Incremental Prognostic Value of Hippocampal Metabolic Activity for Sudden Cardiac Death in Patients With Heart Failure With Reduced Ejection Fraction

Abstract

Background: Hippocampal injury is linked to cognitive impairment and cardiac events in patients with heart failure with reduced ejection fraction. However, the predictive value of hippocampal metabolic activity (HMA) for sudden cardiac death (SCD) and the underlying mechanisms remain unclear. This study aimed to evaluate the independent and incremental predictive value of HMA over myocardial scar for SCD-related events in patients with heart failure with reduced ejection fraction and to explore the potential mechanism.

Methods: This prospective study included 270 patients with heart failure with reduced ejection fraction undergoing gated single-photon emission computed tomography and brain/cardiac ¹⁸F-fluorodeoxyglucose positron emission tomography. HMA and myocardial scarring were evaluated. The independent and incremental predictive values of HMA for SCD-related events were explored via the log-rank test, Cox model, C statistic, global χ² test, and net reclassification improvement. Mediation analysis was used to examine the direct and indirect effects of HMA on SCD-related events by setting left ventricular electrical instability as a mediator.

Results: Overall, 34 (12.6%; median follow-up, 4.3 years) patients experienced SCD-related events. Functional myocardial scarring and HMA were strongly associated with SCD-related events (all P<0.01). HMA resulted in significant incremental improvements in prognostic value (χ² increased by 6.598), discrimination (C statistics of 0.760 versus 0.715), and reclassification (net reclassification improvement, 25.6%) for SCD-related events. HMA was associated with SCD-related events via corrected QT interval.

Conclusions: Functional myocardial scarring and HMA had strong independent prognostic value in the risk stratification of SCD-related events in patients with heart failure with reduced ejection fraction. Additionally, HMA had incremental prognostic value based on myocardial scarring. HMA impairment may induce SCD-related events, partly mediated by corrected QT interval.

Keywords: 18F‐FDG PET/CT; heart failure; hippocampus; myocardial scarring; sudden cardiac death.

Figure 1. Kaplan‐Meier survival curves of adverse events.

A, Patients with a high scar burden (>28 %LV) showed a significantly increased rate of SCD‐related events compared with patients with a low scar burden (≤28 %LV). B, Survival analysis revealed that patients with low HMA (≤0.7114) had a higher rate of SCD‐related events than patients with high HMA (>0.7114). C, Survival analysis indicated no significant difference in the incidence of all‐cause mortality between patients with low and high scar burdens. D, Patients with low HMA exhibited a significantly increased incidence of all‐cause mortality compared with patients with high HMA. Adjusted covariables included age, sex, LVEF, hypertension, diabetes, and hyperlipidemia. HMA indicates hippocampal metabolism activity; HR_adj, adjusted hazard ratio; LV, left ventricle; LVEF, left ventricular ejection fraction; and SCD, sudden cardiac death.

Figure 2. Kaplan‐Meier survival curves of SCD‐related events.

Survival analysis indicated no significant difference in the incidence of SCD‐related events between patients with low and high HMA among those with low scar burden. In patients with a high scar burden, the incidence of SCD‐related events was higher in those with low HMA than in those with high HMA. The incidence of SCD‐related events was higher in patients with high scar burden and low HMA than in those with low scar burden and high HMA. The threshold for low and high scar burden was obtained using receiver operating characteristic curve analysis. The threshold for high and low metabolism is the mean±SD. Adjusted covariables included age, sex, LVEF, hypertension, diabetes, and hyperlipidemia. HMA indicates hippocampal metabolism activity; HR_adj, adjusted hazard ratio; LV, left ventricle; LVEF, left ventricular ejection fraction; and SCD, sudden cardiac death.

Figure 3. Left ventricular dyssynchrony in patients with high or low HMA.

Whether in the total population (A), patients with high scar burden (B), or patients with low scar burden (C), those with low HMA exhibited longer left ventricular electrical instability than those with high HMA. In addition, HMA was significantly correlated with QTc interval. HMA indicates hippocampal metabolism activity; LV, left ventricle; and QTc, corrected QT.

Figure 4. Left ventricular dyssynchrony/HMA in patients with low or high scar burden.

A, Left ventricular dyssynchrony (QTc, BW, SD, and entropy) was greater in patients with high scar burden than in patients with low scar burden. B, Hippocampal activity was significantly higher in patients with high scar burden than in patients with low scar burden. BW indicates bandwidth; HMA, hippocampal metabolism activity; LV, left ventricle; and QTc, crrected QT.