Background: Recent work by the authors showed that over half of the incidences of hypoglycemia in hospitalized diabetics involved patients with CKD, and that 18% of patients with diabetes and end-stage renal disease (ESRD) had a hypoglycemic episode while hospitalized. The importance of reducing inpatient hypoglycemia is reflected in the Centers for Medicare and Medicaid Services (CMS) new electronic clinical quality measure (eCQM) for reporting inpatient episodes of severe hypoglycemia, which begins in 2023. We sought to characterize hypoglycemic episodes in patients with ESRD in hopes of helping to identify ways to reduce hypoglycemia in this high risk group.
Methods: Using HCA Healthcare’s clinical data warehouse of 186 hospitals, we performed a retrospective cohort analysis of diabetic patients with ESRD who were hospitalized from July 2020 through June 2022. Patients were excluded if they were missing demographic data, had a diagnosis of hypoglycemia due to medication overdose, or had been on an insulin drip during hospitalization. Hypoglycemia was defined as having a point-of-care (POC) glucose level < 70 mg/dL, and severe hypoglycemia was defined as having a POC glucose level < 40 mg/dL.
Results: A total of 10,388 patient encounters were analyzed. The median age of patients studied was 63 years. 14.44% had a hypoglycemic episode (N=1,499); 16.34% of those were severe hypoglycemia (N=245). 35.89% of hypoglycemic episodes occurred in the morning, 25.48% in the afternoon, 9.34% in the evening, and 29.29% at night. The median glucose level of patients with a severe hypoglycemic episode was 124 mg/dL vs 107 mg/dL for patients without hypoglycemia. The median maximum glucose level of patients with a severe hypoglycemic episode was 275 mg/dL vs 183 mg/dL for patients without hypoglycemia. Logistic regression analysis for severe hypoglycemia showed increased risk with higher maximum insulin glargine doses on the day of hypoglycemia (OR 1.46, p-value 0.01), higher maximum doses of short-acting insulin (OR 3.83, p-value < 0.01) and higher minimum doses of short-acting insulin (OR 4.07, p-value < 0.01). Higher total glargine doses the day prior to hypoglycemia was also associated with an increased chance of severe hypoglycemia (OR 1.18, p-value < 0.01). Patients with sepsis were also at increased risk of severe hypoglycemia (OR 3.53, p-value < 0.01). Patient age, sex, and BMI were not strongly associated with increased risk of severe hypoglycemia.
Conclusions: Several risk factors for severe hypoglycemic episodes in hospitalized diabetic patients with ESRD were higher doses of short-acting insulin the day of, higher doses of glargine the day prior, and sepsis. Additionally, patients with hypoglycemia had higher POC glucose levels, suggesting that overcorrecting for elevated glucose with high insulin doses may account for these findings and explain some episodes of hypoglycemia. This aligns with studies in the ambulatory setting showing that patients with decreasing renal function require significantly less insulin. We have implemented a text-based EHR intervention during the ordering of insulin glargine that reminds providers that ESRD patients are at increased risk of hypoglycemia and may need lower dosing than normal. Post interventional data will be analyzed for decreases in hypoglycemic episodes after more time has elapsed.