Lay Summary

We studied whether use of antidepressant and antipsychotic medications were associated with lipid levels (total, low-, and high-density lipoprotein (L/HDL-C) cholesterol and triglycerides) in almost half a million people, and if CYP2D6 and CYP2C19 genes influence any such associations. We found each antidepressant studied was associated with worse levels of each lipid, antipsychotics were associated with lower HDL-C and higher triglycerides, and that the CYP2C19 intermediate metaboliser phenotype may be protective for some lipids in people taking the antidepressant sertraline. These results are important because adverse lipid levels contribute to poor health and death in people with mental illnesses and lipids are not currently routinely monitored in people prescribed antidepressants, despite widespread use.

Background

Dyslipidaemia is an important cardiovascular risk factor for people with severe mental illness, contributing to premature mortality. The link between antipsychotics and dyslipidaemia is well-established, whilst evidence on antidepressants is mixed. Using UK Biobank data, we studied if antidepressant/antipsychotic use was associated with lipid levels, and if variation in CYP2D6 and CYP2C19 genes influences lipids in people taking these medications.

Methods

In this population-based study, we reviewed self-reported prescription medications to identify participants taking antidepressants/antipsychotics. Total cholesterol, LDL-C, HDL-C and triglycerides were derived from blood samples. CYP2D6 and CYP2C19 metabolic phenotypes were assigned from genetic data. Linear regression investigated associations between antidepressant/antipsychotic use with each lipid and the influence of genetic metabolic phenotypes in each medication group, adjusted for key covariates.

Results

Of 469,739 participants, 36,043 took antidepressants (53% female, 94% white ethnicity, median age 58, 17% taking cholesterol lowering medications) and 3,255 took antipsychotics (58% female, 92% white ethnicity, median age 57, 27% taking cholesterol lowering medications). Significant associations were found between use of each amitriptyline, fluoxetine, citalopram/escitalopram, sertraline, paroxetine, and venlafaxine with higher total cholesterol, LDL-C, and triglycerides and lower HDL-C, compared to participants not taking each medication. Venlafaxine was associated with the worst lipid profile (total cholesterol, mean difference: 0·21 mmol/L, 95% confidence interval [CI]: 0·17 to 0·26, p<0·001). Antipsychotic use was significantly associated with lower HDL-C and higher triglycerides. In participants taking sertraline, CYP2C19 intermediate metabolisers had higher HDL-C (0·05 mmol/L 95% CI: 0·01 to 0·09, p=0·007) and lower triglycerides (-0·17 mmol/L 95% CI: -0·29 to 0·05, p=0·007), compared to normal metabolisers. Post-hoc analyses of sertraline suggested these genetic effects were apparent only in participants not taking concomitant cholesterol lowering medications.

Conclusion

Antidepressants are significantly associated with adverse lipid profiles, potentially warranting baseline and regular monitoring of lipids. This sample was cross-sectional, we next seek to replicate findings in longitudinal data. Further research should investigate why the CYP2C19 intermediate metaboliser phenotype may be protective for HDL-C and triglycerides in people taking sertraline.