Diabetic ketoacidosis (DKA) is characterised by the triad of hyperglycaemia, ketosis and metabolic acidosis.1 The incidence of DKA admissions to intensive care in Australia and New Zealand increased five-fold to 5.3 per 100,000 people between 2000 and 2013 and was associated with mortality of 1.4 %.2 The majority of DKA cases occur in persons with type 1 diabetes, however DKA in those with type 2 diabetes is increasingly recognised, representing around a third of DKA hospitalisations in Sweden, even after accounting for euglycemic DKA associated with sodium glucose cotransporter-2 inhibitor (SGLT2i) use.3

The management objectives in DKA includes correction of hypokalaemia, hydration, treatment of underlying cause and insulin administration to inhibit ketoacid production and correct hyperglycaemia.4 Typically, intravenous insulin infusions are administered in DKA management due to its rapid onset and titration ability.4,5

Recently, safety concerns have been raised regarding intensive early correction of hyperglycaemia (blood glucose level [BGL] <10 mmol/L) in DKA presentations to the intensive care unit (ICU) in Australia and New Zealand in a study by Mårtensson et al.6 This study concluded that intensive early correction of hyperglycaemia in the first 24 h was independently associated with a higher risk of hypoglycaemia, hypo-osmolarity and death compared with partial early correction of hyperglycaemia.6

Currently, plasma glucose reduction rates of 2.8–4.2 mmol/L/h (50–75 mg/dL/h) and 3.0 mmol/h are recommended by DKA management guidelines from the USA and UK, respectively.1,5 These guidelines also advocate for fixed rate intravenous insulin infusions (FRIII) based on body weight. However, it should be recognised there are no head-to-head comparisons between FRIII versus BGL-based variable rate intravenous insulin infusions (VRIII) regimens on rates of early correction of hyperglycaemia, hypoglycaemia and hypokalaemia outcomes.4

Currently, there is no Australia-wide consensus regarding hyperglycaemia correction rates or the mode of intravenous insulin administration for management of DKA. Additionally, a recent review has highlighted that there are major deficiencies in evidence for the optimal management of DKA.4 To date, there are no studies that have attempted to link adverse outcomes in DKA patients admitted to Australian Hospitals, to hyperglycaemia correction rates and link outcomes to a particular insulin infusion protocol.

Another important consideration in DKA management, is potassium replacement. There is a lack of evidence regarding optimal potassium replacement, with guidelines recommending between 20 and 40 mmol potassium combined with 1 l of fluid.4

The purpose of this retrospective study was to evaluate the biochemical and clinical outcomes of patients with DKA managed in either the ICU or ward with a VRIII algorithm, with considerably reduced insulin rates compared with FRIII weight-based strategies suggested by aforementioned international guidelines.