MON-663 Prandial Insulin Dosing Based on Carbohydrate Content Does Not Significantly Improve Glycemic Control in Hospitalized Patients with Diabetes

Abstract

Introduction: Diabetes mellitus (DM) is a highly prevalent concern in hospital medicine. In-hospital hyperglycemia and hypoglycemia are common. Hospitalized patients have variable and unpredictable amounts of carbohydrate content in their meals. We hypothesized that order sets that allow for flexible dosing of prandial insulin using an insulin to carbohydrate ratio (ICR) would provide superior blood glucose control in the complex hospital environment. Methods: We performed a retrospective review of electronic medical records from a single university hospital. The hospital implemented new order sets allowing prandial insulin dosing based on an ICR and inpatient nursing staff received necessary training. We compared glucose levels over the first three days after admission in hospitalized patients prescribed prandial insulin either as a fixed dose or based on an ICR. Patients on fixed dose insulin were selected from a time 3 months prior to the implementation of ICR order sets, to avoid any bias in patient selection. Patients on ICR dosing were selected from 3 months after implementation of ICR order sets, to allow adequate time for initial implementation. Our inclusion criteria included patients admitted to both medicine and surgery services, with Type 1 or Type 2 DM, age between 18-80. Exclusion criteria included transplant patients, patients on insulin infusion, pregnancy, steroid use and dialysis. 65 patients were included in each group. Outcome measures included average blood glucose over 72 hours, fasting and postprandial hypoglycemia (<70 mg/dL) and hyperglycemia incidence (>180 mg/dL). Results: Average glucose over 3 days was 167 +/- 39 mg/dL and 162 +/- 33 mg/dL and did not differ between groups. However, a higher percentage of glucose values were in target range (70-180 mg/dL) in the fixed dosing group (67.9%) compared to ICR (62.5%, p=0.018). The incidence of hypoglycemia was low and did not differ between groups (1.2% in both). However, patients had more overall hyperglycemia with ICR dosing (36% vs 31%, p=0.018), particularly pre-lunch hyperglycemia (52% vs. 38%, p=0.007). Fasting glucose was similar between groups. Conclusions: In conclusion, our study demonstrated that prandial insulin administered based on ICR did not improve overall glycemic control or reduce incidence of hypoglycemia in hospitalized patients with diabetes. In fact, overall and post-meal rates of hyperglycemia were generally higher with ICR dosing. This could be related to inaccurate counting of carbohydrates or delayed timing of insulin administration as it was given after the food was consumed. Additional studies are needed to further evaluate the effectiveness of flexible meal dosing and the impact on patient satisfaction and staff workload.