Predicting the Need for an Implantable Cardioverter Defibrillator Using Cardiac Metaiodobenzylguanidine Activity Together with Plasma Natriuretic Peptide Concentration or Left Ventricular Function

Patient Population

The study population consisted of 54 consecutive patients (mean age, 53 y; 40 men and 14 women) who met the American Heart Association/American College of Cardiology guidelines for ICD implantation (14) (Table 1). ICD implantation (Micro Jewel II, GEM, and GEM II; Medtronics Inc. or VENTAK MINI IV; Guidant Inc.) was done using a standard (subclavian) venous approach. Informed consent for ICD implantation and participation with prospective follow-up in this study was obtained on the basis of the guidelines of the ethics committee of our hospital. Clinical ventricular arrhythmias were confirmed by Holter monitors or by a standard electrocardiogram before admission of the patient at our hospital. In addition, all patients underwent Holter recordings during admission before ICD implantation. In 35 of the 54 patients, the dual-chamber pacing mode was used because of the presence of a combined atrioventricular block or atrial fibrillation. Table 1 shows the underlying cardiac diseases that were responsible for the lethal ventricular arrhythmias: 9 patients had idiopathic dilated cardiomyopathy, 6 had idiopathic ventricular fibrillation, 16 had secondary cardiomyopathy, 6 had arrhythmogenic right ventricular cardiomyopathy, 7 had Brugada syndrome, 2 had long-QT syndrome, and 8 had hypertrophic cardiomyopathy. Thus, primary or secondary cardiomyopathies were found in 39 patients; no organic heart disease was found in patients with Brugada syndrome, idiopathic ventricular fibrillation, or long-QT syndrome. The LVEF was assessed by 2-dimensional echocardiography within 2 wk of cardiac MIBG imaging. The mean LVEF was 48% ± 17%.

Electrocardiographic Examinations

The following electrophysiologic tests were performed before ICD implantation. Signal-averaged electrocardiography and an electrophysiologic study were performed on 40 and 41 of the 54 patients, respectively, to establish the diagnosis of lethal ventricular tachyarrhythmias and to determine whether ICD implantation was indicated. A ventricular tachyarrhythmia induction test was done in the standard manner and included burst pacing and double or triple paired pacing with a minimal coupling interval of 180 ms from both apical and outflow sites of the right ventricle. Ventricular tachycardia (VT) was defined as 3 or more consecutive ventricular extrasystoles; sustained VT was defined as VT lasting for at least 30 s. A positive late potential on signal-averaged electrocardiography was defined as the presence of at least 2 of the following 3 criteria: a filtered QRS interval of more than 114 ms, a root mean square of less than 20 μV, and a low-amplitude signal of more than 38 ms.

Cardiac MIBG Imaging

Within 2 wk of ICD implantation but at least 3 d after electrocardiographic examinations, cardiac MIBG imaging was done using 111 MBq of ¹²³I-MIBG with a high specific activity (Daiichi Radioisotope Laboratories). Cardiac planar and tomographic MIBG images were obtained with the patient fasting and at rest 30 min (early) and 4 h (late) after intravenous injection of ¹²³I-MIBG using a γ-camera equipped with a low-energy general-purpose collimator. The view was anterior. Cardiac ¹²³I-MIBG activity was quantified as heart-to-mediastinum (HMR) ratio by manually setting the region of interest on the upper mediastinum and the whole cardiac region using visual guidance—a technique that had previously been reported to have a high reproducibility (4–7). ¹²³I-MIBG washout kinetics from the heart were also calculated as washout rate using a polar map technique with tomographic data because of the elimination of background activity, as previously reported (4–7).

Measurements of Plasma Concentration of BNP

Plasma concentration of BNP and atrial natriuretic peptide (ANP) were measured on the same day as or a few days after cardiac MIBG imaging. Samples used for the measurements were transferred to chilled disposable tubes containing aprotinin (500 kallikrein inactivator units/mL) and immediately placed on ice and centrifuged at 4°C. The concentration was measured by a specific immunoradiometric assay using a commercial kit (Shionogi), as previously reported (7); the kit contained 2 antihuman BNP monoclonal antibodies.

Follow-up Protocol

After discharge, all patients were prospectively followed at least every 3 mo for a mean of 15 mo at the outpatient pacemaker clinic of our university hospital, with the primary endpoint being lethal events consisting of cardiac death or appropriate ICD shock triggered by lethal arrhythmias (Table 1). According to documentation on appropriate ICD shock during follow-up, the patients were classified into 2 groups: a group with appropriate ICD shock or sudden cardiac death (group A) and a group with neither (group B). Standard drug therapy for arrhythmias, background diseases, or heart failure that had been used at study entry was continued as long as there were no adverse effects (Table 2). Antiarrhythmic agents, such as amiodarone, sotalol, class I antiarrhythmic drugs, or β-blockers, were used. Diuretics, digoxin, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, or β-blockers were used for heart failure, when indicated. An appropriate ICD discharge was defined as an electrical event occurring in response to sustained lethal VT or fibrillation, including defibrillation, cardioversion, and antitachycardia pacing, and was identified by electrocardiograms stored in the built-in memory system of the ICD.

Statistical Analysis

Statistical values are shown as mean ± SD. Mean values were compared between the 2 groups using the unpaired t test, and prevalence was compared using the χ² test. A P value of less than 0.05 was considered significant. Receiver operating characteristics were analyzed to determine optimal cutoffs for key parameters (i.e., late HMR, LVEF, and BNP) for predicting lethal arrhythmic events (Fig. 1). Hazard ratios and 95% confidence intervals for the univariate model were computed using major clinical parameters. Multivariate analysis with a Cox proportional hazards model was also done using significant parameters identified by univariate analysis. These analyses were performed using a computer software program, the SPSS statistical program package (SPSS, version 11.0; SPSS Inc.).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Receiver operating characteristic curves of cardiac late MIBG activity (HMR), LVEF, and BNP, indicating that optimal cutoffs for identifying lethal arrhythmic events are 1.95 for HMR, 50% for LVEF, and 187 pg/mL for BNP. Dotted arrows show suboptimal values (*) of LVEF (35%) and BNP (110 pg/mL) that were used for subgroup analysis, for which results are shown in Figure 4. EF = ejection fraction.

Sudden Cardiac Death and Cardiac Sympathetic Nerve Function

Several studies have shown that altered autonomic function is responsible for ventricular arrhythmias or sudden cardiac death. Earlier studies using cardiac neurotransmitter imaging (15–17) also demonstrated that impaired uptake of norepinephrine analogs was related to severe ventricular tachyarrhythmias. However, it was still necessary to determine whether altered cardiac MIBG activity is associated with fatal cardiac episodes due to ventricular tachyarrhythmias. At an advanced stage of heart failure, the plasma norepinephrine level and the firing of autonomic nerves in skeletal muscles are markedly increased, whereas heart rate variability and cardiac sympathetic innervation and function are greatly depressed (18–20). These complex changes in autonomic nervous function show that sympathetic innervation acts differently at different functional sites under various conditions, indicating the need for direct assessment of cardiac autonomic innervation. In the present study, the occurrence of documented ICD shock due to fatal arrhythmias during follow-up correlated with cardiac MIBG activity, left ventricular function, and plasma BNP level. These 3 variables were found to contribute to identification of patients at high risk for sudden cardiac death (Tables 2 and 3). Electrophysiologic indices, including the inducibility of lethal ventricular tachyarrhythmias and signal-averaged electrocardiography parameters, were not useful in helping to select patients who had the greatest risk for sudden cardiac death. Arora et al. (20) showed prospectively that appropriate ICD shock most likely occurs when both cardiac MIBG activity and heart rate variability are highly depressed. Microvolt T-wave alternans has recently been noted not only for its usefulness in arrhythmia risk assessment but also for its potential usefulness in the identification of patients who can benefit most from prophylactic ICD therapy (21–26). This method is simple, practical, and probably cost-effective (22). An increase in repolarization instability assessed by T-wave alternans precedes ventricular tachyarrhythmias culminating in sudden cardiac death (23). T-wave alternans, however, is one of the indices of repolarization instability and has a complex spectrum that varies dynamically over short intervals. In addition to differences in measurements and definitions of T-wave alternans (26), averaged measurements are possibly affected by various factors, including artifacts caused by baseline wander due to perspiration, motion, and respiratory oscillation; heart rate or rhythm disturbance; systemic or central (noncardiac) autonomic tone; circulating catecholamines; drugs used; and underlying cardiac diseases (23,24). It remains to be established whether, like an upsurge in T-wave alternans or other forms of repolarization instabilities preceding spontaneous arrhythmic events in humans (23,25), an averaged T-wave alternans frequency assessed by ambulatory recording is directly related to and can predict arrhythmic death in a prospective cohort study. On the other hand, cardiac MIBG activity is highly specific for sympathetic innervation in the myocardium per se and is reproducible because of the specific active transport, storage, and release systems at nerve endings. Because microvolt T-wave alternans testing, however, has high negative predictive powers, there is a need to investigate whether the presented scintigraphic technique has an incremental prognostic value in combination with T-wave alternans testing.

The mechanisms underlying impaired cardiac presynaptic sympathetic innervation that are responsible for lethal arrhythmias remain to be investigated. The changes in pre- and postsynaptic autonomic function in the failing heart are probably dependent on the progression of heart failure and structural myocardial damage (27). Cardiac MIBG activity assessed several hours after tracer injection has been shown to be related to myocardial norepinephrine content, left ventricular systolic function, and New York Heart Association functional status. Both ischemic and nonischemic etiologies can be responsible for the impairment of cardiac MIBG activity (5). In patients with organic heart disease, the ICD shock group had a significant increase in cardiac MIBG washout kinetics, compared with those who had no ICD shock. This finding is consistent with accelerated norepinephrine release and decreased uptake at presynaptic nerve endings, leading to spillover of norepinephrine (MIBG) in patients with moderate heart failure (28). In this group, an increase in cardiac sympathetic dynamics may facilitate arrhythmogenicity by mechanisms involving a reentry circuit, automaticity, or triggered activity. The heterogeneous degeneration process in viable but failing myocardial tissue is also related to exacerbation of electrical instability and causes denervation supersensitivity to sympathetic stimulation that precipitates ventricular arrhythmias (29,30). Thus, both functional and structural impairments of cardiac sympathetic innervation assessed by MIBG activity are likely to be related to lethal arrhythmic events.

Further Risk Stratification of ICD Patients Using Cardiac Sympathetic Activity

Cardiac MIBG activity was closely related both to a low event rate and to a high event rate depending on HMR, but it was not related to an intermediate event rate. These data suggest the limited prognostic efficacy of 1 variable and highlight the necessity of using combinations of other clinical parameters to risk-stratify patients at an intermediate risk for arrhythmia death. The data presented here showed that combined use of LVEF or BNP with cardiac MIBG activity could more effectively evaluate the need for ICD treatment in patients who have moderate impairment of cardiac sympathetic innervation. Triggers of lethal tachyarrhythmias or sudden cardiac death are multifactorial. Such triggers include the etiology of heart failure, cardiac systolic function, hemodynamics, plasma BNP level, concomitant drug use, and noncardiac complications. Patients at an intermediate risk for lethal arrhythmic events are most likely affected by these background factors. Sheldon et al. (31) reported that the patients at highest risk of death who benefit most from ICD therapy could be identified on the basis of age, LVEF, and New York Heart Association status. Our previous studies (4–7) have also shown that reduced cardiac MIBG activity is an independent determinant of long-term prognosis and has better prognostic value when used in combination with age, LVEF, history of previous myocardial infarction, New York Heart Association status, presence of noncardiac complications, such as renal dysfunction or diabetes mellitus, and plasma BNP level. These findings support the rationale for the use of several different cardiac indices that are related to lethal cardiac events. Overall, LVEF was less associated with lethal arrhythmic events in combination with cardiac MIBG activity than was BNP level when combined with cardiac MIBG activity in this study. When cardiac MIBG was moderately reduced, however, LVEF effectively identified less than 35% of patients at an increased risk for events. The present and previous results indicate that a further large-scale investigation is needed to establish the most effective method for risk stratification using significant clinical parameters in combination with cardiac MIBG indices and to show the incremental predictive powers.

Limitations

The present results revealed not only the pathophysiologic implications of impaired cardiac sympathetic innervation in patients at risk for lethal cardiac events or sudden death but also the feasibility for further clarifying the risk of lethal outcomes in patients treated with ICD implantation. Besides BNP and late HMR, the washout rate of MIBG activity and a history of prior myocardial infarction were close to being statistically significant (Table 4). Our previous study (4) showed the prognostic values of prior myocardial infarction when combined with late HMR. Other investigators (8,11,12) have demonstrated the possibility that washout rate is a prognostic marker. Given our encouraging results and previous findings, a large-scale, long-term study should be done to reach more definitive conclusions regarding the efficacies of the presented techniques. In particular, it is critical to find a method that has a maximal negative predictive value, because underuse, as well as overuse, of the ICD has to be avoided so as to minimize the possibility of fatal outcomes without ICD therapy and to maximize the cost-effectiveness of this device. To accomplish this requirement, long-term follow-up and careful management of patients are necessary. In addition, the indications for ICD implantation need to be appropriately revised. There is also a need to increase the availability of cardiac ¹²³I-MIBG imaging. The noninvasive features and simple imaging protocol, as well as the quantification of cardiac sympathetic innervation, of cardiac ¹²³I-MIBG imaging could contribute to its widespread clinical application in patients with ICD implantation who have heart failure and are at a mild-to-moderate risk for lethal cardiac events. Finally, because there are several factors that may affect cardiac MIBG activity, such as electrophysiologic provocation testing with cardioversion, drug treatment (8–13), and progression of underlying cardiac disease, repeated imaging may be necessary for more precise estimation of the activity during a long-term follow-up.