Heart rate variability and nonlinear analysis of heart rate signals of patients during their stay in a multidisciplinary intensive care unit

Neuroautonomic modulation of heart rate was assessed in patients admitted to a multidisciplinary ICU. Metrics used were performed in 47 patients and included power spectral analysis, approximate entropy (ApEn) estimation and detrended fluctuation analysis (DFA). The resulting values were compared with a previously validated measure of severity of illness, the sequential organ failure assessment score (SOFA). We found decreased variability and approximate entropy and increased DFA scaling exponent /spl alpha/ in patients with the highest SOFA, incidence of death and length of stay. Correlations concerning /spl alpha/ exponent were stronger for surgical patients with past medical history while decreased ApEn was associated with increased morbidity of surgical patients without medical history.