Determining rhythmicity and determinism of temperature curves in septic and non-septic critically ill patients...

A few studies have demonstrated that critically ill patients exhibit circadian deregulation and reduced complexity of different time series, such as temperature.

In this prospective study, we enrolled 21 patients divided into three groups: group A (N = 10) included subjects who had septic shock at the time of ICU entry, group B (N = 6) included patients who developed septic shock during ICU stay, and group C consisted of 5 non-septic critically ill patients. Core body temperature (CBT) was recorded for 24 h at a rate of one sample per hour (average of CBT for that hour) and during different occasions: upon ICU entry and exit in groups A and C and upon entry, septic shock development, and exit in group B. Markers of circadian rhythmicity included mean values, amplitude that is the difference between peak and mean values, and peak time. Furthermore, recurrence quantification analysis (RQA) was employed for assessing different markers of complexity of temperature signals. Patients from group C exhibited higher temperature amplitude upon entry (0.45 ± 0.19) in relation with both groups A (0.28 ± 0.18, p < 0.05) and B (0.32 ± 0.13, p < 0.05). Circadian features did not differ within all groups. Temperature amplitude in groups B and C upon entry was negatively correlated with SAPS II (r = - 0.72 and - 0.84, p < 0.003) and APACHE II scores (r = - 0.70 and - 0.63, p < 0.003), respectively, as well as duration of ICU and hospital stay in group B (r = - 0.62 and - 0.64, p < 0.003) and entry SOFA score in group C (r = - 0.82, p < 0.003). Increased periodicity of CBT was found for all patients upon exit related to entry in the ICU. Different RQA features indicating periodic patterns of change of entry CBT were negatively correlated with severity of disease and length of ICU stay for all patients.

Increased temperature rhythmicity during ICU entry was related with lower severity of disease and better clinical outcomes, whereas the more deterministic CBT patterns were found in less critically ill patients with shorter ICU stay.