Insulin use in type 2 diabetes and the risk of dementia: A comparative population-based cohort study.

<p><strong>Objective:</strong> Evidence of an increased dementia risk with insulin use in type 2 diabetes (T2DM) is weakened by confounding by indication and disease severity. Herein we reassess this association, while accounting for confounding through design and analysis. </p> <p><strong>Research Design and</strong> <strong>Methods:</strong> Using administrative healthcare data from British Columbia, Canada we identified patients diagnosed with T2DM in 1998-2016. To adjust for confounding by diabetes severity through design, we compared new users of insulin to new users of a non-insulin class, both from a restricted cohort of those who previously received two non-insulin antihyperglycemic classes. We further adjusted for confounding using: (1) conventional multivariable adjustment and (2) inverse probability of treatment weighting (IPTW) based on the high-dimensional propensity score algorithm. The hazard ratio [HR] (95% confidence interval [CI]) of dementia was estimated using cause-specific hazards models with death as a competing risk.</p> <p><strong>Results:</strong> The analytical comparative cohort included 7,863 insulin vs 25,230 non-insulin users. At baseline, insulin users were more likely to have worse health indicators. A total of 78 dementia events occurred over a median (IQR) follow-up of 3.9 (5.9) years among insulin users and 179 events occurred over 4.6 (4.4) years among non-insulin users. The HR (95% CI) of dementia for insulin use vs non-insulin use was 1.68 (1.29-2.20) before adjustment and 1.39 (1.05-1.86) after multivariable adjustment, which was further attenuated to 1.14 (0.81-1.60) after IPTW weighting. </p> <p><strong>Conclusions:</strong> Among persons with T2DM previously exposed to two non-insulin antihyperglycemic medications, no significant association was observed between insulin use and all-cause dementia.  </p>