Predictive Model of All-Cause Death in Patients with Heart Failure using Heart Rate Variability

Background: Short and long-duration heart rate variability (HRV) data from Holter monitoring could identify predictors of all-cause death in heart failure (HF) patients.Objectives: To build a predictive model of all-cause death in patients with HF using HRV.Methods: Retrospective study including patients with suspected or confirmed HF who were admitted for decompensated HF or syncope that underwent Holter monitoring. In analysis of augmented sympathetic tonus, we evaluated the lowest HRV in nonoverlapping 10-minutes periods throughout 24h continuous electrocardiographic signal recording (short HRV variables). Variables with p<0.01 were included in a multivariate Cox regression model to determine the occurrence of the all-cause death. Variables with statistical significance in Cox regression were chosen to build the predictive model.Results: A total of 116 patients were included, mean age of 71.9 +/- 16.3 years, 45.7% men, mean follow-up of 2.83 +/- 1.27 years. Thirty-nine deaths occurred (33.6%). By comparing survivors vs. non-survivors, the variables that showed statistical significance were lowest SDNN, lowest rMSSD, age and left ventricular ejection fraction (LVEF). In Cox regression, independent predictors of all-cause death were: age>69 years (HR 3.95, 95%CI 1.64-9.52); LVEF <= 57% (HR 4.70, 95%CI 2.38-9.28) and lowest rMSSD <= 12ms (HR 5.54, 95%CI 2.04-15.08). An integer value was assigned to each variable. Score<3 showed AUC=0.802 (95%CI 0.72-0.87).Conclusion: In HF patients hospitalized for decompensated HF or syncope, independent long-term predictors of allcause death were age, LVEF, and 10-minutes rMSSD. These findings indicate that even brief moments of high sympathetic tone can impact survival, specifically in the elderly and patients with HF with reduced ejection fraction.