Tight glycemic control may favor fibrinolysis in patients with sepsis.

Abstract

To investigate whether tight glycemic control, in patients with sepsis, may restore a normal fibrinolysis by lowering plasminogen activator inhibitor (PAI)-1 levels. Prospective randomized clinical trial. Three Italian university hospital intensive care units. Ninety patients with severe sepsis/septic shock. Patients were randomized to receive either tight glycemic control (treatment group, target glycemia, 80-110 mg/dL) or conventional glycemic control (control group, target glycemia, 180-200 mg/dL). Inflammation, coagulation, and fibrinolysis markers were assessed, along with Sepsis-related Organ Failure Assessment scores, >28 days. In the whole population, at enrolment, inflammation and coagulation were activated in >80 of 90 patients, whereas fibrinolysis, as assessed by PAI-1 activity and concentration, was impaired in only 34 patients. The extent of the inflammatory reaction or of the coagulation activation was unrelated to outcome. In contrast, 90-day mortality rate of the 34 patients in whom fibrinolysis was definitely inhibited at study entry was twice that of the 56 patients in whom fibrinolysis was intact (44% vs. 21%, p = 0.02). After randomization, during the study, daily glycemia averaged 112 +/- 23 mg/dL in the treatment group and 159 +/- 31 mg/dL in controls (p < 0.001), with total daily administered insulin 57 +/- 59 IU and 36 +/- 44 IU, respectively (p < 0.001). A small, but significant, enhancement of fibrinolysis could be observed in the treatment group, as indicated by the time course of PAI-1 activity (p < 0.001), PAI-1 concentration (p = 0.004), and plasmin-antiplasmin complexes (p < 0.001). Morbidity, rated with the Sepsis-related Organ Failure Assessment score, became significantly lower (p = 0.03) in the treatment group. Fibrinolysis inhibition, in severe sepsis/septic shock, seems to have a relevant pathogenetic role. In this context, tight glycemic control seems to reduce, with time, the fibrinolytic impairment and morbidity.