Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The association of apolipoprotein E genotype with cholesterol levels and atherosclerosis: three randomized clinical trials
(USC Thesis Other)
The association of apolipoprotein E genotype with cholesterol levels and atherosclerosis: three randomized clinical trials
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE ASSOCIATION OF APOLIPOPROTEIN E GENOTYPE WITH CHOLESTEROL LEVELS AND ATHEROSCLEROSIS: THREE RANDOMIZED CLINICAL TRIALS By Seung Sook You A Thesis Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree MASTER OF SCIENCE (BIOSTATISTICS) May 2015 Copyright 2015 Seung Sook You ii ACKNOWLEDGEMENTS I deeply appreciate everyone who has helped me through the completion of my Master’s thesis. I would like to express my gratitude and special appreciation to my committee chair, Dr. Wendy Mack, for her endless guidance, support, and wisdom throughout the research process. She always managed to find time in her busy schedule to answer my questions, and I could not have finished my thesis without her help. I would also like to express my deepest gratitude to my committee members, Dr. Hooman Allayee and Dr. Howard Hodis, for their assistance and invaluable feedback. A special thanks also goes to my family. I could not have succeeded without the priceless support and help of you all: Mom, Dad, Lisa, Billy, Brian, and Mocci. Also, I would like to give special thanks to Min, who always supported and encouraged me whenever the future started to seem uncertain. Words cannot express my appreciation and love for you all. Thank you, everyone! iii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ii LIST OF TABLES iv ABSTRACT vi CHAPTER 1: INTRODUCTION 1 CHAPTER 2: MATERIALS AND METHODS 2.1. Study Population and Design 2.2. Assessment of Atherosclerosis Progression 2.3. Laboratory Measurements 2.4. Statistical Analysis 4 4 7 8 9 CHAPTER 3: RESULTS 3.1. Baseline Characteristics 3.2. Association of APOE Genotype with Cholesterol Levels (All Trials Combined) 3.3. Association of APOE Genotype with Cholesterol Levels (Excluding ELITE Trial Participants) 3.4. Association of APOE Genotype with CIMT (All Trials Combined) 3.5. Association of APOE Genotype with CIMT (Excluding ELITE Trial Participants) 11 11 24 31 37 44 CHAPTER 4: DISCUSSION 50 REFERENCES 54 iv LIST OF TABLES Table 1 Baseline Characteristics by Three Trials 12 Table 2 APOE Genotype Frequencies (All Trials Combined) 14 Table 3 Total, HDL, and LDL Cholesterol Levels, and CIMT by APOE Genotypes (All Trials Combined) 16 Table 4 P-values for Comparison of Total, HDL, LDL Cholesterol Levels, CIMT, and CIMT Progression within the E2 and E4 Genotype Groups (All Trials Combined) 17 Table 5 Total, HDL, and LDL Cholesterol Levels, and CIMT by APOE Genotypes (Excluding ELITE Trial Participants) 18 Table 6 P-values for Comparison of Total, HDL, LDL Cholesterol Levels, CIMT, and CIMT Progression within the E2 and E4 Genotype Groups (Excluding ELITE Trial Participants) 19 Table 7 Baseline Characteristics by APOE Genotype Groups (All Trials Combined) 21 Table 8 APOE Genotype Frequencies (Excluding ELITE Trial Participants) 22 Table 9 Baseline Characteristics by APOE Genotype Groups (Excluding ELITE Trial Participants) 23 Table 10 Cholesterol Levels at Baseline and During Trial Follow-up (All Trials Combined) 26 Table 11 Comparison of Total Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 27 Table 12 Comparison of LDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 28 Table 13 Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 29 Table 14 Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotypes (All Trials Combined) 30 Table 15 Cholesterol Levels at Baseline and During Trial Follow-up (Excluding ELITE Trial Participants) 32 v Table 16 Comparison of Total Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 33 Table 17 Comparison of LDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 34 Table 18 Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 35 Table 19 Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotypes (Excluding ELITE Trial Participants) 36 Table 20 CIMT Measurement at Baseline and During Trial Follow-up (All Trials Combined) 39 Table 21 Overall Comparison of CIMT (mm) by APOE Genotype Groups (All Trials Combined) 40 Table 22 Comparison of CIMT (mm) by APOE Genotype Groups: Stratification by Gender (All Trials Combined) 41 Table 23 Overall Comparison of CIMT Progression (mm/year) by APOE Genotype Groups (All Trials Combined) 42 Table 24 Comparison of CIMT Progression (mm/year) by APOE Genotype Groups: Stratification by Gender (All Trials Combined) 43 Table 25 CIMT Measurement at Baseline and During Trial Follow-up (Excluding ELITE Trial Participants) 45 Table 26 Overall Comparison of CIMT (mm) by APOE Genotype Groups (Excluding ELITE Trial Participants) 46 Table 27 Comparison of CIMT (mm) by APOE Genotype Groups in Females (Excluding ELITE Trial Participants) 47 Table 28 Overall Comparison of CIMT Progression (mm/year) by APOE Genotype Groups (Excluding ELITE Trial Participants) 48 Table 29 Comparison of CIMT Progression (mm/year) by APOE Genotype Groups in Females (Excluding ELITE Trial Participants) 49 vi ABSTRACT Background and Purpose – Apolipoprotein E (APOE) plays an important role in cholesterol metabolism, which can influence the development of atherosclerosis. We evaluated if APOE genotype affects cholesterol levels, subclinical atherosclerosis as assessed by carotid artery intima-media thickness (CIMT) measurements, and the progression of atherosclerosis through repeated CIMT measurements. Methods – Study subjects, aged 32 to 88 years with no clinical evidence of cardiovascular disease, participated in three randomized clinical trials. Six hundred four females and 180 males were divided into three groups according to their APOE genotype: E2, including E2/E2, E2/E3, and E2/E4 genotypes; E3, including E3/E3 genotype; and E4, including E3/E4 and E4/E4 genotypes. Results – The E2 allele was significantly associated with lower total and low-density lipoprotein (LDL) cholesterol levels, whereas the E4 group exhibited higher total and LDL cholesterol levels. APOE genotype variations did not significantly affect high- density lipoprotein (HDL) cholesterol levels. Among all female participants, the E2 allele group had significantly lower CIMT levels than E3/E3 genotype carriers. After excluding the Early versus Late Intervention Trial with Estradiol (ELITE) trial participants, we observed no significant difference in CIMT levels among the three genotype groups. In males, APOE genotype was not significantly associated with CIMT level. Finally, variations in APOE genotype were not significantly related to the rate of CIMT progression in male or female participants. Conclusions – The present study indicated that the E2 allele was associated with low total and LDL cholesterol levels, while E3/E4 and E4/E4 genotype carriers had higher vii total and LDL cholesterol levels. Female E2 allele carriers were associated with lower atherosclerosis level compared to female E3/E3 carriers; however, exclusion of ELITE trial participants negated this association. In male, atherosclerosis level did not significantly differ across the APOE groups. Additionally, the APOE genotype was not associated with HDL cholesterol levels or atherosclerosis progression in middle-aged individuals with no clinical evidence of cardiovascular disease. 1 CHAPTER 1: INTRODUCTION Approximately 83.6 million Americans suffered from cardiovascular disease (CVD) in 2010 [1]. While CVD-related deaths have declined from 2000 to 2010, they remain the number one cause of death in the United States [1]. The most common cause of CVD is a complex multifactorial condition called atherosclerosis. Atherosclerosis is characterized by the accumulation of lipids and fibrous deposits of plaques in the arterial walls. These plaques may be due to excessive inflammation and fibro-proliferative reactions in response to various types of injuries and stimuli, such as elevated blood cholesterol levels, high blood pressure, smoking, excessive weight, diabetes, and a family history of heart disease [2–4]. The accumulation of plaque can narrow the arteries, leading to reduced blood supply to vital organs, such as the heart, brain, and intestines [5]. Development of plaque is initiated by the arterial accumulation and oxidation of low-density lipoproteins (LDLs) [6]. Plaque also includes extracellular lipids, such as cholesterol and cholesteryl esters, macrophages, smooth muscle cells, and connective tissue products [7]. Apolipoprotein E (APOE) has an important role in transporting lipids from one tissue or cell type to another [8]. The interactions of this protein with LDL receptors make it an important determinant of cholesterol metabolism [8]. The human APOE gene is located on chromosome 19 and has three major alleles, designated as E2, E3, and E4 depending on the amino acids present at residues 112 and 158. E3 contains a cysteine at residue 112 and an arginine at residue 158, the E2 allele contains cysteine residues at both sites, and E4 contains arginine at both sites [9]. In addition, E3 is the most common isoform, present in more than half of the population, whereas the range of E4 and E2 frequencies 2 are about 5%–35% and 1%–5%, respectively [8]. In most populations, E4 frequency exceeds E2 frequency with few exceptions [8]. LDL receptors are cell-surface proteins that transport cholesterol from plasma into cells [10]. Residues 136–150 of APOE directly interact with LDL receptors, while residue 158 affects the receptor-binding region indirectly [8]. APOE isoforms differ in their capability to bind LDL receptors. For example, E4 binds LDL receptors with a slightly higher affinity than E3, whereas E2 binds LDL receptors with a 2% lower binding affinity than E3 and E4 [8, 11]. Due to various receptors binding affinity, APOE has roles in the metabolism of cholesterol [12, 13]. An extensive meta-analysis (86,067 disease-free participants in 82 studies) by Bennet et al. demonstrated decreased plasma LDL cholesterol levels in E2 heterozygous or homozygous individuals compared to individuals with the E3/E3, E3/E4, or E4/E4 genotypes [26]. In contrast, E4 has been associated with higher levels of LDL cholesterol [26]. Similarly, in the Framingham Heart Study, E2 carriers had lower LDL cholesterol levels compared to the E3/E3, E3/E4, and E4/E4 genotypes; however, the presence of an E4 allele was associated with increased LDL cholesterol. Furthermore, in females, the E2/E2 and E2/E3 genotypes were associated with lower carotid intimal-medial thickness (CIMT), as a marker of atherosclerosis, compared to E3/E3 genotype carriers. Additionally, the E4 allele has been associated with increased internal CIMT in diabetic men [15]. Although several studies have investigated the association between APOE genotype and atherosclerosis, the results are inconclusive, as some studies could not confirm a 3 difference in atherosclerosis across the genotypes. Because APOE genetic variants are associated with different lipid metabolism rates, we used data from three randomized clinical trials to investigate the effect of APOE genetic variants on cholesterol levels and CIMT, a validated measure of subclinical atherosclerosis, in a racially diverse population of middle-aged people free of CVD. We used longitudinally measured cholesterol levels and CIMT to evaluate these associations. 4 CHAPTER 2: MATERIALS AND METHODS 2.1. Study Population and Design The study population was derived from three randomized, double-blind, placebo- controlled trials: the B-Vitamin Atherosclerosis Intervention Trial (BVAIT), the Women’s Isoflavone Soy Health (WISH), and the Early versus Late Intervention Trial with Estradiol (ELITE). All three trials were conducted and completed at the Atherosclerosis Research Unit, University of Southern California. 506 participants were randomized in BVAIT to test the effect of high-dose B-Vitamin supplementation on subclinical atherosclerosis progression. The participants were men and postmenopausal women ≥40 years old with fasting total homocysteine (tHcy) ≥8.5 µmol/L and no clinical signs/symptoms of cardiovascular disease (CVD). Exclusion criteria were fasting triglycerides >500mg/dL, diabetes mellitus or fasting serum glucose >126 mg/dL, systolic blood pressure (SBP) ≥160 mm Hg and/or diastolic blood pressure (DBP) ≥100 mm Hg, untreated thyroid disease, creatinine clearance <70 mL/min, life threatening disease with prognosis <5 years, or >5 alcoholic drinks daily. Vital signs, clinical events, diet, and non-study medication and supplement/nutraceutical used were ascertained every 3 months during participants’ clinic visits. Every 6 months, laboratory determinations were performed [16]. 350 participants were randomized in the WISH trial to test the effect of isoflavone soy protein (ISP) on the progression of subclinical atherosclerosis in postmenopausal women. The subjects were postmenopausal women without vaginal bleeding >1 year and serum estradiol <20 pg/mL and no clinical signs and symptoms of CVD. Exclusion criteria were clinical signs, symptoms, or a personal history of CVD, diabetes mellitus or 5 fasting serum glucose >126 mg/dL, fasting triglycerides >500 mg/dL, SBP ≥160 mm Hg and/or DBP ≥100 mm Hg, untreated thyroid disease, serum creatinine >2 mg/dL, life threatening disease with prognosis <5 years, or >5 alcoholic drinks daily or substance abuse, taking menopausal hormone therapy or soy, nut, or related food allergies. Clinic visits occurred every month for the first 6 months and then every other month for the remainder of the trial. Every 6 months, laboratory determinations were performed [17]. 643 participants were randomized in ELITE to test the differential effects of oral 17beta-estradiol treatment in early and late postmenopausal women with no clinical signs and symptoms of CVD. The subjects were postmenopausal women with a serum level of total 17beta-estradiol <25 pg/mL and absence of vaginal bleeding for at least 6 months or bilateral oophorectomy. Exclusion criteria were use of menopausal hormone therapy within the preceding 1 month, history or evidence of CVD, diabetes mellitus, or fasting serum glucose >140mg/dL, uncontrolled hypertension (DBP >110 mmHg or SBP >150 mmHg), fasting plasma triglyceride level >500 mg/dL, untreated thyroid disease, renal insufficiency (serum creatinine >2.0mg/dL), life threatening disease, liver disease, history of deep vein thrombosis or pulmonary embolism, history of breast cancer, and inability to participate in outcome assessments. Clinic visits occurred every month for the first 6 months and every 2 months thereafter. Every 6 months, laboratory determinations were performed [18]. In the three trials, 1499 subjects were randomized. ISP [19] and oral 17beta-estradiol treatment [20] can affect cholesterol levels. As such, we excluded participants in the treatment group from the analysis (175 in the WISH). Since the ELITE trial is still blinded to the treatment allocation, the treatment group was not identified. We excluded 6 participants who were taking cholesterol-lowering medications or estrogen-containing hormone medications during the trials (242 in the BVAIT, 52 in the WISH, and 214 in the ELITE). One hundred and seventy-five participants in the treatment group and 508 participants who were taking cholesterol-lowering medications or estrogen-containing hormones medications were excluded from the data analyses. APOE genotype data were not available for 32 participants, leaving a total of 784 subjects included in these analyses. 7 2.2. Assessment of Atherosclerosis Progression CIMT is a measure of subclinical atherosclerosis [21]. CIMT measures the thickness of the inner two layers (intima and media) of the carotid artery. When carotid ultrasounds are carefully obtained to enhance longitudinal reproducibility, CIMT can be measured as often as needed and it has been shown consistently to predict future vascular events [21]. High-resolution B-mode ultrasound images of the right common carotid artery were obtained with a 7.5-MHz linear array transducer attached to a Toshiba SSH 140A ultrasonography system (Toshiba Corp, Tokyo, Japan) for BVAIT and ELITE trials and an Advanced Technology Laboratories, Inc. (ATL) Apogee ultrasound system (Bothell, WA) for WISH. Participants were examined supine with the head rotated 45 degrees towards the left side. Imaging was done in the plane parallel to the neck with the jugular vein lying immediately above the common carotid artery [22]. The transducer was then rotated 90 degrees around the central line of the transverse image of the stacked jugular vein and carotid artery to obtain a longitudinal image while the stacked position of the vessels was maintained [23]. All images contained internal anatomical landmarks for reproducing probe angulation. All examinations were recorded with the electrocardiogram signal and these techniques have resulted in significant reductions in measurement variability [17]. For all three trials, CIMT measurements were performed at baseline and repeated every 6 months during trial follow-up. 8 2.3. Laboratory Measurements After a fast of at least 8 hours, blood samples were obtained to measure cholesterol levels and DNA extraction for APOE genotype. Total plasma cholesterol and triglyceride levels were measured using an enzymatic method standardized by the Centers for Disease Control and Prevention Standardization Program. High-density lipoprotein (HDL) levels were determined after lipoproteins containing apolipoprotein B were precipitated in whole plasma by using heparin manganese chloride [24]. LDL cholesterol values were calculated by using the Friedewald equation [25]. Genomic DNA was extracted from the buffy coat using DNeasy isolation kits (Qiagen, Valencia, California). APOE genotype was defined by two SNPs, rs429358 and rs7412 at codon positions 112 and 158, respectively. 9 2.4. Statistical Analysis Participant demographic characteristics included age, gender, race, marital status, and education level. Basic clinical variables assessed were smoking status, SBP, DBP, and body mass index (BMI). Differences in continuous variables across APOE genotypes were tested by analysis of variance (ANOVA) for normally distributed variables and the Kruskal-Wallis test for non-normally distributed variables. Differences in categorical variables across APOE genotype groups were tested by the Chi-square test. The distributions of cholesterol levels and CIMT were tested for normality using histograms and normality plots. We used the generalized estimating equation model with a compound symmetry covariance structure to determine the association of the APOE genotype with cholesterol levels: total cholesterol, HDL cholesterol, and LDL cholesterol. The model included as independent variables the APOE genotype and the follow-up time, measured as months since randomization. Follow-up time was treated as a categorical variable. A mixed model with random effects for intercept and slope and an unstructured covariance structure was used to assess the association of the APOE genotype with level and rate of change of CIMT. The model included the APOE genotype, follow-up time treated as a continuous variable (measured as years since randomization), and the APOE genotype x follow-up time interaction term. The interaction term tested whether the APOE genotype influenced the rate of change in CIMT. The covariance matrix for the generalized estimating equation models was selected using the minimum quasi-likelihood information criterion (QIC), comparing autoregressive, compound symmetry, and unstructured covariance structures. The 10 minimum Akaike Information Criterion (AIC) was used to determine a covariance structure selecting from autoregressive, compound symmetry, and unstructured covariance structures in the mixed model. Due to the unknown treatment code of the ELITE trial, the analyses were repeated after exclusion of all 427 participants enrolled in the ELITE trial. The regression models contained the baseline covariates: age as a continuous variable, gender, and race categorized as White, Black, Hispanic, Asian, or Other. Two- way interaction terms between the APOE genotype and age, gender, or race were added and tested to determine if the associations between APOE genotype and response variables differed with respect to age, gender, or race. When there was evidence of any interaction effects, analyses were stratified into the subgroups. Factors such as smoking status, BMI, DBP, and SBP are known correlates of cholesterol levels and atherosclerosis, and these factors were included in the models as possible confounders. However, smoking status did not alter the magnitude of the estimated associations of the APOE genotypes on the response variables. Furthermore, the correlation between DBP and SBP was high (r = 0.70). Therefore, only BMI and SBP were included as confounders in the models. Statistical analysis was performed with the Statistical Analysis Software (SAS 9.3). The level of statistical significance was set at a two-sided P ≤0.05 for all analyses, except interaction terms, for which P ≤0.10 was considered significant. 11 CHAPTER 3: RESULTS 3.1. Baseline Characteristics In total, 784 subjects were included in the analysis of cholesterol levels. One subject had CIMT measurements that were more than 12 times the standard deviation from the CIMT mean (2.1250 and 2.1470 mm). Because this outlier could contribute an artificially high influence on the association between APOE genotype and CIMT, this subject was excluded from CIMT analyses; a total of 783 subjects were included in CIMT analyses. The demographic and clinical characteristics of the BVAIT, ELITE, and WISH trial participants included in this analysis are shown in Table 1. By trial design, only female participants were randomized in the ELITE and WISH trials. The mean age of the participants was 59.5 years, and the participants were primarily women (77%), and primarily non-Hispanic white (66.2%). The most common genotype in all three trials was the E3/E3 genotype, found in 60.3% of participants. 12 Table 1 – Baseline Characteristics by Three Trials Variable BVAIT a ELITE b WISH a Total n (%) 241 (30.7) 427 (54.5) 116 (14.80) 784 Age, mean (SD) 59.7 (10.1) 59.1 (7.0) 60.4 (6.9) 59.5 (8.1) Gender, n (%) Female Male 61 (25.3) 180 (74.7) 427 (100.0) 0 (0.0) 116 (100.0) 0 (0.0) 604 (77.0) 180 (23.0) Race, n (%) White Black Hispanic Asian Other 152 (63.0) 39 (16.2) 24 (10.0) 24 (10.0) 2 (0.8) 288 (67.5) 45 (10.5) 58 (13.6) 36(8.4) 0 (0.0) 79 (68.1) 4 (3.5) 16 (13.8) 12 (10.3) 5 (4.3) 519 (66.2) 88 (11.2) 98 (12.5) 72 (9.2) 7 (0.9) Marital status, n (%) Single Married Divorced or widowed 20 (8.3) 163 (67.9) 52 (23.8) 33 (7.7) 243 (56.9) 151 (35.4) 11 (9.5) 65 (56.0) 40 (34.5) 64 (8.2) 471 (60.1) 248 (31.7) Education, n (%) ≤ High school > High school 13 (5.4) 228 (94.6) 17 (4.0) 410 (96.0) 2 (1.7) 114 (98.3) 32 (4.1) 752 (95.9) Smoking status, n (%) Current Former Never smoked 12 (5.0) 148 (61.9) 79 (33.1) 17 (4.0) 163 (38.2) 247 (57.8) 2 (1.7) 44 (37.9) 70 (60.4) 31 (4.0) 286 (36.6) 465 (59.4) Blood pressure (mmHg), mean (SD) SBP DBP 126.0 (13.3) 78.9 (70) 114.7 (10.6) 73.5 (6.1) 117.5 (12.1) 73.9 (6.4) 118.6 (12.7) 75.2 (6.8) BMI (kg/m 2 ), mean (SD) 28.1 (5.0) 26.7 (5.4) 26.6 (5.2) 27.1 (5.3) Lipids at baseline (mg/dL), mean (SD) Total cholesterol LDL cholesterol HDL cholesterol Triglycerides 208.2 (32.3) 130.0 (28.6) 54.2 (13.2) 120.3 (62.3) 224.5 (34.3) 137.9 (31.9) 67.6 (18.1) 95.6 (46.7) 220.8 (30.8) 137.1 (30.5) 62.4 (16.7) 106.6 (48.8) 219.0 (34.0) 135.3 (30.9) 62.7 (17.6) 104.8 (53.4) CIMT at baseline (mm), mean (SD) 0.76 (0.16) 0.76 (0.10) 0.82 (0.09) 0.77 (0.13) 13 Table 1 (Continued) – Baseline Characteristics by Three Trials Variable BVAIT a ELITE b WISH a Total APOE genotype, n (%) E2/2 E2/3 E2/4 E3/3 E3/4 E4/4 4 (1.7) 34 (14.1) 1 (0.4) 150 (62.2) 44 (18.3) 8 (3.3) 3 (0.7) 47 (11.0) 7 (1.7) 257 (60.2) 100 (23.4) 13 (3.0) 0 (0.0) 12 (10.4) 2 (1.7) 66 (56.9) 36 (31.0) 0 (0.0) 7 (0.9) 93 (11.9) 10 (1.3) 473 (60.3) 180 (22.9) 21 (2.7) APOE group, n (%) c E2 E3 E4 39 (16.2) 150 (62.2) 52 (21.6) 57 (13.4) 257 (60.2) 113 (26.4) 14 (12.1) 66 (56.9) 36 (31.0) 110 (14.1) 473 (60.3) 201 (25.6) a Exclude subjects who were taking medications for cholesterol-lowering or estrogen- containing hormones during trials or were randomized to the active treatment group. b Exclude subjects who were taking medications for cholesterol-lowering or estrogen- containing hormones during trial. c E2: E2/E2, E2/E3, and E2/E4 E3: E3/E3 E4: E3/E4 and E4/E4 14 Table 2 summarizes the frequencies of the APOE genotypes stratified by gender. The E2/E2 genotype frequency in females (n = 4) was 0.6%, the E2/E2 genotype frequency in males (n = 3) was 1.7%, and the E2/E4 genotype in males (n = 1) was 0.6%. Because there were few subjects with the E2/E2 (n = 7) and E2/E4 (n = 10) genotypes, APOE genotypes were broadly categorized into three groups for analysis: E2 (E2/E2, E2/E3, and E2/E4), E3 (E3/E3), and E4 (E3/E4 and E4/E4). The E3 group was defined as the reference group. A study by Bennet et al. showed that cholesterol levels in the E2/E4 genotype were more similar to those of the E2/E2 and E2/E3 genotypes than the cholesterol levels in the E3/E4 and E4/E4 genotypes, providing further rationale that the E2/E4 genotype be included in the E2 group [26]. Table 2 – APOE Genotype Frequencies (All Trials Combined) E2/E2 E2/E3 E2/E4 E3/E3 E3/E4 E4/E4 Female, n (%) 4 (0.6) 68 (11.3) 9 (1.5) 363 (60.1) 145 (24.0) 15 (2.5) Male, n (%) 3 (1.7) 25 (13.9) 1 (0.6) 110 (61.1) 35 (19.4) 6 (3.3) Total, n (%) 7 (0.9) 93 (11.8) 10 (1.3) 473 (60.3) 180 (23.0) 21 (2.7) 15 As shown in Table 3, while the mean total cholesterol and LDL cholesterol levels of the E2/E4 genotype were closer to the mean total and LDL cholesterol levels of the E2/E2 and E2/E3 genotypes than to the mean levels of the E3/E4 and E4/E4 genotypes, there were no notable differences in the mean CIMT values between these groups. As shown in Tables 4, there were no significant differences in total cholesterol levels, LDL cholesterol levels, CIMT, or the CIMT progression rate within the E2 group (E2/E2, E2/E3, and E2/E4) or the E4 group (E3/E4 and E4/E4). Even after excluding the ELITE trial participants, the conclusions remained the same (Tables 5 and 6). HDL cholesterol levels were significantly different among a subset of APOE genotypes within the E2 genotype group (E2/E2 and E2/E3, p < 0.0001) across follow-up visits in all trials (Table 4, p = 0.04). After excluding the ELITE trial participants, HDL cholesterol levels were also significantly different among the E4 genotypes (E3/E4 and E4/E4, p = 0.008) after adjusting for follow-up time, age, race, gender, trial, BMI, and SBP (Table 6, p = 0.03). Analyses comparing HDL cholesterol levels were therefore repeated within the specific APOE genotypes. 16 Table 3 – Total, HDL, and LDL Cholesterol Levels, and CIMT by APOE Genotypes (All Trials Combined) E2/E2 E2/E3 E2/E4 E3/E3 E3/E4 E4/E4 Number of subjects: Cholesterol levels 7 93 10 473 180 21 Number of total cholesterol level measurements obtained 55 856 80 4047 1482 198 Total cholesterol (mg/dL), mean (SD) 178.5 (36.5) 203.6 (31.1) 183.8 (27.5) 215.8 (30.9) 220.6 (32.8) 225.7 (31.3) Number of HDL cholesterol level measurements obtained 55 856 80 4047 1482 198 HDL cholesterol (mg/dL), mean (SD) 53.2 (7.8) 67.3 (21.2) 59.6 (14.3) 66.3 (19.5) 65.8 (18.0) 61.3 (18.8) Number of LDL cholesterol level measurements obtained 55 852 80 4041 1479 198 LDL cholesterol (mg/dL), mean (SD) 95.4 (29.9) 112.9 (29.8) 104.4 (23.0) 129.0 (27.2) 134.6 (30.8) 139.5 (29.8) Number of subjects: CIMT 7 93 10 473 179 21 Number of CIMT measurements obtained 54 841 79 3971 1453 196 CIMT (mm), mean (SD) 0.70 (0.06) 0.76 (0.11) 0.74 (0.06) 0.78 (0.12) 0.79 (0.14) 0.75 (0.10) 17 Table 4 – P-values for Comparison of Total, HDL, LDL Cholesterol Levels, CIMT, and CIMT Progression within the E2 and E4 Genotype Groups (All Trials Combined) Total cholesterol HDL cholesterol LDL cholesterol CIMT CIMT Progression E2 a E4 b E2 E4 E2 E4 E2 E4 E2 E4 Model 1 0.13 c 0.36 c 0.04 c 0.27 c 0.28 c 0.35 c 0.38 d 0.26 d 0.17 d 0.94 d Model 2 0.18 0.24 0.07 0.30 0.33 0.30 0.56 0.45 0.17 0.93 Model 3 0.18 0.23 0.13 0.22 0.25 0.30 0.53 0.46 0.17 0.94 All values are expressed as p-value. Significant p-values (< 0.05) are shown in bold. a E2: Compares mean levels among E2/E2, E2/E3, and E2/E4 genotypes b E4: Compares mean levels among E3/E4 and E4/E4 genotypes c P-values corresponding to cholesterol levels were generated using generalized estimating equation models with a compound symmetry covariance structure d P-values corresponding to CIMT and CIMT progression were generated using mixed models with an unstructured covariance structure Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 18 Table 5 – Total, HDL, and LDL Cholesterol Levels, and CIMT by APOE Genotypes (Excluding ELITE Trial Participants) E2/E2 E2/E3 E2/E4 E3/E3 E3/E4 E4/E4 Number of subjects: Cholesterol levels 4 46 3 216 80 8 Number of total cholesterol level measurements obtained 33 357 20 1569 533 70 Total cholesterol (mg/dL), mean (SD) 167.8 (31.8) 207.6 (37.2) 183.3 (45.5) 210.7 (31.6) 215.5 (32.0) 226.1 (34.1) Number of HDL cholesterol level measurements obtained 33 357 20 1569 533 70 HDL cholesterol (mg/dL), mean (SD) 53.8 (7.8) 57.4 (16.1) 48.8 (11.2) 56.8 (15.7) 58.0 (16.6) 50.1 (10.6) Number of LDL cholesterol level measurements obtained 33 353 20 1563 530 70 LDL cholesterol (mg/dL), mean (SD) 90.4 (26.1) 122.8 (31.7) 113.4 (30.9) 130.8 (27.2) 135.6 (27.3) 143.9 (29.8) Number of subjects: CIMT 4 46 3 216 79 8 Number of CIMT measurements obtained 33 347 19 1535 519 70 CIMT (mm), mean (SD) 0.68 (0.06) 0.76 (0.12) 0.81 (0.05) 0.77 (0.13) 0.80 (0.14) 0.71 (0.12) 19 Table 6 – P-values for Comparison of Total, HDL, LDL Cholesterol Levels, CIMT, and CIMT Progression within the E2 and E4 Genotype Groups (Excluding ELITE Trial Participants) Total cholesterol HDL cholesterol LDL cholesterol CIMT CIMT Progression E2 a E4 b E2 E4 E2 E4 E2 E4 E2 E4 Model 1 0.18 c 0.29 c 0.69 c 0.11 c 0.18 c 0.40 c 0.45 d 0.08 d 0.82 d 0.47 d Model 2 0.15 0.11 0.55 0.33 0.20 0.28 0.52 0.12 0.83 0.47 Model 3 0.13 0.12 0.19 0.03 0.20 0.26 0.58 0.09 0.83 0.46 All values are expressed as p-value. Significant p-values (< 0.05) are shown in bold. a E2: Compares mean levels among E2/E2, E2/E3, and E2/E4 genotypes b E4: Compares mean levels among E3/E4 and E4/E4 genotypes c P-values corresponding to cholesterol levels were generated using generalized estimating equation models with a compound symmetry covariance structure d P-values corresponding to CIMT and CIMT progression were generated using mixed models with an unstructured covariance structure Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 20 Table 7 illustrates the baseline participant characteristics, stratified by the three APOE genotype groups. Demographic and clinical characteristics, including age (p = 0.48), gender (p = 0.47), race (p = 0.10), marital status (p = 0.22), education (p = 0.70), smoking history (p = 0.68), SBP (p = 0.39), DBP (p = 0.18), and BMI (p = 0.19) did not significantly differ across the APOE groups. 21 Table 7 – Baseline Characteristics by APOE Genotype Groups (All Trials Combined) Variable E3 a E2 E4 P-value b n (%) 473 (60.4) 110 (14.0) 201 (25.6) N/A Age, mean (SD) 59.3 (8.0) 60.3 (8.3) 59.4 (8.1) 0.48 Gender, n (%) Female Male 363 (76.7) 110 (23.3) 81 (73.6) 29 (26.4) 160 (79.6) 41 (20.4) 0.47 Race, n (%) White Black Hispanic Asian Other 309 (65.3) 47 (9.9) 62 (13.1) 51 (10.8) 4 (0.9) 75 (68.2) 14 (12.7) 9 (8.2) 9 (8.2) 3 (2.7) 135 (67.2) 27 (13.4) 27 (13.4) 12 (6.0) 0 (0.0) 0.10 Marital status, n (%) Single Married Divorced or widowed 45 (9.5) 275 (58.3) 152 (32.2) 10 (9.1) 69 (62.7) 31 (28.2) 9 (4.5) 127 (63.2) 65 (32.3) 0.22 Education, n (%) ≤ High school > High school 19 (4.0) 454 (96.0) 6 (5.5) 104 (94.5) 7 (3.5) 194 (96.5) 0.70 Smoking history, n (%) Current Former Never smoked 16 (3.4) 171 (36.2) 285 (60.4) 7 (6.4) 40 (36.7) 62 (56.9) 8 (4.0) 75 (37.3) 118 (58.7) 0.68 Blood pressure (mmHg), mean (SD) SBP DBP 118.5 (13.0) 75.3 (7.0) 120.1 (11.9) 76.1 (6.5) 118.0 (12.5) 74.6 (6.8) 0.39 0.18 BMI (kg/m 2 ), mean (SD) 27.1 (5.5) 28.0 (5.6) 26.7 (4.5) 0.19 a E3: E3/E3 E2: E2/E2, E2/E3, and E2/E4 E4: E3/E4 and E4/E4 b Group comparisons used ANOVA for continuous variables or the Chi-square test for categorical Variables. DBP and BMI are compared by the Kruskal-Wallis test. 22 Tables 8 and 9 summarize the characteristics of the 357 participants remaining after exclusion of the ELITE trial participants, stratified by APOE genotype groups. No significant differences in baseline characteristics were observed among the three APOE groups. Table 8 – APOE Genotype Frequencies (Excluding ELITE Trial Participants) E2/E2 E2/E3 E2/E4 E3/E3 E3/E4 E4/E4 Female, n (%) 1 (0.6) 21 (11.9) 2 (1.1) 106 (59.9) 45 (25.4) 2 (1.1) Male, n (%) 3 (1.7) 25 (13.9) 1 (0.6) 110 (61.1) 35 (19.4) 6 (3.3) Total, n (%) 4 (1.1) 46 (12.9) 3 (0.8) 216 (60.5) 80 (22.4) 8 (2.3) 23 Table 9 – Baseline Characteristics by APOE Genotype Groups (Excluding ELITE Trial Participants) Variable E3 a E2 E4 P-value b n (%) 216 (60.5) 53 (14.8) 88 (24.7) N/A Age, mean (SD) 59.7 (9.1) 61.6 (9.3) 59.5 (9.5) 0.36 Gender, n (%) Female Male 106 (49.1) 110 (50.9) 24 (45.3) 29 (54.7) 47 (53.4) 41 (46.6) 0.63 Race, n (%) White Black Hispanic Asian Other 133 (61.6) 27 (12.5) 26 (12.0) 26 (12.0) 4 (1.9) 39 (73.6) 6 (11.3) 3 (5.7) 2 (3.7) 3 (5.7) 59 (67.0) 10 (11.4) 11 (12.5) 8 (9.1) 0 (0.0) 0.19 Marital status, n (%) Single Married Divorced or widowed 22 (10.2) 130 (60.5) 63 (29.3) 4 (7.6) 35 (66.0) 14 (26.4) 5 (5.7) 63 (71.6) 20 (22.7) 0.43 Education, n (%) ≤ High school > High school 8 (3.7) 208 (96.3) 3 (5.7) 50 (94.3) 4 (4.6) 84 (95.4) 0.80 Smoking history, n (%) Current Former Never smoked 7 (3.3) 73 (33.9) 135 (62.8) 3 (5.8) 18 (34.6) 31 (59.6) 4 (4.5) 32 (36.4) 52 (59.1) 0.90 Blood pressure (mmHg), mean (SD) SBP DBP 123.1 (13.9) 77.6 (7.2) 125.3 (12.7) 77.5 (7.1) 122.5 (13.0) 76.5 (7.2) 0.45 0.54 BMI (kg/m 2 ), mean (SD) 27.7 (5.4) 28.1 (4.5) 27.1 (4.5) 0.48 a E3: E3/E3 E2: E2/E2, E2/E3, and E2/E4 E4: E3/E4 and E4/E4 b Group comparisons used ANOVA for continuous variables or the Chi-square test for categorical Variables. DBP and BMI are compared by the Kruskal-Wallis test. 24 3.2. Association of APOE Genotype with Cholesterol Levels (All Trials Combined) The 784 participants had a mean (range) of 24.2 (1–83) months of follow-up in the E3 group, 24.9 (1–78) months of follow-up in the E2 group, and 24.2 (1–76) months of follow-up in the E4 group (p = 0.66). Mean values of total cholesterol, LDL cholesterol, and HDL cholesterol levels at baseline and during trial follow-up are summarized in Table 10. Although total cholesterol and LDL cholesterol levels decreased in all the APOE genotype groups during the trials, HDL cholesterol levels increased in the APOE groups from baseline to the end of the trial. Tables 11–13 summarize the mean difference in cholesterol levels among the three APOE genotype groups. Overall, APOE genotype was significantly associated with total cholesterol and LDL cholesterol levels (all p < 0.001) (Tables 11 and 12). Model 1 analyses adjusted only for follow-up time; model 2 analyses additionally adjusted for age, race, gender, and trial; model 3 further adjusted for BMI and SBP. In models 1-3, subjects with the E2 allele had significantly lower mean total cholesterol levels (all p < 0.001) and lower mean LDL cholesterol levels (all p < 0.001) than subjects with the E3/E3 genotype. In contrast, subjects in the E4 group had significantly higher mean total cholesterol and LDL cholesterol levels than the E3/E3 genotype carriers in models 1-3. Comparisons of HDL cholesterol levels among the three APOE genotype groups are summarized in Table 13. When only follow-up time was adjusted, HDL cholesterol levels did not significantly differ among the APOE genotype groups (p = 0.90). With additional adjustments for age, race, gender, trial, BMI, and SBP, the non-significant result remained unchanged. Table 14 summarizes the association of HDL cholesterol levels 25 with specific APOE genotypes. APOE genotype was not significantly associated with HDL cholesterol levels in models 1 – 3 (all p > 0.05). Two-way interaction terms for APOE group by age, APOE group by gender, and APOE group by race were tested simultaneously for cholesterol levels in model 3. However, there was no significant interaction between APOE genotype groups with age, gender, or race, for total cholesterol levels (p = 0.81, 0.18, and 0.21 respectively), LDL cholesterol levels (p = 0.67, 0.12, and 0.85 respectively), and HDL cholesterol levels (p = 0.67, 0.24, and 0.36 respectively). 26 Table 10 – Cholesterol Levels at Baseline and During Trial Follow-up (All Trials Combined) Baseline (mg/dL), mean (SD) Overall (mg/dL), mean (SD) Change (mg/dL), mean (SD) E3 a E2 E4 E3 E2 E4 E3 E2 E4 Total cholesterol 219.7 (33.2) 204.0 (32.7) 225.3 (34.3) 215.8 (30.9) 200.6 (32.0) 221.2 (32.7) -4.6 (24.6) -3.3 (22.8) -2.6 (25.0) LDL cholesterol 136.1 (29.2) 117.8 (30.0) 142.9 (31.8) 129.0 (27.2) 111.2 (29.6) 135.2 (30.7) -7.2 (23.2) -4.7 (20.0) -6.4 (22.8) HDL cholesterol 63.1 (17.7) 63.1 (17.8) 61.5 (17.1) 66.3 (19.5) 65.9 (20.5) 65.3 (18.2) 1.9 (10.3) 1.5 (9.2) 3.3 (9.7) a E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 201): E3/E4 and E4/E4 27 Table 11 – Comparison of Total Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -14.7 (3.1) 5.6 (2.4) 216.1 (1.3) 201.4 (2.8) 221.7 (2.1) < 0.001 < 0.001 0.019 < 0.001 Model 2 d -14.9 (3.1) 5.2 (2.3) 214.1 (3.1) 199.2 (3.9) 219.3 (3.5) < 0.001 < 0.001 0.023 < 0.001 Model 3 d -14.7 (3.1) 5.1 (2.3) 213.9 (3.1) 199.1 (3.9) 219.0 (3.5) < 0.001 < 0.001 0.027 < 0.001 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 201): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 28 Table 12 – Comparison of LDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -16.9 (2.8) 6.2 (2.2) 128.9 (1.2) 112.0 (2.6) 135.1 (1.9) < 0.001 < 0.001 0.004 < 0.001 Model 2 d -17.2 (2.8) 6.2 (2.2) 131.0 (2.8) 113.8 (3.5) 137.3 (3.3) < 0.001 < 0.001 0.004 < 0.001 Model 3 d -17.5 (2.8) 6.5 (2.2) 130.9 (2.9) 113.4 (3.5) 137.4 (3.3) < 0.001 < 0.001 0.003 < 0.001 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 201): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 29 Table 13 – Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -0.58 (1.9) -0.58 (1.4) 65.69 (0.86) 65.11 (1.7) 65.11 (1.2) 0.90 0.76 0.68 > 0.99 Model 2 d -0.55 (1.7) -1.33 (1.3) 59.00 (1.4) 58.45 (1.9) 57.67 (1.6) 0.57 0.75 0.29 0.68 Model 3 d 0.37 (1.6) -2.00 (1.2) 59.21 (1.4) 59.59 (1.8) 57.25 (1.5) 0.21 0.81 0.09 0.19 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 201): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 30 Table 14 – Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotypes (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a E3/E3 b E2/E2 E2/E3 E2/E4 E3/E4 E4/E4 E3/E3 E2/E2 E2/E3 E2/E4 E3/E4 E4/E4 Model 1 c -11.37 (2.2) 0.78 (2.1) -5.75 (4.7) -0.13 (1.5) -4.45 (3.7) 65.69 (0.9) 54.33 (2.1) 66.47 (1.9) 59.94 (4.7) 65.56 (1.3) 61.24 (3.6) 0.13 Model 2 c -9.63 (3.8) 0.88 (1.9) -7.79 (4.7) -1.01 (1.3) -4.11 (3.2) 59.14 (1.4) 49.51 (4.0) 60.02 (2.1) 51.35 (4.5) 58.13 (1.7) 55.04 (3.3) 0.16 Model 3 c -6.01 (2.7) 1.45 (1.7) -5.54 (4.3) -1.60 (1.2) -5.00 (2.7) 59.34 (1.4) 53.33 (3.0) 60.79 (2.0) 53.80 (4.1) 57.73 (1.6) 54.34 (2.9) 0.09 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b E3/E3 (n = 473), E2/E2 (n = 7), E2/E3 (n = 93), E2/E4 (n = 10), E3/E4 (n = 180), E4/E4 (n = 21) c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 31 3.3. Association of APOE Genotype with Cholesterol Levels (Excluding ELITE Trial Participants) The 357 participants had a mean (range) of 18.0 (1–63) months of follow-up in the E3 group, 18.6 (1–59) months of follow-up in the E2 group, and 17.2 (1–58) months of follow-up in the E4 group (p = 0.22). Mean values of total cholesterol, LDL cholesterol, and HDL cholesterol levels at baseline and during trials are presented in Table 15. Similar to LDL cholesterol and HDL cholesterol levels for all the trials, mean total cholesterol levels decreased and HDL cholesterol levels increased from baseline in all the APOE groups. However, mean total cholesterol levels increased in E3 carriers, whereas decreases were observed in E2 and E4 carriers. As shown in Table 16, in models 1-3, participants in the E2 group had significantly lower total cholesterol levels than those in the E4 (p = 0.015, 0.014, and 0.014, respectively), but not the E3 group (p = 0.14, 0.11, and 0.11, respectively). Excluding ELITE trial participants, the relationship between the APOE genotype groups and LDL cholesterol levels, and the relationship between the APOE genotype and HDL cholesterol levels remained. LDL cholesterol levels in the E2 group were significantly lower than in the E3 and E4 groups (Table 17). HDL cholesterol levels did not significantly among the three APOE genotype group or among the specific APOE genotypes (Tables 18 and 19). 32 Table 15 – Cholesterol Levels at Baseline and During Trial Follow-up (Excluding ELITE Trial Participants) Baseline (mg/dL), mean (SD) Overall (mg/dL), mean (SD) Change (mg/dL), mean (SD) E3 a E2 E4 E3 E2 E4 E3 E2 E4 Total cholesterol 211.2 (31.5) 204.8 (35.4) 219.8 (31.0) 210.7 (31.6) 203.3 (39.0) 216.8 (32.4) 0.33 (22.8) -1.1 (23.8) -2.0 (22.9) LDL Cholesterol 132.1 (28.4) 121.2 (31.9) 139.3 (28.5) 130.8 (27.2) 119.7 (32.4) 136.5 (27.7) -0.78 (21.2) -0.86 (21.3) -1.76 (21.6) HDL Cholesterol 56.5 (14.9) 57.4 (13.4) 57.0 (16.1) 56.8 (15.7) 56.7 (15.5) 57.1 (32.4) 0.46 (7.8) 0.71 (9.8) 0.20 (7.6) a E3 (n = 216): E3/E3 E2 (n = 53): E2/E2, E2/E3, and E2/E4 E4 (n = 88): E3/E4 and E4/E4 33 Table 16 – Comparison of Total Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -7.6 (5.2) 6.0 (3.4) 212.6 (2.0) 205.0 (4.9) 218.7 (3.0) 0.044 0.14 0.08 0.02 Model 2 d -8.2 (5.2) 5.7 (3.3) 214.6 (3.4) 206.4 (5.5) 220.3 (4.0) 0.044 0.11 0.08 0.01 Model 3 d -8.2 (5.1) 5.5 (3.3) 214.7 (3.4) 206.5 (5.5) 220.2 (4.0) 0.046 0.11 0.09 0.01 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 88): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 34 Table 17 – Comparison of LDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -11.5 (4.3) 6.2 (2.9) 133.1 (1.7) 121.6 (4.0) 139.3 (2.5) 0.001 0.007 0.03 0.0001 Model 2 d -12.4 (4.3) 6.1 (2.9) 134.6 (3.1) 122.3 (4.6) 140.7 (3.6) 0.001 0.004 0.04 < 0.001 Model 3 d -12.3 (4.2) 6.2 (2.9) 134.6 (3.1) 122.3 (4.6) 140.7 (3.6) 0.001 0.004 0.04 < 0.001 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 88): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 35 Table 18 – Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotype Groups (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d 0.87 (2.13) 0.13 (1.89) 56.18 (1.02) 57.05 (1.97) 56.31 (1.62) 0.92 0.68 0.95 0.76 Model 2 d 0.70 (2.01) -0.48 (1.62) 55.44 (1.5) 56.14 (2.1) 54.96 (1.8) 0.88 0.73 0.77 0.61 Model 3 d 0.98 (1.9) -0.89 (1.5) 55.39 (1.4) 56.37 (2.0) 54.50 (1.7) 0.67 0.61 0.55 0.38 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b P-values were generated from generalized estimating equation models with a compound symmetry covariance structure for comparison between groups. c E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 88): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 36 Table 19 – Comparison of HDL Cholesterol Levels (mg/dL) by APOE Genotypes (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a E3/E3 b E2/E2 E2/E3 E2/E4 E3/E4 E4/E4 E3/E3 E2/E2 E2/E3 E2/E4 E3/E4 E4/E4 Model 1 c -3.29 (3.7) 1.20 (2.3) 1.32 (9.6) 0.80 (2.0) -6.51 (3.5) 56.19 (1.0) 52.90 (3.5) 57.39 (2.1) 57.51 (9.5) 56.99 (1.7) 49.68 (3.4) 0.58 Model 2 c -2.45 (3.5) 1.02 (2.2) -0.08 (7.7) -0.21 (1.7) -2.93 (3.1) 55.43 (1.5) 52.98 (3.7) 56.44 (2.3) 55.35 (7.3) 55.22 (1.9) 52.50 (3.4) 0.91 Model 3 c -4.07 (2.6) 1.58 (2.0) -1.75 (7.4) -0.42 (1.6) -5.21 (2.7) 55.40 (1.5) 51.33 (2.8) 56.97 (2.2) 53.65 (7.1) 54.98 (1.8) 50.19 (2.9) 0.43 a P-values were generated from generalized estimating equation models with a compound symmetry covariance structure b E3/E3 (n = 216), E2/E2 (n = 4), E2/E3 (n = 46), E2/E4 (n = 3), E3/E4 (n = 80), E4/E4 (n = 8) c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 37 3.4. Association of APOE Genotype with CIMT (All Trials Combined) The 783 participants had a mean (range) of 23.4 (1–82) months of follow-up in the E3 group, 24.0 (1–77) months of follow-up in the E2 group, and 23.5 (1–75) months of follow-up in the E4 group (p = 0.71). At baseline, E2 carriers had lower mean CIMT measurements than the E3 and E4 APOE type carriers and subjects in the E4 group had the largest mean CIMT value (Table 20). The average change in CIMT of all the three groups increased over trial follow-up, consistent with atherosclerosis progression. Table 21 summarizes the overall association of the APOE genotype groups with CIMT. After adjusted for follow-up time, age, race, gender, trial, BMI, and SBP, significant differences in CIMT value were detected across the APOE genotype groups (p = 0.01). In model 3, the mean CIMT was significantly lower in E2 carriers than in E3 (p = 0.002) and E4 group carriers (p = 0.01). Additionally, the presence of a statistically significant interaction (p = 0.07) between gender and the APOE genotype group suggested that gender should be analyzed separately. Table 22 summarizes the effect of the APOE genotype on mean CIMT levels stratified by gender. In males, the CIMT values did not significantly differ among the three APOE genotype groups after adjustment for follow-up time (p = 0.68). This null finding remained with further adjustments, approaching statistical significance when adjusted for age, race, gender, trial, BMI, and SBP (p = 0.06). In females, the effect of the APOE genotype on CIMT measurements was significantly different from that observed in males. In model 1 analysis with adjustments for follow-up time, the E2 group showed significantly lower CIMT values than the E3 and E4 groups (p = 0.04). After additional adjustments, compared with the E3 group, the 38 E2 group still showed significantly lower CIMT values (in model 2: p = 0.007, in model 3: p = 0.005); however, there was no significant difference between the values of E2 and E4 groups (in model 2: p = 0.14, in model 3: p = 0.12). CIMT values did not significantly differ between the E3 and E4 groups in any model. Two-way interactions between APOE group by age and APOE group by race were stratified by gender and tested simultaneously in model 3. No evidence of an interaction effect between APOE genotype groups with age (female p = 0.87; male p = 0.97, respectively) or race (female p = 0.96; male p = 0.86, respectively) was observed in either females or males. Tables 23 and 24 present the association between APOE genotype group and CIMT progression. In the mixed effects regression model, there was no statistically significant interaction of the APOE group x time effect in overall or in gender-stratified analysis, indicating that the rate of change in CIMT over the trial follow-up did not differ among APOE genotype groups. In males, the E2 group had a higher rate of CIMT progression whereas the E4 group had a lower rate of CIMT progression than the other two groups. In contrast, in females, the E4 genotype group had the fastest progression of CIMT and the E2 group showed the lowest rate of CIMT progression among the three APOE groups. However, the differences in the rates of CIMT progression were not significant as shown in Table 24 (all p > 0.05 in models 1–3). 39 Table 20 – CIMT Measurement at Baseline and During Trial Follow-up (All Trials Combined) Baseline (mm), mean (SD) Overall (mm), mean (SD) Change (mm), mean (SD) E3 a E2 E4 E3 E2 E4 E3 E2 E4 CIMT (mm) 0.77 (0.12) 0.75 (0.10) 0.78 (0.12) 0.78 (0.12) 0.75 (0.10) 0.78 (0.12) 0.012 (0.029) 0.011 (0.023) 0.014 (0.028) a E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 200): E3/E4 and E4/E4 40 Table 21 – Overall Comparison of CIMT (mm) by APOE Genotype Groups (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -0.025 (0.013) 0.0048 (0.0098) 0.78 (0.0054) 0.76 (0.011) 0.79 (0.0082) 0.08 0.046 0.63 0.03 Model 2 d -0.031 (0.011) -0.0035 (0.0088) 0.79 (0.010) 0.76 (0.013) 0.79 (0.012) 0.02 0.005 0.69 0.03 Model 3 d -0.034 (0.011) -0.0016 (0.0085) 0.79 (0.0099) 0.76 (0.013) 0.79 (0.011) 0.01 0.002 0.85 0.01 a P-values were generated using mixed models with an unstructured covariance structure b P-values were generated using mixed models with an unstructured covariance structure for comparison between groups c E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 200): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 41 Table 22 – Comparison of CIMT (mm) by APOE Genotype Groups: Stratification by Gender (All Trials Combined) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Female Model 1 d -0.033 (0.013) -0.001 (0.010) 0.79 (0.006) 0.76 (0.012) 0.79 (0.008) 0.04 0.012 0.93 0.03 Model 2 d -0.031 (0.012) -0.012 (0.009) 0.78 (0.010) 0.75 (0.013) 0.77 (0.012) 0.02 0.007 0.19 0.14 Model 3 d -0.031 (0.011) -0.012 (0.009) 0.77 (0.009) 0.74 (0.012) 0.76 (0.011) 0.01 0.005 0.18 0.12 Male Model 1 d 0.001 (0.030) 0.023 (0.027) 0.75 (0.014) 0.76 (0.027) 0.78 (0.023) 0.68 0.97 0.39 0.54 Model 2 d -0.030 (0.028) 0.029 (0.025) 0.76 (0.029) 0.73 (0.037) 0.79 (0.034) 0.20 0.29 0.24 0.07 Model 3 d -0.038 (0.027) 0.038 (0.024) 0.76 (0.028) 0.73 (0.036) 0.80 (0.033) 0.06 0.16 0.11 0.02 a P-values were generated using mixed models with an unstructured covariance structure b P-values were generated using mixed models with an unstructured covariance structure for comparison between groups c Female E3(n = 363): E3/E3, E2(n = 81): E2/E2, E2/E3, and E2/E4, E4(n = 160): E3/E4 and E4/E4 Male E3(n = 110): E3/E3, E2(n = 29): E2/E2, E2/E3, and E2/E4, E4(n = 40): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 42 Table 23 – Overall Comparison of CIMT Progression (mm/year) by APOE Genotype Groups (All Trials Combined) 𝜷 (SE) Mean Estimates (SE) Overall P-value a E3 b E2 E4 E3 E2 E4 Model 1 c -0.00013 (0.0010) 0.0014 (0.00083) 0.0051 (0.00045) 0.0050 (0.00092) 0.0065 (0.00070) 0.23 Model 2 c -0.00014 (0.0010) 0.0013 (0.00083) 0.0052 (0.00045) 0.0050 (0.00092) 0.0065 (0.00070) 0.23 Model 3 c -0.00015 (0.0010) 0.0013 (0.00083) 0.0052 (0.00045) 0.0050 (0.00092) 0.0065 (0.00069) 0.23 a P-values were generated using mixed models with an unstructured covariance structure b E3(n = 473): E3/E3 E2(n = 110): E2/E2, E2/E3, and E2/E4 E4(n = 200): E3/E4 and E4/E4 c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 43 Table 24 – Comparison of CIMT Progression (mm/year) by APOE Genotype Groups: Stratification by Gender (All Trials Combined) 𝜷 (SE) Mean Estimates (SE) Overall P-value a E3 b E2 E4 E3 E2 E4 Female Model 1 c -0.00060 (0.0013) 0.0016 (0.00099) 0.0062 (0.00055) 0.0056 (0.0011) 0.0078 (0.00082) 0.19 Model 2 c -0.00060 (0.0013) 0.0016 (0.00099) 0.0062 (0.00055) 0.0056 (0.0011) 0.0078 (0.00082) 0.19 Model 3 c -0.00061 (0.0013) 0.0016 (0.00099) 0.0062 (0.00055) 0.0056 (0.0011) 0.0078 (0.00082) 0.19 Male Model 1 c 0.0016 (0.0013) -0.00049 (0.0012) 0.0018 (0.00062) 0.0034 (0.0012) 0.0014 (0.0010) 0.40 Model 2 c 0.0016 (0.0013) -0.00049 (0.0012) 0.0018 (0.00062) 0.0034 (0.0012) 0.0014 (0.0010) 0.40 Model 3 c 0.0016 (0.0013) -0.00048 (0.0012) 0.0019 (0.00062) 0.0034 (0.0012) 0.0014 (0.0010) 0.41 a P-values were generated using mixed models with an unstructured covariance structure b Female E3(n = 363): E3/E3, E2(n = 81): E2/E2, E2/E3, and E2/E4, E4(n = 160): E3/E4 and E4/E4 Male E3(n = 110): E3/E3, E2(n = 29): E2/E2, E2/E3, and E2/E4, E4(n = 40): E3/E4 and E4/E4 c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 44 3.5. Association of APOE Genotype with CIMT (Excluding ELITE Trial Participants) The 356 participants had a mean (range) of 17.9 (1–61) months of follow-up in the E3 group, 18.3 (1–57) months of follow-up in the E2 group, and 17.6 (1–55) months of follow-up in the E4 group (p = 0.73). As seen for the total sample, participants in the E2 group had the smallest mean CIMT, E4 carriers had the largest mean CIMT value, and the mean change in CIMT increased (Table 25). Tables 26 and 28 summarize the overall effect of the APOE genotype groups on the CIMT value and progression rate after excluding the ELITE trial participants. Overall, no significant differences were detected in CIMT or CIMT progression across the three APOE groups (all p > 0.05 in models 1–3). As in the analysis with all subjects, the effect of the APOE genotype on the CIMT value and its progression was separately tested in females and males. Because the data for males after excluding the ELITE trial were the same as those after including the ELITE trial and all data for males were obtained from the BVAIT trial, we tested the association of the APOE genotype only in females (n = 177). As shown in Table 27, the APOE genotype groups had no significant effects on CIMT in females. Even after an adjustment for follow-up time, age, race, trial, BMI, and SBP, there was no significant difference in CIMT among the three APOE groups (p = 0.68). The association between APOE genotype groups and CIMT progression in females are shown in Table 29. No significant differences in CIMT progression across the three APOE groups were observed in females (all p = 0.19 in models 1–3). 45 Table 25 – CIMT Measurement at Baseline and During Trial Follow-up (Excluding ELITE Trial Participants) Baseline (mm), mean (SD) Overall (mm), mean (SD) Change (mm), mean (SD) E3 a E2 E4 E3 E2 E4 E3 E2 E4 CIMT (mm) 0.77 (0.13) 0.75 (0.12) 0.79 (0.14) 0.77 (0.13) 0.76 (0.12) 0.79 (0.14) 0.004 (0.013) 0.006 (0.015) 0.005 (0.014) a E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 87): E3/E4 and E4/E4 46 Table 26 – Overall Comparison of CIMT (mm) by APOE Genotype Groups (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -0.017 (0.020) 0.023 (0.016) 0.77 (0.009) 0.76 (0.018) 0.80 (0.014) 0.16 0.37 0.15 0.07 Model 2 d -0.027 (0.017) 0.013 (0.014) 0.79 (0.012) 0.76 (0.018) 0.80 (0.015) 0.13 0.12 0.38 0.04 Model 3 d -0.031 (0.017) 0.014 (0.014) 0.79 (0.011) 0.76 (0.017) 0.81 (0.015) 0.07 0.07 0.32 0.02 a P-values were generated using mixed models with an unstructured covariance structure b P-values were generated using mixed models with an unstructured covariance structure for comparison between groups c E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 87): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 47 Table 27 – Comparison of CIMT (mm) by APOE Genotype Groups in Females (Excluding ELITE Trial Participants) 𝜷 (SE) LS Mean Estimates (SE) Overall P-value a P-value b E3 c E2 E4 E3 E2 E4 E3 vs.E2 E3 vs.E4 E2 vs.E4 Model 1 d -0.037 (0.024) 0.020 (0.018) 0.80 (0.010) 0.76 (0.021) 0.82 (0.015) 0.10 0.13 0.28 0.03 Model 2 d -0.019 (0.020) 0.001 (0.016) 0.78 (0.011) 0.76 (0.019) 0.78 (0.016) 0.64 0.36 0.96 0.39 Model 3 d -0.017 (0.020) -0.006 (0.015) 0.77 (0.011) 0.76 (0.018) 0.77 (0.016) 0.68 0.40 0.69 0.63 a P-values were generated using mixed models with an unstructured covariance structure b P-values were generated using mixed models with an unstructured covariance structure for comparison between groups c E3(n = 106): E3/E3 E2(n = 24): E2/E2, E2/E3, and E2/E4 E4(n = 47): E3/E4 and E4/E4 d Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 48 Table 28 – Overall Comparison of CIMT Progression (mm/year) by APOE Genotype Groups (Excluding ELITE Trial Participants) 𝜷 (SE) Mean Estimates (SE) Overall P-value a E3 b E2 E4 E3 E2 E4 Model 1 c -0.017 (0.020) 0.023 (0.016) 0.0027 (0.00051) 0.0040 (0.0010) 0.0040 (0.00081) 0.26 Model 2 c -0.027 (0.017) 0.013 (0.014) 0.0027 (0.00051) 0.0041 (0.0010) 0.0040 (0.00081) 0.27 Model 3 c 0.0014 (0.0011) 0.0013 (0.00096) 0.0027 (0.00051) 0.0041 (0.0010) 0.0040 (0.00081) 0.28 a P-values were generated using mixed models with an unstructured covariance structure b E3(n = 216): E3/E3 E2(n = 53): E2/E2, E2/E3, and E2/E4 E4(n = 87): E3/E4 and E4/E4 c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 49 Table 29 – Comparison of CIMT Progression (mm/year) by APOE Genotype Groups in Females (Excluding ELITE Trial Participants) 𝜷 (SE) Mean Estimates (SE) Overall P-value a E3 b E2 E4 E3 E2 E4 Model 1 c 0.0012 (0.0018) 0.0026 (0.0015) 0.0036 (0.00081) 0.0049 (0.0017) 0.0062 (0.0012) 0.19 Model 2 c 0.0013 (0.0018) 0.0026 (0.0015) 0.0036 (0.00081) 0.0049 (0.0017) 0.0062 (0.0012) 0.19 Model 3 c 0.0013 (0.0018) 0.0026 (0.0015) 0.0036 (0.00081) 0.0049 (0.0017) 0.0063 (0.0012) 0.19 a P-values were generated using mixed models with an unstructured covariance structure b E3(n = 106): E3/E3 E2(n = 24): E2/E2, E2/E3, and E2/E4 E4(n = 47): E3/E4 and E4/E4 c Reference group Model 1: adjusted for follow-up time Model 2: adjusted for follow-up time, age, race, gender, and trial Model 3: adjusted for follow-up time, age, race, gender, trial, BMI, and SBP 50 CHAPTER 4: DISCUSSION We evaluated the effect of APOE genotype on the levels of total, LDL, and HDL cholesterol, CIMT as a marker for subclinical atherosclerosis, and progression rates of CIMT with repeated measures from three trials. Because of the unknown oral 17beta- estradiol treatment code, we repeated data analyses after excluding ELITE trial. Our results show that APOE genotype significantly affects total and LDL cholesterol levels after adjustment for follow-up time, age, race, gender, trial, BMI, and SBP. Previous studies have indicated that while E2 carriers exhibit decreased total cholesterol levels, the E4 allele is associated with increased total cholesterol levels [25, 26]. Among all participants in our trials, we observed that total cholesterol levels are significantly lower in E2 carriers, including E2/E2, E2/E3, and E2/E4 genotypes compared to the other two allelic genotypes. Participants in the E4 group, including E3/E4 and E4/E4 carriers, showed higher total cholesterol levels than subjects in the E2 and E3 groups. Additional reports identified a significant association between certain APOE genotype variants and LDL cholesterol levels [27–30]. Similar to these previously reports, our study demonstrated that subjects in the E2 group had the lowest LDL cholesterol levels of all participants while E4 carriers showed the highest LDL cholesterol levels. These observations were consistent even when the ELITE trial was excluded from the analysis. With respect to HDL cholesterol levels, while some studies have demonstrated higher HDL cholesterol levels in middle-aged or elderly E2 carriers than in E4 carriers [31–35], others could not show a significant association between APOE genotype and HDL 51 cholesterol levels [27, 36, 37]. Corroborating these other studies, we did not observe a significant association between APOE genotype and HDL cholesterol levels in participants from all trials, even after exclusion of the ELITE trial. Numerous studies have demonstrated significant effects of APOE genotype on LDL cholesterol levels in the middle-aged and elderly. Some reports found higher CIMT levels in E4 carriers [30–32, 35, 36], but other studies have not replicated these associations [33, 34, 37, 38]. Cattin et al. [32] and Luc Djoussé et al. [37] have investigated the effect of APOE on LDL cholesterol levels and CIMT measurements in middle-aged asymptomatic CHD participants from Europe and North America; respectively, and both studies observed higher LDL cholesterol levels in E4 carriers. Notably, Cattin et al. detected higher CIMT levels in E4 carriers than in E3 carriers after adjusting for age, sex, BMI, total and LDL cholesterol levels, triglycerides, and ratio of LDL to HDL cholesterol, while Luc Djoussé et al. found no significant association between APOE genotype and CIMT after adjusting for similar demographic and clinical factors. Because APOE genotype could influence cholesterol levels due to the differential cholesterol metabolism rates and cholesterol levels can influence CIMT, cholesterol level may be a potential mediator of the association between APOE genotype and CIMT levels in these two studies. Indeed, other studies have considered cholesterol levels as covariates for the association between APOE genotype and CIMT levels. Some observed an association between APOE genotype and CIMT measurements that was statistically significant before adjusting for lipid measures: these associations did not remain significant after adjusting for these factors [30, 36]. Using LDL or HDL cholesterol levels as covariates did not alter the results, indicating that APOE genotype may have a 52 cholesterol independent effect on CIMT measurement. In our study, after adjusting for follow-up time, age, race, gender, trial, BMI, and SBP, female E2 carriers exhibited lower CIMT levels than female with E3/E3 genotype. In contrast, there was no significant association between CIMT and APOE genotype in male participants. Very few studies have explored the potential association between APOE genotypes and the rate of CIMT progression. In our study, there was no significant difference in CIMT progression rate with respect to APOE genotypes in either female or male participants. The mean CIMT in all three genotype groups increased over trial follow-up. After adjusting for covariates, male E2 carriers had a higher mean CIMT progression rate compared to other male carriers, and female E4 carriers had a higher mean rate of CIMT progression. Female E2 carriers had the lowest CIMT progression, consistent with the assessment of CIMT level among the three genotype groups. Oral 17beta-estradiol treatment could affect cholesterol levels and CIMT measurements, leading to differing results among participants in all three trials compared to participants in the BVAIT and WISH trials only. After excluding participants in the ELITE trial, the significant association between total cholesterol levels and the E2 and E4 genotype groups remained, although no significant difference was detected between total cholesterol levels in the E2 and E3 groups. With respect to CIMT, in female participants, we saw no more significant difference in CIMT levels between the E2 and E3 groups after excluding participants in the ELITE trial. Although LDL cholesterol levels in the BVAIT and WISH trials were significantly different among the three genotype groups, no significant association between genotype and CIMT was detected. One possible explanation relies on the relationship between the 53 E2 allele and type III hyperlipidemia, a genetic disorder characterized by the accumulation of remnant lipoproteins in plasma [39]. Although the E2 was associated with lower LDL cholesterol levels in our study, this allele is also associated with increased type III hyperlipidemia and increased risk for premature atherosclerosis [39– 41], thus obscuring the possible relationship between APOE genotype and CIMT levels. However, type III hyperlipidemia, occurring in less than 1 in 10,000 in the general population, is a rare genetic lipoprotein disorder [41] and type III hyperlipidemia cannot explain all of the E2 that was associated with higher CIMT levels. One limitation of our study results from the use of subjects who participated in three separate trials, each of which was designed to address different primary interests and involved varying inclusion and exclusion study criteria. In conclusion, low total and LDL cholesterol levels were characteristic of E2 carriers. In contrast, high total and LDL cholesterol levels were observed in the E4 genotype group. Female E2 carriers had lower carotid atherosclerosis than female E3/E3 subjects, but after excluding the ELITE trial, the APOE genotype group was not significantly associated with CIMT. Additionally, no significant associations between APOE genotype and atherosclerosis were detected in males. Furthermore, atherosclerosis progression rates were not significantly different among APOE genotype groups, regardless of gender. 54 REFERENCES [1] American Heart Association. “Heart Disease and Stroke Statistics 2014 Update: A Report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee.” Circulation 129.3 (2014): e28. [2] Sima, Anca V., Camelia S. Stancu, and Maya Simionescu. “Vascular Endothelium in Atherosclerosis.” Cell and Tissue Research, 335.1 (2009): 191- 203. [3] Altman, Raul. “Risk Factors in Coronary Atherosclerosis Athero-Inflammation: the Meeting Point.” Thrombosis Journal 1.1 (2003): 4. [4] “Who is at Risk for Atherosclerosis?” National Institutes of Health, 2011. <http://www.nhlbi.nih.gov/health/health-topics/topics/atherosclerosis/atrisk#> [5] Harvard Medical School. Harvard Medical School Health Topics A-Z: Atherosclerosis. Norwalk, CT: Belvoir Media Group, LLC, 2011. [6] Flemming, Alexandra. “Preventing Plaque Rupture.” Nature Reviews Drug Discovery, 2008. <http://go.galegroup.com/ps/i.do?id=GALE%7CA192498411&v=2.1&u=usocal_ main&it=r&p=AONE&sw=w&asid=7c3b0e877a3a82bf93a36627de38e680> [7] Wissler, Robert W. “Update on the Pathogenesis of Atherosclerosis.” The American Journal of Medicine 91.1 (1991): S3-S9. [8] Mahley, Robert W. and Stanley C. Rall Jr. “Apolipoprotein E: Far More Than a Lipid Transport Protein.” Annual Review of Genomics and Human Genetics 1.1 (2000): 507-537. [9] Myers, Amanda J. and Charles B. Memeroff. “ApoE: A Risk Factor for Multiple Disorders.” The American Journal of Geriatric Psychiatry 20.7 (2012): 545-548. [10] Beisiegel, Ulrike, Wilfried Weber, Gudrun Ihrke, Joachim Herz, and Keith K. Stanley. “The LDL-Receptor-Related Protein, LRP, is an Apolipoprotein E- binding Protein.” Natural 341 (1989): 162-164. [11] Altenburg, Michael, Jose Arbones-Mainar, Lance Johnson, Jennifer Wilder, & Nobuyo Maeda. “Human LDL Receptor Enhances Sequestration of ApoE4 and VLDL Remnants on the Surface of Hepatocytes but not Their Internalization in Mice.” Arteriosclerosis, Thrombosis, and Vascular Biology. 28.6 (2008): 1104- 1110. [12] Weisgraber, Karl H. "Apolipoprotein E: Structure-Function Relationships." Advances in protein chemistry 45 (1994): 249-302. 55 [13] Pang, Chi Pui, Clement C. Tham, and Dennis Shun Chiu Lam. “Chapter 18. ApoE Polymorphisms and Severity of Open-Angle Glaucoma.” Mechanisms of the Glaucomas: Disease Process and Therapeutic Modalities. Eds. Joyce Tombran- Tink, Colin J. Barnstable, and M. Bruce Shields. Totowa: Humana Press 2008. 289-307. [14] “High Cholesterol and the APOE Gene.” GB HealthWatch, 2003. <http://www.gbhealthwatch.com/GND-High-Cholesterol-APOE.php> [15] Elosua, Roberto, Jose M. Ordovas, L. Adrienne Cupples, Caroline S. Fox, Joseph F. Polak, Philip A. Wolf, Ralph A. D’Agostino, Sr., and Christopher J. O’Donnell. “Association of APOE Genotype with Carotid Atherosclerosis in Men and Women the Framingham Heart Study.” Journal of Lipid Research 45.10 (2004): 1868-1875. [16] Hodis, Howard N., Wendy J. Mack, Laurie Dustin, Peter R. Mahrer, Stanley P. Azen, Robert Detrano, Jacob Selhub, Petar Alaupovic, Chao-ran Liu, Ci-hua Liu, Juliana Hwang, Alison G. Wilcox, and Robert H. Selzer. “High-Dose B Vitamin Supplementation and Progression of Subclinical Atherosclerosis: A Randomized Controlled Trial.” Stroke 40.3 (2009): 730-736. [17] Hodis, Howard N., Wendy J. Mack, Naoko Kono, Stanley P. Azen, Donna Shoupe, Juliana Hwang-Levine, Diana Petitti, Lora Whitfield-Maxwell, Mingzhu Yan, Adrian A. Franke, and Robert H. Selzer. “Isoflavone Soy Protein Supplementation and Atherosclerosis Progression in Healthy Postmenopausal Women: A Randomized Controlled Trial." Stroke 42.11 (2011): 3168-3175. [18] Henderson, Victor W., Jan A. St. John, Howard N. Hodis, Carol A. McCleary, Frank Z. Stanczyk, Roksana Karim, Donna Shoupe, Naoko Kono, Laurie Dustin, Hooman Allayee, and Wendy J. Mack. “Cognition, Mood, and Physiological Concentrations of Sex Hormones in the Early and Late Postmenopause.” Proceedings of the National Academy of Sciences 110.50 (2013): 20290-20295. [19] Potter, Susan M., J. A. Baum, H. Teng, R. J. Stillman, N. F. Shay, and J. W. Erdman Jr.. “Soy Protein and Isoflavones: Their Effects on Blood Lipids and Bone Density in Postmenopausal Women.” The American Journal of Clinical Nutrition 68.6 (1998): 1375S-1379S. [20] Corcoran, Michael P., Alice H. Lichtenstein, Mohsen Meydani, Alice Dillard, Ernst J. Schaefer, and Stefania Lamon-Fava. "The Effect of 17β-estradiol on Cholesterol Content in Human Macrophages is Influenced by the Lipoprotein Milieu." Journal of Molecular Endocrinology 47.1 (2011): 109-117. [21] Hodis, Howard N., Wendy J. Mack, Laurie LaBree, Robert H. Selzer, Chao-ran Liu, Ci-hua Liu, and Stanley P. Azen. "The Role of Carotid Arterial Intima-Media 56 Thickness in Predicting Clinical Coronary Events." Annals of Internal Medicine 128.4 (1998): 262-269. [22] Hodis, Howard N., Wendy J. Mack, Laurie LaBree, Peter R. Mahrer, Alex Sevanian, Chao-ran Liu, Ci-hua Liu, Juliana Hwang, Robert H. Selzer, Stanley P. Azen, and for the VEAPS Research Group. "Alpha-Tocopherol Supplementation in Healthy Individuals Reduces Low-Density Lipoprotein Oxidation but Not Atherosclerosis The Vitamin E Atherosclerosis Prevention Study (VEAPS)." Circulation 106.12 (2002): 1453-1459. [23] Hodis, Howard N., Wendy J. Mack, Roger A. Lobo, Donna Shoupe, Alex Sevanian, Peter R. Mahrer, Robert H. Selzer, Chao-ran Liu, Ci-hua Liu, and Stanley P. Azen. "Estrogen in the Prevention of Atherosclerosis." Annals of Internal Medicine 135.11 (2001): 939-953. [24] Warnick, G. Russell, Thuy Nguyen, and Alegria A. Albers. "Comparison of Improved Precipitation Methods for Quantification of High-Density Lipoprotein Cholesterol." Clinical Chemistry 31.2 (1985): 217-222. [25] Lipid Research Clinics Program. Lipid and Lipoprotein Analysis. Vol. 1, in Manual of Laboratory Operations. Bethesda, MD: DHEW publication No. NIH 75-628, 1974. [26] Bennet, Anna M., Emanuele Di Angelantonio, Zheng Ye, Frances Wensley, Anette Dahlin, Anders Ahlbom, Bernard Keavney, Rory Collins, Björn Wiman, Ulf de Faire, and John Danesh. "Association of Apolipoprotein E Genotypes with Lipid Levels and Coronary Risk." The Journal of the American Medical Association 298.11 (2007): 1300-1311. [27] Hall, K., J. Murrell, A. Ogunniyi, M.Deeg, O. Baiyewu, S. Gao, O. Gureje, J. Dickens, R. Evans, V. Smith-Gamble, F.W. Unverzagt, J. Shen, and H. Hendrie. “Cholesterol, APOE Genotype, and Alzheimer Disease: An Epidemiologic Study of Nigerian Yoruba.” Neurology 66.2 (2006): 223–227.PMC. Web. 5 Feb. 2015. [28] Tao, Qing-Qing, Yan Chen, Zhi-Jun Liu, Yi-Min Sun, Ping Yang, Shen-Ji Lu, Miao Xu, Qin-Yun Dong, Jia-Jun Yang, and Zhi-Ying Wu. "Associations between Apolipoprotein E Genotypes and Serum Levels of Glucose, Cholesterol, and Triglycerides in a Cognitively Normal Aging Han Chinese population." Clinical Interventions in Aging 9 (2014): 1063. [29] Larson, Ilona A., Jose M. Ordovas, Carl DeLuca, James R. Barnard, Giso Feussner, and Ernst J. Schaefer. "Association of Apolipoprotein (Apo) E Genotype with Plasma Apo E Levels." Atherosclerosis 148.2 (2000): 327-335. [30] Elosua, Roberto, Jose M. Ordovas, L. Adrienne Cupples, Caroline S. Fox, Joseph F. Polak, Philip A. Wolf, Ralph A. D’Agostino, Sr., and Christopher J. 57 O’Donnell. “Association of APOE Genotype with Carotid Atherosclerosis in Men and Women the Framingham Heart Study.” Journal of Lipid Research 45.10 (2004): 1868-1875. [31] Haraki, T., T. Takegoshi, C. Kitoh, T. Wakasugi, T. Saga, J.-I. Hirai, T. Aoyama, A. Inazu, and H. Mabuchi. "Carotid Artery Intima‐Media Thickness and Brachial Artery Flow‐Mediated Vasodilation in Asymptomatic Japanese Male Subjects Amongst Apolipoprotein E Phenotypes." Journal of Internal Medicine 252.2 (2002): 114-120. [32] Cattin, L., M. Fisicaro, M. Tonizzo, M. Valenti, G.M. Danek, M. Fonda, P.G. Da Col, S. Casagrande, E. Pincetti, M. Bovenzi, and F. Baralle. "Polymorphism of the Apolipoprotein E Gene and Early Carotid Atherosclerosis Defined by Ultrasonography in Asymptomatic Adults." Arteriosclerosis, Thrombosis, and Vascular Biology 17.1 (1997): 91-94. [33] Slooter, A.J.C., M.L. Bots, L.M. Havekes, A. Iglesias del Sol, M. Cruts, D.E. Grobbee, A. Hofman, C. Van Broeckhoven, J.C.M. Witteman, and C.M. van Duijn. "Apolipoprotein E and Carotid Artery Atherosclerosis The Rotterdam Study." Stroke 32.9 (2001): 1947-1952. [34] Horejsí, B., J. Spácil, R. Ceska, M. Vrablík, T. Haas, and A. Horínek. "The independent Correlation of the Impact of Lipoprotein (a) Levels and Apolipoprotein E Polymorphism on Carotid Artery Intima Thickness." International Angiology: a Journal of the International Union of Angiology 19.4 (2000): 331-336. [35] Paternoster, Lavinia, González NAM, Lewis S, and Sudlow C. "Association between Apolipoprotein E Genotype and Carotid Intima-Media Thickness May Suggest a Specific Effect on Large Artery Atherothrombotic Stroke." Stroke 39.1 (2008): 48-54. [36] Volcik, Kelly A., Ruth Ann Barkley, Richard G. Hutchinson, Thomas H. Mosley, Gerardo Heiss, A. Richey Sharrett, Christie M. Ballantyne, and Eric Boerwinkle. "Apolipoprotein E Polymorphisms Predict Low Density Lipoprotein Cholesterol Levels and Carotid Artery Wall Thickness but not Incident Coronary Heart Disease in 12,491 ARIC Study Participants." American Journal of Epidemiology 164.4 (2006): 342-348. [37] Djoussé, Luc, Richard H. Myers, Michael A. Province, Steven C. Hunt, John H. Eckfeldt, Gregory Evans, James M. Peacock, and R. Curtis Ellison. "Influence of Apolipoprotein E, Smoking, and Alcohol Intake on Carotid Atherosclerosis National Heart, Lung, and Blood Institute Family Heart Study." Stroke 33.5 (2002): 1357-1361. 58 [38] Beilby, John P., Clive C.J. Hunt, Lyle J. Palmer, Caroline M.L. Chapman, Jodi P. Burley, Brendan M. McQuillan, Peter L. Thompson, and Joseph Hung. "Apolipoprotein E Gene Polymorphisms Are Associated With Carotid Plaque Formation but Not With Intima-Media Wall Thickening Results From the Perth Carotid Ultrasound Disease Assessment Study (CUDAS)." Stroke 34.4 (2003): 869-874. [39] Schaefer, Juergen R. "Unraveling Hyperlipidemia Type III (Dysbetalipoproteinemia), Slowly." European Journal of Human Genetics 17.5 (2009): 541. [40] Rubin, Raphael, David S. Strayer, and Emanuel Rubin, eds. Rubin's Pathology: Clinicopathologic Foundations of Medicine. Lippincott Williams & Wilkins, 2008. [41] Sorrentino, Matthew J. Hyperlipidemia in Primary Care: A Practical Guide to Risk Reduction. New York: Humana Press, 2011.
Abstract (if available)
Abstract
Background and Purpose: Apolipoprotein E (APOE) plays an important role in cholesterol metabolism, which can influence the development of atherosclerosis. We evaluated if APOE genotype affects cholesterol levels, subclinical atherosclerosis as assessed by carotid artery intima‐media thickness (CIMT) measurements, and the progression of atherosclerosis through repeated CIMT measurements. ❧ Methods: Study subjects, aged 32 to 88 years with no clinical evidence of cardiovascular disease, participated in three randomized clinical trials. Six hundred four females and 180 males were divided into three groups according to their APOE genotype: E2, including E2/E2, E2/E3, and E2/E4 genotypes
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The effects of hormone therapy on carotid artery intima-media thickness and serum lipids by ApoE4 genotype
PDF
The association of asthma and asthma-related medications on subclinical atherosclerosis: a cross-sectional analysis of four randomized clinical trials
PDF
Relationship between progression of atherosclerosis and coagulation measures in a randomized-controlled trial
PDF
Association of arterial stiffness progression with subclinical atherosclerosis measurements in postmenopausal women
PDF
The associations of inflammatory markers with progression of subclinical atherosclerosis in early and late postmenopausal women
PDF
The effect of vitamin D supplementation on the progression of carotid intima-media thickness and arterial stiffness in elderly African American women: Results of a randomized placebo-controlled trial
PDF
Subclinical atherosclerosis in overweight Latino youth: influence of cardiometabolic risk factors
PDF
Bone mineral density is associated with carotid atherosclerosis in healthy postmenopausal women: a longitudinal analysis of randomized clinical trial data
PDF
Sex hormones and atherosclerosis in postmenopausal women
PDF
Association of bone density and bone metabolism markers: the Women’s Isoflavone Soy Health (WISH) trial
PDF
Association of subclinical atherosclerosis with plasma B-vitamin, cysteine, homocysteine, and cysteinyl glycine in a cardiovascular disease-free population
PDF
Relationship of blood pressure and antihypertensive medications to cognitive change in the BVAIT, WISH, and ELITE clinical trials
PDF
Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial
PDF
Association of oophorectomy with progression of carotid atherosclerosis
PDF
Hormone therapy timing hypothesis and atherosclerosis
PDF
Associations between inflammatory markers and change in cognitive endpoints
PDF
Association between endogenous sex hormone levels and cognition: the Women’s Isoflavone Soy Health (WISH) trial
PDF
The impact of data collection procedures on the analysis of randomized clinical trials
PDF
Effect of estradiol on circulating levels of inflammatory cytokines in postmenopausal women
PDF
Effect of menopausal hormone replacement therapy on lipoprotein particle fractions and association with carotid artery intima-media thickness and grey-scale median, independent measurements of su...
Asset Metadata
Creator
You, Seung Sook
(author)
Core Title
The association of apolipoprotein E genotype with cholesterol levels and atherosclerosis: three randomized clinical trials
School
Keck School of Medicine
Degree
Master of Science
Degree Program
Biostatistics
Publication Date
04/23/2015
Defense Date
04/23/2015
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
APOE,apolipoprotein E genotype,atherosclerosis,cholesterol levels,OAI-PMH Harvest
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Mack, Wendy Jean (
committee chair
), Allayee, Hooman (
committee member
), Hodis, Howard N. (
committee member
)
Creator Email
seungsoy@usc.edu,seungsyou@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-559044
Unique identifier
UC11300395
Identifier
etd-YouSeungSo-3386.pdf (filename),usctheses-c3-559044 (legacy record id)
Legacy Identifier
etd-YouSeungSo-3386.pdf
Dmrecord
559044
Document Type
Thesis
Format
application/pdf (imt)
Rights
You, Seung Sook
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
APOE
apolipoprotein E genotype
atherosclerosis
cholesterol levels