The association between antidepressant treatment and rates of insulin initiation in comorbid depression and type 2 diabetes: A UK electronic health record nested case-control study

Aims: To investigate the association between antidepressant prescribing and the rate of insulin initiation in type 2 diabetes. Methods: Using UK primary care records we completed a nested-case control study in a individuals with comorbid depression and type 2 diabetes. Cases were defined as individuals initiating insulin, controls were individuals remaining on oral antidiabetic medication. We used conditional logistic regression to estimate incident rate ratios (IRR) and the 95% confidence intervals (CI) for the association between antidepressant prescribing and initiating insulin. We adjusted for demographic characteristics, comorbidities, health service and previous medication use. Results: We included 11,862 cases who initiated insulin, and 43,452 controls. Increased rates of insulin initiation were associated with any antidepressant prescription (IRR 3.78, 95% CI 3.53-4.04), longer (24+ months) durations of antidepressant treatment (IRR 5.61, 95% CI 5.23-6.03), and higher numbers (3+) of different antidepressant agents prescribed (IRR 5.72, 95% CI 5.25-6.24). There was no difference between recent and nonrecent antidepressant prescriptions, or between different antidepressant agents. Conclusions: Antidepressant prescribing was highly associated with the initiation of insulin therapy. However, this may not indicate a direct causal effect of the antidepressant medication itself, and may be a marker of more severe depression influencing diabetic control.