Reproducibility of transpulmonary thermodilution measurements in patients with burn shock and hypothermia.

Abstract

Thermal noise, including hypothermia and short-term variations in body temperature, has been reported to influence the accuracy and reproducibility of thermodilution measurements. This variation might theoretically limit the usefulness of this technique in patients with acute burns. We therefore sought to determine the reproducibility of hemodynamic parameters derived from arterial thermal dilution in patients with acute burns and hypothermia. This was a prospective, clinical study of 50 mechanically ventilated patients with burns involving more than 25% of the body surface area that were treated at an eight-bed intensive care burn unit in a university-affiliated hospital. A total of 750 arterial thermodilution measurements were analyzed using the COLD system. Triplicate measurements of the intrathoracic blood volume, cardiac output, total blood volume, and extravascular lung water were performed at regular intervals during the first 48 hours after the thermal injury. Reproducibility was assessed by the coefficient of variation of the triplicate measurements. The correlation of variation was less than 10% at all measurement times for cardiac output, intrathoracic blood volume, and total blood volume. For the extravascular lung water, the coefficient of variation ranged from 9.5% to 12.9%. A maximum of 12.9% was found at 48 hours after burn. No correlation was found between body core temperature and the reproducibility of intrathoracic blood volume index (r = 0.145), cardiac index (r = 0.217), or extravascular lung water index (r = 0.167). The parameters derived from arterial thermodilution show a clinically sufficient reproducibility in patients with acute burns associated with thermal instability.