Abstract
Background: QT dispersion is considered an index of spatial inhomogeneity of repolarization duration and increased dispersion of ventricular repolarization is supposed to increase the risk of ventricular arrhythmia. Circadian variation in QT dispersion was investigated.
Methods: Three different modes of lead selection was used: all 12‐leads (QTdisp 12), only precordial leads (QTdisp 6), and one pair of preselected leads (QTdisp 2) in a 24‐hour Holter recording every fourth hour each comprising 10 consecutive measurements in 54 healthy subjects, 29 patients with coronary artery disease (CAD), and 29 patients with heart failure (HF).
Results: A significant circadian variation was observed in healthy subjects when modes QTdisp 12 and QTdisp 6 were used (Mean ± SD 35.58 ± 16.48 ms; P < 0.0001; and 28.82 ± 16.02 ms; P < 0.0001, respectively), and in patients with CAD (Mean ± SD 37.86 ± 17.87 ms; P < 0.01; and 28.72 ± 17.06 ms; P < 0.0001, respectively), whereas no circadian variation was observed in QTdisp 2. No circadian variation was observed in patients with HF irrespectively of lead selection. Patients with CAD without myocardial infarction (MI) had a circadian variation in QTdisp 12 (Mean ± SD 33.13 ± 14.86 ms; P < 0.05), whereas no circadian variation was observed in patients with MI (Mean ± SD 40.35 ± 18.80 ms; P = NS).
Conclusions: Circadian variation of QT dispersion was detected in healthy subjects and in patients with uncomplicated CAD, but not in those who had suffered a previous MI and in patients with HF. The number of leads among which selection of the longest and shortest QT intervals took place was critical for the disclosure of circadian variation of QT dispersion.
Keywords: circadian variation, QT dispersion, coronary artery disease, heart failure
QT dispersion defined as the difference between the longest and shortest QT interval measured on a standard 12‐lead electrocardiogram is considered to be an expression of inhomogeneity of ventricular repolarization 1 , 2 and a marker of risk for ventricular arrhythmias. 3
Patients with coronary artery disease and heart failure have an increased risk of sudden cardiac death. The onset of acute myocardial infarction and sudden cardiac death have a circadian variation with a peak incidence in the early morning hours. 4 , 5 , 6 Abnormalities of the autonomic nervous system appear to be involved in the genesis of sudden cardiac death. The sympathovagal tone may affect QT variability in healthy subjects 7 and treatment with beta‐blockers in patients with heart failure is associated with a reduction in QT dispersion 8 indicating that sympathetic deactivation may cause decreased QT dispersion.
The circadian variation in QT dispersion has been investigated in healthy subjects, 9 , 10 patients with coronary artery disease, 11 , 12 and congestive heart failure 13 , 14 with conflicting observations. These studies vary in selection of leads and timing of measurements and therefore the results are difficult to compare. We have in a methodological study demonstrated that lead selection and sample length influence circadian variation of QT dispersion. 15
In this study, QT dispersion was measured in Holter recordings from three groups: (A) Healthy subject matched for age and sex with the other groups, (B) Patients with chronic coronary artery disease (CAD), and (C) Patients with heart failure (HF). We used three different modes of lead selection: all 12 leads (QTdisp 12), only precordial leads (QTdisp 6), and one pair of preselected leads selected from a 12‐lead ECG strip recorded at 0300 h (QTdisp 2). The aim was to investigate the profile of circadian variation of QT dispersion in these three groups of patients using the abovementioned modes of lead selection.
METHODS
The reproducibility of manually corrected measurements of QT interval in all available leads done every 30 minutes was analyzed using a paired t‐test demonstrating no significant difference in respect to interobersever or intraobserver variability. We, in addition, demonstrated that QT interval measurement had an interobserver and intraobserver reproducibility with a relative error of 0.4% and 1.0% respectively, whereas calculation of QT dispersion had an interobserver reproducibility with a relative error of 2.8% and intraobserver reproducibility with a relative error of 16.4%.
QT dispersion measurement was performed semiautomatically using a commercially available recording and analysis equipment (Holter for Windows, Rozinn Electronics Inc, Glendale, NY), which recorded electrocardiograms digitally and stored three ECG‐leads continuously and in addition recorded a short 12‐lead ECG segment each minute over a 24‐hour period. The automatic measurements were visually checked by an experienced technician who made appropriate correction of cursor locations and excluded tracings in which the T wave was isoelectric or of too low amplitude for accurate determination of the end point. Only recordings containing eight or more usable leads were accepted.
The end of T wave was defined as the intersection of the isoelectric line of the ECG tracing with the tangent to the inflection point to the descending part of the T wave. In the presence of a prominent U wave or a slurring of the endpoint of the T wave the lead was excluded.
We have previously demonstrated, 15 that measurement every fourth hour during the 24‐hour period is sufficient to demonstrate a circadian variation of QT dispersion in healthy subjects. Each measurement was performed every fourth hour comprising 10 consecutive measurements of QT maximum and QT minimum. At each point of time QT dispersion was calculated.
The analysis of circadian variation of QT intervals, QTdisp 12, QTdisp 6, and QTdisp 2 was based on uncorrected and heart rate corrected values obtained with Bazett's formula. 16
Study Population
Seventy‐six healthy volunteers were included among responders to an invitation letter, which was sent to 1000 randomly selected inhabitants of Hvidovre Municipal area, Copenhagen. Only subjects without known disease and medication were asked to contact the study group. After interview and routine physical examination, one subject was excluded due to the finding of a first degree A‐V block, and 21 were considered unfit for participation due to low amplitude of the T Wave in more than four leads. All included subjects were in sinus rhythm with QRS complexes of normal configuration and width. A total of 54 subjects were included.
Seventy‐nine patients with chronic coronary heart disease participating in the CLARICOR 17 study were asked to participate in Holter recording before entering the CLARICOR study. Fifty patients were considered unfit for participation due to low T wave in more than four leads. All included were in regular sinus rhythm with QRS complexes of normal configuration and width except two subjects who had left bundle branch block. A total of 29 patients with coronary artery disease were included.
Over a period of 3 months, 71 patients with congestive heart failure were consecutively included in the outpatient clinic at Copenhagen University Hospital, Hvidovre Hospital. A medical history was obtained including medication and a physical examination was performed on the same day as the Holter recording began. Forty‐two patients were considered unfit for participation due to low T wave in more than four leads leaving a total of 29 patients with congestive heart failure to be included. Only patients with an echocardiographically established ejection fraction ≤ 40% were asked to participate.
All participants gave informed consent and the regional ethical committee approved the study (KF 01–079/01, 01–080/01).
Statistical Analysis
Results are given as mean ± (SD) unless otherwise stated. Statistical calculation was done using SAS System (SAS 8.02), (SAS Institute Inc, Cary, NC, USA).
Mixed models analysis was used to analyze for circadian variation due to the possibility of this method to take into account specific individual effects (random effects) as well as correlation patterns over time. P < 0.05 was considered as the limit of statistical significance.
RESULTS
In Table 1 the clinical characteristics of all three study groups are listed: (A) 54 healthy subjects, (B) 29 patients with coronary artery disease, and (C) 29 patients with heart failure.
Table 1.
Characteristics of Healthy Subjects, Patients with Coronary Artery Disease (CAD) and Patients with Heart Failure (HF)
| Healthy Subjects | CAD | Heart Failure | |
|---|---|---|---|
| Number | 54 | 29 | 29 |
| Age mean (SD) | 63(9) | 65(8) | 68(7) |
| Age range | 54–81 | 50–81 | 53–84 |
| Female/male | 25/29 | 7/22 | 10/19 |
| MI | 0 | 19 | 17 |
| Angina | 0 | 10 | 4 |
| Arterial hypertension | 0 | 9 | 5 |
| Diabetes mellitus | 0 | 4 | 3 |
| RBBB | 0 | 0 | 2 |
| LBBB | 0 | 2 | 9 |
| Beta‐blockers | 0 | 9 | 11 |
| Diuretics | 0 | 8 | 27 |
| ACE‐inhibitors | 0 | 6 | 19 |
CAD = Coronary artery disease; HF = Heart failure; RBBB = right bundle branch block; LBBB = left bundle branch block.
Table 2 demonstrates the mean ± SD values of QTc maximum, QTc minimum, QTdisp 12, QTdisp 6, and QTdisp 2 in the three groups. A significant difference was found in patients with coronary artery disease compared with healthy subjects with respect to mean values of QTc maximum, QTc minimum, QTdisp 12, and QTdisp 2 (all P < 0.0001), but not with QTdisp 6. Patients with heart failure had a significant difference in mean values of QTc maximum, QTc minimum, QTdisp 12, QTdisp 6, and QTdisp 2 (all P < 0.0001) compared to healthy subjects.
Table 2.
Twenty‐Four‐Hour Mean ± (SD) Values of QTc Maximum, QTc Minimum, and QT Dispersion Measured by Three Different Modes of Lead Selection (QTdisp 12, QTdisp 6, and QTdisp 2) in Healthy Subjects, Patients with Coronary Artery Disease (CAD), and Patients with Heart Failure (HF)
| Qtc Maximum | Qtc Minimum | Qtdisp 12 | Qtdisp 6 | Qtdisp 2 | |
|---|---|---|---|---|---|
| Healthy | 416.62(26.32) | 378.06(24.59) | 35.58(16.48) | 28.82(16.02) | 13.74(16.64) |
| CAD | 433.06(26.85)† | 391.82(22.54)† | 37.86(17.87)† | 28.72(17.06) | 19.57(17.52)† |
| HF | 464.91(40.94)‡ | 408.72(38.32)‡ | 49.97(19.89)‡ | 37.28(18.58)‡ | 27.79(17.47)‡ |
†P < 0.001 versus healthy subjects; ‡ P < 0.0001 versus healthy subjects.
In Table 3 the results of analysis using mixed‐models analysis for circadian variation in QTdisp 12, QTdisp 6, and QTdisp 2 every fourth hour in patients with coronary artery disease, patients with heart failure, and healthy subjects are presented. A significant circadian variation in QTdisp 12 was observed in healthy subjects (P < 0.0001) and in patients with coronary artery disease (P < 0.01), but not in patients with heart failure (P = NS). In the group of healthy subjects and patients with coronary artery disease the mean value of QTdisp 12 at 0400–0500 was significantly higher than those recorded at the five other points of time.
Table 3.
QT Dispersion Measured by Three Different Modes of Lead Selection Every Fourth Hour over 24‐Hours
| Inclusion Period | N | Qtdisp 12 Ms | Qtdisp 6 Ms | Qtdisp 2 Ms |
|---|---|---|---|---|
| Healthy subjects | 54 | |||
| 0400–0500 | 43.24 ± (17.76) | 37.69 ± (17.59) | 14.24 ± (21.42) | |
| 0800–0900 | 35.63 ± (15.64) | 28.63 ± (15.22) | 15.50 ± (16.27) | |
| 1200–1300 | 31.99 ± (14.81) | 25.73 ± (13.50) | 13.44 ± (13.77) | |
| 1600–1700 | 33.19 ± (15.97) | 26.13 ± (15.13) | 13.11 ± (14.46) | |
| 2000–2100 | 33.55 ± (16.49) | 25.42 ± (15.41) | 13.23 ± (13.16) | |
| 0000—0100 | 35.87 ± (15.62) | 30.31 ± (15.35) | 12.91 ± (19.05) | |
| Circadian variation | ||||
| P‐value | <0.0001 | <0.0001 | NS | |
| Patients with coronary artery disease | 29 | |||
| 0400–0500 | 40.37 ± (18.11) | 33.81 ± (18.92) | 20.87 ± (19.99) | |
| 0800–0900 | 40.77 ± (19.04) | 32.58 ± (18.90) | 20.98 ± (16.32) | |
| 1200–1300 | 34.23 ± (16.60) | 24.25 ± (14.14) | 17.04 ± (14.98) | |
| 1600–1700 | 36.35 ± (15.83) | 25.31 ± (14.80) | 19.40 ± (18.88) | |
| 2000–2100 | 34.98 ± (17.29) | 24.79 ± (14.60) | 16.74 ± (13.80) | |
| 0000–0100 | 40.49 ± (18.97) | 31.54 ± (17.52) | 22.40 ± (19.57) | |
| Circadian variation | ||||
| P‐value | <0.01 | <0.0001 | NS | |
| Patients with heart failure | 29 | |||
| 0400–0500 | 50.97 ± (20.36) | 37.72 ± (20.60) | 24.74 ± (17.12) | |
| 0800–0900 | 52.02 ± (21.38) | 38.93 ± (20.07) | 28.76 ± (17.75) | |
| 1200–1300 | 46.80 ± (19.20) | 34.30 ± (16.35) | 27.51 ± (16.75) | |
| 1600–1700 | 49.56 ± (19.45) | 37.87 ± (18.07) | 28.13 ± (17.77) | |
| 2000–2100 | 49.00 ± (18.95) | 35.95 ± (17.67) | 29.49 ± (16.93) | |
| 0000–0100 | 51.50 ± (19.61) | 38.90 ± (18.07) | 28.15 ± (18.20) | |
| Circadian variation | ||||
| P‐value | NS | NS | NS | |
Time‐specific mean values ± (SD) and P‐values of mixed models analysis for circadian variation are presented.
QTdisp 12 = All leads included.
QTdisp 6 = Only precordial leads included.
QT disp 2 = One pair of preselected leads included.
Using precordial leads (QTdisp 6), a significant circadian variation was demonstrated in the group of healthy subjects (P < 0.0001) and in patients with coronary artery disease (P < 0.0001), but not in patients with heart failure (P = NS). In the group of healthy subjects and patients with coronary artery disease, time specific mean values of QTdisp 6 were lower during daytime compared to nighttime, whereas time specific mean values of QTdisp 6 in patients with heart failure were of same magnitude during the whole 24‐hour period.
No significant circadian variation of QT dispersion determined from a pair of preselected leads (QTdisp 2) was seen irrespectively of study group.
The main results did not depend on outliers. Analyses were repeated with exclusion of 10% highest and lowest values without changing the overall results.
Table 4 demonstrates the circadian variation in QT and QTc maximum, QT and QTc minimum, and average QTc intervals using all 12‐leads in healthy subjects, patients with coronary artery disease, and patients with heart failure. In all three groups a circadian variation in QT maximum, QT minimum, and average QT interval was observed (P < 0.0001). A circadian variation in QTc maximum was observed only in patients with coronary artery disease (P < 0.05), whereas no circadian variation was observed in patients with heart failure and healthy subjects. Only in healthy subjects QTc minimum demonstrated a significant circadian variation (P < 0.005), whereas no circadian variation was observed in patients with heart failure and in patients with coronary artery disease. A significantly lower time specific mean value of QTc minimum at 0400–0500 (P < 0.005) compared to the rest of the day was seen in the group of healthy subjects, whereas in patients with coronary artery disease a significantly lower time specific mean value of QTc maximum was observed at 1200–1300 (P < 0.01) compared with rest of the day. Time specific mean values of average QTc intervals using all 12‐leads demonstrated a circadian variation in healthy subjects (P < 0.005) and patients with coronary artery disease (P < 0.005), whereas no circadian variation was observed in patients with heart failure.
Table 4.
Time‐Specific Mean Values ± (SD) of Uncorrected and Heart Rate Corrected QT Maximum Intervals, QT Minimum Intervals and Average QT Intervals Using All 12‐Leads Measured Every Fourth Hour in 10 Consecutive Minutes Are Presented and at the Bottom Line of Each Section the P‐Value for Mixed Model Analysis for Circadian Variation Is Presented
| Study Group (N) | QT Maximum ms | Qtc Maximum ms | QT Minimum ms | Qtc Minimum ms | Average QT Intervals ms | Average QTc Intervals ms |
|---|---|---|---|---|---|---|
| Healthy (54) | ||||||
| 0400–0500 | 416.81 ± (31.11) | 417.08 ± (28.16) | 373.57 ± (29.54) | 373.87 ± (28.03) | 400.18 ± (28.13) | 400.47 ± (25.71) |
| 0800–0900 | 377.77 ± (30.92) | 417.00 ± (26.49) | 342.15 ± (31.10) | 377.49 ± (25.97) | 362.26 ± (29.87) | 399.46 ± (23.27) |
| 1200–1300 | 365.58 ± (25.56) | 413.95 ± (24.91) | 333.59 ± (23.56) | 377.66 ± (22.19) | 350.45 ± (23.83) | 396.69 ± (20.90) |
| 1600–1700 | 371.03 ± (26.85) | 414.45 ± (25.06) | 337.84 ± (28.35) | 377.01 ± (22.46) | 355.80 ± (26.24) | 397.24 ± (21.29) |
| 2000–2100 | 379.09 ± (26.56) | 419.69 ± (25.30) | 345.55 ± (23.54) | 382.49 ± (21.50) | 363.55 ± (23.72) | 402.44 ± (21.38) |
| 0000–0100 | 396.89 ± (29.17) | 417.55 ± (27.52) | 361.02 ± (27.04) | 379.81 ± (25.92) | 381.90 ± (27.25) | 401.75 ± (25.35) |
| Circadian variation | ||||||
| P‐value | <0.0001 | NS | <0.0001 | <0.005 | <0.0001 | <0.005 |
| CAD (29) | ||||||
| 0400–0500 | 426.25 ± (30.48) | 435.94 ± (24.56) | 385.88 ± (26.44) | 394.80 ± (23.00) | 408.72 ± (26.72) | 418.09 ± (21.79) |
| 0800–0900 | 397.29 ± (35.15) | 436.04 ± (28.12) | 356.52 ± (31.90) | 390.98 ± (20.72) | 377.76 ± (32.33) | 414.47 ± (22.66) |
| 1200–1300 | 382.24 ± (32.91) | 427.85 ± (25.90) | 348.01 ± (31.81) | 389.21 ± (21.40) | 365.91 ± (31.00) | 409.44 ± (21.84) |
| 1600–1700 | 383.29 ± (29.73) | 432.81 ± (29.97) | 346.94 ± (29.36) | 391.34 ± (23.88) | 364.85 ± (28.37) | 411.71 ± (24.27) |
| 2000–2100 | 393.12 ± (28.63) | 430.74 ± (24.68) | 358.15 ± (26.69) | 392.34 ± (21.73) | 376.07 ± (26.23) | 412.00 ± (20.88) |
| 0000–0100 | 410.76 ± (33.18) | 434.96 ± (26.67) | 370.28 ± (27.69) | 392.28 ± (24.13) | 393.22 ± (28.80) | 416.40 ± (21.78) |
| Circadian variation | ||||||
| P‐value | <0.0001 | <0.05 | <0.0001 | NS | <0.0001 | <0.005 |
| HF (29) | ||||||
| 0400–0500 | 432.45 ± (41.48) | 462.75 ± (41.14) | 381.48 ± (43.92) | 407.77 ± (41.20) | 408.79 ± (41.15) | 437.24 ± (39.27) |
| 0800–0900 | 415.03 ± (48.24) | 469.50 ± (43.49) | 363.01 ± (46.06) | 410.09 ± (37.39) | 388.97 ± (46.44) | 439.68 ± (38.05) |
| 1200–1300 | 395.58 ± (40.24) | 460.46 ± (40.18) | 348.78 ± (38.88) | 405.76 ± (37.58) | 372.46 ± (38.05) | 433.43 ± (36.56) |
| 1600–1700 | 413.49 ± (42.92) | 464.16 ± (39.30) | 363.93 ± (39.54) | 408.21 ± (33.40) | 389.72 ± (40.46) | 437.20 ± (33.72) |
| 2000–2100 | 411.77 ± (42.85) | 466.73 ± (41.91) | 362.77 ± (41.38) | 411.02 ± (39.72) | 387.57 ± (40.18) | 439.25 ± (38.42) |
| 0000–0100 | 428.69 ± (44.14) | 465.85 ± (39.19) | 377.18 ± (46.53) | 409.44 ± (40.18) | 404.11 ± (43.41) | 438.94 ± (37.11) |
| Circadian variation | ||||||
| P‐value | <0.0001 | NS | <0.0001 | NS | <0.0001 | NS |
Table 5 demonstrates the analysis of circadian variation in QTc maximum, QTc minimum, and QTdisp 12 in patients with coronary artery disease according to previous myocardial infarction (MI) or not. The 24‐hour mean ± SD value of QTdisp 12 was 33.13 ± 14.86 ms in patients with coronary artery disease without myocardial infarction, whereas in patients with coronary artery disease with myocardial infarction the 24 hour mean ± SD value of QTdisp 12 was 40.35 ± 18.80 ms. Patients with coronary artery disease without MI had a significant circadian variation in QTdisp 12 (P < 0.05), whereas patients with MI had no circadian variation in QTdisp 12. Patients with coronary artery disease with myocardial infarction had a circadian variation in QTc maximum (P < 0.05), whereas no circadian variation in QTc maximum was observed in the group without myocardial infarction. No circadian variation in QTc minimum was observed irrespectively of previous myocardial infarction.
Table 5.
Circadian Variation of QTc Maximum, QTc Minimum, and QT Dispersion Using All 12‐Leads (QT disp 12) Every Fourth Hour in Patients with Coronary Heart Disease with or without myocardial Infarction (MI)
| QTc Maximum Ms MI | QTc Minimum Ms MI | QTdisp 12 Ms MI | ||||
|---|---|---|---|---|---|---|
| Yes | No | Yes | No | Yes | No | |
| 0400–0500 | 437.58 ± (27.70) | 432.82 ± (16.77) | 398.24 ± (26.04) | 393.07 ± (19.40) | 40.88 ± (19.06) | 39.41 ± (16.19) |
| 0800–0900 | 440.14 ± (27.41) | 428.24 ± (27.93) | 394.21 ± (19.45) | 390.80 ± (23.58) | 43.67 ± (20.41) | 35.27 ± (14.72) |
| 1200–1300 | 429.81 ± (28.88) | 424.10 ± (18.56) | 390.97 ± (23.80) | 394.22 ± (16.68) | 37.55 ± (17.85) | 27.93 ± (11.65) |
| 1600–1700 | 432.56 ± (31.71) | 433.28 ± (26.50) | 391.75 ± (23.81) | 398.91 ± (23.13) | 38.72 ± (16.23) | 31.83 ± (14.06) |
| 2000–2100 | 432.00 ± (24.76) | 428.33 ± (24.47) | 394.07 ± (22.88) | 397.13 ± (22.01) | 37.39 ± (18.56) | 30.39 ± (13.51) |
| 2400–0100 | 437.93 ± (28.85) | 429.31 ± (20.92) | 395.19 ± (26.10) | 394.92 ± (23.71) | 43.92 ± (19.49) | 33.97 ± (16.14) |
| Circadian variation | ||||||
| P‐value | <0.05 | NS | NS | NS | NS | <0.05 |
Time‐specific mean values and at the bottom P‐values of mixed models analysis for circadian variation are presented.
After a follow‐up of 3.5 years, 12 patients with heart failure had died. The 24‐hour mean ± SD value of QTdisp 12 was 50.83 ± 20.68 ms in the fatal cases, and 49.37 ± 19.30 ms in survivors, which demonstrated no significant difference between the two groups (P = NS). Analysis of the circadian variation in QT disp 12 in patients with heart failure also demonstrated no significant difference between fatal cases and survivors. Similarly, no significant difference of circadian variation of QTdisp 12 was observed between those with or without brandle block, diabetes, sex, or use of beta‐blockers in patients with heart failure. A significant difference was observed when patients with heart failure were subdivided according to ejection fraction (EF): EF ≤ 30, 30 < EF < 45, EF ≥ 45. Patients with heart failure with EF ≤ 30 or EF ≥ 45 had no circadian variation, whereas in the group of patients with heart failure with 30 < EF < 45 a significant circadian variation was observed (P < 0.05).
As it appears in Table 6 no difference in significance was observed in the circadian profile of QTdisp12, QTcdisp 12, QTdisp 6, and QTcdisp 6, QTdisp 2, and QTcdisp 2 comparing the uncorrected and corrected values obtained with Bazett's formula, except in patients with coronary artery disease when we compared the analysis of QTdisp 12 (P < 0.01) with QTcdisp 12 (P = NS).
Table 6.
Uncorrected and Heart Rate Corrected Mean ± (SD) Values of QT Intervals and QT Dispersion in Each of the Three Different Lead Selection Modes (QTdisp 12, QTdisp 6, and QTdisp 2) Including P‐values in the Analysis of Circadian Variation Using Mixed Models
| Healthy Subjects | Coronary Artery Disease | Heart Failure | ||||
|---|---|---|---|---|---|---|
| Mean ± (SD) | P‐Value Circadian Variation | Mean ± (SD) | P‐Value Circadian Variation | Mean ± (SD) | P‐Value Circadian Variation | |
| QTdisp 12 | 35.58 ± (16.48) | <0.0001 | 37.86 ± (17.87) | <0.01 | 49.97 ± (19.89) | NS |
| QTcdisp 12 | 38.56 ± (17.88) | <0.005 | 41.23 ± (19.70) | NS | 56.19 ± (23.07) | NS |
| QTdisp 6 | 28.82 ± (16.02) | <0.0001 | 28.72 ± (17.06) | <0.0001 | 37.28 ± (18.58) | NS |
| QTcdisp 6 | 31.09 ± (17.05) | <0.0001 | 31.11 ± (18.38) | <0.0005 | 41.77 ± (20.92) | NS |
| QTdisp 2 | 13.74 ± (16.64) | NS | 19.57 ± (17.52) | NS | 27.79 ± (17.47) | NS |
| QTcdisp 2 | 14.99 ± (17.83) | NS | 21.07 ± (18.62) | NS | 31.40 ± (19.96) | NS |
| QTmax | 384.50 ± (33.31) | <0.0001 | 398.83 ± (35.31) | <0.0001 | 416.17 ± (44.97) | <0.0001 |
| QTcmax | 416.62 ± (26.32) | NS | 433.06 ± (26.85) | <0.05 | 464.91 ± (40.94) | NS |
| QTmin | 348.95 ± (30.68) | <0.0001 | 360.96 ± (32.03) | <0.0001 | 366.19 ± (44.07) | <0.0001 |
| QTcmin | 378.06 ± (24.59) | <0.005 | 391.82 ± (22.54) | NS | 408.72 ± (38.32) | NS |
| RR | 862.43 ± (154.38) | <0.0001 | 857.85 ± (150.03) | <0.0001 | 815.98 ± (172.36) | <0.0001 |
DISCUSSION
We demonstrated in this study a significant circadian variation in healthy subjects and patients with coronary artery disease, when all 12‐leads (QTdisp 12) or only precordial leads (QTdisp 6) were used, but not when only a preselected pair of leads (QTdisp 2) was used during the 24 hours. No circadian variation of QT dispersion was observed in patients with heart failure irrespectively of lead selection. A significant circadian variation of QTc maximum was demonstrated in patients with coronary artery disease, whereas healthy subjects had a significant circadian variation in QTc minimum. No circadian variation was observed in patients with coronary artery disease and with myocardial infarction, whereas in patients without myocardial infarction a circadian variation was observed. No difference in mean values of QT dispersion or circadian variation of QT dispersion was observed between fatal cases and survivors in the group of patients with heart failure.
Two other studies 11 , 12 have demonstrated that QT dispersion has a circadian variation in patients with coronary artery disease and in healthy subjects. We found that the circadian variation of QT dispersion in healthy subjects was mainly due to variations in QTc minimum, whereas in patients with coronary artery disease the circadian variation of QT dispersion was due to changes in QTc maximum, which both are opposite to the observations done by Yetkin et al. 12 This study 12 included a population of patients with acute ischemia, whereas we studied a group of patients with chronic coronary artery disease. Stierle et al 18 and Yunus et al 19 reported that QT dispersion increases due to a decrease in QT minimum during acute ischemia similar to the findings done by Yetkin et al. 12 We observed an increase of QT dispersion due to an increase in QTc maximum, which could be explained by the fact that we investigated patients with chronic coronary artery disease.
Our study also demonstrated that patients with coronary artery disease without myocardial infarction had a circadian variation in QT dispersion in contrast to, patients with a previous myocardial infarction. Transmural necrosis has been demonstrated to cause regional autonomic denervation. 20 , 21 This regional denervation can increase action‐potential duration, which again can change QT dispersion depending on autonomic tone. Patients with coronary artery disease without myocardial infarction might imitate healthy subjects in respect of demonstrating a circadian variation with an increase of QT dispersion (QTdisp 12) in the morning hours, whereas patients with coronary artery disease with myocardial infarction have an autonomic dysfunction, which might contribute to the loss of circadian variation.
We observed no circadian variation of QT dispersion in patients with heart failure irrespective of which mode of lead selection was used in the measurement of QT dispersion. This is contrary to observations done by Kinoshita et al. 13 and Bonnar et al. 14 Kinoshita et al. 13 demonstrated in 11 patients with a 12‐lead recording that QT dispersion was increased in the afternoon compared to the morning hours, whereas Bonnar et al. 14 described a significant increase in QT dispersion and QTc dispersion from 6 am to 8 am. Bonnar et al. 14 included eight patients with heart failure using 12‐leads measurement in a nine time‐point process, whereas in our investigation 29 patients were included with three different modes of lead selection (QTdisp 12, QTdisp 6, and QTdisp 2) in 6 time periods each comprising 10 consecutive measurements.
We demonstrated no difference in mean values of QT dispersion observed in the fatal cases compared to the survivor group in patients with heart failure. Only patients with a moderately reduced ejection fraction (30% < EF < 45%) had a significant circadian variation. The demonstration of circadian variation observed only in patients with moderately reduced ejection fraction most likely is a random finding due to multiple comparisons.
No difference in mean QT dispersion or circadian variation of QT dispersion was observed between the fatal cases and in the survivor group of patients with heart failure. This is in line with a number of studies, 22 , 23 , 24 demonstrating that QT dispersion had no prognostic value regarding all cause mortality, cardiac mortality, or cardiac arrhythmias in patients with heart failure.
We observed that mean values of QT dispersion were increased in patients with heart failure compared with healthy subjects, which is in accordance with previous reports. 8 , 25 , 26 , 27 , 28 The larger mean values of QT dispersion could be a result of myocardial remodeling resulting in a slower and heterogeneous ventricular repolarization. The remodeling includes changes in left ventricular size and function, which may also influence ventricular repolarization. 29
We, in addition, demonstrated that the significance of circadian variation of QT dispersion did not change, neither when the analysis was performed on uncorrected values nor on corrected values obtained with Bazett's formula, except when all 12‐leads were used in the measurement of QT dispersion in patients with coronary artery disease. We demonstrated that the circadian variation of QT intervals was blunted by heart rate correction formula, which is a finding in agreement with previous study. 30
Limitations: Continuous ECG recording with 12‐lead measurements during 24 hours is difficult, because routine daily activity influences the quality of recording. This could result in omitting of recorded minutes that would otherwise be of interest. In addition, postural changes during Holter monitoring may affect QT dispersion.
CONCLUSION
A circadian variation was observed in healthy subjects and patients with coronary artery disease when all 12‐leads (QTdisp 12) or only precordial leads (QTdisp 6) were used. In healthy subjects a circadian variation of QTc minimum was observed, whereas in patients with coronary artery disease a circadian variation of QTc maximum was observed. Also, a significant circadian variation was observed in patients with coronary artery disease without myocardial infarction, whereas in patients with coronary artery disease with myocardial infarction no circadian variation was observed. In patients with heart failure no circadian variation was observed irrespective of lead selection mode.
REFERENCES
- 1. Day C, McComb J, Campbell R. QT dispersion: An indication of arrhythmia risk in patients with long QT interval. Br Heart J 1990;63:342–344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Zabel M, Lichtlen P, Haverich A, et al Comparison of ECG variables of dispersion of ventricular repolarization with direct myocardial repolarization measurements in human heart. J Cardiovasc Electrophysiol 1998;9:1279–1284. [DOI] [PubMed] [Google Scholar]
- 3. Perkiömäki J, Huikuri H, Koistinen J, et al Heart rate variability and dispersion of QT interval in patients with vulnerability to ventricular tachycardia and ventricular fibrillation after previous myocardial infarction. J Am Coll Cardiol 1997;30:1331–1338. [DOI] [PubMed] [Google Scholar]
- 4. Muller JE, Stone PH, Turi ZG, et al Circadian variation in the frequency of onset of acute myocardial infarction. N Engl J Med 1985;313:1315–1322. [DOI] [PubMed] [Google Scholar]
- 5. Rocco MB, Barry J, Campbell S, et al Circadian variation of transient myocardial ischemia in patients with coronary artery disease. Circulation 1987;75:395–400. [DOI] [PubMed] [Google Scholar]
- 6. Willich SN, Levy D, Rocco MB, et al Circadian variation in the frequency of sudden cardiac death in the Framingham Heart Study Population. Am J Cardiol 1987;60:801–806. [DOI] [PubMed] [Google Scholar]
- 7. Bilan A, Witczak A, Palusiński R, et al Circadian rhythm of the QT interval dispersion in healthy subjects. Correlation with heart rate variability circadian pattern. J Electrocardiol 2005;38:36–42. [DOI] [PubMed] [Google Scholar]
- 8. Bonnar CE, Davie AP, Caruana L, et al QT dispersion in patients with chronic heart failure: β blockers are associated with a reduction in QT dispersion. Heart 1999;81:297–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lee K‐T, Lai W‐T, Chu C‐S, et al Circadian variation og QT dispersion determined by twelve‐lead Holter electrocardiography. Cardiology 2003;100:101–111. [DOI] [PubMed] [Google Scholar]
- 10. Ishida S, Nakagawa M, Fujino T, et al Circadian variation of QT interval dispersion. J Electrocardiol 1997;30(3):205–210. [DOI] [PubMed] [Google Scholar]
- 11. Batur K, Aksöyek S, Oto A, et al Circadian variations of QTc dispersion: Is it a clue to morning increase of sudden cardiac death? Clin Cardiol 1999;22:103–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yetkin E, Şenen K, Ileri M, et al Diurnal variation of QT dispersion in patients with and without coronary artery disease. Angiology 2001;52:5:311–316. [DOI] [PubMed] [Google Scholar]
- 13. Kinoshita O, Wakamatsu M, Tomita T, et al Diurnal variation in QT dispersion in patients with chronic heart failure. Congest Heart Fail 2005;11:5:262–265. [DOI] [PubMed] [Google Scholar]
- 14. Bonnar C, Gillespie N, Mac Fadyen R, et al QT‐dispersion is significantly increased between 6 am and 8 am in heart failure patients‐ A possible role in sudden death. J Am Coll Cardiol 1997;29(SupplA):510A. [Google Scholar]
- 15. Hansen S, Rasmussen V, Larsen K, et al Circadian variation in QT dispersion determined from a 12 lead Holter recording: A methodological study of an age‐ and sex‐stratified group of healthy subjects. Ann Noninvasive Electrocardiol 2007; 12:185–196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bazett JC. An analysis of time relations of electrocardiograms. Heart 1920;7:353–367. [Google Scholar]
- 17. Hansen S, Als‐Nielsen B, Damgaard M, et al The CLARICOR Trial design. Intervention with clarithromycin in patients with stable coronary heart disease. Heartdrug 2001;1:14–19. [Google Scholar]
- 18. Stierle U, Giannitsis E, Sheikhzadeh A, et al Relation between QT dispersion and the extent of myocardial ischemia in patients with three‐vessel coronary artery disease. Am J Cardiol 1998;81:564–568. [DOI] [PubMed] [Google Scholar]
- 19. Yunus A, Gillis A, Traboulsi M, et al Effect of coronary angioplasty on precordial QT dispersion. Am J Cardiol 1997;79:1339–1342. [DOI] [PubMed] [Google Scholar]
- 20. Martins J, Lewis R, Wendt D, et al Subendocardial infarction produces epicardial parasympathetic denervation in canine left ventricle. Am J Physiol 1989;256(3Pt2):859–866. [DOI] [PubMed] [Google Scholar]
- 21. Dae M, Herre J, O'Connel W, et al Scintigraphic assessment of sympathetic innervation after transmural versus nontransmural myocardial infarction. J Am Coll Cardiol 1991;17:6:1416–1423. [DOI] [PubMed] [Google Scholar]
- 22. Brendrup B, Elming H, Jun L, et al For the Diamond Study Group . QT dispersion has no prognostic information for patients with advanced congestive heart failure and reduced left ventricular systolic function. Circulation 2001;103:831–835. [DOI] [PubMed] [Google Scholar]
- 23. Brooksby P, Batin P, Nolan J, et al The relationship between QT intervals and mortality in ambulant patients with chronic heart failure. The United Kingdom Heart Failure Evaluation and Assessment of Risk Trial (UK‐HEART). Eur Heart J 1999;20:1335–1341. [DOI] [PubMed] [Google Scholar]
- 24. Maron B, Leyhe M III, Casey S, et al Assessment of QT dispersion as a prognostic marker for sudden death in a regional nonreferred hypertrophic cardiomyopathy cohort. Am J Cardiol 2001;87:114–115. [DOI] [PubMed] [Google Scholar]
- 25. Mesquita E, Deus F, Guedes C, et al Effects of propranonlol on the QT dispersion in congestive heart failure. Arq Bras Cardiol 1999;73:3:295–298. [DOI] [PubMed] [Google Scholar]
- 26. Peng D, Zhao S, Chen Y, et al Effect of bisoprolol on QT dispersion in patients with congestive heart failure: The etiology‐dependent response. Int J Cardiol 2001;77:141–148. [DOI] [PubMed] [Google Scholar]
- 27. Akbulut M, Özbay Y, Ilkay E, et al Effects of spironolactone and metoprolol on QT dispersion in heart failure. Jpn Heart J 2003;44:681–692. [DOI] [PubMed] [Google Scholar]
- 28. Rocha R, Silva G, Perin E, et al Effects of carvedilol therapy on QT‐interval dispersion in congestive heart failure. Is there a difference in the elderly? Texas Heart Inst J 2003;30:3:176–179. [PMC free article] [PubMed] [Google Scholar]
- 29. Dean JW, Lab MJ. Arrhythmia in heart failure. Role of mechanically induced changes in electrophysiology. Lancet 1989;1:309–312. [DOI] [PubMed] [Google Scholar]
- 30. Smetana P, Batchvarov V, Hnatkova K, et al Circadian rhythm of the corrected QT interval: Impact of different heart rate correction models. Pacing Clin Electrophysiol. 2003;26:383–386. [DOI] [PubMed] [Google Scholar]