Vector autoregressive modeling for analyzing feedback regulation between heart rate and blood pressure

We present a noninvasive technique for analyzing the feedback relationship between short-term fluctuations in blood pressure (BP) and heart rate (HR) in humans. Beat-to-beat BP was monitored for several minutes with a noninvasive BP measuring device, i.e., arterial tonometry, in 16 healthy subjects and 15 patients on hemodialysis (HD), and the recorded systolic BP (SBP) and the pulse wave interval (R-R interval) were fitted to vector autoregressive models according to Akaike's method. The impulse-response function was then simulated to determine the contribution of changes in R-R interval or SBP to the fluctuation of SBP or R-R interval along a time scale. In healthy subjects, when a simulated 1-mmHg impulse was applied to SBP, the R-R interval increased to approximately 3.5 ms with a delay of several beats. In HD patients the same impulse caused essentially no response in the R-R interval. A 1-ms impulse applied to the R-R interval decreased SBP, and the response was not different in healthy subjects and HD patients. In the R-R interval-to-R-R interval response and the SBP-to-SBP response, the exponential-like decay after the initial jump was more rapid in the HD patients than in the healthy subjects. The results demonstrated that this technique can provide information on the feedback regulation between fluctuations in BP and HR in patients with or without autonomic nervous system dysfunction. This technique provides a simple and practical way to estimate autonomic function in clinical medicine.