Comparative Rate Of Amputations Among Type 2 Diabetes Patients Starting Different Sglt2 Inhibitors

Prescribing of SGLT2 inhibitors (SGLT2i) has been on the rise. A 2016 FDA safety alert raised concerns of a possible risk for lower extremity amputations for patients initiating treatment with SGLT2i. We used a large national claims database to compare amputation events among patients newly prescribed one of 3 different SGLT2is: Canagliflozin (Cana), Dapagliflozin (Dapa), Empagliflozin (Empa). Study included adults with a first prescription fill for a SGLT2i between January 2014 and May 2017. We estimated adjusted incident rate ratios for any subsequent non-traumatic amputation over a mean of about 16.7 months of follow up using a multi-variable Poisson regression model adjusting for demographics, insurance, year of start, and clinical covariates (prior non-traumatic amputation and Diabetes Complication Severity Index (DCSI)). Among 43017 new SGLT2i starts, 64% (N=27528) were Cana, 22% (N=9487) were Empa and 14% (N=6002) were Dapa. Mean age was 58.0 ± 11.4 years and 42% were women. Use of Empa increased over time (1.3% of new starts in 2014 to 45.3% in 2017). 14% of Dapa patients had Medicare, compared with 32% of Cana and 35% of Empa. Unadjusted event rates (per 1000 patient years) were 2.81 for Cana, 2.71 for Dapa, and 2.43 for Empa. In the adjusted model, no differences in amputation rates for Cana (IRR 1.26 CI [0.81, 1.97]) or Dapa (IRR 1.31 CI [0.72, 2.36]) when compared to Empa. Significant predictors were prior amputation (IRR 21.44 CI [11.88-36.69]) and DCSI score of 3 or higher (IRR 8.44 CI [5.55-12.82]). Non-traumatic amputation rates for patients on SGLT2is were low and did not differ among Cana, Dapa or Empa. Limitations include short follow-up and non-randomized study design. Understanding the risk of amputation in patients with diabetes related complications, who may otherwise benefit from SGLT2i use, is needed especially given benefit in cardiovascular outcomes and mortality. More research should investigate if patients’ risk of amputation can be predicted prior to initiation. Disclosure S. Hakimian: None. A. Wallia: Research Support; Self; Eli Lilly and Company, Novo Nordisk Inc., UnitedHealth Group. R. Kang: Research Support; Self; UnitedHealth Group. A.J. Cooper: None. M. O’Brien: Research Support; Self; UnitedHealth Group. D.T. Liss: None. S.A. Harris: None. M. Cherupally: None. E.D. Parker: None. R.T. Ackermann: Research Support; Self; National Institute of Diabetes and Digestive and Kidney Diseases, UnitedHealth Group. Funding UnitedHealth Group (SP0036847)