Endocrine Abstracts

Background: Hyperosmolar hyperglycaemic state (HHS) is an acute metabolic complication of diabetes that can lead to significant morbidity and mortality if managed incorrectly. With <1% prevalence, there is limited published literature available on HHS and most management guidelines worldwide are based solely on expert advice and opinions.

Aims: To study the precipitating causes and identify baseline practises of HHS management, to highlight areas for improvement.

Methods: This retrospective study included all patients who meet the diagnostic criteria of HHS from May to November 2021 in six hospitals across the West Midlands region of the United Kingdom. The criteria for HHS diagnosis was defined as serum osmolality >320 mOsmol/kg and glucose >25 mmol/l whereas HHS resolution was defined as either serum osmolality <300 mOsmol/kg, when fixed rate intravenous insulin infusion (FRIII) was stopped or when the clinical team documented resolution time, whichever came earliest. Osmolality was calculated using the formula: [(2x sodium) + (2x potassium) + glucose + urea, all values in mmol/l]. Data regarding precipitating causes, insulin and fluid administration, glucose and osmolality measurements, total duration of HHS and length of admission were collected. The data was then analysed using SPSS 28.0 and results were presented as frequencies, median and interquartile range (IQR) where appropriate.

Results: A total of 31 HHS episodes were identified. From these, 64.5% had the diagnosis of HHS documented in hospital records and 48.4% had serum osmolality measured. The most common precipitating causes were intercurrent illness (38.7%), suboptimal compliance to treatment (12.9%) and new onset of diabetes (9.7%). The median calculated serum osmolality at diagnosis was 343.2 mOs mol/kg (IQR: 330.1–363.9). Patients with HHS received a median of 4500 ml (IQR: 1825–8574) of fluid until resolution. 38.7% of patients had FRIII commenced within the first hour of diagnosis and 54.8% were given basal insulin alongside FRIII. The median duration of HHS was 26.7 hours (IQR: 6.7–46.8) and these people were admitted for a median of 9.0 days (IQR: 4.7–11.8). While the length of stay was similar for HHS across included hospitals, there was a significant difference in HHS duration between sites (P<0.001).

Conclusion: Our findings suggest there is an unmet need to improve the awareness of HHS identification and its management. With the difference in HHS duration between hospitals, there is scope to identify and share best practises to provide improved, uniform clinical care for people with HHS across all centres.