Chronic heart failure (HF) is an important healthcare problem in the developed world, since incidence and prevalence numbers have increased rapidly over recent decades. Recently, the American Heart Association (AHA) Statistics Committee and Stroke Statistics Committee reported that chronic HF affects 5.3 million patients in the US, with approximately 660,000 newly diagnosed patients each year.1 In western Europe, approximately four million patients have chronic HF.2 Moreover, in the US one million hospital admissions for (decompensated) HF have been reported annually, resulting in an exponential increase in direct and indirect costs involved with chronic HF.1 Even though significant progress has been made in prevention, diagnostic and therapeutic strategies for chronic HF, mortality rate still exceeds 50%.
Coronary artery disease (CAD) was found to be the main cause of HF in nearly 70% of patients and, despite rapidly developing treatment options for ischaemic and non-ischaemic HF, long-term prognosis remains poor.3–6 In 7,599 patients with New York Heart Association (NYHA) class II–IV HF, originating from the Candesartan in Heart failure – Assessment of Reduction in Mortality and morbidity (CHARM) study, cardiovascular outcome was studied over a mean follow-up of three years. In patients with a left ventricular ejection fraction (LVEF) below 33%, all-cause and cardiac mortality were 34% and 28%, respectively.6 Even higher mortality rates were reported by Levy et al.4 in the Framingham-based study population; cardiac mortality rate was more than 50%, over a mean follow-up of five years. These high mortality rates and the increasing number of hospitalisations due to (decompensated) HF emphasise the need for better risk stratification.
HF represents a complex clinical syndrome resulting from a reduced cardiac pump function. Progressive deterioration in cardiac function is based upon several interacting pathophysiologic mechanisms; in particular, the neurohormonal system (consisting of the adrenergic nervous system and renin–angiotensin–aldosterone system [RAAS]) plays an important role in chronic HF. Initially, neurohormonal feedback mechanisms tend to compensate the haemodynamic consequences of cardiac dysfunction via positive inotropic and chronotropic effects. However, in a chronic state these neurohormonal effects are detrimental and may cause cardiac hypertrophy and fibrosis, eventually resulting in remodelling. Moreover, desensitisation and downregulation of myocardial β-adrenoceptors and alterations in postsynaptic signalling lead to further decline in cardiac function.7
In addition, a dysfunctional cardiac autonomic nervous system may also be related to the development of ventricular arrhythmias.8–11 In this respect, exact mechanisms have to be elucidated and it has been suggested that regions with impaired innervation may be viable and hypersensitive to catecholamines, resulting in increased automaticity and enhanced triggering.8,12,13 In particular, the border zone of myocardial scar tissue may be predisposed to the development of re-entrant circuits because these regions are viable but may have damaged sympathetic nerves.14–18 The viable peri-infarct region can be (partially) denervated because sympathetic nerve fibres are more vulnerable to ischaemia than cardiomyocytes.19
Consequently, important clinical and prognostic information in HF patients can be provided by assessment of cardiac sympathetic innervation and function. Single-photon emission computed tomography (SPECT) and positron emission tomography (PET) are currently the only two imaging modalities to allow assessment of myocardial sympathetic nerve activation and innervation by visualisation of uptake and storage of radiolabelled neurotransmitters into pre-synaptic nerve endings. Both imaging techniques can be used to assess global and regional myocardial sympathetic nerve innervation.
Several PET tracers have been used to visualise sympathetic neurotransmission, including carbon-11 hydroxyephedrine (HED), C-11 epinephrine and F-18 fluorodopamine. At present, HED is the most frequently used PET tracer and allows absolute quantification of tracer uptake into sympathetic nerve terminals. One of the major advantages of HED is its capability to detect regional abnormalities in myocardial sympathetic innervation. Hartmann and colleagues20 performed an important study that demonstrated regional heterogeneity of impaired cardiac sympathetic innervation in 29 patients with idiopathic dilated cardiomyopathy, as compared with patients with normal cardiac function. In HF patients, PET imaging with HED has also been used to evaluate the prognostic value of sympathetic nerve innervation. In an elegant study, Pietila et al.21 demonstrated that global reduced HED accumulation was an independent predictor of adverse outcome in 46 patients with NYHA class II–III HF.
Most studies on cardiac sympathetic innervation, however, have been performed with 123-iodine metaiodobenzylguanidine (MIBG) scintigraphy. The labelling of MIBG with 123-iodine permits visualisation of myocardial sympathetic neuronal uptake. MIBG is a false neurotransmitter, an analogue of norepinephrine, which uses similar uptake mechanisms in the pre-synaptic nerve terminals as norepinephrine. The tracer is primarily transported into the pre-synaptic nerve terminal by sodium- and ATP-dependent transporters, referred to as uptake-1. In the nerve terminal, no degradation of MIBG takes place, resulting in an accumulation of MIBG with high signal intensity. These signals are used for planar and tomographic (SPECT) imaging. Early planar and SPECT imaging is performed at 10–20 minutes after MIBG administration, whereas delayed planar and SPECT imaging are performed three to four hours after tracer injection. From planar images, global cardiac MIBG uptake can be assessed visually or semi-quantitatively using early and late heart–mediastinum (H/M) ratio (see Figure 1). To calculate H/M ratio, regions of interest are manually drawn over the heart and upper mediastinum and the mean of myocardial counts per pixel is divided by the mean of mediastinal counts per pixel. Another planar-based parameter is the cardiac wash-out rate, which indicates the rate by which MIBG is released from the myocardium between early and delayed imaging. To calculate cardiac wash-out rate, late H/M ratio is subtracted from the early H/M ratio and divided by the early H/M ratio.
These planar-based parameters have been studied extensively in HF patients.22–28 In particular, the prognostic value of global MIBG uptake has been studied by Merlet and colleagues.22,23, Initially, the authors22 studied 90 patients with moderate to severe HF who underwent MIBG scintigraphy, chest X-ray, echocardiography and radionuclide angiography. After follow-up of 27 months, cardiac death and cardiac transplantations were reported in 22 and 10 patients, respectively. H/M ratio on delayed images was found to be the best independent predictor for cardiovascular death. Of note, patients with cardiac transplants were discarded from the survival analysis, whereas the second prognostic study performed by Merlet et al.25 included patients who underwent cardiac transplantation. In this study, 112 patients with NYHA class II–IV HF and LVEF <40% (measured with radionuclide angiography) underwent MIBG scintigraphy, right-sided invasive angiography, and routine clinical care (echocardiography, chest X-ray and peak exercise testing).
After a mean follow-up of 27±20 months, cardiac death was noted in 25 patients, and cardiac transplantation was performed in 19. As illustrated in Figure 2, late H/M ratio was significant lower in patients with major cardiac events than in patients without major cardiac events. Furthermore, the authors demonstrated that the only independent predictors for adverse outcome were reduced myocardial MIBG uptake on delayed planar images and depressed LVEF.
Besides the prognostic value of global MIBG uptake on delayed images, several studies have reported on cardiac wash-out rate as a potential predictor for adverse cardiovascular outcome.29–32 In 59 patients with NYHA class II–IV and dilated cardiomyopathy, the predictive value of MIBG parameters was studied by Momose and co-workers;29 among all planar MIBG parameters, cardiac wash-out rate was the most powerful predictor of cardiac death, over a mean follow-up of 25±13 months. For cardiac wash-out rate, a threshold level of 52% showed optimal predictive value for cardiac death. In addition, similar results were demonstrated by Yamada et al.31 in 65 HF patients. Over a mean follow-up of 34±19 months, cardiac wash-out rate was the only independent predictor for cardiac events. Furthermore, patients with cardiac events showed significant higher myocardial wash-out rate than patients without cardiac events, as illustrated in Figure 3. However, importantly, different threshold levels for cardiac wash-out rate were used in these two studies. Yamada et al.31 used a cut-off value of 27% for myocardial wash-out rate, whereas Momose and co-workers29 used a threshold level of 52%.
Currently, no consensus exists on the optimal cut-off values for early and late H/M ratios, cardiac wash-out rate or the defect score on SPECT imaging for prediction of outcome. Agostini et al.33 performed in this respect an important retrospective study on 290 HF patients from six centres across Europe. According to the study protocol, planar and SPECT examinations were re-analysed in a core laboratory by three independent observers who were blinded to all other data. Over a follow-up of two years, major cardiac events were noted, including cardiac death, cardiac transplantation, potential lethal ventricular arrhythmias or appropriate implantable cardioverter defibrillator (ICD) therapy. The authors showed that patients with a major cardiac event had significantly lower H/M ratio on delayed images than patients without a major cardiac event within the two-year follow-up period. Moreover, for late H/M ratio, the optimal threshold to predict cardiac events was defined at 1.75, and yielded a sensitivity of 84% and specificity of 60%.
The available studies have focused on an index MIBG study and its relation to cardiac outcome. However, several studies used imaging before and after HF therapy and demonstrated improved MIBG uptake after optimisation of HF medication; currently, limited information is available on the predictive value of serial MIBG studies.34–37 Recently, Kasama et al.38 reported on the predictive value of serial MIBG imaging in 208 HF patients with an LVEF under 45%. According to the study design, MIBG scintigraphy was performed at baseline and at six-month follow-up. During a mean follow-up of 5±2 months, cardiac death was reported in 56 patients. The authors reported that a six-month change in wash-out rate >5% was an independent predictor for all-cause cardiac death and sudden cardiac death. Although most prognostic studies used all-cause mortality, cardiac mortality (progressive HF), or cardiac transplantation as end-points, some studies suggested that MIBG imaging can also be used for prediction of ventricular arrhythmias and sudden cardiac death.39–41 Arora et al.39 evaluated the role of MIBG imaging in 17 patients with an ICD. In total, 10 patients received appropriate ICD discharges (reflecting potential lethal ventricular arrhythmias or sudden cardiac death) and these patients showed significant lower global (early) and regional (early and late) MIBG uptake than patients who did not receive appropriate ICD discharges. Furthermore, comparable results were reported by Nagahara and colleagues,41 who evaluated whether MIBG scintigraphy was useful for prediction of appropriate ICD therapy in 54 patients after a mean follow-up of 15 months. Ten patients who received appropriate ICD therapy showed significantly lower late MIBG uptake than patients who did not receive appropriate ICD therapy. Moreover, H/M ratio on delayed images was found to be an independent predictor for appropriate ICD therapy.
MIBG studies have reported on planar imaging as a potential tool for risk stratification in HF patients. Although some innervation studies have employed regional MIBG uptake as a prognostic marker, the role of SPECT imaging for prognostication of HF patients is unclear.42–44 In particular, the role of regional sympathetic denervated myocardium in the genesis of spontaneous ventricular arrhythmias has to be further elucidated.14–18 Bax et al.45 studied whether abnormalities in sympathetic innervation on MIBG SPECT could predict inducibility of ventricular arrhythmias on electrophysiological testing in 50 patients with coronary artery disease (CAD) and an LVEF of 40% or lower. The MIBG SPECT defect score was significantly larger in the 30 patients with a positive electrophysiologic test than in patients without inducible ventricular arrhythmias, as illustrated in Figure 4.
In conclusion, the available studies have shown that cardiac innervation imaging holds great potential for risk stratification and prognostication of HF patients. Recent studies suggest that, particularly, MIBG imaging may play a major role in identifying patients with LV dysfunction at elevated risk of HF death or arrhythmic death. The results of these initial studies are promising, and more studies will help to determine the precise role of innervation imaging in the risk stratification of HF patients.