Endocrine Abstracts

Background: Stress induced hyperglycemia is common during acute illness and is related to poorer clinical outcome.

Objective: To evaluate the correlation between plasma glucose elevation from baseline and clinical outcome during acute illness.

Design: Retrospective, observational cohort of patients presenting to emergency department with plasma glucose concentration >500 mg/dl.

Participants: 661 patients visiting the emergency department (ED) of Taipei veterans general hospital between July 1, 2008 and September 30, 2010 with plasma glucose >500 ml/dl were enrolled.

Measurements: Systolic blood pressure, heart rate, plasma glucose, white blood cell, neutrophil count, hematocrit, blood urea nitrogen, serum creatinine, liver function test, plasma glucose concentration were collected at initial presentation to emergency department. HbA1c data within 6 months was reviewed from our hospital database. GAR derives from plasma glucose divided by HbA1c.

Results: GAR of the deceased is substantially higher than survived (81.0±25.9 vs 67.6±25.0, P<0.001). There is a trend toward increased 90-day mortality in groups with higher GAR (Log-Rank test for trend P<0.0001). In multivariable Cox regression analysis, GAR is significantly related to 90-day mortality (Hazard ratio [HR] per 1-S.D. increase: 1.41, 95% CI: 1.22–1.63, P<0.001), but not plasma glucose (HR per 1- S.D. increase: 0.89, 95% CI: 0.70–1.13, P=0.328). Rate of intensive care unit (ICU) admission and mechanical ventilator use also elevated in higher GAR groups in a graded manner (Linear trend P<0.001 for ICU admission, linear trend P=0.026 for mechanical ventilator use).

Limitation: Randomization study design is not feasible.

Conclusion: GAR independently predicts 90-day mortality, intensive care unit admission, and frequency of mechanical ventilator requirement. It is a better predictor for patient outcome compared to plasma glucose in subjects with extremely high glucose.