Intraindividual variability of lipoprotein(a) in the middle-aged population: the impact of smoking
CC BY 4.0
Intraindividual variability of lipoprotein(a) in the middle-aged population: the impact of smoking
CC BY 4.0
Introduction
Lipoprotein(a) (Lp[a]) is an independent, genetically determined risk factor for cardiovascular diseases (CVDs), including coronary artery disease (CAD) and aortic valve stenosis.1,2 Current European Society of Cardiology guidelines recommend Lp(a) measurement at least once in a lifetime, preferably in the first lipid profile, and interpret its values in the context of absolute global CVD risk.3-5 Individuals with extremely high Lp(a) levels, that is, above 180 mg/dl, may have a risk of atherosclerotic CVD similar to that observed in heterozygous familiar hypercholesterolemia.6,7 About 20%–30% of people worldwide are estimated to have elevated Lp(a) level, defined as that above 30 mg/dl in Europe and 50 mg/dl in the United States.1,3,4 In Polish middle‑aged populations, without evident atherosclerotic CVD, the prevalence of Lp(a) levels above 50 mg/dl was 17.8%.7 Moreover, the European Atherosclerosis Society (EAS) introduced the grey zone category (30–50 mg/dl) for intermediate CVD risk, considering the continuous association between Lp(a) levels and the risk of CVD.1
Although plasma Lp(a) concentrations are 75%–95% heritable and lifestyle or pharmacological interventions have a minor impact on Lp(a) level, a significant intraindividual variability in Lp(a) levels has been recently reported.8-11 Trinder et al8 identified 13% of the 16 017 United Kingdom Biobank participants, mostly of white ethnicity at a mean (SD) age of 58.2 (7.4) years, to have Lp(a) concentrations variability defined as a change above 10 mg/dl during a 4‑year follow‑up. Another report, which set the cutoff variability point at 25% or more, showed that almost 41% of 52 patients randomized to placebo in Lp(a)-lowering trials, including 48 individuals with baseline Lp(a) levels equal to or above 75 nmol/l (roughly ≥30 mg/dl), have reached this cutoff point between measurements performed within 6 months.9 Harb et al10 reported that over 50% of 432 patients without kidney or liver disease (at a median age of 50 years), who were at baseline classified to the grey zone category, were reclassified to low (18.4%) or high‑risk (30.3%) category at 1 year of observation. Since little is known about this phenomenon in Poland, we decided to assess the Lp(a) variability in Polish middle‑aged individuals free of CVD and to identify its determinants.
Patients and methods
We studied 301 consecutive individuals aged 40–65 years, free of CVD, who were enrolled in a local prevention program in southern Poland (“Malopolska coronary artery disease prophylactic program for people over 40 years”) from May to October 2022. The population described in this study is a subset of a group reported previously.7 We evaluated individuals with a positive family history of CAD, defined as 1 or more close relatives diagnosed with CAD before the age of 55 years for men and 65 years for women. These individuals were subjected to extended diagnostic workup, and had laboratory tests repeated 3 months after enrollment. The exclusion criteria were: documented CVD, including peripheral artery disease, prior myocardial infarction, cerebrovascular ischemic events, known chronic kidney disease or liver injury, acute infection, and active cancer. Demographic and clinical characteristics were collected on admission. Obesity was defined as body mass index (BMI) equal to or above 30 kg/m2. Arterial hypertension was established as systolic blood pressure of 140 mm Hg and / or diastolic blood pressure of 90 mm Hg and / or treatment with antihypertensive drugs. Diabetes mellitus was defined according to the American Diabetic Association criteria12. Dyslipidemia was defined as total cholesterol level equal to or above 5 mmol/l, or low‑density lipoprotein cholesterol (LDL‑C) equal to or above 3 mmol/l (definition of hypercholesterolemia), or triglycerides equal to or above 1.7 mmol/l (definition of hypertriglyceridemia), or statin use. The 10‑year risk of the first CVD event was assessed with the Systematic Coronary Risk Evaluation 2 (SCORE2) algorithm.3
Fasting blood samples were drawn at 8 to 10 AM, and Lp(a) level was measured on the same day. Lipid profiles (assessed directly with the homogenous enzymatic colorimetric method) and fasting glucose were assayed by routine laboratory techniques. High‑sensitivity C‑reactive protein level was determined using immunoturbidimetry (Siemens, Marburg, Germany). Fibrinogen concentration was determined using the Clauss method, and serum Lp(a) with immunoturbidimetry (Roche Diagnostic, Mannheim, Germany). According to the manufacturer, coefficient of variation of Lp(a) repeatability is below 2.1%. The risk categories of CVD suggested by EAS were applied.1
Laboratory investigations were performed twice, that is, on admission and after 3 months. Patients diagnosed with hypercholesterolemia and / or hypertriglyceridemia were advised to follow a low‑fat, low‑carbohydrate, guideline‑recommended diet.3 The absolute Δ of Lp(a) was defined as a change in its concentration expressed in mg/dl over time and the relative Δ of Lp(a) was defined as a percentage change over time, encompassing both increased and decreased Lp(a) levels after 3 months from enrollment. Based on the relative Δ of Lp(a), the patients were categorized into 3 groups: with Δ Lp(a) below 10%, 10%–20%, and above 20%. The study was approved by the local Ethics Committee (OIL/KBL33/2022).
Statistical analysis
Continuous variables were expressed as median (interquartile range [IQR]) and categorical variables as number (percentage). Normal distribution was assessed using the Shapiro–Wilk test. Equality of variances was assessed using the Levene test. Data obtained on admission and after 3 months of follow‑up were compared with the Wilcoxon matched pairs test or the paired t test, as appropriate. The Mann–Whitney test was used for non‑normally distributed continuous variables. Differences among the 3 groups were compared using the analysis of variance when normally disturbed or the Kruskal–Wallis test for multiple comparison of non‑normally distributed variables. Categorical variables were analyzed with the χ2 test or the Fisher exact test. The Spearman ρ correlation coefficient was calculated to test the associations between continuous variables. Odds ratios (ORs) were calculated using 2 × 2 contingency Tables, and were presented along with 95% CIs. A 2‑sided P value below 0.05 was considered significant. Statistical analysis was performed using the Statistica software (Version 14.1, TIBCO Software, Palo Alto, California, United States).
Results
Baseline
The study included individuals at a median (IQR) age of 50 (45–55) years (men, 37%) with a median (IQR) baseline Lp(a) level of 7.1 (2.9–25.7) mg/dl (Table 1). Regarding the EAS risk categories based on Lp(a) level, there were 230 participants (76.4%) with a low risk, 20 (6.6%) with an intermediate risk, and 51 individuals (16.9%) were classified into the high cardiovascular risk group, including 3 (1%) with Lp(a) level exceeding 180 mg/dl. There were no baseline differences between the participants with high‑to‑very high cardiovascular risk and the remainder with regard to the demographic, clinical, and laboratory variables, except for higher median (IQR) Lp(a) level (89.6 [72–120.1] mg/dl vs 4.9 [2.9–13.3] mg/dl; P <0.001).
Variable | All participants (n = 301) | Relative change in lipoprotein(a) level | ||||
<10% (n = 160; 53.2%) | 10%–20% (n = 65; 21.6%) | >20% (n = 76; 25.2%) | P value | |||
Values are shown as number (percentage) or median (interquartile range) unless provided otherwise.
a P <0.05 between 10% and 10%–20%
b P <0.05 between 10%–20% and >20%
c P <0.05 between <10% and >20%
Abbreviations: BMI, body mass index; CRP, C‑reactive protein; CVD, cardiovascular disease; DM, diabetes mellitus; HDL, high‑density lipoprotein; LDL, low‑density lipoprotein; SCORE2, Systematic Coronary Risk Evaluation 2; TC, total cholesterol; TG, triglycerides | ||||||
Baseline | ||||||
Age, y | 50 (45–55) | 50 (45–55) | 52 (47–57)a | 49 (45–53)b | 0.02 | |
Men | 112 (37) | 61 (38.1) | 24 (36.9) | 27 (35.5) | 0.93 | |
BMI, kg/m2 | 27.2 (24.7–30.8) | 27.1 (24.8–30.8) | 28 (24.8–31.1) | 26.9 (24.4–30) | 0.53 | |
Risk factors | Obesity | 85 (28.2) | 45 (28.1) | 21 (32.3) | 19 (25) | 0.63 |
Hypertension | 95 (31.6) | 51 (31.9) | 25 (38.4) | 19 (25) | 0.23 | |
Active smoking | 43 (14.2) | 21 (13.1) | 6 (9.2) | 16 (21.1) | 0.11 | |
Former smoking | 61 (20.1) | 30 (18.8) | 11 (16.9) | 20 (26.3) | 0.3 | |
DM | 16 (5.3) | 7 (4.4) | 4 (6.2) | 5 (6.6) | 0.74 | |
Hypercholesterolemia | 216 (71.8) | 112 (70) | 49 (75.4) | 62 (81.6) | 0.71 | |
Hypertriglyceridemia | 67 (22.3) | 38 (23.8) | 14 (21.5) | 15 (19.7) | 0.78 | |
Prior statin use | 20 (6.6) | 12 (7.5) | 4 (6.2) | 4 (5.3) | 0.8 | |
Family history of premature CVD | 137 (45.2) | 72 (45) | 29 (44.6) | 36 (47.4) | 0.89 | |
SCORE2, % | 2.95 (1.5–5.1) | 2.85 (1.5–4.7) | 3.6 (2–6.3)a | 2.7 (1.3–4.8)b | 0.04 | |
Laboratory parameters | Lipoprotein(a), mg/dl | 7.1 (2.9–25.7) | 3.6 (2.9–24.1) | 12 (4.5–33.2)a | 10 (5.4–24)c | 0.006 |
Low‑risk category <30 mg/dl | 230 (76.4) | 122 (76.3) | 47 (72.3) | 61 (80.3) | 0.54 | |
Intermediate risk category 30–50 mg/dl | 20 (6.6) | 8 (5) | 5 (7.7) | 7 (9.2) | 0.45 | |
High‑risk category >50 mg/dl | 51 (16.9) | 30 (18.8) | 13 (20) | 8 (10.5) | 0.22 | |
Fasting glucose, mmol/l | 5.3 (5–5.5) | 5.3 (5–5.5) | 5.3 (5.1–5.5) | 5.3 (5–5.5) | 0.84 | |
TG, mmol/l | 1.14 (0.9–1.64) | 1.17 (0.89–1.66) | 1.13 (0.9–1.45) | 1.13 (0.9–1.53) | 0.59 | |
TC, mmol/l | 5.27 (4.59–5.87) | 5.30 (4.64–5.9) | 5.31 (4.68–5.97) | 5 (4.46–5.8) | 0.48 | |
HDL‑cholesterol, mmol/l | 1.52 (1.25–1.81) | 1.52 (1.24–1.82) | 1.50 (1.26–1.83) | 1.51 (1.24–1.78) | 0.89 | |
LDL‑cholesterol, mmol/l | 3.32 (2.77–3.92) | 3.33 (2.72–3.91) | 3.43 (2.92–3.96) | 3.23 (2.67–3.87) | 0.41 | |
Fibrinogen, g/l | 2.81 (2.53–3.1) | 2.81 (2.5–3.1) | 2.88 (2.53–3.15) | 2.78 (2.57–3.09) | 0.81 | |
CRP, mg/l | 1.3 (0.6–2.4) | 1.2 (0.6–2.5) | 1.6 (0.9–2.7) | 1.1 (0.6–2.3) | 0.23 | |
After 3 months | ||||||
Statin use | 93 (30.9) | 51 (31.9) | 26 (40)b | 16 (21.1) | 0.049 | |
Laboratory parameters | Lipoprotein(a), mg/dl | 6.8 (2.9–25.7) | 3.5 (2.9–25.2) | 10.5 (4.4–29.2)a | 10.2 (5.1–22.5)c | 0.002 |
Low‑risk category <30 mg/dl | 232 (77.1) | 124 (77.5) | 47 (72.3) | 61 (80.3) | 0.52 | |
Intermediate risk category 30–50 mg/dl | 17 (5.6) | 5 (3.1) | 5 (7.7) | 7 (9.2) | 0.12 | |
High‑risk category >50 mg/dl | 52 (17.3) | 31 (19.4) | 13 (20) | 8 (10.5) | 0.2 | |
Fasting glucose, mmol/l | 5.3 (5–5.6) | 5.2 (4.9–5.7) | 5.3 (5–5.5) | 5.3 (5–5.6) | 0.84 | |
TG, mmol/l | 1.17 (0.86–1.64) | 1.22 (0.9–1.67) | 1.06 (0.83–1.68) | 1.11 (0.84–1.59) | 0.38 | |
TC, mmol/l | 5.21 (4.67–5.83) | 5.29 (4.71–5.93) | 5.14 (4.71–5.6) | 5.08 (4.54–5.73) | 0.1 | |
HDL‑cholesterol, mmol/l | 1.55 (1.27–1.79) | 1.57 (1.27–1.83) | 1.47 (1.25–1.75) | 1.55 (1.28–1.75) | 0.84 | |
LDL‑cholesterol, mmol/l | 3.25 (2.77–3.76) | 3.35 (2.71–3.88) | 3.16 (2.84–3.56) | 3.14 (2.77–3.73) | 0.51 | |
Fibrinogen, g/l | 2.92 (2.54–3.25) | 2.88 (2.53–3.22) | 2.92 (2.58–3.26) | 2.93 (2.53–3.25) | 0.8 | |
CRP, mg/l | 1.2 (0.7–2.3) | 1.2 (0.7–2.3) | 1.4 (0.8–2.4) | 1.2 (0.8–3) | 0.37 | |
Women, including 86 aged over 50 years (45.5%), had higher Lp(a) median (IQR) level than men (8 [3.3–34.8] mg/dl vs 5.4 [2.9–17.2] mg/dl; P = 0.014), including 11 (5.8%) with Lp(a) above 100 mg/dl, despite a similar age (50 [45–55] years vs 51 [45–56] years; P = 0.64). Baseline Lp(a) weakly correlated with triglycerides (R = –0.121; P = 0.04), LDL‑C (R = 0.114; P = 0.048), and fibrinogen (R = 0.135; P = 0.02). Hypercholesterolemia occurred in 216 patients (71.8%); however, only 20 (6.6%) were treated with statins prior to enrollment.
Follow‑up
None of the patients was lost to follow‑up. After 3 months, all routine tests yielded results similar to those at baseline, except for 3.9% higher median (IQR) fibrinogen level (2.81 [2.53–3.1] g/l vs 2.92 [2.54–3.25] g/l; P <0.001). The median (IQR) Lp(a) concentration after 3 months was 6.8 (2.9–25.7) mg/dl, and did not differ from baseline (P = 0.14), with the values higher by 62.5% in women than men (7.8 [3.3–32.9] mg/dl vs 4.8 [2.9–18.7] mg/dl; P = 0.014). There was an inverse correlation of Lp(a) with triglycerides (R = –0.139; P = 0.02), but not with LDL‑C and fibrinogen. During the 3 months, 93 individuals (30.9%) were on statins, and as expected, the agents had no impact on Lp(a) levels at 3 months (8.3 [2.9–33.6] mg/dl vs 7.6 [2.9–28.5] mg/dl; P = 0.12).
Of note, there were 141 individuals (47%) with Lp(a) variability equal to or above 10% at 3 months (median absolute Δ of –0.5 [–2.6 to 3.1] mg/dl and relative Δ Lp(a) of –11.4 [–19.6 to 24]%), including 65 patients (46%) with Lp(a) variability between 10% and 20% (median absolute Δ of –0.6 [–2.1 to 0.8] mg/dl and relative Δ Lp(a) of –12.1 [–15.9 to 13.5]%) and 76 (54%) with Lp(a) variability above 20% (median absolute Δ of 0.9 [–2.8 to 4.6] mg/dl and relative Δ Lp(a) of 22.2 [–27.1 to 34.8]%; Table 1). The individuals with Lp(a) variability of 10%–20% were slightly older and had higher SCORE2 than those with Lp(a) variability below 10% (P = 0.034 and P = 0.011) and above 20% (P = 0.048 and P = 0.03; Table 1). We found no associations of Lp(a) variability with active or former smoking (Table 1). However, a combined group of ever smokers (n = 104; 34.6%) more likely had Lp(a) variability over 20% (OR, 2.08; 95% CI, 1.22–3.54; P <0.001). The individuals with Lp(a) variability of 10%–20% and above 20% had higher Lp(a) concentrations at baseline and after 3 months than those with Lp(a) variability below 10% (Table 1). However, there were no associations between Lp(a) variability in the patients who were classified to the grey zone category at baseline and those with Lp(a) levels below 30 mg/dl or above 50 mg/dl (Supplementary material, Figure S1A,B), probably due to limited number of participants. Eleven patients (7.8% with Lp(a) variability equal to or above 10%; 3.7% of all, including 8 women [72.7%]) transitioned the EAS risk category, including 6 individuals (4.3%) with Lp(a) variability equal to or above 10% (2% of all) who exceeded the Lp(a) level of 30 mg/dl, and 5 (3.5%; with Lp(a) variability ≥10%; 1.7% of all) who exceeded Lp(a) level of 50 mg/dl (Supplementary material, Figure S2). Of note, 7 of the 11 individuals who transitioned the EAS risk category (ie, 35% of them were originally classified to the intermediate risk category) moved from the grey zone to the low‑risk (n = 4) or to the high‑risk category (n = 3). The Lp(a) concentration decreased in 6 patients (a median [IQR] decrease of 21.5 [15.5–26.7]%). Among them, 4 transitioned to the low‑risk category and 2 into the grey‑zone, while in 5 individuals the Lp(a) level increased (a median [IQR] increase of 26.6 [25.2–27.4]%), including over 30 mg/dl in 2 patients and over 50 mg/dl in 3 patients (Supplementary material, Table S1).
We found no associations among the relative change in Lp(a) level and most demographic, clinical, or laboratory parameters. However, a decrease in Lp(a) concentration at 3 months was more commonly observed in women, who initially had higher Lp(a) level and were in the grey zone or high‑risk category, while an increase in Lp(a) levels was more frequently observed in men with initially lower Lp(a) levels who transitioned to the high‑risk group.
Discussion
To our knowledge, this study is the first to evaluate the intraindividual Lp(a) variability in middle‑aged patients free of CVD within 3 months. We found temporal variability of Lp(a) level equal to or above 10% in about 50% of the studied cohort, but only 3.7% transitioned the EAS risk category. Given previous studies,8-11 it can be suggested that Lp(a) variability in middle‑aged participants free of CVD is lower (about 7% of those with Lp(a) >10 mg/dl) than in CVD patients, and this variability over short periods is concordant with an increasing number of reports.9,13 However, similarly to a previous report,10 individuals in the grey‑zone of Lp(a) level were more likely to be transitioned to a different risk category.
Enkhmaa et al13 reported that Lp(a) levels can vary by up to 25% within a few months, influenced by factors such as diet or hormonal changes. For example, lowering saturated fat intake over 2 months and replacing it with complex carbohydrates (DELTA1 and DELTA2; Dietary Effects on Lipoproteins and Thrombogenic Activity trials) or monosaturated fatty acids (DELTA2) was associated with 11%–20% elevation in the Lp(a) levels, with LDL‑C reduction by 7%–11%.13 In our population, we also noticed a high prevalence of hypercholesterolemia, concerning approximately 72% of individuals, which is by about 20% higher than reported in the WOBASZ II study of 5690 middle‑age Polish adults.14 The prevalence of elevated LDL‑C may be partially explained by increased family risk of CAD in the studied population; however, with no diagnosis of CVD yet. Moreover, 216 patients (71.8%) had BMI above 25 kg/m2, including 30 obese individuals (13.9%).
We were unable to identify factors associated with larger Lp(a) variability, except for the observation that women more often transitioned the risk category. Previous reports showed mostly that female sex was associated with higher Lp(a) variability,9,10 except for Matta et al11 who found slightly greater variability in men. Harb et al10 observed that women (45% of the studied population; median [IQR] age of 51.3 [31.7–63] years) had higher baseline Lp(a) levels and also higher median (IQR) of absolute change in Lp(a) levels than men (47 [17–81] mg/dl vs 31 [11–66] mg/dl; P <0.01 and 8 [3–7] mg/dl vs 5 [2–12.2] mg/dl; P = 0.01, respectively) during a median of 1 year between measurements. Moreover, Marcovina et al9 showed that women had greater variability than men (13 patients, including 8 women, with Lp(a) levels >30 mg/dl) among individuals with Lp(a) variability equal to or above 25%, while Matta et al11 found that only menopausal women more often had Lp(a) variability over 25%. It has been reported that Lp(a) variability may increase up to 30% in postmenopausal women, while hormonal replacement therapy was associated with a mean 25% reduction in Lp(a) levels.15,16 In our study, the influence of menopause on Lp(a) level is unlikely due to the short‑term follow‑up and to the fact that all the women who changed their EAS risk category were younger than 50 years. Therefore, it is more likely that other factors modifying Lp(a) levels were involved, supporting the need for repeated Lp(a) measurement, especially in borderline cases. However, the observation that Lp(a) levels decreased more often in women, and increased more often in men is intriguing and warrants further investigation on larger cohorts. The finding that higher SCORE2 was associated with lower variability in Lp(a) levels also deserves further studies. Besides sex, age was associated with Lp(a) variability,7-9,15 namely the elderly were more prone to greater variability. However, we did not confirm this association in our cohort, probably due to a relatively narrow age interval (40–65 years). We found higher OR for Lp(a) variability exceeding 20% in the patients with a history of smoking. Although smoking does not directly raise Lp(a) levels,17 it can enhance the inflammatory and atherogenic effects of high Lp(a) levels.3 Therefore, our data suggest that smoking may influence Lp(a) variability, which is particularly important for individuals in the grey zone and warrants further investigation.
Current guidelines do not recommend repeated Lp(a) measurements, as they are not expected to improve risk prediction.1,8 However, growing new evidence emphasizes consideration of repeated Lp(a) measurement in individuals who were primarily classified to the grey zone category, especially those with borderline Lp(a) values and comorbidities influencing Lp(a) levels, which might reduce the possibility of its over- or underestimation.8-11,18 Since elevated Lp(a) level is a recognized CVD risk factor,1 temporal Lp(a) variability can impact risk stratification and proper management, especially now, when several Lp(a)-lowering clinical trials are ongoing and different cutoff values have been used as the inclusion criterion.19 Currently, lifestyle changes and treatment of hypercholesterolemia, hypertension, and prediabetes / diabetes should be implemented in individuals with elevated Lp(a) level in the absence of targeted therapy capable of lowering it; no effect on Lp(a) is produced by statins, and about 20%–30% reduction is provided by proprotein convertase subtilisin/kexin 9 inhibitors.1,7,20
The study limitations should be acknowledged. The sample size was limited; however, representative of middle‑aged Polish patients. We did not evaluate liver and renal function markers, and while the medical history did not indicate any relevant diseases, some abnormalities cannot be excluded, though they are unlikely to affect the results. Moreover, all participants had a positive family history of CVD, thus the results may not reflect the variability of Lp(a) in a healthy population. Lastly, data regarding statin treatment (specific agents and daily doses) were incomplete, making the effects of primary prevention hard to assess in this study.21 However, this shortcoming seems to be of negligible importance, as statin treatment has no influence on Lp(a) levels.1,7,17
In conclusion, our study showed a relatively large intraindividual variability of Lp(a) levels in Polish middle‑aged patients, which might affect CVD risk assessment. The findings suggest that Lp(a) measurements should be repeated, especially in women and individuals in the grey zone. However, the optimal frequency for measuring Lp(a) and the acceptable short‑term variability due to nongenetic factors should be determined in larger cohorts of apparently healthy individuals in the advent of emerging Lp(a)-lowering therapies.
- Kronenberg F, Mora S, Stroes ESG, et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. Eur Heart J. 2022; 43: 3925‑3946. | Crossref
- Kopytek M, Ząbczyk M, Mazur P, et al. Oxidized phospholipids associated with lipoprotein(a) contribute to hypofibrinolysis in severe aortic stenosis. Pol Arch Intern Med. 2022; 132: 16372. | Crossref
- Visseren FLJ, Mach F, Smulders YM, et al. 2021 ESC Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J. 2021; 42: 3227‑3337.
- Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: Executive Summary: a Report of the American College of Cardiology / American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2019; 140: e563‑e595. | Crossref
- Mach F, Baigent C, Catapano AL, et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk. Eur Heart J. 2020; 41: 111‑188.
SUPPLEMENTARY MATERIAL
ARTICLE INFORMATION