Association of acute increases in serum creatinine with subsequent outcomes in patients with type 2 diabetes mellitus treated with sodium–glucose cotransporter 2 inhibitor or dipeptidyl peptidase-4 inhibitor

Abstract Aims The frequency of an acute increase in serum creatinine (sCr) of >30%, following treatment of sodium–glucose cotransporter-2 inhibitors (SGLT2is) and its clinical implications in patients with type 2 diabetes remains unclear. Methods and results We used medical data from a multicentre health care provider in Taiwan and recruited 11 657 and 8117 diabetic patients with baseline/follow-up sCr data available within 12 weeks of SGLT2i and dipeptidyl peptidase-4 inhibitor (DPP4i) treatment from 1 June 2016 to 31 December 2018. Participants receiving SGLT2i or DPP4i were categorized by initial sCr change into three groups: >30% sCr increase, 0–30% increase, or no-sCr increase. Participants receiving SGLT2i were associated with a higher proportion of sCr increase of 0–30% (52.7 vs. 42.6%) but a lower proportion of sCr increase of >30% (5.9 vs. 9.6%) when compared with DPP4i. In contrast to DPP4i, the mean estimated glomerular filtration rate over time became stable after 24 weeks in three categories of sCr increase following SGLT2i initiation. Compared with no sCr increase, an initial sCr increase of >30% was associated with a higher risk of major adverse cardiovascular events {adjusted hazard ratio (aHR): 2.91, [95% confidence interval (95% CI):1.37–6.17]}, heart failure hospitalization (HHF) [aHR:1.91, (95% CI:1.08–3.40)], and composite renal outcome [aHR:1.53, (95% CI:1.05–2.25)] in the SGLT2i group; an initial sCr increase of >30% associated with a higher risk of HHF and composite renal outcome in the DPP4i group after multivariate adjustment. Overall, participants receiving SGLT2i were associated with a lower risk of HHF [aHR:0.64, (95% CI:0.48–0.85)] and composite renal outcomes [aHR:0.40, (95% CI:0.34–0.48)] compared with DPP4i after multivariate adjustment, and the treatment benefit was persistent across three categories of sCr increase (P interaction > 0.05). Conclusion A modest increase in serum creatinine (<30%) was common following SGLT2i initiation, and was not associated with worse clinical outcomes, therefore should not stop therapy prematurely, but a larger increase in creatinine following drug therapy was not typical and should raise concern and review of the patient.