Skip to main content
Download PDF
Download ePub

Advanced serum lipoprotein and glycoprotein profiling for cardiovascular event prediction in type 2 diabetes mellitus: the LIPOCAT study

Cardiovascular Diabetology volume 24, Article number: 88 (2025) Cite this article

Abstract

Background

Traditional risk factors cannot accurately predict cardiovascular events (CVE) in type 2 diabetes (T2D). The LIPOCAT study aimed to prospectively evaluate the clinical utility of advanced lipoprotein characteristics and glycoproteins to predict future cardiovascular events (CVE) in a large cohort of subjects with type 2 diabetes mellitus (T2D).

Methods

From four different Spanish prospective cohorts, a total of 933 T2D subjects were selected to form the LIPOCAT study. Advanced 1H-Nuclear Magnetic Resonance (1H-NMR) analysis included lipoprotein (Liposcale®) and glycoprotein (Glycoscale) profiling. Random forest classification models and Area Under the Receiver Operating Characteristics (AUROC) analysis were used to assess the differential contribution of advanced variables in predicting CVE. Validation was performed using an external cohort.

Results

Out of 933 T2D subjects, 104 reported a CVE during follow-up. Analysis of Liposcale®/Glycoscale uncovered elevations in the circulating VLDL-cholesterol(C), remnant IDL-triglycerides (TG) and LDL-TG in subjects with CVE, along with glycoproteins (Glyc) A and B. Moreover, the incorporation of advanced Liposcale® variables to a base model constructed with traditional risk factors significantly improved the prediction of CVE, as evidenced by 1.5-fold increase in the C statistic (AUROC), reaching AUROC values of 0.756. In the independent validation cohort, similar improvements in AUROC values were observed by adding the advanced variables to the traditional models.

Conclusions

Advanced 1H-NMR analysis revealed previously hidden lipoprotein and glycoprotein characteristics associated with CVE in T2D subjects.

Graphical abstract

Background

Type 2 diabetes mellitus (T2D) is a prevalent chronic condition and a significant global public health concern [1]. Individuals with T2D face an increased risk of cardiovascular disease (CVD) [2]. The ability to accurately predict future cardiovascular events (CVEs) is vital for enabling targeted or precision prevention of this main complication. CVD risk prediction models have been designed to discriminate between individuals at lower or higher future risk, however, current models frequently fail to offer sufficient specificity to be used for clinical purposes, especially in T2D subjects [3]. Therefore, the identification of novel diagnostic markers for CVE in T2D subjects has become a subject of intense research.

Conventional serum lipid analysis frequently lacks the sensitivity to identify subjects with T2D at risk of CVE [4]. In this regard, the use of advanced nuclear magnetic resonance (1H-NMR) approaches, i.e., Liposcale®, has been successful in uncovering such hidden lipoprotein alterations associated with preclinical atherosclerosis in newly diagnosed T2D patients that were not captured by the traditional lipid profile analysis [5]. Therefore, advanced analysis of the lipoprotein profile may allow the identification of hidden candidate biomarkers for a more comprehensive CVD risk prevention strategy.

Clinical chronic, low-grade inflammation also contributes to many of the adverse complications of T2D, including CVD [6]. Indeed, low-grade inflammation is one of the key targets for the clinical management of subjects with T2D to reduce the CVD risk [7]. Studies using Glycoscale 1H-NMR spectra have identified a specific inflammation-responsive signal in serum or plasma samples termed Glyc A [8], which has been positively associated with different cardiometabolic conditions [9]. In line with this, Glyc A has also been identified as a reliable biomarker of cardiometabolic risk [10]. Indeed, Glyc A has been associated with an elevated incidence of CVD in recent independent clinical trials over other traditional CVD risk factors [11,12,13,14,15]. The detection of Glyc A has also been linked to incident T2D, even after adjusting for other traditional risk factors of T2D and hypersensitive C-reactive protein (hsCRP) [16, 17]. Overall, these findings might suggest that Glyc A could be considered as a new independent biomarker of chronic, low-grade inflammation related to cardiovascular burden in T2D subjects.

Taken together, quantitative characteristics of lipoprotein and glycoprotein profiles may be clinically relevant predictive biomarkers for risk assessment of CVE, especially in subjects with T2D, who are more prone to develop atherogenic dyslipidemia and inflammation. Therefore, we aimed to characterize the serum 1H-NMR-lipoprotein and glycoprotein profiles in 933 subjects with T2D collected from four different Spanish prospective population-based cohorts. Furthermore, we also aimed to assess the added predictive value for CVE prediction, if any, provided by the incorporation of advanced, non-traditional lipoprotein (i.e., Liposcale®) and glycoprotein characteristics (i.e., Glycoscale). Finally, we externally validated the CVE prediction models with an independent cohort including 187 individuals with T2D.

Methods

Study population

The LIPOCAT study was prospectively built up using four independent cohorts recruited from several areas of Catalonia in the north-east of Spain. Subjects with T2D matched for age, sex and body mass index (BMI) who had a serum sample available were selected and a follow-up update was performed to determine the cardiovascular events and all-cause mortality. The four cohorts that form the LIPOCAT study were previously described: Peripheral Arterial Disease study (PERART/ARTPER) [18, 19], Chronic Liver Disease-Fibro Scan cohort (FIBROSCAN) [21], Diabetes Mellitus study (DM) [22], DIABIMCAP Study (Carotid Atherosclerosis in Newly Diagnosed Type 2 Diabetic Individuals) [24]. We used an independent cohort, the Di@bet.es [25], to validate the results. The characteristics of the five cohorts are detailed in the Supplementary file 1.

The data collection procedures for all cohorts have been extensively detailed in previous publications [19, 21, 24, 25]. Key clinical variables recorded included age, sex, smoking habit, systolic blood pressure (SBP), diastolic blood pressure (DBP), glycated hemoglobin (HbA1c), and lipid profile. T2D was defined by glucose > 126 mg/dL, HbA1c > 6.5%, ongoing treatment or a medical diagnosis. Dyslipidemia was defined by a cholesterol level > 200 mg/dL, ongoing treatment or a medical diagnosis. Hypertension was defined by SBP > 140 mmHg or DBP > 90 mmHg or treatment or a medical diagnosis. Remnant cholesterol was calculated according to the consensus European Atherosclerosis Society (EAS) formula: remnants cholesterol = triglycerides (TG) * (VLDL-C/VLDL-TG) [31]. Additionally, remnant cholesterol was calculated using variables from Liposcale®, using the formula: remnants cholesterol = VLDL-C + IDL-C. Standardized protocols were followed for the measurement of weight, height, and blood pressure. Blood samples were collected in a fasting state, analyzed using standardized methods to determine biochemical parameters. For the Liposcale® test and the glycoprotein profile, blood samples were collected in a fasting state into EDTA tubes, processed immediately after extraction, and stored at − 80 °C at the biobanks of the participant centers until analysis.

A comprehensive review of medical records, encompassing all available information from the healthcare system, was conducted to identify occurrences of all-cause mortality and CVEs during the follow-up update. Incident cardiovascular endpoints were identified using codes from the International Classification Diseases (ICD-9 and ICD-10). These codes were used to identify and track specific cardiovascular events, procedures, and mortality outcomes throughout the follow-up period. Specific diagnostic codes included: Acute myocardial infarction (410, I21, I22), Angina (413, 411.1, I20), Unstable angina (411.1, I20.0), Ischemic heart disease (410–414, I20, I25 excluding 414.10, 414.11, 414.12, 414.19), Stroke (431, 433, 434, 435, 438, I61, I63, I65, I69, G45, G46), Peripheral arterial disease (443.1, 443.8, 443.9, 444, I73.1, I73.8, I73.9, I74), Heart failure (40,201, 40,211, 40,291, 40,401, 40,403, 40,411, 40,413, 40,491, 40,493, 428, I50, I11.0, I13.0, I13.2). In addition, relevant procedures were monitored, including Coronary revascularization (00.66, 36.0, 36.1, 36.2, 36.3) and Revascularization of other vascular territories (39.50, 39.90, 39.79, 00.60, 00.61, 00.62, 00.63, 00.64, 00.65). For mortality assessment: Cardiovascular death (390–459, I00-I09 excluding 427.5, 435, 446, 459.0) and Overall mortality (001-E999, A00-Y89) were also documented. The events included ischemic heart disease (encompassing any documented diagnosis such as angina pectoris), stroke, heart failure, peripheral artery disease, revascularization procedures, and cardiovascular mortality. To ensure accuracy, CVEs were only confirmed if a treating physician had documented a new diagnosis corresponding to the event after hospital admission or in outpatient medical records. This meticulous approach aimed to maintain the integrity and reliability of the collected data. The study protocol for the LIPOCAT study was approved by the Ethics Committees of University Hospital Germans Trias i Pujol (PI-18–039). All participants were informed about the study protocols and provided their consent to participate.

NMR molecular profiling

Three hundred µL of serum samples were shipped on dry ice to Biosfer Teslab (Reus, Spain) for the NMR analysis which included the lipoprotein profile based on the Liposcale® test [27] and the glycoprotein profile [28].

Serum samples were thawed overnight and prepared for NMR analyses. High-resolution 1H-NMR spectroscopy data was acquired on a Bruker 600 MHz spectrometer using LED Diffusion (Diff) experiments to detect larger molecules such as lipoproteins and glycoproteins, running at 37 °C in quantitative conditions.

The Liposcale® test was used to obtain the composition, the mean size and the number of lipoprotein particles of nine subtypes i.e. the main lipoprotein types (VLDL, LDL and HDL) further subdivided into large, medium and small particles for each type according to their respective size. Using the same NMR spectra, the general measurement of circulating glycoproteins was obtained by deconvoluting the specific region where glycoproteins resonate using analytical functions, and the area was quantified in proportion to the concentration of the acetyl groups of N-acetylglucosamine and N-acetyl galactosamine (Glyc A) and acetyl groups of N-acetylneuraminic acid (Glyc B). We analyzed the 1H-NMR spectral region where glycoproteins resonate (2.15–1.90 ppm) using several analytical functions, following a previously published procedure [28].

Statistical analysis

Continuous variables were tested for normality using the Kolmogorov–Smirnov test. Data are presented as median (25th percentile–75th percentile), and n and percentages for categorical variables. Differences between groups were analyzed using the non-parametric Mann–Whitney test or Student’s parametric t-test for continuous variables and the chi-square test or Fisher’s exact test for categorical variables.

We used random forests to construct the prediction models of CVE. Model parameters were adjusted using fivefold cross-validation, and the models were trained on 80% of the dataset and tested on the remaining 20% of data, based on a strategy that has proved to be more accurate for complex scenarios compared to other multivariate approaches [29]. The first model (Model 1) was performed with the traditional cardiovascular risk factors (CVRFs): sex, age, BMI, hypertension, dyslipidemia and smoking. Model 2 was built on top of Model 1 by further incorporating conventional lipid variables: total cholesterol (Total C), HDL-C, LDL-C, TG and remnant cholesterol using the EAS formula. Model 3, also built upon Model 1, incorporated advanced integrative variables of Liposcale®, including ratios and percentages of quantitative characteristics of lipoprotein metabolism. Specifically, the variables included the VLDL-TG-to-VLDL-C ratio, IDL-TG-to-IDL-C ratio, LDL-TG-to-LDL-C ratio, HDL-TG-to-HDL-C, the percentages of smaller VLDL, LDL and HDL, remnant cholesterol calculated using the Liposcale® formula and incorporating glycoproteins Glyc A and Glyc B. To assess model performance, we computed the accuracy and the area under the ROC curve (AUROC) for all models. Differences between AUROC values were assessed using DeLong’s test for correlated ROC curves, with pairwise comparisons performed across models. Statistical significance was set at p < 0.05. Statistical analyses were performed using the R statistical software version 4.4.0 [30].

Results

Characteristics of the LIPOCAT cohort

In the total cohort of 933 T2D subjects drawn from the four distinct prospective cohorts, 104 individuals experienced at least one CVE during the follow-up period. Given the matching by age, sex and BMI, these variables did not differ between CVE groups i.e. with and without CVEs. Additionally, there were no differences in smoking habits between the groups (Table 1).

Table 1 Clinical variables and incidence of cardiovascular events in the prospective T2D cohort (LIPOCAT)
Full size table

The proportion of T2D subjects receiving antihypertensive medication was significantly higher in those with a CVE than those without. Correspondingly, significant elevations in SBP were observed in the CVE group compared with the non-CVE subjects. The HbA1c levels were significantly higher in the CVE group, indicating suboptimal glycemic control. However, conventional total lipids did not differ between both groups, although total cholesterol was marginally elevated in the CVE group. No changes were observed in the serum concentrations of HDL- and LDL-cholesterol.

During the follow-up period among the CVE group, peripheral arterial disease was the most common event (29.8%), followed by heart failure (22.7%), stroke (21.2%), acute myocardial infarction, (18.3%), angina pectoris (17.3%), ischemic heart disease (16.3%) and coronary revascularization (15.4%). Cardiovascular mortality was 10.6% (Table 1).

Advanced lipoprotein and glycoprotein variables

The advanced analysis of serum lipoproteins using the NMR approach revealed a significant increase in total triglycerides in the CVE group. This increase was mainly due to significant increases of IDL- and LDL-triglycerides and VLDL-triglycerides (p-value = 0.054). Despite total cholesterol not differing between groups, NMR analysis uncovered significant elevations in the VLDL-cholesterol and marginally in the IDL-cholesterol (p-value = 0.058) in the CVE group. Interestingly, the serum concentration of LDL-cholesterol did not differ between groups (Table 2).

Table 2 Advanced lipoprotein variables of the LIPOCAT study
Full size table

The triglyceride-to-cholesterol ratio calculated for each lipoprotein fraction revealed a significant increase only in the LDL, partly attributed to the relatively higher triglyceride content in the serum from subjects in the CVE group.

Regarding the particle size, our data showed that the circulating concentrations of larger HDL were significantly elevated in the CVE group (p-value = 0.007). Consistently, the small HDL-P-to-medium HDL-P ratio was concurrently decreased in the CVE group (p-value = 0.039). Medium VLDL was significantly increased in the CVE group. No significant differences were observed in other cardiovascular surrogates calculated from quantitative characteristics of lipoproteins, but the HDL-C-to-total triglycerides showed a trend (p-value = 0.063) towards being decreased in the CVE group compared with the non-CVE group.

Serum glycoproteins, especially Glyc A, are composite biomarkers of inflammation that can also be detected by [1H]-NMR. Median values of Glyc A and Glyc B were higher in T2D subjects with CVE compared with those without (Table 3).

Table 3 Advanced glycoprotein variables of the LIPOCAT study
Full size table

Contribution of advanced lipoprotein and glycoproteins characteristics to CVE prediction

We developed three models with the overall aim of examining whether the addition of advanced variables (i.e., Liposcale® and Glycoscale) that were differentially changed in the CVE group (i.e., IDL-TG, LDL-TG and diameter, and glycated proteins) (Model 3) could improve CVE prediction over traditional CVRFs such as BMI, age, sex, hypertension, dyslipidemia, smoking habit (Model 1), and traditional CVRFs together with the conventional lipid profile (Model 2) (Fig. 1; Table 4 ). We used random forests to construct the prediction models of CVE.

Fig 1.
figure 1

Random forest analysis for CVE prediction. Model 1: a Random Forest model of cardiovascular risk factors (including sex, age, BMI, hypertension, dyslipidemia, smoking habit) and b Model 2: Random Forest of Model 1 + conventional lipid profile (i.e., total triglycerides, total cholesterol, HDL-C and LDL-C, remnant cholesterol calculated by EAS formula = Triglycerides (TG) * (VLDL-C/VLDL-TG). c Model 3: Random forest of Model 1 + Liposcale® + Glycoproteins variables and remnant cholesterol calculated by the Liposcale® formula = VLDL-C + IDL-C. Yellow dots = traditional cardiovascular risk factors. Orange dots = conventional lipid profile. Blue dots = Liposcale® + Glycoproteins variables

Full size image

Model 1: Designed with traditional CVRFs, random forest analysis identified BMI, age and hypertension as the most important variables for making accurate predictions (scores higher than 75%) (Fig. 1a).

Model 2: We incorporated the conventional lipid profile (i.e., total cholesterol, TG, LDL-C, and HDL-C) and estimated cholesterol remnants (EAS formula) to analyze their added effect on CVE risk. Random forest analysis again identified BMI as the most important variable. This analysis also revealed that age, total cholesterol, LDL-C, HDL-C, remnant cholesterol, and total triglycerides scored within the top quartile (> 75%) on the variable importance scale for accurate CVE prediction (Fig. 1b).

Model 3: We incorporated the advanced characteristics of lipoproteins (Liposcale®) and glycoproteins (Glycoscale) into Model 1. In this model cholesterol remnants were calculated using the Liposcale® formula. BMI, age, Glyc A, Glyc B, and all lipids scored over 75% on the variable importance scale for accurate CVE prediction (Fig. 1c).

The AUROC score calculated in Model 3 was 1.5-fold higher (AUROC: 0.756) compared with Model 1 (0.501) (Δ value = 0.255). Likewise, the balanced accuracy value estimated using Model 1, which was strictly built using only traditional risk factors, was 0.492, much lower than Model 3, which had the addition of Liposcale® and Glycoscale variables (0.709). Supplementary Table 1 details false-positive and false-negative rates for each model. (Table 4).

Table 4 Performance Metrics of Predictive Models for Cardiovascular Events in the LIPOCAT study
Full size table

Internal validation cohort analysis

The addition of Liposcale® and Glycoscale variables to the traditional base model increased the prediction accuracy for CVE, as revealed by the use of confusion matrices. Internal analysis by comparing all models revealed that predictability improved from 25% (Model 1) to 85% (Model 3). Moreover, the rate of false negatives was significantly reduced when adding the advanced variables to the traditional ones (Model 3, 15%) compared to using only the traditional CVRFs (Model 1, 75%). However, the relative rate of false positives increased when using the Model 3 (45%) compared to Model 1 (27%).

The comparison of ROC curves (Fig. 2a)and AUC values (Fig. 2c) supported the added value of using advanced variables to predict CVE in the LIPOCAT cohort. Interestingly, the AUC values calculated in Model 3 were significantly increased compared with those calculated using Model 1 (p-value = 0.004) and Model 2 (p-value = 0.002).

Fig 2.
figure 2

Predictability accuracy and ROC curves in the LIPOCAT cohort. a. ROC curves. b. Confusion matrices. c. AUC curves. Model 1: Traditional CV risk factors (CVRF); Model 2: Traditional CVRF + conventional lipid variables, including remnant cholesterol EAS formula = triglycerides (TG) * (VLDL-C/VLDL-TG). Model 3: Traditional CVRF + Liposcale®, + Glycoscale variables, including remnant cholesterol Liposcale® formula = VLDL-C + IDL-C

Full size image

To assess potential differences in predictive performance across cardiovascular event (CVE) types, we applied Model 3 (traditional CV risk factors, Liposcale®, and Glycoscale markers) to individual CVE subgroups. As shown in Table 5, the model demonstrated moderate to good discrimination across most event types, with (AUROC values ranging from 0.571 to 0.866).

Table 5 Predictive performance of Model 3 across different cardiovascular event subgroups in the LIPOCAT cohort
Full size table

To further evaluate the predictive performance of our models, we conducted additional analyses incorporating family history of cardiovascular disease, glycoproteins (Glyc A and Glyc B), HbA1c levels, and SCORE2 alongside traditional cardiovascular risk factors (CVRF). The inclusion of family history of cardiovascular disease slightly reduced the false negative rate; however, it did not significantly improve the overall model performance, as reflected by unchanged AUC values (0.756, Supplementary Table 2 and Supplementary Fig. 1). Upon adding glycoproteins to the traditional CVRF model (Model 4, Supplementary Fig. 2), the AUC increased to 0.629, indicating improved predictive performance compared to the base model using CVRF alone (Model 1). However, this improvement was not statistically significant in comparison to the model that incorporated both Liposcale® and Glycoscale variables (Model 3). Incorporation of HbA1c into the traditional CVRF and Liposcale® parameters (Model 4, Supplementary Fig. 3) resulted in an increased AUC (0.682) compared to the base model using CVRF alone (Model 1). However, this improvement was less pronounced than that observed in Model 3, which included both Liposcale® and Glycoscale variables. Consequently, the addition of HbA1c did not surpass the predictive performance of the model incorporating glycoproteins, and the overall conclusions of the study remain unchanged. Integrating the SCORE2 risk score into the traditional CVRF model (Model 4, Supplementary Fig. 4) led to an AUC increase to 0.678, reflecting improved predictive performance over CVRF alone. The subsequent addition of the HbA1c variable in Model 5 resulted in a slight enhancement in discrimination, though this did not substantially alter the AUROC values compared to Model 4. Importantly, both models demonstrated inferior predictive accuracy relative to Model 3, which consistently yielded the highest performance. Overall, Model 3, which integrates traditional CVRF with Liposcale® and Glycoscale variables, consistently demonstrated the best performance for CVE prediction across all analyses.

External validation cohort analysis

To validate the results from the LIPOCAT cohort, we tested the models and conducted an analysis using confusion matrices and calculated the AUROC values with an independent cohort comprising 187 subjects, of which 41 had a CVE. The clinical variables are detailed in Supplementary Table 3, while the Liposcale® and Glycoscale variables are provided in Supplementary Table 4. This validation process allowed us to assess the performance and generalizability of the models using a separate dataset.

Consistent with the LIPOCAT cohort, this validation cohort showed that the AUROC score values calculated when adding the Liposcale® and Glycoscale variables to traditional CVRFs were 1.2-fold higher (AUROC: 0.669) compared with the traditional base model (0.540) (Δ value = 0.129) (Fig. 3).

Fig. 3
figure 3

Predictability accuracy and ROC curves in the validation cohort. a. ROC curves. b. confusion matrices. c. AUC curves. Model 1: Traditional CV risk factors; Model 2: Traditional CV risk factors + conventional lipid variables, including remnant cholesterol EAS formula = triglycerides (TG) * (VLDL-C/VLDL-TG). Model 3: Traditional CV risk factor + Liposcale®, + Glycoscale variables, including remnant cholesterol Liposcale® formula = VLDL-C + IDL-C

Full size image

Our analysis also revealed that the balanced accuracy value estimated using the traditional base model, which was strictly built only using traditional CVRFs, was 0.499, much lower than that obtained by adding the Liposcale® and Glycoscale variables (0.631) (Table 6).

Table 6 Performance Metrics of Predictive Models for Cardiovascular Events in the validation cohort
Full size table

The addition of Liposcale® and Glycoscale variables to the traditional base Model 1 also increased the prediction accuracy for CVE in the external validation cohort (Fig. 3a), similar to the internal validation analysis results (Fig. 2a). The estimated accuracy was greater when using Model 3 (74%) compared with Model 1 (20%). The false negatives were also reduced when adding the advanced to the traditional CVRFs (Model 3, 26%) compared with the matrix built solely on traditional factors (Model 1, 80%). However, the relative rate of false positives remained much higher for Model 3 (48%) compared with Model 1 (20%).

The comparison of ROC curves (Fig. 3b) and AUC values (Fig. 3c) supports the added value of using advanced variables to predict CVE in this validation cohort, fully consistent with the data obtained using the LIPOCAT cohort.

Discussion

Conventional clinical lipid/lipoprotein variables do not show the full complexity of the altered lipid metabolism associated with increased cardiovascular burden in subjects with T2D [32]. Conversely, the relationship between advanced lipoprotein characteristics or glycoproteins and atherosclerosis has recently been reported in subjects with T2D [5, 33]. However, their contribution to CVE in this complex metabolic context has not been previously evaluated. In this study, we assessed the predictive potential of quantitative non-traditional variables analyzed by 1H-NMR for CVE across 933 T2D subjects enrolled from different cohorts. Advanced NMR methodologies enable the assessment of hidden, differentially expressed molecules related to lipoprotein metabolism and inflammation, which could be candidate biomarkers of CVE in subjects with T2D.

Our NMR data revealed elevations in total triglycerides, primarily due to significant increases in triglycerides in VLDL, IDL, and LDL, although VLDL-TGs were only marginally increased in the CVE group. These findings suggest impaired lipolysis, which could lead to the accumulation of cholesterol remnants in this group of subjects [31]. In support of this is the evidence showing an accumulation of VLDL and IDL in T2D subjects in the CVE group. In this respect, firstly, VLDL-C, and IDL-C albeit marginally, were increased in the CVE group. Secondly, due to the nature of circulating IDL particles, our [1H]-NMR approach did not classify this class of lipoproteins by size. Despite this limitation, medium and smaller VLDL-P were also increased in the CVE group; thus, their sum could be considered a novel correlate of such remnant lipoprotein. Lastly, the remnant cholesterol levels in the CVE group were significantly elevated compared with the non-CVE group. Moreover, the calculated levels of cholesterol remnants were consistent with those directly determined by the NMR approach. Indeed, our NMR analysis provided a more accurate estimation of serum cholesterol remnant concentrations than previous studies [34, 35], which simply calculated them by subtracting the LDL-C and HDL-C moieties from total cholesterol determined using conventional methods.

Although the atherogenicity of cholesterol remnants, which include both medium-to-smaller range of VLDL and IDL, is an old concept [36], there is renewed interest due to their predictive potential for CVE [31, 37]. The accumulation of cholesterol remnants in circulation is considered a potential risk factor for future CVE in T2D subjects [38]; consistent with our data. Specifically, elevations in circulating IDL are highly proatherogenic [39] and considered a risk factor for CVD [40]. In this context, we identified the IDL-TG-to-IDL-C ratio (Liposcale®) as a variable associated with CVE. However, no significant differences in the ratio were observed between the CVE and non-CVE groups.

Another important finding from our study was the elevated concentrations of LDL-TG in the CVE group compared with the non-CVE group. Our data align with other studies where LDL-TG has been recently identified as a risk factor for CVD [41, 42]. Since CVE parallels the development of atherosclerosis in T2D [43], increased triglyceride content in LDL might also be a potential candidate risk factor of CVE; however, LDL-TG did not appear among the advanced variables significantly linked to CVE.. In our cohort, despite the absence of significant differences in LDL-cholesterol levels, both groups (with and without CVE) exhibited a high proportion of small LDL particles, with over 50% of total LDL belonging to the small subclass (particle size < 21 nm). This lipid profile is characteristic of atherogenic dyslipidemia, a well-recognized contributor to increased cardiovascular risk in patients with type 2 diabetes. Our findings suggest that LDL particle composition, rather than LDL-cholesterol concentration alone, may be a more relevant marker of atherogenicity in this population.

Systemic chronic inflammation, a frequent feature in diabetes [44], is also related to an elevated risk of CVE in T2D subjects [45]. In this context, another set of variables differentially expressed in the CVE group was glycoproteins. Recent research has linked circulating NMR-derived glycoproteins with chronic inflammation in various clinical studies [28, 33], suggesting their potential as systemic biomarkers of inflammation. Supporting this notion, glycoprotein elevations has been associated with atherogenic dyslipidemia in subjects with T2D [33]. Importantly, their added predictive value above other traditional risk factors of CVE was unveiled in this study, along with advanced lipoproteins characteristics.

In the external validation cohort, the predictive performance of models incorporating advanced NMR-based lipid and glycoprotein variables did not significantly exceed that of models using traditional lipid parameters. This contrasts with findings from the LIPOCAT cohort, where NMR markers improved CVE prediction. The differences in event composition between cohorts likely contributed to this discrepancy. The external cohort predominantly experienced stroke and heart failure, with no recorded cases of atherosclerotic cardiovascular disease (e.g., myocardial infarction), where lipid metabolism plays a more critical role. Additionally, the smaller sample size and the absence of mortality events may have limited the statistical power to detect differences in model performance.

Although NMR is currently considered a relatively expensive technique, its scalability and clinical utility make it a valuable tool for cardiovascular risk assessment. It is particularly beneficial for patients with dyslipidemia and residual cardiovascular risk, especially those with metabolic disorders such as diabetes, obesity, and metabolic syndrome. The Spanish Society of Arteriosclerosis (SEA) has identified key patient groups who could benefit most from this technology [46]. A preliminary evaluation estimated that around 0.1% of the Spanish population (~ 48,000 individuals) would be eligible for NMR-based lipoprotein analysis. Despite its cost, a single NMR instrument can process up to 50,000 samples per year ensuring sufficient analytical capacity without requiring additional infrastructure for broad territories. Given that a single plasma sample can provide over 70 molecular parameters related to cardiometabolic and inflammatory risk, NMR profiling represents a powerful tool for improving risk stratification and guiding preventive strategies.

One of the main strengths of this study was the inclusion of a large population of age, sex, and BMI-matched T2D subjects with follow-up of CVE. This eliminated major confounding factors, providing a better evaluation of lipoprotein characteristics uncovered by advanced [1H]-NMR analysis. Another important strength was the improved predictive value provided by incorporating Liposcale® variables to traditional clinical variables linked to CVE, confirmed elegantly in an independent validation cohort. However, our study had some limitations. The study did not examine traditional factors like diabetes duration or kidney disease estimates for their effects on CVE risk. This limitation prevented us from directly comparing our developed models to the SCORE2-diabetes risk prediction algorithms, which are currently used to predict events in reference populations. As these variables were not consistently recorded across cohorts, they may have obscured the predictive value of advanced [1H]-NMR-derived variables in our study. Although our model exhibits a relatively high false-positive rate, its value lies in identifying individuals with a molecular profile associated with elevated cardiovascular risk, even if they have not yet experienced the event. A high-risk classification may indicate that the patient is in a subclinical or pre-event stage, suggesting they are closer to experiencing a cardiovascular event due to their molecular signature. This predictive capacity is particularly valuable for early risk stratification, allowing for timely preventive interventions that could mitigate disease progression and reduce the likelihood of future adverse events.

Conclusions

Advanced 1H-NMR analysis reveal previously hidden lipoprotein and glycoprotein characteristics associated with CVE in T2D subjects. The enhanced prediction of CVE offered by these advanced variables, beyond traditional risk factors, suggests their potential as novel biomarkers for cardiovascular risk stratification in diabetes mellitus.

Data availability

Proposals relating to the data access should be directed to the corresponding authors. To gain access, data requestors will need to sign a data access agreement.

Abbreviations

CVD:

Cardiovascular disease

CVE:

Cardiovascular events

CVRFs:

Cardiovascular risk factors

IDL:

Intermediate-density lipoprotein

LDL:

Low-density lipoprotein

HDL:

High-density lipoprotein

NMR:

1H-nuclear magnetic resonance

T1D:

Type 1 diabetes mellitus

T2D:

Type 2 diabetes mellitus

TG:

Triglycerides

VLDL:

Very-low-density lipoprotein

References

  1. Khan MAB, Hashim MJ, King JK, Govender RD, Mustafa H, Al KJ. Epidemiology of type 2 diabetes—global burden of disease and forecasted trends. J Epidemiol Glob Health. 2020;10:107–11.

    Article  PubMed  PubMed Central  Google Scholar 

  2. Einarson TR, Acs A, Ludwig C, Panton UH. Prevalence of cardiovascular disease in type 2 diabetes: a systematic literature review of scientific evidence from across the world in 2007–2017. Cardiovasc Diabetol. 2018;17:83.

    Article  PubMed  PubMed Central  Google Scholar 

  3. Bachmann KN, Wang TJ. Biomarkers of cardiovascular disease: contributions to risk prediction in individuals with diabetes. Diabetologia. 2018;61:987–95.

    Article  CAS  PubMed  Google Scholar 

  4. Ma CX, Ma XN, Guan CH, Li YD, Mauricio D, Fu SB. Cardiovascular disease in type 2 diabetes mellitus: progress toward personalized management. Cardiovasc Diabetol. 2022;21:74.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Amor AJ, Catalan M, Perez A, Herreras Z, Pinyol M, Sala-Vila A, et al. Nuclear magnetic resonance lipoprotein abnormalities in newly-diagnosed type 2 diabetes and their association with preclinical carotid atherosclerosis. Atherosclerosis. 2016;247:161–9.

    Article  CAS  PubMed  Google Scholar 

  6. Furman D, Campisi J, Verdin E, Carrera-Bastos P, Targ S, Franceschi C, et al. Chronic inflammation in the etiology of disease across the life span. Nat Med. 2019;25:1822–32.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Donath MY, Shoelson SE. Type 2 diabetes as an inflammatory disease. Nat Rev Immunol. 2011;11:98–107.

    Article  CAS  PubMed  Google Scholar 

  8. Otvos JD, Shalaurova I, Wolak-Dinsmore J, Connelly MA, Mackey RH, Stein JH, et al. GlycA: a composite nuclear magnetic resonance biomarker of systemic inflammation. Clin Chem. 2015;61:714–23.

    Article  CAS  PubMed  Google Scholar 

  9. Connelly MA, Otvos JD, Shalaurova I, Playford MP, Mehta NN. GlycA, a novel biomarker of systemic inflammation and cardiovascular disease risk. J Transl Med. 2017;15:219.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Connelly MA, Gruppen EG, Otvos JD, Dullaart RPF. Inflammatory glycoproteins in cardiometabolic disorders, autoimmune diseases and cancer. Clin Chim Acta. 2016;459:177–86.

    Article  CAS  PubMed  Google Scholar 

  11. McGarrah RW, Kelly JP, Craig DM, Haynes C, Jessee RC, Huffman KM, et al. A novel protein glycan-derived inflammation biomarker independently predicts cardiovascular disease and modifies the association of HDL subclasses with mortality. Clin Chem. 2017;63:288–96.

    Article  CAS  PubMed  Google Scholar 

  12. Akinkuolie AO, Buring JE, Ridker PM, Mora S. A novel protein glycan biomarker and future cardiovascular disease events. J Am Heart Assoc. 2014;3: e001221.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Gruppen EG, Riphagen IJ, Connelly MA, Otvos JD, Bakker SJ, Dullaart RP. GlycA, a pro-inflammatory glycoprotein biomarker, and incident cardiovascular disease: relationship with C-reactive protein and renal function. PLoS ONE. 2015;10: e0139057.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Duprez DA, Otvos J, Sanchez OA, Mackey RH, Tracy R, Jacobs DR Jr. Comparison of the predictive value of GlycA and other biomarkers of inflammation for total death, incident cardiovascular events, noncardiovascular and noncancer inflammatory-related events, and total cancer events. Clin Chem. 2016;62:1020–31.

    Article  CAS  PubMed  Google Scholar 

  15. Akinkuolie AO, Glynn RJ, Padmanabhan L, Ridker PM, Mora S. Circulating N-linked glycoprotein side-chain biomarker, rosuvastatin therapy, and incident cardiovascular disease: an analysis from the JUPITER trial. J Am Heart Assoc. 2016; 5.

  16. Akinkuolie AO, Pradhan AD, Buring JE, Ridker PM, Mora S. Novel protein glycan side-chain biomarker and risk of incident type 2 diabetes mellitus. Arterioscler Thromb Vasc Biol. 2015;35:1544–50.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Connelly MA, Gruppen EG, Wolak-Dinsmore J, Matyus SP, Riphagen IJ, Shalaurova I, et al. GlycA, a marker of acute phase glycoproteins, and the risk of incident type 2 diabetes mellitus: PREVEND study. Clin Chim Acta. 2016;452:10–7.

    Article  CAS  PubMed  Google Scholar 

  18. Alzamora MT, Baena-Diez JM, Sorribes M, Fores R, Toran P, Vicheto M, et al. Peripheral Arterial Disease study (PERART): prevalence and predictive values of asymptomatic peripheral arterial occlusive disease related to cardiovascular morbidity and mortality. BMC Public Health. 2007;7:348.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Alzamora MT, Fores R, Baena-Diez JM, Pera G, Toran P, Sorribes M, et al. The peripheral arterial disease study (PERART/ARTPER): prevalence and risk factors in the general population. BMC Public Health. 2010;10:38.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Vicente I, Lahoz C, Taboada M, Garcia A, San Martin MA, Terol I, et al. Prevalence of an abnormal ankle-brachial index in relation to the cardiovascular risk estimated by the Framingham function. Med Clin. 2005;124:641–4.

    Article  Google Scholar 

  21. Caballeria L, Pera G, Arteaga I, Rodriguez L, Aluma A, Morillas RM, et al. High prevalence of liver fibrosis among european adults with unknown liver disease: a population-based study. Clin Gastroenterol Hepatol. 2018;16(1138–45): e5.

    Google Scholar 

  22. Alonso N, Traveset A, Rubinat E, Ortega E, Alcubierre N, Sanahuja J, et al. Type 2 diabetes-associated carotid plaque burden is increased in patients with retinopathy compared to those without retinopathy. Cardiovasc Diabetol. 2015;14:33.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Vilanova MB, Falguera M, Marsal JR, Rubinat E, Alcubierre N, Castelblanco E, et al. Prevalence, clinical features and risk assessment of pre-diabetes in Spain: the prospective Mollerussa cohort study. BMJ Open. 2017;7: e015158.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Catalan M, Herreras Z, Pinyol M, Sala-Vila A, Amor AJ, de Groot E, et al. Prevalence by sex of preclinical carotid atherosclerosis in newly diagnosed type 2 diabetes. Nutr Metab Cardiovasc Dis. 2015;25:742–8.

    Article  CAS  PubMed  Google Scholar 

  25. Soriguer F, Goday A, Bosch-Comas A, Bordiu E, Calle-Pascual A, Carmena R, et al. Prevalence of diabetes mellitus and impaired glucose regulation in Spain: the diabetes study. Diabetologia. 2012;55:88–93.

    Article  CAS  PubMed  Google Scholar 

  26. Classification and Diagnosis of Diabetes. Standards of medical care in diabetes-2022. Diabetes Care. 2022;45:S17–38.

    Article  Google Scholar 

  27. Mallol R, Amigo N, Rodriguez MA, Heras M, Vinaixa M, Plana N, et al. Liposcale®: a novel advanced lipoprotein test based on 2D diffusion-ordered 1H NMR spectroscopy. J Lipid Res. 2015;56:737–46.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Fuertes-Martin R, Taverner D, Vallve JC, Paredes S, Masana L, Correig Blanchar X, et al. Characterization of (1)H NMR plasma glycoproteins as a new strategy to identify inflammatory patterns in rheumatoid arthritis. J Proteome Res. 2018;17:3730–9.

    Article  CAS  PubMed  Google Scholar 

  29. Couronne R, Probst P, Boulesteix AL. Random forest versus logistic regression: a large-scale benchmark experiment. BMC Bioinformatics. 2018;19:270.

    Article  PubMed  PubMed Central  Google Scholar 

  30. team Rc. R: A language and environment for statistical computing. Vienna2021.

  31. Ginsberg HN, Packard CJ, Chapman MJ, Boren J, Aguilar-Salinas CA, Averna M, et al. Triglyceride-rich lipoproteins and their remnants: metabolic insights, role in atherosclerotic cardiovascular disease, and emerging therapeutic strategies-a consensus statement from the European atherosclerosis society. Eur Heart J. 2021;42:4791–806.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Alshehry ZH, Mundra PA, Barlow CK, Mellett NA, Wong G, McConville MJ, et al. Plasma lipidomic profiles improve on traditional risk factors for the prediction of cardiovascular events in type 2 diabetes mellitus. Circulation. 2016;134:1637–50.

    Article  CAS  PubMed  Google Scholar 

  33. Amigo N, Fuertes-Martin R, Malo AI, Plana N, Ibarretxe D, Girona J, et al. Glycoprotein profile measured by a (1)H-nuclear magnetic resonance based on approach in patients with diabetes: a new robust method to assess inflammation. Life. 2021;11:1407.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Jansson Sigfrids F, Dahlstrom EH, Forsblom C, Sandholm N, Harjutsalo V, Taskinen MR, et al. Remnant cholesterol predicts progression of diabetic nephropathy and retinopathy in type 1 diabetes. J Intern Med. 2021;290:632–45.

    Article  CAS  PubMed  Google Scholar 

  35. Shan Y, Wang Q, Zhang Y, Tong X, Pu S, Xu Y, et al. High remnant cholesterol level is relevant to diabetic retinopathy in type 2 diabetes mellitus. Lipids Health Dis. 2022;21:12.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Abdel-Maksoud M, Sazonov V, Gutkin SW, Hokanson JE. Effects of modifying triglycerides and triglyceride-rich lipoproteins on cardiovascular outcomes. J Cardiovasc Pharmacol. 2008;51:331–51.

    Article  CAS  PubMed  Google Scholar 

  37. Hoogeveen RC, Ballantyne CM. Residual cardiovascular risk at low LDL: remnants, lipoprotein(a), and inflammation. Clin Chem. 2021;67:143–53.

    Article  PubMed  Google Scholar 

  38. Huh JH, Roh E, Lee SJ, Ihm SH, Han KD, Kang JG. Remnant cholesterol is an independent predictor of type 2 diabetes: a nationwide population-based cohort study. Diabetes Care. 2023;46:305–12.

    Article  CAS  PubMed  Google Scholar 

  39. Paquette M, Bernard S, Baass A. Dysbetalipoproteinemia is associated with increased risk of coronary and peripheral vascular disease. J Clin Endocrinol Metab. 2022;108:184–90.

    Article  PubMed  Google Scholar 

  40. Nordestgaard BG. Triglyceride-rich lipoproteins and atherosclerotic cardiovascular disease: new insights from epidemiology, genetics, and biology. Circ Res. 2016;118:547–63.

    Article  CAS  PubMed  Google Scholar 

  41. Voros S, Bansal AT, Barnes MR, Narula J, Maurovich-Horvat P, Vazquez G, et al. Bayesian network analysis of panomic biological big data identifies the importance of triglyceride-rich LDL in atherosclerosis development. Front Cardiovasc Med. 2022;9: 960419.

    Article  CAS  PubMed  Google Scholar 

  42. Abbate SL, Brunzell JD. Pathophysiology of hyperlipidemia in diabetes mellitus. J Cardiovasc Pharmacol. 1990;16(Suppl 9):S1-7.

    Article  PubMed  Google Scholar 

  43. Ruscitti P, Cipriani P, Masedu F, Romano S, Berardicurti O, Liakouli V, et al. Increased cardiovascular events and subclinical atherosclerosis in rheumatoid arthritis patients: 1 year prospective single centre study. PLoS ONE. 2017;12: e0170108.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Poznyak A, Grechko AV, Poggio P, Myasoedova VA, Alfieri V, Orekhov AN. The diabetes mellitus-atherosclerosis connection: the role of lipid and glucose metabolism and chronic inflammation. Int J Mol Sci. 2020 Mar 6;21 (5):1835.https://doiorg.publicaciones.saludcastillayleon.es/10.3390/ijms21051835

  45. Sharif S, Van der Graaf Y, Cramer MJ, Kapelle LJ, de Borst GJ, Visseren FLJ, et al. Low-grade inflammation as a risk factor for cardiovascular events and all-cause mortality in patients with type 2 diabetes. Cardiovasc Diabetol. 2021;20:220.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Pintó X, Masana L, Civeira F, Real J, Ibarretxe D, Candas B, Puzo J, Díaz JL, Amigó N, Esteban M, Valdivielso P. Consensus document of an expert group from the Spanish society of arteriosclerosis (SEA) on the clinical use of nuclear magnetic resonance to assess lipoprotein metabolism (Liposcale®). Clin Investig Arterioscler. 2020;32(5):219–29.

    PubMed  Google Scholar 

Download references

Acknowledgements

We gratefully acknowledge the participants and the support provided by the IGTP-HUGTP and IRBLleida Biobanks, integrated into the Spanish National Biobanks Network of ISCIII (PT17/0015/0045 and PT17/0015/0027, respectively). We extend our thanks to Amanda Prowse for her editorial assistance with this manuscript. We also acknowledge the support received from CIBERDEM (CB15/00071), a project of the ISCIIII, as well as the Agencia Estatal de Investigación (MCIN/AEI /https://doiorg.publicaciones.saludcastillayleon.es/10.13039/501100011033) and the European Union’s “NextGeneration EU”/PRTR program, under the "Consolidación Investigadora 2022" action (CNS2022-135559), which provided support to JJ. The Institut de Recerca de l'Hospital de la Santa Creu i Sant Pau, the Institut de Recerca Biomèdica de Lleida, and the Germans Trias i Pujol Research Institute are accredited by the Generalitat de Catalunya as Centres de Recerca de Catalunya (CERCA).

Funding

This work was supported by the Ministerio de Sanidad y Consumo and Instituto de Salud Carlos III through grant PI15/0625 (awarded to DM and EC), co-funded by the FEDER program “Una manera de hacer Europa.” Additional support was provided by Laboratorios Rubió S.A. through an unrestricted grant (LIPOCATV1FEB2018) for 1H-NMR characterization.

Author information

Author notes

  1. Núria Amigó and Esmeralda Castelblanco have contributed equally to this work and share the first authorship.

Authors and Affiliations

  1. Biosfer Teslab, Reus, Spain

    Núria Amigó & Neus Martínez-Micaelo

  2. Department of Basic Medical Sciences, Universitat Rovira I Virgili (URV), Institut d’Investigació Sanitària Pere Virgili (IISPV), Reus, Spain

    Núria Amigó

  3. Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Instituto de Salud Carlos III (ISCIII), Madrid, Spain

    Núria Amigó, Josep Julve, Núria Alonso, Josep Ribalta, Montse Guardiola, Josep Franch-Nadal & Didac Mauricio

  4. Department of Internal Medicine, Endocrinology, Metabolism and Lipid Research Division, Washington University School of Medicine, St. Louis, USA

    Esmeralda Castelblanco

  5. Department of Endocrinology & Nutrition, Hospital de La Santa Creu I Sant Pau, Institut de Recerca Sant Pau, Barcelona, Spain

    Josep Julve & Didac Mauricio

  6. Department of Endocrinology & Nutrition, Hospital Universitari Germans Trias I Pujol, Germans Trias I Pujol Research Institute, Badalona, Spain

    Núria Alonso

  7. Department of Endocrinology & Nutrition, Hospital Universitari Arnau de Vilanova, Institut de Recerca Biomèdica de Lleida (IRBLleida), Lleida, Spain

    Marta Hernández

  8. Department of Medicine and Surgery, Universitat Rovira I Virgili (URV), Institut d’Investigació Sanitària Pere Virgili (IISPV), Reus, Spain

    Josep Ribalta & Montse Guardiola

  9. Unitat de Suport a La Recerca Metropolitana Nord, Institut Universitari d’Investigació en Atenció Primària Jordi Gol (IDIAP Jordi Gol), Institut Català de la Salut, Mataro, Spain

    Pere Torán-Monserrat, Victor Lopez-Lifante, Cecilia Herrero-Alonso & Ingrid Arteaga

  10. Department of Endocrinology & Nutrition, Hospital Clinic Barcelona, Instituto de Investigaciones Biomédicas August Pi I Sunyer (IDIBAPS), Barcelona, Spain

    Emilio Ortega

  11. Centro de Investigación Biomédica en Red de Fisiopatología de la Obesidad y Nutrición (CIBEROBN), Instituto de Salud Carlos III (ISCIII), Madrid, Spain

    Emilio Ortega

  12. DAP-Cat Group, Unitat de Suport a la Recerca de Barcelona, Institut Universitari d’Investigació en Atenció Primària Jordi Gol (IDIAP Jordi Gol), Barcelona, Spain

    Josep Franch-Nadal

  13. Faculty of Medicine, University of Vic/Central University of Catalonia (UVIC/UCC), Vic, Spain

    Didac Mauricio

Authors
  1. Núria Amigó
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Esmeralda Castelblanco
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Josep Julve
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Neus Martínez-Micaelo
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Núria Alonso
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Marta Hernández
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Josep Ribalta
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Montse Guardiola
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Pere Torán-Monserrat
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Victor Lopez-Lifante
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Cecilia Herrero-Alonso
    View author publications

    You can also search for this author inPubMed Google Scholar

  12. Ingrid Arteaga
    View author publications

    You can also search for this author inPubMed Google Scholar

  13. Emilio Ortega
    View author publications

    You can also search for this author inPubMed Google Scholar

  14. Josep Franch-Nadal
    View author publications

    You can also search for this author inPubMed Google Scholar

  15. Didac Mauricio
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

NAm, EC, JF-N., and DM participated and designed the study. NAm, EC, JF-N, and DM conceptualized the study. All authors contributed to researched data; NAm, EC, NM-M, JJ and DM analyzed and interpreted data. All authors contributed to the discussion and reviewed the manuscript; NAm, EC, JJ, wrote the original draft of the manuscript; NAm, EC, JJ, MH, NA, JF-N., and DM reviewed/edited the manuscript; DM supervised the study. All authors read and approved the manuscript.

Corresponding authors

Correspondence to Josep Franch-Nadal or Didac Mauricio.

Ethics declarations

Ethics approval and consent to participate

The study protocol for the LIPOCAT Study was approved by the Ethics Committees of University Hospital Germans Trias i Pujol (PI-18–039). The associated cohorts were previously approved by the Ethics Committees of University Hospital Arnau de Vilanova (CEIC-1571, Diabetes Mellitus study), Ethics and Clinical Investigation Committee of IDIAP Jordi Gol (P16/014, ARTPER; P11/58, FIBROSCAN), Hospital Clinic of Barcelona (Clinic cohort ClinicalTrials.gov Identifier: NCT01898572, DIABIMCAP Study), and Carlos Haya Hospital (Di@bet.es), and all participants were duly informed about the study protocols and provided their consent to participate.

Consent for publication

Not applicable.

Competing interests

N. Amigó is a stock owner of Biosfer Teslab and has a patent of the lipoprotein profiling described in the present manuscript. Prof. Mauricio is a co-author of this study and an Editorial Board member of the Cardiovascular Diabetology journal. He did not participate in handling this manuscript during the submission or review processes. The rest of the authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Amigó, N., Castelblanco, E., Julve, J. et al. Advanced serum lipoprotein and glycoprotein profiling for cardiovascular event prediction in type 2 diabetes mellitus: the LIPOCAT study. Cardiovasc Diabetol 24, 88 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12933-025-02636-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12933-025-02636-5

Keywords

Cardiovascular Diabetology

ISSN: 1475-2840