Author + information
- Received January 19, 2000
- Revision received April 19, 2000
- Accepted June 19, 2000
- Published online November 1, 2000.
- Christie M Ballantyne, MD, FACC∗,* (, )
- J.Alan Herd, MD∗,
- Evan A Stein, MD, PhD†,
- Laura L Ferlic, MS∗,
- J.Kay Dunn, PhD∗,
- Antonio M Gotto Jr., MD, DPhil, FACC‡ and
- Ali J Marian, MD, FACC∗
- ↵*Reprint requests and correspondence:
Dr. Christie M. Ballantyne, Baylor College of Medicine, 6565 Fannin, M.S. A-601, Houston, Texas 77030
We sought to examine the association of apolipoprotein (apo) E genotypes with baseline plasma lipid levels and severity of coronary artery disease (CAD), as well as the response to treatment with fluvastatin in the Lipoprotein and Coronary Atherosclerosis Study (LCAS).
Apo E genotypes have been associated with plasma lipid levels and CAD. However, the influence of apo E genotypes on the response of plasma lipids and CAD progression or regression to statin treatment in patients with mildly to moderately elevated cholesterol remains unknown.
Apo E genotypes were determined by polymerase chain reaction and restriction mapping. Plasma lipids were measured at baseline and 12 weeks after therapy with fluvastatin or placebo in 320 subjects. In 287 subjects, quantitative coronary angiography was performed at baseline and after 2.5 years of treatment.
Subjects with the 3/3 genotype had greater reductions in total cholesterol (20.4% vs. 15.4%, p = 0.01) and low density lipoprotein (LDL) cholesterol (28.8% vs. 22.7%, p = 0.03) than did the subjects with the 3/4 or 4/4 genotype. In contrast, subjects with the 2/3 genotype (n = 10) had a greater increase in high density lipoprotein cholesterol (19.1%) than did the subjects with the 3/3 genotype (4.3%, p = 0.002) and those with the 3/4 or 4/4 genotype (7.0%, p = 0.02). Subjects with the 3/4 or 4/4 genotype had an increased frequency of previous angioplasty, but other measures of baseline CAD severity and baseline lipids did not differ significantly among the genotypes, nor did CAD progression or clinical events.
Although subjects with the ϵ4 allele had less reduction in LDL cholesterol with fluvastatin, they had similar benefit in terms of CAD progression.
Apolipoprotein (apo) E mediates the removal of lipoprotein remnants (chylomicron remnants and intermediate density lipoprotein) from the circulation. The gene encoding apo E has several variants, which are estimated to account for ∼10% of the interindividual variation in total serum cholesterol concentration (1). The three common alleles for apo E, in decreasing frequency, are ϵ3, ϵ4 and ϵ2, which code for the three major isoforms: apo E3 (contains cysteine at position 112 and arginine at position 158), apo E4 (contains arginine at positions 112 and 158) and apo E2 (contains cysteine at positions 112 and 158). In general, apo E2 and apo E4 isoforms have opposing effects on plasma lipids. Compared with the ϵ3 allele, ϵ4 is generally associated with increased plasma cholesterol and low density lipoprotein (LDL) cholesterol, and ϵ2 is associated with decreased plasma cholesterol and LDL cholesterol (2) and increased plasma triglycerides (3). Apo E genotypes have also been associated with the response of plasma lipids to lipid-lowering therapy, but the results have been conflicting. Although some studies report a greater cholesterol reduction in response to diet in individuals with the apo E4 isoform (4,5), others found no association (6,7). Similarly, the association of apo E genotypes with the response of plasma lipids to treatment with 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) remains unclear. Previous studies have been performed, primarily in patients with severe hypercholesterolemia, with conflicting results. Some studies in hyperlipidemic patients have found that those with apo E4 had lesser and those with apo E2 had greater reductions in LDL cholesterol in response to statin therapy (8–11), whereas others have reported no association between apo E genotypes and LDL cholesterol reduction (8,12–17). Apo E genotypes have also been associated with coronary artery disease (CAD), with the ϵ4 allele associated with an increased risk and the ϵ2 allele with a decreased risk for CAD (2).
In this study, we examined the association of apo E genotypes with plasma lipid levels and severity of CAD at baseline, and the response of plasma lipids and coronary atherosclerosis to treatment with fluvastatin in the Lipoprotein and Coronary Atherosclerosis Study (LCAS) population. LCAS included only CAD patients with mild to moderate elevations of LDL cholesterol, as is typical of the majority of patients with CAD.
All subjects provided written, informed consent, and the study was approved by the Institutional Review Board of Baylor College of Medicine. The design and primary results of LCAS have been published previously (18,19). In brief, men and postmenopausal women aged 35 to 75 years who had at least one coronary lesion causing 30% to 75% diameter stenosis on coronary angiography and LDL cholesterol of 2.97 mmol/liter to 4.91 mmol/liter (115 mg/dL to 190 mg/dL) after 10 weeks of lipid-lowering diet therapy were randomized to receive fluvastatin (40 mg) or placebo. Beginning at 12 weeks after randomization, subjects who had a mean prerandomization LDL cholesterol level ≥4.14 mmol/liter (160 mg/dL) despite diet (25% of all randomized subjects) were also given open-label adjunctive cholestyramine therapy, up to 12 g/day, as tolerated (mean 8 g/day). Excluded from the study were individuals who had a myocardial infarction within six months before randomization, a mean fasting plasma triglyceride level >3.39 mmol/liter (300 mg/dL) (>2.82 mmol/liter [250 mg/dL] if assigned cholestyramine), diabetes mellitus requiring insulin or an oral hypoglycemic agent, a fasting blood glucose level >9.4 mmol/liter (170 mg/dL), uncontrolled hypertension, previous coronary artery bypass graft surgery (CABG), atherectomy or coronary stenting.
Genotyping and measurement of plasma lipids
Processing of samples was done by laboratory personnel who had no knowledge of the angiographic and clinical data. Apo E genotyping was performed by polymerase chain reaction and restriction digestion with the HhaI enzyme, as published (20), with the exception that polyacrylamide gel electrophoresis was used to separate the digestion products.
Lipid concentrations were assessed in fasting blood samples drawn at weeks −2, 0 and 12 in a laboratory certified by the U.S. National Heart, Lung, and Blood Institute /Centers for Disease Control and Prevention Part III Program (Medical Research Laboratories, Highland Heights, Kentucky), as previously described (18). To determine the effect of fluvastatin on plasma lipid levels, lipid assessments at 12 weeks after randomization, before initiation of cholestyramine, were used for analysis.
Quantitative coronary angiography
Angiography was performed at baseline (1 to 16 weeks before randomization) and at final follow-up (as close to week 130 as possible), using the Cardiovascular Angiography Analysis System. Qualifying lesions were defined by minimal lumen diameter (MLD) ≥25% of the reference lumen diameter at baseline and MLD ≥0.8 mm less than the reference lumen diameter at either baseline or follow-up. Progression of CAD was assessed by the mean intrasubject per-lesion change in MLD, as well as categorization of subjects as having definite progression, definite regression or mixed angiographic change. Definite progression was defined as one or more qualifying lesion(s) with a decrease in MLD of ≥0.4 mm, including new total occlusions, and no qualifying lesion with an increase in MLD of ≥0.4 mm. Definite regression was defined as one or more qualifying lesion(s) with an increase in MLD of ≥0.4 mm, no qualifying lesion with a decrease in MLD of ≥0.4 mm and no new total occlusion.
Clinical events were defined as percutaneous transluminal coronary angioplasty (PTCA), CABG, definite or probable myocardial infarction, unstable angina pectoris requiring hospital admission and death from any cause.
Baseline lipid values were the average of values at weeks −2 and 0. Because of the potentially opposing effects of ϵ2 and ϵ4, subjects with the 2/4 genotype (n = 6) were excluded from all analyses; the remaining ϵ4 carriers (3/4 and 4/4 genotypes) were grouped together. None of the subjects was homozygous for the ϵ2 allele. Genotypes were compared with respect to baseline characteristics by the chi-square test (for categorical variables), one-way analysis of variance (for continuous variables) or the Kruskal-Wallis test (for the per-patient numbers of qualifying lesions and total occlusions). Analysis of variance, with changes in lipid values (total cholesterol, LDL cholesterol, HDL cholesterol, triglyceride) as the outcome variable and treatment (fluvastatin, placebo) and genotype (2/3, 3/3, 3/4 + 4/4) as the factors used to determine if the impact of treatment was different for the different genotypes. A Bonferroni procedure was performed to identify which of the genotype comparisons accounted for significant interaction effects. Potential differences among the genotypes for baseline MLD and change in MLD by treatment were determined by hierarchical regression analysis. Multinomial logistic regression analysis, with categorical angiographic change (progression, regression, mixed) as the outcome variable and treatment (fluvastatin, placebo) and genotype (2/3, 3/3, 3/4 + 4/4) as the factors used to determine whether the impact of treatment differed among the genotypes. For other angiographic indexes and clinical events, the Fisher exact test was used to compare genotypes stratified by treatment. The time to first clinical event was analyzed by the log-rank test. The MLD is reported as the mean value ± SE; all other continuous variables are presented as the mean value ± SD. All calculations were performed with Stata (Stata Corp., College Station, Texas) or the Statistical Analysis System (SAS Institute, Cary, North Carolina).
Of the 429 subjects in LCAS, 330 had apo E genotype data as well as baseline lipid and baseline angiographic data. After exclusion of subjects with the 2/4 genotype (n = 6), baseline characteristics were compared in 324 subjects (Table 1). Baseline and 12-week lipid values were available for 320 of these subjects, and baseline and follow-up (2.5-year) angiographic data were available for 287 subjects.
Each allele has a unique combination of HhaI fragment sizes. The expected digestion products of the ϵ3 allele were 91 base pairs (bp), 48 bp and 35 bp; those of the ϵ2 allele were 91 bp and 83 bp; and those of the ϵ4 allele were 72 bp, 48 bp, 35 bp and 19 bp. A combination of the aforementioned electrophoretic patterns identified the genotypes. In the entire study group, the distribution of apo E genotypes was 2/4 (n = 6), 2/3 (n = 22), 3/3 (n = 206), 3/4 (n = 88) and 4/4 (n = 8). Therefore, the allele frequencies was 0.04 for ϵ2, 0.79 for ϵ3 and 0.17 for ϵ4, similar to those in subjects in the Framingham Offspring Study (0.08, 0.78 and 0.13, respectively) (21).
Genotypes and plasma lipids
At baseline (Table 1), subjects in the 2/3 genotype group had slightly lower total cholesterol and LDL cholesterol than did the subjects in the 3/3 or 3/4 and 4/4 groups, but the differences were not statistically significant.
At 12-week follow-up (Table 2), there were significant interactions between apo E genotypes and fluvastatin therapy for the percent change in total cholesterol (p = 0.01), LDL cholesterol (p = 0.02) and high density lipoprotein (HDL) cholesterol (p = 0.02). Among the subjects treated with fluvastatin, those in the 3/4 and 4/4 group had significantly smaller mean reductions in total cholesterol (p = 0.01) and LDL cholesterol (p = 0.03) than did subjects in the 3/3 group (Fig. 1). In contrast, those with the 2/3 genotype had a significantly greater mean increase in HDL cholesterol than did those in the 3/3 group (p = 0.002) or 3/4 and 4/4 group (p = 0.02). Only 24.5% of the subjects in the 3/4 and 4/4 group had a ≥30% reduction in LDL cholesterol 12 weeks after treatment with fluvastatin, as compared with 50.0% of subjects with the 2/3 genotype and 52.0% of those with the 3/3 genotype.
Genotypes and coronary atherosclerosis
Subjects in the 3/4 and 4/4 group had a higher frequency of previous PTCA (p = 0.02), as compared with subjects in the 2/3 and 3/3 groups. There were no other significant differences among the genotypes in indexes of severity of CAD at baseline (Table 1).
Regarding apo E genotypes and progression or regression of CAD, modest differences, albeit not statistically significant, were noted (Table 3). Among the subjects receiving placebo, approximately one-half of those in the 3/4 and 4/4 group had progression of CAD, in contrast to only 10% of subjects with the 2/3 genotype. Similarly, the subjects receiving placebo in the 3/4 and 4/4 group had a greater loss in MLD (−0.140 ± 0.037) than did those with the 2/3 genotype (−0.014 ± 0.075).
Overall, there was no significant treatment–genotype interaction between apo E genotype and fluvastatin therapy with respect to progression or regression of CAD. However, there was a tendency toward greater angiographic benefit with fluvastatin in subjects with the 3/4 or 4/4 genotype and less in subjects with the 2/3 genotype, as compared with subjects with the 3/3 genotype. The treatment effect of fluvastatin, calculated as the difference between MLD change with fluvastatin and MLD change with placebo, was somewhat greater in subjects in the 3/4 and 4/4 genotype group—0.092 as compared with 0.077 in subjects in the 3/3 genotype group—but the difference was not statistically significant.
Genotypes and clinical events
In subjects with angiograms that could be evaluated, clinical event rates (Table 3) and time to first event were not significantly different among the genotype groups. These results were similar to those in the overall study group.
The main finding of this study is the association of apo E genotypes with the response of plasma lipids to treatment with fluvastatin. In the LCAS population, subjects with the 3/3 genotype had greater reductions in plasma total cholesterol and LDL cholesterol in response to treatment with fluvastatin than did those with the 3/4 or 4/4 genotype. In contrast, subjects with the 2/3 genotype had a greater increase in HDL cholesterol than did those with the 3/3 genotype and those with the 3/4 or 4/4 genotype. Angiographic assessment of CAD showed a tendency toward greater benefit with fluvastatin on CAD progression, measured by either MLD change or categorical assessment, in subjects with the 3/4 or 4/4 genotype; however, the differences were not statistically significant. Overall, there were no significant genotype–treatment interactions with regard to quantitative angiographic indexes of progression or regression of CAD. Thus, apo E genotypes were associated with the response of plasma lipids to fluvastatin, but not with progression or regression of CAD in the LCAS population.
Apo E genotypes and lipids
LCAS is the first study to examine the association of apo E genotypes with the response of plasma lipids to treatment with a statin in subjects with CAD and mildly to moderately elevated LDL cholesterol, as is commonly seen in patients with CAD. A similar association has been observed in some (8–10), but was not significant in other (8,12–16) studies, which included subjects with higher LDL cholesterol levels, often because of familial hypercholesterolemia. The existing data from these studies and ours suggest that subjects with the ϵ2 allele have a greater LDL cholesterol reduction and subjects with ϵ4 allele have a lesser response to statin therapy. In a meta-analysis of five studies (8–10,12,16), which included a total of 625 subjects, those with the 2/2 or 2/3 genotype had significantly greater and those with the 3/4 or 4/4 genotype had significantly smaller LDL cholesterol reductions in response to statin treatment than did those with the 3/3 genotype (10). In all except one (16) of the studies comparing the effects of statin therapy on lipids among apo E genotypes, the subjects were already following a lipid-lowering diet at “baseline,” which may have confounded the association between the apo E genotypes and the response of plasma lipids to statins (10). In LCAS, there was no observed relation between apo E genotype and response to dietary therapy before randomization (data not shown).
A significant interaction between apo E genotypes and response of HDL cholesterol to treatment with fluvastatin was also observed in the LCAS population. Subjects with the 2/3 genotype had a significantly greater increase in HDL cholesterol levels as compared with those with the 3/3 genotype and those with the 3/4 or 4/4 genotype. This observation is novel, but a similar trend has been reported previously. Korhonen et al. (14) observed a 16% increase in HDL cholesterol in subjects with the 2/3 genotype, as compared with an 8% increase in subjects with the 3/3 genotype and a 1% decrease in subjects with the 4/4 genotype, but the difference was not statistically significant. In LCAS, subjects on fluvastatin with the 2/3 genotype did have slightly higher fasting triglyceride levels at baseline (as might be expected because of the association reported between ϵ2 and triglyceride ), as well as slightly greater reductions in triglyceride levels after 12 weeks of treatment, although the differences were not statistically significant. Patients with higher triglyceride levels and low HDL cholesterol have been shown to have a greater increase in HDL cholesterol with statin therapy (22). Low HDL cholesterol levels were common in the subjects in LCAS (23), as in patients with CAD in general, and mean HDL cholesterol levels in LCAS were lower than those in most other studies that examined apo E genotype and response to statin therapy. Enhanced clearance of remnant particles and a reduction in postprandial lipemia may be related to the greater HDL cholesterol increases with statins observed in both patients with low HDL cholesterol and individuals with the 2/3 genotype.
The frequencies of apo E alleles in the LCAS population were similar to those in the Framingham Offspring Study overall (21). However, there was a higher frequency of subjects with the ϵ3 allele in LCAS (0.79) than in the subset of the Framingham Offspring Study subjects with CAD (0.73), and there was a lower frequency of subjects with the ϵ2 allele in LCAS (0.04) than in the Framingham Offspring Study subjects with CAD (0.09) (21). The LCAS entry criteria may account for the observed differences. In LCAS, exclusion of subjects with triglyceride levels >3.39 mmol/liter (300 mg/dL; >2.82 mmol/liter [250 mg/dL] in subjects assigned cholestyramine) may account for the lower percentage of subjects with the ϵ2 allele, which has been associated with hypertriglyceridemia, and the exclusion of subjects with LDL cholesterol >4.91 mmol/liter (190 mg/dL) may have reduced the percentage of subjects with the ϵ4 allele, which is associated with both hypercholesterolemia (3,24) and CAD (21). The combined result of excluding high LDL cholesterol and high triglyceride levels may have led to the significantly increased frequency of ϵ3, as compared with the Framingham Offspring Study subjects with CAD. Selection of subjects based on specific entry criteria may also account for the lack of a significant association between apo E genotypes and baseline plasma lipid levels.
Apo E binds to lipids, heparan sulfate proteoglycans and lipoprotein receptors (LDL receptor and LDL receptor–related protein) and modulates lipoprotein levels by influencing the clearance rate, lipolytic conversion and very low density lipoprotein (VLDL) triglyceride production (25). Apo E genotypes have been postulated to affect plasma lipid levels because of differences in binding of apo E isoforms to receptors. Apo E2 has markedly decreased binding affinity as compared with apo E3, whereas the binding affinity of apo E3 and apo E4 has generally been found to be the same (26), although some reports suggest that apo E4 has increased affinity as compared with apo E3 (27,28). In addition, apo E isoforms influence the binding of apo E to lipoproteins, with apo E4 preferentially binding to VLDL and apo E3 and apo E2 binding to the smaller, phospholipid-rich HDL (29). In general, ϵ2 has been estimated to decrease LDL cholesterol by 12.5% and to decrease HDL cholesterol by 3.1%, as compared with ϵ3, whereas individuals with ϵ4 generally have higher concentrations of both LDL cholesterol and HDL cholesterol, with estimated increases of 6.4% and 1.9%, respectively (30). In LCAS, subjects with the ϵ4 allele had a mean LDL cholesterol reduction of 22.7%, with only 25% having reductions in LDL cholesterol of ≥30%, whereas individuals homozygous for the ϵ3 allele had a mean LDL cholesterol reduction of 28.8%, with 52% having reductions in LDL cholesterol of ≥30%. One possible hypothesis to explain the smaller reduction in LDL cholesterol in response to statin therapy in subjects with the ϵ4 allele is that HMG-CoA reductase activity is already suppressed in these individuals because of the rapid uptake of lipoprotein remnants. Therefore, individuals with the ϵ4 allele may have a lesser response and may require a higher statin dosage to achieve maximal benefit.
Apo E genotype and CAD
In LCAS, subjects with ϵ4 were significantly more likely to have had PTCA before entry into the study. Although other measures of baseline severity of CAD were not significantly different among the genotypes, all lesions in vessels with previous PTCA were excluded from the analyses; therefore, baseline severity may have been underestimated in these subjects because of the predefined criteria for lesion analysis. In a meta-analysis of five observational angiographic studies enrolling a total of 1,686 individuals, risk for angiographic CAD was significantly increased (odds ratio 1.11) with ϵ4 as compared with ϵ3; CAD risk tended to be lower with the ϵ2 allele (odds ratio 0.76) than with ϵ3, but the difference was not statistically significant (31). Although there were a limited number of subjects with the ϵ2 allele in LCAS, evidence to support the reduced risk for CAD with the ϵ2 allele includes the lower frequency of the ϵ2 allele (0.04) in LCAS than in the Framingham Offspring Study overall (0.08) (21) and the observation that LCAS subjects with the ϵ2 allele had the same severity of CAD but had increased nonlipid risk factors for CAD (increased age, tobacco and diabetes), as compared with other LCAS subjects. In a study of 424 subjects (110 without significant CAD; 118 with one-vessel, 96 with two-vessel and 100 with three-vessel disease), the frequency of the ϵ4 allele was directly associated and the frequency of the ϵ2 allele was inversely associated with the number of diseased vessels (32). Although the correlation was mediated largely through circulating levels of apo B and apo B–containing lipoproteins, the significant association between apo E genotype and CAD severity persisted after controlling for these levels.
LCAS is the first trial of statin therapy to examine the association of apo E genotype with angiographic indexes of progression or regression of CAD. Among LCAS subjects, CAD progression occurred to a similar extent (as measured by mean MLD change) and at a similar frequency (as measured by categorical assessment), regardless of apo E genotype. Subjects with the ϵ2 allele had little progression of CAD, but the sample size was small and the difference was not statistically significant. However, subjects with the 3/4 or 4/4 genotype, who, on the basis of observational epidemiologic data, might be expected to have increased CAD risk, had a somewhat greater treatment effect with fluvastatin, as determined by the difference between MLD change with placebo and MLD change with fluvastatin, as compared with subjects with the 2/3 or 3/3 genotype, but the difference was not statistically significant and the sample size limits the power to assess differences in angiographic indexes between the apo E alleles. In LCAS overall, LDL cholesterol change was not a good predictor of MLD change (33).
Because LCAS was an angiographic trial, it was not designed to detect statistically significant differences in clinical events. As would be expected, no significant differences in event rates or time to first event were detected among the apo E genotype groups. However, in larger studies and in pooled analyses, the ϵ4 allele has been associated with increased risk for CAD events. In a meta-analysis of nine observational studies that measured clinical events in a total of 6,355 individuals, risk for a CAD event was significantly increased, with an odds ratio of 1.26 for individuals with the ϵ4 allele as compared with those with ϵ3 (31). In a nested case–control analysis of 619 participants in the Multiple Risk Factor Intervention Trial (MRFIT), ϵ4 was associated with increased risk for nonfatal myocardial infarction, and particularly for CAD death, even after adjustment for differences in baseline LDL cholesterol, baseline HDL cholesterol, body mass index, smoking and diastolic blood pressure (34). In a substudy of 966 patients with CAD in the Scandinavian Simvastatin Survival Study (4S), individuals with the ϵ4 allele who received placebo had an increased risk for death, with an odds ratio of 1.8, which was abolished by treatment with simvastatin (35).
The completion of the Human Genome Project will greatly accelerate efforts to understand pharmacogenetic interactions that determine both disease course and response to therapy. There are now substantial data that apo E genotypes should be considered in studies of interventions designed to reduce the risk for atherosclerosis. In LCAS, fluvastatin therapy produced significantly greater reductions in total cholesterol and LDL cholesterol in subjects with apo E genotype 3/3 than in subjects with 3/4 or 4/4, and a significantly greater increase in HDL cholesterol in subjects with the 2/3 genotype than in those with 3/3 or in those with 3/4 or 4/4. Although subjects with the ϵ4 allele had significantly less LDL cholesterol reduction with fluvastatin, they had similar benefit on CAD progression.
The authors acknowledge the editorial assistance of Kerrie Jara.
☆ The Lipoprotein and Coronary Atherosclerosis Study was supported by a grant from Novartis Pharmaceuticals Corporation, East Hanover, New Jersey. Drs. Ballantyne and Marian are recipients of Established Investigator Awards from the American Heart Association.
- coronary artery bypass graft surgery
- coronary artery disease
- high density lipoprotein
- 3-hydroxy-3-methylglutaryl coenzyme A
- Lipoprotein and Coronary Atherosclerosis Study
- low density lipoprotein
- minimal lumen diameter
- percutaneous transluminal coronary angioplasty
- very low density lipoprotein
- Received January 19, 2000.
- Revision received April 19, 2000.
- Accepted June 19, 2000.
- American College of Cardiology
- Mahley R.W,
- Rall S.C Jr.
- Davignon J,
- Gregg R.E,
- Sing C.F
- Lopez-Miranda J,
- Ordovas J.M,
- Mata P,
- et al.
- DELTA Research Group,
- Lefevre M,
- Ginsberg H.N,
- Kris-Etherton P.M,
- et al.
- Korhonen T,
- Hannuksela M.L,
- Seppänen S,
- Kervinen K,
- Antero Kesäniemi Y,
- Savolainen M.J
- Hixson J.E,
- Vernier D.T
- Ballantyne C.M,
- Herd J.A,
- Ferlic L.L,
- et al.
- Ordovas J.M,
- Litwack-Klein L,
- Wilson P.W.F,
- Schaefer M.M,
- Schaefer E.J
- Weisgraber K.H,
- Innerarity T.L,
- Mahley R.W
- Bohnet K,
- Pillot T,
- Visvikis S,
- Sabolovic N,
- Siest G
- Mamotte C.D.S,
- Sturm M,
- Foo J.I,
- van Bockxmeer F.M,
- Taylor R.R
- Dong L.-M,
- Weisgraber K.H
- Wilson P.W.F,
- Schaefer E.J,
- Larson M.G,
- Ordovas J.M
- Wang X.L,
- McCredie R.M,
- Wilcken D.E.L
- Ballantyne C.M,
- Herd J.A,
- Ferlic L.L,
- Dunn J.K,
- Stein E.A,
- Gotto A.M Jr.
- Gerdes L.U,
- Gerdes C,
- Kervinen K,
- et al.