Systemic lipid abnormalities and choriocapillaris flow deficits in patients with hard drusen Authors: Xiangfei Xianyu, Huan Xia, Gang Dong Published: 2026 ABSTRACT AbstractThe aim of this study was to explore alterations in serum biomarkers and ocular fundus blood flow in patients with hard drusen, and to examine the associations between these parameters. In this cross-sectional study, 48 patients with clinically OCTA-identified drusen and 51 healthy controls were recruited. All participants underwent serum biochemical testing, including prealbumin (PA), glucose (GLU), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDLC), and low-density lipoprotein cholesterol (LDLC), among others. Superficial and deep vascular density (SVD and DVD), as well as choriocapillaris flow deficits (CCFD), were assessed in all subjects using optical coherence tomography angiography (OCTA). Compared with controls, the drusen group exhibited significantly higher levels of GLU, TC, and LDLC, as well as greater CCFD density and mean size (all p < 0.05) In the drusen group, CCFD density was significantly positively correlated with LIMT, RIMT, and LDLC, while CCFD size was significantly positively correlated with RIMT, TC, and LDLC (all p < 0.05); additionally, TG showed negative correlations with DVDt, DVDs, DVDn, and DVDi, and PA was negatively correlated with DVDt and DVDi (all p < 0.05). This study demonstrated that patients with hard drusen had significantly higher plasma LDLC levels, accompanied by increased CCFD density and mean size, with LDLC levels positively correlated with CCFD severity. These findings provide further evidence linking systemic lipid metabolism abnormalities with choriocapillaris microvascular impairment in the pathogenesis of drusen.Keywords: Drusen, Low-density lipoprotein cholesterol, Choriocapillaris flow deficit, Optical coherence tomography angiography, CholesterolSubject terms: Biomarkers, Diseases, Endocrinology, Medical research Introduction Drusen are hallmark clinicopathological features of age-related macular degeneration (AMD), characterized by extracellular deposits located between the retinal pigment epithelium (RPE) and Bruch’s membrane, and are considered one of the major criteria for AMD staging1. Based on diameter, deposits smaller than 63 μm are defined as hard drusen, while those larger than 63 μm are classified as soft drusen2. Although compositional differences exist between the two types, both are primarily composed of lipoprotein-derived lipids and others, suggesting that lipid metabolism plays a crucial role in drusen formation3,4. Longitudinal studies have shown that small hard drusen observed at baseline may progressively increase in number or progress to larger soft drusen and ultimately to AMD5,6, though the precise trajectory of this conversion within any individual eye remains incompletely characterized. These findings indicate that small hard macular drusen may represent an early and potentially transitional stage in AMD pathogenesis2. AMD is one of the leading causes of visual decline and blindness in the elderly7. Exploring the pathophysiological changes underlying the formation and progression of early drusen may provide new insights into the understanding, prevention, and treatment of AMD. Optical coherence tomography angiography (OCTA), a noninvasive and high-resolution imaging modality, enables rapid evaluation of retinal and choroidal blood flow and thickness, and has been widely applied in AMD research. Several studies have demonstrated that choriocapillaris flow deficits (CCFD) are frequently observed beneath and around drusen8 and that the area of CCFD significantly increases with AMD severity9,10. Even in patients with stable-stage AMD—defined as intermediate AMD eyes that did not progress to late AMD over a 12-month follow-up period—CCFD has been reported to enlarge progressively11, and such choriocapillaris damage may be closely associated with the development of geographic atrophy (GA) in late-stage disease12. More recent OCTA findings also indicate that CCFD is not only correlated with the size and morphology of drusen but may also serve as a predictor of their progression and enlargement13,14. Most previous investigations have focused on serum lipid alterations in individuals with soft drusen, whereas relatively few studies have examined the metabolic characteristics associated with hard drusen. Moreover, few studies have simultaneously addressed both lipid metabolism and fundus blood flow changes in patients with drusen. Therefore, this study aims to investigate hard drusen by comprehensively analyzing serum biochemical parameters—particularly lipid metabolism profiles—together with fundus vascular changes, in order to explore the potential interactions between systemic metabolic abnormalities and local microvascular impairment. Subjects and methods Subjects We recruited 48 patients with one or more hard drusen (< 63 μm) identified on fundus OCTA B-scan and 51 healthy controls at Daye People’s hospital of Hubei Province. Exclusion criteria included: (1) systemic diseases potentially affecting vascular structure (e.g., diabetes, cardiovascular disease), as confirmed by accredited health examination centers reports obtained within the preceding six months; (2) any pre-existing retinal disease ; (3) any history of panretinal photocoagulation; and (4) poor-quality OCTA (image quality score < 7). All included participants presented exclusively with hard drusen (< 63 μm) without any additional AMD features, including pigmentary abnormalities, soft drusen, geographic atrophy, or choroidal neovascularization, in either eye. OCTA images were independently reviewed by two clinical ophthalmologists, and consensus was reached for all cases. Written informed consent was obtained from all participants. The study adhered to the guidelines of the Ethics Committee of Daye People’s hospital of Hubei Province and the tenets of the Declaration of Helsinki. Baseline information was collected from all enrolled participants using a standardized questionnaire, including age, sex, body mass index (BMI), hypertension status (yes/no), physical activity level (inactive/partially meeting/meeting WHO 2020 recommendations of ≥ 150 min/week moderate-intensity or ≥ 75 min/week vigorous-intensity aerobic activity), and annual per capita household income (category 1: <2,800 USD; category 2: 2,800 USD–5,570 USD; category 3: ≥5,570 USD). All ophthalmic examinations were performed by the same experienced ophthalmologist; the diagnosis of hard drusen was confirmed by consensus between two experienced clinicians; and all questionnaire data were collected by the same clinician. Serum All participants underwent serum biochemical testing, with key indicators including total protein (TP), albumin (Alb), globulin (Glb), prealbumin (PA), urea, creatinine (Cr), glucose (GLU), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDLC), low-density lipoprotein cholesterol (LDLC), apolipoprotein A1 (apo A1), and apolipoprotein B (apo B). LDLC cholesterol was measured directly using a homogeneous enzymatic colorimetric assay. In addition, the intima–media thickness of the left internal carotid artery (LIMT) and right internal carotid artery (RIMT) was measured in all subjects. OCTA All participants underwent OCTA imaging (RTVue XR Avanti; Optovue Inc., Fremont, CA, USA). Swept- Domain OCT angiography (SD-OCTA) was performed with a 3 × 3 mm macular scan centered on the fovea. The built-in software automatically segmented the scan into a central foveal (f) region (0–1 mm ring) and a parafoveal region (1–3 mm ring), which was further divided into superior (s), temporal (t), inferior (i), and nasal (n) quadrants. The software then automatically generated superficial vascular density (SVD), deep vascular density (DVD), and retinal thickness (RT) (from the internal limiting membrane to the RPE) for these five regions. For patients with unilateral drusen, OCTA data from the affected eye were analyzed. For those with bilateral drusen(with image quality score ≥ 7), the eye with better image quality was selected; if both eyes had comparable quality, the right eye was included. CCFD En-face images of the choriocapillaris (CC) obtained from OCTA were first preprocessed using standard image correction techniques to minimize motion artifacts and projection effects. The images were subsequently registered and averaged to generate standardized en-face choriocapillaris slabs for analysis. To reduce the influence of superficial retinal vessel shadows, a binary mask of large retinal vessels was generated using a fixed-threshold binarization algorithm and overlaid on the CC images, thereby excluding shadowed regions from subsequent quantification. After preprocessing, the en-face CC images were binarized and segmented using a fuzzy C-means clustering algorithm to identify regions of flow deficit (FD). Flow deficits were defined as areas of reduced flow signal below a fixed threshold within the CC slab, and FDs with an equivalent diameter < 24 μm were excluded from analysis to reduce noise. Two quantitative indices were then calculated: (1) mean FD size, defined as the average area (µm²) of all detected FDs within the analyzed region, and (2) FD density, defined as the percentage of the CC image area occupied by FDs relative to the total area of the analyzed region. The detailed calculation procedures and validation of these methods have been described previously15. Statistical analysis Continuous variables were expressed as mean ± standard deviation (SD) or as median (interquartile range, IQR), whereas categorical and ordinal variables were presented as frequencies (n, %). The normality of continuous variables was assessed using the Shapiro–Wilk test. For normally distributed variables, between-group comparisons were performed using independent-samples t-tests; otherwise, the Wilcoxon rank-sum test was applied. Categorical variables were analyzed using the Pearson χ² test. To control for potential confounding effects of age and sex, analysis of covariance (ANCOVA) was further conducted for continuous variables, with group (drusen vs. control) as the main independent variable and age and sex as covariates. Within-group correlations between hematological parameters and OCTA metrics were assessed using Spearman’s rank correlation coefficients. Correlation matrices were visualized as heatmaps, in which coefficient strength was represented by a color gradient, and significant correlations (p < 0.05) were marked with asterisks. To further compare the strength of correlations between groups, Fisher’s r-to-z transformation was applied to correlation coefficients of the same variable pairs, and the between-group differences were tested for significance. A differential correlation heatmap was generated, with Δr (r_{drusen} – r_{control}) representing the intergroup difference; significant differences were indicated with asterisks. All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant. Data analysis was performed using R software (R Foundation for Statistical Computing, Vienna, Austria; 4.2.3). Results In the drusen group, 48 participants (drusen group, 39 females, 9 males) had drusen identified on OCTA images, while in the control group, 51 participants (control group, 48 females, 3 males) had no drusen. Baseline demographic characteristics of the two groups are summarized in Table 1. Significant differences were observed in age, sex, and BMI (all p < 0.05), with the drusen group being older, having a significantly higher proportion of males, and showing a significantly higher BMI compared with controls. Table 1Baseline characteristics of participants in the two groups.VariableDrusen groupControl groupp_valueSex(female/man)39/948/30.050*Hypertension status (yes)23 (47.9%)23 (45.1%)0.779Income(1/2/3)18 (37.5%)/ 17 (35.4%)/13 (27.1%)21 (41.2%)/26 (51.0%)/4 (7.8%)0.033*Activity level(inactive/partially meeting/meeting )0 (0.0%)/22 (45.8%)/26 (54.2%)1 (2.0%)/16 (31.4%)/34 (66.7%)0.182Age(year)63.00 [59.75, 66.00]59.00 [55.00, 62.50]0.001*BMI24.19 ± 2.1923.23 ± 2.450.043**statistically significant results (p < 0.05).BMI: body mass index. Baseline characteristics of participants in the two groups. *statistically significant results (p < 0.05). BMI: body mass index. Comparisons of serum biochemical parameters and OCTA metrics between groups are presented in Table 2. The drusen group exhibited significantly higher levels of GLU, TC, and LDLC, as well as greater CCFD density and size (all p < 0.05). Table 2Comparison of serum biomarkers and OCT parameters between the two groups.VariableAdjusted mean ± SEp_valueDrusen groupControl groupLIMT0.10 ± 0.000.10 ± 0.000.647RIMT0.10 ± 0.000.10 ± 0.000.866TP70.55 ± 0.9469.29 ± 1.000.246Alb43.67 ± 0.5942.93 ± 0.620.277Glb26.86 ± 0.5826.39 ± 0.620.484PA245.12 ± 7.26241.15 ± 7.720.636Urea5.24 ± 0.235.16 ± 0.250.767Cr66.61 ± 2.2366.50 ± 2.370.965GLU4.85 ± 0.164.38 ± 0.170.012*TC5.13 ± 0.194.70 ± 0.200.050*TG1.44 ± 0.151.70 ± 0.150.125HDLC1.37 ± 0.061.38 ± 0.070.905LDLC3.07 ± 0.122.66 ± 0.120.003*apoA11.37 ± 0.041.38 ± 0.040.951apoB1.06 ± 0.050.96 ± 0.050.067CCFDdensity32.20 ± 0.6929.75 ± 0.730.003*CCFDsize14.22 ± 0.5512.57 ± 0.590.011 *SVDf15.52 ± 1.0213.72 ± 1.090.129SVDt45.45 ± 1.1045.03 ± 1.170.741SVDs47.92 ± 1.0947.64 ± 1.160.824SVDn46.09 ± 1.0745.86 ± 1.140.855SVDi47.74 ± 1.0847.81 ± 1.150.959DVDf27.98 ± 1.3826.30 ± 1.460.290DVDt49.23 ± 0.5849.59 ± 0.620.583DVDs47.91 ± 0.6248.26 ± 0.660.622DVDn49.06 ± 1.0649.19 ± 1.130.917DVDi48.37 ± 0.8548.10 ± 0.910.786RTf246.19 ± 3.14241.96 ± 3.340.245RTt313.41 ± 2.37314.63 ± 2.520.657RTs325.82 ± 2.49326.91 ± 2.650.704RTn324.20 ± 2.67325.09 ± 2.850.776RTi321.61 ± 2.51322.86 ± 2.670.666*statistically significant results (p < 0.05).Abbreviations: LIMT, left internal carotid artery intima-media thickness; RIMT, right internal carotid artery intima-media thickness; TP, total protein; Alb, albumin; Glb, globulin; PA, prealbumin; Cr, creatinine; GLU, glucose; TC, total cholesterol; TG, triglycerides; HDLC, high-density lipoprotein cholesterol; LDLC, low-density lipoprotein cholesterol; apoA1, apolipoprotein A1; apoB, apolipoprotein B; CCFD, choriocapillaris flow deficit; SVD, superficial vascular density; DVD, deep vascular density; RT, retinal thickness; f, foveal; t, temporal; s, superior; n, nasal; i, inferior; SE, standard error. Comparison of serum biomarkers and OCT parameters between the two groups. *statistically significant results (p < 0.05). Abbreviations: LIMT, left internal carotid artery intima-media thickness; RIMT, right internal carotid artery intima-media thickness; TP, total protein; Alb, albumin; Glb, globulin; PA, prealbumin; Cr, creatinine; GLU, glucose; TC, total cholesterol; TG, triglycerides; HDLC, high-density lipoprotein cholesterol; LDLC, low-density lipoprotein cholesterol; apoA1, apolipoprotein A1; apoB, apolipoprotein B; CCFD, choriocapillaris flow deficit; SVD, superficial vascular density; DVD, deep vascular density; RT, retinal thickness; f, foveal; t, temporal; s, superior; n, nasal; i, inferior; SE, standard error. Figure 1A illustrates correlations between OCTA metrics and serum parameters in the drusen group. CCFD density was significantly positively correlated with LIMT, RIMT, and LDLC, while CCFD size was significantly positively correlated with RIMT, TC, and LDLC (all p <0.05). TG was significantly negatively correlated with DVDt, DVDs, DVDn, and DVDi (all p <0.05). PA showed significant negative correlations with DVDt and DVDi (all p <0.05). Heatmaps of Spearman correlation coefficients between hematological parameters (x-axis) and OCTA-derived metrics (y-axis). (A) Correlations in the drusen group (Group 1). (B) Correlations in the healthy control group (Group 2). (C) Differential correlation heatmap showing intergroup differences (Δr = r_Group2 − r_Group1). In panels A and B, colors indicate Spearman’s correlation coefficients (r): red denotes positive correlations, blue denotes negative correlations, and color intensity reflects correlation strength. In panel C, red denotes stronger positive correlations in Group 2 and blue denotes stronger positive correlations in Group 1. Asterisks (*) denote statistically significant correlations or intergroup differences (p < 0.05). Intergroup differences in panel C were tested using Fisher’s r-to-z transformation. Figure 1B shows correlations in the control group. CCFD size was significantly negatively correlated with RIMT, PA, and urea (all p < 0.05). SVDf was significantly positively correlated with TP, Glb, and GLU (all p < 0.05). PA was significantly positively correlated with SVDt and SVDn (all p < 0.05). TP was significantly positively correlated with DVDf (p < 0.05). Urea was positively correlated with RTf, and apoA1 was positively correlated with RTt and RTi (all p < 0.05). Comparative analysis of correlation patterns between groups revealed opposite directions for all significant variable pairs (all p < 0.05) (Fig. 1C). Specifically, RIMT was positively correlated with both CCFD density and CCFD size in the drusen group but negatively correlated in controls. PA was positively correlated with CCFD density and CCFD size in the drusen group, whereas negative correlations were observed in the control group. PA showed negative correlations with SVDt, SVDi, SVDn, DVDt, DVDs, and DVDi in the drusen group, but positive correlations in controls. Alb was negatively correlated with RTi in the drusen group, while positively correlated in controls. Similarly, apoA1 was negatively correlated with RTt and RTi in the drusen group but positively correlated in the control group. Discussion Through a comprehensive analysis of serum biochemical parameters and OCTA parameters in patients with hard drusen, we found that plasma LDLC levels were significantly higher compared with healthy controls, and both the density and size of CCFD were significantly increased in the hard drusen group. Moreover, elevated LDLC levels were positively correlated with increased CCFD density and size in the drusen group. These findings suggest that systemic lipid metabolism abnormalities and choriocapillaris microvascular impairment coexist in patients with hard drusen and may interact with each other in the early stages of AMD. This study found that plasma LDLC levels were significantly higher in patients with hard drusen compared with healthy controls. LDLC is widely used in clinical practice as a surrogate marker of LDL burden16. Our finding of elevated LDLC is consistent with prior literature reporting that LDL levels are generally elevated in AMD patients17,18, and correlation analyses have indicated that increased LDL is significantly associated with AMD risk19. LDL is the major cholesterol transporter and contributes to maintaining cell membrane integrity and photoreceptor function in the retina, thereby being essential for local lipid metabolic homeostasis20. Given the retina’s high metabolic rate, elevated oxygen tension, abundance of polyunsaturated fatty acids, presence of photosensitive pigments, and chronic exposure to light, it is highly susceptible to oxidative stress21,22. In this pro-oxidative microenvironment, LDL is prone to oxidative modification, leading to the formation of oxidized LDL (oxLDL)23. Although oxLDL has not been directly identified in human drusen or Bruch’s membrane specimens, it has been extensively used in AMD-relevant model systems to investigate RPE dysfunction and has been proposed as a plausible contributor to subretinal lipid accumulation21,24; its potential subretinal deposition may partially explain the elevated LDLC levels observed in patients with hard drusen in the present study. Our data showing a positive association between systemic LDL abnormalities and drusen formation can be related to literature demonstrating that oxLDL is a known proinflammatory inducer of oxidative stress21. Oxidative stress and inflammation are widely recognized as key mechanisms in the pathophysiology of AMD (Handa, 2017). Therefore, oxLDL deposition may represent the pathological link between systemic LDL abnormalities and the progression of drusen to AMD4. The Increased plasma OX-LDL levels observed in AMD patients may provide further support for this hypothesis25, though direct confirmation in human ocular tissue is currently lacking. This study demonstrated that both the density and size of CCFD were significantly higher in patients with hard drusen compared with controls. These findings are consistent with previous reports, which described small drusen as corresponding to focal “black hole”-like perfusion voids within the CC layer13 and showed that reduced CC perfusion correlates with the extent of drusen area8. Beyond focal deficits, study has reported that choriocapillaris impairment may extend beyond regions directly underlying drusen to cover adjacent macular areas, suggesting a more diffuse disease process11. Primary CC vascular injury—potentially triggered by inflammation, degenerative mechanisms, or other genetic and non-genetic factors—may lead to RPE ischemia and dysfunction26, thereby impairing its ability to process exogenous lipids and metabolic waste products and promoting drusen deposition. Once CC damage accumulates beyond a certain threshold, drusen tend to form above choroidal regions with sparse vascular lumina10, with endothelial cell loss potentially playing a facilitating role in this process27. The observation that areas of significant CC perfusion impairment often precede drusen enlargement or formation in regions with structurally intact RPE further supports this hypothesis14. However, current evidence remains insufficient to establish CC damage as the primary initiating event of the disease. The photoreceptors, RPE, Bruch’s membrane, and CC are closely interconnected and interact bidirectionally28,29. The continuous accumulation of drusen may impair RPE function by creating an additional metabolic exchange barrier between the RPE and CC10. Moreover, the CC is highly dependent on RPE-derived vascular endothelial growth factor (VEGF); when the overlying RPE becomes dysfunctional, nutritional and survival signals may be attenuated, thereby inducing or aggravating endothelial cell loss and further exacerbating CC deficits30. Accordingly, the exact association between CC impairment and drusen formation requires further investigation. In patients with drusen, elevated LDLC levels were positively correlated with increased CCFD, suggesting a potential direct link between systemic lipid metabolism abnormalities and choriocapillaris damage. In hypercholesterolemic rabbits, lipid deposition has been observed in the suprachoroidal tissue, leading to compression of the CC against Bruch’s membrane and partial collapse of capillary lumina, accompanied by extravasation of red blood cells within the vascular wall, indicative of barrier disruption31. Notably, even after lipid-lowering therapy, these ultrastructural vascular alterations remained largely irreversible, underscoring the potential for permanent vascular injury in a hyperlipidemic environment31, which is in line with our finding that elevated LDLC was associated with persistent CCFD in patients with hard drusen. Furthermore, LDL is highly susceptible to oxidative modification, generating oxLDL that induces reactive oxygen species (ROS) production in endothelial cells (ECs) and smooth muscle cells (SMCs)32. This cascade promotes oxidative stress and endothelial dysfunction, thereby exacerbating microvascular injury23, which may manifest clinically as increased CCFD. An important local mechanism involves RPE-derived apoB- and apoE-containing lipoproteins that accumulate within Bruch’s membrane, driving a pro-inflammatory, atherosclerosis-like remodeling process that impairs choriocapillaris integrity20,33. This RPE-derived pathway operates largely independently of circulating LDL levels, yet systemic hypercholesterolemia may amplify the local lipid burden, potentially accelerating this process. Among the parameters showing significant correlation differences between the two groups, RIMT was positively correlated with both CCFD density and size, and LIMT was positively correlated with CCFD density in the drusen group. As the ophthalmic artery is the first major branch of the internal carotid artery, carotid stenosis and hypoperfusion can impair retinal and choroidal microcirculation34. Previous studies have reported elevated TG levels in AMD patients17,18. In our study, TG levels were negatively correlated with parafoveal DVD in the drusen group, suggesting that elevated TG may affect AMD development partly through microvascular injury. We also found higher TC levels in the drusen group, with TC positively correlated with CCFD size. Elevated TC has been reported to be directly associated with an increased risk of GA35, and GA patients exhibit higher CCFD%36. These findings suggest that TC may influence drusen formation and even GA progression through CC impairment. In addition, higher blood glucose levels were observed in drusen patients. This may be explained by the fact that hyperglycemia can induce microvascular injury and oxidative stress37, both of which are implicated in drusen formation33. Moreover, PA—an inflammatory marker—was negatively correlated with parafoveal DVDt and DVDi, which may demonstrate the link between systemic inflammation and microvascular impairment. This study has some limitations. First, its cross-sectional design does not allow us to establish a causal relationship between systemic lipid abnormalities, choriocapillaris flow deficits, and drusen formation; longitudinal studies are needed to clarify temporal and causal links. Second, due to methodological constraints, we measured LDLC rather than LDL particle number (LDLP). Although LDLC is widely used in clinical practice, combining LDLC and LDLP measurements may provide a more accurate reflection of the LDL-related lipid burden and underlying pathophysiological mechanisms. Third, the exclusion of flow deficits < 24 μm in diameter was based on OCTA-derived intercapillary distance estimates; however, histologic data suggest that true macular intercapillary spacing may be smaller, potentially leading to underestimation of choriocapillaris flow deficit burden. Future studies employing histologically validated thresholds are warranted to improve the accuracy of choriocapillaris flow deficit quantification. Conclusions In conclusion, our study found elevated plasma LDLC levels, along with increased CCFD density and size, and demonstrated a positive correlation between LDLC elevation and CCFD in patients with drusen. These findings suggest that systemic lipid metabolism abnormalities and local choriocapillaris impairment may coexist and interact during the development of drusen. A potential mechanism may involve the oxidative modification of lipids within RPE-derived intraocular lipoprotein particles accumulating in Bruch’s membrane; systemic hypercholesterolemia may amplify this local lipid burden, promoting local inflammation and choriocapillaris injury, and CC impairment may further disturb lipid metabolism and promote drusen formation and enlargement in turn. These results provide new evidence for the pathological interplay between systemic dyslipidemia and local choriocapillaris injury, which may underlie the formation and progression of drusen, and potentially contribute to AMD development. Acknowledgements The authors thank all the patients who participated in this study. Abbreviations Age-related macular degeneration Optical coherence tomography angiography Choriocapillaris flow deficit Choriocapillaris Retinal pigment epithelium Superficial vascular density Deep vascular density Retinal thickness Flow deficit Geographic atrophy Choroidal neovascularization Spectral-domain optical coherence tomography angiography Low-density lipoprotein cholesterol High-density lipoprotein cholesterol Total cholesterol Triglycerides Glucose Prealbumin Total protein Albumin Globulin Creatinine Apolipoprotein A1 Apolipoprotein B Left internal carotid artery intima-media thickness Right internal carotid artery intima-media thickness Body mass index Interquartile range Standard deviation Analysis of covariance Oxidized low-density lipoprotein Reactive oxygen species Endothelial cells Smooth muscle cells Vascular endothelial growth factor Bruch’s membrane Apolipoprotein E Author contributions Xiangfei Xianyu: Conceptualization, Investigation, Writing - Original Draft. Huan Xia: Formal analysis, Data Curation. Gang Dong: Conceptualization, Resources, Writing - Review & Editing. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Data availability The data that support the findings of this study are available from the corresponding author, GD, upon reasonable request. Declarations Competing interests The authors declare no competing interests. Ethics approval and consent to participate Ethics approval for this study was obtained from the “Clinical Research Ethics Committee, Daye People’s hospital of Hubei Province.” Consent for publication The publication of all data in this study was approved by the subjects. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Xiangfei Xianyu and Huan Xia contributed equally to this work. References Associated Data Data Availability Statement The data that support the findings of this study are available from the corresponding author, GD, upon reasonable request.