Sodium-glucose co-transporter-2 inhibitors (SGLT2i) are associated with risk of euglycemic ketoacidosis. Guidelines recommend withholding SGLT2i prior to surgery and considering procedure delay in the presence of ketosis. Literature to support this in setting of routine outpatient colonoscopy is limited. Our aim was to clarify the incidence and range of ketosis in all individuals presenting for elective colonoscopies to help setting guidelines and threshold for concern.

This single-centre prospective study recruited patients ≥18 of age who underwent routine outpatient colonoscopies in a medium metropolitan hospital in Brisbane, Australia between August and November 2021. SGLT2i were withheld for 48 h prior and blood glucose and capillary ketone concentrations were recorded within 90 minutes before procedure commencement.

315 individuals were consecutively recruited; 179 (56.8%) were female. Sixty-nine (21.9%) had a previous diagnosis of type 2 diabetes mellitus (T2DM) and 17 (5.4%) were taking SGLT2i. The mean age was 57.79 (± 15.21). Significant ketone levels defined as >1.0 mmol/L were noted in 41 individuals (13.0%). Of these, 13 (33%) were diabetic with ketosis ranging from 1.0-4.2mmol/L. The range of significant ketosis in the 28 non-diabetics was 1.0-5.7mmol/L. Only a diagnosis of T2DM and increased fasting times (>45 mins) conferred a greater trend towards ketosis risk. Patients with T2DM as a whole were 2.06 times more likely to develop ketosis with or without SGLT2i. This did not reach statistical significance (p = 0.05).

A wide range of periprocedural ketosis commonly occurs in patients undergoing colonoscopies with or without T2DM. This phenomenon is not unique to diabetics or in those on SGLT2i. Hence, previously defined significant ketosis cut-offs are unlikely to be useful in the unique context of colonoscopies. Avoiding procedural delays and early commencement oral intake should be a priority.