Letter to the Editor: Response to Hamblin et al body mass index is inversely associated with capillary ketones at the time of colonoscopy: Implications for SGLT2i use

Hamblin et al.1 in their recent publication examine various associations with ketosis in the colonoscopy population. They make the novel observation that in this high-risk population for SGLT2 inhibitor (SGLT2i) induced ketoacidosis that there is an inverse association between body mass index (BMI) and finger prick ketone measurement in 37 SGLT2i treated individuals and 105 individuals with type 2 diabetes (T2D) undergoing colonoscopy. This observation, albeit in a small cohort, may help identify individuals who are at high risk of ketosis and worse outcomes around colonoscopy. Whilst it is logical that the association between BMI and ketosis observed predominantly in the normal range of ketones will continue into the clinically significant range this cannot be assumed. This is particularly true as Hamblin et al also demonstrate a significant association between BMI and ketosis for normoglycaemic individuals and the non-SGLT2i treated individual with T2DM in their univariate analysis, who historically are thought not to be at risk of ketoacidosis. This assumption may be predicated on the small sample size of the study, with only six individuals being noted to have ketones >0.7 mmol/L and only two individual >1.0 mmol/L. Correctly, Hamblin et al.1 conclude that larger trials examining ketosis in the colonoscopy population are required to support this hypothesis. Prompted by this observation, we performed a similar retrospective multivariate analysis in 94 SGLT2i treated individuals and 484 individuals with T2D not treated with SGLT2i who underwent a colonoscopy between 2019 and 2020 at the Queen Elizabeth Hospital in South Australia. The larger sample size enabled us to perform a sub-analysis directly assessing the factors identified by Hamblin et al.1 and those identified by our own regression analysis and compare these factors in individuals who developed clinically significant ketosis (ketones >1.0 mmol/L) and those who did not.2 In the SGLT2i treated individuals, we observed 17 cases of clinically significant ketosis (>1.0 mmol/L) equating to an incidence of 18%.