Temporal patterns of inpatient hypoglycaemia are treatment-dependent.

Background: Inpatient hypoglycaemia has been well studied around the world, and more tools are being developed to understand and predict hypoglycaemic episodes. Most published data, however, focuses on patient characteristics and predictions of whether a patient would have a hypoglycaemic episode during inpatient stay. There is a paucity of data concerning the timing, as well as types of diabetic medications used, surrounding a hypoglycaemia episode. Objectives: To characterise inpatient hypoglycaemia episodes by time and associated diabetes medications, on top of baseline patient characteristics. Design: Retrospective observational study of 425 hypoglycaemia episodes, over a 2 month period from two general internal medicine wards, in a tertiary medical hospital. Methods: A discrete hypoglycaemic episode is defined as a capillary blood glucose (CBG) reading of <4 mmol/L. Hypoglycaemic episodes were further sub-analysed by dividing them into three time frames - day (0801-1600), evening (1601-2359) and night (0000-0800). Results: In total, 425 hypoglycaemia episodes from 207 patients were analysed. Sulphonylurea (SU), premixed, basal and basal-bolus insulin regimens were associated with 31.8%, 30.4%, 15.1% and 5.9% of the hypoglycaemia episodes, respectively. All agents revealed significant intra-day differences ( Conclusion: There are marked differences in the medications associated with inpatient hypoglycaemia at differing time points. These time points offer insight into appropriate CBG testing timings for different diabetes medications. Hence, stratified monitoring and strategic 3 a.m. testing of CBG for patients on sulphonylurea and basal insulin should be considered in tackling inpatient hypoglycaemia.