Author + information
- Received April 19, 2004
- Revision received February 21, 2005
- Accepted March 15, 2005
- Published online July 19, 2005.
- Alfred E. Buxton, MD, FACC⁎,⁎ (, )
- Michael O. Sweeney, MD, FACC†,
- Mark S. Wathen, MD‡,
- Mark E. Josephson, MD, FACC§,
- Mary F. Otterness, MS∥,
- Elaine Hogan-Miller, PhD, RN∥,
- Alice J. Stark, PhD, RN∥,
- Paul J. DeGroot, MS∥,
- PainFREE Rx II Investigators
- ↵⁎Reprint requests and correspondence to:
Dr. Alfred E. Buxton, Cardiovascular Division, Rhode Island Hospital, 2 Dudley Street, Suite 360, Providence, Rhode Island 02905.
Objectives The aim of this study was to determine whether QRS duration (QRSd) correlates with occurrence of ventricular arrhythmia in patients with coronary disease (CAD) receiving implantable cardioverter-defibrillators (ICDs).
Background A QRSd measured on a standard electrocardiograph (ECG) correlates with total mortality risk in CAD patients at high risk for sudden death; however, the relationship between QRSd and risk of ventricular tachyarrhythmias (ventricular tachycardia/ventricular fibrillation [VT/VF]) is unclear.
Methods PainFREE Rx II was a randomized trial, comparing efficacy of antitachycardia pacing versus shock therapy for VT/VF in patients receiving ICDs. We retrospectively correlated the QRSd and specific ECG conduction abnormalities on the 12-lead ECG at study entry with occurrence of VT/VF in 431 patients with CAD enrolled in the trial.
Results The QRSd was ≤120 ms in 291 of 431 (68%) patients. Left bundle branch block (LBBB) was present in 65 patients, right bundle branch block (RBBB) in 48 patients, and nonspecific intraventricular conduction delay (IVCD) was present in 124 patients. Over 12 months’ follow-up, VT/VF occurred in 95 (22%) patients (22% of patients with QRSd ≤120 ms vs. 23% of patients with QRSd >120 ms, p = NS). Patients with LBBB were less likely to experience at least one VT/VF episode than patients with QRSd <120 ms. Patients with RBBB and nonspecific IVCD did not differ from patients with narrow QRS complexes with regard to occurrence of tachycardias.
Conclusions The QRSd and ECG conduction abnormalities are not useful to predict ICD benefit in patients having the characteristics of our study population. The utility of QRSd to predict VT/VF events in patients with CAD requires further prospective evaluation.
Implantable cardioverter-defibrillators (ICDs) reduce the risk of sudden death and total mortality in patients who have survived spontaneous episodes of potentially lethal ventricular tachyarrhythmias (1). The ICDs reduce mortality as well as risk of sudden death in some patients with chronic coronary artery disease (CAD) that have never experienced symptomatic ventricular tachyarrhythmias (2–4). The results of these trials have led to increasing numbers of patients that may benefit from ICD therapy. Although ICD therapy is highly effective, it is expensive and carries certain risks. Therefore, we must strive to identify characteristics of patients most likely to benefit from this therapy and minimize the number of implants in patients that do not require this therapy.
Clearly, the risk of sudden death and total mortality is not equal among all patients meeting the entry criteria of the reported trials. Previous studies of patients with CAD have identified a number of factors carrying prognostic significance in patients at risk for sudden death. Among the factors related to prognosis are the presence of certain electrocardiographic (ECG) conduction abnormalities and QRS duration (QRSd) on the standard ECG (5). One study suggested that patients with very wide QRS complexes (>150 ms) might derive greater survival benefit from ICD therapy than patients with narrow QRS complexes (4). Although duration of the QRS complex and the presence of bundle branch block are associated with poor prognosis in certain populations, there are no data showing a cause-and-effect relationship between electrocardiographic conduction abnormalities and ventricular tachyarrhythmias. Hence, it is not clear whether basing decisions to implant ICDs on QRSd or conduction abnormalities will be useful clinically.
The purpose of the current study was to examine the relationship between conduction abnormalities and QRSd on the standard ECG and the spontaneous occurrence of ventricular tachyarrhythmias in patients receiving clinically indicated ICDs. We also sought to determine whether patients receiving ICD implantation for primary prevention versus secondary prevention indications differed with regard to the influence of QRSd or specific conduction abnormalities on occurrence of tachyarrhythmias.
This analysis was performed on data from patients with CAD enrolled in the PainFREE Rx II Study (6). This was a randomized trial comparing empiric antitachycardia pacing (ATP) versus shocks as therapy for spontaneous episodes of rapid ventricular tachyarrhythmias in patients receiving ICDs. Patients receiving clinically indicated ICDs were eligible for participation in this trial if they did not have hypertrophic cardiomyopathy, Brugada, or long QT syndromes. A total of 634 patients were enrolled at 42 U.S. centers from January 2001 to March 2002. Of these, 431 had CAD and a standard 12-lead ECG suitable for analysis. Only these 431 patients were included in the current analysis. Two hundred six (48%) of these patients received ICDs for primary prevention of sudden death, and 225 received the ICD for secondary prevention. A target follow-up period of one year was incorporated into the trial design. Patients were treated with ICDs from the GEM family of Medtronic ICDs. In 75% of cases, dual-chamber ICDs were implanted. The ICDs were programmed to uniform VT detection criteria at the time of implantation. Three detection zones were utilized: a “slow VT zone” (cycle length cut-off ≥320 to 360 ms), a “fast VT zone” (<320 to 240 ms), and a “VF zone” (cycle length <240 ms). For Fast VT and VF, the number of intervals to detection was programmed to 18 out of 24 intervals, whereas slow VT required 16 consecutive intervals to detect. The ICD has the capability to store data for the last 150 episodes of ventricular fibrillation (VF) and/or ventricular tachycardia (VT). Each episode stored can contain an electrogram strip up to twenty-five seconds in duration.
Each participant in this trial provided written informed consent according to the institutional committee on human research at each study site. The institutional review board at every site approved the study protocol.
Standard 12-lead ECGs were collected from all patients upon entry to the trial. The ECGs were interpreted by a core laboratory that classified the presence of left bundle branch block (LBBB), right bundle branch block (RBBB), and nonspecific intraventricular conduction delay, as well as measuring the QRSd, using standard criteria (7). The core laboratory physicians scoring the ECGs had no information regarding indication for ICD (primary vs. secondary prevention) or occurrence of tachyarrhythmia episodes.
The QRSd was measured from the earliest onset of the QRS complex to the point where the QRS complex returned to baseline. The lead with the longest QRSd was used for this measurement.
Primary prevention indication for ICD
This group was composed primarily of: 1) patients with CAD, left ventricular ejection fraction (EF) ≤0.40, spontaneous nonsustained ventricular tachycardia (NSVT), and inducible sustained VT; or 2) patients with CAD and EF ≤0.30. Excluded from this group were: 1) patients with a history of spontaneous sustained VT or VF, as well as 2) patients with unexplained syncope in whom sustained VT could be induced.
Secondary prevention indication for ICD
Secondary prevention indication for ICD was defined as: 1) spontaneous VF or cardiac arrest without transient or reversible cause; 2) spontaneous sustained VT; or 3) syncope of unknown etiology and sustained VT or VF inducible by programmed stimulation.
Left bundle branch block was defined by a QRSd ≥120 ms, delayed onset of the intrinsicoid deflection in leads I, V5, V6>50 ms, the presence of a broad monophasic, often notched R-wave in leads I, V5and V6, with rS or QS complexes in lead V1and V2, and ST-T-wave vectors opposite in direction to the major QRS vector (7).
Right bundle branch block was defined by a QRSd ≥120 ms, in association with the presence of rsR′, or rSR′ complexes in V1or V2, delayed onset of the intrinsicoid deflection in V1and V2more than 50 ms, and wide, slurred S waves in leads I, V5, and V6. The ST-T-wave vectors are opposite in direction to the major QRS vector (7).
Intraventricular conduction delay
Intraventricular conduction delay was defined as QRSd ≥110 ms with morphology different than left bundle branch block or right bundle branch block (7).
Sudden cardiac death
A sudden cardiac death was a death within 1 h after onset of acute symptoms, including unwitnessed death that was unexpected and without other apparent cause or death during sleep.
Non-sudden cardiac death
Non-sudden cardiac death was defined as all cardiac deaths not classified as sudden deaths, including cardiac deaths of hospitalized patients on inotropic support.
Non-cardiac deaths were deaths not classified as cardiac.
If insufficient information was available, the classification is “unknown.”
Analysis of end points
An expert panel analyzed all available stored electrograms of VT/VF episodes to verify the ventricular origin of tachycardias. The expert panel classifying electrograms was blinded to the results of the 12-lead ECG analysis. We analyzed the relation of QRSd to events in two ways. The primary analysis was performed dichotomizing patients on the basis of QRSd >120 ms versus ≤120 ms. We also analyzed results using the standard medical definition of QRSd ≥120 ms versus <120 ms.
All deaths were reviewed by an Adverse Events Adjudication Committee, composed of three members blinded to randomization group, to the study objectives, and to QRSd. The committee was provided with the classification of death as specified by the investigator, the patient’s history, and the physician’s dictated death summary. Deaths were classified as sudden, non-sudden cardiac, non-cardiac, or unknown.
Continuous variables are summarized as median, 25th, and 75th percentiles, and are compared using the Wilcoxon rank sum test. Dichotomous variables are described by numbers and percentages and are compared using the Fisher exact test. All tests were performed at the 5% type I error level; however, the Hochberg method was used to adjust the testing levels for demographic comparisons, in order to account for multiple comparisons (8). Logistic regression was used to investigate differences in the likelihood of at least one episode. Kaplan-Meier survival curves were constructed to illustrate time-to-event information between dichotomized QRSd groups for ventricular tachyarrhythmia and death events. Log-rank p values are reported in these cases.
Characteristics of patients enrolled are delineated in Table 1.As expected, most patients had a history of myocardial infarction (MI), and in most cases, the MI had occurred more than six months prior to enrollment in this study. The characteristics of patients who received ICDs for primary versus secondary prevention did not differ significantly. The median EF for both groups was 30%. The median duration of follow-up in this study was 12 months for both patient groups.
Approximately one-third of the patients had a QRSd >120 ms, regardless of the ICD indication (Table 2).Likewise, the distribution of specific conduction abnormalities was similar for primary and secondary prevention patients (Table 2).
The characteristics of spontaneous and induced arrhythmias are presented in Table 3.A substantial minority of patients had a history of atrial tachyarrhythmias. Spontaneous NSVT was documented more frequently in the primary prevention patients. Sustained monomorphic VT was induced more frequently in primary prevention patients.
During follow-up, a total of 95 patients experienced at least one ICD therapy in response to a ventricular tachyarrhythmia. Twenty-three of the 138 (17%) primary prevention patients with a QRSd ≤120 ms experienced an episode. In comparison, 13 of the 68 (19%) primary prevention patients with a QRSd >120 ms experienced an episode (p = NS). Forty of 153 (26%) secondary prevention patients with a QRSd ≤120 ms experienced an episode, and 19 of 72 (26%) secondary prevention patients with a QRSd >120 ms experienced an episode (p = NS). Overall (primary and secondary prevention patients grouped together), the odds of experiencing an episode in patients whose QRSd was >120 ms was 1.072 times that for patients having a QRSd ≤120 ms (95% confidence limits 0.66 to 1.74; p = NS). Similar results were obtained when we compared patients with QRSd ≥120 ms versus <120 ms. Furthermore, we found no difference in time to first ventricular tachyarrhythmic event between QRSd groups (Fig. 1).
Because treatment with ATP occurs with less time delay than shock therapy, it is possible that some episodes of tachycardia treated with ATP as initial therapy could represent “false positive episodes” that would have terminated spontaneously after charging, before shock delivery. Therefore, we performed a similar analysis, counting only episodes of ICD discharge for the 220 patients randomized to shock as initial therapy for fast VT, excluding slow VT (cycle length ≥320 ms) episodes. This analysis provided similar results. Overall, 11 (15%) of 75 patients with a QRSd >120 ms versus 17 (12%) of 145 patients having QRSd ≤120 ms experienced an episode of tachycardia (p = NS; odds ratio 1.29 [95% confidence limits 0.57 to 2.93]).
We also analyzed the relation between QRSd and tachycardia episodes, examining a variety of QRSd cutoffs. We analyzed the sensitivity and specificity of QRSd, ranging from 80 to 190 ms in 10-ms intervals (Table 4).No QRSd cutoff yielded a clinically useful combination of sensitivity or specificity. Similar results were obtained when we examined groups of patients receiving ICDs for primary or secondary prevention individually.
We then evaluated the relation between QRSd and frequency of VT/VF episodes for patients who experienced at least one episode of VT/VF during the follow-up period (Table 5).The density of recorded tachycardia episodes in patients with one or more episodes was not related to QRSd for primary or secondary prevention patients using the >120 and ≤120 ms dichotomization.
We also evaluated the relation between specific conduction abnormalities and the occurrence of VT/VF (Table 6).Patients with LBBB were significantly less likely than patients with QRSd <120 ms to experience at least one tachycardia episode. This was not observed in comparisons of patients with RBBB or patients with IVCD and patients with narrow QRS. Among patients experiencing at least one tachyarrhythmia episode, we found no evidence of a difference in episode density.
We analyzed the relation between mortality and QRSd; significantly more patients died in the QRSd >120 ms group than in the QRSd ≤120 group, 15% (21 of 140) versus 9% (25 of 291) (p = 0.047). When examining the time to the event of either a ventricular tachyarrhythmia or all-cause death, however, no difference was found between the two groups (p = NS). Figure 2depicts the results of this analysis. Table 7details the causes of deaths for the two QRSd groups. The population experienced two sudden cardiac deaths and three deaths of unknown cause. In fact, only three of these five patients died without experiencing a recorded ventricular tachyarrhythmia. Thus, the Kaplan-Meier survival curve for time to the event of either ventricular tachyarrhythmia or SCD or death from unknown causes would be similar to Figure 1.
The major finding of this study is that QRSd does not predict occurrence of VT/VF resulting in ICD therapies. The results were similar for patients who received ICDs for both primary as well as secondary prevention indications. This observation was true whether QRSd was dichotomized at 120 ms or was dichotomized at other cutoffs. Furthermore, whereas the presence of RBBB and IVCD was not associated with the occurrence of at least one VT/VF episode, patients with LBBB were less likely to experience at least one VT/VF episode than patients with narrow QRS complexes. Among patients with at least one episode, however, we did not detect a difference in episode density between those with LBBB and those with narrow QRS complexes. This may have been due to small sample size.
We do not believe these findings to be surprising. We are not aware of data that show any direct link between electrocardiographic conduction abnormalities and the development of VT/VF after MI. The presence of conduction abnormalities is not associated with an increased chance of having inducible sustained VT in patients without prior spontaneous sustained VT (5; The MUSTT Investigators, unpublished data, April 2004). Although patients who experience large MI are more likely to develop an ECG conduction abnormality than those with lesser degrees of myocardial involvement, in some cases, bundle branch block may antedate infarction and bear no relation to CAD. In general, patients who develop bundle branch block as a complication of infarction have larger areas of infarction and increased mortality. Furthermore, LBBB is unlikely to result from anterior MI and may be more likely to reflect the presence of left ventricular hypertrophy. In the MUSTT study (5), patients with LBBB or IVCD (but not RBBB) had a lower ejection fraction and a higher prevalence of heart failure than patients without these abnormalities. The Coronary Artery Surgery Study (CASS) (9) investigators observed that LBBB, but not RBBB, was associated with increased mortality. The MUSTT study (5) also found patients with LBBB or IVCD, but not RBBB, to have increased total mortality. In addition, the risk of arrhythmic death or cardiac arrest was increased in patients with LBBB or IVCD. The increase in total mortality associated with LBBB or IVCD, however, outweighed the increase in arrhythmic death risk, suggesting a lack of a cause-and-effect relation between conduction abnormalities and death. This, in turn, would suggest that basing ICD therapy on the presence of ECG conduction abnormalities is not likely to be a very cost-effective way to reduce mortality of post-infarction patients.
Recent studies have demonstrated a relation between LBBB and left ventricular dyssynchrony, potentially contributing to development of heart failure (10,11). It seems likely that this effect may be at least partly responsible for the increased mortality associated with certain conduction abnormalities in patients with left ventricular dysfunction. Although these abnormalities seem likely to contribute to mortality, there is no evidence, however, that they relate specifically to deaths that result from VT/VF.
The PainFREE Rx II study, upon which this analysis is based, was not designed primarily to answer the question of whether electrocardiographic conduction abnormalities are causally related to the occurrence of ventricular arrhythmias. Therefore, our observations are subject to the limitations of all substudies and cannot be taken as proof that there is no relation between conduction abnormalities and arrhythmia occurrence. The median follow-up of 12 months in this trial is relatively short. Therefore, it is possible that if patients were followed for longer periods, different results might have been obtained. Finally, it must be acknowledged that the total number of patients with specific conduction abnormalities in our study was relatively small, thereby limiting the statistical power of the comparisons.
It is likely that some patient populations with CAD and wide QRS complexes (such as patients with advanced symptomatic heart failure who are candidates for bi-ventricular pacing ICDs) have different mortality risks than our study population. If one asked whether QRSd could determine the occurrence of VT or VF (and thus benefit from a biventricular ICD rather than a biventricular pacemaker) in a patient population with advanced CHF, it is possible that different results would be obtained. Today, however, candidates for biventricular devices are identified primarily by the presence of wide QRS complexes, so this question would be difficult or impossible to answer.
There is no relation between QRSd and VT/VF occurrence in patients with CAD who receive ICDs for primary or secondary prevention of sudden death. Furthermore, with the exception of a modest difference in the proportion of patients experiencing at least one tachycardia episode between patients with LBBB and those with narrow QRS, there was no relation between the occurrence of tachyarrhythmias and specific conduction abnormalities. Thus, there seems to be little basis to recommend using QRSd or specific conduction abnormalities to base decisions regarding utilization of ICD therapy in patients with the characteristics of the study population.
The authors acknowledge the aid of Drs. Gaurang Gandhi, Sameh Khouzam, and Arun Kolli in scoring ECG recordings.
This study was supported by a research grant from Medtronic, Inc. Dr. Buxton receives research grants from Medtronic, Guidant, and St. Jude, and has been a consultant to Medtronic. Drs. Sweeney, Wathen, and Josephson have received research grants from Medtronic. Drs. Sweeney and Josephson have been consultants to Medtronic. Ms. Otterness, Drs. Miller and Stark, and Mr. DeGroot are employees of Medtronic. A list of the PainFREE Rx II Study Sites may be obtained from Elaine Hogan-Miller, Medtronic Inc.
- Abbreviations and Acronyms
- antitachycardia pacing
- coronary artery disease
- ejection fraction
- implantable cardioverter-defibrillator
- left bundle branch block
- myocardial infarction
- Multicenter Unsustained Tachycardia Trial
- QRS duration
- right bundle branch block
- ventricular fibrillation
- ventricular tachycardia
- Received April 19, 2004.
- Revision received February 21, 2005.
- Accepted March 15, 2005.
- American College of Cardiology Foundation
- Zimetbaum P.J.,
- Buxton A.E.,
- Batsford W.,
- et al.
- Wathen M.S.,
- DeGroot P.J.,
- Sweeney M.O.,
- et al.
- Willems J.L.,
- Robles de Medina E.O.,
- Bernard R.,
- et al.
- Hochberg Y.
- Freedman R.,
- Alderman E.,
- Sheffield L.,
- Saporito M.,
- Fisher L.
- Grines L.C.,
- Bashore T.M.,
- Boudoulas H.,
- Olson S.,
- Schafer P.,
- Wooley C.E.
- Ziao H.B.,
- Brecker S.J.,
- Gibson D.G.