Power spectral analysis of cardiovascular variability in critically ill neurosurgical patients

Abstract

Patients with brain damage exhibit a number of changes in heart rate and cardiovascular control. The aim of this study was to relate changes in autonomic cardiovascular control seen in critically ill neurosurgical patients to the quality of subsequent outcome and survival. Prospective, longitudinal, outcome study. Intensive care department of a university teaching hospital. A total of 29 consecutive neurosurgical patients admitted for ≥2 days to the intensive care department with a Glasgow Coma Scale score <13 who needed electrocardiographic and invasive arterial monitoring. Sampling of the electrocardiogram, respiratory rate, and arterial pressure into a personal computer was carried out for ≥60 mins. Power spectral analysis was then applied to the data by using a fast Fourier transformation. Arterial baroreflex sensitivity was determined as the gain of the transfer function between systolic arterial blood pressure and electrocardiograph R-R interval (RRI) variability. All surviving patients were followed up at 3 months postadmission to measure quality of outcome. There were reductions in the total power (p < .01) of RRI variability in those who subsequently died compared with those who survived. This was significant for very low frequency (p < .001) and low-frequency (LF) (p < .05) but not high-frequency (HF) bands (p = .11). Blood pressure variability, however, did not change between groups. Baroreflex sensitivity was 8.7 ± 2.2 msecs/mm Hg for patients with a good later outcome and 4.4 ± 1.5 msecs/mm Hg for patients who subsequently died (p = .03). Patients who recovered to a good quality outcome also had a raised LF/HF ratio in RRI (p = .05). A reduction in the total power variability of RRI and a lowered LF/HF ratio of the RRI are associated with a poor quality recovery or death after neurosurgical illness. A reduction in the baroreflex was specifically associated with death in this patient group.