Positron emission tomography absolute stress myocardial blood flow for risk stratification in nonischemic cardiomyopathy

Introduction Sudden cardiac death is a substantial cause of mortality in patients with cardiomyopathy, but evidence supporting implantable cardioverter-defibrillator (ICD) implantation is less robust in nonischemic cardiomyopathy (NICM) than in ischemic cardiomyopathy. Improved risk stratification is needed. We assessed whether absolute quantification of stress myocardial blood flow (sMBF) measured by positron emission tomography (PET) predicts ventricular arrhythmias (VA) and/or death in patients with NICM. Methods In this pilot study, we prospectively followed patients with NICM (left ventricular ejection fraction <= 35%) and an ICD who underwent cardiac PET stress imaging with sMBF quantification. NICM was defined as the absence of angiographic obstructive coronary stenosis, significant relative perfusion defects on imaging, coronary revascularization, or acute coronary syndrome. Endpoints were appropriate device therapy for VA and all-cause mortality. Subgroup analysis was performed in patients who had no prior history of VA (ie, the primary prevention population). Results We followed 37 patients (60 +/- 14 years, 46% male) for 41 +/- 23 months. The median sMBF was 1.56 mL/g/min (interquartile range: 1.00-1.82). Lower sMBF predicted VA, both in the whole population (hazard ratio [HR] for each 0.1 mL/g/min increase: 0.84, P = .015) and in the primary prevention subset (n = 27; HR for each 0.1 mL/g/min increase: 0.81, P = .049). Patients with sMBF below the median had significantly more VA than those above the median, both in the whole population (P = .004) and in the primary prevention subset (P = .046). Estimated 3-year VA rates in the whole population were 67% among low-flow patients vs 13% among high-flow patients, and 39% vs 8%, respectively, among primary-prevention patients. sMBF did not predict all-cause mortality. Conclusions In patients with NICM, lower sMBF predicts VA. This relationship may be useful for risk stratification for ventricular arrhythmia and decision making regarding ICD implantation.