HDL-Related Parameters and COVID-19 Mortality: The Importance of HDL Function

Abstract

COVID-19, caused by the SARS-CoV-2 coronavirus, emerged as a global pandemic in late 2019, resulting in significant global public health challenges. The emerging evidence suggests that diminished high-density lipoprotein (HDL) cholesterol levels are associated with the severity of COVID-19, beyond inflammation and oxidative stress. Here, we used nuclear magnetic resonance spectroscopy to compare the lipoprotein and metabolic profiles of COVID-19-infected patients with non-COVID-19 pneumonia. We compared the control group and the COVID-19 group using inflammatory markers to ensure that the differences in lipoprotein levels were due to COVID-19 infection. Our analyses revealed supramolecular phospholipid composite (SPC), phenylalanine, and HDL-related parameters as key discriminators between COVID-19-positive and non-COVID-19 pneumonia patients. More specifically, the levels of HDL parameters, including apolipoprotein A-I (ApoA-I), ApoA-II, HDL cholesterol, and HDL phospholipids, were significantly different. These findings underscore the potential impact of HDL-related factors in patients with COVID-19. Significantly, among the HDL-related metrics, the cholesterol efflux capacity (CEC) displayed the strongest negative association with COVID-19 mortality. CEC is a measure of how well HDL removes cholesterol from cells, which may affect the way SARS-CoV-2 enters cells. In summary, this study validates previously established markers of COVID-19 infection and further highlights the potential significance of HDL functionality in the context of COVID-19 mortality.

Keywords: COVID-19; HDL; NMR metabolomics; cholesterol efflux capacity; lipoprotein profiling.

Figure 1

Univariate statistics of NMR-measured parameters in COVID-19 patients compared to non-COVID-19 pneumonia patients. (A) shows a volcano plot with significant parameters, which are lower or higher compared to non-COVID-19 pneumonia patients. In (B), the individual comparisons of the parameters are shown. HDCH, HDL cholesterol; HDFC, HDL-free cholesterol; H1FC, HDL1-free cholesterol; SPC, supramolecular phospholipid composite; p.d.u., procedure defined units. ** p < 0. 01, *** p < 0.001.

Figure 2

Metabolomic assessment of metabolites in COVID-19. (A) Multivariate data analyses of all parameters measured by NMR spectroscopy with orthogonal partial least squares discriminant analyses (OPLS-DA) for differentiation between COVID-19 (red) and non-COVID-19 pneumonia patients (green). (B) Permutation validity test: correlation coefficient R²Y = 0.413 (p = 0.003) and cross-validation score Q² of 0.134 (p = 0.008). (C) Variable of importance projection scores to obtain the contribution of the parameters to the model. (D) Comparison of the most prominent changes between the two groups. SPC, supramolecular phospholipid composite; Glyc, glycoprotein; HDPL, HDL phospholipids; HDCH, HDL cholesterol; HDA1, HDL apolipoprotein A-I; p.d.u., procedure defined units. ** p < 0. 01, *** p < 0.001.

Figure 3

Forest plots showing the alterations observed in lipoprotein parameters between the COVID-19 and non-COVID-19 pneumonia control groups. (A) Main classes of lipoproteins. (B) Lipoprotein particle numbers. (C) Triglyceride distribution in lipoprotein classes. (D) Total cholesterol distribution in lipoprotein classes. (E) Free cholesterol distribution in lipoprotein classes. (F) Phospholipid distribution in lipoprotein classes. (G) Apo-B protein distribution in lipoprotein classes. (H) ApoA protein distribution in HDL classes. The fold changes to non-COVID-19 pneumonia controls and the corresponding confidence intervals were calculated. Mann–Whitney-U-test-derived p-values were corrected according to Benjamini–Hochberg’s procedure. Parameters that remained significant after correction are shown in red.

Figure 4

Heatmap representing the correlations between selected variables with CEC and clinical data. Each cell of the heatmap represents a pairwise Spearman correlation between the two parameters indicated in the respective row and column. Correlations that reached significance after the Bonferroni correction are indicated with the corresponding Spearman correlation coefficient in bold. Non-significant correlations are not highlighted.

Figure 5

Survival hazard ratios (HRs) per 1 SD increase and 95% confidence intervals (CIs) derived from Cox regression analyses. The association of cholesterol efflux capacity with mortality risk has already been published [33] and was adjusted for age, sex, and HDL-C. The other NMR-derived parameters were adjusted for age and sex.