Associations between lipid parameters and depression with the moderating role of inflammation: Network analyses of 87,636 UK Biobank participants

Background

Studies investigating lipid–depression associations have reported inconsistent findings, and few have considered the interrelationships among lipid parameters. Additionally, the moderating role of inflammation in these associations remains unclear. This study aims to (1) assess lipid–depression associations after controlling for other lipid parameters, and (2) explore the moderating effect of inflammation.

Methods

The study included 87,636 participants from the UK Biobank. A pairwise mixed graphical model (MGM) network was constructed in the entire sample to examine the prospective associations between seven lipid parameters and depression. The moderating role of inflammation was explored by: (1) estimating pairwise MGM networks for groups stratified by high-sensitivity C-reactive protein (hs-CRP) levels, (2) conducting network comparison tests (NCTs), and (3) constructing moderated network models (MNMs).

Results

In the full-sample network, edges connecting triglycerides (TG, edge weight = 0.034), high-density lipoprotein cholesterol (HDL-C, edge weight = −0.011), low-density lipoprotein cholesterol (LDL-C, edge weight = −0.009), and lipoprotein(a) (Lp(a), edge weight = 0.012) with depression were identified. The non-elevated and elevated hs-CRP networks exhibited opposite signs for the Lp(a)–depression edge. In NCTs, HDL-C–depression and apolipoprotein A1 (Apo A1)–depression showed significant differences in edge weights. hs-CRP moderates the associations of HDL-C, Apo A1, and Lp(a) with depression in MNMs.

Conclusions

After controlling for other lipid parameters, TG and Lp(a) are positively, while HDL-C and LDL-C are negatively associated with depression in the overall study population. Inflammation moderates the associations of HDL-C, Apo A1, and Lp(a) with depression. Our study provides further evidence for the lipid–depression associations.