Cardiac Autonomic Neuropathy: Comparative performance of Phase-Rectified Signal Averaging methods with Cardiac Autonomic Reflex Tests and Time Domain Heart rate variability.
TL;DR
This study compares PRSA methods with conventional methods and defines cut-off values for AC and DC to diagnose CAN and finds that a decrease in DC value and an increase in AC value are associated with a higher probability of CAN.
Abstract
Background: Phase-Rectified Signal Averaging (PRSA) methods, Deceleration (DC), and Acceleration Capacity (AC), provide a comprehensive assessment of cardiac autonomic function (CAN). There are no published studies comparing PRSA methods with conventional methods such as Time Domain Heart Rate Variability (TD-HRV) or Cardiac Autonomic Reflex Tests (CARTs), nor have they described a cut-off value for AC and DC to distinguish patients with CAN (+ve) from those without CAN (-ve). Our study compares PRSA methods with conventional methods and defines cut-off values for AC and DC to diagnose CAN. Methodology: We studied two cohorts: 126 individuals with normal ventricular function (derivation cohort) and 143 individuals with Left Ventricular Dysfunction (validation cohort). These patients underwent CARTs and supine, resting ECG recordings for 2 to 3 minutes. The patients were categorized as CAN +Ve and CAN -Ve based on TD-HRV parameters and the CARTs. Two different CART criteria were studied: the All-India Institute of Medical Sciences (AIIMS-AFT) criteria and the 2011 Toronto Consensus recommendations. Patients with and without CAN were segregated by AC and DC values, and the methods were compared. The cutoff values for DC and AC were calculated using the ROC curve method from the derivation cohort and verified in the validation cohort. Results: A reduction in DC values and an increase in AC values indicate a higher chance of CAN. The cut-off values of -7 for AC and 7 for DC provide the highest accuracy in detecting CAN prevalence. Both values have an AUC of nearly 0.9. Reclassifying both Cohorts as CAN +Ve based on the derived cut-offs and comparing with the prevalence determined by conventional methods results in kappa values ranging from 0.5 to 0.7. Conclusions: A decrease in DC value and an increase in AC value are associated with a higher probability of CAN. An AC value ≥ -7 and a DC value ≤ 7 indicate good accuracy in identifying CAN.