Prognosis and Risk Stratification: Clinical Indications and Results
Iwona Cygankiewicz
KEY POINTS
The highest risk of sudden cardiac death (SCD) is attributed to survivors of cardiac arrest, and these patients are qualified for implantable cardioverter-defibrillator (ICD) in secondary prevention.
SCD may result from ventricular tachycardia/fibrillation (VT/VF), bradyarrhythmias or pulseless electrical activity.
Risk prediction is focused mainly on predicting tachyarrhythmias because those are “shockable” and cardiac arrest can be prevented by ICD.
Prolonged monitoring is recommended in a variety of cardiac and noncardiac conditions.
Coronary artery disease is the most common condition associated with SCD.
Cardiac channelopathies constitute a heterogeneous group of inherited diseases such as long and short QT syndromes, Brugada syndrome (BS), catecholaminergic polymorphic ventricular tachycardia, early repolarization syndrome, and idiopathic VF. Phenotype ECG presentation in these diseases is variable; therefore, ambulatory ECG monitoring (AECG) monitoring may be useful in the assessment of transient ECG changes suggestive of a disease.
Patients with previously recognized heart diseases are known to show the highest risk of arrhythmic events.
After nearly 40 years of experience with ICD in humans, we are still far from being able to estimate SCD risk and identify patients who will benefit from ICD therapy.
INTRODUCTION
Sudden cardiac death (SCD) is a result of the interplay between genetic predisposition, cardiac substrate, coexisting comorbidities, and environmental factors. Its prediction and prevention remain one of the major concerns of contemporary cardiology. The incidence of SCD varies between 180 000 and 450 000 cases annually.
Differences in reported SCD incidence are based mostly on nonuniform definition of SCD in clinical studies as well as on difficulties in the determination of the exact mechanism of death.1,2 Reporting SCD mechanism becomes even more challenging given that most of these cases occur in out-of-hospital settings, with no witnesses and no postmortem autopsy studies available.
Despite a substantial decline in cardiovascular morbidity and mortality over the last few decades, survival after SCD is very low. Changes in lifestyle, diet, and pharmacotherapy may decrease the probability of SCD; nevertheless, implantation of implantable cardioverter-defibrillator (ICD) has emerged as the only strategy to prevent the occurrence of this final event. No doubts exist as to “secondary prevention” ICD implantation in cardiac arrest victims. On the other hand, selection of candidates for ICD implantation in primary prevention remains challenging.3,4,5
POPULATIONS AT RISK
The highest risk of SCD is attributed to survivors of cardiac arrest, and these patients are qualified for ICD implantation in secondary prevention. Indications for primary prevention of SCD are based on the underlying heart disease that was proven to increase the risk of arrhythmic events. As coronary artery disease remains the most common disease associated with SCD, and is responsible for approximately 80% of cases, postinfarction patients with low left ventricular ejection fraction (EF) account for the majority of primary ICD implantations. Heart failure patients with nonischemic cardiomyopathy and low LVEF constitute the second largest population “at risk.” Hypertrophic cardiomyopathy (HCM), arrhythmogenic right ventricular cardiomyopathy (ARVC), congenital heart diseases, and channelopathies are also known to be associated with an increased risk of life-threatening ventricular arrhythmias. A high incidence of SCD events is reported in a variety of noncardiac diseases such as chronic renal failure, neurologic diseases, and diabetes4,5,6 (Table 12.1). Nevertheless, as addressed by Myerburg et al, the highest absolute number of SCD events occurs in subgroups of patients with no overt known cardiovascular disease affecting cardiac structure or ion functioning.7 Therefore, given the absolute numbers of SCD events, selection of patients with low LVEF for an ICD implantation represents only the tip of the iceberg in terms of patients at risk.
MECHANISM OF DEATH
Sudden cardiac death may result from ventricular tachycardia/fibrillation (VT/VF), bradyarrhythmias or pulseless electrical activity. Small clinical studies based on Holter recordings and/or hospital telemetry showed that VT/VF was responsible for most SCD events. To date, the largest study on SCD events while wearing Holter was published by Bayes de Luna et al.8 In this study, which included Holter recordings of 157 subjects, ventricular arrhythmia degenerating to VF was observed in 62% of cases, torsade de pointes in 13%, and primary VF in 8%. Only 26 patients died from bradyarrhythmia (17%).8 Recent decades brought new data obtained from implantable devices. The Cardiac Arrhythmias and Risk Stratification after Myocardial Infarction (CARISMA) study enrolled 312 patients with a recent myocardial infarction and LVEF less than 40%. Final arrhythmias were recognized on the basis of data from implantable loop recorders (ILRs) implanted 5 to 21 days after acute myocardial infarction. During 2 years of follow-up, 26 patients with ILR died, 9 of them from sudden, 10 from nonsudden, and 7 from noncardiac deaths. ILR recordings documented that tachyarrhythmias were associated with sudden death, whereas bradyarrhythmias and electromechanical dissociation were predominant in nonsudden and noncardiac
deaths. Among SCD victims, VF was observed in 6 cases, whereas bradyarrhythmia only in one of them. On the other hand, data from ILRs showed that primary VF was the main mechanism of death in postinfarction patients with left ventricular dysfunction. Ventricular fibrillation episodes stored in ILR memory were not preceded by VT in any of the cases.9 Data on the mechanism of death in ICD recipients were reported by several investigators. Randomized trials reported SCD in 20% to 35% of deaths. In a study of Duray et al, cardiac arrhythmic death accounted for only 16% of all deaths in 882 ICD patients implanted over 10 years’ time.10
deaths. Among SCD victims, VF was observed in 6 cases, whereas bradyarrhythmia only in one of them. On the other hand, data from ILRs showed that primary VF was the main mechanism of death in postinfarction patients with left ventricular dysfunction. Ventricular fibrillation episodes stored in ILR memory were not preceded by VT in any of the cases.9 Data on the mechanism of death in ICD recipients were reported by several investigators. Randomized trials reported SCD in 20% to 35% of deaths. In a study of Duray et al, cardiac arrhythmic death accounted for only 16% of all deaths in 882 ICD patients implanted over 10 years’ time.10
TABLE 12.1 Techniques Used for Risk Stratification | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Main ICD-guidelines relevant trials are now 15 to 20 years old.4 Since then, significant evolution in pharmacotherapy and revascularization procedures has occurred, changing the clinical picture and survival of postinfarction and heart failure patients. Yet early Holter-based studies suggested that in patients with advanced heart failure, bradyarrhythmias prevail as the predominant mechanism of death. Data from the Metoprolol CR/XL Randomized Intervention Trial in-Congestive Heart Failure (MERIT HF) trial showed that in heart failure patients the mechanism of death depends on hemodynamic impairment. Patients in New York Heart Association (NYHA) Class II to III are more likely die suddenly, whereas heart failure progression is responsible for most of those in NYHA Class IV.11 ICD therapy has become widely acceptable and is the gold standard therapy to prevent the occurrence of SCD in the ventricular tachyarrhythmic mechanism. However, ICD allows for longer living and thus, paradoxically, increases the risk of development of heart failure or dying from other noncardiac diseases. Therefore, as a paradox, ICD presence may convert the risk of dying suddenly into a higher risk of death due to heart failure progression.12
METHODS FOR RISK STRATIFICATION
Risk prediction is focused mainly on predicting tachyarrhythmias because those are “shockable” and cardiac arrest can be prevented by ICD. Life-threatening arrhythmias occur more likely in subjects with underlying structural or electrical heart disease and are preceded by a chain of events encompassing the complex interplay between substrate, triggers, and modulators. Factors known to modulate and/or trigger arrhythmias include autonomic nervous system imbalance, transient ischemia, metabolic and electrolyte imbalance, transient volume overload of ventricles, or even proarrhythmic action of drugs. Therefore, risk stratification targeted toward ventricular tachyarrhythmias should take into account different aspects of “Coumel’s triangle”.3,4,13
According to current guidelines, primary prevention ICD implantation is guided by a unique risk marker—LVEF. The currently used cut-off of LVEF less than 30% to 35% is based on prespecified criteria applied in randomized clinical trials designed to evaluate the influence of ICD on survival in selected “high-risk” patients. These trials defined this cut-off on the basis of clinical data from randomized trials, indicating that LVEF less than 40% and particularly LVEF greater than 30% is a clinical covariate associated with higher overall mortality, cardiovascular mortality, and SCD.4,5 Nevertheless, risk stratification based on LVEF alone is the subject of ongoing debate. Perfect “risk stratifiers” should be reliable, reproducible, and not dependent on the operator, while statistical error in LVEF assessment is estimated at 5% to 10%. Furthermore, reduced left ventricular function is not a specific marker as it predicts both arrhythmic events and heart failure progression.14 Different techniques may be used to assess LVEF—echocardiography, cardiac magnetic resonance (CMR), or radionuclide studies. However, the latter two techniques are not applied in the clinical setting, mainly because of economic constraints. It is important to remember that current guidelines are based on randomized trials that used LVEF based on echocardiographic evaluation. Therefore, echo-based LVEF, despite its technical and clinical limitations, remains an imperfect but still very convenient parameter to assess patients’ risk.
Further, risk stratification takes into account the fact that the outcome of patients is influenced by other factors such as etiology, myocardial fibrosis, cardiac volumes, hemodynamic status, ischemia, ectopic activity, and comorbidities. These abnormalities may be evaluated by various imaging techniques, ECG-based methods to detect autonomic dysfunction and repolarization heterogeneity, invasive provocative studies, biomarkers, and, mostly likely in the future, genetic profiling (Table 12.2).
TABLE 12.2 Populations at the Highest Risk of Sudden Cardiac Death | ||
|---|---|---|
|
Surface ECG and Long-Term ECG Monitoring in Risk Stratification
Electrocardiographic evaluation includes 12-lead surface ECG, signal-averaged ECG, ambulatory Holter recording, exercise tests, and other techniques like baroreceptor sensitivity assessment. Risk markers based on surface ECG or ambulatory ECG monitoring (AECG) may reflect myocardial substrate, as well as triggers and modulators potentially contributing to life-threatening arrhythmia risk. Twelve-lead ECG provides data on specific patterns suggestive of cardiomyopathy or primary electrical diseases. Markers that reflect depolarization (QRS duration, late potentials, QRS fragmentation) and repolarization (QT interval duration, dispersion, T-wave variability, T-wave alternans) are the most commonly used for risk stratification purposes.15 A broad QRS complex is associated with an increased risk of cardiovascular mortality; however, it is not specific for SCD. Signal-averaged ECG (SAECG) reveals the presence of late potentials, reflecting heterogeneity of myocardial conduction. Even though the presence of SAECG predicted arrhythmic events in coronary patients, SAECG is nowadays used mostly as a risk marker in patients with ARVC.16 In recent years, more and more attention has been paid to QRS fragmentation. This QRS abnormality is believed to represent delay in ventricular conduction most likely because of the presence of fibrosis and myocardial fibrosis. Its prognostic value in predicting arrhythmias and ICD discharges was documented in various populations, including ischemic, nonischemic, and hypertrophic cardiomyopathy, as well as in patients with channelopathies.17
Surface ECG provides mostly data on static measures, whereas dynamic changes can be easily assessed in longer recordings. Long-term AECG monitoring allows for the evaluation of heart rate, ventricular ectopy, and parameters reflecting autonomic nervous tone as well as dynamicity of repolarization. Given that SCD is preceded by dynamic changes leading to final arrhythmia, AECG can be considered a comprehensive tool for identifying and quantifying risk factors in SCD victims. The effects of the autonomic nervous system on the heart could be evaluated by quantifying average heart rate, heart rate variability (HRV), and heart rate turbulence (HRT). Heterogeneity of repolarization can be reflected by measurements of T-wave alternans (TWA) or QT variability and dynamics.18,19,20
Frequent ventricular ectopy and episodes of nonsustained ventricular tachycardia (NSVT) are the most commonly Holter-based risk markers for risk of SCD. Increased
ventricular ectopy, defined as greater than 10 ventricular premature beats (VPBs) per hour, was one of the oldest risk stratifiers. It was documented to be associated with increased mortality rate in postinfarction patients as well as with a 6-fold higher risk of SCD within the first 6 months after myocardial infarction. Studies that included postinfarction patients treated with angioplasty and wide use of β-blockers did not confirm the results of studies performed in the thrombolytic era.21,22,23 The presence of NSVT episodes in AECG is considered a part of risk scores or indicates the need for further stratification by means of electrophysiologic study. Nevertheless, the use of NSVT as a risk marker is limited by high day-to-day variability in its occurrence. Therefore, current guidelines recommend longer, 48-hour ECG recordings to be used for risk stratification purposes.3,4,18
ventricular ectopy, defined as greater than 10 ventricular premature beats (VPBs) per hour, was one of the oldest risk stratifiers. It was documented to be associated with increased mortality rate in postinfarction patients as well as with a 6-fold higher risk of SCD within the first 6 months after myocardial infarction. Studies that included postinfarction patients treated with angioplasty and wide use of β-blockers did not confirm the results of studies performed in the thrombolytic era.21,22,23 The presence of NSVT episodes in AECG is considered a part of risk scores or indicates the need for further stratification by means of electrophysiologic study. Nevertheless, the use of NSVT as a risk marker is limited by high day-to-day variability in its occurrence. Therefore, current guidelines recommend longer, 48-hour ECG recordings to be used for risk stratification purposes.3,4,18
Heart rate variability measures are currently available in most commercial Holter systems. Evaluation of HRV parameters in risk stratification was introduced into electrocardiology in the early eighties. Heart rate variability indirectly evaluates autonomic nervous system tone based on beat-to-beat RR interval changes. Increased sympathetic tone and decreased vagal activity are correlated with progression of heart failure and a higher risk of arrhythmias. Different techniques are used for HRV assessment. Frequency and time domain as well as nonlinear techniques are the most commonly studied. Time domain measures such as standard deviation of all normal-to-normal RR intervals (SDNN), pNN50, rMSSD are usually derived from long-term recordings, whereas frequency domain and nonlinear parameters can be based on shorter recordings. The detailed methodology of HRV measurement is summarized in a report of the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (ESC/NASPE)24 and in a joint position statement by e-Cardiology Working Group of the ESC/European Heart Rhythm Association (EHRA)/Asia Pacific Heart Rhythm Society (APHRS).25 Over the last 40 years, several clinical studies have noted a prognostic role of HRV in predicting cardiovascular risk. Most of the studies documented a clear relationship between depressed HRV and increased cardiovascular mortality and heart failure progression. Several studies on patients with postinfarction and congestive heart failure linked low SDNN (the most frequently studied HRV parameter) to a high risk of overall death. Nevertheless, patients with impaired HRV at enrollment are more likely to die from heart failure progression than from arrhythmic causes, and correlation with sudden death events is rather weak or absent.26 Randomized trials in patients stratified for ICD implantation based on depressed HRV failed to demonstrate the usefulness of this measure in predicting benefit from ICD therapy.27 On the other hand, simplified evaluation of HRV based on measures from cardiac implantable electronic devices (CIED) memory has been documented as a useful tool in predicting heart failure progression and response to cardiac resynchronization therapy in heart failure patients.28
Heart rate turbulence reaction describes HRV triggered by internal stimulus such as VPB. In healthy or low-risk subjects, this baroreceptor-mediated response of the sinus node to premature ventricular beats is composed of an early acceleration and subsequent deceleration phase. In high-risk patients, HRT is weak or completely missing. As a risk stratifier, HRT combines three independent measures of risk: HRV, VPBs, and baroreceptor sensitivity. Changes in RR intervals following VPBs are subtle and require dedicated software for calculation. Detailed HRT methodology is summarized in the consensus document sponsored by the International Society for Holter and Noninvasive Electrocardiology (ISHNE).19 Abnormal HRT parameters—turbulence slope and turbulence onset—have been documented as predictors of all-cause and cardiovascular mortality in various subsets of patients, including myocardial infarction, heart failure, or other cardiac and noncardiac diseases such as diabetes, obstructive sleep apnea, or connective tissue diseases. Data from modern populations of postinfarction and heart failure patients from FINGER,
REFINE, CARISMA, MUSIC, and GISSI trials documented that abnormal HRT parameters, particularly abnormal turbulence slope, predicted total mortality and sudden death and/or arrhythmic events during follow-up.29,30,31,32 Interestingly, as shown in a combined analysis from REFINE and CARISMA studies, absence of an increase in turbulence slope (TS) values in the early postinfarction period, indicating lack of positive autonomic remodeling, was associated with a 7- to 10-fold higher risk of life-threatening arrhythmias during a 2-year follow-up.29 Abnormal HRT seems to be particularly useful in identifying high-risk patients among those with LVEF greater than 30%33,34 and is frequently evaluated in risk scores that combine assessment of arrhythmogenic substrate (low LVEF), repolarization instability (eg, TWA), and abnormal autonomic nervous system tone (HRV). Combination of abnormal HRT with a novel HRV measure—deceleration capacity (SAF—severe autonomic failure)—was documented as a potent risk marker in postinfarction patients.35
REFINE, CARISMA, MUSIC, and GISSI trials documented that abnormal HRT parameters, particularly abnormal turbulence slope, predicted total mortality and sudden death and/or arrhythmic events during follow-up.29,30,31,32 Interestingly, as shown in a combined analysis from REFINE and CARISMA studies, absence of an increase in turbulence slope (TS) values in the early postinfarction period, indicating lack of positive autonomic remodeling, was associated with a 7- to 10-fold higher risk of life-threatening arrhythmias during a 2-year follow-up.29 Abnormal HRT seems to be particularly useful in identifying high-risk patients among those with LVEF greater than 30%33,34 and is frequently evaluated in risk scores that combine assessment of arrhythmogenic substrate (low LVEF), repolarization instability (eg, TWA), and abnormal autonomic nervous system tone (HRV). Combination of abnormal HRT with a novel HRV measure—deceleration capacity (SAF—severe autonomic failure)—was documented as a potent risk marker in postinfarction patients.35
Macroscopic beat-to-beat changes in morphology, amplitude, and even polarity of the T wave have been observed in ECG recordings of patients with Prinzmetal angina, long QT syndrome, or acute coronary syndromes. These changes in surface ECG were known to be associated with the occurrence of ventricular arrhythmias. Risk stratification by means of microscopic TWA during an exercise treadmill test has been widely tested over the last two decades.20,36,37 Despite initial positive results of such risk stratification current guidelines do not recommend TWA testing in an acute phase of myocardial infarction.4 Evaluation of TWA during exercise requires elevated and stabilized heart rate, which results in a large number of undetermined tests. On the other hand, it is known that arrhythmic events may be triggered by mental stress and may occur in the early morning period. Therefore, evaluation of TWA from ambulatory ECG recordings emerged as a valuable tool in predicting ventricular arrhythmias. The Modified Moving Average (MMA) method for TWA analysis is the most widely studied. Higher levels of TWA indicate greater risk of cardiovascular mortality and ventricular arrhythmias. TWA greater than or equal to 47 µV in standard precordial leads is considered abnormal and TWA greater than 60 µV is severely abnormal. Up-to-date MMA-based TWA analyses on AECG have been proven to predict cardiovascular mortality and SCD in several clinical trials encompassing over 2000 patients, most of them ischemic, and those with impaired LVEF. Abnormal TWA is frequently combined in risk stratification, with other measures reflecting autonomic nervous tone.20,36,37
Not only beat-to-beat changes in the amplitude of the T wave but also discrete changes in the length and morphology of the T wave have been considered to be markers of arrhythmia. Various measures of QT variability and dynamics have been studied. Long-term Holter recordings enable dynamic evaluation of QT behavior. Different methodologic approaches have been proposed to evaluate the rate dependence of the QT interval. The most common are (a) the circadian profile of the rate-corrected QT interval (QTc); (b) long-term evaluation of the QT-RR relationship; and (c) the QT variability index.38,39 Some of these methods have been implemented on commercial Holter systems, and they are becoming available for routine clinical use. Even though clinical studies have documented a higher risk of arrhythmias in patients with increased QT variability or abnormal adaptation to heart rate (QT dynamics), there are no randomized trials documenting the role of QT assessment in guiding ICD therapy.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
