Introduction

The way to get vitamin D was from diet, dietary supplements or sunlight and metabolized to 25-hydroxyvitamin D [25(OH)D] in liver. Serum level of 25(OH)D could be used to determine the vitamin D nutritional status1,2. Conventionally, 25(OH)D was transferred to active form, 1,25-dihydroxyvitamin D [1,25(OH)2D], in kidney for regulating the metabolism of calcium, phosphorus and bone1,3. Recent evidence suggests that vitamin D also has systemic nonskeletal effects that include modulation of cancer cell progression, the cardiovascular system, and multiple autoimmune diseases through the genomic and nongenomic (rapid-response) cellular signaling by regulation of vitamin D receptor (VDR), which is the binding site of the 1,25(OH)2D. VDR has been found in a variety of cells, including cardiomyocytes, endothelial cells, activated T cells, as well as breast, colon, and prostate cancer cells3,4. The active form of vitamin D—1,25(OH)2D may modulate the tissue matrix metalloproteinase (MMP) and anti-proteinase inhibitors (TIMP) system, which was found to be involved in atherosclerosis, whereas MMP level was decreased after vitamin D supplement5,6.

Current evidence regarding the association between vitamin D deficiency and atherosclerosis is contradictory and inconclusive7,8,9. Lowered vitamin D level has been reported to be one of risk factors for coronary heart disease in patients with type 2 diabetes mellitus10, chronic kidney disease11, and human immunodeficiency virus (HIV)-infected African Americans12,13 Vitamin D deficiency was also related to obesity, higher Framingham risk scores, waist circumference, and body mass index (BMI) in different ethnicities14,15,16. On the contrary, in a large longitudinal community-based cohort of 6459 participants, no association between vitamin D concentration and adverse cardiovascular events was found. Instead, vitamin D deficiency was reported to be associated with higher parathyroid hormone (PTH ≥ 65 pg/mL) level, increased left ventricular mass, and risk for incident heart failure17. Vitamin D supplementation has been inconsistently shown to reduce cardiovascular risk factors in a meta-analysis of randomized controlled studies, after adjusting blood pressure, lipid profile, insulin resistance, and vascular function18,19,20. Moreover, vitamin D supplementation has been suggested to decrease the risk of heart failure, but not that of myocardial infarction or stroke21.

Since current evidence regarding the association between vitamin D deficiency and cardiovascular disease/metabolic disorder is contradictory and inconclusive, in this large-scale observational study, we investigated the relationship between serum vitamin D concentration and subclinical atherosclerosis in elderly Asian population of the I-Lan Longitudinal Aging Study (ILAS), which included subjects aged 50 years and above without known active disease22. This is currently the largest study to investigate vitamin D deficiency and subclinical cardiovascular disease in an elderly Asian population.

Results

A total of 1798 participants were analyzed and divided into two groups according to the serum 25(OH)D level (sufficient, > 20 ng/mL and deficient, ≤ 20 ng/mL). Mean age of the study population was 62.0 (55.9–71.1) years old, and 604 subjects (33.6%) were identified as having serum 25(OH)D ≤ 20 ng/mL. Subjects with serum 25(OH)D deficiency were more likely to be younger, female, and less smokers; and they had lower prevalence of hypertension, dyslipidemia, and lipid-lowering agent usage. In addition, they had higher estimated glomerular filtration rate (eGFR) and iPTH concentration as well as lower serum uric acid level, hip BMD, and cIMT. Baseline characteristics of the study cohort are summarized in Table 1.

Table 1 Baseline characteristics of patients grouped by serum 25(OH)D levels.
Full size table

Table 2 lists the correlation coefficients between cIMT and serum 25(OH)D concentration with other variables. Age, male gender, current smoker, hypertension, antihypertensive drug usage, and serum uric acid level were positively correlated with both cIMT and serum 25(OH)D. Meanwhile, serum HDL level, eGFR, and total body fat were negatively correlated with both cIMT and serum 25(OH)D concentration. Metabolic syndrome, BMI, FBG, and HOMA-IR were negatively related to serum 25(OH)D concentration but positively associated with cIMT. Additionally, serum LDL, TG, and iPTH levels were negatively correlated with serum 25(OH)D concentration but were not associated with cIMT (p < 0.01).

Table 2 Correlation coefficients of the serum 25(OH)D concentration and the carotid intima-media thickness with other cardiovascular risk factors.
Full size table

Compared to subjects with serum 25(OH)D sufficiency, patients with serum 25 (OH)D deficiency were associated with lower risk of subclinical atherosclerosis (crude OR: 0.63, 95% CI 0.50–0.81, p < 0.001), according to univariate analysis (Table 3, model 1). After adjusting for age, gender, smoking, BMI, diabetes mellitus, serum 25(OH)D < 20 ng/mL was not associated with increased risk of subclinical atherosclerosis (adjusted OR: 0.84, 95% CI 0.61–1.14, p = 0.256) (Table 3, model 2), and similar findings were observed after adjusting variables that significantly affect serum 25(OH)D concentration including age, gender, smoking, BMI, diabetes mellitus and eGFR (adjusted OR: 0.93, 95% CI 0.69–1.24, p = 0.599) (Table 3, model 3). Subgroup analyses of the association between serum 25(OH)D < 20 ng/mL and subclinical atherosclerosis are presented in Table 4. In multivariate regression model, the results showed no significant association between 25(OH)D < 20 ng/mL and subclinical atherosclerosis in different subgroups, including age ≥ 65 years old, male gender, current smoker, other co-morbidities such as hypertension, diabetes, dyslipidemia, metabolic syndrome, and higher BMI.

Table 3 Univariate and multivariate logistic regression analysis for vitamin D3 (VitD3) deficiency and the risk of subclinical atherosclerosis (defined as carotid intima-media thickness, cIMT > 0.8 mm).
Full size table
Table 4 Subgroup analysis of the association of serum 25(OH)D deficiency and subclinical arthrosclerosis (carotid intima-media thickness, cIMT > 0.8 mm).
Full size table

Discussion

Main finding

In this large observational cohort study composed of 1798 Asians older than 50 years , prevalence of vitamin D deficiency was 33.6% (n = 604). After adjusting the traditional variables associated with vitamin D deficiency and further subgroup analyses, the results showed that vitamin D deficiency was not an independent risk factor for subclinical atherosclerosis. Significant association observed in the univariate analysis may be confounded by gender and age difference or comorbidities.

Current evidence of vitamin D deficiency and atherosclerosis

Vitamin D deficiency and cIMT have been studied in healthy populations and various specific groups, but inconsistent results were found, even within populations. In a prior study composed of patients with type 1 diabetes, deficiency of circulating 25(OH)D (< 20 ng/mL) was associated with a 3.3-fold increased odds ratio of arterial calcification than serum 25(OH)D sufficient group (≥ 30 ng/mL) after adjusting age, gender, and hours of daylight exposure23. However, contrary outcome was found in another cohort study of 1193 subjects with type 1 diabetes mellitus, which suggested that lower serum 25(OH)D level was related to lower mean coronary arterial calcium24. In those with type 1 diabetes mellitus with mean age around 30–35 years old, previous studies have reported no significant association between the circulating serum 25(OH)D concentration and cIMT24,25. In a cross-sectional cohort study of 654 relatively healthy participants without history of coronary heart disease, serum 25(OH)D concentration was significantly inversely related to the internal cIMT but not the common cIMT26 and similar finding was found in another study composed of HIV-infected patients27.

Inconsistent results were also found in different meta-analysis. The enrolled studies in meta-analysis had significant heterogeneity although vitamin D deficiency (serum 25(OH)D < 20 ng/mL) was found to be associated with a higher cIMT; besides, in female patients with a higher BMI, waist circumference, or eGFR, vitamin D deficiency was associated with lower cIMT28,29. These findings indicate that the relationship between vitamin D deficiency and subclinical atherosclerosis is not consistent. In the present study, the association of serum 25(OH)D and subclinical atherosclerosis could vary by age, gender, and underlying diseases. After the covariates were adjusted, serum 25(OH)D deficiency was not an independent risk factor of subclinical atherosclerosis although positive correlation between age and serum 25(OH)D existed.

Relationship between serum 25(OH)D and atherosclerosis: contributor or bystander?

Active form of Vitamin D—[1,25(OH)D] may influence arterial stiffness directly via regulation of nuclear vitamin D receptor (VDR) or indirect pathway that 1,25(OH)D may relate to blood pressure controlled by suppression of Renin–Angiotensin–Aldosterone System (RAAS), in VDR-knockout mice30,31. The relationship between vitamin D deficiency and subclinical atherosclerosis may be confounded by patient’s age and physical characteristics. A nationwide study in Taiwan revealed that subjects with higher vitamin D level were younger; had lower education status, BMI, and vegetable intake; and higher level of work-related physical activity and fish intake32. The vitamin D deficiency group in our cohort was significantly younger, lesser comorbidities with hypertension, diabetes, dyslipidemia, antihypertensive drug usage, and lipid-lowering agent usage; but it had higher prevalence of subjects with metabolic syndrome, and increase in BMI, HOMA-IR and total body fat. Thus, association between vitamin D deficiency and cIMT became insignificant after adjusting age and gender in our study.

Study limitations

There are several limitations of the present study that must be acknowledged. First, due to the cross-sectional study design, causal relationship between vitamin D and atherosclerosis could not be fully assessed. In addition, some confounding factors, such as duration of sun exposure, vitamin D supplementation, betel-quid chewing were not taken into account in the present study. Third, this is a cross-section cohort study and was lack of CV outcomes of long-term follow-up like longitudinal study; thus, we could not investigate the impact of insufficient serum 25(OH)D level on risk of CVD.

Conclusions

Prevalence of vitamin D deficiency (defined by serum level of 25(OH)D) was approximately one-third in the elderly Asian population in this study. Our findings suggested that vitamin D deficiency is not an independent predictor of subclinical atherosclerosis in elderly Asians after adjusting confounding factors. Future studies to address the association between serum 25(OH)D and the cIMT need to be explored carefully.

Methods

Study population

The present study used the sample data from the I-Lan Longitudinal Aging Study (ILAS), which is a research cohort of community-dwelling adults aged more than 50 years old who were randomly recruited through household registration records33. The survey enrolled 1839 community-dwelling older adults from August 2011 to August 2013. The participants were excluded from this study if any of the following conditions were met: (1) unable to cooperate or communicate with study investigators; (2) declined to grant consent; (3) currently institutionalized34; (4) had any known active disease, such as active cancer, sepsis, heart failure, chronic obstructive pulmonary disease, or functional dependence; (5) had a life-expectancy of less than 6 months; and (6) planned to leave I-Lan county. This study was conducted by Institutional Review Board of National Yang-Ming University approval. All participants provided written informed consent. All methods were performed in accordance with the relevant guidelines and regulations.

Anthropometry and demographic measurements

The demographic and medical details of the participants were collected by the research nurse, who also performed the anthropometric measurements of the patients, including weight, height, BMI, and waist circumference. Brachial blood pressure was measured with a mercury sphygmomanometer after the subjects had rested for at least 15 min. Basic medical histories, including underlying diseases, medication, and whether or not the current smokers (including no smoking, current smoker or quitted smoking), were obtained from personal interviews and recorded medical notes. Metabolic syndrome was defined according to the criteria proposed by Taiwan’s Ministry of Health and Welfare, with more than three of the following risk determinants: (1) waist circumference > 90 cm for men or > 80 cm for women; (2) systolic blood pressure ≥ 130 mmHg, diastolic blood pressure ≥ 85 mmHg, or taking antihypertensive agents; (3) high-density lipoprotein (HDL) < 40 mg/dL for men or < 50 mg/dL for women34; (4) triglyceride (TG) ≥ 150 mg/dL; (5) fasting blood glucose (FBG) ≥ 100 mg/dL or taking antihyperglycemic agents35,36. The body composition, including the total fat mass and the bone mineral density (BMD), was measured by whole-body dual-energy X-ray absorptiometry. The total body fat percentage was calculated as the total fat mass divided by the total body mass times 10037.

Laboratory examinations

After fasting for at least 10 h, peripheral blood samples were collected at 7–9 AM to determine the concentrations of hemoglobin A1c, FBG, total cholesterol, HDL, low-density lipoprotein (LDL), TG, uric acid, and high-sensitivity C-reactive protein (hsCRP) by using an automatic analyzer (ADVIA 1800; Siemens, Malvern, PA, USA). Insulin resistance was quantified with the homeostatic model assessment (HOMA-IR) using the following formula: HOMA-IR = glucose (mmol/L) × insulin (μU/mL)/22.538.

Serum concentration of 25(OH)D were determined by using the commercially available LIAISON 25 OH vitamin D total assay from DiaSorin (DiaSorin vit D; DiaSorin Inc., Northwestern Avenue, MN, USA). The assay range was 4.0–150 ng/mL. The intra- and interassay coefficients were 2.9–5.5% and 6.3–7.9%, respectively. Intact parathyroid hormone (iPTH) were measured by chemiluminescence immunoassays (ADVIA Centaur, Siemens, USA)39. According to the recommendation of the US Preventive Services Task Force, vitamin D deficiency was defined as a VitD3 concentration less than 20 ng/mL40.

Assessment of carotid intima-media thickness (cIMT) and subclinical atherosclerosis

The cIMT was measured using a high-resolution, broad-width, linear array transducer (LOGIQ 400 PRO; GE, Cleveland, OH, USA) at the level of the common carotid artery. All examinations were carried out by the same trained technician, who measured the arteries including the proximal to distal parts of the bilateral common carotid arteries on longitudinal views33. The mean cIMT was defined as the average of the right-side and left-side cIMT values. Subclinical atherosclerosis was defined as a mean cIMT > 0.81 mm; this value has been demonstrated to be associated with an increased risk of stroke and coronary events41,42.

Statistical analysis

All continuous descriptive variables are expressed as the median with interquartile range due to the non-normal distribution, and categorical variables are expressed as numbers (percentages). Comparison between two groups was analyzed by the Mann Whitney u test for continuous variables, and Fisher’s exact test for categorical variables. Pearson’s correlation test was used to assess the correlations between serum 25(OH)D and cIMT with other variables. Logistic regression analysis was performed to assess the relationships between the 25(OH)D concentration and subclinical atherosclerosis. Variables significantly correlated to the cIMT were further entered into multivariate regression analysis. To investigate the association of 25(OH)D on subclinical atherosclerosis modified by different comorbidities, we performed subgroup analyses stratified by age, gender, smoking status, hypertension, diabetes, dyslipidemia, metabolic syndrome, and BMI. Statistical analyses were performed using SPSS version 22.0 (SPSS Inc., Chicago, IL, USA). P values < 0.05 were considered to be statistically significant.