Assessment of arterial and cardiopulmonary baroreflex gains from simultaneous recordings of spontaneous cardiovascular and respiratory variability

OBJECTIVES: In usual models of cardiovascular regulation, arterial pressure drives RR interval through a simple baroreflex, and the influence of respiration is dismissed. We examined the applicability of a trivariate autoregressive model to obtain separate values of the gain of the arterial and non-arterial, i.e. cardiopulmonary, components of the lumped baroreflex, employing spontaneous RR interval, systolic arterial pressure and respiration variability. DESIGN: We studied 30 normal subjects (age 37 +/- 1 years), both at rest and during standing, a condition known to enhance sympathetic activity while reducing venous return. Electrocardiogram was obtained by telemetry, arterial pressure by Finapres and respiration with a piezoelectric respiratory belt Data were acquired with a PC and processed with an ad hoc Windows program. METHODS: We employed an additive and a linear multivariate approach to approximate overall gain of the arterial pressure-heart beat period baroreflex (alphalumped) and of its arterial (alphaart) and non-arterial, i.e. cardiopulmonary (alphacp), components, from continuous beat-by-beat series of RR interval, systolic arterial pressure variability and respiration, without using any non-physiological intervention. RESULTS: The overall baroreflex gain at rest (alphalumped = 23.7 +/- 3.4 ms/mmHg) was subdivided into arterial (alphaart = 5.2 +/- 1.0 ms/mmHg) and cardiopulmonary (alphacp = 18.5 +/- 3.2ms/mmHg) components. During active orthostatism, alphaluumped was diminished to 10.0 +/- 2.2 ms/ mmHg. In addition, standing selectively reduced alphacp to 4.8 +/- 1.3 ms/mmHg, while alphaart was not significantly changed. CONCLUSIONS: A trivariate autoregressive model, that considers explicitly the influence of respiration, can subdivide overall, lumped, arterial pressure-heart period baroreflex gain, into two separate components, alphaart and alphacp. Only the latter is reduced by active orthostatism