Kelli Rumbaugh, PharmD, BCPS, BCCCP
Vanderbilt University Medical Center
Nashville
Disclosure information not submitted.
Sara Scott, PharmD
PGY2 Oncology Pharmacy Resident
The University of Kansas Health System, United States
Disclosure information not submitted.
Title: Risk Factors for Hypoglycemia during Insulin Infusions in Critically Ill Surgical Patients
Background: Continuous insulin infusions are commonly used for glycemic control in critically ill patients. Several studies of examined the safety and efficacy of insulin infusions targeting blood glucoses (BG) of 80-110 mg/dL; however, there is a lack of data looking at the safety of using insulin infusions targeting euglycemia. The purpose of our study is to examine the risk factors for severe hypoglycemia (SH) in critically ill surgical patients on an insulin infusion with a target blood glucose of 70-150mg/dL
Methods: This is a single-center, retrospective, IRB approved, cohort study conducted from 1/1/18 - 7/31/20. Patients were included if they were ≥ 18 years, admitted to the cardiac or surgical ICU, and given an insulin infusion. Patients were excluded for an admission diagnosis of DKA. The primary outcome was the incidence of SH defined as BG < 40mg/dL. Secondary outcomes included incidence of hyperglycemia, glucose variability, glucose sourse, and mortality. Logistic regression was used to examine risk factors for SH and mortality. Statistical analysis was performed using SPSS v27.
Results: A total of 4,556 patients and 262,250 BGs were included. There were 328 BGs (0.1%) that were < 40mg/dL, and 195 patients (4.3%) had at least one severe hypoglycemic event. The hypoglycemic patients were more likely to have diabetes (38.5% vs 31.5%,p=0.04), higher peak lactate (5 vs 2.9, p< 0.001), require vasopressors (88.7% vs 68.8%,p< 0.001), CKD (32.8% vs 23.2%,p=0.002), and AKI (80% vs 47.3%,p< 0.001) compared to patients who did not experience SH. There was no difference in glucose source (p=0.09), and glucose variability (p=0.06). A majority of BGs were 71-150 mg/dL (70.5%). Independent predictors for SH were peak lactate (OR 1.09, 95%CI 1.06-1.13), AKI (OR 2.857, 95%CI 1.95-4.19), duration of insulin infusion (OR 1.1, 95%CI 1.06-1.14), and history of DM (OR 1.45, 95%CI 1.05-2.02). SH was an independent predictor for mortality (OR 1.91, 95%CI 1.27-2.88).
Conclusion: Insulin infusions can be used to maintain euglycemia in critically ill surgical patients and result in a low incidence of SH. However, caution should be given to patients with renal impairment, a history of DM, higher severity of illness, and require long durations of insulin infusion as they may be at an increased risk for SH.