Objective: We aimed to characterize SGLT2 inhibitor (SGLT2i) -associated post-operative diabetic ketoacidosis (DKA) in type 2 diabetes patients.

Methods: We analyzed a territory-wide cohort of type 2 diabetes patients who underwent operations during 2015-2020. Patients who had SGLT2i prescriptions within 6 months before operations were assigned to SGLT2i group while others to control group. Inverse probability treatment weighting with propensity scores was used to balance the baseline covariates between two groups. Post-operative DKA was defined by DKA within 30 days post-operatively.

Results: In total, 151488 type 2 diabetes adults undergoing elective (3260 SGLT2i users; 105540 controls) and emergency operations (1087 SGLT2i users; 416controls) were included. SGLT2i users who underwent emergency operations had higher incidence of post-operative DKA (24.5/1000 person-years) than elective operations (1.0/1000 person-years; p<0.001) . SGLT2i exposure significantly elevated incidence of post-operative DKA (elective: incidence rate ratio [IRR]=5.57, p<0.001; emergency: IRR=5.37, p<0.001) . Among SGLT2i users, apart from emergency surgery, HbA1c>8% (64mmol/mol) and baseline insulin use increased the risk of post-operative DKA. SGLT2i users who developed post-operative DKA had worse outcomes (mechanical ventilation, intensive care and length of hospitalization; p<0.05) than those who did not. However, implementation of alert in Clinical Management System about perioperative precaution in SGLT2i users did not reduce the risk of post-operative DKA.

Conclusion: Preoperative SGLT2i use was associated with increased risks of post-operative DKA in type 2 diabetes, higher among patients who underwent emergency operations than elective ones. Clinicians should be vigilant regarding the high-risk group. Effective strategies to mitigate risks of SGLT2i-associated post-operative DKA are urgently needed given the associated morbidities.


D.T.W.Lui: None. T.Wu: None. X.Liu: None. C.Au: None. C.Lee: None. M.Fung: None. Y.Woo: None. K.C.Tan: None. C.K.Wong: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.