Abstract
Background
Type 2 diabetes mellitus (T2DM) increases the risk of cardiovascular disease (CVD), largely by alterations in the blood lipids and the metabolism of circulating lipoproteins (LPs). We studied whether the presence of additional risk factors, such as hypertension, or CVD itself, is associated with further alterations in the LP profiles in individuals with T2DM.
Methods
We performed LP profiling using 1H NMR spectroscopy and quantified 65 parameters in 393 healthy controls (HC) and in 390 T2DM patients with and without cardiovascular comorbidities. Univariate and multivariate analyses were used to assess alterations in LPs in diabetic patients.
Results
Triglycerides in all major LP classes, as well as particle numbers of very low-density lipoproteins (VLDL) and intermediate-density lipoproteins (IDL) were increased in T2DM compared to HC. In contrast, particle numbers of low-density lipoproteins (LDL) and high-density lipoproteins (HDL) were reduced, suggesting slower lipolytic conversion of IDL to LDL and impaired clearance of triglyceride-enriched HDL. Univariate and multivariate analyses converged in identifying distinct LP profiles associated with T2DM, while differences between patients with and without hypertension or CVD were minor, indicating that T2DM is the primary factor driving LP dysregulation. T2DM, with or without cardiovascular comorbidities, also causes differential disruption of the correlation structure among LPs.
Conclusions
T2DM is associated with major alterations in LP metabolism independent of hypertension or CVD. Thus, early lipid management in T2DM is important to mitigate CVD risk. Further research is needed to elucidate how T2DM progresses to CVD in relation to atherogenic LPs.
Graphical abstract
Supplementary Information
The online version contains supplementary material available at 10.1186/s12933-025-03039-2.
Keywords: Type 2 diabetes mellitus, Lipoprotein metabolism, Dyslipidemia, Cardiovascular disease risk, 1H NMR spectroscopy
Research insights
What is currently known about this topic?
Type 2 diabetes mellitus (T2DM) raises plasma VLDL and lowers HDL, increasing cardiovascular disease (CVD) risk. Early lipid management in T2DM is crucial to mitigate CVD risk.
What is the key research question?
Does hypertension or CVD comorbidity further alter lipoprotein metabolism in T2DM?
What is new?
We found dysregulation in both major lipoprotein fractions and smaller, atherogenic subfractions.
CVD comorbidities are linked to minor changes in cholesterol and ApoB in LDL and HDL subfractions.
CVD comorbidities in T2DM were also linked to disrupted correlations among lipoproteins.
How might this study influence clinical practice?
Lipoproteins dysregulations we identified prove new mechanistic insights into T2DM.
Changes in lipoprotein levels may be promising targets for early lipid therapy to reduce CVD risk.
Introduction
Type 2 Diabetes Mellitus (T2DM) is a chronic metabolic disorder characterized by hyperglycemia due to peripheral insulin resistance and inadequate pancreatic insulin secretion [1]. T2DM has many microvascular and macrovascular complications which significantly affect quality of life [2] and increase risk of death, particularly from cardiovascular disease (CVD) [3, 4]. Alterations in lipid and lipoprotein (LP) metabolism in T2DM are linked to insulin resistance and are believed to be involved in the etiology of CVD [5, 6].
T2DM is associated with elevated particle numbers of very low-density lipoproteins (VLDL) and reduced numbers of high-density lipoproteins (HDL) [7, 8]. The increase in VLDL is primarily due to enhanced hepatic secretion and impaired clearance driven by insulin resistance. Elevated VLDL leads to increased conversion to intermediate-density lipoproteins (IDL) through LP lipase-mediated lipolysis. Studies also indicate that low-density lipoproteins (LDL) particles in T2DM exhibit a smaller and denser phenotype [9]. These small dense LDL particles have a higher propensity to penetrate the arterial wall and are more prone to oxidative modification, thereby promoting atherogenesis[10]. Furthermore, both the triglyceride (TG) and the cholesterol (Chol)) contents of LPs differ between patients with T2DM and healthy individuals, contributing to hypertriglyceridemia, hypercholesterolemia, and decreased HDL-mediated reverse Chol transport [11–14]. All these alterations significantly contribute to increased CVD risk in patients with T2DM.
Hypertension is also associated with insulin resistance and is approximately twice as frequent in patients with T2DM compared with healthy individuals, and further increases risk of CVD [15]. It is not clear whether this is because of more extensive alterations in the blood lipid profile [16–19]. Besides traditional lipid analysis, assessing compositional alterations in LP subfractions is crucial for understanding the mechanisms linking T2DM with CVD and informing risk assessment and management strategies [20]. Here, we analyzed 65 plasma LP parameters in a large group of T2DM patients with and without hypertension and coronary heart disease (CHD), and in healthy controls (HC), by using a recently-developed model based on proton nuclear magnetic resonance (1H NMR) spectroscopy [21]. We also investigated alterations in co-variation among different LPs to identify changes in LP metabolism likely to be mechanistically linked to T2DM.
Materials and methods
Study population
This study recruited patients diagnosed with T2DM from the Republican Specialized Scientific and Practical Medical Center of Endocrinology in Tashkent (Uzbekistan) between July 2021 and August 2022. Exclusion criteria included current cancer diagnosis, pregnancy or lactation, and confirmed type 1 or gestational diabetes mellitus. A total of 390 T2DM patients (later referred to as DA group) aged 20–88 years were enrolled, of whom 88 had no hypertension or CHD (DO subgroup), 105 T2DM patients had hypertension (DH subgroup), and 197 had both hypertension and CHD (DHC subgroup). Recruitment of healthy control (HC) group was conducted between July 2021 and July 2022 at the Center for Advanced Technologies in Tashkent (Uzbekistan) through social media and local newspapers. For the HC group, 393 healthy individuals (normal blood glucose concentrations, normal blood pressure, no chronic medical conditions, and no regular medication use) were recruited. Clinical, anthropometric, and medical history data were collected at the time of hospital admission for T2DM patients or during visits to the recruitment center for HC. Each participant completed a detailed questionnaire probing for demographic information and lifestyle factors, including smoking and alcohol consumption. Fasting blood samples were collected between 6 and 10 AM, centrifuged in EDTA-containing vacutainers, and plasma was stored at −80 °C until analysis. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethical Committee of the Ministry of Public Health of the Republic of Uzbekistan (No. 6/11–1694). All participants provided written informed consent.
Chemicals and reagents
All chemicals and reagents were procured from Sigma-Aldrich (Søborg, Denmark). These included deuterium oxide (D2O, 99.9% atom D), monobasic sodium phosphate (NaH2PO4, ≥ 99%), dibasic sodium phosphate (Na2HPO4, ≥ 98%), sodium salt of 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid (TSP, 98% atom D, ≥ 98%), and sodium azide (NaN3, ≥ 99.5%). Purified water was obtained using a Millipore lab water system (Merck KGaA, Darmstadt, Germany) equipped with a 0.22 μm filter membrane.
Sample collection and preparation for 1H NMR analysis
Blood plasma samples were thawed at room temperature and aliquots of 350 μL were carefully mixed with an equal volume of phosphate buffer in 2.0 mL Eppendorf tubes. The phosphate buffer was prepared as previously described [22], using NaH2PO4 and Na2HPO4 solutions. A 600 µL portion of the plasma-buffer mixture was then transferred into 5 mm OD (103.5 mm length) SampleJet tubes (Bruker BioSpin, Germany). Sample preparation and measurements were randomized, and pooled control human blood plasma samples were analyzed at regular intervals throughout the entire measurement sequence.
1H NMR spectral data acquisition and processing
The 1H NMR spectra of plasma samples were acquired as previously described [21]. Briefly, the Bruker Avance III 600 MHz NMR spectrometer was used to acquire 1D NOESY spectra with water suppression (noesygppr1d) with 32 scans, collected into 131,072 data points, with a spectral width of 30 ppm, a 90° pulse, a recycle delay (d1) of 4 s, and a mixing time of 0.01 s. All NMR spectra were imported into the SigMa software [23], and concentrations of LPs were predicted using 0.6–1.4 ppm region of the 1H NMR spectra, representing methyl protons of Chol and fatty acids and methylene protons of fatty acids, using partial least squares (PLS) regression, as described previously [21].
Statistical data analysis
To assess potential confounding factors, continuous variables (e.g., age, body mass index (BMI)) were analyzed using Student’s t-test, while categorical variables (e.g., gender, smoking status) were evaluated with the chi-square test. Differences in plasma LPs between the T2DM patients and HC groups were evaluated using one-way ANOVA with Benjamini–Hochberg’s false discovery rate (FDR-p) correction (5%). An FDR-p value of < 0.05 was considered statistically significant. Multiple linear regression (MLR) and logistic regression analysis (LRA) were used to assess the impact of T2DM on LPs, considering disease status (T2DM versus HC) as the independent variable and adjusting for significant confounders (sex, age, BMI, blood pressure, and alcohol consumption). Confounding effects of medication use were accounted when evaluating differences among T2DM subgroups only (DO, DH, and DHC). Partial least squares-discriminant analysis (PLS-DA) [24] was used to identify key LP classifiers distinguishing T2DM patients from the HC group, following previously described optimization and validation protocol [25]. All statistical analyses were conducted using MATLAB (R2024a, The MathWorks, Inc., Natick, USA) with customized scripts.
Results
Characteristics of the study population
A total of 783 participants were included in the study (Table 1). DA patients were older and had higher BMI, blood glucose (BG), and both systolic (SB) and diastolic blood pressure (DB) compared to HC. Gender distribution and smoking status did not differ significantly, though alcohol consumption was higher in HC. The DO subgroup included more males than DH and DHC, while no differences were found in smoking, alcohol use, BG, or T2DM duration. Age, BMI, SB, DB, and medication use increased progressively across the T2DM subgroups.
Table 1.
Characteristics of the study population
| HC (N = 393) | DA (N = 390) | P | DO (N = 88) | DH (N = 105) | DHC (N = 197) | P | |
|---|---|---|---|---|---|---|---|
| Females | 212 (54%) | 186 (48%) | 0.08 | 25 (28%) | 56 (53%) | 105 (53%) | 2E-4 |
| Age (years) | 43.8 (15.9) | 57.8 (11.6) | 8E-37 | 50.5 (13.1) | 56.3 (11.2) | 61.9 (9) | 8E-15 |
| BMI (kg/m2) | 26.2 (5) | 29.8 (5.1) | 5E-19 | 28.6 (4.3) | 29.8 (4.8) | 30.3 (5.4) | 0.02 |
| Smoking (yes) | 120 (31%) | 122 (31%) | 0.82 | 29 (33%) | 29 (28%) | 64 (32%) | 0.64 |
| Alcohol (yes) | 164 (42%) | 123 (32%) | 0.003 | 30 (34%) | 28 (27%) | 65 (33%) | 0.45 |
| BG (mmol/L) | 4.4 (0.8) | 10.2 (3.7) | 2E-62 | 10.3 (3.8) | 10.3 (3.9) | 10.1 (3.4) | 0.87 |
| SB (mmHg) | 111.8 (15.9) | 126.7 (18.5) | 9E-15 | 118.9 (14.5) | 129.9 (17.9) | 128.6 (19.5) | 4E-05 |
| DB (mmHg) | 74.5 (11.9) | 80.7 (10.1) | 3E-08 | 76.8 (8.9) | 81.4 (10.7) | 82 (9.9) | 2E-4 |
| T2DM (years) | na | 11.2 (8.1) | na | 9.8 (8.4) | 11.2 (6.9) | 11.9 (8.5) | 0.16 |
| Med. (yes) | na | 377 (97%) | na | 78 (89%) | 102 (97%) | 197 (100%) | 5E-06 |
| Hyp. (yes) | na | 300 (77%) | na | 54 (61%) | 85 (81%) | 161 (82%) | 4E-4 |
| Llow. (yes) | na | 138 (35%) | na | 11 (13%) | 26 (25%) | 101 (51%) | 6E-11 |
| Anhyp. (yes) | na | 255 (65%) | na | 29 (33%) | 67 (64%) | 159 (81%) | 4E-14 |
HC = healthy control individuals (N = 393); DA = all patients with Type 2 Diabetes Mellitus (T2DM) (N = 390); DO = diabetic patients without hypertension or coronary heart disease (N = 88); DH = diabetic patients with hypertension (N = 105); DHC = diabetic patients with both hypertension and coronary heart disease (N = 197). Data are presented as mean (standard deviation (SD)) for continuous variables, or as n (%) for categorical variables. P-values were determined using one-way ANOVA or t-test for continuous variables, and the chi-square test for categorical variables. *BMI = body mass index; BG = blood glucose concentration; SB = systolic blood pressure; DB = diastolic blood pressure; Med = regular medication use; Hyp = regular use of hypoglycemic drugs; Llow = regular use of lipid-lowering drugs; Anhyp = regular use of antihypertensive drugs
Lipoprotein profile alterations in type 2 diabetes
Multivariable linear regression (MLR), adjusted for key confounders (age, gender, BMI, systolic/diastolic blood pressure, and alcohol consumption), identified 62 LPs significantly different between HC and DA. Nearly all LPs also showed significant differences between HC and at least one T2DM subgroup (Table 2). After further adjusting for medication use, MLR comparing the three T2DM subgroups (DO, DH, DHC) revealed only six LPs with significant differences. Figure 1A presents results from both LRA and MLR, highlighting two distinct LP clusters on the odds ratio vs. fold change plot, with marker size reflecting effect size. The first cluster includes TG, cholesterol ester (CE), and apolipoprotein B (ApoB) in VLDL and IDL particle numbers, positively associated with T2DM. The second cluster, negatively associated with T2DM, includes Chol, apolipoprotein A1 (ApoA1), ApoB, and phospholipids (Phoslip) in HDL and LDL subfractions (Table 2 and Supplementary Table 1). Similar clustering was observed when LRA and MLR results were jointly analyzed across HC and all T2DM subgroups, with no significant differences among the subgroups (Supplementary Fig. 1).
Table 2.
Concentrations of lipoproteins in healthy individuals and in patients with Type 2 diabetes mellitus
| Lipoproteins | HC, M(SD) | DA, M(SD) | DO, M(SD) | DH, M(SD) | DHC, M(SD) | P | |
|---|---|---|---|---|---|---|---|
| 1 | PlasmaTG | 113.2 (52.9) | 153.9 (72.7) | 140 (68) | 167.5 (82.2) | 152.8 (68.3) | a,b,c,d |
| 4 | LDLTG | 15.7 (5.8) | 17.2 (7.5) | 16.2 (8) | 18.6 (8) | 17 (6.9) | c |
| 5 | VLDLTG | 72.5 (45) | 99.6 (59.8) | 88 (53.9) | 110.3 (68.8) | 99.1 (56.4) | a,c,d |
| 6 | IDLTG | 9.1 (4.6) | 13.3 (6.9) | 12.5 (7.3) | 14.5 (7.9) | 13.1 (6) | a,b,c,d |
| 7 | LDL1TG | 4 (2.1) | 5.2 (2.9) | 4.8 (3.1) | 5.6 (3.2) | 5.1 (2.6) | c,d |
| 8 | HDL2bTG | 1.8 (1.1) | 2.5 (1.7) | 2.2 (1.7) | 2.8 (1.8) | 2.5 (1.6) | a,c,d,e |
| 9 | HDL2aTG | 2.6 (1.7) | 2.3 (1.4) | 2.2 (1.4) | 2.3 (1.4) | 2.4 (1.5) | |
| 10 | PlasmaChol | 183.9 (46.2) | 152.4 (40.7) | 156 (38) | 156.9 (43.8) | 148.3 (39.9) | a,b,c,d |
| 13 | LDLChol | 103.1 (31.3) | 78.2 (29.7) | 83.3 (29.3) | 79.7 (33.1) | 75.2 (27.7) | a,b,c,d |
| 14 | HDLChol | 38.3 (12) | 26.2 (9.2) | 26.5 (9.7) | 25.4 (8.8) | 26.5 (9.1) | a,b,c,d |
| 15 | VLDLChol | 11.9 (8.6) | 16.9 (11.2) | 14.1 (10.3) | 19.1 (12.3) | 17 (10.7) | c,d |
| 16 | IDLChol | 6.7 (3.6) | 8.1 (4) | 7.1 (3.9) | 8.9 (4) | 8.2 (4) | c,d |
| 17 | LDL1Chol | 27.2 (10.8) | 20.9 (10.9) | 22.1 (11) | 21.9 (11.9) | 19.8 (10.3) | a,b,c,d |
| 18 | LDL2Chol | 14.8 (6.3) | 9.7 (5.7) | 11 (5.4) | 10.1 (6.5) | 8.9 (5.4) | a,b,c,d,f |
| 19 | LDL3Chol | 14.8 (5.6) | 10.4 (4.8) | 10.8 (4.8) | 10.8 (5.2) | 10 (4.6) | a,b,c,d |
| 20 | LDL4Chol | 15.7 (4.9) | 12.1 (4.3) | 12.8 (4.4) | 12.3 (4.6) | 11.6 (4) | a,b,c,d |
| 21 | LDL5Chol | 11.8 (4.6) | 9.5 (3.9) | 9.2 (3.5) | 10.1 (4.2) | 9.4 (3.9) | a,b,c,d |
| 22 | HDL2bChol | 14.2 (7.2) | 7.5 (5.1) | 8.1 (5.2) | 7.5 (5.5) | 7.3 (4.9) | a,b,c,d |
| 23 | HDL2aChol | 17.6 (3.8) | 13.3 (3.3) | 13.6 (3.3) | 12.6 (3.5) | 13.5 (3.2) | a,b,c,d,g |
| 24 | HDL3Chol | 19.1 (3.4) | 17.3 (4.1) | 17.6 (3.8) | 17 (4.3) | 17.4 (4.1) | a,c,d |
| 25 | PlasmaFreeChol | 52.4 (11.5) | 46.7 (11.1) | 46.3 (10.4) | 48.6 (11.8) | 45.9 (10.9) | a,b,c,d |
| 26 | VLDLFreeChol | 7.1 (4) | 10.7 (5.2) | 9.5 (5) | 11.7 (5.8) | 10.7 (5) | a,b,c,d |
| 27 | HDL2bFreeChol | 5 (2.1) | 2.9 (1.6) | 3.1 (1.9) | 2.9 (1.7) | 2.8 (1.5) | a,b,c,d |
| 28 | PlasmaPhoslip | 203.6 (45.2) | 183.4 (45.2) | 181.9 (38.9) | 187.7 (50.1) | 181.7 (45.1) | a,b,c,d |
| 31 | LDLPhoslip | 69.1 (21.3) | 52.8 (20.7) | 56 (19.9) | 53.7 (23.3) | 50.9 (19.4) | a,b,c,d |
| 32 | HDLPhoslip | 67.9 (18.1) | 53.8 (17.2) | 52.4 (17.3) | 53 (17.7) | 54.8 (16.9) | a,b,c,d |
| 33 | VLDLPhoslip | 22.9 (13.4) | 32.6 (17.2) | 29.1 (16.2) | 35.9 (19.2) | 32.4 (16.3) | a,b,c,d |
| 34 | LDL1Phoslip | 13.3 (8.8) | 9.1 (6.5) | 9.4 (6.3) | 9.4 (7.9) | 8.9 (5.8) | a,b,c,d |
| 35 | LDL2Phoslip | 7.4 (3.9) | 5.2 (3.3) | 5.6 (3.2) | 5.6 (3.9) | 4.8 (3) | a,c,d |
| 36 | LDL3Phoslip | 10 (3.3) | 7.5 (3) | 7.7 (2.9) | 7.8 (3.3) | 7.3 (2.9) | a,b,c,d |
| 37 | LDL5Phoslip | 7.5 (2.7) | 7.5 (2.7) | 7 (2.4) | 8 (2.7) | 7.5 (2.7) | a |
| 38 | HDL2bPhoslip | 15.7 (11.2) | 10.8 (9.1) | 11.5 (9.9) | 10.9 (10.1) | 10.4 (8.1) | a,b,c,d |
| 39 | HDL2aPhoslip | 36.2 (9) | 35.6 (10.7) | 36 (9.4) | 35 (11) | 35.7 (11.1) | |
| 40 | PlasmaApoA1 | 121 (21.7) | 100.3 (20.6) | 101.5 (19.1) | 98.2 (22.3) | 100.8 (20.3) | a,b,c,d |
| 43 | HDLApoA1 | 101.2 (21.4) | 81.8 (17.3) | 82.8 (16.8) | 80.8 (18) | 81.9 (17.3) | a,b,c,d |
| 44 | HDL2bApoA1 | 16 (9.9) | 9.3 (7.9) | 9.5 (8.6) | 9.8 (8.9) | 8.9 (7) | a,b,c,d |
| 45 | HDL2aApoA1 | 36.5 (8.2) | 31.7 (7.9) | 32.2 (7.1) | 30.5 (8) | 32 (8.1) | a,b,c,d |
| 46 | HDL3ApoA1 | 54.4 (11.1) | 46.2 (9.6) | 46.2 (9.3) | 45.5 (9.3) | 46.7 (10) | a,b,c,d |
| 47 | PlasmaApoB | 93.7 (21.8) | 86.3 (21) | 87.4 (21.2) | 89.4 (22.1) | 84.2 (20.2) | a,d |
| 50 | LDLApoB | 75.3 (18.4) | 63 (17.3) | 65.6 (17.4) | 64.5 (18.9) | 60.9 (16.1) | a,b,c,d |
| 51 | VLDLApoB | 5.8 (3) | 8.1 (3.5) | 7.2 (3.4) | 8.7 (3.7) | 8.2 (3.3) | a,b,c,d |
| 52 | IDLApoB | 6 (1.7) | 7.8 (2.1) | 7.6 (2.3) | 8.2 (2.2) | 7.7 (2) | a,b,c,d |
| 53 | LDL1ApoB | 17.2 (5.1) | 15.5 (5.7) | 16.5 (5.6) | 16 (6.2) | 14.9 (5.4) | a,d |
| 54 | LDL2ApoB | 9.4 (3.3) | 6.9 (3.5) | 7.6 (3.2) | 7 (3.9) | 6.4 (3.3) | a,b,c,d,f |
| 55 | LDL3ApoB | 9.6 (3.4) | 7.7 (3) | 7.7 (3) | 8 (3.1) | 7.5 (2.9) | a,b,c,d |
| 56 | LDL4ApoB | 9.5 (3.2) | 7.5 (2.8) | 7.5 (2.7) | 7.8 (3) | 7.3 (2.6) | a,b,c,d |
| 57 | LDL5ApoB | 7.7 (3) | 6.4 (2.7) | 6 (2.4) | 6.8 (3.2) | 6.4 (2.6) | a,b,c,d |
| 58 | LDL6ApoB | 9.9 (8.5) | 10.7 (9) | 9.6 (7.1) | 11.9 (10.2) | 10.5 (9.1) | |
| 59 | PlasmaCE | 219.2 (56.7) | 176.7 (51.6) | 183.4 (48.6) | 181.1 (55.3) | 171.3 (50.5) | a,b,c,d |
| 60 | VLDLCE | 12.1 (6.9) | 17 (8.8) | 15.3 (8.5) | 18.7 (9.5) | 17 (8.3) | a,b,c,d |
| 61 | IDLCE | 7.6 (2.9) | 9.5 (3.6) | 9.1 (3.8) | 10 (3.4) | 9.4 (3.6) | a,b,c,d |
| 62 | LDLCE | 107.5 (39.1) | 83 (36.8) | 87.5 (37.1) | 86.2 (39.1) | 79.4 (35.2) | a,b,c,d |
| 63 | LDL1CE | 25.4 (13.1) | 19.6 (12.5) | 21.1 (12) | 21.3 (14.4) | 18.1 (11.6) | a,b,c,d |
| 64 | LDL3CE | 16.3 (5.8) | 11.7 (5.9) | 12.3 (6.1) | 12.1 (6.4) | 11.2 (5.4) | a,b,c,d |
| 65 | HDL2bCE | 17.3 (8.4) | 11.9 (7.5) | 12.5 (7.1) | 12.5 (7.9) | 11.3 (7.6) | a,b,c,d |
Concentrations of LPs (mg/dL) in healthy control individuals (HC, N = 393), all patients with Type 2 Diabetes Mellitus (DA, N = 393), diabetic patients without hypertension or coronary heart disease (DO, N = 88), diabetic patients with hypertension (DH, N = 105), and diabetic patients with both hypertension and coronary heart disease (DHC, N = 197). Data are presented as mean (M) (standard deviation (SD)). P-values were determined by multiple linear regression (MLR) analysis after adjusting for confounders. The significance threshold was set at P < 0.05 after false discovery rate (5%) correction. * a: significant difference between HC and DA; b: HC vs. DO; c: HC vs. DH; d: HC vs. DHC; e: DO vs. DH; f: DO vs. DHC; g: DH vs. DHC. TG = triglycerides; Chol = cholesterol; FreeChol = free cholesterol; CE = cholesterol ester; Phoslip = phospholipids; ApoA1 = apo apolipoprotein A1; ApoB = apolipoprotein B; VLDL = very low-density LP; IDL = intermediate-density LP; LDL = low-density LP; HDL = high-density LP. **Concentrations of LP molecules in main and subfraction are omitted from the table, for a complete list see Supplementary Table 1
Fig. 1.
Blood plasma LPs differentiating healthy individuals (HC, N = 393) from patients with Type 2 Diabetes Mellitus (DA, N = 390). A: Scatter plot of adjusted odds ratios versus mean fold changes (FC) between DA and HC groups. Each dot represents a LP variable, with dot size proportional to its effect size (%) as determined by multiple linear regression (MLR) analysis after adjusting for confounders. Odds ratios were obtained from logistic regression analysis (LRA). LP names are shown only for variables that met the significance threshold (P < 0.05 after false discovery rate (5%) correction) in both LRA and MLR analyses. The color code insert indicates whether LP molecules are free in blood plasma or belong to one of the four major (sub)fractions: VLDL, IDL, LDL, or HDL. B: Results from partial least squares-discriminant analysis (PLS-DA) used to identify key LP classifiers distinguishing DA from HC individuals. Normalized variable importance in projection (VIP) scores and the selectivity ratio plot highlight 14 LP variables consistently selected during iterative PLS-DA model optimization. 1—Specificity versus Sensitivity plot shows the area under the curve (AUC) of the receiver operating characteristic curve from the optimized PLS-DA model, developed using a training set (TR—70% of individuals from both groups) and tested on a test set (TS—30%). LV = number of latent variables used to build the model; Error% = misclassification rate for TR and TS. The lower-right panel shows results of a permutation test (2000), indicating model validity. AUC values for the true datasets (TR: 0.91; TS: 0.92) were significantly higher than those for the permuted datasets (TR: 0.57; TS: 0.61)
Triglyceride enrichment was one of the most consistent alterations across diabetic groups. Compared to HC, DA showed markedly elevated PlasmaTG (154 vs. 113 mg/dL), with this increase persisting in all diabetic subgroups. VLDLTG and IDLTG were also elevated, peaking in DH, though differences between DO, DH, and DHC were not significant, suggesting that VLDL and IDL triglyceride enrichment is a core feature of T2DM rather than its comorbidities. HDL2bTG was modestly elevated in DA and increased progressively across subgroups, reaching significance between DO and DHC, while HDL2aTG remained unchanged. LDL and LDL1 triglyceride levels were stable in DA overall, yet significantly elevated in DH and DHC compared to HC. These elevations likely result from increased hepatic VLDL production driven by insulin resistance and free fatty acid flux, coupled with impaired lipolysis [26]. This leads to VLDL and IDL accumulation and sustained triglyceride enrichment in HDL2b and LDL1, indicating disrupted LP remodeling and increased atherogenic risk.
Cholesterol concentrations were reduced across all LPs in DA and all diabetic subgroups compared to HC. VLDLChol and IDLChol showed no increase from HC to DA or DO. Among diabetic subgroups, only minor declines were observed in LDL2Chol (DO to DHC) and HDL2aChol (DH to DHC), possibly reflecting greater lipid-lowering therapy use in DHC. The most striking reduction was a 50% decrease in HDL2bChol in DA, underscoring the loss of cardioprotective HDL. These findings suggest that T2DM is linked to both overall cholesterol depletion and HDL particle remodeling. Notably, reductions were already evident in DO and not further worsened by hypertension or CHD, indicating early onset. The uniform decline across LPs may result from impaired hepatic synthesis or reduced intestinal absorption [27]. Specifically, reduced HDLChol likely reflects impaired reverse cholesterol transport [28].
Free cholesterol levels in plasma and HDL2b decreased in DA, while VLDLFreeChol increased compared to HC, a pattern consistent across all diabetic subgroups (DO, DH, DHC), with no significant differences among them. This redistribution from HDL to TG-rich VLDL suggests impaired reverse cholesterol transport and enhanced hepatic VLDL secretion, likely driven by insulin resistance and reduced LP clearance [29]. Elevated VLDLFreeChol reflects cholesterol retention in remnant particles, while reduced HDL2bFreeChol points to diminished HDL-mediated cholesterol efflux in T2DM. Cholesteryl ester (CE) levels mirrored trends in total and free cholesterol, showing broad reductions across all LPs in DA and diabetic subgroups compared to HC. However, VLDLCE and IDLCE were significantly elevated in DA, unlike VLDLChol and IDLChol, with no differences among diabetic subgroups. This suggests that CE depletion is independent of comorbidities. The rise in VLDLCE and IDLCE may reflect increased cholesterol ester transfer protein (CETP)-mediated transfer of CE from HDL to TG-rich LPs, contributing to a more atherogenic LP profile in diabetes [30].
Phospholipid levels in LPs followed the patterns of Chol and CE, showing reductions in DA and all diabetic subgroups compared to HC. The steepest declines occurred in LDL1–LDL3 and HDL2b (up to 30%). LDL5Phoslip and HDL2aPhoslip remained unchanged. In contrast, VLDLPhoslip consistently increased from HC to DA and progressively across DO, DH, and DHC, reflecting enhanced hepatic VLDL output and impaired lipolysis, which shift phospholipids toward triglyceride-rich remnant particles in insulin-resistant T2DM [31].
Like most other LP molecules, ApoA1 levels were decreased both in plasma and in HDL main and in HDL subfractions consistently across all T2DM subgroups compared to HC. The largest reduction was observed in HDL2bApoA1 levels from being around 9 mg/dL while HC group contained mean concentration of 16 mg/dL. Given the central role of ApoA1 in maintaining HDL function, ApoA1 reduction, especially in large HDL2b particles, suggests impaired reverse cholesterol transport via ABCA1 and diminished HDL functionality in T2DM. This loss affects both the quantity and quality of HDL, weakening its atheroprotective role. The consistent decline across diabetic subgroups highlights ApoA1 as a key marker of HDL dysfunction, contributing to lipid accumulation and increased vascular risk. The same trend of decrease was observed for ApoB as in ApoA1, where its concentrations were reduced in plasma, and in LDL and in all LDL subfractions (LDL1-LDL5), except LDL6ApoB, in all T2DM groups compared to HC. In contrast, VLDL and IDL levels of ApoB were consistently increased in all T2DM groups consistently from being around 6 mg/dL in HC for both particles to nearly 8 mg/dL in diabetic patients. The pattern indicates a shift of ApoB from smaller, cholesterol-rich LDL to larger, triglyceride-rich VLDL and IDL particles, reflecting impaired remnant catabolism. Increased VLDLTG, VLDLCE, and VLDLApoB suggest hepatic overproduction or reduced clearance. This redistribution lowers LDL-associated ApoB, potentially underestimating atherogenic risk in T2DM. The remodeling toward triglyceride-rich remnants highlights a more atherogenic phenotype and emphasizes the need for subclass-specific ApoB profiling in diabetic dyslipidemia.
Discriminating type 2 diabetes from healthy individuals
To identify the cluster LPs with the strongest classification power for distinguishing T2DM patients from HC, a PLS-DA-based iterative variable selection was performed. Using a training set comprising 70% of T2DM and 70% of HC participants, we conducted variable selection by retaining only LPs with significant contributions based on their VIP scores and selectivity ratios. After 1000 iterations, only LPs consistently selected in all cases were retained. The final classification model for DA versus HC included 14 LPs and demonstrated high performance, with an AUC of 0.92 on the independent test set (30% from each group; Fig. 1B). The model’s validity was confirmed by a permutation test (2000). The selected LPs aligned with MLR and LRA results, with IDLApoB being the strongest discriminator, followed by HDL-associated Chol, FreeChol, and ApoA1, as well as total plasma or main-fraction levels of ApoA1, CE, and phospholipids. Similar classification accuracy was achieved for HC versus DO, DH, and DHC comparisons, with the same 14 LPs consistently among the top classifiers (Supplementary Fig. 1). However, no valid PLS-DA models could be developed to distinguish the diabetic subgroups from each other (Supplementary Fig. 1).
Altered correlations between lipoproteins in patients with type 2 diabetes
Apart from studying differences in LPs concentrations between T2DM and HC groups using MLR or identifying T2DM-associated LPs through LRA, we further investigated alterations in the correlation structure among LPs. This complementary approach offers orthogonal insights into the reorganization of lipid metabolism in diabetic dyslipidemia. Our analysis revealed that approximately 10% of positive correlations between LP pairs observed in HC were attenuated in the DA group, while the proportion of significant negative correlations nearly doubled (2.7–5.2%) (Supplementary Fig. 2). Comparable patterns were observed across all diabetic subgroups (DO, DH, and DHC), with the most pronounced disruption in the DO group, where more than 30% of previously significant correlations present in HC were lost. Concurrently, the number of significant negative correlations nearly tripled.
We further examined LP pairs exhibiting the most pronounced disruptions in correlation structure between the HC and T2DM groups or among diabetic subgroups. Strong positive correlations observed in HC between HDL2bLCE and HDLChol (R = 0.75) or HDL2aChol (R = 0.71) were attenuated in all diabetic subgroups, particularly in DH (R = 0.04–0.22) (Fig. 2). Similar patterns emerged among other HDL subfractions; for instance, the robust correlation between HDL2bPhoslip and HDL2aChol in HC (R = 0.80) declined across all diabetic groups, with the sharpest drop in DH (R = 0.20). Likewise, the correlation between HDL2bPhoslip and HDL2aApoA1 in HC (R = 0.71) was markedly reduced in DH and DHC (R = 0.29 and 0.27, respectively). Notably, HDL2aChol and HDLPhoslip both showed moderate correlations with plasma CE in HC (R = 0.46–0.50), which were lost in DO (R = –0.03), but remained significant though slightly weakened in DH and DHC (R = 0.36–0.42) (Supplementary Fig. 3). Intriguingly, weak correlations in HC between HDL2aPhoslip and HDL3Chol (R = 0.24), and between HDL2aTG and IDLTG (R = 0.27), progressively increased across diabetic subgroups, doubling in DHC. A similar trend was observed between HDL3Chol and four VLDL variables (VLDLApoB, VLDLCE, VLDLPhoslip, VLDLFreeChol), with correlations increasing from HC (R = –0.04 to 0.17) to DA (R = 0.30–0.43), reaching highest values in DH (R = 0.45–0.52). Likewise, the lack of correlation between HDL2bCE and IDLCE in HC (R = 0.07) evolved into a strong correlation in DHC (R = 0.51). LDL subfractions also showed disrupted patterns in T2DM. Strong correlations in HC between LDL1Phoslip and HDL2aTG, LDL3CE and LDL5Phoslip, and LDL2ApoB and MainFracPhoslip (R = 0.47–0.63) were consistently reduced across diabetic subgroups (R = –0.13 to 0.16). Surprisingly, notable negative correlations emerged: LDL3CE, weakly or non-correlated with VLDLChol (R = 0.18) and LDLTG (R = 0.29) in HC, exhibited strong negative correlations in DO (R = –0.40), indicating a significant shift in covariation. A similar but more pronounced trend was observed between IDLCE and HDL2aChol.
Fig. 2.
Disrupted correlation structure of blood plasma LPs in patients with Type 2 Diabetes Mellitus. Scatter plots show the correlation between the concentrations (mg/dL) of specific LP pairs in healthy controls (HC, N = 393), all patients with Type 2 Diabetes Mellitus (DA, N = 393), diabetic patients without hypertension or coronary heart disease (DO, N = 88), diabetic patients with hypertension (DH, N = 105), and diabetic patients with both hypertension and coronary heart disease (DHC, N = 197). Each panel displays the Pearson correlation coefficient (R), with asterisks (*) indicating statistically significant correlations (P < 0.05) within the respective group
Discussion
Nearly all LPs (58 out of 65) were significantly different between diabetic patients (DA group) and healthy controls (HC group), consistent with the characteristic dyslipidemia of T2DM, marked by elevated VLDL and IDL particles and triglycerides, and reduced HDL particles and HDLChol levels (Fig. 3). HDL2b particles were depleted in cholesterol and enriched in triglycerides, indicating altered composition and impaired function of HDL in T2DM. This aligns with evidence that triglyceride enrichment accelerates HDL catabolism, thereby lowering HDLChol levels [32]. Increased VLDL-related ApoB, triglycerides, and cholesterol reflect heightened systemic fatty acid availability, enhanced hepatic de novo lipogenesis, and impaired LP catabolism [33]. The latter is evident in elevated IDLApoB and reduced particle numbers of most LDL subclasses, suggesting that the conversion of IDL to LDL is a rate-limiting step in the VLDL → IDL → LDL cascade. This likely promotes formation of small, dense, highly atherogenic LDL particles [34].
Fig. 3.
Overview of major lipoprotein alterations in patients with T2DM compared with healthy controls
One of the novelties of our study is the minimal differences observed in LP profiles among the three diabetic subgroups (DO, DH, and DHC). After FDR correction of MLR multiple comparison results, only five LP variables remained significantly different between at least two subgroups. Among triglycerides, key LPs differentiating HC from DA, only the main fraction and HDL2b subfraction showed significantly higher triglyceride levels in DH compared to DO, while the remaining TG variables did not differ across subgroups. Additionally, LDL2Chol and LDL2ApoB, large and less dense atherogenic particles, were moderately lower in DH than in DO, likely due to greater medication use. The smallest observed difference was reduced HDL2aChol in DH compared to DHC. These minimal subgroup differences likely reflect a shared metabolic dysregulation in T2DM, with subtle variations such as lower LDL2Chol and LDL2ApoB in DH, and reduced HDL2aChol in DH versus DHC, possibly attributable to increased use of lipid-lowering or antihypertensive medications and/or subgroup-specific modulation of LP metabolism [31]. Previous studies examining traditional blood lipid profiles found more pronounced abnormalities in T2DM patients with hypertension than those without, though without adjusting for confounders [16, 17]. Other studies reported better lipid profiles in T2DM patients with CVD compared to those without, potentially reflecting more intensive lipid-lowering therapy [18], while some reported mixed outcomes (e.g., higher TG and HDLChol, but lower LDLChol in T2DM with CVD) [19]. However, none evaluated the comprehensive LP profile as done here. Our findings suggest that T2DM itself is the primary driver of LP dysregulation, and that hypertension and CHD do not further exacerbate these alterations but instead contribute to CVD risk through mechanisms independent of LP metabolism.
Another intriguing finding from our study is the altered correlation patterns between LPs in T2DM (Fig. 3). The disrupted correlation structure driven by insulin resistance–mediated dysregulation of hepatic lipid metabolism, which enhances hepatic VLDL overproduction and impairs LP remodeling by LP lipase (LPL) and hepatic lipase (HL), leading to triglyceride enrichment and accelerated clearance of HDL particles [35]. Normally, little apoC-ΙΙΙ and apoE are associated with HDL particles, but a rise in the content of these apolipoproteins in HDL in diabetes can promote triglyceride enrichment of HDL and its rapid removal from plasma, thereby lowering its ability for reverse cholesterol transport, and adversely affect its antiatherogenic properties [36, 37]. This remodeling process compromises the structural integrity and function of HDL, thereby weakening its physiological correlation with other HDL subfractions and cholesterol content. Subsequently, impaired lipolytic conversion of VLDL to LDL results in the accumulation of IDL and small dense LDL particles, disrupting the typical LDL–IDL–HDL metabolic continuum and giving rise to inverse correlations not observed in normoglycemic individuals [27]. For example, a rise in the apoC-III content in LDL in diabetes modifies LDL structure, enhancing the propensity of apoB100 to bind proteoglycans, potentially leading to LDL accumulation in the subendothelial space of the vascular wall, facilitating the process of atherogenesis [38]. The emergence of negative correlations, especially between HDL and IDL components, and within LDL fractions, may reflect competitive or compensatory shifts in lipid flux through divergent metabolic pathways, a hallmark of advanced diabetic dyslipidemia.
Our results are internally consistent and indicate that T2DM is the primary driver of dysregulated LP profiles and the associated increase in cardiovascular risk. These findings carry important clinical implications, suggesting that dyslipidemia management in T2DM should target the lipid and LP abnormalities inherent to diabetes itself, independent of other risk factors such as hypertension or existing CVD. This highlights the importance of initiating appropriate lipid-lowering therapy early in the course of T2DM, regardless of comorbid hypertension or CHD.
The consistent reduction in HDL-related variables and increase in VLDL-related parameters indicate that standard lipid panels may inadequately reflect the full extent of dyslipidemia in T2DM. Expanded LP profiling, including subfractions and molecular components such as apolipoproteins and phospholipids, may offer more precise biomarkers for disease monitoring. The altered correlation patterns among LPs suggest a complex, diabetes-driven reorganization of lipid metabolism. This supports the need for integrated metabolic treatment strategies that not only lower LDL cholesterol but also enhance HDL functionality and reduce VLDL and triglyceride levels. Further mechanistic studies on the role of insulin resistance, inflammation, and oxidative stress in modulating LP metabolism could deepen our understanding of diabetic dyslipidemia. Longitudinal studies are also warranted to investigate how these alterations on LP profiles evolve over time and contribute to cardiovascular outcomes in T2DM.
This study has limitations, including a relatively small sample size within T2DM subgroups, which may limit statistical power and generalizability. The cross-sectional design also precludes causal inference. Future studies with larger, multi-center cohorts and longitudinal frameworks are necessary to validate and expand upon these findings.
Conclusion
This study demonstrates significant alterations in LP profiles in T2DM, characterized by elevated triglycerides and VLDL, and reduced cholesterol and HDL, independent of hypertension or CVD. Notably, the disruption extended to smaller, atherogenic subfractions, underscoring the depth of metabolic imbalance in T2DM. The presence of cardiovascular comorbidities introduced only modest additional changes, mainly affecting cholesterol and ApoB content in LDL and HDL subfractions. Furthermore, a correlation analysis revealed a significant breakdown in the physiological relationships among lipoproteins in T2DM, which was further altered in those with cardiovascular complications. Our findings suggest that T2DM alone is the principal driver of lipoprotein abnormalities, while cardiovascular comorbidities contribute minimally to additional compositional changes. These results underscore the need for early lipid intervention to mitigate cardiovascular risk. Lipoprotein profiling in diabetic patients may improve risk stratification and guide early lipid-lowering interventions. Future longitudinal studies are warranted to confirm these findings and explore causal mechanisms.
Supplementary Information
Acknowledgements
Data was generated though accessing research infrastructure at UCPH, including FOODHAY (Food and Health Open Innovation Laboratory, Danish Roadmap for Research Infrastructure). Komila Makhkamova is gratefully acknowledged for her invaluable assistance in creating Fig. 3. Figure 3 was created with BioRender.com.
Abbreviations
- ANOVA
Analysis of variance
- ApoA1
Apolipoprotein A1
- ApoB
Apolipoprotein B
- BMI
Body mass index
- CE
Cholesteryl esters
- CHD
Coronary heart disease
- Chol
Cholesterol
- CI
Confidence interval
- CVD
Cardiovascular disease
- DA
All patients with type 2 diabetes mellitus (T2DM), with or without comorbidities
- DH
Diabetic patients with hypertension
- DHC
Diabetic patients with hypertension and coronary heart disease
- DO
Diabetic patients without hypertension or coronary heart disease
- FDR
False discovery rate
- FreeChol
Free (unesterified) cholesterol
- HC
Healthy controls
- HDL
High-density lipoprotein
- IDL
Intermediate-density lipoprotein
- LDL
Low-density lipoprotein
- LRA
Logistic regression analysis
- LP
Lipoprotein
- NMR
Nuclear magnetic resonance
- OR
Odds ratio
- Phoslip
Phospholipids
- PLS-DA
Partial least squares discriminant analysis
- SD
Standard deviation
- T2DM
Type 2 diabetes mellitus
- TG
Triglycerides
- VLDL
Very low-density lipoprotein
Author contributions
B.K, and D.D. acquired funding and were involved in the conception and design of the study. Z.K., contributed to sample collection and recruitment of patients. E.T., D.K., S.N., and V.T., S.A., B.S.S., and B.K. were involved in sample preparation, data acquisition, data cleaning, and data analysis. B.K. wrote the manuscript. F.M., and S.B.E. contributed to the interpretation of the results. All authors contributed to the discussion, edited the manuscript, and approved the final manuscript. B.K. and D.D. are the guarantors of this work and, as such, have full access to all the data included in the study and take responsibility for the integrity of the data and the accuracy of the results.
Funding
Open access funding provided by Copenhagen University. Ministry of Higher Education,Science and Innovations of the Republic of Uzbekistan,REP 25112021-127,University of Copenhagen (Data + project,Strategy 2013 funds)
Data availability
The data generated during and/or analyzed in the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The study was conducted in accordance with the principles outlined in the Declaration of Helsinki and approved by the Ethical Committee of the Ministry of Public Health of the Republic of Uzbekistan (Approval No. 6/11–1694). Written informed consent was obtained from all participants prior to their inclusion in the study.
Consent for publication
This manuscript is not under consideration elsewhere and will not be submitted to another journal until a final decision regarding its acceptability has been made.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Batista TM, Haider N, Kahn CR. Defining the underlying defect in insulin action in type 2 diabetes. Diabetologia. 2021;64(5):994–1006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ma C-X, Ma X-N, Guan C-H, Li Y-D, Mauricio D, Fu S-B. Cardiovascular disease in type 2 diabetes mellitus: progress toward personalized management. Cardiovasc Diabetol. 2022;21(1):74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pearson-Stuttard J, Bennett J, Cheng YJ, Vamos EP, Cross AJ, Ezzati M, et al. Trends in predominant causes of death in individuals with and without diabetes in England from 2001 to 2018: an epidemiological analysis of linked primary care records. Lancet Diabetes Endocrinol. 2021;9(3):165–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Harding JL, Shaw JE, Peeters A, Davidson S, Magliano DJ. Age-specific trends from 2000–2011 in all-cause and cause-specific mortality in type 1 and type 2 diabetes: a cohort study of more than one million people. Diabetes Care. 2016;39(6):1018–26. [DOI] [PubMed] [Google Scholar]
- 5.Brunzell JD, Davidson M, Furberg CD, Goldberg RB, Howard BV, Stein JH, et al. Lipoprotein management in patients with cardiometabolic risk: consensus statement from the American diabetes association and the American college of cardiology foundation. Diabetes Care. 2008;31(4):811–22. [DOI] [PubMed] [Google Scholar]
- 6.Luciani L, Pedrelli M, Parini P. Modification of lipoprotein metabolism and function driving atherogenesis in diabetes. Atherosclerosis. 2024;394:117545. [DOI] [PubMed] [Google Scholar]
- 7.Chehade JM, Gladysz M, Mooradian AD. Dyslipidemia in type 2 diabetes: prevalence, pathophysiology, and management. Drugs. 2013;73(4):327–39. [DOI] [PubMed] [Google Scholar]
- 8.Zhang P, Gao J, Pu C, Zhang Y. Apolipoprotein status in type 2 diabetes mellitus and its complications (Review). Mol Med Rep. 2017;16(6):9279–86. [DOI] [PubMed] [Google Scholar]
- 9.Gupta A, Gupta R. Current understanding of diabetic dyslipidemia: a review. J Indian Inst Sci. 2023;103(1):287–307. [Google Scholar]
- 10.Rizzo M, Berneis K. Low-density lipoprotein size and cardiovascular risk assessment. QJM: An Int J Med. 2006;99(1):1–14. [DOI] [PubMed] [Google Scholar]
- 11.Sokooti S, Flores-Guerrero JL, Heerspink HJL, Connelly MA, Bakker SJL, Dullaart RPF. Triglyceride-rich lipoprotein and LDL particle subfractions and their association with incident type 2 diabetes: the PREVEND study. Cardiovasc Diabetol. 2021;20(1):156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jomard A, Osto E. High density lipoproteins: metabolism, function, and therapeutic potential. Front Cardiovasc Med. 2020. 10.3389/fcvm.2020.00039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hernáez Á, Soria-Florido MT, Schröder H, Ros E, Pintó X, Estruch R, et al. Role of HDL function and LDL atherogenicity on cardiovascular risk: a comprehensive examination. PLoS ONE. 2019;14(6):e0218533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cho K-H. The current status of research on high-density lipoproteins (HDL): a paradigm shift from HDL quantity to HDL quality and HDL functionality. Int J Mol Sci. 2022;23(7):3967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sowers JR, Epstein M, Frohlich ED. Diabetes, hypertension, and cardiovascular disease. Hypertension. 2001;37(4):1053–9. [DOI] [PubMed] [Google Scholar]
- 16.Oyelola OO, Ajayi AA, Babalola RO, Stein EA. Plasma lipids, lipoproteins, and apolipoproteins in Nigerian diabetes mellitus, essential hypertension, and hypertensive-diabetic patients. J Natl Med Assoc. 1995;87(2):113–8. [PMC free article] [PubMed] [Google Scholar]
- 17.Jha J, Malhotra V, Jha O, Gupta S. Study of lipid profile in diabetes mellitus with and without hypertension. J Basic Appl Res Biomed. 2022;5(1):31–6. [Google Scholar]
- 18.Monteiro AM, Palma I. Lipid profile and persistent lipid abnormalities in diabetic patients–a retrospective study. Revista Portuguesa de Endocrinologia, Diabetes e Metabolismo. 2016;11(2):197–201. [Google Scholar]
- 19.Leiter LA, Lundman P, da Silva PM, Drexel H, Jünger C, Gitt AK. Investigators obotD: persistent lipid abnormalities in statin-treated patients with diabetes mellitus in Europe and Canada: results of the dyslipidaemia international study. Diabet Med. 2011;28(11):1343–51. [DOI] [PubMed] [Google Scholar]
- 20.Syvänne M, Taskinen M-R. Lipids and lipoproteins as coronary risk factors in non-insulin-dependent diabetes mellitus. Lancet. 1997;350:S20–3. [DOI] [PubMed] [Google Scholar]
- 21.Khakimov B, Hoefsloot HCJ, Mobaraki N, Aru V, Kristensen M, Lind MV, et al. Human blood lipoprotein predictions from H-1 NMR spectra: protocol, model performances, and cage of covariance. Anal Chem. 2022;94(2):628–36. [DOI] [PubMed] [Google Scholar]
- 22.Centelles SM, Hoefsloot HCJ, Khakimov B, Ebrahimi P, Lind MV, Kristensen M, et al. Toward reliable lipoprotein particle predictions from NMR spectra of human blood: an interlaboratory ring test. Anal Chem. 2017;89(15):8004–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Khakimov B, Mobaraki N, Trimigno A, Aru V, Engelsen SB. Signature mapping (SigMa): an efficient approach for processing complex human urine H-1 NMR metabolomics data. Anal Chim Acta. 2020;1108:142–51. [DOI] [PubMed] [Google Scholar]
- 24.Ståhle L, Wold S. Partial least squares analysis with cross‐validation for the two‐class problem: a monte carlo study. J Chemom. 1987. 10.1002/cem.1180010306. [Google Scholar]
- 25.Khakimov B, Poulsen SK, Savorani F, Acar E, Gurdeniz G, Larsen TM, et al. New nordic diet versus average Danish diet: a randomized controlled trial revealed healthy long-term effects of the New Nordic diet by GC-MS blood plasma metabolomics. J Proteome Res. 2016;15(6):1939–54. [DOI] [PubMed] [Google Scholar]
- 26.Hirano T. Pathophysiology of diabetic dyslipidemia. J Atheroscler Thromb. 2018;25(9):771–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Krauss RM. Lipids and lipoproteins in patients with type 2 diabetes. Diabetes Care. 2004;27(6):1496–504. [DOI] [PubMed] [Google Scholar]
- 28.Shiu SW, Wong Y, Tan KC. Pre-β1 HDL in type 2 diabetes mellitus. Atherosclerosis. 2017;263:24–8. [DOI] [PubMed] [Google Scholar]
- 29.Chait A, Ginsberg HN, Vaisar T, Heinecke JW, Goldberg IJ, Bornfeldt KE. Remnants of the triglyceride-rich lipoproteins, diabetes, and cardiovascular disease. Diabetes. 2020;69(4):508–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Dullaart RP, de Vries R, Kwakernaak AJ, Perton F, Dallinga-Thie GM. Increased large VLDL particles confer elevated cholesteryl ester transfer in diabetes. Eur J Clin Invest. 2015;45(1):36–44. [DOI] [PubMed] [Google Scholar]
- 31.Vergès B. Pathophysiology of diabetic dyslipidaemia: where are we? Diabetologia. 2015;58(5):886–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lamarche B, Uffelman KD, Carpentier A, Cohn JS, Steiner G, Barrett PH, et al. Triglyceride enrichment of HDL enhances in vivo metabolic clearance of HDL apo A-I in healthy men. J Clin Invest. 1999;103(8):1191–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Adiels M, Taskinen MR, Packard C, Caslake MJ, Soro-Paavonen A, Westerbacka J, et al. Overproduction of large VLDL particles is driven by increased liver fat content in man. Diabetologia. 2006;49(4):755–65. [DOI] [PubMed] [Google Scholar]
- 34.Ginsberg HN, MacCallum PR. The obesity, metabolic syndrome, and type 2 diabetes mellitus pandemic: Part I. Increased cardiovascular disease risk and the importance of atherogenic dyslipidemia in persons with the metabolic syndrome and type 2 diabetes mellitus. J Cardiometab Syndr. 2009;4(2):113–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Adiels M, Olofsson S-O, Taskinen M-R, Borén J. Overproduction of very low-density lipoproteins is the hallmark of the dyslipidemia in the metabolic syndrome. Arterioscler Thromb Vasc Biol. 2008;28(7):1225–36. [DOI] [PubMed] [Google Scholar]
- 36.Sorokin AV, Patel N, Abdelrahman KM, Ling C, Reimund M, Graziano G, et al. Complex association of apolipoprotein E-containing HDL with coronary artery disease burden in cardiovascular disease. JCI Insight. 2022. 10.1172/jci.insight.159577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jensen MK, Aroner SA, Mukamal KJ, Furtado JD, Post WS, Tsai MY, et al. High-density lipoprotein subspecies defined by presence of apolipoprotein C-III and incident coronary heart disease in four cohorts. Circulation. 2018;137(13):1364–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hiukka A, Ståhlman M, Pettersson C, Levin M, Adiels M, Teneberg S, et al. ApoCIII-enriched LDL in type 2 diabetes displays altered lipid composition, increased susceptibility for sphingomyelinase, and increased binding to biglycan. Diabetes. 2009;58(9):2018–26. [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
Data Availability Statement
The data generated during and/or analyzed in the current study are available from the corresponding author upon reasonable request.