Abstract
Background: Approximate entropy (ApEn) measures complexity of time-series data. A larger value of ApEn indicates higher randomness, and vice versa for a smaller value. Recently, ApEn of the body temperature curve of ICU patients with septic multiple organ dysfunction syndrome (MODS) has been reported to be inversely correlated with sequential organ failure assessment (SOFA) score, and to possibly predict ICU survival. Objectives: We performed a retrospective study to validate the findings by using the stored data in our general ICU. Methods: Adult septic patients with MODS, admitted to our ICU during the past 18 months, and in whom bladder temperature was continuously monitored at 30-min interval for more than 4 days, were retrospectively enrolled. ApEn was calculated on 68 bladder temperature data points over the preceding 33 hours at every 9:00 am. SOFA score was also determined at 9:00 am. Results: Thirty patients were included. Twenty nine of 30 patients showed an inverse correlation between ApEn and SOFA, 10 of which were statistically significant. A significant difference was seen in ApEn, maximum-SOFA score, and maximum-max-lactate level between survived (n=25) and expired (n=5) cases. However, logistic analysis showed that only maximum-SOFA score was a significant factor associated with ICU outcome. Conclusion: ApEn cannot predict the outcome of ICU patients with septic MODS.
