Abstract
Objective
We examined the distribution of low-density lipoprotein cholesterol (LDL-C) levels and compared the hazards of myocardial infarction (MI), stroke, and their composite according to LDL-C concentration in young adults.
Methods
This was a nationwide, population-based study. From the Korean National Health Insurance Service database (2002–2022), we identified adults aged 20–39 years without prior MI or stroke who underwent at least 1 health examination between 2009 and 2012 (n=6,458,006). The hazards of MI, stroke, and their composite were evaluated across LDL-C levels.
Results
During a median follow-up of 12.6 years, 47,107 MIs, 28,536 strokes, and 73,897 composite events occurred. The hazard of outcomes showed a J-shaped association with LDL-C levels, with the lowest hazard observed around 90–110 mg/dL. Above this range, the hazard increased progressively with rising LDL-C levels. In analyses based on LDL-C categories, levels ≥130 mg/dL were associated with significantly increased hazards of outcomes, with the highest hazards observed at ≥160 mg/dL. The elevated hazards in the higher LDL-C categories were more pronounced in individuals with abdominal obesity and current smokers for all 3 outcomes, and in those with higher body mass index categories, low high-density lipoprotein-cholesterol, and hypertriglyceridemia for MI and the composite outcome.
Conclusion
Among young adults, cardiovascular risk demonstrated a J-shaped association with LDL-C levels. Risks increased progressively beyond approximately 90–110 mg/dL, especially in the presence of other cardiometabolic risk factors. These findings underscore the importance of lipid monitoring and cardiovascular risk assessment even in young adults, particularly those with concurrent risk factors.
Keywords: Young adult, Low-density lipoprotein cholesterol, Myocardial infarction, Stroke, Cardiovascular disease
INTRODUCTION
Elevated low-density lipoprotein cholesterol (LDL-C) is a well-established contributor to the development of atherosclerotic cardiovascular disease (ASCVD), and lipid-lowering therapy, including statin treatment, has been shown to significantly reduce ASCVD risk.1,2 Evidence from randomized trials and meta-analyses consistently demonstrates that lowering LDL-C proportionally decreases ASCVD events and cardiovascular mortality.1,2 Accordingly, contemporary guidelines emphasize intensive LDL-C reduction, particularly for middle-aged or older adults with high ASCVD risk or established ASCVD.3,4
However, most primary prevention strategies targeting LDL-C have been derived from studies focused on individuals aged 40–75 years. As a result, there is limited evidence and no clear consensus regarding optimal LDL-C management in young adults aged <40 years, especially those with moderately elevated LDL-C levels (e.g., 130–190 mg/dL) who do not have familial hypercholesterolemia (FH) or diagnosed ASCVD.5 For this younger population, particularly individuals without FH or diabetes, substantial uncertainty remains regarding whether and when to initiate lipid-lowering therapy.
Despite this, dyslipidemia is not uncommon among young adults. U.S. data from 2015–2018 indicate that 8.5% of men and 4.9% of women aged 20–34 years had total cholesterol ≥240 mg/dL or were receiving lipid-lowering therapy, with these figures increasing to 18.2% in men and 8.9% in women aged 35–44 years.6 Similarly, analysis of the 2016–2022 Korea National Health and Nutrition Examination Survey (KNHANES) showed that the prevalence of hyper-LDL-cholesterolemia—defined as LDL-C ≥160 mg/dL or use of lipid-lowering medication—was 6.9% in men and 5.0% in women in their 20s, rising to 13.3% and 8.2%, respectively, in their 30s.7 Considering that the age-standardized prevalence of hypercholesterolemia (defined as total cholesterol ≥240 mg/dL or lipid-lowering therapy) in Korea has increased by more than 2.5-fold from 2007 to 2022,7 the disease burden attributable to elevated LDL-C among young adults is expected to continue increasing.
The absence of tailored treatment recommendations for young adults with moderately elevated LDL-C may contribute to the underutilization of lipid-lowering agents, even among those at elevated cardiovascular risk. A recent observational study examining individuals with first myocardial infarction (MI) reported that fewer than half of patients under 55 years had met eligibility criteria for statin therapy according to the 2018 American Heart Association/American College of Cardiology guidelines before their event.8 Given their longer life expectancy and the societal implications of early cardiovascular events, identifying LDL-C thresholds that best support risk mitigation and guide preventive strategies in this age group is critically important.
Therefore, we examined the distribution of LDL-C levels and compared the hazards of incident MI, stroke, and their composite (MI or stroke) across LDL-C concentrations in young adults (aged 20–39 years) without baseline MI and/or stroke.
MATERIALS AND METHODS
1. Data sources
For this nationwide, longitudinal, population-based cohort study, we used data from the Korean National Health Insurance Service (KNHIS) from January 2002 to December 2022. The KNHIS, operated by the Korean government, is the single nationwide insurer that provides universal health coverage to all residents.9,10 The KNHIS maintains a comprehensive database that integrates information on healthcare utilization, health-screening results, and eligibility data.11 This database includes demographic characteristics, monthly insurance premiums (used as a proxy for household income level), residential area, diagnostic codes based on the International Classification of Diseases, 10th Revision (ICD-10), prescription records, medical procedures, outpatient visits, hospital admissions for the entire Korean population, and dates of death. The KNHIS also operates a standardized national health screening program, recommending examinations at least every 2 years. These screenings, performed exclusively at KNHIS-certified hospitals, include information on smoking status, alcohol consumption, physical activity, anthropometric measurements (height, weight, and waist circumference [WC]), blood pressure (BP), fasting plasma glucose (FPG), lipid profiles, liver function tests, and estimated glomerular filtration rate. Additional details on the KNHIS database have been described in prior publications.9,10,11
The Institutional Review Board (IRB) of Samsung Medical Center approved this study (SMC 2024-05-043). The IRB granted an informed consent exemption because researchers received anonymous, de-identified KNHIS data with all personal identifiers removed.
2. Study cohort
We identified individuals aged 20–39 years who underwent at least 1 health examination between 2009 and 2012. Baseline was defined as the date of the health examination within this period. We excluded individuals with baseline triglyceride levels >400 mg/dL, those with missing data in at least 1 variable, and those who had ICD-10 code claims for MI (I21–I22) and/or stroke (I63–I64) at or before baseline. We also excluded individuals who died or developed outcomes (MI or stroke) within 1 year after baseline data collection (Supplementary Fig. 1).
3. Outcomes and follow-up
The outcomes of interest were incident MI, stroke, and their composite (MI or stroke). MI was defined as at least 1 hospitalization claim with ICD-10 codes I21–I22 or at least 2 outpatient claims with the same codes.12,13,14 Stroke was defined using ICD-10 codes I63–I64 during hospitalization and required accompanying claims for brain magnetic resonance imaging or computed tomography.13,14 The study population was followed from baseline until the date of death, occurrence of the outcome event, or December 31, 2022, whichever occurred first.
4. Measurements and definitions
Information on smoking status, alcohol intake, and physical activity was obtained from self-administered questionnaires. Definitions for alcohol use, regular exercise, low-income state, and the presence of diabetes mellitus (DM), hypertension, chronic kidney disease (CKD), and body mass index (BMI) are provided in Supplementary Table 1 with corresponding references. Concomitant medications were identified based on prescription records at baseline and during the year preceding baseline. The Charlson Comorbidity Index (CCI) was calculated using a previously validated approach,15 based on diagnostic codes described in an earlier study.16 Blood samples were collected after an overnight fast. LDL-C levels were estimated using the Friedewald formula,17 based on total cholesterol, high-density lipoprotein cholesterol (HDL-C), and triglyceride concentrations.
5. Statistical analyses
All analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC, USA). Statistical significance was defined as a 2-sided p-value <0.05. Participants were classified into 6 LDL-C categories: <55, 55–69, 70–99, 100–129, 130–159, and ≥160 mg/dL, based on treatment targets recommended in previous guidelines.4,18,19 Baseline characteristics were evaluated across the 6 LDL-C categories. Continuous variables were expressed as mean ± standard deviation (SD) for normally distributed data or as geometric means with 95% confidence intervals (CIs) for skewed distributions. Categorical variables were summarized as counts and percentages. Outcome incidence rates were calculated by dividing the number of events by the total follow-up duration in person-years. Cumulative incidence rates across LDL-C groups were assessed using Kaplan–Meier curves, with differences evaluated via the log-rank test. Hazard ratios (HRs) and 95% CIs for outcome incidence were estimated according to the 6 LDL-C categories. In addition to the crude (unadjusted) model, 4 adjusted models were constructed. Model 1 included age, sex, low-income state, BMI, current smoking, heavy alcohol intake, and regular exercise. Model 2 included the variables in Model 1 and DM, hypertension, and statin and fibrate use. Model 3 included the variables in Model 2 and CCI. Model 4 included the variables in Model 3 and triglyceride and HDL-C levels. Restricted cubic spline curves were generated to evaluate adjusted HRs and 95% CIs for outcomes according to LDL-C levels in the total population and in statin users and non-users.
We also compared HRs (95% CIs) for outcome incidence across the 6 LDL-C categories within several subgroups, and calculated p-values for interaction. Subgroups were defined by sex, age group (20–29 or 30–39 years), presence of DM, hypertension, CKD, abdominal obesity, BMI categories (<23.0, 23.0–24.9, or ≥25.0 kg/m2), low HDL-C (<40 mg/dL in men and <50 mg/dL in women), hypertriglyceridemia (triglyceride ≥150 mg/dL or fibrate use), statin use, current smoking, heavy alcohol consumption, and regular exercise.
6. Sensitivity analyses
To account for competing risks from all-cause mortality, we estimated sub-distribution HRs according to the 6 LDL-C categories using the Fine and Gray method,20 adjusting for the same potential confounders included in the main analyses.
RESULTS
1. Distribution of LDL-C levels and baseline characteristics
A total of 6,458,006 individuals aged 20–39 years were included (Supplementary Fig. 1). Among them, 1,217,491 (18.85%) had LDL-C ≥130 mg/dL, and 289,339 (4.48%) had LDL-C ≥160 mg/dL (Fig. 1, Table 1). LDL-C levels were <70 mg/dL in 644,269 (9.98%) participants, and only 177,189 (2.74%) had LDL-C <55 mg/dL.
Fig. 1. Distribution of LDL-C levels among the study population of young adults.
LDL-C, low-density lipoprotein cholesterol.
Table 1. Baseline characteristics according to the ranges of LDL-C levels.
| Variables | Total (n=6,458,006) | Ranges of LDL-C (mg/dL) | p-value | ||||||
|---|---|---|---|---|---|---|---|---|---|
| <55 (n=177,189) | ≥55 and <70 (n=467,080) | ≥70 and <100 (n=2,344,925) | ≥100 and <130 (n=2,251,321) | ≥130 and <160 (n=928,152) | ≥160 (n=289,339) | ||||
| LDL-C (mg/dL) | 105.01±32.76 | 44.60±9.81 | 63.25±4.17 | 86.00±8.37 | 113.11±8.46 | 141.58±8.30 | 183.23±60.18 | <0.0001 | |
| Age (yr) | 30.84±4.99 | 29.73±5.16 | 29.39±5.09 | 30.05±5.03 | 31.20±4.87 | 32.25±4.6 | 32.90±4.37 | <0.0001 | |
| Age <30 yr | 2,730,756 (42.28) | 91,852 (51.84) | 255,217 (54.64) | 1,153,743 (49.20) | 883,666 (39.25) | 276,775 (29.82) | 69,503 (24.02) | <0.0001 | |
| Sex (male) | 3,819,888 (59.15) | 97,899 (55.25) | 233,982 (50.09) | 1,217,063 (51.90) | 1,378,518 (61.23) | 667,145 (71.88) | 225,281 (77.86) | <0.0001 | |
| Smoking history | <0.0001 | ||||||||
| Never-smoker | 3,559,671 (55.12) | 91,692 (51.75) | 272,508 (58.34) | 1,396,181 (59.54) | 1,230,435 (54.65) | 443,410 (47.77) | 125,445 (43.36) | ||
| Past-smoker | 667,230 (10.33) | 14,535 (8.20) | 37,466 (8.02) | 208,471 (8.89) | 245,526 (10.91) | 120,560 (12.99) | 40,672 (14.06) | ||
| Current-smoker | 2,231,105 (34.55) | 70,962 (40.05) | 157,106 (33.64) | 740,273 (31.57) | 775,360 (34.44) | 364,182 (39.24) | 123,222 (42.59) | ||
| Alcohol consumption | <0.0001 | ||||||||
| Nondrinker | 2,445,449 (37.87) | 56,020 (31.62) | 168,313 (36.04) | 908,559 (38.75) | 865,409 (38.44) | 341,132 (36.75) | 106,016 (36.64) | ||
| Moderate drinker | 3,384,825 (52.41) | 91,558 (51.67) | 241,701 (51.75) | 1,211,762 (51.68) | 1,184,457 (52.61) | 500,046 (53.88) | 155,301 (53.67) | ||
| Heavy drinker | 627,732 (9.72) | 29,611 (16.71) | 57,066 (12.22) | 224,604 (9.58) | 201,455 (8.95) | 86,974 (9.37) | 28,022 (9.68) | ||
| Regular exercise | 830,936 (12.87) | 22,806 (12.87) | 59,040 (12.64) | 296,923 (12.66) | 292,744 (13.00) | 121,580 (13.10) | 37,843 (13.08) | <0.0001 | |
| Low-income state | 1,374,878 (21.29) | 47,062 (26.56) | 117,885 (25.24) | 536,493 (22.88) | 456,015 (20.26) | 167,536 (18.05) | 49,887 (17.24) | <0.0001 | |
| Body weight (kg) | 65.39±13.48 | 62.50±12.85 | 61.36±12.34 | 62.67±12.71 | 66.27±13.4 | 70.33±13.56 | 73.01±13.46 | <0.0001 | |
| BMI (kg/m2) | 22.96±3.59 | 22.01±3.40 | 21.74±3.24 | 22.18±3.33 | 23.23±3.54 | 24.38±3.64 | 25.22±3.67 | <0.0001 | |
| Waist circumference (cm) | 77.41±9.96 | 75.25±9.48 | 74.19±9.15 | 75.22±9.41 | 78.10±9.83 | 81.37±9.83 | 83.68±9.69 | <0.0001 | |
| Systolic BP (mmHg) | 117.62±13.13 | 117.24±13.17 | 115.68±12.78 | 115.97±12.76 | 117.97±13.04 | 120.47±13.30 | 122.42±13.76 | <0.0001 | |
| Diastolic BP (mmHg) | 73.70±9.41 | 73.46±9.50 | 72.32±9.17 | 72.56±9.13 | 73.94±9.34 | 75.68±9.58 | 77.09±9.92 | <0.0001 | |
| FPG (mg/dL) | 90.71±15.98 | 90.20±16.96 | 89.00±14.51 | 89.37±13.93 | 90.92±15.45 | 93.04±18.67 | 95.57±23.92 | <0.0001 | |
| Total cholesterol (mg/dL) | 184.12±33.40 | 135.42±22.38 | 144.73±17.49 | 164.55±17.43 | 191.57±17.16 | 221.13±17.44 | 259.35±28.73 | <0.0001 | |
| Triglyceride (mg/dL) | 94.70 (94.66–94.75) | 103.14 (102.81–103.47) | 85.56 (85.41–85.71) | 84.82 (84.75–84.88) | 96.39 (96.32–96.46) | 112.90 (112.78–113.02) | 128.42 (128.19–128.65) | <0.0001 | |
| HDL-C (mg/dL) | 57.19±18.18 | 64.38±30.30 | 60.39±18.08 | 58.42±16.73 | 56.25±16.57 | 54.23±16.46 | 54.50±30.23 | <0.0001 | |
| AST (IU/L) | 21.43 (21.42–21.44) | 21.28 (21.24–21.32) | 20.15 (20.13–20.17) | 20.33 (20.32–20.34) | 21.58 (21.57–21.59) | 23.50 (23.48–23.52) | 25.72 (25.68–25.76) | <0.0001 | |
| ALT (IU/L) | 19.41 (19.40–19.42) | 18.06 (18.01–18.11) | 16.44 (16.41–16.46) | 17.03 (17.02–17.04) | 19.90 (19.89–19.92) | 24.54 (24.51–24.57) | 29.71 (29.64–29.79) | <0.0001 | |
| GGT (IU/L) | 23.13 (23.12–23.14) | 23.45 (23.36–23.53) | 19.97 (19.93–20.01) | 20.13 (20.12–20.15) | 23.53 (23.51–23.55) | 29.44 (29.40–29.48) | 36.17 (36.07–36.27) | <0.0001 | |
| Statin | 48,688 (0.75) | 2,781 (1.57) | 3,554 (0.76) | 11,163 (0.48) | 11,896 (0.53) | 9,993 (1.08) | 9,301 (3.21) | <0.0001 | |
| Ezetimibe | 2,265 (0.04) | 210 (0.12) | 215 (0.05) | 554 (0.02) | 479 (0.02) | 410 (0.04) | 397 (0.14) | <0.0001 | |
| Fibrate | 8,589 (0.13) | 518 (0.29) | 674 (0.14) | 2,419 (0.10) | 2,755 (0.12) | 1,547 (0.17) | 676 (0.23) | <0.0001 | |
| CKD | 145,502 (2.25) | 3,284 (1.85) | 9,625 (2.06) | 52,332 (2.23) | 52,188 (2.32) | 21,428 (2.31) | 6,645 (2.30) | <0.0001 | |
| CCI | <0.0001 | ||||||||
| 0 | 4,709,129 (72.92) | 127,716 (72.08) | 339,218 (72.63) | 1,708,848 (72.87) | 1,644,639 (73.05) | 677,853 (73.03) | 210,855 (72.87) | ||
| 1 | 1,182,606 (18.31) | 32,650 (18.43) | 85,957 (18.40) | 430,147 (18.34) | 411,167 (18.26) | 169,897 (18.30) | 52,788 (18.24) | ||
| ≥2 | 566,271 (8.77) | 16,823 (9.49) | 41,905 (8.97) | 205,930 (8.78) | 195,515 (8.68) | 80,402 (8.66) | 25,696 (8.88) | ||
| Diabetes mellitus | 117,904 (1.83) | 4,488 (2.53) | 7,313 (1.57) | 31,523 (1.34) | 38,919 (1.73) | 23,797 (2.56) | 11,864 (4.10) | <0.0001 | |
| Hypertension | 471,376 (7.30) | 14,016 (7.91) | 25,735 (5.51) | 128,562 (5.48) | 166,514 (7.40) | 96,839 (10.43) | 39,710 (13.72) | <0.0001 | |
Values are presented as number (%), mean ± standard deviation, or geometric mean (95% confidence interval).
LDL-C, low-density lipoprotein cholesterol; BMI, body mass index; BP, blood pressure; FPG, fasting plasma glucose; HDL-C, high-density lipoprotein cholesterol; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; CKD, chronic kidney disease; CCI, Charlson Comorbidity Index.
Baseline characteristics by LDL-C category are presented in Table 1. The mean age of the study population was 30.84±4.99 years, and 59.15% were men. The proportion of statin users was 0.75%. Among individuals with LDL-C ≥55 mg/dL, higher LDL-C categories showed increasing trends in the proportion of men and of individuals in their 30s, as well as in age, BMI, WC, systolic and diastolic BP, FPG, and liver enzyme levels (aspartate aminotransferase, alanine aminotransferase, and gamma-glutamyl transferase). The prevalence of hypertension and DM was numerically lower in the 70–99 mg/dL group compared with the <55 and 55–69 mg/dL groups; however, among individuals with LDL-C ≥70 mg/dL, the prevalence of these comorbidities increased with advancing LDL-C categories. In contrast, HDL-C levels showed a decreasing trend as LDL-C increased.
2. Incidence of outcomes according to LDL-C levels
During a median follow-up of 12.6 years, there were 47,107 MI cases, 28,536 stroke cases, and 73,897 composite events. The corresponding person-years of follow-up were 79,373,457.3 for MI, 79,437,225.3 for stroke, and 79,260,049.9 for the composite outcome. Kaplan–Meier curves for cumulative incidence across LDL-C categories are shown in Fig. 2. Individuals with LDL-C ≥160 mg/dL had the highest cumulative incidence of MI, stroke, and the composite outcome, followed by those with LDL-C 130–159 mg/dL.
Fig. 2. Kaplan-Meier estimates of cumulative incidence of myocardial infarction, stroke, and their composite according to the ranges of LDL-C levels (mg/dL).
LDL-C, low-density lipoprotein cholesterol.
Fully adjusted HRs (95% CIs) by LDL-C category demonstrated a J-shaped association for MI, stroke, and their composite (Table 2). Using <55 mg/dL as the reference, individuals in the 70–99 mg/dL group had significantly lower hazards for all outcomes, while those in the 55–69 mg/dL group had lower hazards for stroke and the composite outcome. Conversely, individuals with LDL-C ≥160 mg/dL had significantly higher hazards for all outcomes, and those with LDL-C 130–159 mg/dL had increased hazards for MI and the composite outcome.
Table 2. HRs and 95% CIs for the incidence of myocardial infarction, ischemic stroke, and their composite according to the ranges of baseline LDL-C.
| Variables | Ranges of LDL-C (mg/dL) | |||||||
|---|---|---|---|---|---|---|---|---|
| <55 (n=177,189) | ≥55 and <70 (n=467,080) | ≥70 and <100 (n=2,344,925) | ≥100 and <130 (n=2,251,321) | ≥130 and <160 (n=928,152) | ≥160 (n=289,339) | |||
| Composite of myocardial infarction and stroke | ||||||||
| No. of events | 2,001 | 4,354 | 22,047 | 25,071 | 14,003 | 6,421 | ||
| Follow-up duration (person-years) | 2,166,036.3 | 5,708,196.6 | 28,739,935.1 | 27,671,934.2 | 11,423,660.1 | 3,550,287.5 | ||
| Incidence rate (per 1,000 person-years) | 0.92 | 0.76 | 0.77 | 0.91 | 1.23 | 1.81 | ||
| HR (95% CI) for the incidence | ||||||||
| Crude model | 1 (Ref.) | 0.827 (0.784, 0.872) | 0.830 (0.793, 0.868) | 0.977 (0.933, 1.022) | 1.318 (1.258, 1.381) | 1.947 (1.852, 2.047) | ||
| Model 1 | 1 (Ref.) | 0.895 (0.848, 0.943) | 0.853 (0.815, 0.893) | 0.873 (0.834, 0.914) | 1.021 (0.974, 1.070) | 1.369 (1.301, 1.440) | ||
| Model 2 | 1 (Ref.) | 0.920 (0.873, 0.971) | 0.894 (0.854, 0.936) | 0.922 (0.881, 0.965) | 1.074 (1.024, 1.126) | 1.406 (1.337, 1.479) | ||
| Model 3 | 1 (Ref.) | 0.921 (0.874, 0.971) | 0.896 (0.856, 0.938) | 0.925 (0.884, 0.968) | 1.079 (1.030, 1.132) | 1.420 (1.349, 1.493) | ||
| Model 4 | 1 (Ref.) | 0.945 (0.897, 0.997) | 0.932 (0.890, 0.976) | 0.965 (0.922, 1.011) | 1.123 (1.071, 1.178) | 1.471 (1.398, 1.547) | ||
| Myocardial infarction | ||||||||
| No. of events | 1,274 | 2,807 | 13,888 | 15,710 | 9,028 | 4,400 | ||
| Follow-up duration (person-years) | 2,169,209.1 | 5,714,495.6 | 28,773,947.3 | 27,711,200.9 | 11,445,166.8 | 3,559,437.6 | ||
| Incidence rate (per 1,000 person-years) | 0.59 | 0.49 | 0.48 | 0.57 | 0.79 | 1.24 | ||
| HR (95% CI) for the incidence | ||||||||
| Crude model | 1 (Ref.) | 0.838 (0.784, 0.895) | 0.821 (0.775, 0.870) | 0.961 (0.908, 1.017) | 1.333 (1.257, 1.414) | 2.092 (1.965, 2.226) | ||
| Model 1 | 1 (Ref.) | 0.906 (0.848, 0.968) | 0.851 (0.803, 0.901) | 0.873 (0.825, 0.925) | 1.057 (0.996, 1.121) | 1.513 (1.420, 1.611) | ||
| Model 2 | 1 (Ref.) | 0.928 (0.868, 0.991) | 0.884 (0.835, 0.937) | 0.914 (0.863, 0.968) | 1.103 (1.039, 1.170) | 1.544 (1.449, 1.644) | ||
| Model 3 | 1 (Ref.) | 0.928 (0.869, 0.992) | 0.887 (0.837, 0.939) | 0.917 (0.866, 0.971) | 1.109 (1.045, 1.176) | 1.559 (1.464, 1.661) | ||
| Model 4 | 1 (Ref.) | 0.955 (0.894, 1.021) | 0.926 (0.874, 0.981) | 0.961 (0.907, 1.018) | 1.157 (1.090, 1.228) | 1.621 (1.522, 1.727) | ||
| Stroke | ||||||||
| No. of events | 790 | 1,650 | 8,639 | 9,923 | 5,335 | 2,199 | ||
| Follow-up duration (person-years) | 2,170,690.66 | 5,718,185.6 | 28,791,054.8 | 27,730,000.2 | 11,458,480.3 | 3,568,813.7 | ||
| Incidence rate (per 1,000 person-years) | 0.36 | 0.29 | 0.30 | 0.36 | 0.47 | 0.62 | ||
| HR (95% CI) for the incidence | ||||||||
| Crude model | 1 (Ref.) | 0.794 (0.730, 0.865) | 0.824 (0.766, 0.886) | 0.979 (0.911, 1.053) | 1.271 (1.179, 1.370) | 1.682 (1.550, 1.824) | ||
| Model 1 | 1 (Ref.) | 0.861 (0.791, 0.938) | 0.839 (0.780, 0.903) | 0.853 (0.793, 0.917) | 0.947 (0.878, 1.021) | 1.127 (1.038, 1.224) | ||
| Model 2 | 1 (Ref.) | 0.895 (0.822, 0.974) | 0.893 (0.830, 0.960) | 0.917 (0.852, 0.986) | 1.012 (0.939, 1.092) | 1.171 (1.079, 1.271) | ||
| Model 3 | 1 (Ref.) | 0.896 (0.823, 0.975) | 0.895 (0.832, 0.963) | 0.920 (0.855, 0.989) | 1.017 (0.943, 1.097) | 1.181 (1.088, 1.282) | ||
| Model 4 | 1 (Ref.) | 0.916 (0.841, 0.997) | 0.925 (0.859, 0.995) | 0.953 (0.886, 1.025) | 1.052 (0.975, 1.135) | 1.217 (1.121, 1.321) | ||
Crude model: unadjusted; Model 1: adjusted for age (continuous variable), sex, low-income state, body mass index, current smoking, heavy alcohol consumption, and regular exercise; Model 2: adjusted for variables in model 1 and diabetes mellitus, hypertension, and statin and fibrate use; Model 3: adjusted for variables in model 2 and Charlson Comorbidity Index; Model 4: adjusted for variables in model 3 and triglyceride and high-density lipoprotein cholesterol levels.
HR, hazard ratio; CI, confidence interval; LDL-C, low-density lipoprotein cholesterol.
Similar J-shaped associations were observed when LDL-C was modeled as a continuous variable using restricted cubic splines in the total population and in statin non-users, with the lowest hazard occurring at approximately 90–110 mg/dL (Fig. 3). Among statin users, no clear increase in hazard was seen at the lowest LDL-C levels for MI, and a J-shaped association was not apparent; instead, MI hazard rose progressively with higher LDL-C levels.
Fig. 3. Adjusted HRs (95% CIs) for incident myocardial infarction, stroke, and their composite according to the LDL-C levels. (A) Total population, (B) statin non-users, and (C) statin users.
Curves represent HRs adjusted for age, sex, low-income state, body mass index, current smoking, heavy alcohol intake, regular exercise, diabetes mellitus, hypertension, and statin and fibrate use. Solid lines represent hazard ratios estimated using restricted cubic spline regression, and the shaded areas indicate 95% CIs.
HR, hazard ratio; CI, confidence interval; LDL-C, low-density lipoprotein cholesterol.
3. Subgroup analyses
Hazards of outcomes by LDL-C category were examined across subgroups defined by multiple clinical and lifestyle factors (Table 3). For all 3 outcomes, the increased hazard in the highest LDL-C categories was more pronounced in individuals with abdominal obesity (p for interaction: 0.0002 for MI and composite outcome; 0.0365 for stroke) and in current smokers (p for interaction: <0.0001 for MI and composite outcome; 0.0184 for stroke). The higher hazards of MI and the composite outcome in elevated LDL-C categories were also more prominent in individuals with a higher BMI category (p for interaction: both <0.0001), low HDL-C (p for interaction: <0.0001 for MI; 0.0010 for composite outcome), hypertriglyceridemia (p for interaction: both <0.0001), and those without heavy alcohol consumption (p for interaction: 0.0018 for MI; <0.0001 for composite outcome). Moreover, the increased hazards of MI and the composite outcome in the highest LDL-C categories were evident only in males (p for interaction: both <0.0001). Significant interactions by age group (20–29 vs. 30–39 years) were found for all outcomes (p for interaction: <0.0001 for MI; 0.0460 for stroke; 0.0100 for the composite outcome). The paradoxically lower hazards for all outcomes in LDL-C categories 55–129 mg/dL relative to the <55 mg/dL reference group were observed only among participants aged 30–39 years. However, for MI, the increased hazard at LDL-C ≥130 mg/dL was more pronounced in the 30–39-year group, whereas for stroke, the increased hazard at LDL-C ≥130 mg/dL was more prominent among those aged 20–29 years.
Table 3. Adjusted HRs and 95% CIs for myocardial infarction, ischemic stroke, and their composite according to the ranges of baseline LDL-C in subgroups stratified by selected factors.
| Subgroups | Ranges of LDL-C (mg/dL) | p for interaction | |||||||
|---|---|---|---|---|---|---|---|---|---|
| <55 | ≥55 and <70 | ≥70 and <100 | ≥100 and <130 | ≥130 and <160 | ≥160 | ||||
| Composite of myocardial infarction and stroke | |||||||||
| Sex | <0.0001 | ||||||||
| Male | 1 (Ref.) | 0.888 (0.831, 0.948) | 0.877 (0.829, 0.927) | 0.932 (0.881, 0.985) | 1.107 (1.047, 1.172) | 1.475 (1.390, 1.566) | |||
| Female | 1 (Ref.) | 0.970 (0.886, 1.062) | 0.920 (0.849, 0.997) | 0.901 (0.831, 0.977) | 0.957 (0.877, 1.044) | 1.073 (0.967, 1.192) | |||
| Age (yr) | 0.0100 | ||||||||
| 20–29 | 1 (Ref.) | 0.971 (0.888, 1.062) | 0.944 (0.872, 1.022) | 0.947 (0.874, 1.026) | 1.093 (1.004, 1.190) | 1.344 (1.215, 1.486) | |||
| 30–39 | 1 (Ref.) | 0.894 (0.837, 0.955) | 0.869 (0.822, 0.919) | 0.908 (0.859, 0.960) | 1.059 (1.001, 1.121) | 1.400 (1.319, 1.487) | |||
| Diabetes mellitus | 0.1837 | ||||||||
| No | 1 (Ref.) | 0.919 (0.870, 0.970) | 0.896 (0.854, 0.939) | 0.920 (0.877, 0.965) | 1.073 (1.021, 1.127) | 1.397 (1.325, 1.473) | |||
| Yes | 1 (Ref.) | 0.952 (0.776, 1.169) | 0.839 (0.707, 0.995) | 0.956 (0.809, 1.130) | 1.090 (0.919, 1.293) | 1.505 (1.262, 1.795) | |||
| Hypertension | 0.5418 | ||||||||
| No | 1 (Ref.) | 0.910 (0.858, 0.964) | 0.890 (0.846, 0.936) | 0.912 (0.867, 0.959) | 1.069 (1.014, 1.126) | 1.396 (1.319, 1.477) | |||
| Yes | 1 (Ref.) | 0.972 (0.846, 1.117) | 0.905 (0.801, 1.021) | 0.915 (0.811, 1.033) | 1.105 (0.976, 1.250) | 1.380 (1.208, 1.577) | |||
| BMI (kg/m2) | <0.0001 | ||||||||
| <23.0 | 1 (Ref.) | 0.874 (0.815, 0.937) | 0.833 (0.784, 0.886) | 0.835 (0.785, 0.888) | 0.951 (0.890, 1.017) | 1.238 (1.140, 1.344) | |||
| 23.0–24.9 | 1 (Ref.) | 0.895 (0.787, 1.017) | 0.887 (0.795, 0.990) | 0.920 (0.825, 1.026) | 1.088 (0.974, 1.216) | 1.373 (1.219, 1.547) | |||
| ≥25.0 | 1 (Ref.) | 1.022 (0.920, 1.135) | 1.031 (0.943, 1.128) | 1.104 (1.011, 1.205) | 1.279 (1.171, 1.398) | 1.685 (1.539, 1.846) | |||
| Abdominal obesity | 0.0002 | ||||||||
| No | 1 (Ref.) | 0.888 (0.838, 0.940) | 0.856 (0.815, 0.899) | 0.879 (0.836, 0.923) | 1.025 (0.973, 1.079) | 1.352 (1.277, 1.430) | |||
| Yes | 1 (Ref.) | 1.116 (0.966, 1.290) | 1.141 (1.009, 1.291) | 1.200 (1.063, 1.355) | 1.382 (1.222, 1.562) | 1.776 (1.567, 2.013) | |||
| Low HDL-C | 0.0010 | ||||||||
| No | 1 (Ref.) | 0.915 (0.864, 0.969) | 0.874 (0.831, 0.918) | 0.897 (0.854, 0.943) | 1.046 (0.994, 1.101) | 1.369 (1.296, 1.446) | |||
| Yes | 1 (Ref.) | 0.953 (0.829, 1.096) | 1.020 (0.904, 1.151) | 1.083 (0.960, 1.222) | 1.265 (1.118, 1.430) | 1.693 (1.486, 1.929) | |||
| Hypertriglyceridemia | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.960 (0.896, 1.028) | 0.913 (0.859, 0.971) | 0.931 (0.876, 0.990) | 1.064 (0.999, 1.134) | 1.344 (1.254, 1.440) | |||
| Yes | 1 (Ref.) | 0.867 (0.796, 0.945) | 0.905 (0.843, 0.971) | 0.967 (0.902, 1.036) | 1.144 (1.066, 1.227) | 1.518 (1.410, 1.635) | |||
| Statin use | 0.2320 | ||||||||
| No | 1 (Ref.) | 0.909 (0.861, 0.960) | 0.872 (0.832, 0.914) | 0.896 (0.855, 0.939) | 1.046 (0.997, 1.098) | 1.387 (1.317, 1.462) | |||
| Yes | 1 (Ref.) | 0.786 (0.604, 1.024) | 0.831 (0.672, 1.027) | 0.812 (0.656, 1.004) | 0.905 (0.730, 1.122) | 1.120 (0.907, 1.384) | |||
| Current smoking | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.937 (0.870, 1.009) | 0.882 (0.826, 0.941) | 0.877 (0.822, 0.937) | 0.969 (0.905, 1.037) | 1.172 (1.088, 1.262) | |||
| Yes | 1 (Ref.) | 0.887 (0.822, 0.957) | 0.892 (0.836, 0.951) | 0.963 (0.903, 1.026) | 1.176 (1.101, 1.255) | 1.621 (1.513, 1.737) | |||
| Heavy alcohol consumption | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.930 (0.876, 0.987) | 0.909 (0.863, 0.957) | 0.941 (0.893, 0.991) | 1.106 (1.049, 1.167) | 1.457 (1.377, 1.542) | |||
| Yes | 1 (Ref.) | 0.903 (0.804, 1.014) | 0.857 (0.777, 0.945) | 0.863 (0.783, 0.952) | 0.938 (0.845, 1.040) | 1.173 (1.043, 1.319) | |||
| Regular exercise | 0.4907 | ||||||||
| No | 1 (Ref.) | 0.912 (0.861, 0.966) | 0.892 (0.849, 0.937) | 0.923 (0.879, 0.970) | 1.069 (1.016, 1.125) | 1.411 (1.336, 1.490) | |||
| Yes | 1 (Ref.) | 0.972 (0.846, 1.117) | 0.905 (0.801, 1.021) | 0.915 (0.811, 1.033) | 1.105 (0.976, 1.250) | 1.380 (1.208, 1.577) | |||
| CKD | 0.0853 | ||||||||
| No | 1 (Ref.) | 0.924 (0.876, 0.975) | 0.900 (0.859, 0.943) | 0.927 (0.885, 0.971) | 1.077 (1.027, 1.130) | 1.417 (1.346, 1.491) | |||
| Yes | 1 (Ref.) | 0.780 (0.555, 1.097) | 0.670 (0.500, 0.896) | 0.752 (0.563, 1.003) | 0.947 (0.705, 1.273) | 1.026 (0.742, 1.418) | |||
| Myocardial infarction | |||||||||
| Sex | <0.0001 | ||||||||
| Male | 1 (Ref.) | 0.902 (0.831, 0.979) | 0.881 (0.821, 0.945) | 0.942 (0.879, 1.010) | 1.174 (1.093, 1.260) | 1.686 (1.565, 1.816) | |||
| Female | 1 (Ref.) | 0.955 (0.854, 1.069) | 0.879 (0.796, 0.971) | 0.856 (0.774, 0.947) | 0.887 (0.796, 0.990) | 0.948 (0.828, 1.085) | |||
| Age (yr) | <0.0001 | ||||||||
| 20–29 | 1 (Ref.) | 0.975 (0.875, 1.085) | 0.917 (0.834, 1.009) | 0.926 (0.841, 1.019) | 1.042 (0.940, 1.155) | 1.384 (1.228, 1.561) | |||
| 30–39 | 1 (Ref.) | 0.897 (0.825, 0.976) | 0.866 (0.805, 0.930) | 0.907 (0.845, 0.975) | 1.114 (1.036, 1.198) | 1.567 (1.452, 1.691) | |||
| Diabetes mellitus | 0.2643 | ||||||||
| No | 1 (Ref.) | 0.933 (0.871, 0.999) | 0.887 (0.836, 0.942) | 0.913 (0.860, 0.968) | 1.105 (1.040, 1.175) | 1.541 (1.443, 1.645) | |||
| Yes | 1 (Ref.) | 0.828 (0.629, 1.090) | 0.821 (0.657, 1.027) | 0.961 (0.773, 1.194) | 1.060 (0.848, 1.325) | 1.575 (1.253, 1.979) | |||
| Hypertension | 0.1438 | ||||||||
| No | 1 (Ref.) | 0.901 (0.839, 0.968) | 0.860 (0.808, 0.915) | 0.887 (0.833, 0.943) | 1.066 (1.000, 1.137) | 1.483 (1.384, 1.588) | |||
| Yes | 1 (Ref.) | 1.082 (0.905, 1.293) | 1.017 (0.874, 1.184) | 1.074 (0.925, 1.247) | 1.312 (1.127, 1.526) | 1.871 (1.601, 2.187) | |||
| BMI (kg/m2) | <0.0001 | ||||||||
| <23.0 | 1 (Ref.) | 0.868 (0.796, 0.946) | 0.810 (0.751, 0.873) | 0.818 (0.758, 0.883) | 0.954 (0.878, 1.036) | 1.286 (1.162, 1.422) | |||
| 23.0–24.9 | 1 (Ref.) | 0.929 (0.791, 1.090) | 0.877 (0.763, 1.007) | 0.918 (0.800, 1.054) | 1.129 (0.981, 1.299) | 1.485 (1.279, 1.724) | |||
| ≥25.0 | 1 (Ref.) | 1.047 (0.916, 1.198) | 1.069 (0.953, 1.199) | 1.135 (1.014, 1.272) | 1.373 (1.225, 1.539) | 1.964 (1.747, 2.207) | |||
| Abdominal obesity | 0.0002 | ||||||||
| No | 1 (Ref.) | 0.896 (0.835, 0.961) | 0.839 (0.789, 0.892) | 0.866 (0.814, 0.921) | 1.047 (0.982, 1.116) | 1.455 (1.357, 1.561) | |||
| Yes | 1 (Ref.) | 1.122 (0.928, 1.355) | 1.217 (1.035, 1.430) | 1.259 (1.073, 1.476) | 1.498 (1.276, 1.759) | 2.115 (1.796, 2.489) | |||
| Low HDL-C | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.918 (0.855, 0.986) | 0.860 (0.808, 0.915) | 0.884 (0.831, 0.940) | 1.059 (0.993, 1.129) | 1.484 (1.386, 1.589) | |||
| Yes | 1 (Ref.) | 0.987 (0.826, 1.179) | 1.044 (0.894, 1.218) | 1.119 (0.960, 1.305) | 1.407 (1.203, 1.645) | 2.002 (1.698, 2.360) | |||
| Hypertriglyceridemia | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.954 (0.876, 1.039) | 0.882 (0.818, 0.951) | 0.902 (0.836, 0.973) | 1.052 (0.972, 1.137) | 1.408 (1.292, 1.534) | |||
| Yes | 1 (Ref.) | 0.882 (0.792, 0.984) | 0.931 (0.852, 1.019) | 0.996 (0.912, 1.087) | 1.245 (1.138, 1.361) | 1.764 (1.607, 1.936) | |||
| Statin use | 0.0931 | ||||||||
| No | 1 (Ref.) | 0.919 (0.859, 0.983) | 0.863 (0.813, 0.915) | 0.890 (0.840, 0.944) | 1.075 (1.012, 1.142) | 1.525 (1.429, 1.626) | |||
| Yes | 1 (Ref.) | 0.744 (0.518, 1.070) | 0.973 (0.734, 1.291) | 0.878 (0.661, 1.168) | 1.086 (0.818, 1.443) | 1.339 (1.011, 1.772) | |||
| Current smoking | <0.0001 | ||||||||
| No | 1 (Ref.) | 0.958 (0.871, 1.053) | 0.888 (0.817, 0.966) | 0.880 (0.810, 0.957) | 0.982 (0.901, 1.071) | 1.228 (1.117, 1.350) | |||
| Yes | 1 (Ref.) | 0.883 (0.804, 0.969) | 0.864 (0.798, 0.935) | 0.940 (0.869, 1.017) | 1.211 (1.117, 1.312) | 1.811 (1.665, 1.970) | |||
| Heavy alcohol consumption | 0.0018 | ||||||||
| No | 1 (Ref.) | 0.941 (0.873, 1.014) | 0.908 (0.850, 0.969) | 0.940 (0.880, 1.003) | 1.145 (1.071, 1.224) | 1.612 (1.502, 1.731) | |||
| Yes | 1 (Ref.) | 0.902 (0.781, 1.042) | 0.815 (0.721, 0.921) | 0.835 (0.740, 0.944) | 0.936 (0.823, 1.065) | 1.253 (1.085, 1.447) | |||
| Regular exercise | 0.8670 | ||||||||
| No | 1 (Ref.) | 0.916 (0.853, 0.984) | 0.880 (0.827, 0.936) | 0.908 (0.854, 0.966) | 1.097 (1.030, 1.170) | 1.543 (1.442, 1.652) | |||
| Yes | 1 (Ref.) | 0.998 (0.838, 1.189) | 0.909 (0.780, 1.059) | 0.948 (0.814, 1.104) | 1.135 (0.971, 1.327) | 1.545 (1.307, 1.825) | |||
| CKD | 0.0779 | ||||||||
| No | 1 (Ref.) | 0.936 (0.875, 1.001) | 0.892 (0.842, 0.946) | 0.919 (0.867, 0.974) | 1.109 (1.044, 1.177) | 1.559 (1.463, 1.662) | |||
| Yes | 1 (Ref.) | 0.631 (0.409, 0.975) | 0.604 (0.421, 0.866) | 0.729 (0.511, 1.041) | 0.880 (0.611, 1.269) | 1.010 (0.677, 1.508) | |||
| Ischemic stroke | |||||||||
| Sex | 0.1555 | ||||||||
| Male | 1 (Ref.) | 0.847 (0.763, 0.940) | 0.852 (0.780, 0.931) | 0.895 (0.820, 0.977) | 0.985 (0.901, 1.077) | 1.132 (1.028, 1.246) | |||
| Female | 1 (Ref.) | 0.993 (0.858, 1.151) | 0.979 (0.859, 1.115) | 0.963 (0.845, 1.098) | 1.065 (0.927, 1.224) | 1.268 (1.078, 1.491) | |||
| Age (yr) | 0.0460 | ||||||||
| 20–29 | 1 (Ref.) | 0.967 (0.825, 1.133) | 1.008 (0.877, 1.159) | 1.001 (0.870, 1.152) | 1.204 (1.037, 1.398) | 1.269 (1.062, 1.518) | |||
| 30–39 | 1 (Ref.) | 0.870 (0.787, 0.963) | 0.850 (0.780, 0.926) | 0.883 (0.811, 0.961) | 0.956 (0.876, 1.044) | 1.126 (1.026, 1.237) | |||
| Diabetes mellitus | 0.1540 | ||||||||
| No | 1 (Ref.) | 0.884 (0.809, 0.966) | 0.893 (0.827, 0.964) | 0.914 (0.847, 0.987) | 1.004 (0.928, 1.087) | 1.154 (1.058, 1.258) | |||
| Yes | 1 (Ref.) | 1.055 (0.787, 1.415) | 0.853 (0.664, 1.096) | 0.931 (0.729, 1.189) | 1.102 (0.859, 1.414) | 1.328 (1.023, 1.723) | |||
| Hypertension | 0.0953 | ||||||||
| No | 1 (Ref.) | 0.900 (0.817, 0.990) | 0.915 (0.842, 0.995) | 0.930 (0.856, 1.011) | 1.048 (0.961, 1.142) | 1.205 (1.096, 1.325) | |||
| Yes | 1 (Ref.) | 0.904 (0.753, 1.086) | 0.813 (0.698, 0.948) | 0.880 (0.758, 1.022) | 0.902 (0.774, 1.052) | 1.073 (0.911, 1.263) | |||
| BMI (kg/m2) | 0.0723 | ||||||||
| <23.0 | 1 (Ref.) | 0.865 (0.771, 0.970) | 0.850 (0.769, 0.939) | 0.837 (0.757, 0.925) | 0.928 (0.832, 1.035) | 1.121 (0.978, 1.284) | |||
| 23.0–24.9 | 1 (Ref.) | 0.813 (0.663, 0.997) | 0.878 (0.740, 1.043) | 0.903 (0.761, 1.070) | 1.006 (0.845, 1.198) | 1.154 (0.955, 1.395) | |||
| ≥25.0 | 1 (Ref.) | 0.996 (0.847, 1.170) | 0.981 (0.855, 1.125) | 1.062 (0.928, 1.215) | 1.150 (1.004, 1.318) | 1.312 (1.139, 1.511) | |||
| Abdominal obesity | 0.0365 | ||||||||
| No | 1 (Ref.) | 0.851 (0.776, 0.933) | 0.864 (0.798, 0.935) | 0.876 (0.809, 0.948) | 0.969 (0.892, 1.052) | 1.149 (1.047, 1.260) | |||
| Yes | 1 (Ref.) | 1.159 (0.933, 1.440) | 1.053 (0.874, 1.269) | 1.143 (0.951, 1.373) | 1.245 (1.034, 1.498) | 1.372 (1.133, 1.662) | |||
| Low HDL-C | 0.2500 | ||||||||
| No | 1 (Ref.) | 0.891 (0.813, 0.977) | 0.878 (0.811, 0.950) | 0.901 (0.833, 0.975) | 1.012 (0.933, 1.098) | 1.166 (1.066, 1.274) | |||
| Yes | 1 (Ref.) | 0.914 (0.735, 1.136) | 0.979 (0.812, 1.180) | 1.012 (0.840, 1.219) | 1.035 (0.854, 1.253) | 1.239 (1.006, 1.525) | |||
| Hypertriglyceridemia | 0.6066 | ||||||||
| No | 1 (Ref.) | 0.940 (0.841, 1.052) | 0.938 (0.850, 1.036) | 0.951 (0.861, 1.050) | 1.057 (0.953, 1.171) | 1.206 (1.076, 1.350) | |||
| Yes | 1 (Ref.) | 0.853 (0.746, 0.975) | 0.863 (0.773, 0.963) | 0.918 (0.825, 1.023) | 0.996 (0.891, 1.112) | 1.158 (1.029, 1.304) | |||
| Statin use | 0.1987 | ||||||||
| No | 1 (Ref.) | 0.878 (0.805, 0.959) | 0.868 (0.805, 0.936) | 0.885 (0.821, 0.954) | 0.984 (0.910, 1.063) | 1.153 (1.059, 1.255) | |||
| Yes | 1 (Ref.) | 0.847 (0.586, 1.225) | 0.658 (0.482, 0.898) | 0.716 (0.527, 0.974) | 0.698 (0.508, 0.959) | 0.854 (0.625, 1.167) | |||
| Current smoking | 0.0184 | ||||||||
| No | 1 (Ref.) | 0.892 (0.794, 1.003) | 0.861 (0.778, 0.953) | 0.862 (0.779, 0.954) | 0.943 (0.848, 1.048) | 1.092 (0.973, 1.227) | |||
| Yes | 1 (Ref.) | 0.883 (0.780, 1.000) | 0.920 (0.828, 1.021) | 0.977 (0.880, 1.083) | 1.085 (0.976, 1.208) | 1.251 (1.115, 1.404) | |||
| Heavy alcohol consumption | 0.1282 | ||||||||
| No | 1 (Ref.) | 0.903 (0.821, 0.994) | 0.901 (0.830, 0.979) | 0.931 (0.857, 1.011) | 1.037 (0.953, 1.129) | 1.206 (1.100, 1.322) | |||
| Yes | 1 (Ref.) | 0.876 (0.727, 1.056) | 0.880 (0.752, 1.028) | 0.870 (0.744, 1.017) | 0.901 (0.764, 1.063) | 1.009 (0.834, 1.221) | |||
| Regular exercise | 0.2236 | ||||||||
| No | 1 (Ref.) | 0.891 (0.813, 0.977) | 0.895 (0.827, 0.968) | 0.928 (0.858, 1.004) | 1.008 (0.929, 1.094) | 1.182 (1.082, 1.292) | |||
| Yes | 1 (Ref.) | 0.916 (0.734, 1.142) | 0.882 (0.729, 1.067) | 0.853 (0.705, 1.031) | 1.035 (0.852, 1.258) | 1.109 (0.895, 1.374) | |||
| CKD | 0.0432 | ||||||||
| No | 1 (Ref.) | 0.897 (0.823, 0.978) | 0.902 (0.837, 0.971) | 0.927 (0.861, 0.998) | 1.017 (0.942, 1.098) | 1.183 (1.088, 1.285) | |||
| Yes | 1 (Ref.) | 0.820 (0.502, 1.338) | 0.605 (0.396, 0.925) | 0.599 (0.393, 0.912) | 0.836 (0.544, 1.286) | 0.807 (0.498, 1.308) | |||
Adjusted for age (continuous variable), sex, low-income state, body mass index, current smoking, heavy alcohol consumption, regular exercise, diabetes mellitus, hypertension, statin and fibrate use, and Charlson Comorbidity Index.
HR, hazard ratio; CI, confidence interval; LDL-C, low-density lipoprotein cholesterol; BMI, body mass index; HDL-C, high-density lipoprotein cholesterol; CKD, chronic kidney disease.
4. Sensitivity analyses
After accounting for all-cause mortality as a competing risk, the sub-distribution HRs for all outcomes remained consistent with the findings from the main analyses (Supplementary Table 2).
DISCUSSION
In this large-scale, nationwide, longitudinal cohort study of more than 6.4 million young adults aged 20–39 years, we observed a J-shaped association between LDL-C levels and the risk of MI, stroke, and their composite outcome over a median follow-up of 12.6 years. When LDL-C was modeled as a continuous variable, the lowest hazard for all outcomes occurred at LDL-C levels around 90–110 mg/dL, after which hazards increased progressively with higher LDL-C concentrations. When examined using categorical LDL-C groups, levels ≥130 mg/dL were associated with significantly increased hazards compared with the lowest category (<55 mg/dL), with the highest risks consistently observed in those with LDL-C ≥160 mg/dL. These findings were robust across multiple multivariable-adjusted models and were confirmed in competing risk analyses. Importantly, the elevated hazards in the highest LDL-C categories were more pronounced in subgroups with cardiometabolic risk factors, including abdominal obesity, higher BMI, low HDL-C, hypertriglyceridemia, and current smoking.
To our knowledge, this study is the first to describe the full distribution of LDL-C levels in a nationwide cohort of Korean young adults. In our population, nearly one-fifth had at least moderately elevated LDL-C (≥130 mg/dL), and approximately 4.5% had LDL-C ≥160 mg/dL. When individuals with LDL-C ≥160 mg/dL or statin prescriptions were considered together, the proportion increased to about 5.2%. According to the Dyslipidemia Fact Sheet in Korea 2024, based on the 2016–2022 KNHANES data, the prevalence of LDL-C ≥160 mg/dL or use of lipid medications was higher than our estimates, reaching 6.9% and 13.3% in men and 5.0% and 8.2% in women in their 20s and 30s, respectively.7 This discrepancy may partly reflect that our study analyzed participants from an earlier period (2009–2012), during which the prevalence of hypercholesterolemia in Korea was lower and has subsequently risen.7 Additionally, our cohort included only young adults without baseline MI and/or stroke who underwent at least 1 national health screening during 2009–2012. Individuals who did not participate in screening—who may represent a less healthy segment of the population—were not included, which could also contribute to differences in prevalence estimates.
Notably, J-shaped associations with outcome hazards were observed in the total population and among statin non-users, with slightly higher hazards evident at the lowest LDL-C levels. We believe this pattern should be interpreted with caution. In our dataset, individuals with LDL-C <55 mg/dL comprised only a small fraction (2.74%) of the population. Baseline characteristics across LDL-C categories revealed that heavy alcohol consumption was most prevalent in the <55 mg/dL group, and the prevalence of hypertension and DM was slightly higher than in mid-range LDL-C groups. In addition, the proportion of statin users showed a J-shaped distribution across LDL-C levels. Participants with very low LDL-C values likely include individuals with extremely low LDL-C due to underlying unhealthy conditions—such as malnutrition or excessive alcohol use—as well as individuals with intentionally lowered LDL-C resulting from lipid-lowering therapy prescribed for existing clinical risk factors. These features may help explain the modestly elevated hazards observed in the lowest LDL-C category in the total population and the absence of increased MI hazard at the lowest LDL-C levels among statin users.
Our findings provide important evidence regarding the cardiovascular risk associated with moderate to higher LDL-C elevations among young adults in the primary prevention setting. Although current LDL-C management strategies primarily target adults over 40 years of age who are at high ASCVD risk, substantially less evidence exists for individuals under 40 years—especially those without FH or established cardiovascular disease.5 Consequently, the clinical implications of elevated LDL-C in adults aged 20–39 years have remained unclear. In our study, even moderately elevated LDL-C levels (≥130 mg/dL) were associated with significantly increased ASCVD risk, underscoring the need to recognize and appropriately manage such elevations in younger adults.
The association between higher LDL-C categories and the risk of cardiovascular outcomes was more pronounced in young adults with abdominal obesity, current smoking, overweight or obesity, low HDL-C, and hypertriglyceridemia. These findings suggest that actively lowering LDL-C levels in young adults may produce greater benefits in individuals who carry these cardiometabolic risk factors. Therefore, even among young adults without prior ASCVD, clinicians may need to proactively assess and manage LDL-C levels, particularly in those with additional cardiometabolic risks. Similarly, in elderly populations, the association between elevated LDL-C and increased stroke risk has been reported to be stronger in individuals with obesity.21 Regarding smoking, it has been linked to increased triglyceride concentrations and higher levels of small dense LDL-C, along with reduced HDL-C levels,22 suggesting that the adverse cardiovascular effects of elevated LDL-C may be amplified among smokers.
Interestingly, the increased hazards of MI and the composite outcome in the higher LDL-C categories were less evident in heavy alcohol consumers. In individuals with heavy alcohol consumption, the LDL-C/HDL-C ratio has been reported to be lower than in nondrinkers.23 Moreover, alcohol cessation has been associated with increased LDL-C and decreased HDL-C concentrations compared with continued alcohol use, while alcohol initiation has shown the opposite pattern; these changes have been more pronounced at higher levels of alcohol consumption.24 These reductions in LDL-C among heavy drinkers may have contributed to the weaker association between higher LDL-C categories and cardiovascular risk in this subgroup.
Several limitations of our study should be acknowledged. First, given the observational design, definitive causal relationships cannot be established. Nevertheless, to reduce the potential for reverse causality, we excluded individuals with claims for MI and/or stroke at or before baseline and those who died or developed outcomes within 1 year after baseline. Second, because the study population consisted exclusively of individuals of Korean ethnicity, generalizability to other ethnic groups may be limited. Third, LDL-C levels were estimated using the Friedewald formula rather than being directly measured. Although this method correlates well with direct measurement, it may underestimate LDL-C in individuals with elevated triglycerides or low HDL-C.25 To mitigate this limitation, we excluded individuals with baseline triglycerides >400 mg/dL. However, among participants with relatively high triglyceride levels (within the included range) or low HDL-C, the protective effects of lower LDL-C levels and the hazard differences across LDL-C categories may still have been underestimated. Despite this, for MI and the composite outcome, our results indicated that the increased hazards associated with higher LDL-C categories were more pronounced in individuals with low HDL-C or hypertriglyceridemia. Fourth, LDL-C was measured only once at baseline. Given possible variability over time and potential changes in lifestyle or treatment, a single measurement may misclassify longer-term exposure. Fifth, covariates were assessed only at baseline, which prevented systematic evaluation of changes in risk factors, medication use, or lifestyle during follow-up. Sixth, outcome ascertainment relied on hospital-diagnosed MI and stroke; therefore, fatal events occurring before hospital arrival may have been undercaptured, and we did not evaluate cause-specific mortality separately. Finally, among lipid-lowering therapy users, we lacked detailed information on dosage, treatment intensity, and adherence, and thus could not adjust for these factors. However, because statin users represented only 0.75% of the cohort and stratified analyses among non-users showed patterns consistent with the main results, residual confounding from treatment intensity is unlikely to have materially influenced the findings.
Our study has several notable strengths. We used a nationwide cohort database established and maintained by the Korean government, which has been validated for epidemiologic research. Comprehensive data on lifestyle factors, anthropometric measurements, and laboratory parameters were available for more than 6.4 million young adults aged 20–39 years, enabling adjustment for a wide range of potential confounders. The large sample size and lengthy follow-up duration (median: 12.6 years) also allowed us to capture a substantial number of clinical outcomes, even in this relatively young population.
In this real-world, nationwide, large-scale population-based study of young Korean adults aged 20–39 years, a J-shaped association between LDL-C levels and cardiovascular risk was identified. Above approximately 90–110 mg/dL, risks increased progressively with higher LDL-C levels, consistent with findings reported in middle-aged populations. However, mildly higher hazards at the lowest LDL-C levels were not observed for MI among statin users. The association between elevated LDL-C and increased cardiovascular risk was also more pronounced in individuals with cardiometabolic risk factors such as abdominal obesity, general obesity, low HDL-C, hypertriglyceridemia, and current smoking. Given the long-term implications of early-onset dyslipidemia and the rising burden of cardiometabolic conditions in younger populations, our results underscore the importance of early lipid screening and individualized risk assessment. Such efforts may facilitate timely preventive strategies and help reduce the lifetime burden of cardiovascular disease in young adults. Furthermore, beyond LDL-C, additional markers such as non-HDL-C, apolipoprotein B, and lipoprotein(a) may enhance risk stratification in young adults and warrant further investigation in future research.
ACKNOWLEDGEMENTS
This work was performed using the database from the Korean National Health Insurance Service (KNHIS). The National Health Information Database constructed by the KNHIS was used, and the results do not necessarily represent the opinion of the KNHIS.
Footnotes
Funding: This study was supported by the Korean Society of Lipid and Atherosclerosis (KSOLA2024-03-003). The funders had no role in the study design, data collection, data analysis, data interpretation, or writing of the report.
Conflict of Interest: The authors have no conflicts of interest to declare.
Data Availability Statement: The data that support the findings of this study are available from the Korean National Health Insurance Service (KNHIS) but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are, however, available from the corresponding authors upon reasonable request and with permission of the KNHIS.
- Conceptualization: Lee YB, Han K.
- Data curation: Lee KN, Han K.
- Formal analysis: Lee KN, Han K.
- Writing - original draft: Lee YB.
- Writing - review & editing: Lee YB, Lee KN, Han K.
SUPPLEMENTARY MATERIALS
Definitions of the covariates
Subdistribution HRs and 95% CIs for the incidence of myocardial infarction and ischemic stroke, and their composite according to the ranges of baseline LDL-C, accounting for all-cause mortality as a competing event
Flow diagram of the study population.
References
- 1.Silverman MG, Ference BA, Im K, Wiviott SD, Giugliano RP, Grundy SM, et al. Association between lowering LDL-C and cardiovascular risk reduction among different therapeutic interventions: a systematic review and meta-analysis. JAMA. 2016;316:1289–1297. doi: 10.1001/jama.2016.13985. [DOI] [PubMed] [Google Scholar]
- 2.Mhaimeed O, Burney ZA, Schott SL, Kohli P, Marvel FA, Martin SS. The importance of LDL-C lowering in atherosclerotic cardiovascular disease prevention: lower for longer is better. Am J Prev Cardiol. 2024;18:100649. doi: 10.1016/j.ajpc.2024.100649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the management of blood cholesterol: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2019;139:e1082–e1143. doi: 10.1161/CIR.0000000000000625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mach F, Baigent C, Catapano AL, Koskinas KC, Casula M, Badimon L, et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020;41:111–188. doi: 10.1093/eurheartj/ehz455. [DOI] [PubMed] [Google Scholar]
- 5.O’Toole T, Kelsey MD, Shah NP, McGarrah RW, Pagidipati NJ. Eradicating atherosclerosis: should we start statins at younger ages and at lower LDL-Cs. Curr Cardiol Rep. 2022;24:1397–1406. doi: 10.1007/s11886-022-01760-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.National Center for Health Statistics. Health, United States, 2019. Trend tables [Internet] National Center for Health Statistics; 2021. [cited 2025 May 10]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK569311/ [Google Scholar]
- 7.Korean Society of Lipid and Atherosclerosis. Dyslipidemia fact sheet in Korea 2024. Korean Society of Lipid and Atherosclerosis; 2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zeitouni M, Nanna MG, Sun JL, Chiswell K, Peterson ED, Navar AM. Performance of guideline recommendations for prevention of myocardial infarction in young adults. J Am Coll Cardiol. 2020;76:653–664. doi: 10.1016/j.jacc.2020.06.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lee YH, Han K, Ko SH, Ko KS, Lee KU Taskforce Team of Diabetes Fact Sheet of the Korean Diabetes Association. Data analytic process of a nationwide population-based study using National Health Information Database established by National Health Insurance Service. Diabetes Metab J. 2016;40:79–82. doi: 10.4093/dmj.2016.40.1.79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kim MK, Han K, Lee SH. Current trends of big data research using the Korean National Health Information Database. Diabetes Metab J. 2022;46:552–563. doi: 10.4093/dmj.2022.0193. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cheol Seong S, Kim YY, Khang YH, Heon Park J, Kang HJ, Lee H, et al. Data resource profile: the National Health Information Database of the National Health Insurance Service in South Korea. Int J Epidemiol. 2017;46:799–800. doi: 10.1093/ije/dyw253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lee YB, Han K, Kim B, Lee SE, Jun JE, Ahn J, et al. Risk of early mortality and cardiovascular disease in type 1 diabetes: a comparison with type 2 diabetes, a nationwide study. Cardiovasc Diabetol. 2019;18:157. doi: 10.1186/s12933-019-0953-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kim MK, Han K, Koh ES, Kim ES, Lee MK, Nam GE, et al. Blood pressure and development of cardiovascular disease in Koreans with type 2 diabetes mellitus. Hypertension. 2019;73:319–326. [PubMed] [Google Scholar]
- 14.Kim MK, Han K, Kim HS, Park YM, Kwon HS, Yoon KH, et al. Cholesterol variability and the risk of mortality, myocardial infarction, and stroke: a nationwide population-based study. Eur Heart J. 2017;38:3560–3566. doi: 10.1093/eurheartj/ehx585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 16.Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA. New ICD-10 version of the Charlson Comorbidity Index predicted in-hospital mortality. J Clin Epidemiol. 2004;57:1288–1294. doi: 10.1016/j.jclinepi.2004.03.012. [DOI] [PubMed] [Google Scholar]
- 17.Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502. [PubMed] [Google Scholar]
- 18.Rhee EJ, Kim HC, Kim JH, Lee EY, Kim BJ, Kim EM, et al. 2018 Guidelines for the management of dyslipidemia in Korea. J Lipid Atheroscler. 2019;8:78–131. doi: 10.12997/jla.2019.8.2.78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yang YS, Kim HL, Kim SH, Moon MK Committee of Clinical Practice Guideline, Korean Diabetes Association and Clinical Practice Guideline Committee, Korean Society of Lipid and Atherosclerosis. Lipid management in Korean people with type 2 diabetes mellitus: Korean Diabetes Association and Korean Society of Lipid and Atherosclerosis Consensus Statement. J Lipid Atheroscler. 2023;12:12–22. doi: 10.12997/jla.2023.12.1.12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509. [Google Scholar]
- 21.Lee YB, Koo M, Noh E, Hwang SY, Kim JA, Roh E, et al. Myocardial infarction, stroke, and all-cause mortality according to low-density lipoprotein cholesterol level in the elderly, a nationwide study. Diabetes Metab J. 2022;46:722–732. doi: 10.4093/dmj.2021.0225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nakamura M, Yamamoto Y, Imaoka W, Kuroshima T, Toragai R, Ito Y, et al. Relationships between smoking status, cardiovascular risk factors, and lipoproteins in a large Japanese population. J Atheroscler Thromb. 2021;28:942–953. doi: 10.5551/jat.56838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wakabayashi I. Associations between heavy alcohol drinking and lipid-related indices in middle-aged men. Alcohol. 2013;47:637–642. doi: 10.1016/j.alcohol.2013.10.001. [DOI] [PubMed] [Google Scholar]
- 24.Suzuki T, Fukui S, Shinozaki T, Asano T, Yoshida T, Aoki J, et al. Lipid profiles after changes in alcohol consumption among adults undergoing annual checkups. JAMA Netw Open. 2025;8:e250583. doi: 10.1001/jamanetworkopen.2025.0583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lee J, Jang S, Jeong H, Ryu OH. Validation of the Friedewald formula for estimating low density lipoprotein cholesterol: the Korea National Health and Nutrition Examination Survey, 2009 to 2011. Korean J Intern Med. 2020;35:150–159. doi: 10.3904/kjim.2017.233. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Definitions of the covariates
Subdistribution HRs and 95% CIs for the incidence of myocardial infarction and ischemic stroke, and their composite according to the ranges of baseline LDL-C, accounting for all-cause mortality as a competing event
Flow diagram of the study population.