The identification of causal relationships between hemodynamic variables in cardiovascular and cerebrovascular circulation

Abstract

The functional activities of the heart and its vessels in response to autonomic regulatory mechanisms are difficult to identify by monitoring the relationship between systolic blood pressure (SBP) and heart rate (HR) because such an analysis should consider changes in both SBP and diastolic blood pressure (DBP) in cardiovascular circulation. Thus, SBP, diastolic blood pressure (DBP), HR, and stroke volume (SV) should be analyzed via multivariate causal relationship analyses to identify the functional activity of the heart and its vessels in response to autonomic regulatory mechanisms. Subsequently, we consider the coupling between cardiovascular and cerebrovascular circulation. In particular, there are independent cerebral autoregulation (CAR) mechanisms in cerebrovascular circulation that individually control transferred cardiovascular hemodynamic variables such as SBP, DBP, HR, and SV to achieve homeostasis in cerebral blood flow (CBF). Therefore, the systemic blood flow -- which is measured according to pressure, volume, and frequency of beats -- should not be correlated with cerebral hemodynamic variables such as cerebral blood flow velocity (CBFV) and oxyhemoglobin concentration (O2Hb). To identify the functional abilities of intact cerebral autoregulation, we performed multivariate analyses using both systemic and cerebral hemodynamic variables. Head-up tilt tests (HUT) were performed on 15 males between 20 and 30 years of age to induce perturbations with orthostatic stress. During the test, we measured ECG, NIBP, ICG, CBFV, and O2Hb signals, from which we extracted beat-to-beat hemodynamic variables including HR, SBP, DBP, SV, Maximum CBFV (MxFV), Minimum CBFV (MnFV), Maximum O2Hb (MxO2Hb), and Minimum O2Hb (MnO2Hb). The causal relationships between variables were analyzed using extended Granger causality analysis (eGC, Vector autoregressive linear model basis). We also performed a multivariate analysis of the set [SBP – DBP – HR – SV] in the cardiovascular circulation. To assess couplings between systemic and cerebral circulation, combinations consisted of hemodynamic variables comprising both sets of circulatory metrics ([SBP - DBP - HR - SV - MxFV - MnFV] and [SBP - DBP - HR - SV - MxO2Hb - MnO2Hb]). The eGC values derived from each combination were first tested for reliability through bootstrap resampling, and were then analyzed for changes in causal relationships with respect to orthostatic stress using a repeated measures ANOVA test. The significance level was set at 5% and verified by post-hoc comparison with a Bonferroni correction. Analyses of cardiovascular circulation revealed that the causal relationship between HR  DBP was strongly prevalent in the first supine state, and this causality significantly decreased during the head-up tilt. Conversely, HR  SBP was not as strong as HR  DBP, but there was a clear and stable causality without significant difference in the state change. SV  SBP was less than 0.1 in the supine state, but was significantly increased in the head-up tilt state. DBP/SV  HR and DBP  SV each exhibited a causality of less than 0.1, regardless of any state change. Causality with respect to HR was confirmed only in the SBP  HR relationship, which was also stable without any difference in state change. Moreover, the causal relationship between systemic and cerebral circulation showed that the SBP  MxFV/MxO2Hb or DBP  MnFV family in the head-up tilt state had a small, but causal, relationship. However, in all other relationships, we observed a causality of less than 0.1. Although SV  CBFV/O2Hb, the causality of HR  MnFV/MnO2Hb, was not stable, a causal relation was formed without significant changes in orthostatic stress, even though the relations between HR  MnFV/MnO2Hb were not high. In addition, HR  MxO2Hb was significantly decreased during the HUT condition. In previous bivariate analyses using the SBP – HR relation, attention was focused on the effects of the feedback (BRS) loops involved in the SBP  HR relationship and feedforward loop in the HR  SBP relationship in order to identify autonomic nervous system activities only in the heart. In contrast, we identify causal relationships between hemodynamic variables through multivariate analyses using SBP, HR and additional variables of DBP and SV. Thus, the functional activities of parasympathetic and sympathetic arousal that have influence over the heart and its blood vessels were analyzed by identifying changes in causal relationships according to orthostatic stress. In addition, causal coupling between systemic and cerebral hemodynamic variables was generally shown to be relatively low. However, the effects of HR were slightly transferred to cerebral circulation, even though the causality was also low; whereas, the effects of SV were under 0.1 in all interrelationships. Thus, although cerebral circulation was independently controlled by intact cerebral autoregulation, hemodynamic variables that were influenced by parasympathetic nervous system activity were partly transferred to cerebral circulation.