Vitamin D is recognised as the vitamin of sunshine(Reference Holick1). A relationship between vitamin D and CVD has been proposed by observing more incidence of CVD in winter compared with summer in many countries(Reference Holick2, Reference Dunnigan, Harland and Fyfe3), which might be attributable to the protective effect of vitamin D on CVD(Reference Bull and Morton4). Ecological studies have indicated the higher rate of CHD and high blood pressure increment with more distance from the equator, which is also related to low sun exposure and higher prevalence of vitamin D deficiency(Reference Scragg5, Reference Grimes, Hindle and Dyer6). It has been shown that a range of markers including geographic latitude, altitude and season affect vitamin D status which are negatively correlated with cardiovascular morbidity and mortality. Evidence suggests that low levels of vitamin D may contribute to the development of CVD(Reference Rostand7).
On the other hand, a range of CVD risk factors including dyslipoproteinaemia, high blood pressure, reduced glucose tolerance, diabetes and increased inflammatory markers are related to obesity, which has been increasing dramatically during recent decades(Reference Zittermann, Schleithoff and Koerfer8–Reference Bray and Bellanger10). Serum 25-hydroxyvitamin D (25(OH)D) levels negatively correlate with BMI(Reference Sarti and Gallagher11). It is likely that alteration in vitamin D homeostasis affects CVD development in obese individuals(Reference Parikh, Edelman and Uwaifo12). Low levels of 25(OH)D cause higher levels of parathyroid hormone (PTH), which is a known non-traditional CVD risk factor(Reference Wilson, Kannel and Silbershatz13, Reference Krause, Buhring and Hopfenmuller14).
Investigations have shown that hypovitaminosis D has an undesirable effect on total and LDL-cholesterol concentrations(Reference Merke, Hofmann and Goldschmidt15), therefore resulting in high total cholesterol and low apoA-I concentrations. An independent and positive significant correlation has been reported among serum 25(OH)D levels and apoA-I and HDL-cholesterol concentrations. It seems that the effects of vitamin D on the lipid profile are independent of Ca and other unfavourable effects caused by high type 2 diabetes risk(Reference Merke, Hofmann and Goldschmidt15, Reference Merke, Milde and Lewicka16).
Scragg et al. (Reference Scragg, Sowers and Bell17) and Jorde et al. (Reference Jorde, Figenschau and Emaus18) have shown the association between 25(OH)D and blood pressure. Vitamin D affects blood pressure via paracrine and endocrine mechanisms in tissues that are specifically related to high blood pressure including vascular smooth muscle cells(Reference Holick19), the endothelium and cardiomyocytes(Reference Li, Kong and Wei20). Also, 1,25-hydroxyvitamin D is considered as a negative endocrine regulator of the renin–angiotensin system, and sunlight exposure as an indirect marker of vitamin D synthesis in the skin has an inverse correlation with the prevalence of high blood pressure, reducing blood pressure levels(Reference Li, Kong and Wei20–Reference Sambrook, Chen and March22).
As the prevalence of vitamin D deficiency is high in Iranian women(Reference Hashemipour, Larijani and Adibi23, Reference Heshmat, Mohammad and Majdzadeh24) and the burden of overweight and obesity is the fifth leading risk for global deaths(25), which increases the risk of non-communicable diseases including CVD, we evaluated the effect of vitamin D3 supplementation on cardiovascular risk factors in healthy overweight and obese women.
Materials and methods
The present study was performed between November 2009 and April 2010 in the Heart and Vascular Laboratory at the Pharmacology Department of Tehran University of Medical Sciences, Tehran, Iran. We distributed advertisements in the campus and requested all women to participate in the study. The inclusion criteria were as follows: good public health status; age 18–50 years; BMI ≥ 25 kg/m2; free of known osteoporosis; gastrointestinal disease; diabetes mellitus; CVD; renal disease; high blood pressure (>160/90 mmHg). Participants were excluded from the study if they were following a weight-reduction programme, taking weight-loss drugs, having a change in weight more than 3 kg during the last 3 months, pregnant, lactating, smoking, drinking alcohol, taking nutritional supplements, cholesterol and TAG-lowering agents as well as anti-hypertensive agents. At the beginning of the study, sunscreen use was discontinued in all subjects.
A total of eighty-five participants who met the above inclusion criteria were recruited into a double-blind clinical trial. We randomly assigned participants to one of two groups: vitamin D (n 42) or placebo (n 43). Written informed consent was obtained from all subjects. The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients were approved by the Ethics Committee of the Tehran University of Medical Sciences and Iranian Registry of Clinical Trial (registration no. IRCT138809092709N2). This trial was registered at ClinicalTrials.gov as NCT01344161. Of the eighty-five participants who were eligible for the present intervention study, eight subjects dropped out of the follow-up: in the vitamin D group, one patient became pregnant and one patient followed a weight-loss programme. In the placebo group, one patient used oral contraceptive pills and five other patients were unwilling to continue the 12-week examination for personal reasons. Finally, seventy-seven participants completed the study and, of these, thirty-nine were in the vitamin D group and thirty-eight were in the placebo group.
A vitamin D3 supplement (25 μg/d as cholecalciferol; Merck Pharma GmbH) or placebo (25 μg/d as lactose; Merck Pharma GmbH) was given to the participants per month. The period of the intervention was 90 d and the participants took one tablet of vitamin D3 supplement or placebo every day. Compliance with the supplementation was 87 %.
Blood samples were collected from the antecubital vein of each patient after an overnight fast of at least 12 h. After centrifugation for 20 min (3000 g), the serum samples were frozen simultaneously and stored at − 80°C until analysed. In order to eliminate the probable effects of sex hormones on blood lipids, blood sampling was not performed between days 1 and 5 of the menstrual cycle.
Body weight and height were measured on a digital scale (model 763; Seca GmbH & Co, KG) with participants wearing light indoor clothing. Waist and hip circumference were measured using an Ergonomic Circumference Measuring Tape (model 201; Seca GmbH & Co, KG). BMI was calculated by dividing weight (kg) by height (m2). Body composition was assessed by Bioelectrical Impedance Analysis (model 4000, Body Stat Quad Scan; Bodystat). All anthropometric indices were measured by following the WHO standard procedures(26). We assessed physical activity levels using the International Physical Activity Questionnaire(Reference Craig, Marshall and Sjostrom27) and determined the average of metabolic equivalents (MET)-min/week.
Energy, macronutrient, Ca, vitamin D, saturated fat, MUFA and PUFA intakes were estimated using 24 h food recall and validated FFQ(Reference Esfahani, Asghari and Mirmiran28). On a monthly basis, a nutritionist completed the questionnaires by a direct interview. Because the Iranian food composition table is incomplete (limited to only raw materials and a few nutrients)(Reference Azar and Sarkisian29), each food and beverage, only unfortified US food equivalents, was analysed for nutrient intake using Nutritionist IV software (version 4.1; First Databank Division, The Hearst Corporation) to assess macronutrient and micronutrient contents of the foods. The Iranian food composition table was used as an alternative for traditional Iranian food items, such as kashk, which is not included in the Food Composition Tables for Use (United States Department of Agriculture food composition table)(30).
Systolic and diastolic blood pressure was measured in the right or the left arm supported at the heart level of seated participants (Model Gamma G-7; Heine). Means of two measurements taken at 5 and 10 min of rest were used.
25(OH)D was measured using an enzyme immunoassay (Immunodiagnostic Systems Limited). Intra- and inter-assay CV for 25(OH)D were 6·9 and 8·1 %, respectively. Intact PTH was measured using an immunoenzymometric assay (Immunodiagnostic Systems Limited). Intra- and inter-assay CV for intact PTH were 5·5 and 8·3 %, respectively. TAG, total cholesterol, LDL-cholesterol and HDL-cholesterol were measured by a direct colorimetric enzymatic method (Greiner). Intra- and inter-assay CV for TAG were 2·6 and 2·9 %, for total cholesterol were 1·4 and 2·1 %, for LDL-cholesterol were 2·2 and 2·5 %, and for HDL-cholesterol were 2·4 and 2·6 %, respectively. Lipoprotein(a) was measured by an enzyme immunoassay (Mercodia). Intra- and inter-assay CV for lipoprotein(a) were 6·6 and 7·8 %, respectively. Also, we measured apoA-I and apoB-100 using a sandwich technique (ELISA; AlerCHEK). Intra- and inter-assay CV for apoA-I and apoB-100 were 5·6, 6·9 % and 6·1, 7·4 %, respectively.
Continuous variables are expressed as means and standard deviations, and categorical variables as percentage proportions. The change in variables was analysed with ANCOVA to adjust mean differences in variables, in which baseline value, fat mass and waist circumference were used as covariates. P values < 0·05 were considered statistically significant. Relationships between change in 25(OH)D concentrations and cardiovascular markers were evaluated using simple Pearson's correlations. To adjust the analyses for confounding variables, partial correlations were computed for all dependent variables, after adjusting for fat mass and waist circumference as independent variables. The statistical program SPSS (version 16; SPSS, Inc.) was used to perform the analyses.
Participant characteristics are shown in Table 1. Baseline physical characteristics in both groups were similar. The intake of SFA and MUFA decreased in both groups from the beginning of the study up to week 12. The intake of PUFA increased in the vitamin D group and decreased in the placebo group. Despite this, changes in dietary fatty acid intake were not significant between the two groups.
MET, metabolic equivalents; 25(OH)D, 25-hydroxyvitamin D; PTH, parathyroid hormone; LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol.
* Mean values were not significantly different from week 0 between the groups (P>0·05; unpaired t test).
† Mean values were not significantly different from month 0 between the groups (P>0·05; unpaired t test).
‡ To convert 25(OH)D values to ng/ml, divide by 2·5. To convert PTH values to pg/ml, divide by 0·11.
Baseline systolic blood pressure was lower in the intervention group (67·9 (sd 10·1) v. 71·9 (sd 9·1) mmHg). Systolic blood pressure increased in the vitamin D group, but declined in the placebo group (0·51 (sd 12·7) v. − 2·2 (sd 10·2) mmHg; Table 2). Diastolic blood pressure increased in both groups (2·3 (sd 6·7) v. 0·13 (sd 8·3) mmHg). The changes were not statistically significant between the two groups.
MET, metabolic equivalents; PTH, parathyroid hormone; LDL-C, LDL-cholesterol; HDL-C, HDL-cholesterol.
* After 12 weeks.
† An ANCOVA was used to adjust mean differences in all dependent variables.
‡ Mean values remained significantly different between the groups after adjusting for fat mass and waist circumference (P < 0·05).
At baseline, the percentage of subjects with hypovitaminosis D (25(OH)D < 75 nmol/l) and vitamin D sufficiency (25(OH)D ≥ 75 nmol/l) in the vitamin D group was 89·7 and 10·3 %, reaching 53·8 and 46·2 % after supplementation, respectively. Related values in the placebo group were 89·5 and 10·5 % at baseline, reaching 86·8 and 13·2 %, respectively (P < 0·001). Serum intact PTH concentrations declined in the vitamin D group up to 1·2 (sd 0·5) pmol/l. However, serum PTH concentrations increased in the placebo group up to 1·7 (sd 0·8) pmol/l during 12 weeks (P < 0·001). Body fat mass decreased in the vitamin D and placebo groups ( − 2·7 (sd 2) v. − 0·4 (sd 2) kg; P < 0·001, respectively).
Serum TAG concentrations decreased in both groups after 12 weeks, but these alterations were higher in the placebo group compared with in the vitamin D group (0·26 (sd 0·28) v. 0·29 (sd 0·3) mmol/l, respectively). This result was not statistically significant. Baseline serum total cholesterol concentrations in the vitamin D group were lower than those in the placebo group (4·7 (sd 0·6) v. 4·89 (sd 0·8) mmol/l, respectively). After 12 weeks, total cholesterol concentrations increased in the vitamin D group and declined in the placebo group (0·08 (sd 0·56) v. − 0·47 (sd 0·58) mmol/l; P < 0·001), respectively. Similarly, serum LDL-cholesterol concentrations increased in the vitamin D group and decreased in the placebo group (0·13 (sd 0·5) v. − 0·3 (sd 0·5) mmol/l; P < 0·001), respectively. HDL-cholesterol concentrations increased in the vitamin D group and declined in the placebo group (0·07 (sd 0·2) v. − 0·03 (sd 0·2) mmol/l; P = 0·037), respectively. At baseline, serum apoA-I concentrations in the vitamin D group were lower than those in the placebo group (1·63 (sd 0·2) v. 1·78 (sd 0·2) g/l; P = 0·006), respectively. After the intervention, apoA-I concentration increased in the vitamin D group, though it decreased in the placebo group (0·04 (sd 0·39) v. − 0·25 (sd 0·2) g/l; P < 0·001). Serum apoB-100 concentrations declined in both groups, but the changes were not statistically significant. The LDL-cholesterol:apoB-100 ratio increased in the vitamin D group, which indicates less atherogenic properties of LDL-cholesterol particles, whereas this ratio declined in the placebo group indicating that LDL-cholesterol particles were smaller and had higher density (0·11 (sd 0·6) v. − 0·19 (sd 0·3); P = 0·014). The apoA-I:apoB-100 ratio increased in the vitamin D group, but decreased in the placebo group (0·04 (sd 0·3) v. − 0·17 (sd 0·3); P = 0·009).
A positive correlation was observed between changes in serum 25(OH)D concentrations and systolic blood pressure (r 0·24, P = 0·032), which remained statistically significant after adjusting for fat mass and waist circumference (r 0·23, P = 0·044). A positive and significant trend was seen between changes in serum 25(OH)D concentrations and total cholesterol concentrations (r 0·27, P = 0·014; Fig. 1), which did not remain significant after adjusting for fat mass and waist circumference. There was a positive and significant correlation between changes in serum 25(OH)D concentrations and HDL-cholesterol concentrations (r 0·26, P = 0·022; Fig. 2) and between serum 25(OH)D concentrations and apoA-I concentrations (r 0·25, P = 0·023; Fig. 3). The correlation between changes in serum 25(OH)D concentrations and HDL-cholesterol concentrations was statistically significant even after adjusting for fat mass and waist circumference (r 0·25, P = 0·03), but did not remain significant between changes in serum 25(OH)D concentrations and apoA-I concentrations.
The present study is one of the first reports about the effect of vitamin D3 supplementation solely on blood lipids and lipoproteins in healthy overweight and obese women. The present study has shown that although the daily intake of a 25 μg vitamin D3 supplement increases total and LDL-cholesterol concentrations, it has a beneficial effect on HDL-cholesterol, apoA-I concentrations, apoA-I:apo B-100 and LDL-cholesterol:apoB-100 ratios in overweight and obese women.
In the present study, although vitamin D3 supplementation significantly increased 25(OH)D concentrations, some participants in the vitamin D group did not reach sufficient 25(OH)D concentrations. It seems that they may need higher doses or a longer period of time to be supplemented(Reference Aloia, Patel and Dimaano31). 25(OH)D concentrations are known to be an independent predictor of CVD(Reference Tishkoff, Nibbelink and Holmberg32). 25(OH)D levels < 37·5 nmol/l, compared with levels higher than 75–100 nmol/l, are related to an increase in cardiovascular events. In a previous study, men with 25(OH)D levels ≤ 37·5 nmol/l had a higher risk of myocardial infarction (R 2 2·42), compared with those with 25(OH)D levels ≥ 75 nmol/l(Reference Giovannucci, Liu and Hollis33). Moreover, in a prospective cohort study with an average follow-up duration of 7·7 years in 3285 patients, Dobnig et al. (Reference Dobnig, Pilz and Scharnagl34) reported that hazard ratios (HR) for all-cause mortality (HR 2·08, 95 % CI 1·60, 2·70; HR 1·53, 95 % CI 1·17, 2·01, respectively) and those for cardiovascular mortality causes (HR 2·22, 95 % CI 1·57, 3·13; HR 1·82, 95 % CI 1·29, 2·58, respectively) in the two lowest quartiles of 25(OH)D with medians of 19 and 33·2 nmol/l were higher compared with patients in the highest quartiles of 25(OH)D with a median of 71 nmol/l. In the Framingham Offspring Cohort Study with an average follow-up duration of 5·4 years in individuals with high blood pressure, the risk of cardiovascular events increased two times in participants with 25(OH)D concentrations ≤ 37·5 nmol/l compared with those with 25(OH)D concentrations ≥ 37·5 nmol/l(Reference Wang, Pencina and Booth35). Scragg et al. reported that among white patients without high blood pressure from the Third National Health and Nutrition Examination Survey (NHANES III 1988–1994), systolic blood pressure and pulse pressure values in the highest quartile of 25(OH)D ( ≥ 85·7 nmol/l) were lowered by 1·8 and 1·6 mmHg, respectively, compared with the lowest quartile of 25(OH)D ( ≤ 40·4 nmol/l). Also, systolic blood pressure was decreased up to 1·8 and 4·6 mmHg in patients aged >50 and < 50 years, respectively, when 25(OH)D concentrations increased from 20 to 100 nmol/l(Reference Scragg, Sowers and Bell17). In contrast, in a cross-sectional study, the Longitudinal Aging Study Amsterdam, by Snijder et al. (Reference Snijder, Lips and Seidell36), no correlation has been reported between 25(OH)D concentrations and systolic blood pressure or diastolic blood pressure. We did not find any significant change in blood pressure after 12 weeks. Zittermann et al. (Reference Zittermann, Frisch and Berthold37) demonstrated that after 12 months of supplementation with 83 μg vitamin D/d coupled with a weight-reduction programme in overweight patients, systolic and diastolic blood pressure decreased, but only significant effects of time were observed. Additionally, Major et al. (Reference Major, Alarie and Doré38) and Pfeifer et al. (Reference Pfeifer, Begerow and Minne39) reported similar results. PTH and its related peptides can affect cardiovascular cells via receptors(Reference Schluter and Piper40–Reference Somjen, Weisman and Kohen42). PTH is known as a risk factor for CVD, and patients with hyperparathyroidism are more exposed to cardiovascular morbidity and mortality(Reference Schluter and Piper43, Reference Perkovic, Hewitson and Kelynack44). PTH enhances myocyte hypertrophy(Reference Andersson, Rydberg and Willenheimer45) and vascular remodelling(Reference Garcia de la Torre, Wass and Turner46). Moreover, PTH probably has a pre-inflammatory effect by stimulating cytokine release in vascular smooth muscle cells(Reference Martin-Ventura, Ortego and Esbrit47). In the present study, PTH concentrations decreased significantly in the vitamin D group, but increased in the placebo group.
On the other hand, the correlation between vitamin D and serum TAG concentrations has been reported in patients with the end stage of renal disease(Reference Foley, Parfrey and Sarnaak48). Zittermann et al. (Reference Zittermann, Frisch and Berthold37) and Major et al. (Reference Major, Alarie and Doré38) demonstrated that TAG concentrations reduced in the vitamin D group. It has been suggested that vitamin D can reduce hepatic TAG synthesis or secretion by increasing intestinal Ca absorption(Reference Khajehdehi and Taheri49). By this way, vitamin D increases lipolytic activity of heparin and also increases TAG uptake by peripheral tissues(Reference Lacour, Basile and Drueke50). Furthermore, Ca increment in hepatocytes stimulates microsomal TAG transfer protein, which is correlated with VLDL-cholesterol synthesis and secretion. Vitamin D induces intestinal Ca absorption that results in the suppression of Ca increment in hepatocytes and a decrease in VLDL-cholesterol synthesis and/or secretion(Reference Khajehdehi and Taheri49).
In the present study, supplementation with vitamin D3 increased total and LDL-cholesterol concentrations. Chiu et al. (Reference Chiu, Chu and Go51) reported that 25(OH)D levels had a positive correlation with total and LDL-cholesterol concentrations. Zittermann et al. (Reference Zittermann, Frisch and Berthold37) also found a significant vitamin D-dependent increase in LDL-cholesterol. In contrast, in 170 British Bangladeshis, John et al. (Reference John, Noonan and Mannan52) reported a negative correlation between 25(OH)D concentrations and total and LDL-cholesterol concentrations. It seems that vitamin D through an increase in intestinal Ca absorption(Reference Barger-Lux, Heaney and Lanspa53) reduces the formation of insoluble Ca–fat soaps(Reference Denke, Fox and Schulte54), therefore resulting in low Ca in lumen content, a low rate of Ca bonding to bile acids(Reference Reid55, Reference Vaskonen, Mervaala and Sumuvuori56), conversion of cholesterol to bile acids and, finally, cholesterol excretion(Reference Van der Meer, Welberg and Kuipers57). Several studies have demonstrated that serum 25(OH)D concentrations are a strong, independent predictor of apoA-I concentrations(Reference Chiu, Chu and Go51, Reference Malik58, Reference Auwerx, Bouillon and Kesteloot59). Auwerx et al. (Reference Auwerx, Bouillon and Kesteloot59) reported that apoA-I concentrations had a positive correlation with 25(OH)D levels in Belgian men. It has been postulated that in the promoter of the apoA-I gene, there are vitamin D response elements through which 1,25(OH)2D enhances the transcription of the apoA-I gene in human hepatoma cells(Reference Wehmeier, Beers and Haas60).
In conclusion, daily supplementation with 25 μg vitamin D3 by repletion of the body's vitamin D storage results in significant improvement in some cardiovascular risk factors in overweight and obese women.
The authors thank Sahar Dehghani, Ghazaleh Shimi, Saeedeh Nasiri and Majid Goharinejad for the anthropometric, dietary and specimen processing. Dr Mehdi Hedayati and colleagues are acknowledged for the biochemical analyses. The study was supported financially by the Tehran University of Medical Sciences (grant no. 852) and the Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences (grant no. 293). The authors' contributions are as follows: A. S. and M. R. contributed to the study design, data collection, data interpretation and writing of the manuscript; F. S. and F. H. contributed to the study design, recruitment of patients, data collection, review of the original data and their compilation; M. V. contributed to the study design, critical revision of the manuscript for important intellectual content, and final approval of the manuscript; A. H. contributed to the data analysis and interpretation, and critical revision of the manuscript for important intellectual content; M. G. contributed to the data analysis, manuscript revision, and critical revision of the manuscript for important intellectual content. None of the authors has any conflict of interest to declare.