Identification and Management of Glucocorticoid-Induced Hyperglycemia on an Inpatient Malignant Hematology Ward: A Quality Improvement Initiative

Background: Hyperglycemia in malignant hematology inpatients has been associated with increased adverse events. Glucocorticoids (GC) are commonly used to treat hematologic malignancies, increasing the likelihood of hyperglycemia even in patients who do not have a history of diabetes. In particular, glucocorticoids may contribute to postprandial hyperglycemia. Studies identifying quality interventions in this setting are lacking.

Methods: We performed a retrospective review of all admissions to the malignant hematology ward at The Ottawa Hospital between September 1 to November 30, 2012 to document current practices for identifying and managing GC-induced hyperglycemia. Admissions were included if at least one dose of GC was given during hospital stay. We assessed glucose monitoring strategies, glycemic control quality and hyperglycemia therapies during the first 7 days of GC use, and up to 24 hours post discontinuation. Associations between adverse events of infection, readmission or Emergency Room visit within 30 days of GC initiation were assessed using regression analyses.

Results: We identified 77 encounters: the most common diagnosis was acute leukemia (27%), and the most frequent reason for admission to the hematology ward was for autologous hematopoietic stem cell transplantation (14%). Median patient age was 57, median body mass index was 23.6 kg/m², proportion of male patients was 54.5% and median length of stay was 12 days. Of the 77 encounters, 26% of patients were on GC prior to admission, 40% were discharged with a prescription for GC, and 5% had a previous history of GC-induced hyperglycemia. Only 19% of admissions had scheduled point-of-care testing of capillary glucose ordered during the first 7 days of glucocorticoid therapy, while 95% of admissions had at least one glucose measurement performed during hospital stay. Correctional scale insulin and scheduled basal or prandial insulin were only ordered in 14% and 3% of admissions respectively. One patient was assessed by the Diabetic Nurse Educator, and 3 patients were assessed through formal consultation by Endocrinology. Average fasting glucose was < 110 mg/dl in 38%, 110 – 180 mg/dl in 57%, 181 – 252 mg/dl in 3%, and > 252 mg/dl in 2%. At least one extreme hyperglycemic (>252 mg/dl) event occurred in 10% of cases. In the 48 admissions with at least one glucose measured in the non-fasting state, 10% had an average non-fasting glucose greater than 180 mg/dl. No associations between fasting glucose and adverse events were identified.

Conclusion: GC-induced hyperglycemia is common in malignant hematology inpatients. Only 19% of cases had scheduled qid point of care testing of capillary glucose. The American and Canadian Diabetes Associations recommend glycemic monitoring for at least 48 hours in those started on GC, with or without diabetes, identifying a practice gap. Our local needs assessment provides a foundation for future quality improvement interventions to enhance patient care through the development of appropriate screening programs and referral pathways.