT or QTcRTHa. Conclusion: The 2 newer QT-heart rate correction formula should be used when evaluating QT interval at faster heart rates especially those associated with exercise. Key Indexing Terms: QT interval; Heart rate; Exercise. [Am J Med Sci 2016;351(2):133–139.]
INTRODUCTION
A
ssessment of the QT interval on a standard 12-lead electrocardiogram (ECG) is of value in the recognition of a number of conditions.1-3 Because of the effect of heart rate to alter the QT interval, a critical part of the use of this measurement is the adjustment of the QT interval for heart rate. The most widely known approaches to adjust the QT interval for heart rate have been based largely on studies of the resting ECG. The QT interval during physical activity has received comparatively little attention. Yet the ECG is recorded during physical activity and exercise-induced changes in the QT interval have been proposed to be useful in the diagnosis of inherited abnormalities of the long QT syndromes as well as assessment of risk for future cardiac events.4-6 Exercise-based algorithm have been developed for prediction of inherited QT genetics in relatives of individuals with long QT syndromes.7 These efforts underscore the need to have accurate QT formulae to account for heart rate changes during exercise. There are many different QT-heart rate correction (QTc) formulae that attempt to “correct” or adjust for the effect of heart rate on QT interval.8 The most popular
correction formula was proposed by Bazett (QTcBZT),9 and is also frequently used to adjust for the heart rate increases with exercise. The data on the response to exercise using this equation are variable as studies have reported that QTcBZT increases,10 does not change11 or both increases and decreases12 with exercise. Some investigators contend that QTcBZT has too many limitations over the course of exercise to be of utility.12 Yet studies employ QTcBZT to adjust for exercise-induced heart rate changes to identify individuals with genetic predispositions to inherited QT prolongation.5-7 Some investigators have proposed new equations to evaluate individual at rest and at exercise.13 There have been previous attempts to compare QT correction formulae in exercise but these have used older QTc formulae such as those proposed by Bazett and Fridericia (QTcFRD), almost 100 years ago.14-16 Since then, formulae based on larger sample sizes have been proposed. Recent data from very large populations have been used to construct new formulae to adjust the QT interval for heart rate17,18 but these have not been tested in an exercise setting. The objective of this study was to evaluate QT-heart rate correction formulae during exercise specifically to
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
133
Rabkin and Cheng
compare the older and most frequently used Bazett and Fridericia formulae with the newer ones to determine an optimal formula.
MATERIALS AND METHODS Individuals whose baseline ECG showed sinus rhythm, with no interventricular conduction delay (or bundle branch block) or significant ST-T changes (at rest or on exercise) including U waves or T2 complexes that would render QT interval measurement difficult, were included in the study. A 12-lead ECGs recorded at a paper speed of 25 mm/s, before and during exercise testing, following the standard Bruce treadmill protocol. QT measurements were made at specific time points standing before exercise and near completion of each stage of the exercise protocol. The heart rate and QT interval were measured manually from the ECG printout. For each RR interval, a caliper was used to measure the distance from 1 peak of the QRS complex to the next peak. Each recorded RR value was the mean of 2 consecutive RR intervals. For each QT interval, a caliper was used to measure the distance from the start of the QRS complex to the end of the T wave. The end of the T wave was defined as the intersection of a horizontal line representing the baseline of the ECG to the tangent of the downward curve of the T wave at its steepest point. The baseline of the ECG is typically defined at the level of the PR interval. In some cases, the PR interval shortens with exercise making the identification of the level of the PR segment more challenging. In those situations, the beginning of the QRS complex of 2 consecutive QRS complexes was used to define baseline. Each pair of corresponding RR and QT interval was taken from the same complex. The distances measured were in 0.50 mm increments, with an uncertainty of ⫾ 0.25 mm. As 2 readings were taken from each lead, the average of 2 was used in all analysis, The QT interval was almost always measured from lead II. When there was excessive baseline artifact (noise) in lead II, measurements were made from lead I or III. One-half of the measurements were made by one individual (X.C.) and one-half were made by the other person (S.W.R.) to minimize interobserver variability. Of the many formulae that incorporate a heart rate correction for the QT interval, 3 older formulae proposed by Bazett (QTcBZT),9 Fridericia (QTcFRD)14 and Hodges (QTcHDG)19 were compared to 3 newer formulae proposed by Framingham (QTcFRM),20 Dmitrienko et al17 (QTcDMT) and 2 by Rautaharju et al18 (QTcRTHa and QTcRTHb). Rautaharju et al18 proposed 2 formulae (QTcRTHa and QTcRTHb) but only QTcRTHa was used because it was the one without an age or sex correction factor. Several other equations could not be used because of lack of provision of appropriate constants.13 Some formulae were not included because they were equations that were similar to others being used.12 Some other formulae required 2 different constants for
134
men and 2 different ones for women21,22 or require different constants for different heart rates. Data Analysis The data are presented as mean ⫾1 standard deviation. Statistical testing used either analysis of variance Kruskal-Wallis test (KW) or Student t test for paired values. A linear regression model was used to calculate the slope of QTc versus heart rate relationship. The data from each individual was used to calculate the linear relationship and the slope of the relationship was calculated (GraphPad Prism).
RESULTS A total of 40 persons (2 groups of 20 consecutive persons) who fulfilled the ECG entry criteria had their exercise stress tests examined (Table 1). Heart rate in the standing position before commencement of exercise was 74.7 ⫾ 16.9 bpm and there was a significant (KW ¼ 95.4; P o 0.0001) increase in heart rate with exercise that was apparent at each stage of exercise (Figure 1). The (uncorrected) QT interval decreased significantly (KW ¼ 78.8, P o 0.0001) with exercise (Figure 1). The QT intervals adjusted for heart rate were different depending on the QTc formula. QTcBZT increased with exercise while QTcFRD and QTcFRM decreased with exercise. In contrast QTcDMT and QTcRTHa were relatively stable during exercise while QTcHDG was more variable (Figure 2). At rest (standing) before exercise, heart rate was 74.7 ⫾ 16.9 bpm and comparisons of 6 QTc formulae found that the formulae were similar (KW ¼ 6.1, P ¼ 0.30). At the end of stage 1, the heart rate was 97.5 ⫾ 18.6 bpm and there were a significant differences (KS ¼ 27.9; P o 0.0001) between formulae. At the completion of Stage 2, the mean heart rate was 113.6 ⫾ 15.5 bpm and there were significant differences between the correction formulae (KW ¼ 67.2; P o 0.0001). At the completion of Stage 3, the mean heart rate was 133.8 ⫾ 16.2 bpm and there were significant TABLE 1. shows the characteristics of the subjects in the study. Age (years) mean ⫾ SD
54.7 ⫾ 15.1
Sex (M) Risk factors (%) Hypertension Diabetes mellitus Current or past smoker Hyperlipidemia Positive family history Medications (%) Beta blocker (BB) Calcium channel blocker (CCB) Neither BB or CCB ACE inhibitor Diuretics
60% 42% 3% 15% 20% 35% 17% 5% 78% 22% 3%
ACE, angiotensin-converting enzyme; SD, standard deviation.
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
FIGURE 1. This shows the change in heart rate at the completion of each stage of exercise (upper panel) and the corresponding QT interval (lower panel).
differences between the correction formulae (KW ¼ 76.7; P o 0.0001). Focusing on the exercise ECGs at completion of Stages 1, 2 and 3 showed that the differences between QTcBZT and the other correction were highly significantly different (Table 2). Similarly, QTcFRD was significantly difference from other formulae except for QTcFRM, at the lower stages of exercise. QTcHDG was significantly different from the other formulae except for a few formulae at lower stages of exercise. Because the goal of each formula is to produce QTc values that are not correlated with heart rate, that is, the slope of QTc or RR regression being 0, we calculated the linear slope for each of the corrective equations. The
relationship of the QT interval to heart rate was evaluated for each person at the completion of each stage of exercise. The uncorrected QT interval showed the anticipated shortening of QT with increasing heart rate. Although there were some individual differences in linearity for some persons, overall most individuals show a linear relationship between QT and RR interval with the increase in heart rate with exercise. We calculated the slope of the relationship between QT interval and heart rate for each individual during the exercise test and then calculated the group mean and variances (Figure 3). The rank order of the slopes (from the smallest to largest absolute value) was QTcRTHa, QTcDMT, QTcBZT, QTcHDG, QTcFRD and QTcFRM. In this analysis the
FIGURE 2. This shows mean QTc for the different QT correction equations at each stage of exercise. Abbreviations are outlined in the text.
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
135
Rabkin and Cheng
TABLE 2. The difference between older QT correction formulae QTcBZT, QTcFRD and QTcHDG and the newer QTc.
Stage I QTcFRD QTcHDG QTcFRM QTcDMT
QTcBZT
QTcBZT
QTcFRD
Difference
CI
P
Difference
CI
P
Difference
CI
31.9 21.8 31.6 17.0
27.6–36.1 19.5–24.1 26.1–37.0 14.7–19.3
o0.0001 o0.0001 o0.0001 o0.0001
10.1 0.3 14.9
14.3 to 5.8 2.0 to 1.4 16.8 to 12.9 14.8 to 8.9
o0.0001 0.72 o0.0001
9.8 4.8
0.0003 0.0012
o0.0001
1.8
4.8-14.7 7.6 to 2.0 3.8 to 0.2
26.5 to 16.9 0.53 to 2.6 23.3 to 21.2 21.9 to 16.5
o0.0001 0.188 o0.0001
22.7 2.0
17.5-28.0 1.7 to 5.7
o0.0001 0.28
o0.0001
2.5
0.1 to 5.1
0.06
49.6 to 35.5 0.6-4.8 25.6 to 23.2 32.6 to 26.8
o0.0001 0.012 o0.0001 o0.0001
45.3 18.2 12.9
38.5-52.0 11.8-24.5 8.0-17.7
o0.0001 o0.0001 o0.0001
QTcRTHa Stage II QTcFRD QTcHDG
20.0
18.4–21.6
o0.0001
11.9
42.4 20.7
39.2–45.7 17.9–23.5
o0.0001 0.006
21.7
QTcFRM QTcDMT
43.5 22.7
39.1–47.8 21.0–24.5
o0.0001 o0.0001
1.0 19.7
QTcRTHa
23.2
22.5–23.9
o0.0001
19.2
Stage III QTcFRD QTcHDG QTcFRM QTcDMT QTcRTHa
52.9 10.4 55.6 28.5 23.2
50.2–55.6 4.8–16.0 51.5–59.3 27.0–30.0 22.4–24.1
o0.0001 o0.0001 o0.0001 o0.0001 o0.0001
42.6 2.7 24.4 29.7
QTcHDG P
0.076
Difference is the absolute difference, and P is P value from paired t test.
best formula QTcRTHa, was significantly (P o 0.0001) better than QTcBZT or QTcHDG. There was no significant difference between QTcBZT and QTcHDG. Evaluation of the older formulae compared to the newer formulae showed significant differences between the slope of the QT-RR relationship for QTcRTHa or QTcDMT and the other formulae.
DISCUSSION This study is the first to compare, in the setting of exercise-induced increases in heart rate, newer QT correction formulae with older formulae that are in widespread clinical use. We used 2 different assessment approaches—comparison of group data at specific stages of exercise and an individual approach
FIGURE 3. This shows the slope of the relationship between QT interval and heart rate using the data for each individual from the ECG standing and at the completion of stages 1, 2 and 3. The mean ⫾ SD are shown for each of different QT correction equations. The level of statistical significance is shown for the differences between the most commonly used QT correction formulae and the other QTc formulae. SD, standard deviation.
136
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
calculating the slope of the QT-heart rate across each stage of exercise. Both approaches identified significant discrepancies in the calculation of QTc depending on the QTc adjustment formula. Furthermore, newer approaches were found to be more accurate because they are more independent of heart rate influence on QT interval duration, especially at higher heart rates, compared to older approaches. The QT-heart rate adjustment formulae, used herein, were derived from data on the QT intervals from different persons with different heart rates. This kind of analysis includes between person variability of QT assessment as well as the QT-heart rate relationship. The extent to which person-to-person variability influences the development and validity of these formulae can be questioned. Instead the present study evaluated the change in heart rate in the same individual and its impact on QT interval. Our analysis concluded that the newer QTc approaches are better at the task of correction for heart rate than older formulae. Our data provide a potential explanation for the discrepancies in reported QT changes with exercise.10-12,16 The data suggest that exercise does not change QT interval once it has been appropriately corrected for the increase in heart rate. During exercise, uncorrected QT interval decreased. QTcFRD and QTcFRM blunted the decrease but the reduction in QTc was still evident. In contrast, QTcBZT demonstrated an increase in QTc, which became progressively longer with each duration of exercise and increase in heart rate. These discordances in QTc when applied to the same data suggest that QT changes are not a function of exercise but rather are because of the kind of correction approach used. In contrast, newer formulae such as those proposed by Rautaharja et al18 and Dmitrienko et al17 were relatively independent of heart rate during exercise and are significantly differed from both QTcBZT and QTcFRD. Of the older formulae, QTcHDG performed quite well in the exercise evaluation but was significantly different from the 2 newer formulae especially at the higher heart rates. The exercise, used in our study, involves a treadmillstandardized workload for a fixed duration of time. This protocol provides a 3-minute stable workload during which QT interval may attain a steady state value. The exercise stress test (Bruce) protocol used in this study is commonly used and provides a more vigorous stress and therefore heart rate increase to test QTc formulae compared to other exercise protocols, including bicycle ergometers, to assess the impact of exercise on QTc.15,23 Aytemir et al15 compared the older QT-heart rate corrections specifically QTcBZT, QTcFRD, QTcHDG and QTcFRM; and a nomogram in 21 healthy men and women without apparent cardiovascular disease who exercised on a bicycle ergometer. They reported that QTcFRD was the best correction formula.15 Benatar and Decraene16 evaluated the exercise response in children, most were on a bicycle ergometer, and reported
differences between QTcBZT, QTcFRD, QTc HDG and QTcFRM but did not test each formulae for the degree to which it rendered the QT interval independent of heart rate. We found that the newer formulae that were developed using linear regression analysis were better than the older functions. Our data support the recommendations that linear regression functions rather than the Bazett's formula should be used for QT-rate correction and that the method used for rate correction should be identified in ECG analysis reports.24 We found that QTcBZT was a poor correction formulae. Although the Bazett correction approach (QTcBZT) has been criticized for inaccuracies at higher and lower heart rates using between person data from the resting ECG,20,25,26 it has continued to be used perhaps of the absence of a gold standard for heart rate correction. Exercise stress testing to identify individuals with inherited long QT syndromes as well as exercisebased algorithms, developed for prediction of inherited QT genetics in relatives of LQTS probands have used QTcBZT.5-7 Based on our findings the use of this formulae should be reconsidered. Our study also identifies problems with using QTcFRD to correct for heart rate. Puddu et al27 evaluated the resting ECG of 881 middle-aged men from 1 Italian cohort of the 7 countries study, and suggested that the Fridericia's formula is better suited to fit the data than QTcBZT or other formulae. Our data indicate that QTcFRD retains the potential for bias at faster heart rates based on our findings in the same individual when heart rate is increased. We found that the performance of Framingham formula20 was similar to that of Fridericia approach.14 There have been many attempts to derive a formulae to adjust QT interval to make it independent of heart rate. It was not our objective to create a new formula. Rather we sought to determine which of the developed formulae is most likely to be independent of heart rate. Testing the “goodness of fit” of a model for a data set, often uses approaches such as the Akaike Information (AI) criterion. It includes a factor for the number of parameters and the natural log of the likelihood ratio. In our study, there is only 1 parameter for each of the equations namely the RR interval with different kinds of transformation of the RR interval. With k ¼ 1, AI is a direct reflection of the likelihood ratio which is equivalent to the P values which we present. Similarly the modified AI criterion was not used because it adjusts for the sample size, which in our study is constant. There are several limitations of this study that warrant further discussion. First, although it appears that our sample size is small, relative to the recent large data sets that developed formulae for QTc, the number of individuals in this study exceeded (almost double) many other studies on the effect of exercise on QT interval.11,13,15,28,29 Second, all studies in this field are challenged by the absence of a “gold standard” which is incontestably the standard for independence of QT on
OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
THE AMERICAN JOURNAL
137
Rabkin and Cheng
heart rate and against which all formulae can be compared. In addition, there are only limited approaches to approximate the best formula to determine which formula “best” fits the data compared to another and what constitutes a “best fit.” Thus, in 1 approach, we have to select what should be a good approximation of heart rate independence. We selected a criterion proposed and used by others namely that the slope of the relationship between QT and heart rate should approximately 0. Third, the individual data analysis approach limits the number of points from which to construct a regression line as baseline and at most 3 exercise stages are available. However, conclusions both clinically and in clinical trials can often use similar or fewer ECGs for comparison. In addition, ECG evaluation for inherited QT prolongation uses similar numbers of ECGs during exercise testing.5,7 Fourth, we did not correct for QT or RR hysteresis, ie, the QT interval duration is not only dependent on the simultaneously measured RR intervals but also on the prior RR intervals. Hysteresis, however, is difficult to estimate for a single individual within the Bruce exercise protocol. Equations have been developed to minimize hysteresis but these have used exercise protocols of longer duration, at least 5 minutes at a given work load, and are often followed by a longer rest period before exercise is begun again. Some investigators suggest steady state exercise durations of at least 5 minutes, with multiple ECG recordings are necessary to approximate the impact of hysteresis or the impact of preceding RR intervals on the QT measurement.30 Exercise-induced QT or RR hysteresis is especially evident comparing the resting with the exercise portions of the exercise test as “for a given R-R interval, the QT interval duration is shorter during recovery after exercise than during exercise.”31 Our study deliberately did not examine the recovery stage of exercise because of concerns about the potential impact of QT or RR hysteresis. Regardless within the context of our study different formulae were applied at a standard time at each stage of exercise so that we examine the comparative ability of the different formulae rather than studying the process of QT or RR hysteresis or the attained heart rate or the speed at which the heart rate is reached during exercise. Fifth, heart rate and QT interval were measured from 2 QRS complexes at the completion of each stage of exercise. Although an extended time would have allowed more QT measurements, it would violate the principle of assessing the QT-RR interval relationship at a fixed time and workload. Within the context of our study, however, different formulae were applied at a standard time at each stage of exercise so that we examine the comparative ability of the different formulae as each formula was evaluated under the same conditions. Sixth, newer QT correction formulae have more complex transformations of the RR, which some ECG readers may find to be more challenging for manual calculations. This inconvenience is eliminated by computer applications to permit ease of
138
calculation for these formulae. Lastly, QTc calculation during exercise is mostly used for the diagnosis of long QT syndrome. Further studies are necessary in patients with long QT syndrome (congenital or acquired) to test the value of the newer QTc formulae.
CONCLUSION This study evaluated different approaches to assess the QT interval, a valuable marker for drug-induced cardiac toxicity, electrolyte abnormalities and genetic mutations that carry a high risk of sudden death. The data suggest that 2 new QT-heart rate correction formulae specifically QTcDMT and QTcRTHa should be used preferentially rather than older formulae such as those proposed by Bazett or Fridericia especially when evaluating persons with faster heart rates and in the exercise setting.
REFERENCES 1. Rabkin SW. Aging effects on QT interval: implications for cardiac safety of antipsychotic drugs. J Geriatr Cardiol 2014;11:20–5. 2. Burchell HB. The QT interval historically treated. Pediatr Cardiol 1983;4: 139–48. 3. Roden DM. Drug-induced prolongation of the QT interval. N Engl J Med 2004;350:1013–22. 4. Yi G, Crook R, Guo XH, et al. Exercise-induced changes in the QT interval duration and dispersion in patients with sudden cardiac death after myocardial infarction. Int J Cardiol 1998;63:271–9. 5. Aziz PF, Wieand TS, Ganley J, et al. Genotype- and mutation sitespecific QT adaptation during exercise, recovery, and postural changes in children with long-QT syndrome. Circ Arrhythmia Electrophysiol 2011;4: 867–73. 6. Vincent GM, Jaiswal D, Timothy KW. Effects of exercise on heart rate, QT, QTc and QT/QS2 in the Romano-Ward inherited long QT syndrome. Am J Cardiol 1991;68:498–503. 7. Sy RW, van der Werf C, Chattha IS, et al. Derivation and validation of a simple exercise-based algorithm for prediction of genetic testing in relatives of LQTS probands. Circulation 2011;124:2187–94. 8. Rabkin SW, Cheng XB. Nomenclature, categorization and usage of formulae to adjust QT interval for heart rate. World J Cardiol 2015;7: 315–25. 9. Bazett H. An analysis of the time-relations of electrocardiograms. Heart 1920;7:353–67. 10. Yu PN, Bruce RA, Lovejoy FW, et al. Observations on the change of ventricular systole (QT interval) during exercise. J Clin Invest 1950;29: 279–89. 11. Davidowski TA, Wolf S. The QT interval during reflex cardiovascular adaptation. Circulation 1984;69:22–5. 12. Kligfield P, Lax KG, Okin PM. QTc behavior during treadmill exercise as a function of the underlying QT-heart rate relationship. J Electrocardiol 1995; 28(suppl):206–10. http://dx.doi.org/10.1016/S0022-0736(95)80058-1. 13. Sarma JS, Sarma RJ, Bilitch M, et al. An exponential formula for heart rate dependence of QT interval during exercise and cardiac pacing in humans: reevaluation of Bazett's formula. Am J Cardiol 1984;54:103–8. 14. Fridericia LS. The duration of systole in an electrocardiogram in normal humans and in patients with heart disease. 1920. Ann Noninvasive Electrocardiol 2003;8:343–51. 15. Aytemir K, Maarouf N, Gallagher MM, et al. Comparison of formulae for heart rate correction of QT interval in exercise electrocardiograms. Pacing Clin Electrophysiol 1999;22:1397–401. 16. Benatar A, Decraene T. Comparison of formulae for heart rate correction of QT interval in exercise ECGs from healthy children. Heart 2001;86:199–202. 17. Dmitrienko AA, Sides GD, Winters KJ, et al. Electrocardiogram reference ranges derived from a standardized clinical trial population. Drug Inf J 2005;39:395–405.
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
18. Rautaharju PM, Mason JW, Akiyama T. New age- and sex-specific criteria for QT prolongation based on rate correction formulas that minimize bias at the upper normal limits. Int J Cardiol 2014;174:535–40. 19. Hodges M, Salerno D, Erlien D, et al. correction reviewed: evidence that a linear QT correction for heart rate is better. J Am Coll Cardiol 1983;1(2):694. 20. Sagie A, Larson MG, Goldberg RJ, et al. An improved method for adjusting the QT interval for heart rate (the Framingham Heart Study). Am J Cardiol 1992;70:797–801. 21. Adams W. The normal duration of the electrocardiographic ventricular complex. J Clin Invest 1936;15:335–42. 22. Boudoulas H, Geleris P, Lewis RP, et al. Linear relationship between electrical systole, mechanical systole, and heart rate. Chest 1981;80:613–7. 23. Kligfield P, Lax KG, Okin PM. QT interval-heart rate relation during exercise in normal men and women: definition by linear regression analysis. J Am Coll Cardiol 1996;28:1547–55. 24. Rautaharju PM, Surawicz B, Gettes LS, et al. American Heart Association Electrocardi C on CCAC of CFHRS. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part IV: the ST segment, T and U waves, and the QT interval: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias C. J Am Coll Cardiol 2009;53:982–91. 25. Ahnve S. Correction of the QT interval for heart rate: review of different formulas and the use of Bazett's formula in myocardial infarction. Am Heart J 1985;109:568–74. 26. Manion CV, Whitsett TL, Wilson MF. Applicability of correcting the QT interval for heart rate. Am Heart J 1980;99:678.
27. Puddu PE, Jouve R, Mariotti S, et al. Evaluation of 10 QT prediction formulas in 881 middle-aged men from the seven countries study: emphasis on the cubic root Fridericia's equation. J Electrocardiol 1988;21:219–29. 28. Akhras F, Rickards AF. The relationship between QT interval and heart rate during physiological exercise and pacing. Jpn Heart J 1981;22: 345–51. 29. Arrowood JA, Kline J, Simpson PM, et al. Modulation of the QT interval: effects of graded exercise and reflex cardiovascular stimulation. J Appl Physiol 1993;75:2217–23. 30. Jacquemet V, Dube B, Knight R, et al. Evaluation of a subject-specific transfer-function-based nonlinear QT interval rate-correction method. Physiol Meas 2011;32:619–35. 31. Lauer MS, Pothier CE, Chernyak YB, et al. Exercise-induced QT/R-Rinterval hysteresis as a predictor of myocardial ischemia. J Electrocardiol 2006;39:315–23.
From the Department of Medicine (Cardiology) (SWR, XC), University of British Columbia, Vancouver, BC, Canada. Submitted May 28, 2015; accepted September 1, 2015. The authors have no financial or other conflicts of interest to disclose. Correspondence: Simon W. Rabkin, MD, Department of Medicine, Division of Cardiology, University of British Columbia, Level 9 2775 Laurel St, Vancouver, BC, Canada V5Z 1M9. (E-mail: [email protected]).
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
139
Newer QT Correction Formulae to Correct QT for Heart Rate Changes During Exercise Simon W. Rabkin, MD and XinBo (Justin) Cheng, BASc ABSTRACT Background: The QT interval is a marker for drug-induced cardiac toxicity, electrolyte abnormalities and genetic mutations with a high risk of sudden death. Aim: The objective was to determine the optimal QT-heart rate correction when heart rate is increased. Materials and Methods: A total of 40 persons had QT interval measured before at the end of each stage of a Bruce protocol. Currently used heart rate correction formulae (QTc) were compared to recently proposed QTc formulae derived from large population studies. Results: Comparing the data at each stage of exercise found that QTc using the Bazett formula (QTcBZT) increased with exercise while the QTc proposed by Fridericia (QTcFRD) and by Framingham (QTcFRM) decreased with exercise. In contrast QTc proposed by Dmitrienko (QTcDMT) and Rautaharju (QTcRTHa) were relatively constant despite the increase in heart rate during exercise, whereas QTc proposed by Hodges (QTcHDG) was more variable. With exercise, the differences between QTcBZT or QTcFRD and the other correction formulae became greater and highly significant. Next, the slope of QTc or RR regression was calculated for each individual during the exercise test. The rank order of the slopes (from the smallest to largest absolute value) was QTcRTHa, QTcDMT, QTcBZT, QTcHDG, QTcFRD and QTcFRM. Furthermore the slope of the QT/heart rate relationship was significantly (P o 0.0001) different between the older formulae proposed by Bazett or Fridericia compared to the newer formulae QTcDMT or QTcRTHa. Conclusion: The 2 newer QT-heart rate correction formula should be used when evaluating QT interval at faster heart rates especially those associated with exercise. Key Indexing Terms: QT interval; Heart rate; Exercise. [Am J Med Sci 2016;351(2):133–139.]
INTRODUCTION
A
ssessment of the QT interval on a standard 12-lead electrocardiogram (ECG) is of value in the recognition of a number of conditions.1-3 Because of the effect of heart rate to alter the QT interval, a critical part of the use of this measurement is the adjustment of the QT interval for heart rate. The most widely known approaches to adjust the QT interval for heart rate have been based largely on studies of the resting ECG. The QT interval during physical activity has received comparatively little attention. Yet the ECG is recorded during physical activity and exercise-induced changes in the QT interval have been proposed to be useful in the diagnosis of inherited abnormalities of the long QT syndromes as well as assessment of risk for future cardiac events.4-6 Exercise-based algorithm have been developed for prediction of inherited QT genetics in relatives of individuals with long QT syndromes.7 These efforts underscore the need to have accurate QT formulae to account for heart rate changes during exercise. There are many different QT-heart rate correction (QTc) formulae that attempt to “correct” or adjust for the effect of heart rate on QT interval.8 The most popular
correction formula was proposed by Bazett (QTcBZT),9 and is also frequently used to adjust for the heart rate increases with exercise. The data on the response to exercise using this equation are variable as studies have reported that QTcBZT increases,10 does not change11 or both increases and decreases12 with exercise. Some investigators contend that QTcBZT has too many limitations over the course of exercise to be of utility.12 Yet studies employ QTcBZT to adjust for exercise-induced heart rate changes to identify individuals with genetic predispositions to inherited QT prolongation.5-7 Some investigators have proposed new equations to evaluate individual at rest and at exercise.13 There have been previous attempts to compare QT correction formulae in exercise but these have used older QTc formulae such as those proposed by Bazett and Fridericia (QTcFRD), almost 100 years ago.14-16 Since then, formulae based on larger sample sizes have been proposed. Recent data from very large populations have been used to construct new formulae to adjust the QT interval for heart rate17,18 but these have not been tested in an exercise setting. The objective of this study was to evaluate QT-heart rate correction formulae during exercise specifically to
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
133
Rabkin and Cheng
compare the older and most frequently used Bazett and Fridericia formulae with the newer ones to determine an optimal formula.
MATERIALS AND METHODS Individuals whose baseline ECG showed sinus rhythm, with no interventricular conduction delay (or bundle branch block) or significant ST-T changes (at rest or on exercise) including U waves or T2 complexes that would render QT interval measurement difficult, were included in the study. A 12-lead ECGs recorded at a paper speed of 25 mm/s, before and during exercise testing, following the standard Bruce treadmill protocol. QT measurements were made at specific time points standing before exercise and near completion of each stage of the exercise protocol. The heart rate and QT interval were measured manually from the ECG printout. For each RR interval, a caliper was used to measure the distance from 1 peak of the QRS complex to the next peak. Each recorded RR value was the mean of 2 consecutive RR intervals. For each QT interval, a caliper was used to measure the distance from the start of the QRS complex to the end of the T wave. The end of the T wave was defined as the intersection of a horizontal line representing the baseline of the ECG to the tangent of the downward curve of the T wave at its steepest point. The baseline of the ECG is typically defined at the level of the PR interval. In some cases, the PR interval shortens with exercise making the identification of the level of the PR segment more challenging. In those situations, the beginning of the QRS complex of 2 consecutive QRS complexes was used to define baseline. Each pair of corresponding RR and QT interval was taken from the same complex. The distances measured were in 0.50 mm increments, with an uncertainty of ⫾ 0.25 mm. As 2 readings were taken from each lead, the average of 2 was used in all analysis, The QT interval was almost always measured from lead II. When there was excessive baseline artifact (noise) in lead II, measurements were made from lead I or III. One-half of the measurements were made by one individual (X.C.) and one-half were made by the other person (S.W.R.) to minimize interobserver variability. Of the many formulae that incorporate a heart rate correction for the QT interval, 3 older formulae proposed by Bazett (QTcBZT),9 Fridericia (QTcFRD)14 and Hodges (QTcHDG)19 were compared to 3 newer formulae proposed by Framingham (QTcFRM),20 Dmitrienko et al17 (QTcDMT) and 2 by Rautaharju et al18 (QTcRTHa and QTcRTHb). Rautaharju et al18 proposed 2 formulae (QTcRTHa and QTcRTHb) but only QTcRTHa was used because it was the one without an age or sex correction factor. Several other equations could not be used because of lack of provision of appropriate constants.13 Some formulae were not included because they were equations that were similar to others being used.12 Some other formulae required 2 different constants for
134
men and 2 different ones for women21,22 or require different constants for different heart rates. Data Analysis The data are presented as mean ⫾1 standard deviation. Statistical testing used either analysis of variance Kruskal-Wallis test (KW) or Student t test for paired values. A linear regression model was used to calculate the slope of QTc versus heart rate relationship. The data from each individual was used to calculate the linear relationship and the slope of the relationship was calculated (GraphPad Prism).
RESULTS A total of 40 persons (2 groups of 20 consecutive persons) who fulfilled the ECG entry criteria had their exercise stress tests examined (Table 1). Heart rate in the standing position before commencement of exercise was 74.7 ⫾ 16.9 bpm and there was a significant (KW ¼ 95.4; P o 0.0001) increase in heart rate with exercise that was apparent at each stage of exercise (Figure 1). The (uncorrected) QT interval decreased significantly (KW ¼ 78.8, P o 0.0001) with exercise (Figure 1). The QT intervals adjusted for heart rate were different depending on the QTc formula. QTcBZT increased with exercise while QTcFRD and QTcFRM decreased with exercise. In contrast QTcDMT and QTcRTHa were relatively stable during exercise while QTcHDG was more variable (Figure 2). At rest (standing) before exercise, heart rate was 74.7 ⫾ 16.9 bpm and comparisons of 6 QTc formulae found that the formulae were similar (KW ¼ 6.1, P ¼ 0.30). At the end of stage 1, the heart rate was 97.5 ⫾ 18.6 bpm and there were a significant differences (KS ¼ 27.9; P o 0.0001) between formulae. At the completion of Stage 2, the mean heart rate was 113.6 ⫾ 15.5 bpm and there were significant differences between the correction formulae (KW ¼ 67.2; P o 0.0001). At the completion of Stage 3, the mean heart rate was 133.8 ⫾ 16.2 bpm and there were significant TABLE 1. shows the characteristics of the subjects in the study. Age (years) mean ⫾ SD
54.7 ⫾ 15.1
Sex (M) Risk factors (%) Hypertension Diabetes mellitus Current or past smoker Hyperlipidemia Positive family history Medications (%) Beta blocker (BB) Calcium channel blocker (CCB) Neither BB or CCB ACE inhibitor Diuretics
60% 42% 3% 15% 20% 35% 17% 5% 78% 22% 3%
ACE, angiotensin-converting enzyme; SD, standard deviation.
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
FIGURE 1. This shows the change in heart rate at the completion of each stage of exercise (upper panel) and the corresponding QT interval (lower panel).
differences between the correction formulae (KW ¼ 76.7; P o 0.0001). Focusing on the exercise ECGs at completion of Stages 1, 2 and 3 showed that the differences between QTcBZT and the other correction were highly significantly different (Table 2). Similarly, QTcFRD was significantly difference from other formulae except for QTcFRM, at the lower stages of exercise. QTcHDG was significantly different from the other formulae except for a few formulae at lower stages of exercise. Because the goal of each formula is to produce QTc values that are not correlated with heart rate, that is, the slope of QTc or RR regression being 0, we calculated the linear slope for each of the corrective equations. The
relationship of the QT interval to heart rate was evaluated for each person at the completion of each stage of exercise. The uncorrected QT interval showed the anticipated shortening of QT with increasing heart rate. Although there were some individual differences in linearity for some persons, overall most individuals show a linear relationship between QT and RR interval with the increase in heart rate with exercise. We calculated the slope of the relationship between QT interval and heart rate for each individual during the exercise test and then calculated the group mean and variances (Figure 3). The rank order of the slopes (from the smallest to largest absolute value) was QTcRTHa, QTcDMT, QTcBZT, QTcHDG, QTcFRD and QTcFRM. In this analysis the
FIGURE 2. This shows mean QTc for the different QT correction equations at each stage of exercise. Abbreviations are outlined in the text.
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
135
Rabkin and Cheng
TABLE 2. The difference between older QT correction formulae QTcBZT, QTcFRD and QTcHDG and the newer QTc.
Stage I QTcFRD QTcHDG QTcFRM QTcDMT
QTcBZT
QTcBZT
QTcFRD
Difference
CI
P
Difference
CI
P
Difference
CI
31.9 21.8 31.6 17.0
27.6–36.1 19.5–24.1 26.1–37.0 14.7–19.3
o0.0001 o0.0001 o0.0001 o0.0001
10.1 0.3 14.9
14.3 to 5.8 2.0 to 1.4 16.8 to 12.9 14.8 to 8.9
o0.0001 0.72 o0.0001
9.8 4.8
0.0003 0.0012
o0.0001
1.8
4.8-14.7 7.6 to 2.0 3.8 to 0.2
26.5 to 16.9 0.53 to 2.6 23.3 to 21.2 21.9 to 16.5
o0.0001 0.188 o0.0001
22.7 2.0
17.5-28.0 1.7 to 5.7
o0.0001 0.28
o0.0001
2.5
0.1 to 5.1
0.06
49.6 to 35.5 0.6-4.8 25.6 to 23.2 32.6 to 26.8
o0.0001 0.012 o0.0001 o0.0001
45.3 18.2 12.9
38.5-52.0 11.8-24.5 8.0-17.7
o0.0001 o0.0001 o0.0001
QTcRTHa Stage II QTcFRD QTcHDG
20.0
18.4–21.6
o0.0001
11.9
42.4 20.7
39.2–45.7 17.9–23.5
o0.0001 0.006
21.7
QTcFRM QTcDMT
43.5 22.7
39.1–47.8 21.0–24.5
o0.0001 o0.0001
1.0 19.7
QTcRTHa
23.2
22.5–23.9
o0.0001
19.2
Stage III QTcFRD QTcHDG QTcFRM QTcDMT QTcRTHa
52.9 10.4 55.6 28.5 23.2
50.2–55.6 4.8–16.0 51.5–59.3 27.0–30.0 22.4–24.1
o0.0001 o0.0001 o0.0001 o0.0001 o0.0001
42.6 2.7 24.4 29.7
QTcHDG P
0.076
Difference is the absolute difference, and P is P value from paired t test.
best formula QTcRTHa, was significantly (P o 0.0001) better than QTcBZT or QTcHDG. There was no significant difference between QTcBZT and QTcHDG. Evaluation of the older formulae compared to the newer formulae showed significant differences between the slope of the QT-RR relationship for QTcRTHa or QTcDMT and the other formulae.
DISCUSSION This study is the first to compare, in the setting of exercise-induced increases in heart rate, newer QT correction formulae with older formulae that are in widespread clinical use. We used 2 different assessment approaches—comparison of group data at specific stages of exercise and an individual approach
FIGURE 3. This shows the slope of the relationship between QT interval and heart rate using the data for each individual from the ECG standing and at the completion of stages 1, 2 and 3. The mean ⫾ SD are shown for each of different QT correction equations. The level of statistical significance is shown for the differences between the most commonly used QT correction formulae and the other QTc formulae. SD, standard deviation.
136
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
calculating the slope of the QT-heart rate across each stage of exercise. Both approaches identified significant discrepancies in the calculation of QTc depending on the QTc adjustment formula. Furthermore, newer approaches were found to be more accurate because they are more independent of heart rate influence on QT interval duration, especially at higher heart rates, compared to older approaches. The QT-heart rate adjustment formulae, used herein, were derived from data on the QT intervals from different persons with different heart rates. This kind of analysis includes between person variability of QT assessment as well as the QT-heart rate relationship. The extent to which person-to-person variability influences the development and validity of these formulae can be questioned. Instead the present study evaluated the change in heart rate in the same individual and its impact on QT interval. Our analysis concluded that the newer QTc approaches are better at the task of correction for heart rate than older formulae. Our data provide a potential explanation for the discrepancies in reported QT changes with exercise.10-12,16 The data suggest that exercise does not change QT interval once it has been appropriately corrected for the increase in heart rate. During exercise, uncorrected QT interval decreased. QTcFRD and QTcFRM blunted the decrease but the reduction in QTc was still evident. In contrast, QTcBZT demonstrated an increase in QTc, which became progressively longer with each duration of exercise and increase in heart rate. These discordances in QTc when applied to the same data suggest that QT changes are not a function of exercise but rather are because of the kind of correction approach used. In contrast, newer formulae such as those proposed by Rautaharja et al18 and Dmitrienko et al17 were relatively independent of heart rate during exercise and are significantly differed from both QTcBZT and QTcFRD. Of the older formulae, QTcHDG performed quite well in the exercise evaluation but was significantly different from the 2 newer formulae especially at the higher heart rates. The exercise, used in our study, involves a treadmillstandardized workload for a fixed duration of time. This protocol provides a 3-minute stable workload during which QT interval may attain a steady state value. The exercise stress test (Bruce) protocol used in this study is commonly used and provides a more vigorous stress and therefore heart rate increase to test QTc formulae compared to other exercise protocols, including bicycle ergometers, to assess the impact of exercise on QTc.15,23 Aytemir et al15 compared the older QT-heart rate corrections specifically QTcBZT, QTcFRD, QTcHDG and QTcFRM; and a nomogram in 21 healthy men and women without apparent cardiovascular disease who exercised on a bicycle ergometer. They reported that QTcFRD was the best correction formula.15 Benatar and Decraene16 evaluated the exercise response in children, most were on a bicycle ergometer, and reported
differences between QTcBZT, QTcFRD, QTc HDG and QTcFRM but did not test each formulae for the degree to which it rendered the QT interval independent of heart rate. We found that the newer formulae that were developed using linear regression analysis were better than the older functions. Our data support the recommendations that linear regression functions rather than the Bazett's formula should be used for QT-rate correction and that the method used for rate correction should be identified in ECG analysis reports.24 We found that QTcBZT was a poor correction formulae. Although the Bazett correction approach (QTcBZT) has been criticized for inaccuracies at higher and lower heart rates using between person data from the resting ECG,20,25,26 it has continued to be used perhaps of the absence of a gold standard for heart rate correction. Exercise stress testing to identify individuals with inherited long QT syndromes as well as exercisebased algorithms, developed for prediction of inherited QT genetics in relatives of LQTS probands have used QTcBZT.5-7 Based on our findings the use of this formulae should be reconsidered. Our study also identifies problems with using QTcFRD to correct for heart rate. Puddu et al27 evaluated the resting ECG of 881 middle-aged men from 1 Italian cohort of the 7 countries study, and suggested that the Fridericia's formula is better suited to fit the data than QTcBZT or other formulae. Our data indicate that QTcFRD retains the potential for bias at faster heart rates based on our findings in the same individual when heart rate is increased. We found that the performance of Framingham formula20 was similar to that of Fridericia approach.14 There have been many attempts to derive a formulae to adjust QT interval to make it independent of heart rate. It was not our objective to create a new formula. Rather we sought to determine which of the developed formulae is most likely to be independent of heart rate. Testing the “goodness of fit” of a model for a data set, often uses approaches such as the Akaike Information (AI) criterion. It includes a factor for the number of parameters and the natural log of the likelihood ratio. In our study, there is only 1 parameter for each of the equations namely the RR interval with different kinds of transformation of the RR interval. With k ¼ 1, AI is a direct reflection of the likelihood ratio which is equivalent to the P values which we present. Similarly the modified AI criterion was not used because it adjusts for the sample size, which in our study is constant. There are several limitations of this study that warrant further discussion. First, although it appears that our sample size is small, relative to the recent large data sets that developed formulae for QTc, the number of individuals in this study exceeded (almost double) many other studies on the effect of exercise on QT interval.11,13,15,28,29 Second, all studies in this field are challenged by the absence of a “gold standard” which is incontestably the standard for independence of QT on
OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
THE AMERICAN JOURNAL
137
Rabkin and Cheng
heart rate and against which all formulae can be compared. In addition, there are only limited approaches to approximate the best formula to determine which formula “best” fits the data compared to another and what constitutes a “best fit.” Thus, in 1 approach, we have to select what should be a good approximation of heart rate independence. We selected a criterion proposed and used by others namely that the slope of the relationship between QT and heart rate should approximately 0. Third, the individual data analysis approach limits the number of points from which to construct a regression line as baseline and at most 3 exercise stages are available. However, conclusions both clinically and in clinical trials can often use similar or fewer ECGs for comparison. In addition, ECG evaluation for inherited QT prolongation uses similar numbers of ECGs during exercise testing.5,7 Fourth, we did not correct for QT or RR hysteresis, ie, the QT interval duration is not only dependent on the simultaneously measured RR intervals but also on the prior RR intervals. Hysteresis, however, is difficult to estimate for a single individual within the Bruce exercise protocol. Equations have been developed to minimize hysteresis but these have used exercise protocols of longer duration, at least 5 minutes at a given work load, and are often followed by a longer rest period before exercise is begun again. Some investigators suggest steady state exercise durations of at least 5 minutes, with multiple ECG recordings are necessary to approximate the impact of hysteresis or the impact of preceding RR intervals on the QT measurement.30 Exercise-induced QT or RR hysteresis is especially evident comparing the resting with the exercise portions of the exercise test as “for a given R-R interval, the QT interval duration is shorter during recovery after exercise than during exercise.”31 Our study deliberately did not examine the recovery stage of exercise because of concerns about the potential impact of QT or RR hysteresis. Regardless within the context of our study different formulae were applied at a standard time at each stage of exercise so that we examine the comparative ability of the different formulae rather than studying the process of QT or RR hysteresis or the attained heart rate or the speed at which the heart rate is reached during exercise. Fifth, heart rate and QT interval were measured from 2 QRS complexes at the completion of each stage of exercise. Although an extended time would have allowed more QT measurements, it would violate the principle of assessing the QT-RR interval relationship at a fixed time and workload. Within the context of our study, however, different formulae were applied at a standard time at each stage of exercise so that we examine the comparative ability of the different formulae as each formula was evaluated under the same conditions. Sixth, newer QT correction formulae have more complex transformations of the RR, which some ECG readers may find to be more challenging for manual calculations. This inconvenience is eliminated by computer applications to permit ease of
138
calculation for these formulae. Lastly, QTc calculation during exercise is mostly used for the diagnosis of long QT syndrome. Further studies are necessary in patients with long QT syndrome (congenital or acquired) to test the value of the newer QTc formulae.
CONCLUSION This study evaluated different approaches to assess the QT interval, a valuable marker for drug-induced cardiac toxicity, electrolyte abnormalities and genetic mutations that carry a high risk of sudden death. The data suggest that 2 new QT-heart rate correction formulae specifically QTcDMT and QTcRTHa should be used preferentially rather than older formulae such as those proposed by Bazett or Fridericia especially when evaluating persons with faster heart rates and in the exercise setting.
REFERENCES 1. Rabkin SW. Aging effects on QT interval: implications for cardiac safety of antipsychotic drugs. J Geriatr Cardiol 2014;11:20–5. 2. Burchell HB. The QT interval historically treated. Pediatr Cardiol 1983;4: 139–48. 3. Roden DM. Drug-induced prolongation of the QT interval. N Engl J Med 2004;350:1013–22. 4. Yi G, Crook R, Guo XH, et al. Exercise-induced changes in the QT interval duration and dispersion in patients with sudden cardiac death after myocardial infarction. Int J Cardiol 1998;63:271–9. 5. Aziz PF, Wieand TS, Ganley J, et al. Genotype- and mutation sitespecific QT adaptation during exercise, recovery, and postural changes in children with long-QT syndrome. Circ Arrhythmia Electrophysiol 2011;4: 867–73. 6. Vincent GM, Jaiswal D, Timothy KW. Effects of exercise on heart rate, QT, QTc and QT/QS2 in the Romano-Ward inherited long QT syndrome. Am J Cardiol 1991;68:498–503. 7. Sy RW, van der Werf C, Chattha IS, et al. Derivation and validation of a simple exercise-based algorithm for prediction of genetic testing in relatives of LQTS probands. Circulation 2011;124:2187–94. 8. Rabkin SW, Cheng XB. Nomenclature, categorization and usage of formulae to adjust QT interval for heart rate. World J Cardiol 2015;7: 315–25. 9. Bazett H. An analysis of the time-relations of electrocardiograms. Heart 1920;7:353–67. 10. Yu PN, Bruce RA, Lovejoy FW, et al. Observations on the change of ventricular systole (QT interval) during exercise. J Clin Invest 1950;29: 279–89. 11. Davidowski TA, Wolf S. The QT interval during reflex cardiovascular adaptation. Circulation 1984;69:22–5. 12. Kligfield P, Lax KG, Okin PM. QTc behavior during treadmill exercise as a function of the underlying QT-heart rate relationship. J Electrocardiol 1995; 28(suppl):206–10. http://dx.doi.org/10.1016/S0022-0736(95)80058-1. 13. Sarma JS, Sarma RJ, Bilitch M, et al. An exponential formula for heart rate dependence of QT interval during exercise and cardiac pacing in humans: reevaluation of Bazett's formula. Am J Cardiol 1984;54:103–8. 14. Fridericia LS. The duration of systole in an electrocardiogram in normal humans and in patients with heart disease. 1920. Ann Noninvasive Electrocardiol 2003;8:343–51. 15. Aytemir K, Maarouf N, Gallagher MM, et al. Comparison of formulae for heart rate correction of QT interval in exercise electrocardiograms. Pacing Clin Electrophysiol 1999;22:1397–401. 16. Benatar A, Decraene T. Comparison of formulae for heart rate correction of QT interval in exercise ECGs from healthy children. Heart 2001;86:199–202. 17. Dmitrienko AA, Sides GD, Winters KJ, et al. Electrocardiogram reference ranges derived from a standardized clinical trial population. Drug Inf J 2005;39:395–405.
THE AMERICAN JOURNAL VOLUME 351
OF THE
MEDICAL SCIENCES
NUMBER 2
February 2016
QT and Exercise
18. Rautaharju PM, Mason JW, Akiyama T. New age- and sex-specific criteria for QT prolongation based on rate correction formulas that minimize bias at the upper normal limits. Int J Cardiol 2014;174:535–40. 19. Hodges M, Salerno D, Erlien D, et al. correction reviewed: evidence that a linear QT correction for heart rate is better. J Am Coll Cardiol 1983;1(2):694. 20. Sagie A, Larson MG, Goldberg RJ, et al. An improved method for adjusting the QT interval for heart rate (the Framingham Heart Study). Am J Cardiol 1992;70:797–801. 21. Adams W. The normal duration of the electrocardiographic ventricular complex. J Clin Invest 1936;15:335–42. 22. Boudoulas H, Geleris P, Lewis RP, et al. Linear relationship between electrical systole, mechanical systole, and heart rate. Chest 1981;80:613–7. 23. Kligfield P, Lax KG, Okin PM. QT interval-heart rate relation during exercise in normal men and women: definition by linear regression analysis. J Am Coll Cardiol 1996;28:1547–55. 24. Rautaharju PM, Surawicz B, Gettes LS, et al. American Heart Association Electrocardi C on CCAC of CFHRS. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part IV: the ST segment, T and U waves, and the QT interval: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias C. J Am Coll Cardiol 2009;53:982–91. 25. Ahnve S. Correction of the QT interval for heart rate: review of different formulas and the use of Bazett's formula in myocardial infarction. Am Heart J 1985;109:568–74. 26. Manion CV, Whitsett TL, Wilson MF. Applicability of correcting the QT interval for heart rate. Am Heart J 1980;99:678.
27. Puddu PE, Jouve R, Mariotti S, et al. Evaluation of 10 QT prediction formulas in 881 middle-aged men from the seven countries study: emphasis on the cubic root Fridericia's equation. J Electrocardiol 1988;21:219–29. 28. Akhras F, Rickards AF. The relationship between QT interval and heart rate during physiological exercise and pacing. Jpn Heart J 1981;22: 345–51. 29. Arrowood JA, Kline J, Simpson PM, et al. Modulation of the QT interval: effects of graded exercise and reflex cardiovascular stimulation. J Appl Physiol 1993;75:2217–23. 30. Jacquemet V, Dube B, Knight R, et al. Evaluation of a subject-specific transfer-function-based nonlinear QT interval rate-correction method. Physiol Meas 2011;32:619–35. 31. Lauer MS, Pothier CE, Chernyak YB, et al. Exercise-induced QT/R-Rinterval hysteresis as a predictor of myocardial ischemia. J Electrocardiol 2006;39:315–23.
From the Department of Medicine (Cardiology) (SWR, XC), University of British Columbia, Vancouver, BC, Canada. Submitted May 28, 2015; accepted September 1, 2015. The authors have no financial or other conflicts of interest to disclose. Correspondence: Simon W. Rabkin, MD, Department of Medicine, Division of Cardiology, University of British Columbia, Level 9 2775 Laurel St, Vancouver, BC, Canada V5Z 1M9. (E-mail: [email protected]).
THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES Copyright © 2016 Southern Society for Clinical Investigation. Published by Elsevier Inc. All rights reserved. VOLUME 351 NUMBER 2 February 2016 www.amjmedsci.com www.ssciweb.org
139