OBJECTIVES: : The mechanisms by which correcting hyperglycemia with exogenous insulin improves mortality and morbidity in critically ill patients remain unclear. We designed this study to test the hypothesis that relative endogenous insulin deficiency is associated with adverse outcomes in critical illness related to hyperglycemia.

DESIGN: : Prospective controlled animal study.

SETTING: : University research laboratory.

SUBJECTS: : Male C57BL/6J mice, 8-12 wks old.

INTERVENTIONS: : Spontaneously breathing mice were instrumented with chronic indwelling arterial and venous catheters. After a postoperative recovery period, endotoxemia was initiated with intra-arterial lipopolysaccharide (1 mg/kg) in the presence of dextrose infusion (100 microL/hr). Insulin secretion was blocked with diazoxide (2.5-30 mg/kg/day). Mice were monitored continuously for 48 hrs with blood sampled serially for blood glucose and plasma insulin determinations.

MEASUREMENTS AND MAIN RESULTS: : In both saline- and glucose-infused mice, lipopolysaccharide administration induced transient hemodynamic instability without significant impact on mortality. In the saline-infused group, lipopolysaccharide administration caused a transient reduction in blood glucose and in circulating insulin. However, in glucose-infused mice, lipopolysaccharide induced a large and unexpected increase in circulating insulin without significant alteration in blood glucose. Blockade of insulin secretion in response to lipopolysaccharide in the presence of exogenous glucose precipitated marked hyperglycemia and resulted in >90% mortality. In a subanalysis of animals matched for the degree of hyperglycemia, nonsurvivors had markedly lower insulin levels compared with survivors (3.5 +/- 0.8 ng/dL vs. 9.3 +/- 1.4 ng/dL; p < .004).

CONCLUSIONS: : Endogenous insulin deficiency in the face of hyperglycemia is associated with mortality in a mouse model of lipopolysaccharide-induced critical illness.