Assessment of Underlying Etiology and Cardiac Sympathetic Innervation to Identify Patients at High Risk of Cardiac Death

MATERIALS AND METHODS

Scintigraphic Imaging and Quantification of Cardiac MIBG Activity

Planar and SPECT imaging was performed 30 min (early) and 3–4 h (late) after an intravenous injection of 111 MBq ¹²³I-MIBG (Daiichi Radioisotope Labs, Ltd., Tokyo, Japan), with a specific activity of 1,665–2,035 MBq/mg, under resting and fasting conditions. Planar data were acquired from an anterior view using a gamma camera equipped with a low-energy, general-purpose collimator, with a 20% window centered on 159 keV. SPECT data were then acquired from a 45° left posterior oblique angle to a 45° right anterior oblique angle over a 180° arc in 36 preset sampling angles for 30 s per angle. The data were stored in a 64 × 64 matrix (7). After reconstruction of short-axis slices using a backprojection method with a Shepp–Logan filter, a polar map was generated to calculate washout kinetics of myocardial ¹²³I-MIBG activity by avoiding contamination by background activity. In this study, however, no SPECT data could be obtained for 15 (20%) of the 76 ISM patients and for 6 (11%) of the 56 ICM patients because of substantially reduced (not visually identifiable) cardiac uptake of ¹²³I-MIBG, even in early planar images. Therefore, for these patients, only data obtained from planar ¹²³I-MIBG imaging were used for further analysis. For patients who had been prescribed oral drugs before imaging, drug administration was withdrawn transiently for at least a half-day and then was restarted after the study. As previously described (7), myocardial ¹²³I-MIBG activity was quantified as a heart-to-mediastinum ratio (H/M) by manual setting of an 11 × 11 pixel region of interest on cardiac and upper mediastinal areas of early and late planar images by nuclear medicine staff without knowledge of the patient’s clinical data but with a highly reproducible technique of data sampling (y = 0.29 + 0.99x; r = 0.996; n = 40; P < 0.0001). When manual tracing was difficult because of absence of myocardial ¹²³I-MIBG uptake, a planar perfusion image from an anterior view was used as a reference to identify the cardiac region (Fig. 1). Resting perfusion imaging with ²⁰¹Tl (111 MBq), ^99mTc-sestamibi (600 MBq), or ^99mTc-tetrofosmin (600–740 MBq) was performed 10 min or 30–60 min after the tracer injection. Planar data were obtained from an anterior view using a gamma camera equipped with a low-energy, high-resolution, parallel-hole collimator (ZLC 75/Scintipac 2400 or ZLC 7500/Scintipac 700; Shimadzu, Tokyo, Japan, or Starcam 4000XC/T; General Electric Medical Systems, Milwaukee, WI), followed by SPECT imaging with the same protocol as that used for ¹²³I-MIBG imaging. The washout rate of cardiac ¹²³I-MIBG activity was calculated, using SPECT data and a polar map technique, as the percentage of change from early to late images without a physical decay correction: washout rate = [100 × (early ¹²³I-MIBG activity − late ¹²³I-MIBG activity)/early ¹²³I-MIBG activity] (7). Briefly, depending on patient heart size, 12–16 short-axis slices were selected to generate early and late polar maps. For each pixel in the polar maps, washout rate was calculated using early and late ¹²³I-MIBG data to obtain the mean value that was used as a washout rate for each patient.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Planar early and late ¹²³I-MIBG and ²⁰¹Tl images of 2 typical patients, 72-y-old man with ISM and sudden cardiac death (A) and 65-y-old woman with ICM and death from heart failure (B). Both early and late cardiac ¹²³I-MIBG uptakes are tremendously reduced in well-perfused hearts; in particular, late H/M was <1.5 in both patients.

RESULTS

The mean follow-up intervals were 54 ± 31 mo for the ISM group and 55 ± 29 mo for the ICM group (Table 1). During follow-up, 47 cardiac and 11 noncardiac deaths were recorded (Table 1). In the ISM group, the cardiac deaths included 13 deaths from heart failure, 9 sudden cardiac deaths, 1 death from arrhythmia (uncontrollable ventricular tachycardia and fibrillation), and 5 deaths from myocardial infarction. In the ICM group, 16 deaths from heart failure, 2 sudden cardiac deaths, and 1 death from arrhythmia were documented. Figure 1 shows planar ¹²³I-MIBG and ²⁰¹Tl images of 2 typical patients from the ISM and ICM groups who died from cardiac conditions. Despite well-perfused myocardium (as determined by rest thallium, sestamibi, or tetrofosmin imaging), both early and late cardiac ¹²³I-MIBG uptakes were considerably reduced to <1.5 as a late H/M. Although the ISM group had a greater prevalence of sudden cardiac death and death from arrhythmia (13%) than did the ICM group (5.4%), and the ICM group had a greater prevalence of death from heart failure (29%) than did the ISM group (17%) (Table 1), the differences were not statistically significant. The ISM and ICM groups had comparable early and late H/Ms of ¹²³I-MIBG activity, but both of these were significantly (P < 0.05) lower than those of the control group: 1.95 ± 0.41 and 1.79 ± 0.38, respectively, for the ISM group; 1.91 ± 0.43 and 1.71 ± 0.41, respectively, for the ICM group; and 2.30 ± 0.17 and 2.38 ± 0.29, respectively, for the control group (Table 1). Compared with the control group, both the ICM group and the ISM group had a significantly greater washout rate of ¹²³I-MIBG activity, and the washout rate of the ICM group was significantly greater than that of the ISM group: 42% ± 13% for the ICM group, 34% ± 13% for the ISM group, and 20% ± 9% for the control group (P < 0.001) (Table 1).

In the ISM group, patients who died from cardiac conditions had a significantly greater LVDd and LVDs, and a smaller percentage of fractional shortening, than did surviving patients (Table 2). In the ICM group, patients who died from cardiac conditions and those who survived had a comparable NYHA class and cardiac function (Table 2). In both the ISM and the ICM groups, patients who died from cardiac conditions tended to have a smaller H/M of ¹²³I-MIBG activity than did patients who survived: 1.67 ± 0.34 and 1.84 ± 0.37, respectively (P = 0.057), for the ISM group; 1.56 ± 0.35 and 1.79 ± 0.40, respectively (P = 0.042), for the ICM group (Table 2). ICM patients who died from cardiac conditions, however, had a significantly greater washout rate (46% ± 10%) than did ICM patients who survived (38% ± 13%) and ISM patients with lethal cardiac events (31% ± 13%). For ISM patients, there was no significant difference in washout between those who survived and those who had cardiac death. Five ISM patients who died of acute myocardial infarction tended to have greater late MIBG activity (1.79 ± 0.39) than did ISM patients who died of pump failure (1.64 ± 0.34) or who had sudden cardiac death (1.66 ± 0.35), although the statistical powers were not significant (P = 0.715). There was also no significant difference in clinical variables, cardiac function, or other MIBG data between ISM patients who died of acute myocardial infarction or pump failure and ISM patients who had sudden cardiac death.

Table 3 summarizes the overall results of prognosis analysis for 69 ISM patients after 7 patients who had died from noncardiac causes had been excluded to more specifically identify the prognostic powers of clinical, scintigraphic, and cardiac function parameters. Multivariate discriminant analysis using the Cox proportional hazards model showed that late H/M was the most powerful independent predictor of a lethal clinical outcome, with a Wald χ² value of 18.6502, an odds ratio of 0.0050, and a 95% confidence interval of 0.0004–0.0555 (P = 0.0000), compared with other significant variables, such as early H/M, atrial fibrillation, NYHA class, and the male sex (Table 3). Likewise, when 52 ISM patients with an LVEF < 40% were analyzed, late H/M had the greatest Wald χ² value (16.9424), with an odds ratio of 0.0061 and a 95% confidence interval of 0.00054–0.0694 (P = 0.0000), compared with other significant variables (Table 4). Table 5 summarizes the overall results of prognosis analysis for 52 ICM patients after exclusion of 4 patients with noncardiac causes of death. Multivariate discriminant analysis revealed that late H/M was the only significant independent predictor of cardiac death for the 52 ICM patients (Wald χ² value, 5.3394; odds ratio, 0.1326; 95% confidence interval, 0.0239–0.7359 [P = 0.0208]) and when analysis was restricted to 42 ICM patients with an LVEF < 40% (Table 5).

In the current population, several candidate thresholds of late H/M using the mean value and SD of late H/M in cardiac death patients (Table 2) were evaluated by Kaplan–Meier analysis to identify an upper limit of late H/M for differentiating patients who are at high risk from those who are not. For all patients, the upper cutoff values of late H/M were 1.82 (P = 0.0321; log rank = 4.59), which was the mean value plus 0.5 SD in ISM patients who died from cardiac conditions, and 1.82 (P = 0.0255; log rank = 4.99), which was the mean value plus 0.75 SD in ICM patients who died from cardiac conditions (Table 6). The 5-y survival rates of patients with versus those without a late H/M < 1.82 were 56% versus 83% for the ISM group and 51% versus 86% for the ICM group (Fig. 2). When only patients with an LVEF < 40% were considered, however, the upper limit of late H/M for identification of high-risk patients was 1.50 (P = 0.0358; log rank = 4.41) for ISM patients and 2.02 (P = 0.0050; log rank = 7.86) for ICM patients (Table 7). In this case, the 5-y survival rates of ISM patients with versus those without a late H/M < 1.50 were 40% versus 68% and the 5-y survival rates of ICM patients with versus those without a late H/M < 2.02 were 48% versus 100% (Fig. 2). Figure 3 shows the annual rate of cardiac death when each patient group was divided into 4 subgroups. Except for the ICM subgroup with an LVEF of 40% or more and a late H/M > 2.02, which had only 3 patients, the subgroups with the greatest annual rates of cardiac death—18.2%/y for the ISM group and 11.9%/y for the ICM group—were those with both an LVEF < 40% and a late H/M less than identified thresholds. However, in ICM patients with an LVEF < 40%, there was no significant difference in clinical or cardiac function variables between the subgroups with and without a reduced late H/M.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Kaplan–Meier survival curves produced by identified cutoff values of late H/M in patients with ISM (A) and ICM (B). In both ISM and ICM groups, patients with late H/M < 1.82, as identified by overall analysis, show significantly lower survival rates than those with late H/M ≥ 1.82. Furthermore, when patients with LVEF < 40% are considered, ischemic patients with late H/M < 1.50 and idiopathic patients with late H/M < 2.02 show significantly lower survival rates than each counterpart.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Annual rate of cardiac death when each group was divided into 4 subgroups based on LVEF of 40% and on identified thresholds of ¹²³I-MIBG activity. For LVEF of <40% and late H/M less than identified thresholds (1.5 for ischemic and 2.02 for idiopathic), annual rates of cardiac death are greatest, 18.2%/y for ischemic group and 11.9%/y for idiopathic group, when subgroup with LVEF of 40% or more and late H/M > 2.02 are excluded because of small number of patients (n = 3).

DISCUSSION

Our study cohort comprised patients with ISM and ICM because these are the most prevalent cardiomyopathies responsible for cardiac death and heart transplantation. Of the clinical indices analyzed for both cardiomyopathies, cardiac ¹²³I-MIBG activity had the strongest independent long-term prognostic value. Overall, the fact that the ability of late ¹²³I-MIBG uptake to identify patients at the highest risk for cardiac death was comparable for the 2 cardiomyopathies strongly suggests that, for both, a lethal clinical outcome involves impaired cardiac sympathetic innervation.

We performed both 30-min and 3- to 4-h postinjection imaging in this study to assess the integrity of presynaptic nerve terminals and uptake 1 function (19–21) and to calculate ¹²³I-MIBG kinetics. The ¹²³I-MIBG washout rate may be not only a significant marker of cardiac sympathetic drive (5,10) or norepinephrine (22) but also useful for ascertaining the mechanism of reduced cardiac ¹²³I-MIBG activity. Both ISM and ICM patients, particularly ICM patients with lethal outcomes, had significantly higher ¹²³I-MIBG washout rates than did control subjects (Tables 1 and 2), supporting the previous findings of accelerated sympathetic tone or norepinephrine spillover in heart failure patients. Neither ¹²³I-MIBG washout rate nor early ¹²³I-MIBG activity was, however, shown to be a significant independent predictor in recent studies (7,23), probably because an early ¹²³I-MIBG uptake reflects only integrity of presynaptic nerve terminals and uptake 1 function whereas late ¹²³I-MIBG activity includes overall information about neuronal functions from uptake to release through the storage system at nerve terminals (4,19). Augmented cardiac sympathetic drive or ¹²³I-MIBG washout rate is not specific for heart failure and does not necessarily result in a later ¹²³I-MIBG defect (5–15). In this study, therefore, the late-phase ¹²³I-MIBG abnormality was shown to be a most powerful and independent predictor for cardiac death.

Several mechanisms are likely to be responsible for a lethal clinical outcome from later reduction of cardiac ¹²³I-MIBG activity. Cardiac sympathetic neurons can be anatomically lost because of ischemic injury or myocardial scarring and because of the degenerative cardiomyopathy process. Distal denervation induced by proximal neuron damage or myocyte damage has been observed in patients with myocardial ischemia or acute infarction (14,24–26). Uptake 1 function may be impaired at nerve endings, which are an ATP-requiring system, in failing hearts because intracellular ATP content is depleted in the failing or ischemic myocardium (20,26,27) and sympathetic nerves are more susceptible to ischemia than are myocytes (14,24,25). These possibilities can elucidate both initial and delayed cardiac ¹²³I-MIBG defects without an increased ¹²³I-MIBG release. Functional abnormality, such as altered ¹²³I-MIBG kinetics, may also explain reduction of late cardiac ¹²³I-MIBG uptake. The greater washout rate in ICM patients who died from a cardiac condition, in comparison with surviving ICM patients and ISM patients, is likely to have been responsible for the further reduction in late cardiac ¹²³I-MIBG activity (Table 2). The possible mechanisms behind the augmented ¹²³I-MIBG kinetics are impaired norepinephrine storage at the cardiac synapse, augmented spillover because of an accelerated sympathetic drive, and relatively impaired efficiency of the reuptake mechanism compared with neuronal release of norepinephrine (MIBG) (11,22). Thus, both initial deficits in cardiac sympathetic activity and concurrent functional abnormalities may be responsible for a considerable reduction in cardiac ¹²³I-MIBG activity.

SPECT imaging was used in this study for precise calculation of washout kinetics without contamination by increased background activity. Lung ¹²³I-MIBG activity decreases considerably over several hours independently of cardiac disease (28). Especially when initial cardiac ¹²³I-MIBG uptake is markedly low, this value may be exaggerated. ¹²³I-MIBG washout rate assessed by planar imaging may thus have been overestimated as a prognostic marker in other studies, even though ¹²³I-MIBG washout kinetics could be a sensitive marker of cardiac sympathetic drive and therapeutic manipulations (17,18). When H/M of ¹²³I-MIBG activity was <1.5, however, it was extremely difficult to see ¹²³I-MIBG uptake well enough to reconstruct SPECT images. Considerably reduced ¹²³I-MIBG uptake was observed in 21 patients (17%), preventing us from obtaining data on tracer washout and regional abnormalities. In this context, SPECT imaging may be less efficient than quantitative planar imaging for risk stratification in heart failure patients (3,7,23). ¹²³I-MIBG kinetics may be more reliable as an index of augmented cardiac sympathetic nerve drive or impaired presynaptic vesicular storage at a relatively early or compensatory stage of heart failure, rather than as a long-term prognostic marker.

Despite careful selection of patients in this study, selection bias may not have been eliminated completely because ICM patients had greater functional abnormality and ischemic patients were older (Table 1). Clinical characteristics, such as age, sex, and coronary risk, and therapeutic strategies can also modify the timing and outcome of cardiac events in coronary artery disease. Nevertheless, multivariate analysis clearly showed that the prognostic value of impaired cardiac sympathetic function was identical in ISM and ICM patients in terms of the statistically independent and most powerful predictive value of late ¹²³I-MIBG activity for identifying patients most at risk for cardiac death. Circadian patterns of autonomic nervous system function parallel the timing of ischemia, myocardial infarction, arrhythmia, and sudden cardiac death. Sudden death from either ischemic or nonischemic cardiac disease appears to have a similar timing in heart failure patients, irrespective of the underlying cardiac conditions (29–32). These findings and our current results strongly suggest that altered cardiac sympathetic function has pathophysiologic and prognostic implications in both ISM and ICM processes. The precise etiologic relationship between cardiac sympathetic innervation and cardiac death was not determined by this study, but several possible explanations can be offered. Impaired cardiac sympathetic innervation is closely related to hemodynamic and thrombogenic states leading to plaque rupture and thrombus formation, to arrhythmogenesis from sympathovagal imbalance and denervation supersensitivity (25), to endothelial dysfunction, and to downregulated adrenoceptor function responsible for impaired inotropic and chronotropic reserves (23,33–35). The upper limit of late H/M (1.82) for differentiating high-risk patients from others was identical for both the ISM and the ICM groups. However, when LVEF was <40%, the H/M threshold was higher for ICM patients (2.02) than for ISM patients (1.50) (Fig. 3). Although the precise mechanisms of these differences are not clear from this study, underlying cardiac function, as well as autonomic function abnormality, is likely to still modify ultimate clinical outcome, and impairment of cardiac sympathetic innervation may have more crucial implications for prognosis in ICM patients when cardiac function is severely depressed.

This study had several limitations. It provided long-term follow-up results using comparable patient numbers in ICM and ISM groups when compared with earlier studies (3,7,23). However, how the current results would apply to patients at high risk for cardiac death from other diseases and to patients treated with currently accepted therapies remains to be determined. This study was designed before angiotensin-converting enzyme inhibitors, β-blockers, and implantable cardioverter defibrillators came into general use in heart failure patients, as well as before social acceptance of heart transplantation in our country (36). The current data are limited but may reflect the true predictive value of the imaging technique for lethal cardiac events without the confounding effects of more aggressive treatment. Prognostic data from the technique should be used for developing more effective therapeutic interventions for high-risk patients (37,38). Our recent study revealed potential for combining the use of ¹²³I-MIBG uptake and heart rate variability analysis to select patients at increased risk for sudden death (6). Cohen-Solal et al. (23) reported a close correlation between late ¹²³I-MIBG uptake and peak-exercise oxygen uptake. Kaye et al. (39), using an invasive method, showed greater prognostic value for cardiac sympathetic innervation than for whole-body indices. Thus, altered cardiac sympathetic nerve activity appears not only to be a consequence of degenerative disease processes but also to have etiologic implications for lethal clinical outcomes. More work is needed to better understand the interactions between myocardial injury and noncardiac effects, such as neurohormonal markers (40), pulmonary function, and peripheral factors, responsible for fatal clinical outcomes. Finally, sudden cardiac death is possibly attributable to lethal arrhythmias (6,37,38). The precise mechanisms of sudden cardiac death and the correlation between cardiac ¹²³I-MIBG activity and lethal arrhythmias, however, remain to be revealed.

CONCLUSION

Independent of cardiac function parameters and ¹²³I-MIBG washout kinetics, impaired ¹²³I-MIBG activity is the most powerful long-term predictor for cardiac death in both ISM and ICM patients. Combined assessment of cardiac function and the ¹²³I-MIBG activity threshold is more likely to identify an increased risk of cardiac death in both ISM and ICM patients. These findings suggest that impaired cardiac presynaptic function has pathophysiologic implications in the occurrence of lethal cardiac events in both ISM and ICM patients. The underlying etiology of cardiac dysfunction, however, may affect the threshold of ¹²³I-MIBG activity for the differentiation of high-risk patients.