Optimal values for oxygen transport during hypothermia in sepsis and ARDS

Mild hypothermia (33 degrees C to 35.5 degrees C) is reported to improve oxygenation and survival in patients with lung failure (1). Although hypermetabolism may account for about 50% of the ventilatory demand in ARDS patients, the concept of reducing oxygen consumption (VO2) by lowering metabolic rate, has only recently gained attention (2). Our study was aimed to test whether mild hypothermia established by continuous veno-venous haemofiltration (CVVHF), could optimize values for oxygen kinetics in ARDS patients. Overall, we recruited 27 patients with ARDS and sepsis. Prior initiation of CVVHF patients had to meet the following criteria: a) Murray score > 2.5, and hypoxaemia with PaO2/FIO2 < 200, b) hyperthermia of > 38 degrees C, c) cardiovascular instability requiring inotropic support. Evaluation of cardio-respiratory data was performed within four different phases (I = before, II + III during and IV = after CVVHF) every 6 hours. Core temperature as derived from the thermistor of pulmonary artery catheter was aimed to be between 35.0 degrees C and 36.5 degrees C. Optimal values for oxygen delivery (DO2) (> 550 mL/min/m2) and VO2 (> 160 mL/min/m2) were defined according to Shoemaker and achieved by fluid loading, transfusion and inotropic support (3). Septic shock occurred in 10 of 14 nonsurvivors (nons) and 2 of 13 survivors (surv). Mean values for DO2 and VO2 were calculated at different body temperature ranges. While at 37 degrees C DO2 was identical between surv and nons, (663 +/- 128 versus 666 +/- 127 means +/- SD) moderate hypothermia led to a small decrease of DO2 in surv and a significant decrease in nons (632 +/- 134 versus 605 +/- 128 mL/min/m2) at 35 degrees C. Concerning VO2 during hypothermia, there was a significant drop in nonsurvivors while in survivors the decrease was less pronounced. We could demonstrate a decrease in DO2 and VO2 during mild hypothermia during CVVHF. However, decreases in nonsurvivors were more pronounced than in survivors. These results suggest that the inability to achieve optimal values for DO2 and VO2 during mild hypothermia induced by CVVHF could serve as a prognostic sign for fatal outcome. Although oxygen consumption is decreased during hypothermia, hypoxaemia may result due to alterations of the oxygen transport on a cellular basis. The relationship between oxygen transport and temperature during CVVHF therefore deserves further studies.