Skip to main content

Some NLM-NCBI services and products are experiencing heavy traffic, which may affect performance and availability. We apologize for the inconvenience and appreciate your patience. For assistance, please contact our Help Desk at info@ncbi.nlm.nih.gov.

The American Journal of Clinical Nutrition logoLink to The American Journal of Clinical Nutrition
. 2017 Mar 29;105(6):1283–1290. doi: 10.3945/ajcn.116.137158

Choline and its metabolites are differently associated with cardiometabolic risk factors, history of cardiovascular disease, and MRI-documented cerebrovascular disease in older adults1,2

Annie J Roe 3,6, Shucha Zhang 5, Rafeeque A Bhadelia 4, Elizabeth J Johnson 3, Alice H Lichtenstein 3, Gail T Rogers 3, Irwin H Rosenberg 3, Caren E Smith 3, Steven H Zeisel 5, Tammy M Scott 3,*
PMCID: PMC5445668  PMID: 28356272

Abstract

Background: There is a potential role of choline in cardiovascular and cerebrovascular disease through its involvement in lipid and one-carbon metabolism.

Objective: We evaluated the associations of plasma choline and choline-related compounds with cardiometabolic risk factors, history of cardiovascular disease, and cerebrovascular pathology.

Design: A cross-sectional subset of the Nutrition, Aging, and Memory in Elders cohort who had undergone MRI of the brain (n = 296; mean ± SD age: 73 ± 8.1 y) was assessed. Plasma concentrations of free choline, betaine, and phosphatidylcholine were measured with the use of liquid-chromatography–stable-isotope dilution–multiple-reaction monitoring–mass spectrometry. A volumetric analysis of MRI was used to determine the cerebrovascular pathology (white-matter hyperintensities and small- and large-vessel infarcts). Multiple linear and logistic regression models were used to examine relations of plasma measures with cardiometabolic risk factors, history of cardiovascular disease, and radiologic evidence of cerebrovascular pathology.

Results: Higher concentrations of plasma choline were associated with an unfavorable cardiometabolic risk-factor profile [lower high-density lipoprotein (HDL) cholesterol, higher total homocysteine, and higher body mass index (BMI)] and greater odds of large-vessel cerebral vascular disease or history of cardiovascular disease but lower odds of small-vessel cerebral vascular disease. Conversely, higher concentrations of plasma betaine were associated with a favorable cardiometabolic risk-factor profile [lower low-density lipoprotein (LDL) cholesterol and triglycerides] and lower odds of diabetes. Higher concentrations of plasma phosphatidylcholine were associated with characteristics of both a favorable cardiometabolic risk-factor profile (higher HDL cholesterol, lower BMI, lower C-reactive protein, lower waist circumference, and lower odds of hypertension and diabetes) and an unfavorable profile (higher LDL cholesterol and triglycerides).

Conclusion: Choline and its metabolites have differential associations with cardiometabolic risk factors and subtypes of vascular disease, thereby suggesting differing roles in the pathogenesis of cardiovascular and cerebral large-vessel disease compared with that of small-vessel disease.

Keywords: cardiometabolic risk, cerebrovascular disease, choline, older adults, magnetic resonance imaging, nutrition

INTRODUCTION

Choline is an essential nutrient necessary for the structural integrity and signaling of cells and is a precursor to the neurotransmitter acetylcholine (1). Much attention has been given to the potential role of choline in cardiovascular disease. Although the main focus has been on the microbial metabolism of phosphatidylcholine and the production of the proatherosclerotic metabolite trimethylamine N-oxide (2), there have also been links between choline and more-traditional cardiometabolic risk factors. The biosynthesis of phosphatidylcholine in the liver is critical for the synthesis and secretion of VLDLs (3, 4) and is an essential component of lipoproteins. In addition, supplementation with the choline metabolite betaine has been reported to increase plasma LDL-cholesterol concentrations relative to the effect of a placebo and to increase the total-cholesterol:HDL-cholesterol ratio (5).

Choline may be oxidized to form the methyl donor betaine for the conversion of homocysteine to methionine. Hyperhomocysteinemia has been associated with increased risk of cardiovascular disease (6). As part of the one-carbon metabolism pathway, choline, folate, and vitamins B-12 and B-6 intersect at the point of homocysteine methylation. Both betaine and 5-methyltetrahydrofolate can be used to remethylate homocysteine, whereas vitamin B-6 is involved in the transmethylation of homocysteine (7). Thus, a fluctuation in the concentration of any of these nutrients has the potential to alter the utilization and concentrations of the other nutrients, and all nutrients should be taken into account when assessing the relations between choline and cardiovascular or cerebrovascular health.

Both choline and betaine can be obtained from food or synthesized de novo. Konstantinova et al. (8) showed differing associations between plasma choline and betaine and components of metabolic syndrome with an unfavorable cardiovascular disease risk profile being associated with high-choline and low-betaine plasma concentrations. To our knowledge, whether this same relation holds true for cerebrovascular disease has not previously been evaluated. In this study, we evaluated the associations of plasma choline and choline-related compounds with cardiometabolic risk factors, history of cardiovascular disease, and cerebrovascular pathology.

METHODS

Subjects and study design

The study involved a secondary analysis of a subset of the NAME (Nutrition, Aging, and Memory in Elders)7 cohort who had undergone MRI and cognitive testing and had archived plasma (n = 296) (Supplemental Figure 1). The NAME study’s design, subject recruitment, and data-collection methodology have been previously described in detail (9). Briefly, the NAME study was a cross-sectional study that was designed to identify relations between micronutrient status and cognitive capacity and cerebrovascular pathology in older adults. The study cohort consisted of community-based elders who were living in the greater Boston, Massachusetts, area, and were aged ≥60 y. Subjects were visited in their homes by trained research staff who administered a neuropsychological testing battery. Fasting blood samples (overnight fast; 0000−0800) and a range of health and behavior questionnaires and anthropometric and blood pressure measures were obtained in the participants’ homes by trained research staff. MRI was done at Tufts Medical Center, Boston, Massachusetts. The study was approved by the Tufts University Health Sciences Campus Institutional Review Board.

Plasma choline and related compounds

Fasting plasma samples that were collected between 2003 and 2007 were stored in a freezer at −80°F. Plasma aliquots were shipped on dry ice to the University of North Carolina Chapel Hill Nutrition Research Institute for the assay of free choline, betaine, and phosphatidylcholine. The lipid fraction was separated with the use of standardized methods (10). Aqueous and organic compounds were analyzed and quantified with the use of liquid-chromatography–stable-isotope dilution–multiple-reaction monitoring–mass spectrometry according to previously published methods (11).

Cardiometabolic risk factors and cardiovascular disease

Plasma HDL-cholesterol, LDL-cholesterol, triglyceride, and glucose concentrations were assessed with the use of Beckman Coulter AU400 standard operating procedures (Beckman Coulter American Inc.). Plasma total homocysteine concentrations were determined with the use of HPLC with fluorescence detection (12). Folate and vitamin B-12 concentrations were measured with the use of the Quantaphase II radioassay kit (Bio-Rad Laboratories). Pyridoxal-P (PLP; vitamin B-6) concentrations were determined with the use of the tyrosine decarboxylase apoenzyme method (13). High-sensitivity C-reactive protein was measured with the use of a solid-phase, 2-site, chemiluminescent, immunometric assay on the IMMULITE 2000 immunoassay analyzer (Siemens Healthcare Diagnostics) with intra-assay and interassay CVs of 3.5% and 5%, respectively. Diabetes status was defined with the use of guidelines from the American Diabetes Association (14). Hypertension was assessed according to the guidelines of the Joint National Committee on Prevention (15). BMI (in kg/m2) was calculated as weight divided by height squared. Cardiovascular disease was self-reported and included histories of congestive heart failure, coronary heart disease, and myocardial infarction.

Cerebrovascular disease and MRI

Cranial MRI was performed with a 1.5-T Symphony superconducting magnet (Siemens). The MRI protocol included 1) T1-weighted axial images, 2) intermediate and T2-weighted conventional spin-echo axial images, and 3) fluid-attenuation inversion recovery (FLAIR) turbo spin-echo axial images. Analyze image-analysis software (1986–2004; Biomedical Imaging Resource) was used to quantify brain segmentation and a region-of-interest analysis. Analysts worked with de-identified data.

White-matter hyperintensities (WMHIs), which are indicators of vascular pathology, were defined as hyperintense changes on intermediate-intensity/FLAIR and T2-weighted images with no corresponding T1 abnormality. Quantitative measurements involved histogram-analysis procedures as previously described (16). High-resolution images were reviewed with the use of primary and secondary trained analysts under the supervision of a neuroradiologist. The interrater reliability was r = 0.92 (P < 0.001).

Brain infarcts were identified and measured with the use of a neuroradiologist to quantify subclinical and clinical cardiovascular disease. Large-vessel infarcts were defined as cortical infarcts of any size and subcortical infarcts >1.5 cm in maximum dimension, whereas small-vessel infarcts were defined as subcortical infarcts from 3 mm to 1.5 cm. To qualify for an infarct, the lesion had an increased signal on T2 and FLAIR images. For subcortical infarcts in white matter, additional low-signal on T1-weighted images was required to be considered as an infarct (17).

Statistical methods

Data are presented as means ± SDs or frequencies (percentages). Multiple linear regression models and logistic regression models were used to examine relations between plasma concentrations of choline and choline-related metabolites and cardiometabolic risk factors, B vitamins, and odds of cardiometabolic-related diseases and cerebrovascular pathology. Independent variables were examined as continuous variables and as categorical data (tertiles) when the outcome variable was dichotomous. When continuous data were not normally distributed, values were log transformed before the analysis. We assessed linear trends across categories of plasma concentrations of choline and choline-related metabolites by assigning each participant the median value for the category and modeling this value as a continuous variable.

Three regression models were used to assess relations while controlling for 1) basic covariates [age, sex, smoking, and drinking status (self-reported alcohol and cigarette use defined as never, current, or past)]; 2) one-carbon metabolism covariates (age, sex, smoking and drinking status, plasma total homocysteine, PLP, and folate), and 3) lipid metabolism covariates (age, sex, smoking and drinking status, LDL cholesterol, HDL cholesterol, triglycerides, and BMI). Subjects with missing covariate data or outliers for plasma choline, betaine, or phosphatidylcholine concentrations were excluded from the analyses (betaine: n = 2; phosphatidylcholine: n = 3) (see Supplemental Figure 1 for details). Results were considered significant at P ≤ 0.05.

RESULTS

Population characteristics

The mean ± SD age was 73 ± 8.1 y, and participants were predominantly women (73%) (Table 1). The mean BMI was in the obese range (30.7 ± 7.7), and there was a high prevalence of hypertension (84%). The attained level of education was moderate with 33% of subjects being high-school graduates, and 44% of subjects having attended at least some college. The mean betaine concentration was greater in men than in women, whereas the mean phosphatidylcholine concentration was greater in women than in men (P ≤ 0.05) (Table 2).

TABLE 1.

Population characteristics

Variable n Value
Age, y 296 73.6 ± 8.1 (59.0–103.0)1
BMI, kg/m2 280 30.7 ± 7.7 (15.4–75.0)
Plasma
 HDL cholesterol, mg/dL 295 50.6 ± 15.0 (21.0–122.0)
 LDL cholesterol, mg/dL 291 108.3 ± 35.3 (13.6–271.4)
 Triglycerides, mg/dL 295 144.6 ± 97.8 (41.0–905.0)
 Glucose, mg/dL 295 111.4 ± 36.8 (51.0–440.0)
 Folate, ng/mL 293 15.3 ± 10.6 (2.4–109.6)
 PLP,2 nm/L 295 83.0 ± 81.0 (8.3–492.6)
 Vitamin B-12, pg/mL 294 575.9 ± 330.4 (166.0–3444.0)
 Total homocysteine, μmol/L 295 11.8 ± 5.5 (4.5–44.3)
Sex, F, % 296 72.6
Hypertension, % 291 84.2
Diabetes, % 295 30.5
Self-reported cardiovascular disease, % 287 35.2
Smoking status (self-reported cigarette use), % 294
 Never 30.6
 Past 45.2
 Current 24.2
Drinking status (self-reported alcoholic beverage intake), % 294
 Never 10.5
 Past 47.0
 Current 42.5
Education, % 296
 None to fourth grade 1.4
 Fifth to eighth grades 7.5
 9th to 11th grades 14.3
 12th grade or high school graduate 32.6
 Some college or bachelor’s degree 36.7
 Graduate school 7.5
1

Mean ± SD; range in parentheses (all such values).

2

PLP, pyridoxal-P.

TABLE 2.

Plasma concentration of choline and choline-related compounds1

Men Women All
Choline, μM 10.3 ± 3.8 9.6 ± 3.2 9.7 ± 3.4
Betaine, μM 46.1 ± 25.2 38.8 ± 15.3 40.8 ± 18.92
Phosphatidylcholine, μM 2030 ± 418 2315 ± 417 2237 ± 4352
1

All values are means ± SDs. n = 296 (81 men and 215 women).

2

Significant difference between men and women, P ≤ 0.05 (independent t test).

Plasma choline and cardiometabolic risk factors

LDL cholesterol and triglyceride were inversely associated with plasma betaine and positively associated with plasma phosphatidylcholine concentrations (Table 3). HDL cholesterol was inversely associated with plasma choline and positively associated with phosphatidylcholine concentrations. There were no significant relations between blood glucose and plasma choline and choline-related metabolite concentrations. Plasma total homocysteine was positively associated with the plasma choline concentration. BMI was positively associated with plasma choline and inversely associated with the plasma phosphatidylcholine concentration. Waist circumference and the concentration of C-reactive protein were both inversely associated with the plasma phosphatidylcholine concentration.

TABLE 3.

Multiple linear regression coefficients for cardiometabolic risk factors in relation to plasma concentrations of choline, betaine, and phosphatidylcholine1

Choline
Betaine
Phosphatidylcholine
n β ± SE P n β ± SE P n β ± SE P
Plasma
 LDL cholesterol 287 −0.05 ± 0.06 0.40 285 −0.13 ± 0.06 0.04 284 0.0002 ± 0.00002 <0.0001
 HDL cholesterol 291 −0.11 ± 0.05 0.03 289 −0.06 ± 0.05 0.28 288 0.0001 ± 0.00002 <0.0001
 Triglycerides 291 0.07 ± 0.09 0.46 289 −0.31 ± 0.09 0.001 288 0.0002 ± 0.00003 <0.0001
 Glucose 291 −0.05 ± 0.05 0.31 289 −0.07 ± 0.05 0.14 288 −0.00001 ± 0.00002 0.40
 tHcy 291 0.64 ± 0.16 <0.0001 289 −0.20 ± 0.16 0.23 288 −0.00009 ± 0.00006 0.12
BMI 277 0.13 ± 0.04 0.002 275 0.01 ± 0.04 0.85 275 −0.00006 ± 0.00001 <0.0001
CRP 291 0.43 ± 0.23 0.06 289 −0.03 ± 0.23 0.90 288 −0.0002 ± 0.00008 0.01
Waist circumference 252 0.04 ± 0.30 0.20 250 −0.02 ± 0.03 0.56 250 −0.17 ± 0.05 0.001
1

Multiple linear regression analysis was adjusted for age, sex, smoking status, and drinking status. All variables except for phosphatidylcholine were log transformed for the statistical analysis. CRP, C-reactive protein; tHcy, total homocysteine.

Plasma choline and B vitamins

A multiple linear regression was used to assess potential relations between plasma B-vitamin and choline and choline-related metabolite concentrations while controlling for age and sex (Table 4). Plasma PLP was positively associated with the phosphatidylcholine concentration. Plasma folate had a weak positive relation with choline and betaine concentrations. There were no significant relations between plasma vitamin B-12 and plasma choline, betaine, or phosphatidylcholine concentrations.

TABLE 4.

Multiple linear regression coefficients for plasma concentrations of B vitamins in relation to plasma concentrations of choline, betaine, and phosphatidylcholine1

Choline
Betaine
Phosphatidylcholine
n β ± SE P n β ± SE P n β ± SE P
PLP2 295 0.06 ± 0.16 0.70 293 0.16 ± 0.16 0.31 292 0.0001 ± 0.00005 0.05
Vitamin B-12 294 −0.08 ± 0.09 0.33 292 −0.02 ± 0.09 0.80 291 −0.00005 ± 0.00003 0.12
Folate 293 0.31 ± 0.11 0.004 291 0.32 ± 0.11 0.003 290 0.000004 ± 0.00004 0.90
1

Multiple linear regression analysis was adjusted for age and sex. All variables except for phosphatidylcholine were log transformed for the statistical analysis.

2

PLP, pyridoxal-P.

Plasma choline and associations with cardiometabolic-related disease

Odds of self-reported cardiovascular disease were greater in subjects with plasma choline concentrations in the second or third tertile than in the first tertile after controlling for age, sex, and smoking and alcohol use (Table 5, model 1). These relations remained significant when controlling for plasma total homocysteine, PLP, and folate concentrations (model 2) as well as for BMI and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations (model 3). The linear trend across tertiles was also significant in all models, which indicated a positive association.

TABLE 5.

Multivariate-adjusted ORs for cardiometabolic-related disease risk factors across tertile categories of plasma choline-related compounds1

Choline
Betaine
Phosphatidylcholine
Tertile (μM range)
Tertile (μM range)
Tertile (μM range)
n 1 (3.6–8.1) 2 (8.12–10.4) 3 (10.42–26.9) P-trend n 1 (13.6–33.3) 2 (33.32–43.66) 3 (43.662–239.4) P-trend n P-trend 1 (1093.2–2044.7) 2 (2044.8–2383.0) 3 (2383.1–3856.5)
CVD
 Model 1 283 1.00 (—)2 2.34 (1.20, 4.58) 3.26 (1.67, 6.36) 0.001 281 1.00 (—) 1.08 (0.58, 2.04) 1.24 (0.65, 2.37) 0.51 280 1.00 (—) 0.83 (0.45, 1.54) 0.66 (0.34, 1.27) 0.21
 Model 2 280 1.00 (—) 2.28 (1.16, 4.48) 3.08 (1.55, 6.13) 0.002 278 1.00 (—) 1.12 (0.59, 2.15) 1.24 (0.62, 2.34) 0.59 277 1.00 (—) 0.89 (0.48, 1.67) 0.75 (0.38, 1.48) 0.41
 Model 3 263 1.00 (—) 2.35 (1.15, 4.78) 3.06 (1.50, 6.22) 0.002 261 1.00 (—) 1.21 (0.62, 2.39) 1.23 (0.61, 2.48) 0.57 261 1.00 (—) 0.74 (0.34, 1.59) 0.44 (0.15, 1.26) 0.12
Hypertension
 Model 1 287 1.00 (—) 0.59 (0.28, 1.25) 2.25 (0.85, 5.93) 0.15 285 1.00 (—) 1.10 (0.50, 2.42) 1.25 (0.54, 2.89) 0.60 284 1.00 (—) 0.89 (0.35, 2.22) 0.42 (0.18, 0.98) 0.03
 Model 2 284 1.00 (—) 0.54 (0.25, 1.17) 1.59 (0.58, 4.33) 0.56 282 1.00 (—) 1.15 (0.51, 2.61) 1.11 (0.45, 2.70) 0.81 281 1.00 (—) 0.98 (0.38, 2.56) 0.43 (0.18, 1.06) 0.04
 Model 3 267 1.00 (—) 0.59 (0.27, 1.31) 1.93 (0.71, 5.26) 0.30 265 1.00 (—) 1.08 (0.47, 2.48) 1.25 (0.50, 3.11) 0.63 265 1.00 (—) 0.68 (0.22, 2.08) 0.27 (0.06, 1.15) 0.05
DM
 Model 1 291 1.00 (—) 0.75 (0.39, 1.43) 0.91 (0.48, 1.72) 0.75 289 1.00 (—) 1.26 (0.67, 2.35) 0.51 (0.25, 1.03) 0.06 288 1.00 (—) 0.86 (0.46, 1.59) 0.39 (0.19, 0.79) 0.01
 Model 2 289 1.00 (—) 0.72 (0.37, 1.38) 0.82 (0.42, 1.60) 0.55 287 1.00 (—) 1.22 (0.64, 2.30) 0.43 (0.21, 0.89) 0.02 286 1.00 (—) 0.88 (0.47, 1.65) 0.41 (0.20, 0.83) 0.01
 Model 3 272 1.00 (—) 0.61 (0.31, 1.24) 0.73 (0.36, 1.46) 0.36 270 1.00 (—) 1.13 (0.57, 2.23) 0.47 (0.22, 1.00) 0.05 270 1.00 (—) 0.96 (0.44, 2.12) 0.39 (0.12, 1.22) 0.09
1

Model 1 was a logistic regression that was adjusted for age, sex, smoking status, and drinking status. Model 2 was a logistic regression that was adjusted as for model 1 and for plasma (log) total homocysteine, (log) pyridoxal-P, and (log) folate. Model 3 was a logistic regression that was adjusted as for model 1 and for plasma (log) LDL cholesterol, (log) HDL cholesterol, and (log) triglycerides and (log) BMI. CVD, cardiovascular disease; DM, diabetes mellitus.

2

OR; 95% CI in parentheses (all such values).

Odds of hypertension were lower in subjects with plasma phosphatidylcholine concentrations in the third tertile than in the first tertile when controlled for age, sex, smoking, and alcohol use. These relations did not remain significant when plasma total homocysteine, PLP, and folate concentrations were included (model 2) or when BMI and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations were added to the model (model 3). However, the linear trend across tertiles was significant in all models, which indicated an inverse association.

Odds of diabetes were lower in subjects with betaine concentrations in the third tertile than in the first tertile after controlling for age, sex, smoking and alcohol use, and plasma total homocysteine, PLP, and folate concentrations (model 2). The linear trend across tertiles was also significant in this model and after controlling for BMI and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations (model 3), which indicated an inverse relation between plasma betaine and the presence of type 2 diabetes. Odds of diabetes were also lower in subjects with plasma phosphatidylcholine concentrations in the third tertile than in the first tertile when controlling for age, sex, and smoking and alcohol use (model 1). These relations remained significant when controlling for plasma total homocysteine, PLP, and folate concentrations (model 2) but not after controlling for BMI and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations (model 3). The linear trend across tertiles was also significant in models 1 and 2, which indicated an inverse relation.

Plasma choline and cerebrovascular pathology

Relations between plasma choline and choline-related compounds and MRI volumetric measurements are presented in Tables 6 and 7. There were no significant associations between concentrations of plasma choline-related compounds and the volume of WMHIs. The odds of a small-vessel infarct were lower in individuals with plasma choline concentrations in the second or third tertiles than in the first tertile after controlling for age, sex, smoking and alcohol use, and plasma total homocysteine, PLP, and folate concentrations (model 2). These relations were NS after controlling for age, sex, smoking and alcohol use, BMI, and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations (model 3). The linear trend across tertiles was also significant in models 1 and 2 (P ≤ 0.05), which indicated an inverse relation between plasma choline concentrations and the presence of cerebral small-vessel infarcts.

TABLE 6.

Multiple linear regression coefficients for the volume of white-matter hyperintensities in relation to plasma concentrations of choline, betaine, and phosphatidylcholine1

Choline
Betaine
Phosphatidylcholine
n β ± SE P n β ± SE P n β ± SE P
Model 1 292 0.08 ± 0.21 0.70 290 −0.04 ± 0.21 0.86 289 0.00005 ± 0.00008 0.48
Model 2 289 0.10 ± 0.23 0.68 287 −0.03 ± 0.22 0.88 286 0.00006 ± 0.00008 0.45
Model 3 272 0.24 ± 0.29 0.73 270 −0.03 ± 0.23 0.89 270 0.00009 ± 0.0002 0.58
1

Model 1 was a logistic regression that was adjusted for age, sex, smoking status, and drinking status. All variables except for phosphatidylcholine were log transformed for the statistical analysis. Model 2 was a logistic regression that was adjusted as for model 1 and for plasma (log) total homocysteine, (log) pyridoxal-P, and (log) folate. Model 3 was a logistic regression that was adjusted as for model 1 and for plasma (log) LDL cholesterol, (log) HDL cholesterol, and (log) triglycerides and (log) BMI.

TABLE 7.

Multivariate-adjusted ORs for the presence of cerebrovascular infarct across tertile categories of plasma choline related compounds1

Choline
Betaine
Phosphatidylcholine
Tertile (μM range)
Tertile (μM range) Tertile (μM range)
n 1 (3.6–8.1) 2 (8.12–10.4) 3 (10.42–26.9) P-trend n 1 (13.6–33.3) 2 (33.32–43.66) 3 (43.662–239.4) P-trend n 1 (1093.2–2044.7) 2 (2044.75–2383.0) 3 (2383.1–3856.5) P-trend
SVINF
 Model 1 287 1.00 (—)2 0.44 (0.21, 0.90) 0.51 (0.25, 1.02) 0.05 285 1.00 (—) 0.80 (0.40, 1.60) 0.57 (0.27, 1.18) 0.13 284 1.00 (—) 0.55 (0.26, 1.15) 0.89 (0.44, 1.82) 0.81
 Model 2 284 1.00 (—) 0.41 (0.20, 0.87) 0.44 (0.21, 0.93) 0.03 282 1.00 (—) 0.83 (0.40, 1.70) 0.63 (0.29, 1.37) 0.25 281 1.00 (—) 0.57 (0.27, 1.21) 0.90 (0.43, 1.89) 0.83
 Model 3 267 1.00 (—) 0.48 (0.23, 1.02) 0.52 (0.25, 1.11) 0.08 265 1.00 (—) 0.77 (0.37, 1.59) 0.53 (0.24, 1.15) 0.10 265 1.00 (—) 0.43 (0.18, 1.05) 0.55 (0.18, 1.70) 0.42
LVINF
 Model 1 287 1.00 (—) 3.16 (1.07, 9.34) 3.29 (1.12, 9.66) 0.04 285 1.00 (—) 1.29 (0.52, 3.18) 0.76 (0.28, 2.02) 0.54 284 1.00 (—) 0.60 (0.24, 1.49) 0.87 (0.35, 2.14) 0.78
 Model 2 284 1.00 (—) 3.30 (1.11, 9.84) 3.73 (1.24, 11.16) 0.02 282 1.00 (—) 1.48 (0.59, 3.76) 0.90 (0.33, 2.49) 0.79 281 1.00 (—) 0.61 (0.24, 1.54) 0.89 (0.35, 2.25) 0.83
 Model 3 267 1.00 (—) 4.15 (1.22, 14.05) 5.74 (1.69, 19.50) 0.01 265 1.00 (—) 1.06 (0.39, 2.87) 0.72 (0.25, 2.04) 0.50 265 1.00 (—) 0.77 (0.25, 2.41) 0.94 (0.22, 4.02) 1.00
1

Model 1 was a logistic regression that was adjusted for age, sex, smoking status, and drinking status. Model 2 was a logistic regression that was adjusted as for model 1 and for plasma (log) total homocysteine, (log) pyridoxal-P, and (log) folate. Model 3 was a logistic regression that was adjusted as for model 1 and for plasma (log) LDL cholesterol, (log) HDL cholesterol, and (log) triglycerides and (log) BMI. LVINF, large-vessel infarct; SVINF, small-vessel infarct.

2

OR; 95% CI in parentheses (all such values).

Conversely, odds of a large-vessel infarct were greater in individuals with plasma choline concentrations in the second or third tertile than in the first tertile after controlling for age, sex, smoking and alcohol use, and plasma total homocysteine, PLP, and folate concentrations (model 2) as well as when controlling for age, sex, smoking and alcohol use, BMI, and plasma LDL-cholesterol, HDL-cholesterol, and triglyceride concentrations (model 3). The linear trend across tertiles was also significant in all models (P ≤ 0.05), which indicated a positive association between plasma choline concentrations and the presence of large-vessel infarcts.

DISCUSSION

In a population of community-dwelling elders, choline and its metabolites had differential associations with cardiometabolic risk factors and subtypes of vascular disease. We showed that higher plasma choline concentrations were associated with an unfavorable cardiometabolic risk-factor profile (lower HDL cholesterol, higher total homocysteine, and higher BMI) and greater odds of large-vessel cerebral vascular disease or a history of cardiovascular disease but lower odds of small-vessel cerebral vascular disease. Conversely, higher plasma betaine concentrations were associated with a favorable cardiometabolic risk-factor profile (lower LDL cholesterol and triglycerides) and lower odds of diabetes. Higher plasma phosphatidylcholine concentrations were associated with characteristics of both a favorable cardiometabolic risk-factor profile (higher HDL cholesterol, lower BMI, lower C-reactive protein, lower waist circumference, and lower odds of hypertension and diabetes) and an unfavorable profile (higher LDL cholesterol and triglycerides). Our findings that plasma phosphatidylcholine concentrations were associated with both favorable and unfavorable cardiometabolic risk factors (higher HDL cholesterol, LDL cholesterol, and triglycerides) were not unexpected because phosphatidylcholine biosynthesis in the liver is involved in the assembly and secretion of lipoprotein particles. Any disruption would result in the impairment of hepatic VLDL- and HDL-cholesterol secretions and an increase in hepatic steatosis (18).

In general, our findings were consistent with previous reports of the relation between choline and its metabolites and cardiovascular disease and metabolic disorders. Higher betaine concentrations were associated with a favorable risk-factor profile, and higher choline concentrations were associated with an unfavorable risk-factor profile (8). Plasma choline and betaine concentrations were comparable with those previously reported (8, 1922). Compared with the overall US population aged ≥65 y, the NAME population has a higher percentage of women, more college graduates, higher prevalences of diabetes and hypertension, and higher reported current alcohol consumption and cigarette use (23). In addition, the NAME population consisted of community-dwelling elders in contrast with patient populations who have been studied by other authors. For example, a study of a cohort with acute coronary syndrome concluded that high and low plasma betaine concentrations were associated with increased risks of secondary heart failure and acute myocardial infarction, respectively (24). A second study in the same cohort concluded that lower plasma betaine concentrations were associated with higher concentrations of triglycerides and non-HDL cholesterol (25). In contrast, in the NAME population, although there was an inverse relation between plasma betaine and both LDL-cholesterol and triglyceride concentrations, there was no significant relation between plasma betaine concentrations and odds of cardiovascular or cerebrovascular disease. However, our findings were similar to those of a large-cohort study of patients with stable angina pectoris that observed no association between plasma betaine and acute myocardial infarction or all-cause mortality (26).

Minor discrepancies between our results and findings from some other studies may have been partially due to whether the analyses were based on plasma concentrations or dietary intakes of choline and its metabolites. Dietary choline does not account fully for both the endogenous and exogenous pathways; however, circulating choline and metabolites may not be sensitive markers of choline status (27) because disturbances in circulating choline metabolites arise early in disease development (28), and lipid overload (indicated by elevated BMI) may perturb choline metabolism, thereby resulting in reduced plasma betaine concentrations (20). For example, we showed that subjects with highest plasma phosphatidylcholine concentrations had lower odds of type 2 diabetes, which somewhat countered the findings of a brief report on dietary intake data from 3 ongoing cohorts in the United States that showed an increase in the risk of type 2 diabetes that was associated with increased dietary choline from phosphatidylcholine (29). However, our data were consistent with recent prospective studies from both a Norwegian sample of patients with suspected stable angina pectoris (30) and an American general-population sample (31) that showed an inverse relation between plasma betaine and the presence of type 2 diabetes. In addition, analyses of 4 other cohorts showed no relation between dietary intake of betaine or choline and the incidence of peripheral artery disease (32), incidence of coronary heart disease (33), or cardiovascular disease risk (34). However, an analysis of a large prospective cohort with stable angina pectoris showed increased risk of acute myocardial infarction in individuals with plasma choline concentrations in the highest quartile than in the lowest quartile (26). Although this relation was only shown in smokers, in the NAME population, subjects with choline concentrations in the upper 2 tertiles had significantly greater odds of self-reported cardiovascular disease than did subjects with plasma choline concentrations in the lower tertile in all models (which included smoking status as a covariate).

The possibility cannot be ruled out that the relation between choline and vascular disease is different when considering the location of blood vessels, cerebrovascular or cardiovascular, as well as the size of the blood vessels. This possibility may have been the case in the NAME population in whom we showed that the odds of a small-vessel infarct were inversely related with the plasma choline concentration, whereas the odds of a large-vessel infarct were positively related with the plasma choline concentration. Similarly, a longitudinal study showed a significant inverse relation between dietary choline intake and the volume of cerebral WMHIs, which is a measure of myelin integrity that has been associated with small-vessel disease (35). Furthermore, results from a prospective cohort study of African Americans recently reported that higher dietary choline intake was associated with lower risk of incident ischemic stroke (36). These results support the concept that the relation between choline and its related compounds and vascular disease is dependent on the vessel location and size. Coronary artery disease is a large-vessel disease, whereas the volume of WMHIs is associated with small-vessel disease. It is also possible that this discrepancy may have arisen because of different kinds of bias that were introduced in an elderly population with high rates of diabetes and hypertension (which would have put them at increased stroke risk), because patients who were dying from a large-vessel infarct may have been underrepresented in the cross-sectional analysis, or because cerebral infarctions that are related to large-vessel disease were related to embolic rather than thrombotic events.

The major limitations of this study were its cross-sectional nature and specific subject population, which prevented causal conclusions or the generalization of results and may have caused the study to be more prone to bias than a prospective study would have been. We were also limited by the plasma metabolites that were measured, including measurements of plasma dimethylglycine [which is an indicator of the methylation capacity from choline (37)] and trimethylamine N-oxide [which is produced when bacteria in the large intestines digest dietary phosphatidylcholine and has been associated with cardiovascular disease (2)], that would have provided a more complete picture of choline metabolism. Single nucleotide polymorphisms in genes that are involved in choline metabolism may modify these relations and have been shown to be associated with odds of cerebrovascular disease (38) or cardiovascular disease (39) as well to alter the susceptibility to choline deficiency (40, 41).

In conclusion, our findings that choline and its metabolites have differential associations with cardiometabolic risk factors and subtypes of vascular disease suggest differing roles in the pathogenesis of small-and large-vessel cardiovascular and cerebrovascular disease. These results help to provide metabolic insight into the processes that contribute to or protect from these disease outcomes and suggest a need to further explore these associations to better understand the underlying factors.

Acknowledgments

We thank Gayle Petty and the Nutrition Evaluation Laboratory at the Jean Mayer USDA Human Nutrition Research Center on Aging at Tufts University for analytical contributions.

The authors’ responsibilities were as follows—AJR: designed the research, analyzed the data, performed the statistical analysis, and wrote the manuscript; SZ: conducted the research; RAB: designed and conducted the research; EJJ and AHL: designed the research and wrote the manuscript; GTR: performed the statistical analysis; IHR: designed the research, provided essential materials, and wrote the manuscript; CES: designed the research; SHZ: provided essential materials and wrote the manuscript; TMS: designed and conducted the research and wrote the manuscript; and all authors: read and approved the final manuscript. SHZ is the founder of Nutrigene Sciences LLC, which is a company in which he owns stock equity. SHZ serves on the Scientific Advisory Board of Metabolon, Genovive, and Dupont Nutrition. None of these organizations had a financial interest in the outcomes of this article, but Genovive might indirectly benefit from the article. SHZ has been an expert witness in a court case that involved a diet supplement that contains choline. There was no direct benefit to either party in the case, but either party might indirectly benefit from the article. SHZ has received research support from Nestlé Nutrition and support from Hershey Inc. that were not related to this article. SHZ has received support from the Egg Nutrition Center on the topic of choline and trimethylamine N-oxide, which was not directly related to this article but was relevant to cardiometabolic risks and choline. The remaining authors reported no conflicts of interest related to the study.

Footnotes

7

Abbreviations used: FLAIR, fluid-attenuation inversion recovery; NAME, Nutrition Aging, and Memory in Elders; PLP, pyridoxal-P; WMHI, white-matter hyperintensity.

REFERENCES

  • 1.Zeisel SH, Blusztajn JK. Choline and human nutrition. Annu Rev Nutr 1994;14:269–96. [DOI] [PubMed] [Google Scholar]
  • 2.Rak K, Rader DJ. Cardiovascular disease: the diet-microbe morbid union. Nature 2011;472:40–1. [DOI] [PubMed] [Google Scholar]
  • 3.Vance DE. Role of phosphatidylcholine biosynthesis in the regulation of lipoprotein homeostasis. Curr Opin Lipidol 2008;19:229–34. [DOI] [PubMed] [Google Scholar]
  • 4.Yao ZM, Vance DE. The active synthesis of phosphatidylcholine is required for very low density lipoprotein secretion from rat hepatocytes. J Biol Chem 1988;263:2998–3004. [PubMed] [Google Scholar]
  • 5.Olthof MR, van Vliet T, Verhoef P, Zock PL, Katan MB. Effect of homocysteine-lowering nutrients on blood lipids: results from four randomised, placebo-controlled studies in healthy humans. PLoS Med 2005;2:e135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Wald DS, Law M, Morris JK. Homocysteine and cardiovascular disease: evidence on causality from a meta-analysis. BMJ 2002;325:1202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.King WD, Ho V, Dodds L, Perkins SL, Casson RI, Massey TE. Relationships among biomarkers of one-carbon metabolism. Mol Biol Rep 2012;39:7805–12. [DOI] [PubMed] [Google Scholar]
  • 8.Konstantinova SV, Tell GS, Vollset SE, Nygard O, Bleie O, Ueland PM. Divergent associations of plasma choline and betaine with components of metabolic syndrome in middle age and elderly men and women. J Nutr 2008;138:914–20. [DOI] [PubMed] [Google Scholar]
  • 9.Scott TM, Peter I, Tucker KL, Arsenault L, Bergethon P, Bhadelia R, Buell J, Collins L, Dashe JF, Griffith J, et al. The Nutrition, Aging, and Memory in elders (NAME) study: design and methods for a study of micronutrients and cognitive function in a homebound elderly population. Int J Geriatr Psychiatry 2006;21:519–28. [DOI] [PubMed] [Google Scholar]
  • 10.Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol 1959;37:911–7. [DOI] [PubMed] [Google Scholar]
  • 11.Koc H, Mar MH, Ranasinghe A, Swenberg JA, Zeisel SH. Quantitation of choline and its metabolites in tissues and foods by liquid chromatography/electrospray ionization-isotope dilution mass spectrometry. Anal Chem 2002;74:4734–40. [DOI] [PubMed] [Google Scholar]
  • 12.Araki A, Sako Y. Determination of free and total homocysteine in human plasma by high-performance liquid chromatography with fluorescence detection. J Chromatogr 1987;422:43–52. [DOI] [PubMed] [Google Scholar]
  • 13.Shin YS, Rasshofer R, Friedrich B, Endres W. Pyridoxal-5′-phosphate determination by a sensitive micromethod in human blood, urine and tissues; its relation to cystathioninuria in neuroblastoma and biliary atresia. Clin Chim Acta 1983;127:77–85. [DOI] [PubMed] [Google Scholar]
  • 14.Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183–97. [DOI] [PubMed] [Google Scholar]
  • 15.The sixth report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. Arch Intern Med 1997;157:2413–46. Erratum in: Arch Intern Med 1998;158:573. [DOI] [PubMed] [Google Scholar]
  • 16.DeCarli C, Maisog J, Murphy DG, Teichberg D, Rapoport SI, Horwitz B. Method for quantification of brain, ventricular, and subarachnoid CSF volumes from MR images. J Comput Assist Tomogr 1992;16:274–84. [DOI] [PubMed] [Google Scholar]
  • 17.Bryan RN, Wells SW, Miller TJ, Elster AD, Jungreis CA, Poirier VC, Lind BK, Manolio TA. Infarctlike lesions in the brain: prevalence and anatomic characteristics at MR imaging of the elderly–data from the Cardiovascular Health Study. Radiology 1997;202:47–54. [DOI] [PubMed] [Google Scholar]
  • 18.Cole LK, Vance JE, Vance DE. Phosphatidylcholine biosynthesis and lipoprotein metabolism. Biochim Biophys Acta 2012;1821:754–61. [DOI] [PubMed] [Google Scholar]
  • 19.Nurk E, Refsum H, Bjelland I, Drevon CA, Tell GS, Ueland PM, Vollset SE, Engedal K, Nygaard HA, Smith DA. Plasma free choline, betaine and cognitive performance: the Hordaland Health Study. Br J Nutr 2013;109:511–9. [DOI] [PubMed] [Google Scholar]
  • 20.Yan J, Winter LB, Burns-Whitmore B, Vermeylen F, Caudill MA. Plasma choline metabolites associate with metabolic stress among young overweight men in a genotype-specific manner. Nutr Diabetes 2012;2:e49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Wallace JM, McCormack JM, McNulty H, Walsh PM, Robson PJ, Bonham MP, Duffy ME, Ward M, Molloy AM, Scott JM, et al. Choline supplementation and measures of choline and betaine status: a randomised, controlled trial in postmenopausal women. Br J Nutr 2012;108:1264–71. [DOI] [PubMed] [Google Scholar]
  • 22.Eussen SJ, Ueland PM, Clarke R, Blom HJ, Hoefnagels WH, van Staveren WA, de Groot LC. The association of betaine, homocysteine and related metabolites with cognitive function in Dutch elderly people. Br J Nutr 2007;98:960–8. [DOI] [PubMed] [Google Scholar]
  • 23.Administration on Aging, Administration for Community Living, US Department of Health and Human Services. [Internet]. A profile of older Americans: 2014. [cited 2016 Mar 28]. Available from: https://aoa.acl.gov/aging_statistics/profile/index.aspx. [Google Scholar]
  • 24.Lever M, George PM, Elmslie JL, Atkinson W, Slow S, Molyneux SL, Troughton RW, Richards AM, Frampton CM, Chambers ST. Betaine and secondary events in an acute coronary syndrome cohort. PLoS One 2012;7:e37883. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Lever M, George PM, Atkinson W, Molyneux SL, Elmslie JL, Slow S, Richards AM, Chambers ST. Plasma lipids and betaine are related in an acute coronary syndrome cohort. PLoS One 2011;6:e21666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Schartum-Hansen H, Pedersen ER, Svingen GF, Ueland PM, Seifert R, Ebbing M, Strand E, Bleie O, Nygard O. Plasma choline, smoking, and long-term prognosis in patients with stable angina pectoris. Eur J Prev Cardiol 2015;22:606–14. [DOI] [PubMed] [Google Scholar]
  • 27.Fischer LM, daCosta KA, Kwock L, Stewart PW, Lu TS, Stabler SP, Allen RH, Zeisel SH. Sex and menopausal status influence human dietary requirements for the nutrient choline. Am J Clin Nutr 2007;85:1275–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bae S, Ulrich CM, Neuhouser ML, Malysheva O, Bailey LB, Xiao L, Brown EC, Cushing-Haugen KL, Zheng Y, Cheng TY, et al. Plasma choline metabolites and colorectal cancer risk in the Women’s Health Initiative observational study. Cancer Res 2014;74:7442–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Li Y, Wang DD, Chiuve SE, Manson JE, Willett WC, Hu FB, Qi L. Dietary phosphatidylcholine intake and type 2 diabetes in men and women. Diabetes Care 2015;38:e13–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Svingen GF, Schartum-Hansen H, Pedersen ER, Ueland PM, Tell GS, Mellgren G, Njolstad PR, Seifert R, Strand E, Karlsson T, et al. Prospective associations of systemic and urinary choline metabolites with incident type 2 diabetes. Clin Chem 2016;62:755–65. [DOI] [PubMed] [Google Scholar]
  • 31.Walford GA, Ma Y, Clish C, Florez JC, Wang TJ, Gerszten RE, Diabetes Prevention Program Research Group. Metabolite profiles of diabetes incidence and intervention response in the diabetes prevention program. Diabetes 2016;65:1424–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Bertoia ML, Pai JK, Cooke JP, Joosten MM, Mittleman MA, Rimm EB, Mukamal KJ. Plasma homocysteine, dietary B vitamins, betaine, and choline and risk of peripheral artery disease. Atherosclerosis 2014;235:94–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Bidulescu A, Chambless LE, Siega-Riz AM, Zeisel SH, Heiss G. Usual choline and betaine dietary intake and incident coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) study. BMC Cardiovasc Disord 2007;7:20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Dalmeijer GW, Olthof MR, Verhoef P, Bots ML, van der Schouw YT. Prospective study on dietary intakes of folate, betaine, and choline and cardiovascular disease risk in women. Eur J Clin Nutr 2008;62:386–94. [DOI] [PubMed] [Google Scholar]
  • 35.Poly C, Massaro JM, Seshadri S, Wolf PA, Cho E, Krall E, Jacques PF, Au R. The relation of dietary choline to cognitive performance and white-matter hyperintensity in the Framingham Offspring Cohort. Am J Clin Nutr 2011;94:1584–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Millard HR, Musani SK, Dibaba DT, Talegawkar SA, Taylor HA, Tucker KL, Bidulescu A. Dietary choline and betaine; associations with subclinical markers of cardiovascular disease risk and incidence of CVD, coronary heart disease and stroke: the Jackson Heart Study. Eur J Nutr 2016. Aug 22 (Epub ahead of print; DOI: 10.1007/s00394-016-1296-8). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Obeid R. The metabolic burden of methyl donor deficiency with focus on the betaine homocysteine methyltransferase pathway. Nutrients 2013;5:3481–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Lai CQ, Parnell LD, Troen AM, Shen J, Caouette H, Warodomwichit D, Lee YC, Crott JW, Qiu WQ, Rosenberg IH, et al. MAT1A variants are associated with hypertension, stroke, and markers of DNA damage and are modulated by plasma vitamin B-6 and folate. Am J Clin Nutr 2010;91:1377–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Weisberg IS, Park E, Ballman KV, Berger P, Nunn M, Suh DS, Breksa AP 3rd, Garrow TA, Rozen R. Investigations of a common genetic variant in betaine-homocysteine methyltransferase (BHMT) in coronary artery disease. Atherosclerosis 2003;167:205–14. [DOI] [PubMed] [Google Scholar]
  • 40.Zeisel SH. Gene response elements, genetic polymorphisms and epigenetics influence the human dietary requirement for choline. IUBMB Life 2007;59:380–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.da Costa KA, Kozyreva OG, Song J, Galanko JA, Fischer LM, Zeisel SH. Common genetic polymorphisms affect the human requirement for the nutrient choline. FASEB J 2006;20:1336–44. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from The American Journal of Clinical Nutrition are provided here courtesy of American Society for Nutrition

RESOURCES