Background: Development of dementia is usually considered as a chronic process, and type 2 diabetes increases dementia risk. Our recent active-comparator new-user cohort study found an association of sodium-glucose cotransporter 2 (SGLT2) inhibitors vs dipeptidyl peptidase-4 (DPP4) inhibitors with a 20% lower dementia risk; however, it has not been examined whether longer SGLT2 inhibitor exposure correlates with greater risk reduction.
Objectives: This follow-up analysis aims to quantify the associations of different durations of cumulative SGLT2 inhibitor use with incident dementia.
Methods: The analysis used administrative databases housed at ICES. Residents of Ontario, Canada older than 66 years from July 1st, 2016 to March 31th, 2021 were selected. New users of an SGLT2 inhibitor or a DPP4 inhibitor who were free of either drug class in the past year entered this active-comparator new-user cohort. Incident dementia was defined using a validated algorithm. A 1-year lag period from cohort entry was applied to address protopathic bias and disease latency, and thus those with censoring or incident dementia within 1 year of cohort entry were excluded. Propensity scores were used to adjust for covariates before cohort entry. Cumulative SGLT2 inhibitor use as a time-dependent exposure variable was classified into no use (reference), “0-1 year”, “1-2 year”, “2-3 year”, and “>3 year”, and a propensity-score weighted time-dependent Cox regression model was used to calculate adjusted hazard ratios (aHRs) and confidence intervals (CIs) for incident dementia.
Results: There were 36,513 SGLT2 inhibitor new users and 70,390 DPP4 inhibitor new users at baseline, and individuals were followed for maximum 4.78 years and over a mean of 2.80 years from drug initiation. SGLT2 inhibitor use for 0-1 year did not show a significant association (aHR 0.90, 95% CI 0.77-1.04; 399 events over 50,870 person years), compared to no SGLT2 inhibitor use (1,740 events over 151,050 person years). SGLT2 inhibitor use of 1-2 years (aHR 0.77, 95% CI 0.65-0.92; 244 events over 51,111 person years), of 2-3 years (aHR 0.56, 95% CI 0.47-0.66; 219 events over 27,911 person years), and of more than 3 years (aHR 0.51, 95% CI 0.41-0.63; 129 events over 18,338 person years) showed associations with a lower dementia risk. Effect sizes increased with longer exposure durations, and the overall trend for different durations was significant (χ2=75.6, DF=4, p< 0.001).
Conclusions: The analysis provides new evidence that longer SGLT2 inhibitor use was associated with larger reduction of dementia risk in older adults with diabetes. SGLT2 inhibitor use of more than 3 years was associated with a 49% lower risk.
