BACKGROUNDThe lowering of body temperature is a common, almost reflexive step in the daily care of septic shock patient. However, the effect of different magnitudes of fever control on the outcome of refractory septic patients with a very poor outcome is controversial and has yet to be explored.

METHODSThis prospective trial examined sixty-five refractory septic shock patients with a core temperature higher than 38.5°C. Patients were randomly assigned to a group achieving a "low temperature" range (LT group: 36.0 - 37.5°C) or to a group achieving a "high temperature" range (HT group: 37.5 - 38.3°C) by physical methods including a water-flow cooling blanket and ice packs. A target core temperature was achieved in 1 - 2 hours post-treatment, and maintained for 72 hours. Averaged values of core temperature as well as hemodynamic, respiratory, and laboratory variables were analyzed at baseline and during the first 72 hours after fever control.

RESULTSThirty-four (52.31%) patients were assigned to the LT group and thirty-one (47.69%) patients were assigned to the HT group. The mean core temperature was significantly lower in the LT group than in the HT group (36.61 vs. 37.85°C, respectively; P < 0.0001). The average heart rate (HR) (75.5 vs. 91.9 beats/min, respectively; P < 0.0001) and the mean cardiac output (CO) (5.35 vs. 6.45 L/min, respectively; P = 0.002) were also statistically significant lower in the LT group than in the HT group. The averaged serum lactate level was significantly higher in the LT group compared to the HT group (5.59 vs. 2.82 mmol/L, respectively; P = 0.008). Fibrinogen and activated partial thromboplatin time were also different between the two groups. The 28 days mortality was significantly higher in the LT group than in the HT group (61.8 vs. 25.8%, respectively; P = 0.003). A Cox-regression model analysis showed that mean core temperature during the 72 h period was an independent predictor of 28 days mortality (odds ratio (OR) = 0.42, 95%CI 0.25, 0.6; P = 0.001).

CONCLUSIONControlling fever to a lower range (36.0 - 37.5°C) may be harmful to patients with refractory septic shock by worsening tissue perfusion, compared to controlling it within a higher range (37.5 - 38.3°C). An understanding of the mechanisms responsible for these observations requires further investigation.