Abstract
Background: Covid-19 and diabetes have a bidirectional relationship [1], with dyslipidaemia further aggravating disease severity [2]. This study investigates dyslipidaemia in diabetes with prior Covid-19 infection.
Methods: Nuclear Magnetic Resonance (NMR) and Liquid Chromatography-Mass Spectrometry (LC–MS) profiling of blood samples from 251 individuals yielded six data blocks with 1,110 metabolic variables, including 112 lipoproteins and 937 lipids, with 34 cytokines quantified by a multiplex immunoassay. A total 194 individuals with complete data were included and classified into four groups for analysis: control (n = 11), diabetes (n = 27), post-Covid (≥30 days since SARS-CoV-2 infection, n = 78) and diabetes with post-Covid (n = 78). Data were modelled using OnPLS for feature selection [3], followed by OPLS-DA to reveal metabolic alterations [4].
Results: OPLS-DA effectively discriminated diabetes from control (CV-AUROC = 0.63), identifying 102 altered metabolites, including 94 lipids (p < 0.05), and diabetes with post-Covid from post-Covdi (CV-AUROC = 0.83), identifying 614 altered metabolites, including 542 lipids and 54 lipoproteins (p < 0.05), revealing a broader dyslipidaemia in diabetes with post-Covid. Notably, increased small dense LDL cholesterol, phospholipids and their particle numbers, along with decreased HDL cholesterol, a marked increase in polyunsaturated triacylglycerols and elevated diacylglycerols and sphingolipids were prominently observed in diabetes with post-Covid, particularly in those with prior severe acute Covid-19 and persistent Covid-19 symptoms. These small-dense LDL particles were correlated with cytokines, such as TNF-alpha, IL-1beta and IL-17A (p < 0.05).
Conclusion: Long-term adverse effects of Covid-19 on people with diabetes are characterized by worsened atherogenic dyslipidaemia, associated with increased insulin resistance, chronic inflammation and risk of potentially debilitating and costly long-term organ complications, such as atherosclerotic cardiovascular disease.