ABSTRACT
BACKGROUND: Diabetic ketoacidosis (DKA) is an acute metabolic condition, sometimes requiring admission to an intensive care unit (ICU). AIMS: To investigate the outcomes of DKA patients admitted to a hospital with restricted ICU capacity. METHODS: We included all DKA patients above age 18 who were admitted to a tertiary hospital during 2004-2017. We conducted multivariate logistic regression analysis adjusted for ICU bed availability to analyse parameters associated with ICU admission, and a composite outcome of mortality, DKA recurrence and mechanical ventilation. RESULTS: Among 382 DKA patients in our cohort, 94 (24.6%) were admitted to the ICU. The in-hospital mortality was 4.7%. Low bicarbonate (<10 mmoL/L) and pH (<7) levels at presentation were associated with ICU admission (P < 0.001 for both). In multivariate models availability of beds in the ICU was not associated with ICU admission, mortality or DKA recurrence of any type. CONCLUSION: In a setting of limited ICU capacity, DKA treatment does not necessarily require admission to the ICU. When the rising rates of diabetes mellitus and the associated elevated rates of DKA are taken into account, our results highlight the importance of including step-down units when devising local protocols for care of these patients.