Different estimation methods of spontaneous baroreflex sensitivity have different predictive value in heart failure patients

OBJECTIVE: Several methods have been developed so far to estimate cardiac baroreflex sensitivity (BRS) through the analysis of spontaneous fluctuations of systolic arterial pressure (SAP) and R-R interval. Their relative performance in predicting cardiac mortality in heart failure patients is not known. The aim of this study was to compare the prognostic value of a set of representative indexes of spontaneous BRS in these patients. METHODS: We studied 228 stable, moderate-to-severe heart failure patients with reduced ejection fraction, in sinus rhythm, who underwent an 8-min supine recording of ECG, arterial blood pressure and respiration during paced breathing (15 breaths/min). BRS was estimated according to the following methods: sequence (BRSSeq); nonparametric transfer function in the low-frequency band (BRSTF_NP-LF); parametric spectral computed in the low-frequency and high-frequency bands (BRSPS-LF and BRSPS-HF); parametric transfer function computed in the low-frequency and high-frequency bands (BRSTF_P-LF and BRSTF_P-HF); model-based closed loop (BRSCL); and bivariate phase-rectified signal averaging (BRSPRSA). RESULTS: During a median follow-up of 36 months, 45 patients experienced a cardiac event. Only BRSTF_NP-LF, BRSPS-LF, BRSTF_P-LF and BRSPRSA were significantly associated with the outcome (P < 0.01), and statistical significance remained (P ≤ 0.03) after adjusting for clinical covariates. BRSTF_NP-LF and BRSPRSA also significantly improved the risk classification. CONCLUSION: This study shows that different spontaneous BRS indexes have different predictive value in patients with heart failure. It also shows that the prognostic information of BRS estimates is linked to SAP and RR oscillations in the low-frequency band.