Phenotypic Clustering Identifies Heterogeneous Cardiovascular Risk Among Patients with Elevated Lipoprotein(a)

Phenotypic Clustering Identifies Heterogeneous Cardiovascular Risk Among Patients with Elevated Lipoprotein(a) | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Phenotypic Clustering Identifies Heterogeneous Cardiovascular Risk Among Patients with Elevated Lipoprotein(a) Hyung Joon Joo, Soon Joon Hong, Cheol Woong Yu, Seung Yong Shin, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9065291/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 15 You are reading this latest preprint version Abstract Lipoprotein(a) [Lp(a)] is an established cardiovascular risk enhancer, yet fixed concentration thresholds may not fully capture the heterogeneity of cardiovascular risk among individuals with elevated levels. We retrospectively analyzed 17,653 patients with Lp(a) measurements from three tertiary hospitals (2017–2024). After exclusions, 4,320 patients with Lp(a) ≥ 50 mg/dL underwent k-means clustering identified two phenotypic groups based on demographic, comorbidity, laboratory, and medication variables. Cluster validation using elbow, shilhouette, and NbClust consensus methods supported a two-cluster solution. Cluster 1 consisted of older, male-predominant patients with a higher cardiometabolic burden and lower renal function, whereas Cluster 2 included younger, female-predominant patients with fewer comorbidities and relatively treatment-naïve dyslipidemia. Despite similar Lp(a) levels, 3-year major adverse cardiovascular events (MACE) occurred more frequently in Cluster 1 than Cluster 2 (8.9% vs 2.0%, log-rank p < 0.01). In multivariable Cox models, Cluster 1 was associated with higher MACE risk compared with the Lp(a) < 30 mg/dL reference group (HR 1.39, 95% CI, 1.13–1.72), whereas Cluster 2 showed no significant risk difference (HR 1.08 95% CI 0.69–1.68). These findings suggest that phenotypic clustering of high-Lp(a) patients delineates subgroups with distinct cardiovascular risk profiles. Incorporating phenotype-guided risk assessment may refine cardiovascular risk stratification beyond fixed Lp(a) thresholds. Health sciences/Biomarkers Health sciences/Cardiology Health sciences/Diseases Health sciences/Medical research Health sciences/Risk factors Lipoprotein(a) Cardiovascular Disease Cluster Analysis Phenotype Risk Assessment Figures Figure 1 Figure 2 Figure 3 Introduction Lipoprotein(a) [Lp(a)] has emerged as an independent, genetically determined risk factor for atherosclerotic cardiovascular disease (ASCVD). Previous Mendelian-randomization study and large clinical cohort study reported a significant increase in cardiovascular events with the elevated Lp(a) concentration 1 – 3 . However, the incidence of cardiovascular events differs across trials and cohorts, suggesting that co-existing clinical factors may modify Lp(a)-mediated risk. Machine-learning methodologies offer a hypothesis-free approach to disentangle such heterogeneity. In the RED-CARPET study, Zhang et al. first applied k-means clustering to the patients with Lp(a) ≥ 50 mg/dL and identified four phenogroups: a dyslipidemia cluster, a cluster of aged women, a cluster of men with unhealthy lifestyle, and a cluster by anemia, renal insufficiency and hypercoagulability 4 . Although Lp(a) concentrations were comparable, the risk of ASCVD varied markedly; men with unhealthy lifestyle had the highest risk relative a low-Lp(a) reference group, whereas the anemia/renal insufficiency cluster showed no significant excess risk. This finding highlights the importance of multi-dimensional phenotyping over single-analyte risk stratification. However, whether such phenotypic subgroups provide independent prognostic value beyond conventional cardiovascular risk factors remains uncertain, and the reproducibility of these phenotypes across different populations has not been well established. Furthermore, most prior studies have focused on cross-sectional phenotyping rather than evaluating long-term clinical outcomes. Therefore, in the present study we applied unsupervised k-means clustering to a large contemporary, unselected Korean patient with elevated Lp(a) cohort and examined the association between identified phenotypes and 3-year major adverse cardiovascular events (MACE). By integrating phenotypic clustering with time-to-event analysis, this study aims to determine whether clinically meaningful heterogeneity exists within the high-Lp(a) population beyond concentration-based theresholds. Methods Study Design and Data Sources This retrospective observational study was based on de-identified data from the Observational Medical Outcomes Partnership Common Data Model (OMOP-CDM) database and electronic health records (EHRs) of three tertiary academic medical centers in Korea. The OMOP-CDM framework, provided by the Observational Health Data Sciences and Informatics (OHDSI) collaboration, served as the standardized structure for organizing EHRs. Clinical data, including diagnoses based on the International Classification of Diseases, 10th Revision (ICD-10), medications, and laboratory results, were mapped to unique concept identifiers in the OMOP-CDM. All data were stored on Microsoft SQL servers and assessed using direct SQL queries. This study adhered to the principles of the Declaration of Helsinki and was approved by the institutional Review Board (IRB No. 2025AN0429). The requirement for informed consent was waived due to the retrospective nature of the study and use of de-identified data. Study Population The dataset included adult patients (age ≥ 18 years) who underwent serum Lp(a) testing between January 2017 and December 2024. A total of 33,954 patients with Lp(a) measurements were initially screened. Exclusion criteria were history of myocardial infarction (n = 2,556) or stroke (n = 4,834), any cancer (n = 3,537), thyroid disease or hormone therapy (n = 5,399), chronic viral infection (n = 506), extremely abnormal liver injury (n = 19). Patients who experienced a major adverse cardiovascular event (MACE) within the first 30 days of follow-up (n = 752) or were lost to follow-up during this period (n = 1,396) were subsequently excluded. After these exclusions, 17,653 patients remained in the final cohort. For clustering analysis, patients with Lp(a) levels ≥ 50 mg/dL were included (n = 4,320). Serum Lipoprotein(a) Measurement Lp(a) testing was conducted as part of routine clinical practice using the Randox Lp(a) assay (Catalog No. LP3403, Randox Laboratories, Crumlin, United Kingdom) on a Beckman Coulter AU5800 analyzer. Daily internal control assays maintained the precision with intra- and inter-assay coefficients of variation between 2.4% and 3.2%. Lp(a) risk categories were defined based on the EAS consensus statement: high risk (Lp(a) ≥ 50 mg/dL), low risk (Lp(a) < 30 mg/dL), and intermediate risk for values between 30 and 50 mg/dL. Clinical Variables and Outcomes Demographic, socio-economic status, lifestyle behaviors, comorbidities, medication use, and laboratory data were extracted and defined. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Serum creatinine-based estimated glomerular filtration rate (eGFR) was estimated using the CKD-EPI equations, truncated at a maximum of 150 mL/min/1.73m 2 to mitigate physiological overestimation. The primary outcome was the time-to-occurrence of MACE, defined as a composite of cardiovascular death, new-onset myocardial infarction, stroke, heart failure hospitalization or coronary revascularization. New-onset myocardial infarction was defined as hospital admission for chest pain, angina, or dyspnea accompanied by cardiac marker levels exceeding the 99th percentile of the upper reference limit. New-onset stroke was identified by a new diagnosis of stroke or confirming acute, subacute, or recent cerebral infarction on brain MRI. Heart failure hospitalization was defined as a hospital admission following an emergency department visit, accompanied by elevated N-terminal pro B-type natriuretic peptide or brain natriuretic peptide levels, and requiring intravenous furosemide administration 5 , 6 . Time-to-event calculated from 30 days after the date of Lp(a) measurement to event occurrence or last known follow-up, truncated at three years. Clinical outcomes were identified using a combination of diagnostic codes, laboratory measurements, and treatment records within the OMOP-CDM framework. This algorithm-based approach has been widely used in observational EHR studies to improve the accuracy of event ascertainment. When available, laboratory thresholds and treatment criteria were incorporated to increase specificity of outcome definitions and reduce misclassification. Statistical analysis Unsupervised k-means clustering was applied to patients with Lp(a) ≥ 50 mg/dL. Thirty-three clinical and biochemical variables were standardized prior to clustering. The optimal number of clusters was determined via elbow method, average silhouette width, and NbClust consensus. For clustering analysis, patients with missing values among selected clustering variables were excluded. In subsequent analysis, multiple imputations by chained equations were employed using the predictive mean matching method. Categorical variables are expressed as frequencies and percentages, and continuous variables are expressed as mean with standard deviations. Continuous variables were compared across groups using one-way analysis of variance (ANOVA) for normally distributed data and the Kruskal-Wallis test for non-normally distributed data. Post hoc pairwise comparisons were performed with Bonferroni correction or Dunn’s test as appropriate. Categorical variables were compared using the chi-square test or Fisher’s exact test. Time-to-event outcomes were analyzed using the Kaplan-Meier method to estimate cumulative incidence rates over the follow-up period, with censoring applied at the time of death or the last available follow-up. To evaluate clinical outcomes across clusters, multivariable Cox proportional hazards models were constructed. Final model selection was performed through backward selection using Akaike Information Criterion. All analyses were conducted using R version 4.3.1. Key packages included cluster, Nbclust, factoextra, mice, randomForest, and survival. Plots were generated using ggplot2. To enhance the robustness of clustering results, multiple cluster validation approaches were applied, including the elbow method, average silhouette width, and consensus recommendations from the NbClust package. Principal component analysis was additionally used to visually assess cluster separation in reduced dimensional space. These complementary approaches were used to mitigate the risk of overfitting or arbitrary cluster selection and to ensure that the chosen cluster solution reflected stable and interpretable phenotypic patterns within the dataset. Results Cluster validation and identification of two phenotypes Cluster validation indices revealed heterogenous recommendations, with the majority suggesting two clusters as optimal (Supplemental Fig. 1). The elbow plot of within-cluster variance showed an inflection at two clusters, and silhouette analysis demonstrated better cohesion and separation for the two-cluster solution (mean silhouette width, 0.09) compared with three clusters (0.07). Principal component analysis was used to visualize the distribution of cluster assignments for k = 2, k = 3, and k = 4. The four- and three-cluster models showed substantial overlap among clusters, while the two-cluster model demonstrated the clearer separation between groups along the first principal component (13.1% of variance) (Fig. 1 ). These findings support the selection of two clusters as the most interpretable and distinct phenotypic classification for high Lp(a) patients. These two phenotypic clusters were subsequently used for further comparative and outcome analyses. Determinants of cluster differentiation To identify the clinical features most strongly associated with cluster assignment, feature importance was evaluated using a random forest classification model (Fig. 2 A). The top ten features contributing to cluster differentiation included total cholesterol, HDL-cholesterol, LDL-cholesterol, hypertension, diabetes mellitus, sex, dyslipidemia, alcohol consumption, smoking, and glucose. Notably, lipid parameters comprised a substantial proportion of the most influential predictors. A radar plot of standardized mean values for these key features (Fig. 2 B) demonstrated that Cluster 1 was characterized by higher prevalence of hypertension and diabetes mellitus, along with more frequent alcohol consumption and smoking. In contrast, Cluster 2 demonstrated higher levels of total cholesterol, LDL-cholesterol, and HDL-cholesterol, consistent with a drug-naïve dyslipidemia-dominant phenotype. It suggests that traditional cardiovascular risk factors could play a role in distinguishing between these clusters. Baseline clinical characteristics of study population Baseline characteristics were compared among Group 1 and Group 2 (both derived from k-means clustering of patients with high Lp(a) ≥ 50 mg/dL), and Group 3 and Group 4, which were defined according to Lp(a) thresholds of 30–50 mg/dL and < 30 mg/dL, respectively (Table 1 ). Significant differences were observed across the four groups in demographic characteristics, lifestyle factors, comorbidities, laboratory findings, and medication use ( p < 0.01 for most comparisons). Table 1 Baseline characteristics of the study population stratified by Lp(a)-based groups and k-means clustering Variable Group 1 (Cluster 1, Lp(a) ≥ 50) (n = 1,249) Group 2 (Cluster 2, Lp(a) ≥ 50) (n = 1,106) Group 3 (Lp(a) 30–50) (n = 2,029) Group 4 (Lp(a) < 30) (n = 13,269) p- value Age, years 67.3 ± 11.1 59.4 ± 12.9 62.5 ± 13.3 60.4 ± 14.5 < 0.01 Male sex 760 (60.8%) 354 (32.0%) 1,001 (49.3%) 7,434 (56.0%) < 0.01 Current smoking 751 (60.1%) 274 (24.8%) 892 (44.0%) 5,627 (42.4%) < 0.01 Alcohol consumption 804 (64.4%) 312 (28.2%) 926 (45.6%) 6,037 (45.5%) < 0.01 Low socioeconomic status 80 (6.4%) 40 (3.6%) 125 (6.2%) 605 (4.6%) < 0.01 Body mass index, kg/m² 25.3 ± 3.5 23.9 ± 3.4 24.5 ± 3.5 24.8 ± 3.6 < 0.01 Hypertension 1,145 (91.7%) 597 (54.0%) 1,408 (69.4%) 9,124 (68.8%) < 0.01 Diabetes mellitus 775 (62.0%) 192 (17.4%) 752 (37.1%) 5,069 (38.2%) < 0.01 Dyslipidemia 1,213 (97.1%) 790 (71.4%) 1,638 (80.7%) 10,395 (78.3%) < 0.01 Chronic kidney disease 199 (15.9%) 16 (1.4%) 159 (7.8%) 791 (6.0%) < 0.01 Lp(a), mg/dL 89.4 ± 41.5 83.9 ± 34.2 38.4 ± 5.6 11.6 ± 7.2 < 0.01 Hemoglobin, g/dL 13.5 ± 1.6 13.6 ± 1.3 13.7 ± 1.5 13.9 ± 1.5 < 0.01 WBC, ×10³/µL 6.8 ± 1.8 5.9 ± 1.6 6.3 ± 1.8 6.3 ± 1.7 < 0.01 Platelet count, ×10³/µL 228.0 ± 57.2 244.4 ± 56.7 234.4 ± 58.0 234.7 ± 57.9 < 0.01 Creatinine, mg/dL 0.96 ± 0.39 0.76 ± 0.18 0.85 ± 0.30 0.85 ± 0.26 < 0.01 eGFR, mL/min/1.73m² 82.0 ± 23.2 95.5 ± 15.3 90.3 ± 20.0 92.7 ± 19.0 < 0.01 AST, U/L 28.9 ± 15.2 25.7 ± 8.0 27.4 ± 14.9 28.2 ± 17.1 < 0.01 ALT, U/L 27.0 ± 20.7 22.9 ± 13.3 25.2 ± 19.6 27.1 ± 21.5 < 0.01 Total bilirubin, mg/dL 0.72 ± 0.31 0.71 ± 0.31 0.73 ± 0.30 0.76 ± 0.42 < 0.01 Albumin, g/dL 4.3 ± 0.4 4.4 ± 0.3 4.3 ± 0.4 4.4 ± 0.3 < 0.01 Total cholesterol, mg/dL 145.5 ± 30.9 191.9 ± 37.2 168.9 ± 42.1 167.8 ± 42.2 < 0.01 LDL-cholesterol, mg/dL 77.5 ± 25.1 113.4 ± 34.4 96.3 ± 36.0 94.6 ± 35.9 < 0.01 HDL-cholesterol, mg/dL 48.4 ± 11.3 59.1 ± 12.3 53.6 ± 13.3 52.2 ± 12.8 < 0.01 Triglyceride, mg/dL 125.6 ± 58.3 113.1 ± 53.6 116.0 ± 56.4 127.3 ± 66.6 < 0.01 Glucose, mg/dL 115.5 ± 24.4 101.9 ± 15.5 108.6 ± 20.9 109.1 ± 21.2 < 0.01 hs-CRP, mg/L 1.46 ± 1.92 1.09 ± 1.47 1.32 ± 1.70 1.26 ± 1.63 < 0.01 PT-INR 0.99 ± 0.08 0.97 ± 0.06 0.98 ± 0.07 0.98 ± 0.07 < 0.01 aPTT, sec 34.3 ± 4.1 33.6 ± 3.8 34.1 ± 4.1 34.0 ± 3.9 < 0.01 Antiplatelet use 782 (62.6%) 262 (23.7%) 798 (39.3%) 4,456 (33.6%) < 0.01 Antihypertensive use 930 (74.5%) 413 (37.3%) 1,027 (50.6%) 6,774 (51.1%) < 0.01 Antidiabetic use 513 (41.1%) 116 (10.5%) 503 (24.8%) 3,356 (25.3%) < 0.01 Statin use 1,097 (87.8%) 625 (56.5%) 1349 (66.5%) 8,345 (62.9%) < 0.01 Ezetimibe use 585 (46.8%) 302 (27.3%) 617 (30.4%) 3,440 (25.9%) < 0.01 PCSK9 inhibitor use 11 (0.9%) 5 (0.5%) 5 (0.2%) 13 (0.1%) < 0.01 Data are presented as mean ± standard deviation or number (percentage). p -values were calculated using one-way ANOVA or Kruskal-Wallis tests for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Group 1 and Group 2 were identified through k-means clustering of patients with Lp(a) ≥ 50 mg/dL. Group 3 includes patients with Lp(a) levels between 30 and 50 mg/dL, and Group 4 includes patients with Lp(a) < 30 mg/dL. Lp(a), lipoprotein(a); LDL, WBC, white blood cell; eGFR, estimated glomerular filtration rate; AST, aspartate aminotransferase; ALT, alanine aminotransferase; LDL, low-density lipoprotein; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; PT-INR, prothrombin time-international normalized ratio; aPTT, activated partial thromboplastin time. Group 1 (n = 1,249) had the highest mean age (67.3 ± 11.1 years), a predominance of males (60.8%), and the highest prevalence of cardiometabolic comorbidities, including hypertension (91.7%), diabetes mellitus (62.0%), dyslipidemia (97.1%), and chronic kidney disease (15.9%). Renal function was reduced (eGFR 82.0 ± 23.2 mL/min/1.73m 2 ), and lipid levels were the lowest among all groups (total cholesterol 145.5 ± 30.9 mg/dL, LDL-cholesterol 77.5 ± 25.1 mg/dL), reflecting a heavily treated high-risk phenotype, supported by high statin (87.8%) and antiplatelet (62.6%) use. Group 2 (n = 1,106) was the youngest group (59.4 ± 12.9 years), predominantly female (68.0%), with the lowest prevalence of hypertension (54.0%), diabetes mellitus (17.4%), and chronic kidney disease (1.4%). This group exhibited the highest lipid levels (total cholesterol 191.9 ± 37.2 mg/dL, LDL-cholesterol 113.4 ± 34.4 mg/dL, HDL-cholesterol 59.1 ± 12.3 mg/dL) and lower statin (56.5%) and antiplatelet (23.7%) use, consistent with a relatively drug-naïve dyslipidemia phenotype. Group 3 (n = 2,029), defined by Lp(a) levels of 30–50 mg/dL, showed intermediate clinical characteristics, with a mean age of 62.5 ± 13.3 years, moderate prevalence of hypertension (69.4%), diabetes mellitus (37.1%), and dyslipidemia (80.7%), and lipid levels between those of Groups 1 and 2. Group 4 (n = 13,269), defined by Lp(a) < 30 mg/dL, represented the majority of the cohort and demonstrated similar intermediate characteristics; mean age 60.4 ± 14.5 years, hypertension 68.8%, diabetes mellitus 38.2%, and total cholesterol 167.8 ± 42.2 mg/dL. Renal function was well preserved (eGFR 92.7 ± 19.0 mL/min/1.73m 2 ). Post hoc analyses (Bonferroni-adjusted Dunn test) confirmed that most variables, including age, body mass index, comorbidities, lipid profiles, glucose, and renal function, differed significantly between groups. These findings confirm that the k-means derived high Lp(a) clusters represent distinct clinical phenotypes: Cluster 1 is characterized by older age, increased cardiometabolic disease burden, polypharmacy, and lower renal function, whereas Cluster 2 is predominantly a young, female, relatively drug-naïve dyslipidemia group. Clinical outcomes according to cluster and Lp(a) level During the follow-up period (725.7 ± 386.8 days [median: 857 days]), the incidence of MACE differed significantly across the four groups (overall log-rank p < 0.05, Fig. 3 ). Group 1(Cluster 1, Lp(a) ≥ 50 mg/dL) exhibited the highest cumulative MACE incidence (8.9%), followed by Group 3 (Lp(a) 30–50 mg/dL, 4.8%), Group 4 (Lp(a) < 30 mg/dL, 3.7%), and Group 2 (Cluster 2, Lp(a) ≥ 50 mg/dL, 2.0%). Pairwise log-rank testing confirmed that all inter-group comparisons were statistically significant, with the largest hazard separation observed between Group 1 (Cluster 1) and Group 2 (Cluster 2). For secondary outcomes, significant overall differences between groups were observed for cardiovascular death ( p = 0.01), new-onset myocardial infarction ( p < 0.01), new-onset stroke ( p = 0.02), new-onset heart failure hospitalization ( p < 0.01), coronary revascularization ( p < 0.01), and all-cause death ( p < 0.01) (Supplemental Fig. 2). The incidence of new-onset myocardial infarction was highest in Group 1 (2.4%), followed by Group 3 (1.7%), Group 4 (0.7%), and Group 2 (0.2%). New-onset heart failure hospitalization occurred most frequently in Group 1 (1.9%), with marked lower rates in the other groups (≤ 0.8%). Similarly, coronary revascularization was performed most often in Group 1 (5.2%), compared with 2.5% in Group 3, 1.8% in Group 4, and 0.9% in Group 2. All-cause mortality was significantly higher in Group 1 (1.3%) compared with the other groups (≤ 0.8%), with the largest difference observed between Group 1 and Group 4 ( p < 0.01). For cardiovascular death, although the absolute number of events was low across all groups, the incidence was highest in Group 1 (0.2%) and lowest in Group 2 (0%). Overall, these findings indicate that, among patients with high Lp(a), those classified as Group 1 (Cluster 1)—characterized by older age, higher prevalence of comorbidities, and greater medication use—exhibited substantially higher risks of adverse cardiovascular outcomes compared to both Group 2 (Cluster 1) and the other low Lp(a) groups (Group 3 and 4). Notably, Group 2 (Cluster 2) patients (younger, relatively drug-naïve dyslipidemia phenotype) had lowest event rates. Multivariable Cox Proportional Hazard Analysis In the multivariable Cox proportional hazards regression model, several clinical characteristics were proposed as independent risk predictors of 3-year MACE (Table 2 ). Older age (HR [95%CI] 1.03 [1.03–1.04]), male sex (HR [95%CI] 1.42 [1.19–1.70]), smoking (HR [95%CI] 1.45 [1.10–1.91]), alcohol consumption (HR [95%CI] 4.02 [2.95–5.49]), and low socio-economic status (HR [95%CI] 1.90 [1.52–2.39]) were each associated with higher 3-year MACE risk ( p < 0.01 for all). Hypertension (HR [95%CI] 1.44 [1.15–1.81]), dyslipidemia (HR [95%CI] 1.86 [1.36–2.54]), higher LDL-cholesterol (HR [95%CI] 1.01 [1.00-1.01]), elevated hs-CRP (HR [95%CI] 1.06 [1.02–1.09]), and higher creatinine (HR [95%CI] 1.42 [1.04–1.94]) were also significant predictors. Table 2 Multivariable Cox Proportional Hazards Regression for 3-Year MACE Variable HR (95% CI) p -value Age (per year) 1.034 (1.027–1.042) < 0.01 Male sex (vs. female) 1.418 (1.186–1.696) < 0.01 Smoking 1.451 (1.101–1.912) < 0.01 Alcohol consumption 4.023 (2.949–5.488) < 0.01 Low socio-economic status 1.903 (1.518–2.386) < 0.01 Body mass index (per kg/m²) 0.980 (0.958–1.003) 0.08 Hypertension 1.440 (1.148–1.805) < 0.01 Diabetes mellitus 1.098 (0.920–1.309) 0.30 Dyslipidemia 1.860 (1.363–2.538) < 0.01 Chronic kidney disease 1.120 (0.814–1.541) 0.49 Total cholesterol (per mg/dL) 0.999 (0.995–1.003) 0.72 LDL-cholesterol (per mg/dL) 1.005 (1.001–1.010) 0.02 HDL-cholesterol (per mg/dL) 0.995 (0.988–1.003) 0.24 Triglyceride (per mg/dL) 1.001 (1.000–1.002) 0.14 Glucose (per mg/dL) 1.002 (0.999–1.006) 0.19 Creatinine (per mg/dL) 1.422 (1.040–1.945) 0.03 hs-CRP (per mg/L) 1.056 (1.019–1.095) < 0.01 Group 1 (vs Group 4) 1.391 (1.127–1.718) < 0.01 Group 2 (vs Group 4) 1.078 (0.692–1.679) 0.74 Group 3 (vs Group 4) 1.182 (0.948–1.472) 0.14 Hazards ratios (HR) and 95% confidence intervals (CI) were derived from pooled Cox proportional hazards regression across multiple imputed datasets (Rubin’s rules). When stratified by cluster phenotype, using Group 4 as the reference, Group 1 (Cluster 1) patients had a significantly higher hazard (HR [95%CI] 1.39 [1.13–1.72]; p < 0.01). whereas Group 2 (Cluster 2) and 3 did not differ significantly from Group 4. These findings suggest that the phenotypic grouping captured by the clustering methodology has prognostic relevance, driven largely by the Cluster 1 profile (Group 1). Moreover, despite the markedly elevated Lp(a) levels (≥ 50 mg/dL) in Group 2 (Cluster 2), the risk of MACE did not differ significantly from the low-Lp(a) (< 30 mg/dL) reference Group 4 (HR [95%CI] 1.08 [0.69–1.68]; p = 0.74), suggesting that even within the high-Lp(a) population, a clinically distinct low-risk subgroup may be identifiable. Discussion In a large, contemporary cohort, we applied unsupervised k-means clustering to patients with elevated Lp(a) (≥ 50 mg/dL) and showed two reproducible phenogroups with distinct clinical profiles and outcomes. Cluster 1 was characterized by older age, male predominance, high prevalence of cardiometabolic comorbidities; Cluster 2, included younger, predominantly female patients with a relatively treatment-naïve dyslipidemia profile. Despite comparable Lp(a) levels, 3-year MACE rates differed markedly (8.9% vs 2.0%, p < 0.01). In multivariable Cox models, Cluster 1 independently predicted higher MACE risk compared with the low-Lp(a) reference group, while Cluster 2 showed risk similar to the reference. These findings support the prognostic value of phenotype-based clustering beyond concentration cut-offs and reveal clinically meaningful heterogeneity within the high-Lp(a) population. Importantly, the elevated risk observed in Cluster 1 likely reflects the combined influence of multiple cardiometabolic risk factors rather than the isolated effect of Lp(a) concentration alone. This observation underscores a key concept in Lp(a)-related cardiovascular risk: the clinical impact of elevated Lp(a) may depend substantially on the surrounding biological and clinical context. In this framework, Lp(a) may act as a risk amplifier within high-risk phenotypes rather than functioning as a uniform risk determinant across all individuals with elevated levels. Lp(a) has emerged as a major residual cardiovascular risk factor and is designated a “risk enhancer” in current guideline 7 . The Copenhagen City Heart Study reported a 2.6-fold increase in myocardial infarction among individuals with Lp(a) >95th percentile compared to those < 22nd percentile 8 . The 2010 European Atherosclerosis Society consensus proposed a 50 mg/dL threshold, aligned with approximately the 80th percentile of Danish distribution 5 . Analyses from the UK Biobank further demonstrated a continuous, log-linear relationship between Lp(a) concentrations and atherosclerotic cardiovascular disease risk, reinforcing its dose-response association 9 . Translational evidence from the FOURIER trial showed greater absolute event reduction with PCSK9 inhibition among patients with above-median Lp(a) despite optimal statin therapy 10 . Nonetheless, the present study indicates that a single fixed threshold (≥ 50 mg/dL) does not uniformly convey equivalent risk, highlighting the need to contextualize Lp(a) within broader biological and clinical phenotypes. Emerging evidence suggests that the pathogenicity of Lp(a) is not determined by plasma concentration alone but is modulated by the surrounding milieu. First, apo(a) isoform size appears to influence risk independently of concentration, with smaller isoform conferring higher event rates—potentially via higher oxidized phospholipid burden 11 . Apo(a) size has been also associated to attenuated Lp(a) reduction with PCSK9 inhibition, suggesting biological subgroup heterogeneity 12 . Second, the cholesterol cargo of Lp(a) can inflate calculated LDL-cholesterol; accounting for Lp(a)-cholesterol has been proposed to mitigate misclassification 13 . Third, inflammatory status further modifies Lp(a)-related risk—e.g., in familial hypercholesterolemia, Lp(a) predicted adverse events primarily in the presence of elevated C-reactive protein 14 . These mechanisms align with our observation that the comorbidity- and inflammation-enriched phenotype (Cluster 1) concentrated on events, whereas the relatively favorable phenotype (Cluster 2) did not, despite equally high Lp(a). Recent longitudinal data also show that Lp(a) changes track with clinical and laboratory profiles, further supporting a context-dependent risk model 15 . Unsupervised clustering approaches have gained interest in lipidology and cardiovascular risk stratification 16 , 17 . In Lp(a) research, prior studies have primarily demonstrated cross-sectional associations or surrogate markers. Saraiva et al. stratified 661 patients by Lp(a) and observed higher cardiovascular event rates in the high-Lp(a) cluster 18 . The RED-CARPET analysis identified four high-Lp(a) phenogroups, with the “male/unhealthy lifestyle” cluster showing the most pronounced atherosclerotic cardiovascular risk association 4 . Extending these observations, the present study integrates clustering with time-to-event analysis in a large real-world cohort, thereby demonstrating the independent prognostic significance of phenotype-based stratification among patients with elevated Lp(a). These data have distinct clinical implications. A threshold-only strategy risks over-treating biologically low-risk individuals (e.g., Cluster 2) and under-recognizing high-risk phenotypes (e.g., Cluster 1). Phenotype-guided strategies—incorporating inflammation (hs-CRP), apoB/LDL-cholesterol (with consideration of Lp(a)-cholesterol), renal function, and lifestyle factors—may more precisely target intensified therapy, including PCSK9 inhibitors and, prospectively, Lp(a)-targeted agents. Thus, even for a causal risk factor, the observed clinical effect size is contingent on patient context, supporting a shift from population-based to milieu-adjusted risk assessment. This study has several limitations. First, the retrospective design using electronic health record data may introduce residual confounding and selection bias despite multivariable adjustment. In particular, Lp(a) testing was performed in routine clinical practice rather than through systematic screening, which may have resulted in a cohort enriched for patients with higher baseline cardiovascular risk. Second, clustering analyses are inherently data-driven and may be influenced by variable selection and data structure. Although multiple validation approaches were applied, external validation in independent cohorts will be necessary to confirm the reproducibility of these phenotypes. Third, the study population was derived from tertiary academic centers in Korea, which may limit the generalizability of findings to other healthcare settings or ethnic populations. In conclusion, phenotypic clustering of patients with elevated Lp(a) reveals distinct, prognostically relevant subgroups and supports integrating multidimensional patient profiling—beyond Lp(a) concentration alone—into cardiovascular risk assessment and therapeutic decision-making. Declarations Author Contributions H.J.J. and S.Y.S. conceived the study. H.J.J., S.J.H., and C.W.Y. collected and curated the data. H.J.J. performed statistical analyses. H.J.J. drafted the manuscript. E.J.K. and S.Y.S. critically revised the manuscript. All authors reviewed and approved the final manuscript. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Due to institutional regulations and patient privacy protection, the raw data cannot be publicly shared. Competing interests The authors declare no competing interests. Acknowledgement We sincerely acknowledge the invaluable support and contributions of the Cardiovascular Data Science Research Group of the Department of Cardiology at Korea University. Their dedication and expertise were instrumental in the successful completion of this study. Sources of Funding This research was supported from the Medical data-driven hospital support project through the Korea Health Information Service(KHIS), funded by the Ministry of Health & Welfare, and the Ministry of Science and ICT, Korea, under the ICT Challenge and Advanced Network of HRD program (IITP-2025-RS-2022-00156439) supervised by the Institute of Information & Communications Technology Planning & Evaluation. References Doherty, S. et al. Lipoprotein(a) as a Causal Risk Factor for Cardiovascular Disease. Curr. Cardiovasc. Risk Rep. 19 , 8 (2025). Emerging Risk Factors Collaboration. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA 302 , 412–423 (2009). Nordestgaard, B. G. et al. Lipoprotein(a) as a cardiovascular risk factor: current status. Eur. Heart J. 31 , 2844–2853 (2010). Zhang, S. et al. Phenomapping of subgroups in high-Lp(a) patients: a data-driven cluster analysis in RED-CARPET study. Clin Res. Cardiol (2025). Richards, A. M. Biomarkers in Acute Heart Failure - Cardiac And Kidney. Card Fail. Rev. 1 , 107–111 (2015). Çelik, A. et al. How to Use Natriuretic Peptides in Patients with Heart Failure with Non-Reduced Ejection Fraction? Anatol. J. Cardiol. 27 , 308–318 (2023). Kronenberg, F. et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. Eur. Heart J. 43 , 3925–3946 (2022). Kamstrup, P. R., Tybjaerg-Hansen, A., Steffensen, R. & Nordestgaard, B. G. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. JAMA 301 , 2331–2339 (2009). Patel, A. P. et al. Lp(a) (Lipoprotein[a]) Concentrations and Incident Atherosclerotic Cardiovascular Disease: New Insights From a Large National Biobank. Arterioscler. Thromb. Vasc Biol. 41 , 465–474 (2021). O'Donoghue, M. L. et al. Lipoprotein(a), PCSK9 Inhibition, and Cardiovascular Risk. Circulation 139 , 1483–1492 (2019). Saleheen, D. et al. Apolipoprotein(a) isoform size, lipoprotein(a) concentration, and coronary artery disease: a mendelian randomisation analysis. Lancet Diabetes Endocrinol. 5 , 524–533 (2017). Blanchard, V. et al. The size of apolipoprotein (a) is an independent determinant of the reduction in lipoprotein (a) induced by PCSK9 inhibitors. Cardiovasc. Res. 118 , 2103–2111 (2022). Willeit, P. et al. Low-Density Lipoprotein Cholesterol Corrected for Lipoprotein(a) Cholesterol, Risk Thresholds, and Cardiovascular Events. J. Am. Heart Assoc. 9 , e016318 (2020). Tada, H. et al. Synergistic effect of lipoprotein (a) and C-reactive protein on prognosis of familial hypercholesterolemia. Am. J. Prev. Cardiol. 12 , 100428 (2022). Joo, H. J. et al. Predictors of lipoprotein(a) variability in clinical practice and their impact on cardiovascular risk. Lipids Health Dis. 24 , 250 (2025). Cho, A. R., Heo, S. J., Han, T. & Kwon, Y. J. Cluster-Based Analysis of Lipid Profiles and Inflammation in Association With Cardiovascular Disease Incidence and Mortality: A 17.5-Year Longitudinal Study. J. Clin. Hypertens. (Greenwich) . 27 , e70035 (2025). Sharma, A. et al. Cluster Analysis of Cardiovascular Phenotypes in Patients With Type 2 Diabetes and Established Atherosclerotic Cardiovascular Disease: A Potential Approach to Precision Medicine. Diabetes Care . 45 , 204–212 (2022). Saraiva, M. et al. Unveiling patient profiles associated with elevated Lp(a) through an unbiased clustering analysis. Front. Cardiovasc. Med. 12 , 1546351 (2025). Additional Declarations No competing interests reported. Supplementary Files GraphicalAbstract1.tif Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 06 Apr, 2026 Reviews received at journal 05 Apr, 2026 Reviews received at journal 03 Apr, 2026 Reviews received at journal 31 Mar, 2026 Reviewers agreed at journal 27 Mar, 2026 Reviews received at journal 25 Mar, 2026 Reviewers agreed at journal 24 Mar, 2026 Reviewers agreed at journal 23 Mar, 2026 Reviewers agreed at journal 22 Mar, 2026 Reviewers agreed at journal 22 Mar, 2026 Reviewers invited by journal 16 Mar, 2026 Editor invited by journal 16 Mar, 2026 Editor assigned by journal 09 Mar, 2026 Submission checks completed at journal 09 Mar, 2026 First submitted to journal 08 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9065291","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":607154195,"identity":"1a75795b-33bc-48f7-8074-3acd365605a0","order_by":0,"name":"Hyung Joon Joo","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"prefix":"","firstName":"Hyung","middleName":"Joon","lastName":"Joo","suffix":""},{"id":607154197,"identity":"842deb59-1cf1-4264-86f7-d4269eddce94","order_by":1,"name":"Soon Joon Hong","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"prefix":"","firstName":"Soon","middleName":"Joon","lastName":"Hong","suffix":""},{"id":607154202,"identity":"a6c3d50a-a4d9-484c-b552-0bd8830b83a6","order_by":2,"name":"Cheol Woong Yu","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"prefix":"","firstName":"Cheol","middleName":"Woong","lastName":"Yu","suffix":""},{"id":607154204,"identity":"795b94d4-0c2d-411b-ad78-17ad071edf75","order_by":3,"name":"Seung Yong Shin","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIie3RMQrCMBSA4VcCugRcLZV6hYZCR72KoRCXgo4FBStCu+he8RSeQCWQLjlABRcRvEBXESOIixJ1c8g/Pvh4Dx6AyfSPNQEQeICtLH3OkFbgB3HRXPxEAPx6zr4kXWfGq+HwQBfLMz3FcQca2Rb5sW5LSzAn9850uWKcSBlCU/YQldrDogBhj9P1qp/a01TdVALaJVoyqKo72ewLRa4TaH8mETiK+FZeE/Y04eApQrWkZMGduNachSQRBSaSzoiO1PPwVOELV68U5JiMR65bcG7ryOtaAOsnYDKZTKY33QBslkkgNiA0zQAAAABJRU5ErkJggg==","orcid":"","institution":"Korea University","correspondingAuthor":true,"prefix":"","firstName":"Seung","middleName":"Yong","lastName":"Shin","suffix":""},{"id":607154209,"identity":"71313435-e10c-436f-855a-bdf5ccf31340","order_by":4,"name":"Eung Ju Kim","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"prefix":"","firstName":"Eung","middleName":"Ju","lastName":"Kim","suffix":""}],"badges":[],"createdAt":"2026-03-08 15:39:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9065291/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9065291/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104994289,"identity":"50b57865-2850-4b32-a286-111e2c718767","added_by":"auto","created_at":"2026-03-19 15:59:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":670334,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrincipal component plots of K-means clustering assignments.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrincipal component analysis plots showing the distribution of high Lp(a) patients classified by k-means clustering with (A) two clusters, (B) three clusters, and (C) four clusters. Each data point represents an individual projected onto the first two principal components (Dim1 and Dim 2), with colors denoting cluster assignments and 95% confidence ellipses indicating boundaries.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9065291/v1/0e171c6870179c931d32ba02.png"},{"id":104994285,"identity":"c068f56b-fa2c-40b7-a417-5a62c5554ccc","added_by":"auto","created_at":"2026-03-19 15:59:41","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":397954,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKey features contributing to cluster differentiation and cluster profiles\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Top ten clinical features in differentiating clusters, based on mean decrease in accuracy from random forest analysis. (B) Radar plot of standardized clinical profiles, illustrating phenotypic differences between the two clusters.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9065291/v1/148197db3c27fd9cf13c77b1.png"},{"id":104994291,"identity":"58ecd887-b3b6-4107-bbe4-ec46f3ab7bc9","added_by":"auto","created_at":"2026-03-19 15:59:43","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":354026,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKaplan-Meier curves for major adverse cardiovascular events according to phenotypical clusters and Lp(a) level\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTime-to-first event curves stratified by four groups for MACE over 3 years. Overall survival differences were evaluated with the log-rank test, with pairwise comparisons performed between groups.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9065291/v1/81d1acd040164d52410d1348.png"},{"id":105035059,"identity":"89f21c53-d0df-423a-a856-5b65695c5c74","added_by":"auto","created_at":"2026-03-20 07:25:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2096219,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9065291/v1/1101d0b2-195d-43d9-9a94-f245e982e187.pdf"},{"id":104994288,"identity":"9615a816-7090-4650-91c4-665103ea58cb","added_by":"auto","created_at":"2026-03-19 15:59:42","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2801396,"visible":true,"origin":"","legend":"","description":"","filename":"GraphicalAbstract1.tif","url":"https://assets-eu.researchsquare.com/files/rs-9065291/v1/01bbd2b6421dcb74166d9ce1.tif"}],"financialInterests":"No competing interests reported.","formattedTitle":"Phenotypic Clustering Identifies Heterogeneous Cardiovascular Risk Among Patients with Elevated Lipoprotein(a)","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLipoprotein(a) [Lp(a)] has emerged as an independent, genetically determined risk factor for atherosclerotic cardiovascular disease (ASCVD). Previous Mendelian-randomization study and large clinical cohort study reported a significant increase in cardiovascular events with the elevated Lp(a) concentration \u003csup\u003e\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. However, the incidence of cardiovascular events differs across trials and cohorts, suggesting that co-existing clinical factors may modify Lp(a)-mediated risk.\u003c/p\u003e \u003cp\u003eMachine-learning methodologies offer a hypothesis-free approach to disentangle such heterogeneity. In the RED-CARPET study, Zhang et al. first applied k-means clustering to the patients with Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL and identified four phenogroups: a dyslipidemia cluster, a cluster of aged women, a cluster of men with unhealthy lifestyle, and a cluster by anemia, renal insufficiency and hypercoagulability \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Although Lp(a) concentrations were comparable, the risk of ASCVD varied markedly; men with unhealthy lifestyle had the highest risk relative a low-Lp(a) reference group, whereas the anemia/renal insufficiency cluster showed no significant excess risk. This finding highlights the importance of multi-dimensional phenotyping over single-analyte risk stratification.\u003c/p\u003e \u003cp\u003eHowever, whether such phenotypic subgroups provide independent prognostic value beyond conventional cardiovascular risk factors remains uncertain, and the reproducibility of these phenotypes across different populations has not been well established. Furthermore, most prior studies have focused on cross-sectional phenotyping rather than evaluating long-term clinical outcomes. Therefore, in the present study we applied unsupervised k-means clustering to a large contemporary, unselected Korean patient with elevated Lp(a) cohort and examined the association between identified phenotypes and 3-year major adverse cardiovascular events (MACE). By integrating phenotypic clustering with time-to-event analysis, this study aims to determine whether clinically meaningful heterogeneity exists within the high-Lp(a) population beyond concentration-based theresholds.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Data Sources\u003c/h2\u003e \u003cp\u003eThis retrospective observational study was based on de-identified data from the Observational Medical Outcomes Partnership Common Data Model (OMOP-CDM) database and electronic health records (EHRs) of three tertiary academic medical centers in Korea. The OMOP-CDM framework, provided by the Observational Health Data Sciences and Informatics (OHDSI) collaboration, served as the standardized structure for organizing EHRs. Clinical data, including diagnoses based on the International Classification of Diseases, 10th Revision (ICD-10), medications, and laboratory results, were mapped to unique concept identifiers in the OMOP-CDM. All data were stored on Microsoft SQL servers and assessed using direct SQL queries. This study adhered to the principles of the Declaration of Helsinki and was approved by the institutional Review Board (IRB No. 2025AN0429). The requirement for informed consent was waived due to the retrospective nature of the study and use of de-identified data.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eStudy Population\u003c/h3\u003e\n\u003cp\u003eThe dataset included adult patients (age\u0026thinsp;\u0026ge;\u0026thinsp;18 years) who underwent serum Lp(a) testing between January 2017 and December 2024. A total of 33,954 patients with Lp(a) measurements were initially screened. Exclusion criteria were history of myocardial infarction (n\u0026thinsp;=\u0026thinsp;2,556) or stroke (n\u0026thinsp;=\u0026thinsp;4,834), any cancer (n\u0026thinsp;=\u0026thinsp;3,537), thyroid disease or hormone therapy (n\u0026thinsp;=\u0026thinsp;5,399), chronic viral infection (n\u0026thinsp;=\u0026thinsp;506), extremely abnormal liver injury (n\u0026thinsp;=\u0026thinsp;19). Patients who experienced a major adverse cardiovascular event (MACE) within the first 30 days of follow-up (n\u0026thinsp;=\u0026thinsp;752) or were lost to follow-up during this period (n\u0026thinsp;=\u0026thinsp;1,396) were subsequently excluded. After these exclusions, 17,653 patients remained in the final cohort. For clustering analysis, patients with Lp(a) levels\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL were included (n\u0026thinsp;=\u0026thinsp;4,320).\u003c/p\u003e\n\u003ch3\u003eSerum Lipoprotein(a) Measurement\u003c/h3\u003e\n\u003cp\u003eLp(a) testing was conducted as part of routine clinical practice using the Randox Lp(a) assay (Catalog No. LP3403, Randox Laboratories, Crumlin, United Kingdom) on a Beckman Coulter AU5800 analyzer. Daily internal control assays maintained the precision with intra- and inter-assay coefficients of variation between 2.4% and 3.2%. Lp(a) risk categories were defined based on the EAS consensus statement: high risk (Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL), low risk (Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30 mg/dL), and intermediate risk for values between 30 and 50 mg/dL.\u003c/p\u003e\n\u003ch3\u003eClinical Variables and Outcomes\u003c/h3\u003e\n\u003cp\u003eDemographic, socio-economic status, lifestyle behaviors, comorbidities, medication use, and laboratory data were extracted and defined. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Serum creatinine-based estimated glomerular filtration rate (eGFR) was estimated using the CKD-EPI equations, truncated at a maximum of 150 mL/min/1.73m\u003csup\u003e2\u003c/sup\u003e to mitigate physiological overestimation.\u003c/p\u003e \u003cp\u003eThe primary outcome was the time-to-occurrence of MACE, defined as a composite of cardiovascular death, new-onset myocardial infarction, stroke, heart failure hospitalization or coronary revascularization. New-onset myocardial infarction was defined as hospital admission for chest pain, angina, or dyspnea accompanied by cardiac marker levels exceeding the 99th percentile of the upper reference limit. New-onset stroke was identified by a new diagnosis of stroke or confirming acute, subacute, or recent cerebral infarction on brain MRI. Heart failure hospitalization was defined as a hospital admission following an emergency department visit, accompanied by elevated N-terminal pro B-type natriuretic peptide or brain natriuretic peptide levels, and requiring intravenous furosemide administration \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. Time-to-event calculated from 30 days after the date of Lp(a) measurement to event occurrence or last known follow-up, truncated at three years.\u003c/p\u003e \u003cp\u003eClinical outcomes were identified using a combination of diagnostic codes, laboratory measurements, and treatment records within the OMOP-CDM framework. This algorithm-based approach has been widely used in observational EHR studies to improve the accuracy of event ascertainment. When available, laboratory thresholds and treatment criteria were incorporated to increase specificity of outcome definitions and reduce misclassification.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eUnsupervised k-means clustering was applied to patients with Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL. Thirty-three clinical and biochemical variables were standardized prior to clustering. The optimal number of clusters was determined via elbow method, average silhouette width, and NbClust consensus.\u003c/p\u003e \u003cp\u003eFor clustering analysis, patients with missing values among selected clustering variables were excluded. In subsequent analysis, multiple imputations by chained equations were employed using the predictive mean matching method.\u003c/p\u003e \u003cp\u003eCategorical variables are expressed as frequencies and percentages, and continuous variables are expressed as mean with standard deviations. Continuous variables were compared across groups using one-way analysis of variance (ANOVA) for normally distributed data and the Kruskal-Wallis test for non-normally distributed data. Post hoc pairwise comparisons were performed with Bonferroni correction or Dunn\u0026rsquo;s test as appropriate. Categorical variables were compared using the chi-square test or Fisher\u0026rsquo;s exact test.\u003c/p\u003e \u003cp\u003eTime-to-event outcomes were analyzed using the Kaplan-Meier method to estimate cumulative incidence rates over the follow-up period, with censoring applied at the time of death or the last available follow-up. To evaluate clinical outcomes across clusters, multivariable Cox proportional hazards models were constructed. Final model selection was performed through backward selection using Akaike Information Criterion. All analyses were conducted using R version 4.3.1. Key packages included cluster, Nbclust, factoextra, mice, randomForest, and survival. Plots were generated using ggplot2.\u003c/p\u003e \u003cp\u003eTo enhance the robustness of clustering results, multiple cluster validation approaches were applied, including the elbow method, average silhouette width, and consensus recommendations from the NbClust package. Principal component analysis was additionally used to visually assess cluster separation in reduced dimensional space. These complementary approaches were used to mitigate the risk of overfitting or arbitrary cluster selection and to ensure that the chosen cluster solution reflected stable and interpretable phenotypic patterns within the dataset.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eCluster validation and identification of two phenotypes\u003c/h2\u003e \u003cp\u003eCluster validation indices revealed heterogenous recommendations, with the majority suggesting two clusters as optimal (Supplemental Fig.\u0026nbsp;1). The elbow plot of within-cluster variance showed an inflection at two clusters, and silhouette analysis demonstrated better cohesion and separation for the two-cluster solution (mean silhouette width, 0.09) compared with three clusters (0.07). Principal component analysis was used to visualize the distribution of cluster assignments for k\u0026thinsp;=\u0026thinsp;2, k\u0026thinsp;=\u0026thinsp;3, and k\u0026thinsp;=\u0026thinsp;4. The four- and three-cluster models showed substantial overlap among clusters, while the two-cluster model demonstrated the clearer separation between groups along the first principal component (13.1% of variance) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). These findings support the selection of two clusters as the most interpretable and distinct phenotypic classification for high Lp(a) patients. These two phenotypic clusters were subsequently used for further comparative and outcome analyses.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eDeterminants of cluster differentiation\u003c/h3\u003e\n\u003cp\u003eTo identify the clinical features most strongly associated with cluster assignment, feature importance was evaluated using a random forest classification model (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The top ten features contributing to cluster differentiation included total cholesterol, HDL-cholesterol, LDL-cholesterol, hypertension, diabetes mellitus, sex, dyslipidemia, alcohol consumption, smoking, and glucose. Notably, lipid parameters comprised a substantial proportion of the most influential predictors. A radar plot of standardized mean values for these key features (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB) demonstrated that Cluster 1 was characterized by higher prevalence of hypertension and diabetes mellitus, along with more frequent alcohol consumption and smoking. In contrast, Cluster 2 demonstrated higher levels of total cholesterol, LDL-cholesterol, and HDL-cholesterol, consistent with a drug-na\u0026iuml;ve dyslipidemia-dominant phenotype. It suggests that traditional cardiovascular risk factors could play a role in distinguishing between these clusters.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eBaseline clinical characteristics of study population\u003c/h2\u003e \u003cp\u003eBaseline characteristics were compared among Group 1 and Group 2 (both derived from k-means clustering of patients with high Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL), and Group 3 and Group 4, which were defined according to Lp(a) thresholds of 30\u0026ndash;50 mg/dL and \u0026lt;\u0026thinsp;30 mg/dL, respectively (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Significant differences were observed across the four groups in demographic characteristics, lifestyle factors, comorbidities, laboratory findings, and medication use (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01 for most comparisons).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics of the study population stratified by Lp(a)-based groups and k-means clustering\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGroup 1\u003c/p\u003e \u003cp\u003e(Cluster 1, Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;1,249)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGroup 2\u003c/p\u003e \u003cp\u003e(Cluster 2, Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;1,106)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGroup 3\u003c/p\u003e \u003cp\u003e(Lp(a) 30\u0026ndash;50)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;2,029)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGroup 4\u003c/p\u003e \u003cp\u003e(Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30)\u003c/p\u003e \u003cp\u003e(n\u0026thinsp;=\u0026thinsp;13,269)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cem\u003ep-\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e67.3\u0026thinsp;\u0026plusmn;\u0026thinsp;11.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e59.4\u0026thinsp;\u0026plusmn;\u0026thinsp;12.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e62.5\u0026thinsp;\u0026plusmn;\u0026thinsp;13.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e60.4\u0026thinsp;\u0026plusmn;\u0026thinsp;14.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale sex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e760 (60.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e354 (32.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,001 (49.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7,434 (56.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCurrent smoking\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e751 (60.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e274 (24.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e892 (44.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5,627 (42.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlcohol consumption\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e804 (64.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e312 (28.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e926 (45.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6,037 (45.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow socioeconomic status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80 (6.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40 (3.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e125 (6.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e605 (4.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody mass index, kg/m\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25.3\u0026thinsp;\u0026plusmn;\u0026thinsp;3.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23.9\u0026thinsp;\u0026plusmn;\u0026thinsp;3.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e24.5\u0026thinsp;\u0026plusmn;\u0026thinsp;3.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24.8\u0026thinsp;\u0026plusmn;\u0026thinsp;3.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1,145 (91.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e597 (54.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,408 (69.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9,124 (68.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes mellitus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e775 (62.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e192 (17.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e752 (37.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5,069 (38.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDyslipidemia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1,213 (97.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e790 (71.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,638 (80.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10,395 (78.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChronic kidney disease\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e199 (15.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (1.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e159 (7.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e791 (6.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLp(a), mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e89.4\u0026thinsp;\u0026plusmn;\u0026thinsp;41.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e83.9\u0026thinsp;\u0026plusmn;\u0026thinsp;34.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38.4\u0026thinsp;\u0026plusmn;\u0026thinsp;5.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e11.6\u0026thinsp;\u0026plusmn;\u0026thinsp;7.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHemoglobin, g/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.5\u0026thinsp;\u0026plusmn;\u0026thinsp;1.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.6\u0026thinsp;\u0026plusmn;\u0026thinsp;1.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWBC, \u0026times;10\u0026sup3;/\u0026micro;L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6.8\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.9\u0026thinsp;\u0026plusmn;\u0026thinsp;1.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6.3\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6.3\u0026thinsp;\u0026plusmn;\u0026thinsp;1.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet count, \u0026times;10\u0026sup3;/\u0026micro;L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e228.0\u0026thinsp;\u0026plusmn;\u0026thinsp;57.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e244.4\u0026thinsp;\u0026plusmn;\u0026thinsp;56.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e234.4\u0026thinsp;\u0026plusmn;\u0026thinsp;58.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e234.7\u0026thinsp;\u0026plusmn;\u0026thinsp;57.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCreatinine, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.96\u0026thinsp;\u0026plusmn;\u0026thinsp;0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.76\u0026thinsp;\u0026plusmn;\u0026thinsp;0.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.85\u0026thinsp;\u0026plusmn;\u0026thinsp;0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.85\u0026thinsp;\u0026plusmn;\u0026thinsp;0.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eeGFR, mL/min/1.73m\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e82.0\u0026thinsp;\u0026plusmn;\u0026thinsp;23.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95.5\u0026thinsp;\u0026plusmn;\u0026thinsp;15.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90.3\u0026thinsp;\u0026plusmn;\u0026thinsp;20.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e92.7\u0026thinsp;\u0026plusmn;\u0026thinsp;19.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAST, U/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e28.9\u0026thinsp;\u0026plusmn;\u0026thinsp;15.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25.7\u0026thinsp;\u0026plusmn;\u0026thinsp;8.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27.4\u0026thinsp;\u0026plusmn;\u0026thinsp;14.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e28.2\u0026thinsp;\u0026plusmn;\u0026thinsp;17.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eALT, U/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e27.0\u0026thinsp;\u0026plusmn;\u0026thinsp;20.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.9\u0026thinsp;\u0026plusmn;\u0026thinsp;13.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e25.2\u0026thinsp;\u0026plusmn;\u0026thinsp;19.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e27.1\u0026thinsp;\u0026plusmn;\u0026thinsp;21.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal bilirubin, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.72\u0026thinsp;\u0026plusmn;\u0026thinsp;0.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.71\u0026thinsp;\u0026plusmn;\u0026thinsp;0.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.73\u0026thinsp;\u0026plusmn;\u0026thinsp;0.30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.76\u0026thinsp;\u0026plusmn;\u0026thinsp;0.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlbumin, g/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.4\u0026thinsp;\u0026plusmn;\u0026thinsp;0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.3\u0026thinsp;\u0026plusmn;\u0026thinsp;0.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.4\u0026thinsp;\u0026plusmn;\u0026thinsp;0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal cholesterol, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e145.5\u0026thinsp;\u0026plusmn;\u0026thinsp;30.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e191.9\u0026thinsp;\u0026plusmn;\u0026thinsp;37.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e168.9\u0026thinsp;\u0026plusmn;\u0026thinsp;42.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e167.8\u0026thinsp;\u0026plusmn;\u0026thinsp;42.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDL-cholesterol, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77.5\u0026thinsp;\u0026plusmn;\u0026thinsp;25.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e113.4\u0026thinsp;\u0026plusmn;\u0026thinsp;34.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e96.3\u0026thinsp;\u0026plusmn;\u0026thinsp;36.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e94.6\u0026thinsp;\u0026plusmn;\u0026thinsp;35.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHDL-cholesterol, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e48.4\u0026thinsp;\u0026plusmn;\u0026thinsp;11.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e59.1\u0026thinsp;\u0026plusmn;\u0026thinsp;12.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e53.6\u0026thinsp;\u0026plusmn;\u0026thinsp;13.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e52.2\u0026thinsp;\u0026plusmn;\u0026thinsp;12.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglyceride, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e125.6\u0026thinsp;\u0026plusmn;\u0026thinsp;58.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e113.1\u0026thinsp;\u0026plusmn;\u0026thinsp;53.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e116.0\u0026thinsp;\u0026plusmn;\u0026thinsp;56.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e127.3\u0026thinsp;\u0026plusmn;\u0026thinsp;66.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlucose, mg/dL\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e115.5\u0026thinsp;\u0026plusmn;\u0026thinsp;24.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e101.9\u0026thinsp;\u0026plusmn;\u0026thinsp;15.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e108.6\u0026thinsp;\u0026plusmn;\u0026thinsp;20.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e109.1\u0026thinsp;\u0026plusmn;\u0026thinsp;21.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehs-CRP, mg/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.46\u0026thinsp;\u0026plusmn;\u0026thinsp;1.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.09\u0026thinsp;\u0026plusmn;\u0026thinsp;1.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.32\u0026thinsp;\u0026plusmn;\u0026thinsp;1.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.26\u0026thinsp;\u0026plusmn;\u0026thinsp;1.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePT-INR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.99\u0026thinsp;\u0026plusmn;\u0026thinsp;0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.97\u0026thinsp;\u0026plusmn;\u0026thinsp;0.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.98\u0026thinsp;\u0026plusmn;\u0026thinsp;0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.98\u0026thinsp;\u0026plusmn;\u0026thinsp;0.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eaPTT, sec\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e34.3\u0026thinsp;\u0026plusmn;\u0026thinsp;4.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33.6\u0026thinsp;\u0026plusmn;\u0026thinsp;3.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e34.1\u0026thinsp;\u0026plusmn;\u0026thinsp;4.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e34.0\u0026thinsp;\u0026plusmn;\u0026thinsp;3.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntiplatelet use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e782 (62.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e262 (23.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e798 (39.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4,456 (33.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntihypertensive use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e930 (74.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e413 (37.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,027 (50.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6,774 (51.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntidiabetic use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e513 (41.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e116 (10.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e503 (24.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3,356 (25.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStatin use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1,097 (87.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e625 (56.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1349 (66.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8,345 (62.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEzetimibe use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e585 (46.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e302 (27.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e617 (30.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3,440 (25.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePCSK9 inhibitor use\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11 (0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5 (0.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13 (0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eData are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation or number (percentage). \u003cem\u003ep\u003c/em\u003e-values were calculated using one-way ANOVA or Kruskal-Wallis tests for continuous variables and chi-square or Fisher\u0026rsquo;s exact tests for categorical variables. Group 1 and Group 2 were identified through k-means clustering of patients with Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL. Group 3 includes patients with Lp(a) levels between 30 and 50 mg/dL, and Group 4 includes patients with Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30 mg/dL. Lp(a), lipoprotein(a); LDL, WBC, white blood cell; eGFR, estimated glomerular filtration rate; AST, aspartate aminotransferase; ALT, alanine aminotransferase; LDL, low-density lipoprotein; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; PT-INR, prothrombin time-international normalized ratio; aPTT, activated partial thromboplastin time.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eGroup 1 (n\u0026thinsp;=\u0026thinsp;1,249) had the highest mean age (67.3\u0026thinsp;\u0026plusmn;\u0026thinsp;11.1 years), a predominance of males (60.8%), and the highest prevalence of cardiometabolic comorbidities, including hypertension (91.7%), diabetes mellitus (62.0%), dyslipidemia (97.1%), and chronic kidney disease (15.9%). Renal function was reduced (eGFR 82.0\u0026thinsp;\u0026plusmn;\u0026thinsp;23.2 mL/min/1.73m\u003csup\u003e2\u003c/sup\u003e), and lipid levels were the lowest among all groups (total cholesterol 145.5\u0026thinsp;\u0026plusmn;\u0026thinsp;30.9 mg/dL, LDL-cholesterol 77.5\u0026thinsp;\u0026plusmn;\u0026thinsp;25.1 mg/dL), reflecting a heavily treated high-risk phenotype, supported by high statin (87.8%) and antiplatelet (62.6%) use.\u003c/p\u003e \u003cp\u003eGroup 2 (n\u0026thinsp;=\u0026thinsp;1,106) was the youngest group (59.4\u0026thinsp;\u0026plusmn;\u0026thinsp;12.9 years), predominantly female (68.0%), with the lowest prevalence of hypertension (54.0%), diabetes mellitus (17.4%), and chronic kidney disease (1.4%). This group exhibited the highest lipid levels (total cholesterol 191.9\u0026thinsp;\u0026plusmn;\u0026thinsp;37.2 mg/dL, LDL-cholesterol 113.4\u0026thinsp;\u0026plusmn;\u0026thinsp;34.4 mg/dL, HDL-cholesterol 59.1\u0026thinsp;\u0026plusmn;\u0026thinsp;12.3 mg/dL) and lower statin (56.5%) and antiplatelet (23.7%) use, consistent with a relatively drug-na\u0026iuml;ve dyslipidemia phenotype.\u003c/p\u003e \u003cp\u003eGroup 3 (n\u0026thinsp;=\u0026thinsp;2,029), defined by Lp(a) levels of 30\u0026ndash;50 mg/dL, showed intermediate clinical characteristics, with a mean age of 62.5\u0026thinsp;\u0026plusmn;\u0026thinsp;13.3 years, moderate prevalence of hypertension (69.4%), diabetes mellitus (37.1%), and dyslipidemia (80.7%), and lipid levels between those of Groups 1 and 2. Group 4 (n\u0026thinsp;=\u0026thinsp;13,269), defined by Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30 mg/dL, represented the majority of the cohort and demonstrated similar intermediate characteristics; mean age 60.4\u0026thinsp;\u0026plusmn;\u0026thinsp;14.5 years, hypertension 68.8%, diabetes mellitus 38.2%, and total cholesterol 167.8\u0026thinsp;\u0026plusmn;\u0026thinsp;42.2 mg/dL. Renal function was well preserved (eGFR 92.7\u0026thinsp;\u0026plusmn;\u0026thinsp;19.0 mL/min/1.73m\u003csup\u003e2\u003c/sup\u003e).\u003c/p\u003e \u003cp\u003ePost hoc analyses (Bonferroni-adjusted Dunn test) confirmed that most variables, including age, body mass index, comorbidities, lipid profiles, glucose, and renal function, differed significantly between groups. These findings confirm that the k-means derived high Lp(a) clusters represent distinct clinical phenotypes: Cluster 1 is characterized by older age, increased cardiometabolic disease burden, polypharmacy, and lower renal function, whereas Cluster 2 is predominantly a young, female, relatively drug-na\u0026iuml;ve dyslipidemia group.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eClinical outcomes according to cluster and Lp(a) level\u003c/h2\u003e \u003cp\u003eDuring the follow-up period (725.7\u0026thinsp;\u0026plusmn;\u0026thinsp;386.8 days [median: 857 days]), the incidence of MACE differed significantly across the four groups (overall log-rank \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Group 1(Cluster 1, Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL) exhibited the highest cumulative MACE incidence (8.9%), followed by Group 3 (Lp(a) 30\u0026ndash;50 mg/dL, 4.8%), Group 4 (Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30 mg/dL, 3.7%), and Group 2 (Cluster 2, Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL, 2.0%). Pairwise log-rank testing confirmed that all inter-group comparisons were statistically significant, with the largest hazard separation observed between Group 1 (Cluster 1) and Group 2 (Cluster 2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFor secondary outcomes, significant overall differences between groups were observed for cardiovascular death (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.01), new-onset myocardial infarction (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), new-onset stroke (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.02), new-onset heart failure hospitalization (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), coronary revascularization (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), and all-cause death (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Supplemental Fig.\u0026nbsp;2). The incidence of new-onset myocardial infarction was highest in Group 1 (2.4%), followed by Group 3 (1.7%), Group 4 (0.7%), and Group 2 (0.2%). New-onset heart failure hospitalization occurred most frequently in Group 1 (1.9%), with marked lower rates in the other groups (\u0026le;\u0026thinsp;0.8%). Similarly, coronary revascularization was performed most often in Group 1 (5.2%), compared with 2.5% in Group 3, 1.8% in Group 4, and 0.9% in Group 2. All-cause mortality was significantly higher in Group 1 (1.3%) compared with the other groups (\u0026le;\u0026thinsp;0.8%), with the largest difference observed between Group 1 and Group 4 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01). For cardiovascular death, although the absolute number of events was low across all groups, the incidence was highest in Group 1 (0.2%) and lowest in Group 2 (0%).\u003c/p\u003e \u003cp\u003eOverall, these findings indicate that, among patients with high Lp(a), those classified as Group 1 (Cluster 1)\u0026mdash;characterized by older age, higher prevalence of comorbidities, and greater medication use\u0026mdash;exhibited substantially higher risks of adverse cardiovascular outcomes compared to both Group 2 (Cluster 1) and the other low Lp(a) groups (Group 3 and 4). Notably, Group 2 (Cluster 2) patients (younger, relatively drug-na\u0026iuml;ve dyslipidemia phenotype) had lowest event rates.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMultivariable Cox Proportional Hazard Analysis\u003c/h2\u003e \u003cp\u003eIn the multivariable Cox proportional hazards regression model, several clinical characteristics were proposed as independent risk predictors of 3-year MACE (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Older age (HR [95%CI] 1.03 [1.03\u0026ndash;1.04]), male sex (HR [95%CI] 1.42 [1.19\u0026ndash;1.70]), smoking (HR [95%CI] 1.45 [1.10\u0026ndash;1.91]), alcohol consumption (HR [95%CI] 4.02 [2.95\u0026ndash;5.49]), and low socio-economic status (HR [95%CI] 1.90 [1.52\u0026ndash;2.39]) were each associated with higher 3-year MACE risk (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01 for all). Hypertension (HR [95%CI] 1.44 [1.15\u0026ndash;1.81]), dyslipidemia (HR [95%CI] 1.86 [1.36\u0026ndash;2.54]), higher LDL-cholesterol (HR [95%CI] 1.01 [1.00-1.01]), elevated hs-CRP (HR [95%CI] 1.06 [1.02\u0026ndash;1.09]), and higher creatinine (HR [95%CI] 1.42 [1.04\u0026ndash;1.94]) were also significant predictors.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMultivariable Cox Proportional Hazards Regression for 3-Year MACE\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHR (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (per year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.034 (1.027\u0026ndash;1.042)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale sex (vs. female)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.418 (1.186\u0026ndash;1.696)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmoking\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.451 (1.101\u0026ndash;1.912)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAlcohol consumption\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.023 (2.949\u0026ndash;5.488)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow socio-economic status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.903 (1.518\u0026ndash;2.386)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody mass index (per kg/m\u0026sup2;)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.980 (0.958\u0026ndash;1.003)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.440 (1.148\u0026ndash;1.805)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes mellitus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.098 (0.920\u0026ndash;1.309)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.30\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDyslipidemia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.860 (1.363\u0026ndash;2.538)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChronic kidney disease\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.120 (0.814\u0026ndash;1.541)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal cholesterol (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.999 (0.995\u0026ndash;1.003)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.72\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLDL-cholesterol (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.005 (1.001\u0026ndash;1.010)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHDL-cholesterol (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.995 (0.988\u0026ndash;1.003)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglyceride (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.001 (1.000\u0026ndash;1.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlucose (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.002 (0.999\u0026ndash;1.006)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCreatinine (per mg/dL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.422 (1.040\u0026ndash;1.945)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.03\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ehs-CRP (per mg/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.056 (1.019\u0026ndash;1.095)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup 1 (vs Group 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.391 (1.127\u0026ndash;1.718)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup 2 (vs Group 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.078 (0.692\u0026ndash;1.679)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.74\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGroup 3 (vs Group 4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.182 (0.948\u0026ndash;1.472)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eHazards ratios (HR) and 95% confidence intervals (CI) were derived from pooled Cox proportional hazards regression across multiple imputed datasets (Rubin\u0026rsquo;s rules).\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eWhen stratified by cluster phenotype, using Group 4 as the reference, Group 1 (Cluster 1) patients had a significantly higher hazard (HR [95%CI] 1.39 [1.13\u0026ndash;1.72]; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01). whereas Group 2 (Cluster 2) and 3 did not differ significantly from Group 4. These findings suggest that the phenotypic grouping captured by the clustering methodology has prognostic relevance, driven largely by the Cluster 1 profile (Group 1). Moreover, despite the markedly elevated Lp(a) levels (\u0026ge;\u0026thinsp;50 mg/dL) in Group 2 (Cluster 2), the risk of MACE did not differ significantly from the low-Lp(a) (\u0026lt;\u0026thinsp;30 mg/dL) reference Group 4 (HR [95%CI] 1.08 [0.69\u0026ndash;1.68]; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.74), suggesting that even within the high-Lp(a) population, a clinically distinct low-risk subgroup may be identifiable.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn a large, contemporary cohort, we applied unsupervised k-means clustering to patients with elevated Lp(a) (\u0026ge;\u0026thinsp;50 mg/dL) and showed two reproducible phenogroups with distinct clinical profiles and outcomes. Cluster 1 was characterized by older age, male predominance, high prevalence of cardiometabolic comorbidities; Cluster 2, included younger, predominantly female patients with a relatively treatment-na\u0026iuml;ve dyslipidemia profile. Despite comparable Lp(a) levels, 3-year MACE rates differed markedly (8.9% vs 2.0%, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01). In multivariable Cox models, Cluster 1 independently predicted higher MACE risk compared with the low-Lp(a) reference group, while Cluster 2 showed risk similar to the reference. These findings support the prognostic value of phenotype-based clustering beyond concentration cut-offs and reveal clinically meaningful heterogeneity within the high-Lp(a) population. Importantly, the elevated risk observed in Cluster 1 likely reflects the combined influence of multiple cardiometabolic risk factors rather than the isolated effect of Lp(a) concentration alone. This observation underscores a key concept in Lp(a)-related cardiovascular risk: the clinical impact of elevated Lp(a) may depend substantially on the surrounding biological and clinical context. In this framework, Lp(a) may act as a risk amplifier within high-risk phenotypes rather than functioning as a uniform risk determinant across all individuals with elevated levels.\u003c/p\u003e \u003cp\u003eLp(a) has emerged as a major residual cardiovascular risk factor and is designated a \u0026ldquo;risk enhancer\u0026rdquo; in current guideline \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. The Copenhagen City Heart Study reported a 2.6-fold increase in myocardial infarction among individuals with Lp(a) \u0026gt;95th percentile compared to those \u0026lt;\u0026thinsp;22nd percentile \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. The 2010 European Atherosclerosis Society consensus proposed a 50 mg/dL threshold, aligned with approximately the 80th percentile of Danish distribution \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Analyses from the UK Biobank further demonstrated a continuous, log-linear relationship between Lp(a) concentrations and atherosclerotic cardiovascular disease risk, reinforcing its dose-response association \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Translational evidence from the FOURIER trial showed greater absolute event reduction with PCSK9 inhibition among patients with above-median Lp(a) despite optimal statin therapy \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Nonetheless, the present study indicates that a single fixed threshold (\u0026ge;\u0026thinsp;50 mg/dL) does not uniformly convey equivalent risk, highlighting the need to contextualize Lp(a) within broader biological and clinical phenotypes.\u003c/p\u003e \u003cp\u003eEmerging evidence suggests that the pathogenicity of Lp(a) is not determined by plasma concentration alone but is modulated by the surrounding milieu. First, apo(a) isoform size appears to influence risk independently of concentration, with smaller isoform conferring higher event rates\u0026mdash;potentially via higher oxidized phospholipid burden \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Apo(a) size has been also associated to attenuated Lp(a) reduction with PCSK9 inhibition, suggesting biological subgroup heterogeneity \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Second, the cholesterol cargo of Lp(a) can inflate calculated LDL-cholesterol; accounting for Lp(a)-cholesterol has been proposed to mitigate misclassification \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Third, inflammatory status further modifies Lp(a)-related risk\u0026mdash;e.g., in familial hypercholesterolemia, Lp(a) predicted adverse events primarily in the presence of elevated C-reactive protein \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. These mechanisms align with our observation that the comorbidity- and inflammation-enriched phenotype (Cluster 1) concentrated on events, whereas the relatively favorable phenotype (Cluster 2) did not, despite equally high Lp(a). Recent longitudinal data also show that Lp(a) changes track with clinical and laboratory profiles, further supporting a context-dependent risk model \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eUnsupervised clustering approaches have gained interest in lipidology and cardiovascular risk stratification \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. In Lp(a) research, prior studies have primarily demonstrated cross-sectional associations or surrogate markers. Saraiva et al. stratified 661 patients by Lp(a) and observed higher cardiovascular event rates in the high-Lp(a) cluster \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. The RED-CARPET analysis identified four high-Lp(a) phenogroups, with the \u0026ldquo;male/unhealthy lifestyle\u0026rdquo; cluster showing the most pronounced atherosclerotic cardiovascular risk association \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Extending these observations, the present study integrates clustering with time-to-event analysis in a large real-world cohort, thereby demonstrating the independent prognostic significance of phenotype-based stratification among patients with elevated Lp(a).\u003c/p\u003e \u003cp\u003eThese data have distinct clinical implications. A threshold-only strategy risks over-treating biologically low-risk individuals (e.g., Cluster 2) and under-recognizing high-risk phenotypes (e.g., Cluster 1). Phenotype-guided strategies\u0026mdash;incorporating inflammation (hs-CRP), apoB/LDL-cholesterol (with consideration of Lp(a)-cholesterol), renal function, and lifestyle factors\u0026mdash;may more precisely target intensified therapy, including PCSK9 inhibitors and, prospectively, Lp(a)-targeted agents. Thus, even for a causal risk factor, the observed clinical effect size is contingent on patient context, supporting a shift from population-based to milieu-adjusted risk assessment.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, the retrospective design using electronic health record data may introduce residual confounding and selection bias despite multivariable adjustment. In particular, Lp(a) testing was performed in routine clinical practice rather than through systematic screening, which may have resulted in a cohort enriched for patients with higher baseline cardiovascular risk. Second, clustering analyses are inherently data-driven and may be influenced by variable selection and data structure. Although multiple validation approaches were applied, external validation in independent cohorts will be necessary to confirm the reproducibility of these phenotypes. Third, the study population was derived from tertiary academic centers in Korea, which may limit the generalizability of findings to other healthcare settings or ethnic populations.\u003c/p\u003e \u003cp\u003eIn conclusion, phenotypic clustering of patients with elevated Lp(a) reveals distinct, prognostically relevant subgroups and supports integrating multidimensional patient profiling\u0026mdash;beyond Lp(a) concentration alone\u0026mdash;into cardiovascular risk assessment and therapeutic decision-making.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eH.J.J. and S.Y.S. conceived the study. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eH.J.J., S.J.H., and C.W.Y. collected and curated the data. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eH.J.J. performed statistical analyses. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eH.J.J. drafted the manuscript. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eE.J.K. and S.Y.S. critically revised the manuscript. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are available from the corresponding author upon reasonable request. Due to institutional regulations and patient privacy protection, the raw data cannot be publicly shared.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe sincerely acknowledge the invaluable support and contributions of the Cardiovascular Data Science Research Group of the Department of Cardiology at Korea University. Their dedication and expertise were instrumental in the successful completion of this study.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eSources of Funding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was supported from the Medical data-driven hospital support project through the Korea Health Information Service(KHIS), funded by the Ministry of Health \u0026amp; Welfare, and the Ministry of Science and ICT, Korea, under the ICT Challenge and Advanced Network of HRD program (IITP-2025-RS-2022-00156439) supervised by the Institute of Information \u0026amp; Communications Technology Planning \u0026amp; Evaluation.\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eDoherty, S. et al. Lipoprotein(a) as a Causal Risk Factor for Cardiovascular Disease. \u003cem\u003eCurr. Cardiovasc. Risk Rep.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 8 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEmerging Risk Factors Collaboration. Lipoprotein(a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. \u003cem\u003eJAMA\u003c/em\u003e \u003cb\u003e302\u003c/b\u003e, 412\u0026ndash;423 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNordestgaard, B. G. et al. Lipoprotein(a) as a cardiovascular risk factor: current status. \u003cem\u003eEur. Heart J.\u003c/em\u003e \u003cb\u003e31\u003c/b\u003e, 2844\u0026ndash;2853 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, S. et al. Phenomapping of subgroups in high-Lp(a) patients: a data-driven cluster analysis in RED-CARPET study. \u003cem\u003eClin Res. Cardiol\u003c/em\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRichards, A. M. Biomarkers in Acute Heart Failure - Cardiac And Kidney. \u003cem\u003eCard Fail. Rev.\u003c/em\u003e \u003cb\u003e1\u003c/b\u003e, 107\u0026ndash;111 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e\u0026Ccedil;elik, A. et al. How to Use Natriuretic Peptides in Patients with Heart Failure with Non-Reduced Ejection Fraction? \u003cem\u003eAnatol. J. Cardiol.\u003c/em\u003e \u003cb\u003e27\u003c/b\u003e, 308\u0026ndash;318 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKronenberg, F. et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. \u003cem\u003eEur. Heart J.\u003c/em\u003e \u003cb\u003e43\u003c/b\u003e, 3925\u0026ndash;3946 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKamstrup, P. R., Tybjaerg-Hansen, A., Steffensen, R. \u0026amp; Nordestgaard, B. G. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. \u003cem\u003eJAMA\u003c/em\u003e \u003cb\u003e301\u003c/b\u003e, 2331\u0026ndash;2339 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel, A. P. et al. Lp(a) (Lipoprotein[a]) Concentrations and Incident Atherosclerotic Cardiovascular Disease: New Insights From a Large National Biobank. \u003cem\u003eArterioscler. Thromb. Vasc Biol.\u003c/em\u003e \u003cb\u003e41\u003c/b\u003e, 465\u0026ndash;474 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO'Donoghue, M. L. et al. Lipoprotein(a), PCSK9 Inhibition, and Cardiovascular Risk. \u003cem\u003eCirculation\u003c/em\u003e \u003cb\u003e139\u003c/b\u003e, 1483\u0026ndash;1492 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaleheen, D. et al. Apolipoprotein(a) isoform size, lipoprotein(a) concentration, and coronary artery disease: a mendelian randomisation analysis. \u003cem\u003eLancet Diabetes Endocrinol.\u003c/em\u003e \u003cb\u003e5\u003c/b\u003e, 524\u0026ndash;533 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlanchard, V. et al. The size of apolipoprotein (a) is an independent determinant of the reduction in lipoprotein (a) induced by PCSK9 inhibitors. \u003cem\u003eCardiovasc. Res.\u003c/em\u003e \u003cb\u003e118\u003c/b\u003e, 2103\u0026ndash;2111 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilleit, P. et al. Low-Density Lipoprotein Cholesterol Corrected for Lipoprotein(a) Cholesterol, Risk Thresholds, and Cardiovascular Events. \u003cem\u003eJ. Am. Heart Assoc.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e, e016318 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTada, H. et al. Synergistic effect of lipoprotein (a) and C-reactive protein on prognosis of familial hypercholesterolemia. \u003cem\u003eAm. J. Prev. Cardiol.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 100428 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoo, H. J. et al. Predictors of lipoprotein(a) variability in clinical practice and their impact on cardiovascular risk. \u003cem\u003eLipids Health Dis.\u003c/em\u003e \u003cb\u003e24\u003c/b\u003e, 250 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCho, A. R., Heo, S. J., Han, T. \u0026amp; Kwon, Y. J. Cluster-Based Analysis of Lipid Profiles and Inflammation in Association With Cardiovascular Disease Incidence and Mortality: A 17.5-Year Longitudinal Study. \u003cem\u003eJ. Clin. Hypertens. (Greenwich)\u003c/em\u003e. \u003cb\u003e27\u003c/b\u003e, e70035 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSharma, A. et al. Cluster Analysis of Cardiovascular Phenotypes in Patients With Type 2 Diabetes and Established Atherosclerotic Cardiovascular Disease: A Potential Approach to Precision Medicine. \u003cem\u003eDiabetes Care\u003c/em\u003e. \u003cb\u003e45\u003c/b\u003e, 204\u0026ndash;212 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaraiva, M. et al. Unveiling patient profiles associated with elevated Lp(a) through an unbiased clustering analysis. \u003cem\u003eFront. Cardiovasc. Med.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 1546351 (2025).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Lipoprotein(a), Cardiovascular Disease, Cluster Analysis, Phenotype, Risk Assessment","lastPublishedDoi":"10.21203/rs.3.rs-9065291/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9065291/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLipoprotein(a) [Lp(a)] is an established cardiovascular risk enhancer, yet fixed concentration thresholds may not fully capture the heterogeneity of cardiovascular risk among individuals with elevated levels. We retrospectively analyzed 17,653 patients with Lp(a) measurements from three tertiary hospitals (2017\u0026ndash;2024). After exclusions, 4,320 patients with Lp(a)\u0026thinsp;\u0026ge;\u0026thinsp;50 mg/dL underwent k-means clustering identified two phenotypic groups based on demographic, comorbidity, laboratory, and medication variables. Cluster validation using elbow, shilhouette, and NbClust consensus methods supported a two-cluster solution. Cluster 1 consisted of older, male-predominant patients with a higher cardiometabolic burden and lower renal function, whereas Cluster 2 included younger, female-predominant patients with fewer comorbidities and relatively treatment-na\u0026iuml;ve dyslipidemia. Despite similar Lp(a) levels, 3-year major adverse cardiovascular events (MACE) occurred more frequently in Cluster 1 than Cluster 2 (8.9% vs 2.0%, log-rank p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). In multivariable Cox models, Cluster 1 was associated with higher MACE risk compared with the Lp(a)\u0026thinsp;\u0026lt;\u0026thinsp;30 mg/dL reference group (HR 1.39, 95% CI, 1.13\u0026ndash;1.72), whereas Cluster 2 showed no significant risk difference (HR 1.08 95% CI 0.69\u0026ndash;1.68). These findings suggest that phenotypic clustering of high-Lp(a) patients delineates subgroups with distinct cardiovascular risk profiles. Incorporating phenotype-guided risk assessment may refine cardiovascular risk stratification beyond fixed Lp(a) thresholds.\u003c/p\u003e","manuscriptTitle":"Phenotypic Clustering Identifies Heterogeneous Cardiovascular Risk Among Patients with Elevated Lipoprotein(a)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-19 15:59:31","doi":"10.21203/rs.3.rs-9065291/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-04-06T10:46:25+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-06T03:54:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-03T11:38:48+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-31T09:57:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"227511677466155798684950317461369612679","date":"2026-03-27T08:21:07+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-25T04:42:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"237135271011280379204994251649678510225","date":"2026-03-24T14:22:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"217629791236059293002463820046401945400","date":"2026-03-23T05:41:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"140642388447385975549146488090797403338","date":"2026-03-22T16:58:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"296614077378466034028232291051856661788","date":"2026-03-22T15:48:06+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-16T08:01:34+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-16T05:46:54+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-09T11:37:41+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-09T11:36:50+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-03-08T15:26:46+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f862d179-2d97-4b10-b321-b6cbcd04f3e6","owner":[],"postedDate":"March 19th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":64606284,"name":"Health sciences/Biomarkers"},{"id":64606285,"name":"Health sciences/Cardiology"},{"id":64606286,"name":"Health sciences/Diseases"},{"id":64606288,"name":"Health sciences/Medical research"},{"id":64606290,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2026-04-29T10:53:20+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-19 15:59:31","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9065291","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9065291","identity":"rs-9065291","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}