Abstract
Background
Despite optimal low-density lipoprotein cholesterol (LDL-C) management, residual vascular risk remains high in patients with type 2 diabetes (T2D). Remnant cholesterol (RC) is an atherogenic lipid fraction, yet its longitudinal association with diabetic retinopathy (DR) remains insufficiently characterised. We aimed to evaluate whether longitudinal RC measurements independently predict new-onset DR.
Materials and Methods
We recruited 1,580 Chinese adults with T2D and no DR from a community-based prospective cohort established between 2017 and 2019. Participants were followed annually until 2023. Association between RC and incident DR was evaluated using time-dependent Cox regression models. We specifically examined this association among participants who achieved clinical targets for conventional lipid panel.
Results
We observed 304 incident DR cases during five follow-up visits. Each one-unit increase in natural log-transformed RC was associated with a 26% higher risk of new-onset DR (adjusted hazard ratio [HR] 1.26, 95% confidence interval [CI] 1.08–1.47). Elevated RC (≥0.8 mmol/L) was associated with a 50% increased DR risk (aHR 1.50, 95% CI 1.11–2.03). These associations persisted among participants meeting guideline-recommended targets for triglycerides, LDL-C, and high-density lipoprotein cholesterol. Such associations were more prominent in women and patients with longer duration of T2D (p for interaction <0.05).
Conclusions
Elevated RC is an independent risk factor for new-onset DR and represents an important contributor to residual risk in patients with otherwise optimal lipid profiles. These findings suggest that RC merits consideration for integration into routine metabolic monitoring to improve DR risk stratification, particularly among female patients and those with long-standing diabetes.
Keywords: Remnant cholesterol, diabetic retinopathy, type 2 diabetes, residual risk, longitudinal study
Introduction
The global prevalence of diabetes is projected to nearly triple by 2050, reaching an estimated 1.31 billion individuals [1]. This surge is expected to drive a parallel increase in diabetic retinopathy (DR), a leading cause of vision loss among working-age adults that affects approximately one-third of the diabetic population [2–4]. Without substantial improvements in its prevention and treatment, the prevalence and burden of DR will continue to rise with the age population and the ongoing diabetes epidemic globally. In this context, identifying modifiable risk factors for DR and targeting high-risk populations becomes essential for informing clinical practice and public health recommendations.
Dyslipidaemia is a common metabolic abnormality in diabetes. It has been shown that disorders of lipid metabolism are closely linked to the pathogenesis of insulin resistance and type 2 diabetes (T2D) [5–8]. Beyond its prevalence, dyslipidaemia serves as a key driver in accelerating diabetic vascular complications, including DR [9–11]. Non-conventional lipid markers, such as remnant cholesterol (RC), have gained increasing attention as a valuable supplement to traditional lipid profile for refining risk assessment. RC represents the cholesterol content of triglyceride-rich lipoprotein remnants, including very-low-density lipoprotein, intermediate-density lipoprotein, and chylomicron remnants [12]. These particles are larger and more numerous than low-density lipoprotein particles and carry a greater cholesterol load per particle [13]. RC is increasingly recognised as an independent driver of cardiovascular disease (CVD) and adverse cardiovascular outcomes. The atherogenic potential of these remnants may be attributed to their increased capacity for arterial wall penetration and accumulation, as well as pro-inflammatory properties [13,14]. RC may therefore represent a significant marker of residual vascular risk.
Although the association between RC and macrovascular disease is increasingly recognised, its role in diabetic microvascular complications, particularly DR, remains insufficiently defined. Current evidence is largely based on cross-sectional observations, which are limited by their inability to establish temporal precedence or capture the effects of long-term lipid fluctuations [15–17]. To address the knowledge gap regarding the role of RC in the management of T2D, we aimed to investigate the longitudinal association between RC levels and the risk of new-onset DR in a prospective cohort.
Materials and methods
Study design
Our investigation was based on the Guangzhou Diabetic Eye Study (GDES), a prospective community-based cohort in Guangzhou, one of China’s major metropolitan cities. A detailed protocol was published elsewhere [18]. Between November 2017 and December 2019, the cohort recruited over 2,900 community-dwelling adults aged 30–80 years with clinically diagnosed T2D from primary care settings. Participants were systematically referred through a regional network of community health centres to Zhongshan Ophthalmic Centre, a nationally recognised, tertiary ophthalmic hospital, where standardised examinations including dilated, ophthalmic assessment were conducted annually through December 2023.
Participants
According to Chinese Diabetes Society guidelines, T2D was diagnosed by primary care physicians when at least one of the following criteria was met: fasting plasma glucose ≥ 7.0 mmol/L, 2-h postprandial glucose during an oral glucose tolerance test ≥ 11.1 mmol/L, or glycated haemoglobin (HbA1c) ≥ 6.5% [19]. Individuals with type 1 or gestational diabetes, major systemic conditions, and cognitive impairment were excluded. Full inclusion and exclusion criteria are shown in the study flowchart (Supplementary Figure S1). Of the 2,963 participants enrolled at baseline, 2,165 were free of DR. We excluded participants with incomplete baseline data on DR, RC, or covariates (n = 140), and those who did not attend any subsequent follow-up visit for DR assessment (n = 445). Thus, a total of 1,580 eligible participants were included in the closed-cohort analysis (Supplementary Figure S1).
Assessment of diabetic retinopathy
The primary outcome was the first occurrence of ophthalmologist-diagnosed DR during follow-up. All participants underwent seven-standard field colour stereoscopic fundus photography after pupil dilation using a digital retinal camera (Canon CR-2) at baseline and each follow-up visit. The severity of DR was graded independently by two trained ophthalmologists according to the modified Airlie House Classification and the Early Treatment Diabetic Retinopathy Study (ETDRS) retinopathy severity scheme [20]. Disagreements (<8%) were adjudicated by a senior ophthalmologist. Incident DR was determined according to the grade of the more severely affected eye.
Assessment of remnant cholesterol
Overnight fasting venous blood was collected by a trained nurse at the Zhongshan Ophthalmic Centre following routine clinical procedures, after all participants completed an instructed fast of at least 8 h. Enzymatic colorimetric assays performed on a Cobas 8000 analyzer were used for biochemical analysis to quantify total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), LDL-C, and triglycerides (TG). RC was calculated as TC minus HDL-C minus LDL-C. This practical formula, based on routinely measured lipid parameters, has been widely used in previous epidemiological studies to avoid the cost and technical challenges associated with direct RC measurement [15–17,21–23]. Elevated RC was defined as ≥ 0.8 mmol/L, in accordance with the Chinese expert consensus on non-traditional blood lipid parameters and the recommendations of the European Atherosclerosis Society [24,25].
Assessment of covariates
Face-to-face questionnaires were administered by centrally trained interviewers to collect data on sociodemographic characteristics, lifestyle behaviours, medical history, and medication use. Participants reported whether they had used hypolipidaemic drugs, oral antidiabetic agents, or insulin during the past week. Physical parameters were measured according to a standardised protocol, with participants wearing light clothing and no shoes. Body weight and height were measured using an automated anthropometric scale, and body mass index (BMI) was calculated as weight in kilograms divided by height in metres squared. Waist circumference (WC) was measured at the midpoint between the lowest palpable rib and the top of the iliac crest to the nearest 0.1 cm. Obesity was defined as BMI ≥ 28 kg/m2 with WC ≥ 90.0 cm in men or ≥ 85.0 cm in women [26]. After participants remained seated for at least 10 min, blood pressure (BP) was measured on the arm with the higher reading by a calibrated automated sphygmomanometer. Two readings were obtained 1–2 min apart, and the average was used in subsequent analyses. HbA1c and C‑reactive protein (CRP), a widely used clinical marker of inflammation, were measured by standardised laboratory methods. Hypertriglyceridaemia was defined as TG ≥ 2.3 mmol/L [27].
Statistical analysis
Categorical variables are presented as n (%), and continuous variables as mean (standard deviation [SD]) or median (interquartile range [IQR]; 25th–75th percentiles), according to distribution. Participant characteristics were analysed according to baseline RC groups. Between-group differences were assessed using two-sample t tests, nonparametric Wilcoxon rank sum tests, or chi-square tests, as appropriate. To assess potential attrition bias, we compared the baseline characteristics of participants included in the final analysis with those excluded because of non-attendance at any follow-up visit. Incidence rates of DR were presented per 100 person-years. Owing to a markedly skewed distribution, RC was natural log-transformed to mitigate the influence of outliers and facilitate interpretation (Supplementary Figure S2). Primary analyses were conducted in the overall participants and stratified by sex, given established sex differences in lipid metabolism [28,29]. Dose-response relationships were evaluated using restricted cubic splines (RCS), with three knots placed at the 10th, 50th, and 90th percentiles of the natural log-transformed RC distribution [30].
Time-dependent Cox proportional hazards models were constructed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of RC with incident DR at follow-up. RC and all covariates, except for age, sex, education level, and diabetes duration, were treated as time-varying variables. Model 1 was adjusted for age and sex. Model 2 was further adjusted for education level, diabetes duration, cigarette smoking, alcohol drinking, HbA1c, CRP, obesity, hypertension, hypertriglyceridaemia, and use of antidiabetic and hypolipidaemic drugs. The last observation carried forward method was applied to impute missing values for time-varying covariates, which were minimal (less than 3% for demographic variables and less than 1% for laboratory parameters). The proportional hazards assumption for model fit was tested using the scaled Schoenfeld residuals, and multicollinearity was assessed via variance inflation factors (all <2). To further evaluate the incremental predictive value of elevated RC beyond conventional risk factors, we performed receiver operating characteristic (ROC) analyses and compared the areas under the curve (AUC) between the base model and the model incorporated with RC, both overall and stratified by sex.
In accordance with Chinese diabetes management guidelines [31], participants were categorised according to whether they met recommended targets for key lipid parameters: TG (<1.7 mmol/L), HDL-C (≥1.0 mmol/L in men; and ≥1.3 mmol/L in women), and LDL-C (<2.6 mmol/L). To explore potential effect modifiers, time-dependent Cox regression analyses were conducted within prespecified subgroups defined by age, sex, smoking, drinking, HbA1c, diabetes duration, baseline comorbidities (hypertension, obesity, and hypertriglyceridaemia), and antidiabetic medication use. Statistical interactions were assessed by including a cross-product term between RC and each stratifying variable in the regression model.
The robustness of the main findings was assessed in sensitivity analyses. Participants who had incident DR within the first follow-up visit were excluded to account for possible reverse causality bias. We also restricted the analysis to participants not taking hypolipidaemic drugs to minimise its influence on the association between RC and DR. Analyses were conducted using SAS (version 9.4) and R (version 4.0.2). Results were considered statistically significant when p < 0.05.
Ethics approval and consent to participate
In accordance with the Declaration of Helsinki (2013), the study was approved by the Zhongshan Ophthalmic Centre Medical Ethics Committee, Sun Yat-Sen University (Ref: 2017KYPJ094). All study participants provided written consent at enrolment, and data were anonymised by removing all subject identifiers before data analysis.
Results
Study population and baseline characteristics
The study participants were aged 34 to 80 years, with a mean age of 64.66 ± 7.41 years. The mean follow-up duration, calculated from baseline to the diagnosis of DR or the participant’s last follow-up assessment, whichever occurred first, was 3 years (SD 1.18 years). During follow-up, 304 participants (118 men and 186 women) developed DR. At baseline, elevated RC was present in 20.57% (325/1,580) of participants. Key baseline characteristics were comparable between participants included in the final analysis and those excluded due to non-attendance at any follow-up visit (Supplementary Table S1).
Table 1 compares baseline characteristics of participants according to RC status. Participants with elevated RC were younger (mean age, 63.74 vs. 64.90 years, p = 0.012) and had a lower education level (p = 0.018). Sex distribution, current smoking, and regular alcohol drinking did not differ significantly between the two groups (all p > 0.05). Compared to the normal RC group, the elevated RC group had significantly higher prevalences of dyslipidaemia (98.46% vs. 84.14%, p < 0.001) and hypertriglyceridaemia (94.15% vs. 23.90%, p < 0.001). The prevalence of hypertension and duration of diabetes were comparable between groups. Participants with elevated RC had higher BMI, WC, BP, TC, TG, and HbA1c levels, but lower HDL-C levels than those in the normal RC group (all p < 0.05). No significant differences between groups were observed in LDL-C levels or in the use of antidiabetic or hypolipidaemic drugs.
Table 1.
Baseline characteristics of participants with normal and elevated remnant cholesterol.
| Variables | Normal RC (n = 1255) | Elevated RC (n = 325) | p Value |
|---|---|---|---|
| RC, mmol/L, median (IQR) | 0.36 (0.18, 0.54) | 1.08 (0.91, 1.47) | <0.001 |
| Socio-demographics | |||
| Age, years, mean (SD) | 64.90 (7.39) | 63.74 (7.43) | 0.012 |
| Women, n (%) | 739 (58.88) | 186 (57.23) | 0.590 |
| Education level, n (%) | 0.018 | ||
| Junior secondary school or below | 347 (27.65) | 112 (34.46) | |
| Senior secondary school | 598 (47.65) | 152 (46.77) | |
| College or above | 310 (24.70) | 61 (18.77) | |
| Lifestyle | |||
| Current smoking, n (%) | 149 (11.87) | 41 (12.62) | 0.714 |
| Regular drinking, n (%) | 97 (7.73) | 35 (10.77) | 0.078 |
| Medical history | |||
| Diabetes duration, years, median (IQR) | 7.02 (3.16, 12.93) | 6.55 (3.47, 10.84) | 0.107 |
| Hypertension, n (%) | 686 (54.66) | 166 (51.08) | 0.248 |
| Dyslipidaemia, n (%) | 1056 (84.14) | 320 (98.46) | <0.001 |
| Hypertriglyceridaemia, n (%) | 300 (23.90) | 306 (94.15) | <0.001 |
| Antidiabetic drugs use, n (%) | 300 (23.90) | 84 (25.85) | 0.467 |
| Oral antidiabetic drug, n (%) | 122 (9.72) | 37 (11.38) | 0.374 |
| Insulin, n (%) | 185 (14.74) | 49 (15.08) | 0.879 |
| Hypolipidaemic drugs use, n (%) | 42 (3.35) | 13 (4.00) | 0.567 |
| Clinical parameters | |||
| BMI, kg/m², mean (SD) | 24.40 (3.26) | 25.33 (3.03) | <0.001 |
| WC, cm, mean (SD) | 85.38 (9.40) | 88.28 (8.32) | <0.001 |
| SBP, mmHg, mean (SD) | 132.00 (17.76) | 134.30 (18.52) | 0.038 |
| DBP, mmHg, mean (SD) | 69.50 (10.19) | 71.71 (9.94) | <0.001 |
| HbA1c, %, mean (SD) | 6.75 (1.18) | 7.08 (1.27) | <0.001 |
| TC, mmol/L, mean (SD) | 4.77 (1.05) | 5.27 (1.06) | <0.001 |
| TG, mmol/L, median (IQR) | 1.67 (1.22, 2.29) | 4.09 (3.13, 5.25) | <0.001 |
| LDL-C, mmol/L, mean (SD) | 3.04 (0.97) | 2.96 (0.95) | 0.175 |
| HDL-C, mmol/L, mean (SD) | 1.36 (0.38) | 0.98 (0.23) | <0.001 |
| CRP, mg/L, median (IQR) | 1.42 (0.69, 2.62) | 1.81 (0.96, 3.33) | <0.001 |
Note: Data are presented as mean (SD) for normally distributed continuous variables, median (IQR) for non-normally distributed continuous variables, and n (%) for categorical variables. Elevated RC was defined as ≥ 0.8 mmol/L. Participants with and without elevated RC at baseline were compared using the two‑sample t test or Wilcoxon rank sum test for continuous variables and the chi‑square test for categorical variables, as appropriate. RC, remnant cholesterol; IQR, interquartile range; SD, standard deviation; BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycated haemoglobin; TC, total cholesterol; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; CRP, C-reactive protein.
Associations between RC and incident DR
During follow-up, the overall incidence of new-onset DR was 6.53 per 100 person-years, compared with 7.82 and 6.22 per 100 person-years among participants with and without elevated RC at baseline, respectively (Table 2). In multivariable time-dependent Cox models, each 1-unit increase in ln(RC) was associated with a 26% higher risk of new-onset DR in the total cohort (aHR 1.26, 95% CI 1.08–1.47, p = 0.004) after full adjustment for potential confounders. When analysed categorically, individuals in the elevated RC group had a 50% greater risk of developing DR compared to those in the normal RC group (aHR 1.50, 95% CI 1.11–2.03, p = 0.009). Sex-stratified analyses showed significant associations between RC and DR in women but not in men, when RC was analysed as both a continuous and categorical variable. Among women, each 1-unit increase in ln(RC) was associated with a 37% higher risk of DR (aHR 1.37, 95% CI 1.12–1.69, p = 0.003). In the categorical analysis, women with elevated RC had a 48% higher risk of DR than those with normal RC (aHR 1.48, 95% CI 1.02–2.16, p = 0.040). Sensitivity analyses excluding participants who had new-onset DR during the first year of follow-up or used hypolipidaemic drugs yielded consistent results, supporting the robustness of these associations (Supplementary Tables S2-S3). RCS analyses showed a linear dose-response relationship (p for nonlinearity = 0.220) (Figure 1) with sex-specific patterns (p for interaction = 0.018) (Figure 2). The curves indicated a positive association in women but not in men. In ROC curve analyses, incorporating RC into the base model improved the AUC in all participants (from 0.602 to 0.609, p = 0.043) and in females (from 0.616 to 0.629, p = 0.015) (Supplementary Figure S3).
Table 2.
Incidence rate of diabetic retinopathy at follow-up and association of remnant cholesterol with new-onset diabetic retinopathy.
| Variables | Cases, n | Incidence ratea | Crude model | Model 1 | Model 2 | |||
|---|---|---|---|---|---|---|---|---|
| HR (95%CI) | p Value | aHR (95%CI) | p Value | aHR (95%CI) | p Value | |||
| All participants | 304 | 6.53 | ||||||
| Per one-unit increase | 1.21 (1.07, 1.37) | 0.002 | 1.21 (1.07, 1.37) | 0.003 | 1.26 (1.08, 1.47) | 0.004 | ||
| RC group | ||||||||
| Normal RC | 234 | 6.22 | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 70 | 7.82 | 1.34 (1.06, 1.70) | 0.015 | 1.33 (1.05, 1.69) | 0.018 | 1.50 (1.11, 2.03) | 0.009 |
| Women | 186 | 6.78 | ||||||
| Per one-unit increase | 1.35 (1.15, 1.59) | <0.001 | 1.35 (1.15, 1.60) | <0.001 | 1.37 (1.12, 1.69) | 0.003 | ||
| RC group | ||||||||
| Normal RC | 141 | 6.34 | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 45 | 8.64 | 1.48 (1.10, 2.00) | 0.009 | 1.49 (1.11, 2.01) | 0.009 | 1.48 (1.02, 2.16) | 0.040 |
| Men | 118 | 6.18 | ||||||
| Per one-unit increase | 1.01 (0.83, 1.22) | 0.938 | 1.00 (0.83, 1.21) | 0.992 | 1.09 (0.86, 1.39) | 0.479 | ||
| RC group | ||||||||
| Normal RC | 93 | 6.05 | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 25 | 6.68 | 1.12 (0.75, 1.67) | 0.591 | 1.09 (0.73, 1.63) | 0.678 | 1.57 (0.94, 2.65) | 0.087 |
Note: RC was examined as both a continuous measure (per one-unit increase in natural log-transformed RC) and a binary indicator (elevated RC [≥ 0.8 mmol/L] vs. normal RC [< 0.8 mmol/L]) in relation to incident DR using time-dependent Cox proportional hazards models. The crude model included time-updated RC measurements without covariate adjustment. Model 1 was adjusted for age and sex, both treated as fixed covariates. Model 2 was further adjusted for education level and diabetes duration as fixed covariates, and for cigarette smoking, alcohol drinking, glycated haemoglobin, C-reactive protein, obesity, hypertension, hypertriglyceridaemia, and the use of antidiabetic and hypolipidaemic drugs as time-varying covariates. aper 100 person-years. RC, remnant cholesterol; DR, diabetic retinopathy; HR, hazard ratio; CI, confidence interval; aHR, adjusted hazard ratio.
Figure 1.
Dose-response relationship of remnant cholesterol with incident diabetic retinopathy. Note: The dose-response relationship was examined using restricted cubic splines (RCS), with three knots placed at the 10th, 50th, and 90th percentiles of the natural log‑transformed remnant cholesterol (RC) distribution in the overall cohort and in sex-stratified analyses. Solid lines represent the fitted curves, and shaded areas indicate the 95% confidence interval (CI) bands. The blue histogram shows the distribution of ln(RC). Analyses used fully adjusted time‑dependent Cox proportional hazards models, with age, sex, education level, and diabetes duration treated as fixed covariates, and cigarette smoking, alcohol drinking, glycated haemoglobin, C‑reactive protein, obesity, hypertension, hypertriglyceridaemia, and use of antidiabetic and hypolipidaemic drugs treated as time‑varying covariates. DR, diabetic retinopathy; RC, remnant cholesterol; HR, hazard ratio; CI, confidence interval; aHR, adjusted hazard ratio.
Figure 2.
Associations of remnant cholesterol with incident diabetic retinopathy across patient subgroups. Notes: Analyses used fully adjusted time‑dependent Cox proportional hazards models, with age, sex, education level, and diabetes duration treated as fixed covariates, and cigarette smoking, alcohol drinking, glycated haemoglobin, C‑reactive protein, obesity, hypertension, hypertriglyceridaemia, and use of antidiabetic and hypolipidaemic drugs treated as time‑varying covariates. P values for interaction were calculated by including cross-product terms between remnant cholesterol and the relevant stratification variable in the fully adjusted models. RC, remnant cholesterol; DR, diabetic retinopathy; aHR, adjusted hazard ratio; CI, confidence interval; HbA1c, glycated haemoglobin.
The association between RC and DR was further investigated across categories defined by conventional lipid profiles (Table 3). The positive association between RC and DR persisted among participants whose conventional lipid levels met guideline-recommended targets, including LDL-C < 2.6 mmol/L, HDL-C ≥ 1.0 mmol/L in men or ≥ 1.3 mmol/L in women, or TG < 1.7 mmol/L, suggesting that RC may contribute independently to the development of DR.
Table 3.
Association of remnant cholesterol with incident diabetic retinopathy by levels of conventional lipid parameters.
| Crude model | Model 1 | Model 2 | ||||
|---|---|---|---|---|---|---|
| HR (95%CI) | p Value | aHR (95%CI) | p Value | aHR (95%CI) | p Value | |
| TG < 1.7 mmol/L | ||||||
| Per one-unit increase | 1.40 (1.12, 1.74) | 0.003 | 1.40 (1.12, 1.75) | 0.003 | 1.39 (1.10, 1.74) | 0.006 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 1.91 (1.05, 3.47) | 0.035 | 1.91 (1.04, 3.50) | 0.036 | 2.12 (1.15, 3.93) | 0.017 |
| TG ≥ 1.7 mmol/L | ||||||
| Per one-unit increase | 1.08 (0.88, 1.32) | 0.459 | 1.07 (0.88, 1.30) | 0.492 | 1.11 (0.89, 1.39) | 0.359 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 1.23 (0.91, 1.64) | 0.174 | 1.22 (0.91, 1.63) | 0.190 | 1.38 (0.98, 1.93) | 0.064 |
| LDL-C < 2.6 mmol/L | ||||||
| Per one-unit increase | 1.20 (0.99, 1.47) | 0.067 | 1.21 (0.99, 1.48) | 0.067 | 1.37 (1.04, 1.80) | 0.025 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 1.17 (0.79, 1.73) | 0.426 | 1.19 (0.81, 1.77) | 0.376 | 1.47 (0.87, 2.47) | 0.149 |
| LDL-C ≥ 2.6 mmol/L | ||||||
| Per one-unit increase | 1.22 (1.04, 1.43) | 0.015 | 1.22 (1.04, 1.43) | 0.013 | 1.22 (1.01, 1.47) | 0.038 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 1.46 (1.08, 1.97) | 0.014 | 1.47 (1.09, 1.99) | 0.012 | 1.58 (1.08, 2.31) | 0.018 |
| HDL-C ≥ 1.0/1.3 mmol/L (Men/Women) | ||||||
| Per one-unit increase | 1.59 (1.23, 2.05) | <0.001 | 1.62 (1.25, 2.09) | <0.001 | 1.77 (1.31, 2.41) | <0.001 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 2.07 (1.29, 3.33) | 0.003 | 2.16 (1.34, 3.49) | 0.002 | 2.56 (1.47, 4.47) | <0.001 |
| HDL-C < 1.0/1.3 mmol/L (Men/Women) | ||||||
| Per one-unit increase | 1.06 (0.91, 1.24) | 0.432 | 1.03 (0.88, 1.21) | 0.675 | 1.03 (0.85, 1.25) | 0.792 |
| RC group | ||||||
| Normal RC | 1 (reference) | 1 (reference) | 1 (reference) | |||
| Elevated RC | 1.16 (0.88, 1.53) | 0.295 | 1.11 (0.84, 1.48) | 0.464 | 1.21 (0.84, 1.73) | 0.303 |
Note: Stratified analyses were performed according to the status of conventional lipid parameters, defined as normal or abnormal based on clinical cutoffs. RC was examined as both a continuous measure (per one-unit increase in natural log-transformed RC) and a binary indicator (elevated RC [≥ 0.8 mmol/L] vs. normal RC [<0.8 mmol/L]) in relation to incident DR using time-dependent Cox proportional hazards models. The crude model included time-updated RC measurements without covariate adjustment. Model 1 was adjusted for age and sex, both treated as fixed covariates. Model 2 was further adjusted for education level and diabetes duration as fixed covariates, and for cigarette smoking, alcohol drinking, glycated haemoglobin, C-reactive protein, obesity, hypertension, hypertriglyceridaemia, and the use of antidiabetic and hypolipidaemic drugs as time-varying covariates. RC, remnant cholesterol; DR, diabetic retinopathy; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; HR, hazard ratio; CI, confidence interval; aHR, adjusted hazard ratio.
Subgroup analysis
Subgroup analyses were performed to assess the consistency of the association between RC and new-onset DR across various baseline characteristics. The positive association observed in the primary analysis remained broadly consistent across subgroups. In multivariable-adjusted time-dependent Cox models, interactions between RC and the stratifying variables were evaluated. In addition to the observed effect modification by sex, a statistically significant interaction was identified for diabetes duration (p for interaction = 0.012). Specifically, the association between RC and DR was more pronounced among participants with a longer disease duration (≥ 10 years; aHR 1.54, 95% CI 1.16–2.03) than among those with a shorter duration. No significant effect modification was detected for other stratifying variables, including age, smoking status, drinking status, HbA1c level, comorbidities, or glucose-lowering medication use (all p for interaction > 0.05) (Figure 2).
Discussion
Main findings
In a prospective, community-based cohort of Chinese patients with T2D followed for approximately 3 years, we demonstrated a significant longitudinal association between elevated RC and increased risk of incident DR, particularly among women and those with longer diabetes duration. Notably, this association remained robust among participants with otherwise optimal levels of conventional lipid panel. These findings highlight RC as an independent biomarker for DR risk stratification, particularly in addressing residual microvascular risk.
Relationship with other studies
The wide variation in reported DR incidence largely reflects differences in assessment protocols and cohort characteristics. Our observed DR incidence rate was within the upper range (2–13%) previously documented for Asian populations [32]. This relatively high incidence may reflect our cohort profile, notably its older age and longer diabetes duration, with over one-third of patients having a disease history exceeding 10 years, inherently placing them at increased risk of DR.
In our analysis, HDL-C was significantly and inversely associated with RC, which was consistent with findings from an analysis of a large US population survey [33]. Previous studies conducted among the US population also reported a discordance of LDL-C with RC, which may reflect differences in the metabolism and therapeutic responses of LDL particles and triglyceride-rich remnants [34,35]. In contrast, we did not observe a significant association between LDL-C and RC in our study population. It is worth noting that the discordance does not necessarily imply a significant inverse association; rather, it indicates that RC may remain elevated across a range of LDL-C concentrations, including among individuals with low or well-controlled LDL-C. Variation across studies may be attributable to differences in population characteristics, lipid-lowering treatment, and analytical methods [36]. Our findings therefore add to the existing evidence suggesting that reliance on LDL-C alone may be insufficient to assess residual lipid-related risk in clinical practice.
Large-scale epidemiological studies have established that elevated RC is associated with the incidence and progression of CVD in both the general population and individuals with diabetes [34,37,38]. RC may exert a stronger pro-atherogenic effect than LDL-C [34]. A cohort study based on Copenhagen and UK populations revealed that elevated RC levels were attributable to approximately one-fifth of atherosclerotic CVD events among patients with diabetes [38]. Current evidence regarding the relationship between RC and DR remains limited. A recent cohort study reported an association between RC and sight-threatening DR in patients with type 1 diabetes [39]. In T2D, most available evidence derives from cross-sectional studies. Among these, studies conducted in northern Chinese and Taiwanese populations demonstrated that elevated RC levels were significantly associated with the presence of DR, both before and after adjustment for multiple confounding factors [16,17]. A more recent cross-sectional study involving 6,535 patients with T2D further suggested that RC may serve as a risk factor for DR, with particularly strong associations among those with diabetes duration exceeding 7 years [15]. Similarly, we identified a stronger association among individuals with longer diabetes duration, suggesting that the pathogenic influence of RC on DR may increase as the disease progresses. To the best of our knowledge, only one previous longitudinal study, conducted in a UK population, has reported a significant association between baseline RC and incident DR [40]. Our study is among the first to confirm this relationship while also accounting for time-varying metabolic parameters and lifestyle behaviours. Furthermore, RC maintained positively associated with DR incidence among T2D patients with normal levels of conventional lipid panel. This suggests that RC may represent an independent lipid-related risk factor and that RC merits consideration for integration into routine clinical practice to improve risk stratification in patients with T2D.
Given established sex differences in lipid metabolism [28,29], the role of RC in DR development warrants separate examination in men and women. We observed a notable sex difference in RC-related DR risk, with women exhibiting a significantly higher risk than men. Although evidence on sex-specific associations between RC and DR remains limited, existing evidence indicates that RC appears to be more strongly correlated with metabolic abnormalities in women than in men [41]. Two vascular health studies similarly identified distinct sex-specific patterns of association. One investigation of postmenopausal women demonstrated that elevated RC levels increased the incidence of both coronary artery disease and diabetes [42]. Likewise, an analysis of US adult population identified elevated RC as a risk factor for severe abdominal aortic calcification exclusively in women [43]. As diabetic complications often share underlying pathological mechanisms [44], these findings support the sex-specific association observed in our study. Beyond differences in body morphology and adipose tissue distribution, women experience distinct lipid metabolic changes during pregnancy, lactation, and menopause, resulting in greater cumulative cholesterol exposure than men [45]. Social and healthcare-related factors may also contribute, as differences in healthcare delivery and utilisation between the sexes may partly explain women’s greater excess risk of vascular complications [45,46]. Collectively, these findings underscore the importance of greater attention to RC levels in women, particularly in diabetes management and vascular complication prevention.
The pathophysiological links between RC and incident DR remain incompletely understood. Morphological evidence indicates that high RC levels are associated with wider retinal venules and an elevated vascular fractal dimension [15]. A leading hypothesis centres on the pro-inflammatory effects of RC. Rich in cholesterol and characterised by prolonged circulatory residence, RC exerts potent pro-inflammatory effects that may induce endothelial dysfunction and subsequent retinal vasodilation [15,47,48]. Furthermore, hyperglycaemia may also act synergistically with RC to promote retinal ischaemia and hypoxia, resulting in compensatory vasodilation and shunting. Together, these pathways may accelerate the onset of early-stage DR [49,50]. Further research is warranted to elucidate the complex mechanisms linking RC and the development of DR.
Implications for research and clinical practice
Against a backdrop of increasing DR incidence driven by demographic and epidemiological trends, improving risk prediction and early intervention becomes a public health priority. Conventional risk factors leave substantial residual risk unexplained [51], underscoring the need for improved risk stratification tools. In this context, our findings are the first to demonstrate the potential value of RC as a complementary indicator to conventional lipid parameters for refining DR risk assessment in T2D. Lifestyle modification and pharmacological management among individuals with elevated RC may help mitigate their risk of DR and other diabetes-related complications [52], particularly in high-risk subgroups such as women and those with long diabetes duration. RC can also be estimated from a standard lipid panel, offering a practical advantage for broader use in resource-limited settings without additional costs and facilitating early risk identification in primary care. To translate this potential into practice, large-scale prospective studies and interventional trials are needed to determine whether reducing RC improves clinical outcomes in patients with T2D. Such evidence will be crucial for developing individualised risk management pathways and advancing the precision prevention of diabetic microvascular complications.
Limitations of the study
Our study has several limitations which merit consideration. First, RC was not directly measured by ultracentrifugation or nuclear magnetic resonance spectroscopy due to resource constraints. However, these direct measurement techniques have limited clinical applicability, as they do not capture the full spectrum of plasma RC and involve complex procedures and high costs [47]. RC was instead estimated from conventional lipid profiles using direct LDL-C measurement, a method considered to provide more accurate estimates than the Friedewald equation where the calculation of LDL-C is known to be influenced by TG levels [47,53]. Second, residual confounding by unmeasured factors cannot be ruled out, including environmental exposures, physical activity, diet, healthcare utilisation, and medication-related factors such as specific drug classes, dosages, and medication adherence. Third, the specific reasons for non-attendance at follow-up visits were not systematically recorded; therefore, we could not determine whether loss to follow-up was due to withdrawal, illness, or other reasons. However, participants included in the final analysis and those excluded due to non-attendance were comparable in most baseline characteristics. Finally, as all participants were Chinese adults with T2D, caution is warranted when extrapolating these findings to other populations. Their relevance to different racial or ethnic groups should be examined in future studies.
Conclusion
Elevated RC may contribute independently to the development of DR, including among patients whose conventional lipid levels met guideline-recommended targets. Incorporating RC into routine metabolic monitoring may therefore improve DR risk stratification, particularly among female patients and individuals with long-standing diabetes.
Supplementary Material
Acknowledgements
We wish to acknowledge the local Health Commission (Bureau) for their community liaison support in establishing the Generalist-Specialist Alliance. We also thank all collaborating researchers and frontline clinical staff who contributed to this study.
Funding Statement
This work was supported by the National Natural Science Foundation of China [grant 72574247], the Basic and Applied Basic Research Foundation of Guangdong Province [grant 2026A1515011411] and the Open Research Funds of the State Key Laboratory of Ophthalmology [grants 303060202400377 and 303060202400362]. The study sponsor or funder was not involved in the design of the study, the collection, analysis, and interpretation of data, or the writing of the report, and did not impose any restrictions regarding the publication of the report.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
References
- 1.GBD 2021 Diabetes Collaborators. Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2023;402(10397):203–234. doi: 10.1016/s0140-6736(23)01301-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yau JWY, Rogers SL, Kawasaki R, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care. 2012;35(3):556–564. doi: 10.2337/dc11-1909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.ElSayed NA, Aleppo G, Aroda VR, et al. Retinopathy, neuropathy, and foot care: standards of care in diabetes-2023. Diabetes Care. 2023;46(Suppl 1):S203–S215. doi: 10.2337/dc23-S012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Leasher JL, Bourne RRA, Flaxman SR, et al. Global Estimates on the number of people blind or visually impaired by diabetic retinopathy: a meta-analysis from 1990 to 2010. Diabetes Care. 2016;39(9):1643–1649. doi: 10.2337/dc15-2171. [DOI] [PubMed] [Google Scholar]
- 5.Scharfmann R, Rachdi L, Ravassard P.. Concise review: in search of unlimited sources of functional human pancreatic beta cells. Stem Cells Transl Med. 2013;2(1):61–67. doi: 10.5966/sctm.2012-0120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wilson PWF, Meigs JB, Sullivan L, et al. Prediction of incident diabetes mellitus in middle-aged adults: the framingham offspring study. Arch Intern Med. 2007;167(10):1068–1074. doi: 10.1001/archinte.167.10.1068. [DOI] [PubMed] [Google Scholar]
- 7.Seo MH, Bae JC, Park SE, et al. Association of lipid and lipoprotein profiles with future development of type 2 diabetes in nondiabetic Korean subjects: a 4-year retrospective, longitudinal study. J Clin Endocrinol Metab. 2011;96(12):E2050–4. doi: 10.1210/jc.2011-1857. [DOI] [PubMed] [Google Scholar]
- 8.Kechagia I, Barkas F, Liberopoulos E, et al. Association between simple, combined lipid markers and 20-year cumulative incidence of type 2 diabetes: the ATTICA cohort study (2002-2022). Lipids Health Dis. 2024;23(1):413. doi: 10.1186/s12944-024-02383-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nelson AJ, Rochelau SK, Nicholls SJ.. Managing dyslipidemia in type 2 diabetes. Endocrinol Metab Clin North Am. 2018;47(1):153–173. doi: 10.1016/j.ecl.2017.10.004. [DOI] [PubMed] [Google Scholar]
- 10.Wu L, Parhofer KG.. Diabetic dyslipidemia. Metabolism. 2014;63(12):1469–1479. doi: 10.1016/j.metabol.2014.08.010. [DOI] [PubMed] [Google Scholar]
- 11.Horton WB, Barrett EJ.. Microvascular dysfunction in diabetes mellitus and cardiometabolic disease. Endocr Rev. 2021;42(1):29–55. doi: 10.1210/endrev/bnaa025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chait A, Ginsberg HN, Vaisar T, et al. Remnants of the triglyceride-rich lipoproteins, diabetes, and cardiovascular disease. Diabetes. 2020;69(4):508–516. doi: 10.2337/dbi19-0007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yang XH, Zhang BL, Cheng Y, et al. Association of remnant cholesterol with risk of cardiovascular disease events, stroke, and mortality: a systemic review and meta-analysis. Atherosclerosis. 2023;371:21–31. doi: 10.1016/j.atherosclerosis.2023.03.012. [DOI] [PubMed] [Google Scholar]
- 14.Borén J, Chapman MJ, Krauss RM, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease: pathophysiological, genetic, and therapeutic insights: a consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J. 2020;41(24):2313–2330. doi: 10.1093/eurheartj/ehz962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen S, Xu Y, Chen B, et al. Remnant cholesterol is correlated with retinal vascular morphology and diabetic retinopathy in type 2 diabetes mellitus: a cross-sectional study. Lipids Health Dis. 2024;23(1):75. doi: 10.1186/s12944-024-02064-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Pan W, Han Y, Hu H, et al. The non-linear link between remnant cholesterol and diabetic retinopathy: a cross-sectional study in patients with type 2 diabetic mellitus. BMC Endocr Disord. 2022;22(1):326. doi: 10.1186/s12902-022-01239-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Shan Y, Wang Q, Zhang Y, et al. High remnant cholesterol level is relevant to diabetic retinopathy in type 2 diabetes mellitus. Lipids Health Dis. 2022;21(1):12. doi: 10.1186/s12944-021-01621-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zhang S, Chen Y, Wang L, et al. Design and baseline data of the diabetes registration study: Guangzhou Diabetic Eye Study. Curr Eye Res. 2023;48(6):591–599. doi: 10.1080/02713683.2023.2182745. [DOI] [PubMed] [Google Scholar]
- 19.Weng J, Ji L, Jia W, et al. Standards of care for type 2 diabetes in China. Diabetes Metab Res Rev. 2016;32(5):442–458. doi: 10.1002/dmrr.2827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Early Treatment Diabetic Retinopathy Study Research Group.. Grading diabetic retinopathy from stereoscopic color fundus photographs - an extension of the modified airlie house classification: ETDRS Report Number 10. Ophthalmology. 2020;127(4s):S99–s119. doi: 10.1016/j.ophtha.2020.01.030. [DOI] [PubMed] [Google Scholar]
- 21.Kaltoft M, Langsted A, Nordestgaard BG.. Triglycerides and remnant cholesterol associated with risk of aortic valve stenosis: mendelian randomization in the Copenhagen General Population Study. Eur Heart J. 2020;41(24):2288–2299. doi: 10.1093/eurheartj/ehaa172. [DOI] [PubMed] [Google Scholar]
- 22.Jang SY, Kang M, Song E, et al. Remnant cholesterol is an independent risk factor for the incidence of chronic kidney disease in newly-diagnosed type 2 diabetes: a nationwide population-based study. Diabetes Res Clin Pract. 2024;210:111639. doi: 10.1016/j.diabres.2024.111639. [DOI] [PubMed] [Google Scholar]
- 23.Fu L, Tai S, Sun J, et al. Remnant cholesterol and its visit-to-visit variability predict cardiovascular outcomes in patients with type 2 diabetes: findings from the ACCORD cohort. Diabetes Care. 2022;45(9):2136–2143. doi: 10.2337/dc21-2511. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nordestgaard BG, Langsted A, Mora S, et al. Fasting is not routinely required for determination of a lipid profile: clinical and laboratory implications including flagging at desirable concentration cut-points-a joint consensus statement from the European Atherosclerosis Society and European Federation of Clinical Chemistry and Laboratory Medicine. Eur Heart J. 2016;37(25):1944–1958. doi: 10.1093/eurheartj/ehw152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Specialist Committee on Laboratory Medicine of Chinese Association of Integritive Medicine. Chinese expert consensus on non-traditional blood lipid parameters to control the risks of arteriosclerotic cardiovascular disease. Chin J Prev Med. 2022;56(4):405–421. doi: 10.3760/cma.j.cn112150-20211130-01106. [DOI] [PubMed] [Google Scholar]
- 26.Pan X-F, Wang L, Pan A.. Epidemiology and determinants of obesity in China. Lancet Diabetes Endocrinol. 2021;9(6):373–392. doi: 10.1016/s2213-8587(21)00045-0. [DOI] [PubMed] [Google Scholar]
- 27.Joint Committee on the Chinese Guidelines for Lipid Management. Chinese guidelines for lipid management (2023). Chin J Cardiol. 2023;51(3):221–255. doi: 10.3760/cma.j.cn112148-20230119-00038. [DOI] [PubMed] [Google Scholar]
- 28.Yang G, Schooling CM.. Investigating sex-specific associations of lipid traits with type 2 diabetes, glycemic traits and sex hormones using Mendelian randomization. Cardiovasc Diabetol. 2023;22(1):3. doi: 10.1186/s12933-022-01714-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Pikkemaat M, Woodward M, Af Geijerstam P, et al. Lipids and apolipoproteins and the risk of vascular disease and mortality outcomes in women and men with type 2 diabetes in the ADVANCE study. Diabetes Obes Metab. 2024;26(12):5669–5680. doi: 10.1111/dom.15935. [DOI] [PubMed] [Google Scholar]
- 30.Harrell F. Regression modeling strategies: with applications to linear models, logistic and ordinal regression, and survival analysis. New York: Springer Nature Switzerland AG; 2015. p. 26–28. [Google Scholar]
- 31.Chinese Diabetes Society, Zhu D, Guo L.. Guideline for the prevention and treatment of diabetes mellitus in China (2024 edition). Chinese J Diab Mellitus. 2025;17(1):16–139. doi: 10.3760/cma.j.cn115791-20241203-00705. [DOI] [Google Scholar]
- 32.Sabanayagam C, Banu R, Chee ML, et al. Incidence and progression of diabetic retinopathy: a systematic review. Lancet Diabetes Endocrinol. 2019;7(2):140–149. doi: 10.1016/s2213-8587(18)30128-1. [DOI] [PubMed] [Google Scholar]
- 33.Meng W, Wang C, Xu Z, et al. Association between remnant cholesterol and peripheral artery disease from multiple perspectives. Atherosclerosis. 2026;417:120784. doi: 10.1016/j.atherosclerosis.2026.120784. [DOI] [PubMed] [Google Scholar]
- 34.Quispe R, Martin SS, Michos ED, et al. Remnant cholesterol predicts cardiovascular disease beyond LDL and ApoB: a primary prevention study. Eur Heart J. 2021;42(42):4324–4332. doi: 10.1093/eurheartj/ehab432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wang Y, Li N, Shao X, et al. The discordance of remnant cholesterol and low-density lipoprotein cholesterol as a predictor of diabetes, diabetic microvascular diseases, and cardiovascular disease. Diabetol Metab Syndr. 2026;18(1):65. doi: 10.1186/s13098-025-02039-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Stürzebecher PE, Katzmann JL, Laufs U.. What is ‘remnant cholesterol’? Eur Heart J. 2023;44(16):1446–1448. doi: 10.1093/eurheartj/ehac783. [DOI] [PubMed] [Google Scholar]
- 37.Elías-López D, Wadström BN, Vedel-Krogh S, et al. Impact of remnant cholesterol on cardiovascular risk in diabetes. Curr Diab Rep. 2024;24(12):290–300. doi: 10.1007/s11892-024-01555-1. [DOI] [PubMed] [Google Scholar]
- 38.Wadström BN, Pedersen KM, Wulff AB, et al. One in five atherosclerotic cardiovascular disease events in individuals with diabetes attributed to elevated remnant cholesterol. Diabetes Metab Res Rev. 2024;40(8):e70005. doi: 10.1002/dmrr.70005. [DOI] [PubMed] [Google Scholar]
- 39.Jansson Sigfrids F, Dahlström EH, Forsblom C, et al. Remnant cholesterol predicts progression of diabetic nephropathy and retinopathy in type 1 diabetes. J Intern Med. 2021;290(3):632–645. doi: 10.1111/joim.13298. [DOI] [PubMed] [Google Scholar]
- 40.Liu J, Ren R, Li Y, et al. Relationship between remnant cholesterol and the risk of diabetic microvascular complications in type 2 diabetes: an observational study. Preprint (Version 1) Available at Research Square. 22 May 2025. doi: 10.21203/rs.3.rs-6676665/v1. [DOI]
- 41.Zou Y, Kuang M, Zhong Y, et al. Remnant cholesterol can identify individuals at higher risk of metabolic syndrome in the general population. Sci Rep. 2023;13(1):5957. doi: 10.1038/s41598-023-33276-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Zhang Y, Song K, Bi S, et al. Higher remnant cholesterol increases the risk of coronary heart disease and diabetes in postmenopausal women. Front Endocrinol. 2024;15:1475933. doi: 10.3389/fendo.2024.1475933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Yan L, Hu X, Wu S, et al. Gender-specific correlations between remnant cholesterol and severe abdominal aortic calcification in American adults. Front Endocrinol. 2024;15:1415424. doi: 10.3389/fendo.2024.1415424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Laakso M. Heart in diabetes: a microvascular disease. Diabetes Care. 2011;34 (Suppl 2):S145–S149. doi: 10.2337/dc11-s209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Roeters van Lennep JE, Tokgözoğlu LS, Badimon L, et al. Women, lipids, and atherosclerotic cardiovascular disease: a call to action from the European Atherosclerosis Society. Eur Heart J. 2023;44(39):4157–4173. doi: 10.1093/eurheartj/ehad472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.de Ritter R, de Jong M, Vos RC, et al. Sex differences in the risk of vascular disease associated with diabetes. Biol Sex Differ. 2020;11(1):1. doi: 10.1186/s13293-019-0277-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Sandesara PB, Virani SS, Fazio S, et al. The forgotten lipids: triglycerides, remnant cholesterol, and atherosclerotic cardiovascular disease risk. Endocr Rev. 2019;40(2):537–557. doi: 10.1210/er.2018-00184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Varbo A, Benn M, Tybjærg-Hansen A, et al. Elevated remnant cholesterol causes both low-grade inflammation and ischemic heart disease, whereas elevated low-density lipoprotein cholesterol causes ischemic heart disease without inflammation. Circulation. 2013;128(12):1298–1309. doi: 10.1161/circulationaha.113.003008. [DOI] [PubMed] [Google Scholar]
- 49.Saldívar E, Cabrales P, Tsai AG, et al. Microcirculatory changes during chronic adaptation to hypoxia. Am J Physiol Heart Circ Physiol. 2003;285(5):H2064–71. doi: 10.1152/ajpheart.00349.2003. [DOI] [PubMed] [Google Scholar]
- 50.Yim-Lui Cheung C, Wong TY, Lamoureux EL, et al. C-reactive protein and retinal microvascular caliber in a multiethnic Asian population. Am J Epidemiol. 2010;171(2):206–213. doi: 10.1093/aje/kwp357. [DOI] [PubMed] [Google Scholar]
- 51.Jampol LM, Glassman AR, Sun J.. Evaluation and care of patients with diabetic retinopathy. N Engl J Med. 2020;382(17):1629–1637. doi: 10.1056/NEJMra1909637. [DOI] [PubMed] [Google Scholar]
- 52.Sebastian-Valles F, Montes Muñiz Á, Marazuela M.. Remnant cholesterol: from pathophysiology to clinical implications in type 1 diabetes. Endocrines. 2025;6(3):46. doi: 10.3390/endocrines6030046. [DOI] [Google Scholar]
- 53.Nguyen SV, Nakamura T, Uematsu M, et al. Remnant lipoproteinemia predicts cardiovascular events in patients with type 2 diabetes and chronic kidney disease. J Cardiol. 2017;69(3):529–535. doi: 10.1016/j.jjcc.2016.04.011. [DOI] [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 datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.