Study patients

This retrospective study comprised two groups of patients with a history of a transmural MI more than one month ago. The first study group consisted of consecutive patients referred to Helsinki University Central Hospital for an electrophysiological evaluation after successful resuscitation from out-of-hospital cardiac arrest not associated with a new MI (no appearance of new Q waves in the ECG and no significant increase in the MB-fraction of the creatinine kinase enzyme) (VF group). The inclusion criteria were: sudden onset of symptoms indicating cardiac arrest, the first documented rhythm after the beginning of symptoms was VF, the patient had definite coronary artery disease in coronary angiography without evidence of any other significant heart disease, the patient had recovered neurologically, and the patient had an available 24-hour Holter recording while ambulatory after the resuscitation. We excluded patients who had a known episode of sustained ventricular tachycardia. The median time from MI to VF was 54 months and from VF to the Holter recording 25 days.

The second study group consisted of post-MI patients without a history of syncope or sustained ventricular tachyarrhythmias that were referred for coronary angiography because of stable coronary artery disease (control group). Patients in the control group were matched with patients in the VF group with regard to the extent of coronary artery disease, the MI localization, and the left ventricular ejection fraction. In all study patients class I and III antiarrhythmic drugs were discontinued at least five half-lives before the study and subjects with a history of use of amiodarone were excluded. Only patients in sinus rhythm and without a complete bundle branch block were included. None of the patients had an implanted cardioverter defibrillator before participating in the study protocol.

According to the clinic's practice that time all patients in the VF group underwent an electrophysiologic study. The study protocol included drive-train stimuli from two right ventricular sites with two drive cycle lengths up to three extrastimuli. Of the 37 patients with VF, 15 (41%) were inducible to sustained monomorphic ventricular tachycardia.

In coronary angiograms, a luminal stenosis more than 50% was considered significant. Either left ventriculography or biplane echocardiography was used to determine the left ventricular ejection fraction. The localization of the previous MI was based on myocardial wall motion abnormalities and Q waves in the ECG. All patients underwent a bicycle exercise test, and the test was considered positive for exercise ischemia if there was a depression of the ST segment > 0.1 mV measured at 80 ms after the J point and an anginal chest pain.

The Ethical Review Committee of the institute approved the study, and informed consent was obtained from all participants.

Holter recordings and analyses

All study subjects underwent a 2-channel 24-hour ECG-recording (Marquette 8500, Marquette Electronics Inc., Milwaukee, WI, USA). The recordings were first analyzed with a Marquette 8000 Holter analysis system (version 5.8 software) to label the QRS complexes to normal, ventricular extrasystoles, or aberrant complexes. The ECG data were manually edited to exclude artefacts, and only recordings containing eligible data for more than 85% of the recording time were accepted into the study.

Measurement of QT intervals

The ECG signal was processed by over sampling, by subtracting the fitted spline baseline, and by retriggering. Signal-to-noise ratio was improved by continuously averaging five successive QRST complexes. By use of the method previously described in detail we computed QT intervals for each normal-labelled QRS complex in the 24-hour recording (7). In QT interval measurements, bifid T waves exhibiting a time interval of < 0.15 s between the highest peak and the later lower peak were calculated into the QT duration; otherwise, the later lower peak was not included into the QT duration (10). All T-waves irrespective of the polarity were accepted if the amplitude was < −0.1 mV or > 0.1 mV. All measurements were obtained using modified lead V5 representing preferentially the left ventricular repolarization.

Data analyses and definitions

The rate dependence of QT intervals was analyzed by plotting the QT values against the preceding respective RR intervals as described earlier in detail (7). Using diurnal QT/RR plots we calculated the QT/RR slope of all diurnal QT intervals (diurnal QT/RR slope; Figure 1, panel A). Our diurnal QT/RR slope is the first suggested method to measure QT interval rate dependence from Holter recordings and was introduced by Merri et al. (11). We also analyzed the rate dependence of QT intervals by computing the median values of QT intervals against RR intervals in RR steps of 10 ms. Heart rates had to be stable for 60 s with RR interval variation < 10%. From median QT/RR curves after 60 s of stable heart rates we computed the mean QT/RR slope using successive RR intervals of 400 ms at RR intervals from 600 to 1 000 ms or up to 1 200 ms if sufficient data at shorter RR intervals was not present (Figure 1, panel B). We developed this method earlier to measure the rate dependence of QT intervals in patients with the long QT syndrome (7). We also calculated the maximal QT/RR slope using successive RR intervals of 100 ms (Figure 1, panel C). This measure was included because it differentiated affected long QT syndrome patients from unaffected subjects in a previous study (7). Wake and sleep periods were determined using each individual's heart rate curve and corresponding wake and sleep values for mean QT/RR slope (daytime QT/RR slope and night-time QT/RR slope, respectively) were calculated. The baseline QT interval adjusted for heart rate (QTc) was measured in lead V5 from 12-lead ECGs using the Bazett formula.

Dynamic QT/RR relationship in post-myocardial infarction patients with and without cardiac arrest

All authors

Jere Järvenpää, Lasse Oikarinen, Petri Korhonen, Heikki Väänänen, Lauri Toivonen & Matti Viitasalo

https://doi.org/10.3109/14017431.2010.490950

Published online:

17 November 2010

Figure 1. Data from a control patient showing specifications of the QT/RR slopes. (Panel A) Continuously averaged five successive QT intervals plotted against respective preceding RR intervals. The line shows the diurnal QT/RR slope. (Panel B and Panel C) The median QT/RR curve after 60 s of stable heart rates in RR steps of 10 ms with the line showing the mean QT/RR slope using successive RR intervals of 400 ms at RR intervals from 620 to 1020 ms (B) and with the maximal QT/RR slope with successive RR intervals of 100 ms (C).

Display full size

Statistical analyses

Continuous variables are presented as mean ± SD. Differences in the normally distributed variables were assessed using the t-test. Mann-Whitney U test and Wilcoxon Signed Ranks test were used as nonparametric tests when appropriate. Proportions were compared by χ² test. The correlation between continuous variables was calculated using Spearman's correlation coefficients. All tests were two-sided and a probability value of p < 0.05 was considered statistically significant. SPSS 14.0 (SPSS Inc, Chicago, Ill, USA) was used for data analysis.

Clinical characteristics

Thirty-seven patients met the inclusion criteria for the VF group and 30 for the control group. Clinical and ECG characteristics of the study patients are presented in Table I. The baseline QTc intervals were similar in the study groups. During the 24-hour EGC recordings, heart rate levels were similar in both groups, whereas VF patients showed lower vagally mediated heart rate variability (HRV) measures than control patients (Table I).

QT/RR relationship

QT/RR slope values calculated from all diurnal QT intervals were similar in the study groups (Table II). However, QT/RR slopes calculated from QT intervals measured at stable heart rates differed between the groups with the control patients showing steeper mean and maximal QT/RR slopes than VF patients (Table II). When we analyzed only those patients receiving beta blocking medication in each study group, the difference between groups in mean as well as in maximal QT/RR slope measurements remained similar (data not shown). Both study groups showed steeper mean QT/RR slope during daytime than during sleep (Table II). To evaluate possible confounding effects of gender on the QT/RR slope values we analyzed the slope values in men only in each study group. Among men the mean QT/RR slope values in the control and in the VF group were 0.20 ± 0.09 and 0.15 ± 0.09 , respectively (p=0.04) and the corresponding maximal QT/RR slope values were 0.42 ± 0.22 and 0.33 ± 0.18, respectively (p=0.01).

Relation of QT/RR slopes, exercise ischemia, localization of MI, ejection fraction, and measurements of heart rate variability

To assess the potential influence of myocardial ischemia and localization of previous MI on the QT/RR slope values, we divided the entire study population into corresponding two subgroups on the basis of a positive or negative exercise test result for ischemia as well as on the basis of anterior or inferior location of previous MI. We observed no significant differences in the QT/RR slope measurements between the subgroups. Within the whole study population, none of the QT/RR slope measures correlated with left ventricular ejection fraction (data not shown). The mean QT/RR slope correlated with the standard deviation of successive RR intervals (SDNN) (r=0.39, p <0.001) and the maximal QT/RR slope correlated with the low to high frequency ratio (LF/HF ratio) of HRV (r=0.37, p=0.003). In the VF group, daytime QT/RR slope correlated with the LF/HF ratio (r=0.399, p=0.016) whereas night-time QT/RR slope correlated with SDNN (r=0.455, p=0.006). In the control group, no statistically significant corresponding correlations were observed.

T-wave polarity and the QT/RR relation

During 24-hour recordings, positive T-waves only, negative T-waves only and both positive and negative T-waves were present in 17, eight and five patients in the control group and in 15, 14 and eight patients in the VF group (p=0.45), respectively. To evaluate the influence of T-wave polarity on QT/RR slope measurements we divided the entire study population into three subgroups on the basis of positive T-waves only, negative T-waves only, or both positive and negative T-waves. The diurnal as well as the mean QT/RR slope values were similar between the three subgroups whereas maximal QT/RR slopes were steeper in cases with both positive and negative T-waves (data not shown). The intervindividual variability of diurnal, mean and maximal QT/RR slope values were greatest in cases with both positive and negative T-waves (data not shown). Figure 2 displays the complexity of QT/RR slope measurements in cases with both positive and negative T-waves.

Dynamic QT/RR relationship in post-myocardial infarction patients with and without cardiac arrest

All authors

Jere Järvenpää, Lasse Oikarinen, Petri Korhonen, Heikki Väänänen, Lauri Toivonen & Matti Viitasalo

https://doi.org/10.3109/14017431.2010.490950

Published online:

17 November 2010

Figure 2. Two QT/RR data plots in two different patients exhibiting both positive and negative T-waves. (A) A two-part diurnal QT/RR data plot with positive T-waves and negative T-waves yielding the lower and higher parts of the plot, respectively. The diurnal QT/RR slope (continuous line) is higher than the corresponding slopes of the part of the plot with positive T-waves only or with negative T-waves only (dotted lines). (B) A two-part diurnal QT/RR data plot with positive T-waves and negative T-waves yielding the higher and lower parts of the plot, respectively. The diurnal QT/RR slope (continuous line) is lower than the corresponding slope of the part of the plot with positive T-waves only or with the negative T-waves only (dotted lines). Note that the plot with positive T-waves may result in lower (A) or higher (B) QT/RR slope than the plot with negative T-waves. Representative ECG signals are shown below the plots. Note also that the slopes are influenced by the number of data points in each part of the plot (point density not shown).

Display full size

Main findings

The present results showed different QT/RR relationship between post-MI patients with and without a history of VF. Patients without a history of VF exhibited steeper QT/RR relation than patients who had experienced VF.

Measurement of QT/RR slopes

In this study among post-MI patients we observed steeper mean QT/RR slopes and steeper maximal QT/RR slopes determined at stable heart rates in patients without a history of VF than in patients with a history of VF. On the other hand, QT/RR slopes calculated from all diurnal QT intervals on beat-to-beat basis were similar in the study groups. In a post-MI patient population substantial number of subjects do not exhibit T-waves with positive polarity in lead V5 or show T-waves with both positive and negative polarity arguing against the use of only positive T-waves in the QT/RR slope measurements. The present results suggest that in cases with both positive and negative T-waves the use of mean QT/RR slopes determined at stable heart rates may express the QT dynamicity better than the use of slopes derived from all diurnal QT intervals. In addition, the maximal QT/RR slope using successive RR intervals of 100 ms was especially sensitive to the polarity differences of T-waves.

Using 24-hour Holter recordings Merri et al. evaluated first the dynamic relation between ventricular repolarization duration and cycle length by calculating the best-fit linear regression equation between these two parameters on beat-to-beat basis (11). Later, the QT/RR relation has also been evaluated after converting the 24-hour ECG recording into median QRST-templates obtained at 30 s intervals (4), or averaging only beats preceded by a period of at least one minute stable heart rate (12), and finally calculating the QT/RR slope values from virtual QRST-templates. Methods using either QRST- templates at for example 30 s intervals, or methods averaging beats preceded by a period of at least one minute stable heart rate and calculating few QRST templates may combine the information of different types of T-waves in QT/RR slope calculation. These two approaches, however, use virtual T-waves possibly obscuring the arrhythmogenically important abrupt dynamicity of T-waves (13). The present mean QT/RR slope method includes both positive and negative T-waves and averages continuously five consecutive beats only allowing the influence of dynamic changes on the mean QT/RR slope values. In addition, because the linear QT/RR relationship becomes curvilinear at heart rates higher than 100 beats per minute (14), we limited the calculation of the QT/RR slope values between 50 to 100 beats per minute as earlier recommended also by Maison-Blanche (9).

T-wave morphology

The present results showed that the QT interval rate relation depends on the polarity of T-waves with patients exhibiting both positive and negative T-waves showing steepest maximal QT/RR slopes. Thus, the management of T-waves with different polarity in the analysis of the QT interval rate dependence may modify the results in post MI patients. Recently, new measures of the T-wave morphology has been shown to both reveal drug-induced repolarization abnormalities (15) and provide prognostic information in general population better than baseline QT interval (16).

QT/RR relation in post-MI patients susceptible to life-threatening ventricular tachyarrhythmias

Previously Extramiana et al. observed steeper QT/RR relation in post-MI patients presenting with sustained ventricular tachycardia compared with post-infarction patients without sustained ventricular arrhythmias (17). They averaged beats preceded by a period of at least one minute stable heart rate and calculated mean QRST templates at RR intervals of 20 ms, which typically yielded little more than 10 templates from one 24-hour recording (12). Chevalier et al. reported an increased diurnal QT/RR slope to be an independent predictor of sudden death after myocardial infarction. They calculated the QT/RR slopes using median QRST-templates obtained at 30 s intervals (2880 templates during a 24-hour recording) and excluded T-waves < 0.15 mV (1). Szydlo et al. found steeper QT/RR relationship in post-MI patients with a history of sustained ventricular tachycardia or VF than in patients without a history of ventricular tachyarrhythmias (2). They calculated the QT/RR slopes using all diurnal QT intervals and included T-waves “satisfactory for QT analysis”. Previous studies concerning the QT/RR relationship in post-MI patients with ventricular tachyarrhythmias have included also patients with sustained ventricular tachycardia (without having VF) who show steep QT/RR slopes (17) and, according to our experience, exhibit also both positive and negative T-waves (unpublished data). We excluded patients who had clinical episodes of sustained ventricular tachycardia to focus on those patients in whom ventricular tachyarrhythmia likely degenerates rapidly to fibrillation. Thus, differences between previous and the present studies both in the methodology and in the study populations may explain why previous studies showed steeper QT/RR slope values in the arrhythmia patient groups.

The present unexpected finding of lower mean QT/RR slope in post-MI patients susceptible to VF corresponds to previous findings reported by Tavernier et al. in patients with VF without structural heart disease (18) and by Fujiki et al. in patients with idiopathic VF (19). It is important to note that patients without a structural heart disease do usually exhibit positive T-waves only and thus the differences in the T-wave polarity may not play any role in QT/RR slope measurements among those patient groups.

In this study both patient groups showed a day-to-night difference in the mean QT/RR slope values indicating a preserved autonomic control of the QT/RR relationship (5). However, although VF patients showed a diminished autonomic activity measured as the HRV, we observed stronger autonomic modulation of the QT/RR relationship in the VF group compared to controls. An analogous finding with an exaggerated autonomic modulation of QT intervals in post-MI patients showing diminished autonomic activity has been reported also previously (20). The location and the extent of myocardial injury as well as the myocardial ischemic potential were similar in the study groups. Moreover, in spite of the exaggerated autonomic modulation the QT/RR slopes were lower in the VF group. This suggests possible heritable differences in the cardiac ionic channel properties between the study groups encouraging future molecular genetic research in post-MI arrhythmias.

Study limitations

This study was a retrospective analysis of 24-hour ECG recordings in a limited number of post-MI patients with a history of VF not associated with a new MI after these patients were successfully resuscitated from VF. Therefore, we can not generalize our results to the entire population of patients with a history of VF in different clinical settings. We did not perform programmed stimulation or long term follow-up among control patients. Therefore, it is not certain, whether the control group is one without vulnerability to ventricular tachyarrhythmias. In addition, we can not rule out the possibility that the QT/RR relationship may be different after the resuscitation. We did not measure QT/RR slopes using virtual QRST templates obtained at specified time intervals or extracted using selective beat averaging. Therefore, we can not compare the classical QT/RR slope analyses with the mean QT/RR slope approach used in this study. We emphasize that our method measures different aspects of the QT/RR relationship than classical QT/RR analyses with the present results giving also different message as compared with previous QT/RR slope studies. The QT/RR slope measurements used in this study showed interindividual variation within the study groups limiting the use of these measurements in evaluating arrhythmia risk in patients after MI. The study population consisted of predominantly male patients.