A pooled analysis of multicenter cohort studies of ¹²³I-mIBG imaging of sympathetic innervation for assessment of long-term prognosis in heart failure

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.