Temporal and spatial dispersion of human body temperature during deep hypothermia

Clinical temperature management remains challenging. Choosing the right sensor location to determine the core body temperature is a particular matter of academic and clinical debate. This study aimed to investigate the relationship of measured temperatures at different sites during surgery in deep hypothermic patients. In this prospective single-centre study, we studied 24 patients undergoing cardiothoracic surgery: 12 in normothermia, 3 in mild, and 9 in deep hypothermia. Temperature recordings of a non-invasive heat flux sensor at the forehead were compared with the arterial outlet temperature of a heartlung machine, with the temperature on a conventional vesical bladder thermistor and, for patients undergoing deep hypothermia, with oesophageal temperature. Using a linear model for sensor comparison, the arterial outlet sensor showed a difference among the other sensor positions between 0.54 and 1.12C. The 95 confidence interval ranged between 7.06 and 8.82C for the upper limit and 8.14 and 10.62C for the lower limit. Because of the hysteretic shape, the curves were divided into phases and fitted into a non-linear model according to time and placement of the sensors. During cooling and warming phases, a quadratic relationship could be observed among arterial, oesophageal, vesical, and cranial temperature recordings, with coefficients of determination ranging between 0.95 and 0.98 (standard errors of the estimate 0.691.12C). We suggest that measured surrogate temperatures as indices of the cerebral temperature (e.g. vesical bladder temperature) should be interpreted with respect to the temporal and spatial dispersion during cooling and rewarming phases.