#4246: Use of low-dose quetiapine and the risk of diabetes: A nation-wide cohort study using a high-dimensional propensity score

Background: Quetiapine is associated with metabolic disturbances when used in high doses for treatment of schizophrenia. The use of low-dose quetiapine for other psychiatric conditions has become increasingly common due to its sedative-hypnotic and anxiolytic properties. It is not known to what degree low doses of quetiapine is associated with an increased metabolic risk.

Objectives: To investigate the association between use of low-dose quetiapine and incident diabetes.

Methods: We conducted a nation-wide, active comparator, new user cohort study of low-dose quetiapine and the risk of diabetes using benzodiazepine-related drugs (BZRDs) and selective serotoninreuptake inhibitors (SSRIs) as active comparators. Users of quetiapine and comparator drugs were identified in the Danish National Prescription Registry and matched 1:1 using a high-dimensional propensity score. Cases of incident diabetes were identified by first use of antidiabetic medications, first diagnosis of diabetes, or first HbA1cmeasurement of more than or equal to 48 mmol/mol. We computed incidence rate ratios (IRR) using both intention-to-treat and ontreatment approaches.

Results: We identified 76,735 users of low-dose quetiapine in the Danish population between 1998 and 2018. From these, we were able to match 29,869 with BZRD-users and 19,844 with SSRI-users. Comparing the unmatched cohorts, we found a slightly increased risk of diabetes after 5 years both in comparison with BZRDs (IRR=1.06, 95%CI: 1.00-1.13) and SSRIs (IRR=1.19, 95%CI: 1.10-1.27). However, this difference was not present after matching (IRR-BZRD=0.82, 95%CI: 0.75-0.89 and IRR-SSRI=1.03, 95%CI: 0.93-1.14). Moreover, the on-treatment analyses found a reduced risk of diabetes versus BZRDs (IRR 0.58, 95%CI: 0.42-0.80) but an increased risk versus SSRIs (IRR=1.76, 95%CI: 1.27-2.40). Supplementary on-treatment analysis identified an increased risk of diabetes with BZRDs versus SSRIs (IRR 1.87, 95%CI: 1.72-2.04).

Conclusions: We did not find a substantially increased risk of diabetes among low-dose quetiapine-users, neither in comparison with BZRD-users nor with SSRI-users. The observed pattern is suggestive of a potential diabetogenic effect of BZRD, which needs to be investigated further.