306

Abstract

Introduction: There are 7 million undiagnosed diabetics in the US although the number in the ICU remains poorly defined. The use of hemoglobin A1C is validated for the diagnosis of diabetes and its application presents an opportunity to identify unknown diabetics in the ICU. Hypothesis: To determine the prevalence and characteristics of unknown diabetics in the ICU. Methods: A prospective observational study of 4535 consecutive patients admitted to 8 ICUs in an academic medical center and hybrid academic/community hospital. 3841 (77%) had an A1C drawn during their admission. Patients with A1C >6.5 were examined for a history of anti-hyperglycemic medications/ICD-9 diagnosis of diabetes to determine if they were uncontrolled or unknown diabetics. Patients with A1C <6.5 were examined for history of anti-hyperglycemic medications/ICD-9 diagnosis of diabetes to determine if they were non-diabetics or controlled diabetics. This led to four groups: non-diabetics, controlled diabetics, uncontrolled diabetics and unknown diabetics. Results: 451(12.9%) patients were undiagnosed diabetics,2231 (64.1%) were non-diabetics, 400 (11.5%) were controlled diabetics and 399 (11.5%) were uncontrolled diabetics. The average A1C (blood glucose)was 8.21 (156) for unknown diabetics, 5.48 (131) for the non-diabetics, 5.51 (142) for the controlled diabetics and 8.62 (165) for the uncontrolled diabetics. Glucose readings were significantly different with 75.6% of unknown diabetics within a range of 70-180 compared with 89.8% of non-diabetics, 80.4% of controlled diabetics and 68.2% of uncontrolled diabetics. Unknown diabetics also had more hyperglycemic episodes 23%, than non-diabetics 8.5% and controlled diabetics 17.3% but less than uncontrolled diabetics 29.7%. Unknown diabetics were more likely to need insulin infusions 53%, than non-diabetics 30.8% and controlled diabetics 32.5% but less than uncontrolled diabetics 56.1%. Conclusions: Undiagnosed diabetics represent a significant percentage of the ICU population and significant challenges in glycemic control and ICU resource use. The use of A1C in this population represents a opportunity to identify and engage this patient population.