Author + information
- Received July 17, 2012
- Revision received January 28, 2013
- Accepted February 12, 2013
- Published online July 1, 2013.
- Tomoaki Nakata, MD, PhD∗∗ (, )
- Kenichi Nakajima, MD, PhD†,
- Shohei Yamashina, MD, PhD‡,
- Takahisa Yamada, MD, PhD§,
- Mitsuru Momose, MD, PhD⋮,
- Shu Kasama, MD, PhD¶,
- Toshiki Matsui, MD, PhD#,
- Shinro Matsuo, MD, PhD†,
- Mark I. Travin, MD∗∗ and
- Arnold F. Jacobson, MD, PhD††
- ∗Second (Cardiology) Department of Internal Medicine, Sapporo Medical University School of Medicine, Sapporo, Hokkaido, Japan
- †Nuclear Medicine, Kanazawa University School of Medicine, Kanazawa, Japan
- ‡Department of Cardiovascular Medicine, Toho University Omori Medical Center, Tokyo, Japan
- §Cardiology, Osaka Prefectural General Medical Center, Osaka, Japan
- ⋮Nuclear Medicine, Tokyo Women’s Medical University, Tokyo, Japan
- ¶Cardiology, Cardiovascular Hospital of Central Japan, Shibukawa, Japan
- #Cardiology, Social Insurance Shiga General Hospital, Otsu, Japan
- ∗∗Cardiology and Nuclear Medicine, Montefiore Medical Center, Albert Einstein College of Medicine, Bronx, New York
- ††Medical Diagnostics, GE Healthcare, Princeton, New Jersey
- ↵∗Reprint requests and correspondence:
Dr. Tomoaki Nakata, Second (Cardiology) Department of Internal Medicine, Sapporo Medical University School of Medicine, Hokkaido Prefectural Esashi Hospital (Cardiology), S-1, W-16, Chuo-ku, Sapporo 060-0061, Japan.
Objectives The study objectives were to create a cardiac metaiodobenzylguanidine (mIBG) database using multiple prospective cohort studies and to determine the quantitative iodine-123–labeled mIBG indices for identifying patients with chronic heart failure (HF) at greatest and lowest risk of lethal events.
Background Although the prognostic value of cardiac mIBG imaging in patients with HF has been shown, clinical use of this procedure has been limited. It is required to define universally accepted quantitative thresholds for high and low risk that could be used as an aid to therapeutic decision-making using a large cohort database.
Methods Six prospective HF cohort studies were updated, and the individual datasets were combined for the present patient-level analysis. The database consisted of 1,322 patients with HF followed up for a mean interval of 78 months. Heart-to-mediastinum ratio (HMR) and washout rate of cardiac mIBG activity were the primary cardiac innervation markers. The primary outcome analyzed was all-cause death.
Results Lethal events were observed in 326 patients, and the population mortality rate was 5.6%, 11.3%, and 19.7% at 1, 2, and 5 years, respectively. Multivariate Cox proportional hazard model analysis for all-cause mortality identified age (p < 0.0001), New York Heart Association (NYHA) functional class (p < 0.0001), late HMR of cardiac mIBG activity (p < 0.0001), and left ventricular ejection fraction (LVEF) (p = 0.0029) as significant independent predictors. Analysis of the 512-patient subpopulation with B-type natriuretic peptide (BNP) results showed BNP (p < 0.0001), greater NYHA functional class (p = 0.0002), and late HMR (p = 0.0011) as significant predictors, but LVEF was not. The receiver-operating characteristic–determined threshold of HMR (1.68) identified patients at significantly increased risk in any LVEF category. Survival rates decreased progressively with decreasing HMR, with 5-year all-cause mortality rates >7% annually for HMR <1.25, and <2% annually for HMR ≥1.95. Addition of HMR to clinical information resulted in a significant net reclassification improvement of 0.175 (p < 0.0001).
Conclusions Pooled analyses of independent cohort studies confirmed the long-term prognostic value of cardiac mIBG uptake in patients with HF independently of other markers, such as NYHA functional class, BNP, and LVEF, and demonstrated that categoric assessments could be used to define meaningful thresholds for lethal event risk.
Despite recent advances in pharmacological and nonpharmacological treatment of heart failure (HF), the physiological characteristics that determine the risk profile of an individual patient are often poorly understood. One attribute of cardiac physiology that could have direct application to therapeutic measures used to reduce occurrence of lethal events in patients with HF is the status of the autonomic nervous system.
Autonomic nervous system function is associated with manifestations, progression, and outcomes of chronic HF (1). Despite initial compensatory augmentation for maintenance of systemic hemodynamics and peripheral circulation, long-lasting and excess stimulation of sympathetic nerve function is responsible for adverse events in HF attributable to increased activity of the renin-angiotensin-aldosterone system (2,3). This adrenergic hyperactivity results in increases in tissue oxygen demand, peripheral vascular resistance, plasma volume and oxygen stress, and water–electrolyte imbalance. Other potential consequences of this hyperactivity are myocardial ischemia and arrhythmic events.
Persistent efferent stimulation of sympathetic nerve function in failing hearts impairs efficiency of reuptake, turnover, and storage of norepinephrine (NE) at pre-synaptic nerve endings, resulting in increased spill-over and deficiency of NE stores (4). Cardiac sympathetic function and status of innervation can be assessed noninvasively using neuroimaging with the NE analog, iodine (I)-123–labeled metaiodobenzylguanidine (mIBG), which shares uptake, storage, and release system at nerve endings with NE (5). Increased NE turnover and pre-synaptic NE deficits can be identified as an increased mIBG washout rate (WR) from the heart and decreased mIBG activity quantified as the heart-to-mediastinum ratio (HMR).
Over the past 2 decades, numerous single and multicenter studies (6–11) and a recent meta-analysis using published data (12) have demonstrated the efficacy of cardiac mIBG sympathetic imaging for identifying symptomatic patients with HF most likely to experience lethal events. However, these studies have not provided data sufficient to establish long-term prognostic values and universally accepted levels of uptake that define high- and low-risk patient cohorts. Such definitions would require analyses of a large cohort database including prognostic mIBG markers and other standard clinical variables assessed for HF risk stratification. The present study was designed to create such a cardiac mIBG database by combining patient-level data from multiple prospective HF cohort studies (8,13–19), all of which were performed as physician-initiated clinical research in routine cardiology practice in Japan. The objective was to prospectively analyze the database to confirm short- and long-term prognostic values of quantitative 123I-mIBG indices on myocardial sympathetic innervation imaging and then define uptake levels that identified patients with HF at highest and lowest risk of all-cause mortality.
The present study was designed as a patient-level analysis combining individual prospectively obtained original datasets from multiple prospective HF cohorts studied independently between 1990 and 2009 in Japan (8,13–19). All patients were participants enrolled in prospective observational studies of mIBG cardiac imaging, a procedure that is approved for clinical use in Japan. The entry criteria for the current analyses were as follows: a prospective HF follow-up study; establishment of the HF diagnosis by cardiologists using standard diagnostic criteria (the Framingham criteria); a minimum of 12 months of periodic follow-up for patients who survived; a primary end-point of all-cause mortality; and cardiac mIBG imaging performed after stabilization using contemporary drug therapy for the measurement of HMR and WR. In the present study, only definitively confirmed deaths were considered events, but not aborted events such as recovery from resuscitated cardiac arrest or appropriate implantable cardioverter-defibrillator (ICD) shock.
Cardiac mIBG imaging is a routinely available clinical procedure in Japan (cost covered by the national medical insurance system of Japan) that is recommended in nuclear cardiology practice guidelines for the management and prognostic assessment of HF (20). Standard imaging procedures performed in Japan are as follows: use of 111 MBq of 123I-mIBG with a high specific activity (commercially available in Japan), a gamma camera equipped with a low-energy collimator, and early and delayed imaging, from which HMR and WR are calculated at the time of imaging.
According to guidelines of the ethics committee of each medical center, patient informed consent for participation was obtained in all original cohorts studied. Performance of the present study also was based on ethical committee guidelines for approval to collect data and personal intelligence protection.
Patients with HF from 6 Japanese medical centers provided data for this study. All patients were regularly followed for the mean follow-up interval of 77.6 (range 1 to 175) months with a median interval of 67 months at the outpatient clinic of each facility by cardiologists who were in charge of HF management. Median follow-up intervals in the 6 institutions were 72.0 (range 1 to 175), 73.0 (range 1 to 115), 58.7 (range 9.5 to 91), 83.1 (range 1.6 to 175), 127 (range 1 to 173), and 31.5 (range 10 to 72) months. The primary endpoint was all-cause death, including cardiac deaths (pump failure death, death due to myocardial infarction, and sudden cardiac death) and noncardiac death. Sudden cardiac death was defined as witnessed cardiac arrest and death within 1 h after onset of acute symptoms or unexpected death in patients known to have been well within the previous 24 h. Follow-up data collected as part of the original studies were updated when possible. Clinical outcomes were confirmed regularly by patient medical records or telephone interview in each facility; therefore, follow-up data were available for all patients included in this study.
A total of 1,360 patients were entered into the original cohort studies, but 32 patients (2.35%) were excluded from the database because they were lost to follow-up thereafter. Data from 1,401 mIBG studies in the remaining 1,328 patients were submitted for this study. Six surviving patients were excluded because of a follow-up period <1 year, and results from 73 repeated (not first) mIBG studies also were removed from the analysis. Thus, the database for final analysis consisted of results from a total of 1,322 patients, 71% of whom were male, with a mean age of 61 (range 15 to 97) years; 362 (27.4%) had an ischemic HF cause (275 [20.8%] with prior myocardial infarction), 1,322 had a mean left ventricular ejection fraction (LVEF) of 37% (range 6% to 88%), 298 (23.8%) had diabetes mellitus, 388 (31.0%) had hypertension, and 288 (21.8%) had dyslipidemia. The following patients were not submitted for inclusion in the database: patients with end-stage renal failure requiring dialysis therapy; patients with malignant disorders; patients with neurogenic disorders involving the autonomic nervous system; patients having been treated with tricyclic antidepressant drugs or other drugs that are known to have a sympathomimetic action and interfere with cardiac mIBG uptake; and patients who were scheduled to undergo any cardiac surgery (typically for revascularization rather than device implantation). Thirteen patients (1.2%) had ICDs at the time of mIBG imaging. Plasma B-type natriuretic peptide (BNP) data were available for 512 patients (39%).
Cardiac mIBG imaging
Cardiac mIBG imaging was performed using standard procedures after administration of 111 MBq of high specific activity 123I-mIBG (commercially available in Japan, Daiichi Radioisotope/Fujifilm RI Pharma, Tokyo, Japan) with patients in fasting and resting condition (8,13–19). Anterior planar mIBG images of the chest were obtained using a gamma camera equipped with a low-energy collimator 15 to 30 min (early image) and 3 to 4 h (late image) after the intravenous tracer injection. Cardiac 123I-mIBG activity was quantified as HMR by manually drawing regions of interest on the upper mediastinum and the whole cardiac region by an experienced nuclear medicine technician who was unaware of any clinical data. 123I-mIBG WR from the heart, based on count differences between early and late images, also was calculated (11).
Analysis of the combined database was performed by independent statistical experts (K.N. and S.M.) without participation of the other investigators. Summary values for numeric variables are shown as mean ± SD. Mean values were compared between the 2 groups (fatal events and survivors) using the unpaired t test, and prevalence values were compared using the chi-square test. Receiver-operating characteristic (ROC) analysis was performed to determine the optimal cutoff value of an independent significant parameter. Univariate and multivariate analyses using a Cox proportional hazards model were performed. The multivariate model for time to all-cause mortality was evaluated using a stepwise forward elimination procedure and categorized variables such as age (<30 to ≥80 years, 10-year unit), greater New York Heart Association (NYHA) functional class (I/II vs. III/IV), late HMR (<1.1 to ≥2.0, 0.1 unit), LVEF (<20% to ≥50%, 10% unit), and BNP (<200, 200 to 399, ≥400). Factors considered included demographic data (age, sex), medical history (e.g., diabetes mellitus, hypertension), cardiac medications (e.g., angiotensin-converting enzyme inhibitors, beta-blockers), and measured parameters (LVEF, BNP, HMR). Factors significant at the p <0.05 level were retained in the model. Survival curves of patient subgroups were created by the Kaplan-Meier method and compared using the log-rank test. Analyses based on sequential HMR subgroups were used to define thresholds for different mortality rates to 5 years.
Models without and with inclusion of HMR were compared using ROC curve and net reclassification improvement (NRI) methods for more appropriate identification of low and/or high-risk categories. All analyses were performed using the SAS statistical program package (JMP version 9, SAS, Cary, North Carolina).
During follow-up, 326 patients (24.7%) died. Of these deaths, 263 (81%) were judged to be of cardiac cause as part of the original studies or during the current data review. Table 1 compares clinical backgrounds between groups with and without lethal events. The event group was older and more frequently male, had a history of sustained ventricular tachycardia (VT) and diuretic use, had greater NYHA functional class, had greater BNP level, and had lower LVEF. The event group less frequently had hypertension or dyslipidemia, and used beta-blocking agents. On mIBG imaging, the deceased group had significantly lower late HMR and greater WR than the survivors. There was no significant difference in the prevalence of ischemic cause of HF, diabetes mellitus, or use of renin-angiotensin-aldosterone system inhibitors between the groups.
ROC analysis identified 1.68 and 43% as optimal thresholds of HMR and WR, respectively (Fig. 1), for dichotomizing the population into higher- and lower-risk patients for lethal outcomes. Patients with late HMR <1.68 or WR >43% had significantly (p < 0.0001) lower survival rates (Fig. 2). Likewise, the late HMR and WR values differentiated the high- from the low-risk patients both for sudden cardiac death and for pump failure death at 5 years (Table 2).
According to the mean ± SD value (1.75 ± 0.35) of late HMR in the total population, patients were classified into 3 subgroups that could be characterized as low-, intermediate-, and high-risk (Fig. 3). Patients also could be classified into 3 subgroups using LVEF categories: ≤35%, 36% to 49%, and ≥50%. The subgroup with LVEF ≤35% had significantly poorer survival than the other 2 subgroups (Fig. 4). The late HMR threshold of 1.68 identified high-risk patients in each LVEF-categorized subgroup (Fig. 5).
Univariate Cox analysis identified significant variables, including age, LVEF, BNP, and mIBG parameters (HMR and WR), listed in Table 3. Ischemic cause, renin-angiotensin-aldosterone system inhibitors, and diabetes mellitus were not significant. Multivariate Cox proportional hazards analysis using categorized variables demonstrated age (<30 to ≥80 years, 10-year unit), greater NYHA functional class (I/II vs. III/IV), late HMR (<1.1 to ≥2.0, 0.1 unit), and LVEF (<20% to ≥50%, 10% unit) for the total population as independent significant predictors of lethal events (Table 4). When the subpopulation with BNP was analyzed, categoric BNP (<200, 200 to 399, ≥400), greater NYHA functional class (I/II vs. III/IV), and late HMR (<1.1 to ≥2.0, 0.1 unit) were identified to be significant, but LVEF (<20% to ≥50%, 10% unit) was not.
All-cause mortality progressively decreased with increasing HMR. All-cause mortality rates at 1 year, 2 years, and 5 years for the total population were 5.6%, 11.3%, and 19.7%, respectively (Fig. 6). When late HMR was <1.7, mortality rates at these 3 times were greater than the average, whereas those with HMR >1.7 had below average mortality rates, confirming the findings of the earlier ROC analysis (HMR threshold 1.68). For 5-year mortality, patients with HMR ≤1.2 had an annual death rate >7%, whereas patients with HMR ≥2.0 had an annual death rate <2%.
Table 5 shows the results of NRI analysis of the models using age (decile), LVEF (decile), and NYHA functional class (I/II vs. III/IV) with and without late HMR data at 3 5-year mortality categories of <10%, 10% to 30%, and >30%. Compared with the baseline model without HMR, the model with HMR had net gains in reclassification of 0.073 for patients who died (p = 0.0415), 0.102 for surviving patients (p < 0.0001), and NRI = 0.175 (p < 0.0001).
The present study using multiple cohort studies demonstrated the independent prognostic value of altered cardiac sympathetic function as assessed by categorical and numeric HMR data from cardiac mIBG imaging. Compared with earlier investigations (6–11), including the ADMIRE-HF (AdreView Myocardial Imaging for Risk Evaluation in Heart Failure) study, the present results showed the long-term efficacy of risk assessment by HMR for more than 10 years. The results derived from a large number of patients not only support previous mIBG data, but also strengthen the value of quantitative mIBG markers to identify all-cause mortality risk in patients with HF. The present data also reflect the consistent, high-quality imaging procedures using high specific activity 123I-mIBG in the prospective cohort studies. More important, the results suggest that the HMR could be used to define suitable thresholds for high and low risk to facilitate therapeutic decision making.
Patient-level pooled analysis
Unlike a previous mIBG meta-analysis, which only included calculations based on published aggregate imaging and outcome results (12), this study involved collection of the individual data records for each patient with HF, thereby permitting detailed analyses of the full combined population and selected relevant subpopulations. A particular strength of the present study was the use of multiple different analysis techniques, including ROC curve, Kaplan-Meier survival, Cox proportional hazards, and NRI. Multivariate Cox analysis demonstrated independent prognostic value for mIBG imaging findings beyond that of demographic and medical history variables. NRI confirmed appropriate reclassification of patients who did and did not have lethal events. Examined from these multiple statistical perspectives, the merged data from the participating sites provided consistent evidence of efficacy for cardiac mIBG imaging equivalent to that typically obtained from a multicenter prospective trial. A further measure of the success of the analyses was the consistency of the results for subsets, including those based on different LVEF ranges. Despite the expected limitations inherent in aggregating data originally collected over a 20-year period, the consistency of the results provides validation of the methods used in performance of this pooled patient-level analysis.
Mechanisms of impaired cardiac sympathetic innervation
Cardiac sympathetic imaging can help in understanding the pathophysiological and prognostic implications of alterations of cardiac sympathetic nerve function and innervation in HF. Given that abnormalities of cardiac sympathetic innervation are common in most forms of heart disease, including ischemic and nonischemic causes (6–9,21,22), and medications affecting this system (particularly beta-blockers) are among the most effective therapies for patients with HF and ischemic heart disease, a quantitative assessment of this system should have obvious value as an additional piece of relevant medical information. Although the mIBG imaging technique has been readily available in Japan for 2 decades, the new insights provided by this pooled analysis should expand understanding of the importance of the sympathetic nerve findings. In the same way that assessment of myocardial perfusion provides functional insights to supplement the information about coronary artery disease provided by angiographic procedures, assessment of myocardial sympathetic innervation augments the morphological and mechanical indices on left ventricular dimensions and systolic and diastolic function provided by methods such as echocardiography and cardiac magnetic resonance.
Several investigations have shown that impaired cardiac mIBG activity can improve in response to current medical management using beta-adrenoceptor blockers and inhibitors of the renin-angiotensin-aldosterone system in selected patients (23,24). Therefore, this imaging technique can help to monitor response to therapy and is also likely to predict the effectiveness of such drug treatment and nonpharmacological device therapy in patients with HF (25–28). However, decreased late cardiac mIBG activity and increased WR can reflect the aggregate effect of a number of processes that cannot be individually assessed in vivo, such as reduced number of sympathetic nerve terminals, decreased tyrosine hydroxylase activity, decreased function of NE and vesicular monoamine transporters, and increased neuronal stimulation with increased NE turnover. Impaired sympathetic pre-synaptic activity in conjunction with excess and long-lasting activation of central excitatory inputs not only limits compensatory action to maintain cardiac inotropy, but also induces dysregulation of post-synaptic adrenoceptor signal transduction, potentially producing denervation supersensitivity that may contribute to lethal arrhythmogenicity in failing hearts (29).
Impact on clinical management of patients with HF
Precise risk stratification of patients with HF can contribute to implementation of prophylactic medical therapy against lethal events and appropriate selection of candidates who can benefit most from nonpharmacological device therapy using ICDs (26–28) and/or cardiac resynchronization therapy. LVEF provides little prognostic information for the one-third to one-half of patients with HF with preserved LV systolic function, and the prognostic limitations of LVEF for the prediction of sudden cardiac/arrhythmic death have been noted (30). In this study, the prognostic value of LVEF was reduced when other powerful prognosis indicators such as BNP and cardiac mIBG activity were available, whereas the prognostic value of cardiac mIBG activity was successfully demonstrated in patients with HF with reduced and normal LVEF. It is unlikely but needed in the foreseeable future to establish specific treatments for impaired cardiac sympathetic innervation independently of LVEF. Nevertheless, identification of neuronal deficiencies using mIBG neuroimaging should contribute to more effective use of currently available therapies and therefore better outcomes for patients with HF.
The value of accurate risk stratification should be greatest for patients identified as being of high or low risk for adverse outcomes. In that regard, patients with preserved myocardial sympathetic innervation (HMR ≥1.95) had an annual death rate <2%, whereas those with substantially reduced innervation (HMR <1.25) had an annual death rate >7%, demonstrating the degree of difference in prognosis that can be determined with mIBG imaging. The important observation is that user-defined high and low event rate HMR thresholds can be developed using data appropriate for specified populations. For example, in the Japanese population, an HMR threshold of 1.15 defined an even higher annual mortality rate of >10%. Given that patients with HF are considered for many drugs and invasive and expensive device treatments, there is a great need for flexible prognostic procedures to identify the highest- and lowest-risk patient groups.
Prediction of mode of death is another important clinical issue in HF management. The presented mIBG results did not predict mode of death, but rather identified high and low risk for each mode of death using late HMR or WR, suggesting the mechanistic relevance of impaired cardiac innervation and inappropriately augmented cardiac sympathetic stimulation to HF progression and sudden cardiac events. However, the present results cannot be used to indicate which mode of death will affect a particular high-risk patient. Sudden death per se is an outcome resulting from multifactorial pathologies and several triggers that could be modulated by interaction in various clinical conditions, including diabetic state, kidney function, and anemia (31). Hemodynamic instability enhances tachyarrhythmias via augmentations of afferent and efferent autonomic neuronal regulation and the renin-angiotensin-aldosterone system, leading to ischemia and myocyte injury, all of which in turn produce a vicious cycle resulting in pump failure and lethal arrhythmic events.
The present study did not directly address the influence of ischemia on arrhythmogenicity as a mechanism of sudden death. However, the possibility is supported by our previous findings (28) that impairment of myocardial perfusion is related to lethal arrhythmic events/ICD shocks independently of and in addition to altered cardiac sympathetic innervation. From the previous (26–28) and the present findings, further prospective large-cohort investigations are justified for more precise identification of patients with HF at low or high probability of sudden cardiac death and for establishing more appropriate indication for prophylactic or therapeutic use of ICDs in patient populations categorized by LVEF and background cardiac diseases.
Because only baseline data collected at the time of the original studies were available, the individual datasets did not always include complete records of patient demographics, medical history, concomitant medications, and profiles of excluded patients before enrollment in each original database. Likewise, because of the age of some studies, more recently introduced biomarkers such as BNP were only available for a minority (512 [39%]) of the 1,322 patients. A limitation of the VT classification was that history of “sustained” VT was independently confirmed at each institution, and at the time of mIBG imaging, few patients (n = 13) had ICDs that would have provided more accurate documentation of rhythm disturbances. There is also a lack of sufficient information on ICD therapy during long-term follow-up, which could affect the outcome analyses.
The prevalence of ischemic HF cause was as much as 50% lower than in many western countries (32), but this is representative of the 25% to 35% prevalence of ischemic cause in Japanese patients with HF reported in many cohorts. Nevertheless, the proportion of patients with ischemic HF was nearly identical (27%) in the dead and surviving groups, and HF cause was not significant in the multivariate analyses. The high proportion of patients with nonischemic cardiomyopathy also may have contributed to the lower than expected mortality rate compared with other clinical HF trials (32). Overall, the results are consistent with our previous study (13) and a recent prospective trial (11) that demonstrated comparable prognostic value of cardiac mIBG activity in patients with ischemic HF cause and patients with nonischemic HF cause.
Because almost all the original studies were observational, the new analyses can provide information about prognosis but not the therapeutic implications of these observations. The consistency of the aggregate results does support the desirability of a prospective interventional study to establish appropriate therapeutic strategies in patients with HF identified to be at greatest risk for lethal events by cardiac mIBG imaging. However, even without such a study, the results of the present analyses, in conjunction with those of the recent prospective mIBG trial (11), indicate that HMR values from cardiac mIBG imaging can identify patients with HF who might not be identified from conventional clinical and laboratory assessments as requiring more aggressive or more conservative treatment.
As presented in previous publications and in this study, cardiac mIBG parameters (delayed HMR and WR) have been used to monitor patient outcome and response to medical treatment. However, HMR has not been standardized among publications from various institutions and countries, primarily because of patient-based and methodology-based reasons (33,34). Fortunately, in this study consisting of multiple Japanese prospective cohorts, imaging protocols (including collimator used), dose, and specific activity of 123I-mIBG were nearly uniform (8,11,13–19). Nevertheless, the exact numeric HMR values cannot be readily applied to studies performed in other geographies and using other methodologies. Finally, a cost–benefit analysis is needed to identify specific patients in whom cardiac mIBG imaging in combination with conventional clinical and laboratory information would be most beneficial.
A patient-level pooled analysis combining 6 prospectively obtained HF databases, with 1,322 Japanese patients, clearly demonstrated the long-term prognostic value of altered cardiac sympathetic function as assessed by cardiac mIBG imaging. The continuous numeric and categorical assessment of cardiac mIBG activity using HMR provides information on the probability of survival in patients with HF independently of plasma BNP level and LVEF. By using the HMR values, the cardiologist can define the particular level of risk at which specific therapeutic interventions are deemed appropriate or unnecessary.
The presented multicenter pooled analysis study was performed by the Research Consortium consisting of the following 6 Japanese medical centers, medical statistic experts (K.N. and T.M.), and 2 advisory members (M.I.T. and A.F.J.) in cooperation with the co-researchers, clinical staff, and nuclear medicine technicians at each facility. The authors thank the following co-researchers for their efforts and cooperation in this cohort study: Akiyoshi Hashimoto, MD, PhD, Sapporo Medical University School of Medicine, Sapporo, Hokkaido, Japan; Junichi Yamazaki, MD, PhD, Department of Cardiovascular Medicine, Toho University Omori Medical Center, Tokyo, Japan; Masatake Fukunami, MD, PhD, Osaka Prefectural General Medical Center, Osaka, Japan; Kazutomo Minami, MD, PhD; and Shuichi Ichikawa, MD, PhD, Cardiovascular Hospital of Central Japan, Shibukawa, Gunma, Japan.
Dr. Travin has participated in a GE Healthcare research project. Dr. Jacobson is an employee and stockholder of GE Healthcare, which manufactures metaiodobenzylguanidine in the United States and Europe. Dr. Jacobson participated in this project as a scientist. Aside from Dr. Jacobson’s salary support, GE Healthcare provided no financial support for this study and was not involved in data analysis. 123Iodine-metaiodobenzylguanidine used in this study was commercially provided by Daiichi Radioisotope/Fujifilm RI Pharma, Tokyo, Japan. Other investigators are from academic and/or prefectural institutes or hospitals. All other authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- B-type natriuretic peptide
- heart failure
- heart-to-mediastinum ratio
- implantable cardioverter-defibrillator
- left ventricular ejection fraction
- net reclassification improvement
- New York Heart Association
- receiver-operating characteristic
- ventricular tachycardia
- washout rate
- Received July 17, 2012.
- Revision received January 28, 2013.
- Accepted February 12, 2013.
- American College of Cardiology Foundation
- Triposkiadis F.,
- Karayannis G.,
- Giamouzis G.,
- Skoularigis J.,
- Louridas G.,
- Butler J.
- Eisenhofer G.,
- Friberg P.,
- Rundqvist B.,
- et al.
- Wieland D.M.,
- Wu J.I.,
- Brown L.E.,
- Mangner T.J.,
- Swanson D.P.,
- Beierwaltes W.H.
- Jacobson A.F.,
- Senior R.,
- Cerqueira M.D.,
- et al.
- Wakabayashi T.,
- Nakata T.,
- Hashimoto A.,
- et al.
- Yamazaki J.,
- Muto H.,
- Kabano T.,
- Yamashina S.,
- Nanjo S.,
- Inoue A.
- Matsui T.,
- Tsutamoto T.,
- Maeda K.,
- Kusukawa J.,
- Kinoshita M.
- Kasama S.,
- Toyama T.,
- Kumakura H.,
- et al.
- Yamada T.,
- Shimonagata T.,
- Fukunami M.,
- et al.
- Kyuma M.,
- Nakata T.,
- Hashimoto A.,
- et al.
- Sakata K.,
- Shirotani M.,
- Yoshida H.,
- Kurata C.
- Agostini D.,
- Belin A.,
- Amar M.H.,
- et al.
- Toyama T.,
- Hoshizaki H.,
- Seki R.,
- et al.
- Nagahara D.,
- Nakata T.,
- Hashimoto A.,
- et al.
- Nishisato K.,
- Hashimoto A.,
- Nakata T.,
- et al.
- Inoue H.,
- Zipes D.P.
- Buxton A.E.,
- Lee K.L.,
- Hafley G.E.,
- et al.
- Doi T.,
- Nakata T.,
- Hashimoto A.,
- et al.
- Gheorghiade M.,
- Sopko G.,
- De Luca L.,
- et al.