Author + information
- Received August 16, 2005
- Revision received October 25, 2005
- Accepted November 1, 2005
- Published online May 2, 2006.
- Theodore Chow, MD, FACC⁎,
- Dean J. Kereiakes, MD, FACC⁎,
- Cheryl Bartone, RN⁎,
- Terri Booth, RN⁎,
- Edward J. Schloss, MD, FACC⁎,
- Theodore Waller, MD, FACC⁎,
- Eugene S. Chung, MD⁎,
- Santosh Menon, MD⁎,
- Brahmajee K. Nallamothu, MD, MPH† and
- Paul S. Chan, MD, MSc†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Paul S. Chan, Cardiology (111-A), VA Ann Arbor Healthcare System, 2215 Fuller Road, Ann Arbor, Michigan 48105.
Objectives The purpose of this study was to assess if microvolt T-wave alternans (MTWA) is an independent predictor of mortality in patients with ischemic cardiomyopathy.
Background Microvolt T-wave alternans has been proposed as an effective tool for identifying high-risk patients with ischemic cardiomyopathy who are likely to benefit from implantable cardioverter-defibrillator (ICD) therapy. However, earlier studies have been limited in their ability to control for baseline differences between MTWA-negative and -non-negative (positive and indeterminate) patients.
Methods We enrolled 768 consecutive patients with ischemic cardiomyopathy (left ventricular ejection fraction ≤35%) and no prior history of ventricular arrhythmia. All patients underwent baseline MTWA testing and were classified as MTWA negative or non-negative. Multivariable Cox regression analyses, stratified by ICD status, were used to determine the association between MTWA testing and mortality after adjusting for demographic, clinical, and treatment differences between MTWA-negative and -non-negative patients.
Results We identified 514 (67%) patients with a non-negative MTWA test. After multivariable adjustment, a non-negative MTWA test was associated with a significantly higher risk for all-cause (stratified hazard ratio [HR] = 2.24 [95% confidence interval 1.34 to 3.75]; p = 0.002) and arrhythmic mortality (stratified HR = 2.29 [1.00 to 5.24]; p = 0.049) but not for nonarrhythmic mortality (stratified HR = 1.77 [0.84 to 3.74]; p = 0.13). In subgroup analyses, a non-negative MTWA test was also associated with a higher risk for all-cause mortality in patients with ejection fractions ≤30% (stratified HR = 2.10 [1.18 to 3.73]; p = 0.01) and after excluding those with indeterminate MTWA tests (stratified HR = 2.08 [1.18 to 3.66]; p = 0.01).
Conclusions Microvolt T-wave alternans is a strong and independent predictor of all-cause and arrhythmic mortality in patients with ischemic cardiomyopathy.
Sudden cardiac death (SCD) is the leading cause of mortality in patients with ischemic heart disease and left ventricular dysfunction (1). The second Multicenter Automatic Defibrillator Implantation Trial (MADIT-II) and the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) have shown that implantable cardioverter-defibrillators (ICDs) reduce mortality in patients with ischemic heart disease and left ventricular ejection fractions (LVEF) ≤30% and ≤35%, respectively (2,3). Although ICD implantation may be cost effective overall (4–7), widespread use of ICDs is likely to be constrained by fixed health care budgets. Further risk stratification of the pool of patients considered eligible for ICDs would have broad implications for the healthcare system.
Microvolt T-wave alternans (MTWA) involves the detection of every-other-beat alternations in T-wave morphology. These alternations are thought to represent abnormalities in intracellular calcium handling that may predispose patients to ventricular tachyarrhythmias (8–10). Earlier studies have suggested that MTWA may be a useful, noninvasive method for discriminating between patients with ischemic cardiomyopathy that are at low and high risk for SCD (11,12). Although provocative, these studies were primarily limited in their ability to adjust for baseline differences in clinical characteristics (such as age and LVEF) between patients who tested MTWA negative and non-negative (positive and indeterminate) (13). Without adjusting for such baseline differences, it is unclear whether MTWA adds additional prognostic value over traditional clinical characteristics that are often associated with SCD (13,14).
We therefore evaluated whether MTWA was an independent predictor of mortality in patients with ischemic cardiomyopathy after adjusting for demographics and clinical characteristics including ICD status, age, LVEF, prolonged QRS duration, clinical comorbidities, and medication treatment. We also determined if the prognostic utility of MTWA was related to its ability to predict arrhythmic versus nonarrhythmic mortality.
A multicenter prospective cohort composed of four outpatient cardiology clinics was developed and coordinated by the Ohio Heart and Vascular Center and the Lindner Clinical Trials Center, Cincinnati, Ohio. Consecutive patients with varying LVEF were enrolled between March 2001 and June 2004 during their MTWA test (index date). For this study, we included patients with ischemic heart disease (defined as cardiac catheterization with ≥70% stenosis in at least one coronary vessel, documented myocardial infarction, or a history of coronary revascularization) and LVEF ≤35%. Patients also had to be at least 21 years old, have no history of a prior ventricular arrhythmic event, and be in sinus rhythm at the time of MTWA testing. All patients gave informed consent to registry enrollment and follow-up. The study was approved by the institutional review board at The Christ Hospital, Cincinnati, Ohio.
MTWA testing protocol
At study enrollment, all patients underwent baseline MTWA testing by treadmill exercise (Heartwave System, Cambridge Heart, Bedford, Massachusetts) with elevation of the heart rate to a target level of 120 beats/min. Beta-blockers and non-dihydropyridine calcium-channel blockers were withheld for >24 h before the test. All MTWA tests were interpreted according to standard criteria by an expert reader blinded to patient characteristics and clinical outcomes (15). A positive MTWA test was defined as sustained alternans with an onset heart rate of ≤110 beats/min. A negative MTWA test was defined as the absence of criteria for a positive test with a maximum heart rate of ≥105 beats/min. All other tests were classified as indeterminate. Based on earlier studies (11,12,14,16), we classified both indeterminate and positive tests as “non-negative” during statistical analyses, but also examined MTWA-positive and -indeteminate results separately during secondary analyses.
We collected patient data on demographic and clinical characteristics at the time of enrollment, including: age, gender, LVEF, QRS duration >120 ms, diabetes mellitus, hypertension, symptomatic heart failure, chronic obstructive pulmonary disease, chronic renal insufficiency, peripheral vascular disease, and history of myocardial infarction, stroke, transient ischemic attack, atrial fibrillation, unexplained syncope, or revascularization therapy. Data on baseline medication use of aspirin, angiotensin-converting enzyme inhibitor, angiotensin-receptor blocker, beta-blocker, digoxin, diuretic, class I or III antiarrhythmic agent, statin, and spironolactone were also obtained.
We also collected data on diagnostic testing with Holter monitoring and ICD implantation. We defined nonsustained ventricular tachycardia on Holter monitoring as >100 beats/min for three or more consecutive beats and <30 s. The decision for ICD implantation, including the type of device and programming, was at the discretion of the treating physician.
Primary end points and follow-up
The primary end point for the study was all-cause mortality. Secondary endpoints included cause-specific mortality, individual comparisons between the MTWA-positive, -indeterminate, and -negative groups, and the delivery of appropriate ICD shocks in patients with ICDs. Cause-specific mortality was determined by two study team members blinded to the decedent’s MTWA status and was classified as arrhythmic or nonarrhythmic in etiology using a modified Hinkle-Thaler system (17). Arrhythmic deaths included unwitnessed deaths (if stable when last observed before death), witnessed instantaneous deaths, and deaths as a sequelae of cardiac arrest. In patients with ICDs, ICD shocks were reviewed by a physician blinded to MTWA status to determine their appropriateness. Clinical follow-up for end points was achieved by quarterly office visits, telephone contact with patients, review of office charts, and by an annual query of the National Death Index (18,19).
Because ICDs have been shown to be efficacious in reducing mortality in this population, stratified Cox proportional hazards analyses by ICD status were planned for the study outcomes if no significant interaction existed between the ICD and MTWA variables. The rationale for stratified analyses was to ensure appropriate adjustment of a variable known to significantly affect mortality in patients with ischemic cardiomyopathy (20).
Patients in the study cohort were first divided into those with and without ICDs. Baseline characteristics between those who tested MTWA negative and non-negative in the ICD and non-ICD groups were compared using Student ttests for continuous variables and chi-square tests for categorical variables. Survival curves between the MTWA-negative and -non-negative groups stratified by ICD status were constructed using Kaplan-Meier estimates, and the unadjusted relationship of MTWA testing with mortality was assessed with the stratified log-rank test.
Stratified multivariable Cox proportional hazards modeling was then performed to assess the independent relationship of MTWA with event-free survival. Candidate variables included age, gender, LVEF, QRS duration >120 ms, clinical comorbid conditions (diabetes mellitus, hypertension, symptomatic heart failure, chronic obstructive pulmonary disease, chronic renal insufficiency, peripheral vascular disease, and history of myocardial infarction, stroke, transient ischemic attack, atrial fibrillation, unexplained syncope, or revascularization therapy), and medication treatment (aspirin, angiotensin-converting enzyme inhibitor or angiotensin-receptor blocker, beta-blocker, digoxin, diuretic, class I or III antiarrhythmic agent, statin, and spironolactone). Covariates associated with survival in univariate analyses (p ≤ 0.10) were entered in a stepwise fashion based on a significance level of p ≤ 0.05. Age and LVEF were included in the final model regardless of the level of significance. Potential two-way interactions between covariates were also examined, including an interaction between MTWA result and ICD status to determine if the hazard ratio for all-cause mortality with MTWA differed by ICD group and to ensure that the stratified analysis would be appropriate (20).
Additional analyses were performed to assess whether there were important differences in hazard ratios across prespecified subgroups in the cohort. Consistent with the MADIT-II trial inclusion criteria, we performed stratified multivariable analyses for only those patients with LVEF ≤30%. To determine how the inclusion of patients with indeterminate MTWA results in the MTWA-non-negative group affected our results, we repeated our analyses comparing only those patients who tested MTWA positive or indeterminate with those who tested MTWA negative. Finally, for the subgroup of patients with Holter monitor test results, stratified Cox proportional hazards models were constructed with the addition of nonsustained ventricular tachycardia to the candidate variables.
Finally, we also assessed whether MTWA was an independent predictor of arrhythmic mortality, nonarrhythmic mortality, and appropriate ICD shock therapy (in the ICD group only). In all models, the assumption of proportionality for the Cox proportional hazards models was visually assessed with log-log (survival) versus log (survival time) to ensure parallelism. All statistical analyses were performed with SAS version 9.2 (SAS Institute, Cary, North Carolina).
Baseline characteristics and unadjusted analyses
Of the 768 patients, 514 (67%) had a non-negative MTWA test. Mean follow-up for the entire cohort was 18 ± 10 months, and was similar across MTWA subgroups. Table 1depicts baseline characteristics of the study cohort stratified by ICD status. When compared with patients who tested MTWA negative, patients without ICDs with non-negative MTWA results were significantly older, had lower ejection fractions, were more likely to have evidence of nonsustained ventricular tachycardia on Holter monitoring, were on digoxin at a higher rate, were less likely to be on statins, and were less likely to have a history of myocardial infarction. Patients with ICDs with non-negative MTWA were significantly older and more likely to have diabetes mellitus.
In the non-ICD group, there were 58 deaths (MTWA negative: 15 [8.4%]; MTWA non-negative: 43 [21.8%]), of which 28 were arrhythmic deaths (MTWA negative: 6 [3.4%]; MTWA non-negative: 22 [11.2%]). In the ICD group, there were 41 deaths (MTWA negative: 6 [8.0%]; MTWA non-negative: 35 [11.0%]), of which 14 were arrhythmic deaths (MTWA negative: 3 [4.0%]; MTWA non-negative: 11 [3.5%]) (Table 2).There were also a total of 26 appropriate ICD therapies not associated with death in the ICD group.
Kaplan-Meier event-free survival curves for all-cause mortality comparing those who tested MTWA negative and non-negative, stratified by ICD status, are shown in Figure 1.Event-free survival for MTWA-negative patients was significantly higher than for MTWA-non-negative patients (stratified log-rank statistic = 13.50; p = 0.0002).
Multivariable Cox proportional hazard models
Patients with a non-negative MTWA result had a higher risk of all-cause mortality after multivariable adjustment (stratified hazard ratio [HR]=2.24 [95% confidence interval [CI] 1.34 to 3.75]; p = 0.002) (Table 3).Although a number of interaction terms were examined, none reached statistical significance. Importantly, an HR associated with a non-negative MTWA test did not significantly differ by ICD status (p = 0.47 for interaction between ICD status and MTWA test result). When the results were examined separately for the ICD and non-ICD groups, the hazard ratios for a MTWA-non-negative test were not much different (HR = 1.79 for ICD patients and 2.27 for non-ICD patients), but the findings in the ICD group were not statistically significant. Results were equally robust when only patients with LVEF ≤30% were considered (n = 537; stratified HR = 2.10 [95% CI 1.18 to 3.73] for a non-negative MTWA test; p = 0.01) (Table 3). A non-negative MTWA test also showed a trend toward higher mortality among patients with LVEF between 31% and 35% (n = 231; stratified HR = 2.99 [95% CI 0.93 to 9.68]; p = 0.067). Finally, when we examined only those patients who also had Holter tests performed, a non-negative MTWA remained a significant predictor after adjusting for the presence of nonsustained ventricular tachycardia in addition to other candidate variables (n = 514; stratified HR = 2.31 [95% CI 1.28 to 4.17]; p = 0.005).
Cause-specific mortality and appropriate ICD shocks
Stratified Cox proportional hazards analyses found that a non-negative MTWA test was an independent predictor of arrhythmic mortality (stratified HR = 2.29 [95% CI 1.00 to 5.24]; p = 0.049). However, no significant relationship with MTWA status was found for nonarrhythmic mortality (stratified HR = 1.77 [95% CI 0.84 to 3.74]; p = 0.13) (Table 3). In patients with ICDs, a non-negative MTWA showed a trend toward higher risk for appropriate ICD shocks not associated with death (HR = 3.79 [95% CI 0.88 to 15.91]; p = 0.07). When a combined end point of all-cause mortality and ICD shocks was assessed, a non-negative MTWA was an independent predictor of event-free survival in the ICD group (HR = 2.42 [95% CI 1.07 to 5.48]; p = 0.035).
Comparisons of MTWA-positive or -indeterminate patients with MTWA-negative patients
Of those classified as MTWA non-negative (n = 514), 159 tested MTWA indeterminate. Reasons for an indeterminate test included frequent ectopy (n = 63 [40%]), inability to reach adequate heart rate for MTWA assessment (n = 59 [37%]), excessive background noise (n = 22 [14%]), and unsustained alternans (n = 15 [9%]). When the analysis was limited to only those patients who tested MTWA positive or negative (i.e., indeterminate MTWA patients excluded), a positive MTWA was associated with higher all-cause mortality (n = 609; stratified HR = 2.08 [95% CI 1.18 to 3.66]; p = 0.01) and a trend toward higher arrhythmic mortality (stratified HR = 2.03 [95% CI 0.83 to 4.97]; p = 0.13) (Table 4).Similarly, comparisons between patients who tested MTWA indeterminate and negative showed that an indeterminate MTWA was associated with higher rates of all-cause (n = 413; stratified HR = 2.78 [95% CI 1.55 to 4.99]; p = 0.0006) and arrhythmic (stratified HR = 3.62 [95% CI 1.44 to 9.13]; p = 0.006) mortality. However, comparisons between patients who tested MTWA positive and indeterminate did not show significant differences in all-cause or cause-specific mortality.
This study is the largest series to date to examine the predictive ability of MTWA to risk-stratify patients with ischemic heart disease and left ventricular dysfunction. We found that patients with ischemic cardiomyopathy and a non-negative MTWA test result have a more than two-fold higher risk of all-cause mortality when compared with patients with a negative MTWA result. Although one could argue that the presence of MTWA may simply represent a composite risk profile from a patient’s joint distribution of clinical and comorbid conditions, our study found that MTWA remained an independent predictor even after controlling for demographics, clinical comorbidities, ICD status, and medication treatment. Importantly, we also demonstrated that the increased risk associated with a non-negative MTWA result is mediated primarily through higher rates of arrhythmic mortality.
Prior studies of MTWA testing in patients with ischemic cardiomyopathy have reported more impressive results than the adjusted hazard ratio in this study (11,12). An analysis of 129 patients who met MADIT-II study criteria from two prospective studies found a two-year arrhythmic mortality rate of 0.0% in those who tested MTWA negative and 15.6% in those who tested non-negative (12). Another recent study prospectively followed 177 MADIT-II–eligible patients for two years and found that those who tested MTWA non-negative had a hazard ratio of 4.8 (p = 0.02 vs. MTWA negative) for all-cause mortality (11). In those studies, however, potential differences in baseline characteristics between patients with non-negative and negative MTWA results were not reported or adjusted for during statistical analyses (13). Moreover, the mean age of the study populations was younger (61 and 63 years, respectively) than that of our study population (67 years). Finally, patients in our study were recruited from an unselected outpatient population as opposed to a hospitalized setting, so there may have been differences in baseline SCD risk between study populations.
A recent meta-analysis of MTWA found that a non-negative MTWA test result was associated with a univariate risk ratio for cardiac arrhythmic events of 2.42 (95% CI 1.30 to 4.50) in patients with ischemic cardiomyopathy (14). However, the authors reported that the major limitation of most studies examining MTWA in various patient populations has been the lack of adjustment for known risk factors (such as age and LVEF) or other potential confounders. This has prevented a true assessment of the incremental prognostic utility of MTWA in predicting mortality (13,14). Moreover, many prior studies have reported cardiac arrhythmic events as their primary outcome without providing analyses for all-cause mortality. This leaves open the possibility that competing risks (whereby a non-negative MTWA test predicts arrhythmic mortality but not all-cause mortality) or misclassification bias (for cause-specific mortality) may have confounded earlier results (13). To our knowledge, our study is the first to assess the prognostic utility of MTWA for both all-cause and arrhythmic rates of mortality after adjusting for baseline differences in age and LVEF, as well as for ICD status, medication usage, and clinical variables.
Because of the large sample size in our study, we were able to assess whether an MTWA-indeterminate test was an independent predictor of outcome. Prior studies have combined MTWA-positive and -indeterminate patients into an MTWA-non-negative category without being able to adequately assess whether MTWA-indeterminate patients had similar mortality hazard ratios as MTWA-positive patients when compared with MTWA-negative patients. To our knowledge, our study is the first to specifically examine this in patients with ischemic cardiomyopathy, and we found no significant differences in all-cause or cause-specific mortality risks between the MTWA-indeterminate and -positive groups, suggesting that our decision to combine them into one “non-negative” category in this study was statistically appropriate. In fact, the mortality hazard ratio was nonsignificantly higher for MTWA-indeterminate patients than for MTWA-positive patients. Because the majority of patients (77%) had frequent ectopy or inability to reach adequate heart rate as the primary reason for an indeterminate MTWA result, this may reflect underlying heart substrates or physical deconditioning that may predispose MTWA-indeterminate patients to similar or higher mortality rates than MTWA-positive patients.
Although our study found that MTWA was an independent predictor for all-cause and arrhythmic mortality in the non-ICD group and for the entire study population, there were no significant mortality differences found in the ICD group. This is largely because arrhythmic rates of mortality were much lower for the ICD group (3.4%) than for the non-ICD group (7.6%). As a sensitivity test, when a combined mortality and appropriate ICD shock end point was the outcome for the ICD group, a non-negative MTWA was an independent predictor of events (HR = 2.42 [95% CI 1.07 to 5.48]). This suggests that patients with ICDs who tested MTWA non-negative in our study were exposed to higher baseline mortality and arrhythmic event rates than those who tested MTWA negative.
The potential implication of MTWA testing in patients at high risk for SCD is far reaching. The Centers for Medicare and Medicaid Services has recently extended coverage for prophylactic ICD implantation to the SCD-HeFT– and MADIT-II–eligible population (21). However, this comes at considerable financial cost, and it is likely that ICD therapy will continue to be withheld from many “appropriate” candidates because of barriers to treatment, including lack of hospital and physician resources or a perception that the existing criteria are too broad (22,23). The ability to more selectively target devices in high-risk individuals using a screening test such as MTWA might lead to more complete treatment of those who are most likely to benefit (6,24). Our findings raise the possibility that, in patients with ischemic cardiomyopathy and without prior history of ventricular arrhythmia, those who test MTWA non-negative will have a higher absolute and relative risk of mortality from arrhythmic deaths compared with those who test MTWA negative and may therefore derive more benefit with ICD therapy.
Our findings should be interpreted in the context of the following study design issues. First, there is the potential for residual confounding as with all cohort studies. We did not have patient-level data on certain variables, including New York Heart Association functional class and laboratory data, which may have affected our results. A particular strength of our study, however, was our ability to adjust for ICD status and a number of demographic, clinical, and treatment variables at a level not previously done. Moreover, MTWA can be performed only in patients with sinus rhythm. This is likely to affect only a minority of patients, as only 8% to 15% of patients in the MADIT-II and SCD-HeFT studies were noted to be in atrial fibrillation or atrial flutter.
Microvolt T-wave alternans is a strong and independent predictor of all-cause and arrhythmic mortality in patients with ischemic cardiomyopathy. Microvolt T-wave alternans adds incremental prognostic utility to presently known risk factors for mortality in this population, with significant implications for future risk stratification models.
This study was funded in part by Medtronic. Dr. Chan is supported by a National Institutes of Health Cardiovascular Multidisciplinary Research Training Grant and by the Ruth L. Kirchstein Research Service Award.
- Abbreviations and Acronyms
- confidence interval
- hazard ratio
- implantable cardioverter-defibrillator
- left ventricular ejection fraction
- second Multicenter Automatic Defibrillator Implantation Trial
- microvolt T-wave alternans
- sudden cardiac death
- Sudden Cardiac Death in Heart Failure Trial
- Received August 16, 2005.
- Revision received October 25, 2005.
- Accepted November 1, 2005.
- American College of Cardiology Foundation
- Greenberg H.,
- Case R.B.,
- Moss A.J.,
- Brown M.W.,
- Carroll E.R.,
- Andrews M.L.
- Lab M.J.,
- Lee J.A.
- Shimizu W.,
- Antzelevitch C.
- Armoundas A.A.,
- Tomaselli G.F.,
- Esperer H.D.
- Bloomfield D.M.,
- Steinman R.C.,
- Namerow P.B.,
- et al.
- Gehi A.K.,
- Stein R.H.,
- Metz L.D.,
- Gomes J.A.
- Hohnloser S.H.,
- Klingenheben T.,
- Bloomfield D.,
- Dabbous O.,
- Cohen R.J.
- Hinkle L.E. Jr..,
- Thaler H.T.
- Chan P.S.,
- Hayward R.A.
- (2003) National Death Index User’s Manual (Department of Vital Statistics, Centers for Disease Control, Hyattsville, MD).
- Hosmer D.,
- Lemeshow S.
- ↵Phurrough S, Salive M, Baldwin J, Chin J. Decision Summary: Implantable Cardioverter-Defibrillators (CAG 00157R2). Available at: http://www.cms.hhs.gov/mcd/viewncd.asp?ncd_id=20.4&ncd_version=2&basket=ncd:20.4:2:Implantable+Automatic+Defibrillators. Accessed December 28, 2004