Author + information
- Received January 12, 2001
- Revision received May 25, 2001
- Accepted June 15, 2001
- Published online October 1, 2001.
- Luis A Pires, MD, FACC∗,* (, )
- Michael H Lehmann, MD, FACC†,
- Alfred E Buxton, MD, FACC‡,
- Gail E Hafley, MS§,
- Kerry L Lee, PhD§,
- the Multicenter Unsustained Tachycardia Trial Investigators
- ↵*Reprint requests and correspondence:
Dr. Luis A. Pires, Cardiac Electrophysiology, St. John Hospital and Medical Center, 22101 Moross Road, Detroit, Michigan 48236
The goal of this study was to describe the influence of the clinical setting (in-hospital vs. out-of-hospital) in which nonsustained ventricular tachycardia (NSVT) is discovered on the rate of inducibility of sustained ventricular tachycardia (VT), arrhythmic events and survival in patients with coronary artery disease (CAD) and left ventricular (LV) dysfunction.
In-hospital presentation of sustained VT is independently associated with lower long-term overall survival. The impact of the clinical setting in which NSVT is documented is unknown.
In the Multicenter Unsustained Tachycardia Trial (MUSTT), designed to assess the benefit of randomized antiarrhythmic therapy guided by electrophysiologic testing in patients with asymptomatic NSVT, CAD and LV dysfunction, eligible patients were enrolled irrespective of the setting in which the index arrhythmia was discovered. In this retrospective analysis, we compared the rate of VT inducibility and outcome of MUSTT-enrolled patients with in-hospital versus out-of-hospital presentation of NSVT.
Monomorphic sustained VT was induced in 35% and 28% of the patients whose index NSVT occurred in-hospital and out-of-hospital, respectively (adjusted p = 0.006). Cardiac arrest or death due to arrhythmia at two- and five-year follow-ups were 14% and 28% for untreated patients with in-hospital-identified NSVT and 11% and 21% for the out-of-hospital group (adjusted p = 0.10). Overall mortality rates at two- and five-year follow-ups were 24% and 48% for inpatients and 18% and 38% for outpatients (adjusted p = 0.018). In patients randomized to antiarrhythmic therapy, there was no significant interaction between patient status (in-hospital vs. out-of-hospital) and treatment impact on the rates of total mortality (p = 0.98) and arrhythmic events (p = 0.08).
In patients with CAD and impaired LV function, asymptomatic NSVT identified in-hospital, compared with that identified out-of-hospital, is associated with a higher rate of induction of sustained VT and overall mortality. Therefore, in similar patients, the clinical setting in which NSVT is discovered should be taken into account when formulating patient risk, treatment and clinical trial design.
The presence of nonsustained ventricular tachycardia (NSVT) in patients with coronary artery disease (CAD) and left ventricular (LV) dysfunction is associated with a significantly increased risk for sudden and nonsudden deaths (1–5). Several factors have been identified as important prognostic markers in patients with asymptomatic NSVT (6); however, the importance of the clinical setting (in-hospital vs. out-of-hospital) in which NSVT is documented has not been evaluated in these patients. In patients with sustained ventricular tachyarrhythmias (VT), in-hospital presentation has been shown to be a strong and independent predictor of lower long-term survival (7). For patients who have not yet developed but are at increased risk for life-threatening ventricular arrhythmias, the identification of any additional risk factor might be of even greater significance. Out-of-hospital documentation of NSVT in a stable patient may be a more specific marker of future arrhythmic events, whereas a similar in-hospital identified arrhythmia in a potentially sicker patient may be a less specific indicator of future events. On the other hand, the overall outcome of hospitalized patients may be adversely affected by added comorbidities. Such differences would have important implications with regard to risk stratification, treatment and future trial design involving similar patients.
When the Multicenter Unsustained Tachycardia Trial (MUSTT) was designed, eligible patients with asymptomatic NSVT were enrolled irrespective of where the index arrhythmia was discovered (8). This study describes the influence of the clinical setting (in-hospital vs. out-of-hospital) of the index NSVT on the rates of induction of sustained VT, arrhythmic event rates and total mortality and the impact of inpatient and outpatient status on the treatment outcome of patients enrolled in MUSTT.
The MUSTT investigation was designed to test the benefit of antiarrhythmic treatment (drugs and defibrillator) guided by electrophysiologic testing in reducing the risk of sudden death and cardiac arrest among patients with CAD, LV dysfunction and asymptomatic NSVT. The complete study protocol and results have been published (8–10). At the time of enrollment, details about the characteristics, mode of documentation (e.g., hospital telemetry, Holter monitor) and clinical setting of each patient’s index NSVT were collected. For the current analysis, we reviewed the records of enrolled patients and determined the clinical setting (in-hospital vs. out-of-hospital) in which the patients’ index arrhythmias were discovered. First, any patient hospitalized during the time in which the index NSVT was documented (by any recording method) or whose NSVT was recorded on hospital telemetry was classified as “in-hospital.” Second, any patient with Holter documented NSVT and known not to be hospitalized at the time was classified as “out-of-hospital.” Third, patients whose hospitalization status was not known were classified based on the location (inpatient or outpatient) in which they were initially contacted/screened. Patients who could not be classified based on these steps were excluded from analysis.
Differences in clinical characteristics, VT inducibility during baseline electrophysiologic testing and medical and antiarrhythmic treatment at the time of hospital discharge between patients with in-hospital versus out-of-hospital index NSVT were compared by the Wilcoxon rank-sum test (for continuous variables), the chi-square test or Fisher exact test (for categorical variables). Values for continuous and categorical variables are presented, respectively, as medians (with 25th and 75th percentiles) and percentages. Cumulative event rates were estimated using the Kaplan-Meier method, with the day of patient enrollment as time zero. Adjusted relative risks, expressed as hazard ratios (with 95% confidence intervals), were based on the Cox proportional-hazards model. For VT inducibility, covariates previously identified to be independently associated with inducible sustained VT (11)were used in the model and included history of myocardial infarction (MI), gender, race, angina within six months of enrollment, prior coronary artery bypass graft surgery, other cardiac diseases, LV dyskinesis and the use of beta-adrenergic blocking agents. Similarly, in assessing for arrhythmic events and total mortality, variables found to have an independent impact on outcome were used in the adjusted analysis. The covariates included age, race, gender, prior coronary bypass surgery and percutaneous coronary intervention, prior MI, number of stenotic coronary arteries, intraventricular conduction abnormalities or left bundle branch block, the use of digitalis at baseline, treatment with beta-blockers at hospital discharge, longest episode of NSVT and induction of randomizable VT (monomorphic sustained VT induced by any method of stimulation or polymorphic sustained VT, ventricular flutter and fibrillation induced by one or two extrastimuli).
A total of 2,202 patients were enrolled in the study. Data regarding the clinical setting in which the index NSVT was discovered was available in 2,190 patients (99.5%) and missing in 12 (0.5%). Of the 2,190 patients, 1,631 (74.5%) had their index arrhythmia identified in-hospital, and 559 (25.5%) had their index arrhythmia identified out-of-hospital. Inducibility of sustained VT was compared in all 2,190 patients, and, to avoid potential confounding impact of antiarrhythmic treatment, long-term arrhythmic events (nonfatal sustained VT, cardiac arrest and death due to arrhythmias) and total mortality were assessed among only the 1,780 untreated patients (1,299 in-hospital and 481 out-of-hospital). Untreated patients included those with inducible VT randomized to no antiarrhythmic therapy plus noninducible patients followed in a registry. For patients randomized to antiarrhythmic treatment, we used the Cox model to assess the impact of inpatient versus outpatient status on treatment outcome.
Patients whose index NSVT was discovered in-hospital were older and less often Caucasians than those with out-of-hospital-identified arrhythmia (Table 1). Patients with in-hospital-identified NSVT had greater numbers of diseased coronary arteries, more often had history of congestive heart failure (CHF) and had a more advanced New York Heart Association functional class and a shorter time from most recent MI to enrollment, but the two groups had similar mean LV ejection fractions (0.29 vs. 0.30). The index NSVT was longer (5 complexes vs. 4 complexes, p < 0.001) and had a shorter cycle length (410 ms vs. 440 ms, p ≤ 0.001) in the group with in-hospital-identified as opposed to out-of-hospital-identified NSVT (Table 1). These same patient characteristics and differences were also noted when analysis was confined to the untreated patients.
Baseline arrhythmia inducibility
Baseline electrophysiologic testing was performed 7 days and 42 days (p < 0.001) after discovery of the index NSVT in patients with in-hospital- and out-of-hospital-documented arrhythmia, respectively. Of the total 2,190 patients, nearly half in each group had induction of some form of sustained VT. In-hospital patients more often had inductions of monomorphic sustained VT (35% vs. 28%; unadjusted p = 0.008, adjusted p = 0.006) and any randomizable VT (38% vs. 32%; unadjusted p = 0.02, adjusted p = 0.01) compared with patients that had out-of-hospital documented arrhythmia (Table 2).
Medical and antiarrhythmic therapy
At the time of hospital discharge, more of the untreated patients with in-hospital-identified NSVT were receiving digitalis (58% vs. 50%, p = 0.002), diuretics (60% vs. 53%, p = 0.008) and nitrates (45% vs. 36%, p = 0.001). Fewer in-hospital patients (60% vs. 65%, p = 0.049) were receiving aspirin, and similar numbers of in-hospital and out-of-hospital patients were receiving angiotensin-converting enzyme inhibitors, beta-blockers and calcium channel blockers (Table 3). At the last follow-up, there were no substantial changes in the proportion (and group differences) of patients receiving these medications. Fewer untreated out-of-hospital patients (0.9% vs. 3.1%, p = 0.01) were discharged on antiarrhythmic drugs, and an equal but small number received a defibrillator at the time of hospital discharge (0.2% vs. 0.6%, p = 0.46) or at any time after enrollment (6.3% vs. 7.3%, p = 0.46).
Of the 1,780 untreated patients, those with in-hospital-identified NSVT were followed for a median of 39 months versus 42 months for the out-of-hospital group (p < 0.001). The two- and five-year rates for cardiac arrest or death due to arrhythmia were 14% and 28%, respectively, among in-hospital patients and 11% and 21% for out-of-hospital patients (unadjusted p = 0.004 and adjusted p = 0.10) (Fig. 1and Table 4). Over the five-year follow-up, the incidence of spontaneous sustained VT that did not cause cardiac arrest or death was 6% and 5% for in-hospital and out-of-hospital patients, respectively (p = 0.42). Total mortality rates at two- and five-year follow-ups were 24% and 48% for the group with in-hospital identified NSVT and 18% and 38% for those with out-of-hospital arrhythmia (unadjusted p = 0.0001 and adjusted p = 0.018), respectively (Fig. 2and Table 4). Cardiac deaths accounted for 77% and 76% (p = 0.91) of total deaths in in-hospital and out-of-hospital patients, respectively. Deaths classified as arrhythmic were similar (approximately 50%) in the two groups.
For patients randomized to antiarrhythmic therapy, five-year arrhythmic event rates were similar between patients with in-hospital- and out-of-hospital-identified NSVT whether they were treated with antiarrhythmic drugs (30% vs. 41%, p = 0.31) or a defibrillator (8% vs. 11%, p = 0.66). Five-year total mortality rate was also similar in the two groups treated with antiarrhythmic drugs (56% vs. 48%, p = 0.37), but patients with in-hospital-identified NSVT treated with a defibrillator had a slightly greater five-year mortality rate (27% vs. 14%, p = 0.05). The clinical setting in which the index NSVT was discovered had no significant impact on treatment outcome with respect to total mortality (p = 0.98); however, treatment had a slightly greater impact in reducing arrhythmic events in in-hospital as compared with out-of-hospital patients (p = 0.08).
The major finding of this analysis is that, in patients with CAD, LV dysfunction and asymptomatic NSVT, those whose index arrhythmias are discovered in-hospital are more likely to have inducible, especially monomorphic, sustained VT and a worse long-term prognosis compared with those with out-of-hospital identified arrhythmias. Group differences in VT inducibility and overall survival persisted after adjusting for several baseline predictors. However, there was no difference in the adjusted rates of cardiac arrest and death due to arrhythmia between the groups.
Risk stratification of patients with NSVT
Several factors, most notably LV function, have been used to identify high-risk patients with CAD and NSVT (6). Previous studies, which included patients with either in-hospital- (the majority) or out-of-hospital-identified NSVT, did not evaluate the impact of the clinical setting in which the index arrhythmia was documented on patient outcome (1–5,12–16). In this study, we identified yet another potentially important risk-stratifying factor in patients with CAD, reduced ejection fraction and spontaneous asymptomatic NSVT. Patients with in-hospital-identified NSVT were more likely to have inducible monomorphic sustained VT (35% vs. 28%, adjusted p = 0.006) and had greater five-year total mortality rate (48% vs. 38%, adjusted p = 0.018) than patients with out-of-hospital-identified arrhythmias. These differences persisted after adjusting for several predictive baseline factors; however, other variables unaccounted for in our analysis might explain the results. The shorter median time between the most recent acute MI and enrollment in patients with in-hospital- compared with out-of-hospital-identified NSVT (Table 1)may have accounted for the greater rate of VT induction in the former group (17,18). Moreover, in-hospital patients underwent electrophysiologic testing significantly closer in time (7 vs. 42 days, p < 0.001) to their index arrhythmia (thus, closer to their most recent MI) than out-of-hospital patients. A greater number of patients with in-hospital- as compared with out-of-hospital-identified NSVT had a history of CHF (72% vs. 61%, p = 0.001) and more advanced functional class (p = 0.001). However, because relevant data regarding CHF were collected in only about half the patients, they were not included as variables in our adjusted mortality analysis (Table 4). In a separate adjusted analysis involving only the patients with relevant data, the addition of CHF status as a covariate resulted in a similar adjusted total mortality rate in patients with in-hospital- and out-of-hospital-identified NSVT (p = 0.15). This suggests that a greater prevalence of CHF could explain, at least in part, the higher overall mortality rate in patients with in-hospital-documented NSVT.
Differences in the rates of arrhythmic events could not be explained by inducibility of randomizable sustained VT alone (Table 4). This is consistent with the known influence of many of the analyzed factors, besides inducibility of sustained VT, on the outcome of various patient groups with NSVT and sustained VT (6,12–16,19). Our results do not contradict the earlier MUSTT data, which assessed arrhythmic events and overall survival according to VT inducibility rather than the clinical setting in which the index NSVT was discovered (10).
Clinical setting of NSVT and treatment outcome
Except for a greater five-year total mortality rate among inpatients compared with outpatients treated with defibrillator (27% vs. 14%, p = 0.05), we found no differences in the rates of arrhythmic events and total mortality between inpatients and outpatients treated with either antiarrhythmic drugs or a defibrillator. The similar outcome in treated patients with NSVT diagnosed in-hospital and those with NSVT diagnosed out-of-hospital, unlike untreated patients (Table 4), are likely due to reduced analysis power (reflected in the far smaller number of treated patients) and differences between treated and untreated patients. Whereas the untreated group included those with inducible VT randomized to no antiarrhythmic therapy plus noninducible registry patients, treated patients included, by definition, only those with inducible randomizable VT (9). We also found no evidence suggesting a significant interaction between the setting of the index NSVT and treatment outcome. The apparent slightly greater treatment impact on arrhythmic events in those identified in-hospital can probably be explained by the fact that slightly more inpatients (50% vs. 41%) received a defibrillator, which is consistent with the positive impact of a defibrillator, but not antiarrhythmic drug therapy in treated patients enrolled in MUSTT (9). Therefore, based on current data, attempts at risk stratification and treatment should not be withheld for the lower-risk patients with out-of-hospital-identified NSVT.
Comparison with patients with sustained VT
Our finding of a higher total mortality rate in patients with in-hospital- compared with out-of-hospital-identified NSVT is in agreement with data involving registry patients enrolled in the Antiarrhythmic Versus Implantable Defibrillator (AVID) study, in which in-hospital presentation of sustained VT/ventricular fibrillation was associated with a nearly twofold greater adjusted, two-year mortality rate (7). As with this study population, in-hospital patients enrolled in the AVID trial were apparently sicker than their out-of-hospital counterparts, but survival differences between the groups persisted after adjusting for several baseline mortality predictors (including history of CHF, diabetes and atrial fibrillation) and therapies at the time of hospital discharge. However, considering the importance of primary prevention of sudden cardiac death, which requires proper patient selection based on level of arrhythmic risk, the importance of the clinical setting in which NSVT is identified may be of broader significance when applied to asymptomatic patients who are at risk of developing life-threatening VTs.
With regard to study limitations, foremost is the fact that the patients represent a highly selected group, and the findings are applicable only to those with CAD, LV ejection fraction ≤0.40 and asymptomatic NSVT. Considering that greater than two-thirds of the patients had their index arrhythmia identified in-hospital, there may have been considerable bias in patient selection. In addition to being sicker and, thus, hospitalized, in-hospital patients were probably easier to identify and recruit. Second, in addition to limited data about CHF history, information about other comorbidities, such as history of hypertension and diabetes, were not available, nor were data regarding the primary admitting diagnoses (e.g., cardiac vs. noncardiac) and hospital course of hospitalized patients with in-hospital-documented NSVT collected.
This study shows that, among patients with CAD, LV dysfunction and asymptomatic NSVT, VT inducibility, arrhythmic events and patient survival are greatly influenced by the clinical setting in which the index arrhythmia was discovered. In-hospital documentation of NSVT was independently associated with a greater likelihood of inducibility of sustained VT and lower overall survival. However, the higher arrhythmic event rates seem to reflect differences in the group’s clinical characteristics. Since assigning risk and, thus, antiarrhythmic treatment for this group of patients is based first on the patient’s clinical profile and second on the result of electrophysiologic testing (10), both of which are influenced by the clinical setting of the patient’s index arrhythmia, the setting in which NSVT is discovered should be considered in the decision. It might also be useful to aggressively identify and treat comorbidities (e.g., CHF) in such patients. Future investigations should focus on identifying and, whenever possible, correcting additional factors that may increase a patient’s risk. Furthermore, the clinical setting in which NSVT is documented should also be taken into account when deciding on which patients (and perhaps their individual proportions) to enroll in future clinical trials involving patients similar to those enrolled in MUSTT. Proper interpretation of treatment outcome data requires comparison of patients with a truly similar clinical profile.
☆ Supported by grants from the National Heart, Lung, and Blood Institute (UO1HL45700 and UO1 HL45726), CR Bard, Berlex Laboratories, Boehringer-Ingelheim Pharmaceuticals, Guidant, Knoll Pharmaceutical, Medtronic, Merck, Searle, Ventritex, St. Jude Medical and Wyeth-Ayerst Laboratories.
- Antiarrhythmic Versus Implantable Defibrillator study
- coronary artery disease
- congestive heart failure
- left ventricle or left ventricular
- myocardial infarction
- Multicenter Unsustained Tachycardia Trial
- nonsustained ventricular tachycardia
- ventricular tachycardia
- Received January 12, 2001.
- Revision received May 25, 2001.
- Accepted June 15, 2001.
- American College of Cardiology
- Anderson K.P,
- DeCamilla J,
- Moss A.J
- Kleiger R.E,
- Miller J.P,
- Thanavaro S,
- et al.
- Bigger J.T,
- Weld F.M,
- Rolnitzky L.M
- Denes P,
- Gillis A.M,
- Pawitan Y,
- et al.
- Epstein A.E,
- Powell J,
- Yao Q,
- et al.
- Buxton A.E,
- Hafley G.E,
- Lehmann M.H,
- et al.
- Gomes J.A.C,
- Hariman R.I,
- Kang P.S,
- et al.
- Buxton A.E,
- Marchlinski F.E,
- Flores B.T,
- et al.
- Wilber D.J,
- Olshansky B,
- Moran J.P,
- et al.