To the Editor:
Endothelial dysfunction is a major risk factor for cardiovascular disease, and may contribute to failure to mature arteriovenous fistulas (AVFs).1 Understanding mechanisms and risk factors associated with endothelial dysfunction in patients with advanced chronic kidney disease (CKD) may inform treatment approaches to improve clinical outcomes. High-density lipoprotein (HDL) in patients with CKD lacks typical anti-inflammatory and cholesterol efflux properties, potentially causing endothelial dysfunction; the abnormal lipid composition of HDL particles may partially explain this.2, 3, 4 HDL in CKD is differentially enriched in lipids associated with endothelial cell autophagy in basic science models and may be associated with endothelial dysfunction and fistula maturation.3,5 We aimed to evaluate associations specifically between HDL lipid composition and markers of endothelial dysfunction (brachial artery dilation, pulse-wave velocity, and venous occlusion plethysmography) and fistula maturation in patients with advanced CKD in the Hemodialysis Fistula Maturation Study. We hypothesized that greater relative HDL abundance of ceramides, sphingomyelins with long-chain fatty acids, and saturated and monounsaturated phosphatidylcholines would be associated with endothelial dysfunction, and with failure to mature AVFs.
We isolated HDL particles via ultracentrifugation; lipids were quantified using targeted liquid chromatography tandem-mass spectrometry in patients with advanced CKD and kidney failure undergoing AVF creation. Outcomes included markers of endothelial dysfunction (brachial artery dilation, pulse-wave velocity, and venous occlusion plethysmography) and failure to mature AVF without assistance within 9 months of AVF creation. Associations between HDL lipids (relative to HDL protein) and outcomes were assessed using multiple linear regression with adjustment for potential covariates; full methods are available in Supplementary Materials (Item S1).
Among 308 participants, the mean (±SD) age was 55 (14) years; 32% were women. In total, 65% were receiving dialysis, 47% had cardiovascular disease, and 56% had diabetes. Most participants used statins (53%), although few used fibrates or ezetimibe (4% and 1%, respectively) (Table S1). Unassisted fistula maturation within 9 months of placement occurred in 157 participants (51%). All lipid species were positively correlated with each other (Figure S1). Relative HDL lipid concentrations did not differ significantly between participants with kidney failure receiving dialysis and those with advanced CKD or between those who matured and failed to mature their AVF (Tables S2-S3). HDL lipids (as subclasses or individual lipids) were not significantly associated with any marker of endothelial dysfunction at the false discovery rate <5% (Table 1, Tables S4-S6). We did not observe significant associations between relative HDL abundance of lipid subclasses and AVF maturation (Table S7, Figure 1). In adjusted models, 1 of 37 lipids was associated with failure to mature AVF. However, given the large number of exposures and outcomes, we favored this to be a spurious finding (Table S8).
Table 1.
Association of Markers of Endothelial Dysfunction With Concentration of HDL Lipids (per Standard Deviation [SD] Increment), Adjusted.a
| Flow-mediated dilation (% change from baseline) |
Nitroglycerin-mediated dilation (% change from baseline) |
Carotid-femoral pulse-wave velocity (m/s) |
Carotid-radial pulse-wave velocity (m/s) |
Capacitance (mL/100 mL/10 mm Hg) |
Maximum venous outflow (mL/100 mL/min)/10 mm Hg |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Difference per SD increment | Pb | Difference per SD increment | Pb | Difference per SD increment | Pb | Difference per SD increment | Pb | Difference per SD increment | Pb | Difference per SD increment | Pb | |
| Total ceramides | –0.10 (–0.71 to 0.52) | 0.76 | –0.23 (–1.06 to 0.59) | 0.58 | –0.17 (–0.48 to 0.13) | 0.27 | –0.06 (–0.33 to 0.21) | 0.68 | –0.03 (–0.08 to 0.02) | 0.25 | –0.05 (–0.44 to 0.34) | 0.80 |
| Total sphingomyelins | 0.13 (–0.47 to 0.73) | 0.67 | –0.10 (–0.91 to 0.71) | 0.81 | –0.02 (–0.39 to 0.35) | 0.91 | –0.15 (–0.40 to 0.10) | 0.23 | –0.06 (–0.11 to –0.00) | 0.04 | –0.22 (–0.60 to 0.16) | 0.26 |
| Sphingomyelins with long fatty acids | 0.06 (–0.49 to 0.61) | 0.83 | 0.27 (–0.56 to 1.11) | 0.52 | –0.15 (–0.53 to 0.23) | 0.44 | –0.16 (–0.40 to 0.09) | 0.20 | –0.04 (–0.08 to 0.00) | 0.06 | –0.17 (–0.55 to 0.22) | 0.40 |
| Sphingomyelins with very long fatty acids | 0.19 (–0.44 to 0.82) | 0.56 | –0.50 (–1.24 to 0.23) | 0.18 | 0.14 (–0.21 to 0.50) | 0.43 | –0.11 (–0.34 to 0.13) | 0.38 | –0.06 (–0.12 to –0.01) | 0.03 | –0.24 (–0.61 to 0.14) | 0.22 |
| Total phosphatidylcholines | –0.01 (–0.59 to 0.57) | 0.98 | 0.27 (–0.50 to 1.04) | 0.50 | –0.14 (–0.47 to 0.19) | 0.40 | –0.06 (–0.29 to 0.18) | 0.63 | –0.02 (–0.06 to 0.01) | 0.22 | –0.19 (–0.57 to 0.19) | 0.33 |
| Phosphatidylcholines with 0-1 double bonds | 0.07 (–0.51 to 0.65) | 0.82 | 0.24 (–0.57 to 1.06) | 0.56 | –0.17 (–0.53 to 0.18) | 0.34 | –0.12 (–0.35 to 0.12) | 0.34 | –0.04 (–0.08 to –0.00) | 0.04 | –0.24 (–0.63 to 0.14) | 0.21 |
| Phosphatidylcholines with 2+ double bonds | –0.05 (–0.62 to 0.53) | 0.87 | 0.26 (–0.48 to 1.01) | 0.48 | –0.11 (–0.42 to 0.19) | 0.47 | –0.02 (–0.26 to 0.21) | 0.86 | –0.01 (–0.05 to 0.02) | 0.53 | –0.15 (–0.53 to 0.23) | 0.44 |
| Total lipids | –0.00 (–0.60 to 0.59) | 0.99 | 0.10 (–0.71 to 0.91) | 0.80 | –0.13 (–0.46 to 0.19) | 0.42 | –0.08 (–0.33 to 0.17) | 0.54 | –0.03 (–0.07 to 0.01) | 0.14 | –0.17 (–0.56 to 0.21) | 0.38 |
Adjusted for age, biological sex, body mass index, systolic and diastolic blood pressure, cardiovascular disease, diabetes, smoking status, C-reactive protein levels, use of statins, fibrate levels, ezetimibe levels, nitroglycerine levels, calcium channel blockers, and angiotensin converting enzyme inhibitors or angiotensin receptor blockers.
Reported P values are uncorrected; bold font indicates statistical significance at the false discovery rate <5%.
Figure 1.
Association of HDL lipid subclasses with failure to mature AVF.
HDL lipid composition was not associated with markers of endothelial dysfunction or AVF maturation in participants with advanced CKD or kidney failure undergoing placement of AVF. HDL lipid composition may not be a risk factor for endothelial dysfunction in patients with advanced CKD, contrary to our hypothesis. CKD is associated with altered lipid profiles and HDL lipid composition, including higher levels of ceramides and sphingomyelins acylated to long-chain fatty acids.3,4,6 These lipids have been associated with cardiovascular disease in general population studies and animal models,7 and may be associated with endothelial cell apoptosis and endothelial dysfunction.5 However, we found no associations between HDL lipids and endothelial dysfunction. There are multiple potential explanations. First, we were unable to evaluate all HDL lipids with our targeted approach. We did not evaluate non-HDL lipids or total plasma lipids, which may be associated with endothelial function. Alternatively, lipid molecules may exert different physiologic effects in patients with CKD compared with normal kidney function because of differing physiology. Finally, HDL lipid composition may not be associated with endothelial dysfunction in advanced CKD, although lipids may influence cardiovascular disease risk via other pathways, including inflammation and cardiomyocyte apoptosis.
We similarly found no associations between HDL lipidomics and AVF maturation. Multiple factors (including diabetes, peripheral vascular disease, hypertension, smoking, age, and endothelial dysfunction) have been suggested as predictors of AVF maturation, yet studies remain inconclusive.8 Data from the Hemodialysis Fistula Maturation Study (purposefully designed to evaluate these predictors) did not demonstrate associations of preoperative characteristics or endothelial measures with AVF maturation.9 Further investigation is needed to identify risk factors associated with AVF maturation. This study has several merits; we leveraged a dedicated study of advanced CKD with multiple rigorously-ascertained markers of endothelial dysfunction and used targeted lipid measurements. The cross-sectional nature of the study, low total number of participants, and inability to adjust for residual kidney function are key limitations. Additionally, residual confounding remains possible.
Relative HDL lipid composition was not associated with markers of endothelial dysfunction, or fistula maturation. Further work might consider associations between whole-plasma (or non-HDL) lipidomics and endothelial dysfunction. Alternatively, relative HDL lipid abundance may contribute to other cardiovascular risk pathways in patients with CKD.
Article Information
Authors’ Contributions
Research idea and study design: BL, BRK, IdB, NB; data acquisition: AH, CRC; data analysis and interpretation: BL, LZ, AF, NB; statistical analysis: BL, LZ; supervision or mentorship: RL, NB. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.
Support
The Hemodialysis Fistula Maturation Study was supported by NIDDK grants U01DK082179, U01DK082189, U01DK082218, U01DK082232, U01DK082236, U01DK082222, and U01DK082240. BL was supported by a Ben Lipps Fellowship grant from the American Society of Nephrology, and KL2TR002317 through the National Center for Advancing Translational Sciences.
Financial Disclosure
AH reports grant funding through NHLBI, NIDDK, NIA, and the Partnership for Clean Competition; serves as an Associate Editor for Clinical Chemistry; has received equipment and instrument support from Waters, Inc; and has provided expert testimony to Kilpatrick Townsend & Stockton LLP. IdB reports grant funding through NIDDK, NHLBI, and JDRF; has served as a consultant for Alnylam, AstraZeneca, Bayer, Boeringer-Ingelheim, Lilly, George Clinical, Gilead, Medscape, Novo Nordisk, and Otsuka; and has received equipment and supplies for research from DexCom and Novo Nordisk. BRK reports consulting fees from Launch Therapeutics. CRC reports grant funding through NIDDK, and serves on the editorial boards of BMC Nephrology, Clinical Journal of the American Society of Nephrology, and Clinical Nephrology. BL, LZ, AF, RL, and NB report no relevant disclosures.
Data Sharing
Hemodialysis Fistula Maturation Study data are available on request through the NIDDK. HDL lipid data generated for these analyses will be made available on request for the purposes of replicating findings or for future collaborations upon request from the corresponding author.
Peer Review
Received February 2, 2024. Evaluated by 1 external peer reviewer, with direct editorial input from an Associate Editor and the Editor-in-Chief. Accepted in revised form September 2, 2024.
Footnotes
Figure S1: Correlation of individual lipid species, lipoproteins, and statin use at baseline
Item S1: Supplemental methods and references.
Table S1: Characteristics at Baseline Overall and by Tertiles of Total Ceramide Levels.
Table S2: Relative Median (IQR) HDL Abundance of Individual Lipid Species by Dialysis Treatment Status.
Table S3: Relative Median (IQR) HDL Abundance of Individual Lipid Species by Unassisted Fistula Maturation Status at 9 months After Placement.
Table S4: Association of HDL Lipids With Flow-Mediated Dilation Markers (FMD and NMD), Expressed as % Change (95% CI) in Each Metric per SD-Change in Lipid Levels.
Table S5: Association of HDL Lipids With Pulse-Wave Velocity (PWV) (Carotid-Femoral, and Carotid-Radial), Expressed as M/S Change (95% CI) in Each Metric per SD-Change in Lipid Levels.
Table S6: Association of HDL Lipids With Change in Capacitance Slope and Maximum Venous Outflow Slope (95% CI) per SD-Change in Lipid Levels.
Table S7: Association of Markers of Endothelial Dysfunction With Concentration of HDL Lipids (per Standard Deviation [SD] Increment), Unadjusted.
Table S8: Association Between Relative HDL Abundance of Lipids and Unassisted Fistula Maturation.
Supplementary Materials
Item S1, Figure S1, Tables S1-S8.
References
- 1.Widmer R.J., Lerman A. Endothelial dysfunction and cardiovascular disease. Glob Cardiol Sci Pract. 2014;2014(3):291–308. doi: 10.5339/gcsp.2014.43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ding M., Rexrode K.M. A review of lipidomics of cardiovascular disease highlights the importance of isolating lipoproteins. Metabolites. 2020;10(4):163. doi: 10.3390/metabo10040163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lidgard B., Hoofnagle A.N., Zelnick L.R., et al. High-density lipoprotein lipidomics in chronic kidney disease. Clin Chem. 2023;69(3):273–282. doi: 10.1093/clinchem/hvac216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lidgard B., Bansal N., Zelnick L.R., et al. Evaluation of plasma sphingolipids as mediators of the relationship between kidney disease and cardiovascular events. eBioMedicine. 2023;95 doi: 10.1016/j.ebiom.2023.104765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Choi R.H., Tatum S.M., Symons J.D., Summers S.A., Holland W.L. Ceramides and other sphingolipids as drivers of cardiovascular disease. Nat Rev Cardiol. 2021;18(10):701–711. doi: 10.1038/s41569-021-00536-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mantovani A., Lunardi G., Bonapace S., et al. Association between increased plasma ceramides and chronic kidney disease in patients with and without ischemic heart disease. Diabetes Metab. 2021;47(1) doi: 10.1016/j.diabet.2020.03.003. [DOI] [PubMed] [Google Scholar]
- 7.Mantovani A., Dugo C. Ceramides and risk of major adverse cardiovascular events: A meta-analysis of longitudinal studies. J Clin Lipidol. 2020;14(2):176–185. doi: 10.1016/j.jacl.2020.01.005. [DOI] [PubMed] [Google Scholar]
- 8.Siddiqui M.A., Ashraff S., Carline T. Maturation of arteriovenous fistula: analysis of key factors. Kidney Res Clin Pract. 2017;36(4):318–328. doi: 10.23876/j.krcp.2017.36.4.318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Allon M., Greene T., Dember L.M., et al. Association between preoperative vascular function and postoperative arteriovenous fistula development. J Am Soc Nephrol. 2016;27(12):3788–3795. doi: 10.1681/ASN.2015020141. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Item S1, Figure S1, Tables S1-S8.