Influence of individual characteristics on outcome of glycemic control in intensive care unit patients with or without diabetes mellitus

Abstract

Objective: To clarify the relationship of patient and critical illness characteristics (including any history of diabetes mellitus) to glycemic control with insulin and hospital mortality.

Patients and methods: A case-control descriptive study was performed of patients admitted to a tertiary-care center multidisciplinary closed intensive care unit (ICU) at Mayo Clinic Hospital in Phoenix, Ariz, between January 1, 1999, and December 31, 2003, after implementation of a glycemic management protocol. Hospital mortality, the primary outcome, was examined in nondiabetic and diabetic ICU patients receiving insulin and in patients not requiring insulin (control group).

Results: Of 7285 patients, 2826 (39%) required insulin, 1083 of whom (15% of total) had a history of diabetes mellitus. The control group had a median (10th-90th percentile) glucose level of 118 mg/dL (range, 97-153 mg/dL) and a 5% mortality rate. The median glucose level was 134 mg/dL (range, 110-181 mg/dL) in nondiabetic patients and 170 mg/dL (121-238 mg/dL) in diabetic patients (P<.001), whereas mortality rates were 10% and 6%, respectively (P<.001). Compared with nondiabetic survivors, nondiabetic nonsurvivors had longer periods with glucose levels greater than 144 mg/dL. Diabetic nonsurvivors vs diabetic survivors had longer periods with glucose levels greater than 200 mg/dL. Poor glycemic control in nondiabetic patients was associated with increased insulin requirement and increased mortality. Critical illness characteristics that predicted poor glycemic control were advanced age, history of diabetes, cardiac surgery, postoperative complications, severity of illness, nosocomial infections, prolonged mechanical ventilation, or concurrent medications.

Conclusions: Critical illness characteristics determined glycemic control and clinical outcome in ICU patients. Acute insulin resistance was associated with worse outcomes in nondiabetic patients. Although critical illness characteristics influenced glycemic control, future evaluation of the effect of insulin administration and optimal glycemic control in ICU patients is necessary.