Risk assessment studies use a suite of nominally independent noninvasive heart rate metrics, often brought together in a statistical model to compute a risk score. The ongoing need to noninvasively identify the higher risk patients requiring more invasive investigations/interventions drives the search for better noninvasive predictive metrics, with increased sensitivity. Many varieties of autoregulatory malfunction occur within the cardiovascular system; thus, it seems a daunting challenge to build predictive models that account for all potential modes of failure. Auto-entrainment (AE) methodology was developed to help address this challenge. AE methodology tests intrinsic capacity to maintain a stable and coherent oscillatory dynamic of autoregulatory control via respiratory entrainment of the blood pressure and heart period. Using cardiovascular death (n=18) at follow-up (1.5 years) as the end point, analysis of AE measurements from 148 patients with heart failure revealed 2 parameters significantly predictive of death. Using logistic regression, the magnitude of systolic pulsus alternans measured during AE had predictive sensitivity of 90% (confidence interval, 62%-100% and specificity of 62% (confidence interval, 49%-74%). The capacity to maintain a stable oscillatory dynamic was measured by the fraction of the total RR-interval spectral power contained within the AE-band. This capacity had predictive sensitivity of 73% (confidence interval, 47%-99%) and specificity of 55% (confidence interval, 43%-66%). AE methodology provides a noninvasive platform to assess the integrity of cardiovascular autoregulatory control systems for risk assessment in heart failure patients.
Read full abstract