# Characteristics and outcomes of diabetes emergencies in nonagenarians admitted to ICU: a binational retrospective cohort study

## TL;DR

This study examines diabetes emergencies in nonagenarians admitted to ICU in Australia and New Zealand, finding generally favorable outcomes despite high mortality risks.

## Contribution

The study provides the first detailed description of nonagenarians with diabetic emergencies in ICU, offering insights for evidence-based triage decisions.

## Key findings

- Diabetic ketoacidosis (DKA) patients had lower odds of acute kidney injury compared to hyperosmolar hyperglycemic state (HHS) patients.
- ICU and hospital length-of-stay did not differ significantly between DKA and HHS groups.
- Mortality rates were higher in HHS patients compared to DKA patients, but survival over 48 months did not differ between groups.

## Abstract

Nonagenarian patients admitted to Intensive Care Units (ICU) are expected to rise with an ageing population. However, diagnosis-specific data is lacking in this cohort to guide clinical decisions. This includes diabetic emergencies, namely hyperosmolar hyperglycemic state (HHS) and diabetic ketoacidosis (DKA), which carry substantial morbidity and mortality. We sought to determine the incidence, clinical characteristics, and outcomes of nonagenarians admitted to ICUs across Australia and New Zealand (ANZ) with DKA and HHS to provide foundations for evidence-based prognostication and resource allocation.

We conducted a binational multi-center retrospective analysis of nonagenarian patients admitted to ANZ ICUs with DKA or HHS between 2018-2024. Data was sourced from the Australian and New Zealand Intensive Care Society Centre for Outcome and Resource Evaluation. Our primary outcome was to determine the incidence of nonagenarians ICU admissions for diabetic emergencies. Baseline characteristics, physiological and biochemical data, length-of-stay (LOS) and mortality were compared between groups. Multivariable regression models were used to explore associations between diagnosis, complications, interventions, and ICU and hospital LOS. Cox proportional hazards models assessed effects of diagnosis on ICU and hospital LOS, and mortality.

19,078 nonagenarian patients were admitted to an ANZ ICU during the study period, and 86 (0.45%) were admitted with DKA (55 (64.0%)) or HHS (31 (36.0%)). Pre-admission diabetes related complications were more prevalent in HHS compared to DKA (100% vs 80%, P = 0.004), with no other significant differences in demographic, clinical and biochemical measures. DKA patients had significantly lower odds of developing AKI compared to HHS (adjusted OR 0.29, 95% CI 0.08 to 0.98; P = 0.046). ICU and hospital LOS did not differ significantly between groups. ICU mortality occurred in 1 (1.8%) DKA patient and 3 (9.7%) HHS patients, and in-hospital mortality in 7 (12.7%) and 6 (19.4%) patients, respectively. Survival over 48-months did not differ between groups (log-rank P = 0.790).

Our findings suggest favourable outcomes can be achieved in suitable nonagenarians with reversible endocrine emergencies through ICU admission. We provide the first detailed description of this cohort, where future functional outcomes and quality of life assessment post-ICU admission could further inform triage decision-making.