Healthy or atherosclerotic male subjects were recruited from among in- and out-patients of Korea University College of Medicine in Seoul, Korea, after approval from Biomedical Human Subjects Review Committee. After first screening (n = 340) with a health questionnaire (family history, exercise), subjects were further evaluated with a dietary survey and anthropometry [body mass index (BMI), blood pressure (BP)]; biochemical tests were performed for 192 final subjects. The 192 subjects were divided into atherosclerosis patients (AP, n = 101) or control (n = 91) groups after AP were diagnosed by coronary angiography by a cardiologist. The coronary artery lesion was defined where stricture levels of Lt. main, Lt. anterior descending artery, left circumflex, or right coronary artery were over 50%. To determine whether Hcy is an independent risk factor for CVD, subjects were assigned to one of three Hcy levels, using the European Concerted Action Projects (ECAP) cut-off value.¹³ Dietary surveys were performed using the CANpro 3.0 software (developed by Korean Nutrition Society). Survey results on dietary intakes and biochemistry were statistically analyzed. None of the participants were using vitamin and/or mineral supplements.

Cholesterol and triglycerides (TG) were analyzed by enzymatic Kits (Sigma Co, St Louis, MO, USA).¹⁴,¹⁵ After HDL and LDL were separated by density gradient ultracentrifugation, apo A-I in HDL and apo B-100 in LDL were quantified by Western blotting analysis (antibodies; Santa Cruz, CA, USA). In order to measure Ox-LDL, pure LDL was high-speed centrifugated using KBr, at a density of 1.019-1.063 g/mL. During centrifugation, 0.1% EDTA and 0.02% sodium azoid were added. Phosphate-buffered saline (PBS) was used for dialysis of centrifugated LDL. SDS-polyacrylamine gel electrophoresis (PAGE) was used to verify LDL purification.¹⁶ The purified LDL was dialyzed for a day, and then LPO [malondialdehyde (MDA)] production was analyzed. The thiobarbituric acid (TBA) method (modified Packer and Smith method) was used to obtain correct amount of lipid peroxide with the MDA produced. By using 1,1,3,3-tetraethoxypropane as the standard reagent, comparisons were made between different treatments according to thiobarbituric acid reactive substances (TBARS) values from different amounts of 1 mg protein. Butylhydroxytoluen (BHT), sulfuric acid, and Na₂WO₄ were mixed with LDL, and the mixture was centrifugated. The liquid that rose was taken; NaOH and 1% TBA were added and heated for 60 minutes at 100℃. After centrifugation at 3,000 rpm for 10 minutes, absorbance was read.^16-18 After the protein content of LDL was measured, units of LDL per mg protein were used to express the amount. Total plasma Hcy (sum of free-, protein-bound and mixed disulfide forms) in all samples was determined by the use of "Abbott IMX system Homocysteine E/B3D390, 77-0650/R2" kit (ABBOTT Plasma Total Hcy Diagnostic kit; Abbott, Abbott Park, IL, USA) in one run.¹⁹ By processing Hcy molecules (protein, residual thiol, or sulfide combined) with DTT, a crystalline form of Hcy can be obtained. By applying an S-adenosyl-L-homocysteine (SAH) hydrolase to the crystalline with cholesterol, the amount of Hcy that combines with adenosine to form SAH was measured. For determination of vitamin B₁₂ and folate in serum and plasma, vitamin B₁₂/folate kit (Diagnostic Products Corporation, Los Angeles, CA, USA) was used according to the method of Soild Phase No Boil Dual-couint.²⁰: 200 µL of serum was mixed with sample buffer, and the mixture was left for 30 minutes at room temperature. The above samples were reacted with NaOH/KCN at 37℃, and vitamin B₁₂/folate binder was added into each tube. The supernatant after centrifugation at 2,000 g for 15 minutes was discarded. The radioactivity from ¹²⁵I-folate and ⁵⁷Co-vit B₁₂ in the pellet was measured by Gamma counter Cobra II. For the amount of vitamin B₁₂, pg/mL was multiplied by 0.7378 to convert to pmol/L. For folate, ng/mL was multiplied by 2.266 to convert to pmol/L. The clinical reference value for blood vitatimin B₁₂ is 200-950 pg/mL, and 3.0-17 ng/mL for blood folate.

Using SAS statistical software (Carey, NC, USA), ANOVA and independent t-test were performed. Statistical significance was accepted at p-value < 0.05. Relationships among variables were assessed by correlation analysis. Univariate and multivariate logistic regression analyses were used to obtain the crude and adjusted OR (AOR) and 95% CI.

Average age of the AP group was 56.2 years, while the average age of the control group was 42.4 years. The two groups exhibited no differences in blood pressure and BMI (Table 1). Over 60% of the AP subjects exhibited other comorbidities also, whereas 73% of controls were disease-free. The incidences of hypertension and diabetes were 35% and 16%, respectively, in the AP (data not shown). Smoking and drinking rates were similar in both groups, but the amount and length of the use were much higher in the AP; 47.7% of total AP smoked more than 1 pack of cigarettes/day (control: 32.2%) and 35.5% of the AP drank 2-4 times/week (control: 28.8%). The AP also had higher coffee consumption rate (5 cups/day) than the control (data not shown).

In the AP, blood Hcy and LDL cholesterol levels were higher and apo A-I in HDL and folate were lower, exhibiting characteristic lipid spectrum of atherosclerosis (Table 1). Plasma Hcy and LDL can be viewed as pro-arteriosclerosis factors, and HDL's apo A-I as an anti-arteriosclerosis factor. However, no apparent differences in total cholesterol (T-C), HDL, TG, and apo B-100 were observed between AP and control. Blood Hcy and Ox-LDL were higher in AP, suggesting that Hcy plays a role in LDL oxidation. Blood folate levels were lower in AP, whereas vitamin B₁₂ levels were similar. In order to verify the changes in lipid profiles, resulting from the effects of vitamins and Ox-LDL on Hcy concentrations, subjects were divided into 3 groups depending on Hcy levels: low-Hcy (≤ 6.9 uM/L), normal (7 uM-12 uM/L) and high-Hcy (≥ 12.1 uM/L) (Table 2). Ox-LDL levels were not different between AP and control, but signifi-cantly higher in the normal and high-Hcy groups than in the low-Hcy group (Table 2). The positive association of LDL with Ox-LDL was stronger in high-Hcy subjects than that in total subjects, normal or low Hcy subjects (data not shown). The increased Ox-LDL in the high-Hcy group can be explained as LDL oxidation caused by the increase in blood Hcy. Blood folate levels were significantly different (p < 0.05) among the 3 Hcy groups, and it was negatively correlated with Hcy (Fig. 1). There was no correlation bet-ween vitamin B₁₂ and Hcy even though blood folate and vitamin B₁₂ were positively correlated (data not shown). The blood lipid levels did not change according to Hcy levels and were not correlated with Hcy.

The relative percentages of carbohydrate, protein and fat intakes per total calories were similar in AP and controls, even though AP had lower energy intakes (data not shown), suggesting that AP might have attempted to lose weight, since they had slightly, but not significantly, higher BMIs. Dietary intakes of other B vitamins were not different between the two groups, except vitamin B₁ and B₂ which were lower in AP. Vitamin B₁₂ intake could not be evaluated because of insufficient data. There were no differences in any dietary intakes except vitamin A, β-carotene and vitamin C among the three Hcy groups (Table 2). Hcy was not affected by dietary folate intakes in this study, and the pattern of correlation between dietary folate and Hcy or plasma folate were not obvious (Fig. 1). Therefore, low dietary folate in AP might have not induced hyperhomocysteinemia in this study. However, dietary intakes of vitamin A, and its precursor, and vitamin C were significantly lower with increasing Hcy. The negative correlations between dietary vitamin A (r² = - 0.226) or β-carotene (r² = - 0.233) antioxidants and Ox-LDL were also significant in AP (p < 0.001) compared to no correlations in controls (data not shown).

Using the ECAP cut-off value (≥ 12.1 uM/L), 26.6% of the total Korean male subjects had hyperhomocysteinemia, and 69.3% and 4.2% had normal and low-Hcy levels, respectively. The frequency of atherosclerosis was substantially increased if the blood Hcy was higher than ≥ 12.1 uM/L (OR, 2.887; 95% CI, 1.51-5.53) and this status was stronger after adjusting for age and BMI (AOR, 3.002; 95% CI, 1.27-7.09)(data not shown). To identify links between risk factors for atherosclerosis and high blood Hcy levels, all anthropometric and biochemistry variables and daily nutrient intakes were subjected to logistic regression analysis (Table 3). The AOR for atherosclerosis was 4.420 (95% CI, 1.542-12.663) with high blood Hcy and LDL levels compared to when LDL was not included as a factor (Model I & II). When dietary folate and plasma were added, the AOR of high-Hcy for atherosclerosis was not changed. After dietary vitamin A and β-carotene were included as antioxidants, the AOR of high-Hcy for atherosclerosis was reduced to 4.206 (95% CI, 1.457-12.142). We demonstrated that the most important risk factor for atherosclerosis was LDL and the least important risk factor was plasma folate in hyperhomocysteinemia. When participants were divided into tertiles of mean Hcy and LDL concentrations, 20.7% of subjects in the 3rd tertile of Ox-LDL (276.13-745 uM/mg LDL) came under two combined categories with the highest Hcy (11.89-50 uM/L) and LDL levels (124.81-244 mg/dL) (Fig. 2). The probability of subjects having the highest Ox-LDL while also having low Hcy and low LDL levels was only 3.4%. We found that the greatest progression of atherosclerosis was observed with the highest Ox-LDL simultaneously with both the highest Hcy and LDL concentrations. This means that a high circulating Hcy level was a risk factor for atherosclerosis only when the subject had also elevated LDL levels