Potential pathophysiological role for the vitamin D deficiency in essential hypertension

Abstract

Vitamin D deficiency has been indicated as a pandemic emerging public health problem. In addition to the well-known role on calcium-phosphorus homeostasis in the bone, vitamin D-mediated processes have been recently investigated on other diseases, such as infections, cancer and cardiovascular diseases. Recently, both the discovery of paracrine actions of vitamin D (recognized as “local vitamin D system”) and the link of vitamin D with renin-angiotensin-aldosterone system and the fibroblast growth factor 23/klotho pathways highlighted its active cardiovascular activity. Focusing on hypertension, this review summarizes the more recent experimental evidence involving the vitamin D system and deficiency in the cardiovascular pathophysiology. In particular, we updated the vascular synthesis/catabolism of vitamin D and its complex interactions between the various endocrine networks involved in the regulation of blood pressure in humans. On the other hand, the conflicting results emerged from the comparison between observational and interventional studies emphasize the fragmentary nature of our knowledge in the field of vitamin D and hypertension, strongly suggesting the need of further researches in this field.

Core tip: This review provides a comprehensive and critical analysis of the most recent studies investigating the relationship between vitamin D and essential hypertension. From the both observational and interventional studies, conflicting results have been shown. This review article provides some hypothesis to explain these discrepancies. In addition to the potential bias related to the study design, some pathophysiological explanation was suggested, especially involving the potential role of local vitamin D system as well as the fibroblast growth factor 23/klotho axis. This review aims at suggesting a careful reflection so that future studies might be designed for minimize bias and encompass the complex biology of vitamin D system.

INTRODUCTION

Vitamin D deficiency has recently emerged as a public health problem, affecting almost 50% of the population worldwide[1]. In addition to the reduced exposition to sunlight[2], also genetic and environmental factors have been suggested as a cause of this pandemic, such as pollution, diet, sedentary life style and stress[3]. Moreover, vitamin D is no longer considered as only a pivotal mediator of calcium metabolism and skeletal health, but it also regulates several cell functions, including differentiation and metabolism. This aspect may explain the reason why hypovitaminosis D has been proved to be an independent risk factor for overall mortality in various cohort analyses[4], whereas vitamin D supplementation significantly reduced mortality[5]. Moreover, similar data were collected from different clusters of inflammatory and chronic diseases, such as infections[6], autoimmunity[7], neurodegenerative pathologies[8], as well for cancer[9]. However, a special interest was conferred to the potential relationship between vitamin D and cardiovascular (CV) disorders. Although in human cohorts low vitamin D levels were associated with impaired CV outcomes[10], a causal relationship remains unknown, and the general enthusiasm about the benefits of vitamin D supplementation have been recently replaced by words of caution.

On the other hand, novel topics that might address many question in the field of vitamin D, such as fibroblast growth factor (FGF) 23-klotho axis, non-genomic effects of vitamin D and the paracrine effects of vitamin D (also called “local vitamin D system”) have been identified. In the following paragraphs, we will focus on the mechanisms triggered by vitamin D in arterial hypertension, starting from the complex interplay with the renin-angiotensin-aldosterone system (RAAS) in both basic research and clinical trials.

VITAMIN D SYSTEM AND BLOOD PRESSURE

Vitamin D

In humans, more than 80% of vitamin D requirements is produced through the ultraviolet-B (UVB)-induced conversion of 7-dehydrocholesterol to vitamin D in the skin, whereas only 10%-20% is absorbed with the diet[1]. The photosynthesis of vitamin D evolved over 750 million years ago, first in the phytoplankton and then in early plants and animals[11]. From an evolutionary stand point it is interesting to note that the first living beings synthetizing vitamin D were missing calcific skeleton. This suggests that a new recognized non-metabolic role (called “non-classical effects”) of vitamin D might actually be the oldest. Regardless of the source, vitamin D requires liver hydroxylation [through 25-hydroxylase (CYP2R1 or CYP27A1)] to form 25-hydroxyvitamin D [25(OH) vitamin D or calcidiol], inactive form but used as reference for vitamin D status, because abundant, stable and easier to quantify[1]. In the kidney 25(OH) vitamin D is then hydroxylated to 1,25-dihydroxyvitamin D [1,25(OH)2 vitamin D or calcitriol] the active form of vitamin D [through 25-hydroxyvitamin D-1α-hydroxylase (CYP27B1)]. This latter step is a pivotal effector of calcium homeostasis and thus highly controlled by the up-regulation of parathyroid hormone (PTH) and the suppression of FGF23/klotho axis[12]. Although the exact contribution of extra-renal hydroxylation in determining the circulating levels of 1,25(OH)2 vitamin D is still unknown, it has been recognized also an extra-renal activity of CYP27B1. Finally, the recent identification of a role of vitamin D binding proteins on vitamin D catabolism has further increased complexity of the system[13].

Vitamin D receptor

Vitamin D receptor (VDR) is member of nuclear hormone receptors superfamily. Following binding with 1,25(OH)2 vitamin D, VDR recruits one of the retinoid X receptors (RXR α, β or γ) forming homo- or heterodimers to promote a specific, high-affinity DNA-binding interaction. This transcriptional complex binds to repeated sequences of 6 hexamers [vitamin D response elements (VDRE)] in the promoter region of target gene[1]. VDR is believed to directly or indirectly regulate 3% to 5% of human genome and the different genomic activation of vitamin D in the different cell types involves allosteric influences, VDRE location and epigenetic modification (of both DNA and histones)[14]. In addition, VDR recognizes extra-nuclear ligands including endogenous steroids and other lipophilic compounds[15,16]. Finally, VDR may be expressed also on the cell surface membrane and within mitochondria thus might modulate non-genomic signalling pathways, such as 1,25(OH)2 vitamin D-mediated rapid-response[17]. Vitamin D are deeply involved in several patterns of CV pathophysiology, including vascular inflammation[18] and endothelial dysfunction[19] as observed in patients with chronic kidney disease (CKD)[20] and type 2 diabetes[21] as well as in asymptomatic subjects[22]. For instance, in vitro VDR activation induces nitric oxide production in endothelial cells[23] and improves the angiogenic properties of endothelial progenitor cells[24], while regulates proliferation[25], migration[26], mineralization[27] and thrombogenic protein expression[28] in vascular smooth muscle cells (VSMCs). The recent recognition of specific VDR polymorphisms and genetic susceptibility in pathophysiology of hypertension has further supported these insights[29].

Vitamin D hydroxylases

The gene encoding for CYP27B1 is widespread expressed in various tissue of endodermal, ectodermal and mesenchymal origin. Since even VDR is highly represented in tissues, an autocrine/paracrine vitamin D system has been strongly suggested. In contrast to endocrine vitamin D system, local regulation of 1,25(OH)2 vitamin D levels is independent of PTH expression, but rather relies on environmental factors[30]. CYP27B1 expression in endothelial cell is regulated by pro-inflammatory cytokines[31], in VSMCs is under estrogenic control[32] whereas many signals regulate the expression in monocyte/macrophage, including toll-like receptor[33], interferon-γ[34], FGF23[35] and uremia[36]. Accordingly, CYP27B1^-/- mice develop an hypertensive phenotype, also characterized by increased circulating level of renin, angiotensin (Ang) II and aldosterone, then suppressed by administration of 1,25(OH)2 vitamin D independently of serum levels of calcium or phosphorus[37].

Vitamin D and FGF23/Klotho pathways

Recently, the discovery of FGF23 has extended the complexity of the endocrine network involving the vitamin D system. As vitamin D counter-regulatory hormone, FGF23 suppresses renal synthesis of 1,25(OH)2 vitamin D by inhibiting CYP27B1 and up-regulating CYP24A1. These effects are independent of VDR but require co-factor klotho, essential for FGF23 signal transduction[38].

Overall, 1,25(OH)2 vitamin D and FGF23 are involved in a classical hormonal loop also including PTH. High levels of 1,25(OH)2 vitamin D raise the serum concentrations of both calcium and phosphate. Concomitantly, the feedback by PTH reduces only calcium levels by enhancing its urinary excretion. Increased levels of FGF suppress the expression of sodium-phosphate cotransporter NaPi-2a on renal proximal tubules, thus resulting in increased phosphaturia[39]. Therefore, phosphorus homeostasis might be maintained by 1,25(OH)2 vitamin D via a direct regulation on FGF23 levels.

Thus, the discovery of FGF23 might explain some paradoxical concerns on vitamin D, especially among the ambiguous results of interventional studies. A strong correlation between an increased risk of mortality and high circulating levels of both FGF23 and phosphate has been also reported[40,41], suggesting that there is a threshold in vitamin D supplementation beyond which 1,25(OH)2 vitamin D may have detrimental effects.

For instance, the age-associated suppression of Klotho expression[42] may promote a vitamin D toxicosis during therapeutic supplementation characterized by over-hyperphosphatemia and thus increased cardiovascular risk[43]. Although it is likely a failure of the normal feedback mechanism regulating vitamin D and FGF23, the molecular bases of these clinical features have not been identified yet. Furthermore, Camalier et al[44] recently provided evidence of both rapid and late effects induced by FGF23 on mesenchymal stromal cells, involving cell proliferation and extracellular matrix (ECM) regulation. In addition, Jimbo et al[45] showed that FGF23 promoted osteoblastic differentiation of aortic VSMCs from uremic rats by inducing ERK1/2 phosphorylation pathway. However, it should be noted that these features were shown only in primary rat VSMCs and other studies failed to recognize the relevance of FGF23-Klotho signalling in mouse arteries[46,47].

Ultimately, although further studies in humans are warranted, we agree with Glade M.J., who suggested that there may be an age at which vitamin D deficiency may become life-sustaining, not life-threatening[48].

PATHOPHYSIOLOGICAL PATHWAYS OF VITAMIN D IN HYPERTENSION

Although the effects of vitamin D on blood pressure have been known for several decades, some physiological aspects on the modulation of vascular cells and the vascular tone still remain to be clarified.

RAAS

RAAS plays a pivotal role in maintaining sodium and blood volume homeostasis even by modulating the renal function and blood pressure. RAAS up-regulation was shown to promote the development of hypertension and increased CV risk[49,50].

Salt- and volume-independent RAAS up-regulation (documented by an increase in renin and Ang II levels) was associated with hypertension and cardiac hypertrophy in VDR^-/- mice[51]. Similarly, in wild-type mice, 1,25(OH)2 vitamin D inhibition (through dietary intake of strontium) increased renin expression, while 1,25(OH)2 vitamin D supplementation down-regulated RAAS in a VDR-dependent manner[51].

Also the evidence of a preserved CV function in VDR^-/- mice undergoing RAAS inhibition (using Angiotensin converting enzyme inhibitors or Angiotensin receptor I blockers) confirmed a direct connection between RAAS and vitamin D system[52]. Interestingly, similar results were also reported in CYP27B1^-/- mice[37]. Among the several cross-sectional and prospective studies investigating the association of vitamin D deficiency and hypertension only Forman et al[53] provided a mechanistic role of vitamin D system in the RAAS regulation. Lower 25(OH) vitamin D levels correlated with both higher Ang II at baseline (P = 0.03), and blunted renal plasma flow response to Ang II infusion in a cohort of 184 normotensive subjects treated with high-salt diet. These findings were confirmed in subsequent studies[54,55].

From a molecular point of view, the research group directed by Li et al[52] discovered a direct effect of 1,25(OH)2 vitamin D on renin gene transcription. They identified that vitamin D is capable of suppressing renin gene transcription by a cAMP response element, identified on the promoter region of Ren-1c gene[56]. In addition, the same authors confirmed a central role of active vitamin D by excluding the control of PTH or serum calcium levels on renin expression[57]. On the other hand, Ferder and co-workers have recently proposed a new hypothesis about the dependency instead of complementarity vitamin D system and RAAS. Overturning the classical view, the authors suggested the RAAS-induced inflammatory response as regulator of vitamin D status thus representing the “primum movens” of current vitamin D deficiency pandemic[58]. Anyway, although suggestive, this hypothesis of a reciprocal counter-regulatory effect between vitamin D and RAAS is currently highly speculative. Research models identifying effectors shared by RAAS and vitamin D are still missing[59]. Angiotensin II is a main mediator responsible for adverse vascular remodelling in hypertension[60]. By promoting endothelial dysfunction and vascular permeability, RAAS induces recruitment and activation of inflammatory cells within the vessel wall. This inflammatory behaviour stimulates hyperplasia and hypertrophy of VSMCs, but also their release of pro-inflammatory molecules (VCAM-1, monocyte chemoattractant protein-1, interleukin 6 and 8)[61]. Furthermore, angiotensin II was shown to mediate the shift of VSMCs toward a fibroblast phenotype that alters the ECM composition by suppressing the activity of matrix metalloproteinases and enhancing the production of their inhibitors[62]. Among the intracellular signalling pathways involved in angiotensin II signalling a key role is played by oxidants and their downstream signalling cascades including mitogen-activated protein kinase, protein kinase C, phospholipase A2 and the transcription factors NFκB and activator protein-1[63].

PTH

PTH is a crucial regulator of calcium and phosphate homeostasis, achieved in different ways, such as osteoclast/osteoblast activation, enhancement of intestinal and renal calcium absorption and up-regulation of CYP27B1 expression in the kidney. Although not generally accepted[64], higher PTH concentrations were associated with an increase in several CV risk factors[65], including hypertension[66-76]. Moreover, several cohorts of sporadic primitive hyperparathyroidism were found associated with arterial stiffness[77-84]. The mechanism linking PTH and blood pressure is still unclear and several pathways might be triggered. PTH up-regulates RAAS activity promoting renin release[85,86], but it also directly promotes aldosterone release from adrenal glands[87]. Also the increase of serum calcium PTH may indirectly modulate renin release[88] and aldosterone synthesis[87] in addition to activate VSMC[89]. PTH increases sympathetic activity with additional RAAS activation (increase in renin release and aldosterone secretion)[90] and vascular contractility[90]. Finally, a cellular interaction through the PTH/PTH-related protein receptor expressed on endothelial cells[91], VSMC[92] and inflammatory cells[93] may directly affect the vascular function.

CLINICAL STUDIES

The association between vitamin D levels and blood pressure was previously reported, observing higher blood pressure trends in the winter months and location further from the equator[94]. Many clinical studies have subsequently provided consistent results but this topic is still widely debated, especially after the results observed in the interventional clinical trials.

Cross-sectional studies

A large number of cross-sectional studies investigated the relationship between vitamin D deficiency and blood pressure, as well as the prevalence of hypertension. Table 1 summarizes the studies having 25(OH) vitamin D as reference for the vitamin D status[1].

Table 1.

Cross-sectional studies evaluating vitamin D blood pressure

Ref.	Year	Study design (sample size)	Country (ethnicity) Age	Correlation (lower reference range of 25(OH) vitamin D)	Findings
Snijder et al[67]	2007	Cross-sectional from the LASA (1205 subjects more than 65 yr old)	Netherlands (caucasian) men and women ≥ 65 yr	No (I quartile: < 25 nmol/L)	25(OH) vitamin D was not associated with systolic or diastolic BP or prevalence of hypertension. Instead, PTH correlated with both BP and hypertension incidence
Martins et al[95]	2007	Cross-sectional from the 1988-1994 NHANES (15088 subjects)	United States (Caucasian and African Americans and other) men and women age stratified	Yes (I quartile: < 52.5 nmol/L)	Adjusted inter-quartile analysis showed an increased prevalence of hypertension in in the lower quartile of 25(OH) vitamin D (OR = 1.30, 95%CI; 1.13-1.49; P < 0.05)
Scragg et al[96]	2007	Cross-sectional from the 1988-1994 NHANES (12644 subjects not treated with anti-hypertensive drugs)	United States (Caucasian and African Americans and other) men and women age stratified	Yes (I quintile: < 40 nmol/L)	Adjusted inter-quintile analysis of 25(OH) vitamin D showed significant inverse correlation with both systolic (P < 0.01) and diastolic (P < 0.05) BP. This association was stronger in more than 50 years old and black people
Judd et al[97]	2008	Cross-sectional from the 1988-1992 NHANES (7699 non-hypertensive subjects)	United States (White and black people ) men and women age stratified	Yes (Vitamin D deficiency defined as < 50 nmol/L)	Lower 25(OH) vitamin D concentrations were associated with a higher blood pressure category in white people (P < 0.01) but after adjustment for age the association was no longer significant
Hintzpeter et al[98]	2008	Cross-sectional from GNHIES (4030 adults)	Germany (Caucasian) men and women 18-79 yr	Yes (Vitamin D deficiency defined as < 12 nmol/L[99])	According to 25(OH) vitamin D levels, in multivariate analysis there was a relationship between 25(OH) vitamin D and hypertension bot in men (OR = 0.97, 95%CI: 0.94-0.99; P < 0.05) and in women (OR 0.96, 95%CI: 0.93-0.99; P < 0.05)
Hypponen et al[100]	2008	Cross-sectional from 1958 British birth cohort (6810 subjects)	United Kingdom (Caucasian) men and women 45-47 yr	Yes (I tertile: < 45 nmol/L)	The lower 25(OH) vitamin D tertile was associated with hypertension (OR 0.72, 95%CI: 0.61-0.86; P < 0.01)
Reis et al[101]	2009	Cross-sectional from the 2001-2004 NHANES (3577 non-pregnant adolescents without diagnosed diabetes)	United States (Caucasian and African Americans and other) male and female adolescent 12-19 yr	Yes (I quartile: < 37.5 nmol/L)	25(OH) vitamin D was inversely associated with systolic BP (P < 0.05) also in the adjusted odds ratio for the interquartile comparison (OR = 2.36, 95%CI: 1.33-4.19; P < 0.05)
Pasco et al[102]	2009	Cross-sectional (861 subjects)	Australia (Caucasian) women: 20-92 yr	Yes (I tertile 25(OH)D: < 30 nmol/L)	In this cohort there was a significant inter-tertile difference in mean BP (P < 0.001) as well as in anti-hypertensive medication use (P < 0.01)
Almirall et al[103]	2010	Cross-sectional (237 subjects more than 64 years old)	Spain (Caucasian) men and women 64-93 yr	Yes (cut-off for vitamin D deficiency: < 62.5 nmol/L)	A significant negative association was observed between serum 25(OH) vitamin D levels and both systolic (P < 0.05) and diastolic BP (P < 0.05) also in multivariate analysis
Jorde et al[104]	2010	Cross-sectional from the Tromsø Study (4125 subjects not treated with anti-hypertensive drugs)	Norway (Caucasian) Men and women age stratified	Yes (I quartile: < 41.4 nmol/L)	At adjusted inter-quartile analysis serum 25(OH) vitamin D was inversely correlated with systolic BP (P < 0.01)
Kim et al[105]	2010	Cross-sectional (1330 subjects)	South Korea (Asian)	Yes (I quintile: < 29.7 nmol/L)	At adjusted inter-quintile analysis, both systolic and diastolic BP decreased linearly with increasing of 25(OH) vitamin D (quintile 1-5; P for trend < 0.01). Moreover, inter-quintile comparison of BP had OR of 0.42 (95%CI: 0.24-0.73; P < 0.05)
Zhao et al[106]	2010	Cross-sectional from the 2003-2006 NHANES (5414 subjects not assuming anti-hypertensive drugs)	Men and women < 40 yr United States (Hispanic, Caucasian and African Americans) men and women ≥ 20 yr	Yes (I quintile: < 37.5 nmol/L	Across 25(OH) vitamin D quintiles systolic and diastolic BP decreased linearly and inversely (P < 0.01). Moreover, the prevalence ratio for hypertension was lower in the highest quintile (OR = 0.82, 95%CI: 0.73-0.91; P < 0.05)
Fraser et al[107]	2010	Cross-sectional from the 2001-2006 NHANES (3958 subjects)	United States (Caucasian and African Americans and other) men and women ≥ 20 yr	Yes (linear correlation)	25(OH) vitamin D has an inverse linear correlation with systolic blood pressure in various adjusted models (P < 0.05)
Steinvil et al[108]	2011	Cross-sectional case-control study (34874 subjects of which 8387 hypertensive)	Israel men and women 38-72 yr	Yes (vitamin D deficiency defined as < 37.5 nmol/L)	The age-adjusted OR for hypertension among normal and deficient serum 25(OH) vitamin D was 1.19 (95%CI: 1.09-1.31; P < 0.01) in women, whereas in men there was not statistical difference
Burgaz et al[109]	2011	Cross-sectional from the ULSAM (833 adult men)	Sweden (Caucasian) Men 71 yr	Yes (vitamin D deficiency defined as < 37.5 nmol/L)	Adjusted logistic regression confirmed the association between 25(OH) vitamin D concentration < 37.5 nmol/L and hypertension (OR = 3.3, 95%CI: 1.0-11.0; P < 0.05)
Bhandari et al[110]	2011	Cross-sectional (2722 subjects of which 1415 hypertensive)	United States (Caucasian and African Americans and other) men and women mean age 58.5 yr	Yes (I quartile: < 37.5 nmol/L)	The prevalence rate of hypertension was inversely correlated with serum 25(OH) vitamin D. Inter-quartile comparison showed an adjusted OR of 2.70 (95%CI: 1.41-5.19; P < 0.05)
Pacifico et al[111]	2011	Cross-sectional case-control study (452 children and adolescent of which 304 over-weight/obese and 148 normal weight)	Italy (Caucasian) Male and female children	Yes (I tertile of 1,25(OH)2 vitamin D: < 42.5 nmol/L)	1,25(OH)2 vitamin D was inversely correlated with systolic BP both in the whole population (P < 0.01) and over-weight (P < 0.01) population as well as in control group (P < 0.01). Regardless of model for adjusted analysis, the OR for hypertension among tertile categories had a P value < 0.05.
Williams et al[112]	2011	Cross-sectional from 2003-2006 NHANES (5617 adolescent)	United States (Caucasian and African Americans and other) male and female children 12-19 yr	Yes (linear correlation)	In this cohort, 25(OH) vitamin D showed a linear inverse association with systolic BP in multivariate analysis (P < 0.01).
Forrest et al[113]	2011	Cross-sectional from 2005-2006 NHANES (4495 adults subjects of which 1482 hypertensive)	United States (Caucasian and African Americans and other) Men and women age stratified	Yes (vitamin D deficiency defined < 50 nmol/L[114])	Vitamin D deficiency independently correlated with prevalence of hypertension (P < 0.01).
He et al[70]	2011	Cross-sectional from 2003-2006 NHANES (7561 of which 1849 treated with anti-hypertensive drugs)	United States (Caucasian and African Americans and other) Men and women age stratified	No (I quintile: < 33 nmol/L)	25(OH) vitamin D was inversely associated with systolic BP. However, 25(OH) vitamin D lost its statistical significance in a multivariate analysis including PTH. Instead, PTH maintained a strong correlation with BP in multivariate analysis regardless of covariates.
Dorjgochoo et al[115]	2012	Cross-sectional study from two, population-based, prospective cohort studies (1460 subjects of which 547 hypertensive)	China (Asian) men and women 40-74 yr	Yes (lowers range defined by I quintile 23.5 nmol/L and cut-offs of 37.5 nmol/L[116] and 27.5 nmol/L[117])	Among men cohort, BP was inversely and significantly correlated with 25(OH) vitamin D (P < 0.05). Moreover, prevalence of hypertension was inversely associated with non-deficient status of vitamin D (adjusted OR = 0.29, 95%CI: 0.10-0.82; P < 0.05)
Sakamoto et al[118]	2013	Cross-sectional from the AHS-2 (568 subjects)	United States (equally matched Caucasian and African Americans) men and women 30-95 yr	Yes (vitamin D deficiency defined as < 50 nmol/L)	Regardless of adjusted analysis models, Caucasian people showed a linear inverse correlation between 25(OH) vitamin D and BP (P < 0.05). Also the comparison between vitamin D deficient and non-deficient showed statistical difference (P < 0.05).
Li et al[64]	2012	Cross-sectional (1420 subjects of which 487 hypertensive)	China (Asian) Men and women ≥ 65 yr	No (I quartile: < 42 nmol/L)	Serum 25(OH) vitamin D levels were not associated with risk of hypertension in single and multiple regression models. Similarly, PTH is not independently associated with BP or risk of hypertension
Caro et al[119]	2012	Cross-sectional (219 subjects of which 115 hypertensive)	Puerto Rico (Hispanic) Men and women 21-50 yr	No (cut-off used to define non optimal: 75 nmol/L)	Vitamin D status was not found to be associated with BP
Chan et al[72]	2012	Cross-sectional (939 men aged 65 yr and older)	China (Asian) men ≥ 65 yr (age strafied)	No (I quartile: < 63 nmol/L)	Vitamin D status was not found to be associated with BP. Instead, PTH was directly and independently associated with BP also in multivariate analysis.
Parikh et al[120]	2012	Cross-sectional (701 adolescents)	United States (Caucasian and African Americans) Male and female 14-18 yr	Yes (I tertile: < 54.8 nmol/L)	Serum 25(OH) vitamin D has a linear inverse correlation with both systolic (P < 0.05) and diastolic (P < 0.01) BP. However, in the adjusted analysis only the relationship with systolic BP remained significant.
Sabanayagam et al[121]	2012	Cross-sectional from NHANES III (9215 subjects of which 3712 with pre-hypertension)	United States (Caucasian and African Americans and other) men and women age stratified	Yes (I quartile: < 44.25 nmol/L)	In this cohort the systolic BP are inversely correlated with the vitamin D status (P < 0.05) and lower values of 25(OH) vitamin D were associated with increase prevalence of pre-hypertension (adjusted OR = 1.48, 95%CI: 1.16-1.90; P value for trend < 0.01).
van Ballegooijen et al[122]	2012	Cross-sectional from the Hoorn study (256 subjects)	The Netherlands (Caucasian) men and women 50-75 yr	Yes (I quartile: < 60.8 nmol/L)	In this cohort there was an inverse correlation between 25(OH) vitamin D and both systolic and diastolic BP (P value for trend < 0.01 for both)
Skaaby et al[123]	2012	Cross-sectional 4330 subjects)	Denmark (Caucasian) men and women 30-60 yr	No (I quartile: < 33 nmol/L)	Mean 25(OH) vitamin D levels did not differed between hypertensive and normotensive subjects. There was not increased prevalence of hypertension in vitamin D deficient subjects
Kruger et al[124]	2013	Cross-sectional form the PURE study (291 African women)	All over the world countries (African) women > 47 yr	Yes (vitamin D deficiency defined as < 30 nmol/L[125])	Both systolic and diastolic BP correlated linearly and inversely with serum 25(OH) vitamin D level (P < 0.05 for both). However, only systolic BP maintain statistical significance in multivariate analysis (P < 0.05).
Mateus-Hamdan et al[73]	2013	Cross-sectional (284 geriatric patients of which 106 hypertensive)	France men and women mean age 85 ± 6 yr	No (linear correlation)	Means PTH but not 25(OH) vitamin D levels are significant different in hypertensive compared to normotensive patients.
Ke et al[126]	2013	Cross-sectional from the ATBC (2271 subjects of which 1430 hypertensive)	Finland (Caucasian) men and women 50-69 yr	Yes (I quartile: < 25 nmol/L)	Serum 25(OH) vitamin D level has a significant and inverse association with systolic BP (P < 0.05), also if stratified in groups. Moreover, the lower group was associated with increased prevalence of hypertension in multivariate analysis (P value for trend < 0.05).

LASA: Longitudinal aging study amsterdam; 25(OH)D: Cholecalciferol; BP: Blood pressure; PTH: Parathyroid hormone; NHANES III: Third United States national health and nutrition examination survey; OR: Odds ratio; GNHIES: German national health Interview and examination survey; ULSAM: Uppsala Longitudinal study of adult men; AHS-2: Adventist health study-2; PURE study: Prospective urban and rural epidemiology study; ATBC: Alpha-tocopherol and beta-carotene study.

The most relevant results were acquired from the national health and nutrition examination survey (NHANES), widely representative of non-hospitalized United States civilian population. First Martins et al[95] showed an increased prevalence of hypertension associated with low serum 25(OH) vitamin D levels in 15088 subjects from this cohort. In addition, the very large sample size of this cohort allowed to recognize the inverse relationship between 25(OH) vitamin D and raised blood pressure also in several subgroups (such as African Americans and older people[96,97], children/adolescents[101,112], Hispanic people[113], in addition to observed an increased prevalence of pre-hypertension in 25(OH) vitamin D deficient subjects[121]). Other cross-sectional cohort studies with large sample size supporting these findings were the German National Health Interview and Examination Survey (4030 subjects)[98], the 1958 British birth cohort (6810 subjects)[100], and the Tromsø Study (4125 subjects)[104] as well as the cohorts collected from Israel people (34874 subjects)[108] and Copenhagen population[123]. Other smaller cohorts supporting these insights were collected in Europe[73,103,109,111,122,126], North America[110,118,120,124], Oceania[102] and Asia[72,105,115]. Despite the large numbers of subjects and their worldwide distribution, a clear relationship between vitamin D and blood pressure has not yet been established so far. In fact, among the studies listed in Table 1, seven did not confirm this association[64,67,70,72,73,119,123]. These conflicting results are in accordance with some unanswered questions in the field of vitamin D biology. In fact, despite the standardization of the season of subject recruitment, the latitudes, where studies were carried out, determine a confounding effect related to the pivotal role of sunlight exposure and consequent vitamin D synthesis within the skin[2]. Another potential bias is that differences in serum 25(OH) vitamin D levels might depend on the age. Elderly subjects have a reduced skin synthesis and intestinal absorption of vitamin D in addition to spend less time outdoors, limiting sunlight exposure[127]. Regardless of the latitude and season, only few studies have estimated sun exposure and dietary intake (as well as a possible supplementation) of vitamin D, especially in the elderly population. Moreover, racial differences should be recognized, since the black population correlated with a higher incidence of vitamin D deficiency (and also hypertension), because of their high skin content of melanin[128]. In this regard, it should be emphasized that most of the negative studies were made up from Caucasian[67,73,123] Hispanic[119] and Chinese[64,72] cohorts. Finally, there is still much debate about which cut-off value defines 25(OH) vitamin D deficiency. However, among the results reported in Table 1 most of the studies showed the first quartile or proposed a cut-off closed to 30 nmol/L. In addition, for higher mean 25(OH) vitamin D levels, blood pressure poorly correlated with vitamin D but rather with PTH levels[67,70].

Longitudinal studies

Few studies have investigated the incidence of hypertension in vitamin D-deficient subjects. In addition, no study among them had this aim as a primary outcome, suggesting some potential limitation in the statistical power estimation. In addition, the majority of the cohorts investigated was limited to the Caucasian race and female gender, further limiting the generalizability of the results. However, we believe that the main limitation is represented by the lack of prospective risk evaluation in the elderly. In fact, even if the follow-up is extended over 65 years, this overlap does not recognize the critical alterations in vitamin D metabolism during aging. Taking those important limitations into the account, Table 2 summarizes the most important longitudinal observational studies, starting from the results of health professionals’ follow-up study (HPFS) and the nurse health study (NHS)2.

Table 2.

Longitudinal studies addressing the association between vitamin D and blood pressure

Ref.	Year	Study design and follow-up (sample size)	Country (ethnicity) Age	Correlation (lower reference range of 25(OH) vitamin D)	Findings
Forman et al[129]	2007	Prospective observational nested case-control study from HPFS and NHS-2 4 yr (1811 subjects)	United States (Caucasian) men 47-82 yr women 43-68 yr	Yes (vitamin D deficiency defined as < 37.5 nmol/L[130])	Multivariate RR of incident hypertension among vitamin D deficient subject was 3.18 (95%CI: 1.39-7.29; P < 0.05)
Forouhi et al[131]	2008	Prospective observational from the Ely study 10 yr (534 subject)	United Kingdom (Caucasian) men and women 40-69 yr	No (vitamin D deficiency defined as < 25 nmol/L)	There were not significant changes in BP during the follow-up
Forman et al[132]	2008	Prospective observational nested case-control study from the NHS 2 7 yr (1484 normotensive women)	United States (Caucasian) women: 32-52 yr	Yes (I quartile: < 21 nmol/L)	Median 25(OH) vitamin D were lower in women developing hypertension (P < 0.01). Moreover, interquartile analysis showed significant and inverse correlation between 25(OH) vitamin D and hypertension (OR = 1.66, 95%CI: 1.11-2.48; P value for trend < 0.05)
Jorde et al[104]	2010	Prospective observational from the Tromsø Study 14 yr (4125 subjects not treated with anti-hypertensive drugs)	Norway (Caucasian) men and women 25-84 yr	No (I quartile: < 41.4 nmol/L)	At adjusted analysis, 25(OH) vitamin D did not predict future hypertension or increase in BP: Moreover there was not any association between change in serum 25(OH) vitamin D and BP
Anderson et al[133]	2010	Prospective observational average 1.3 yr (maximum 9.1 yr) (41497 subjects)	United States men and women 34-76 yr	Yes (vitamin D deficiency defined as < 37.5 nmol/L)	Lower 25(OH) vitamin D levels were associated with higher incidence of hypertension (HR = 1.62, 95%CI: 1.48-2.02; P < 0.01)
Griffin et al[134]	2011	Prospective observational from MBHMS 14 yr (559 women)	United States (Caucasian) women 24-44 yr	Yes (vitamin D deficiency defined as < 80 nmol/L)	25(OH) vitamin D insufficiency has an increased risk of systolic hypertension at multivariate analysis (OR = 3.0, 95%CI: 1.01-8.7; P < 0.05)
Margolis et al[135]	2012	Prospective observational from the WHI 7 yr (4863 post-menopausal women)	United States (Caucasian, African, Hispanic, Asian and others) women 50-79 yr	No (I quartile: < 34.4 nmol/L)	There was not significant linear or nonlinear trend in the risk of incident hypertension
Wang et al[136]	2012	Prospective observational form PHS 15.3 yr (1211 normotensive men)	United States men 40-84 yr	Yes (I quartile: < 39.9 nmol/L)	There was significant difference only between I and III quartile (HR = 0.69, 95%CI: 0.50-0.96; P < 0.05)
Skaaby et al[123]	2012	Prospective observational 5 yr (4330 subjects)	Denmark (Caucasian) men and women 30-61 yr	No (I quartile: < 33 nmol/L)	Multivariate logistic regression analyses did not show any association between 25(OH) vitamin D incidence rate of hypertension.
Ke et al[126]	2013	Prospective observational from the ATBC 4 yr (2271 subjects of which 1430 hypertensive)	Finland (Caucasian) men and women 50-69 yr	No (I quartile: < 25 nmol/L)	25(OH) vitamin D did not predict future hypertension.

HPFS: Health professionals’ follow-up study; NHS 2: Nurse heatlh study 2; 25(OH)D: Cholecalciferol; RR: Relative risk; BP: Blood pressure; OR: Odds ratio; HR: Hazard ratio; MBHMS: Michigan bone health and metabolism study; WHI: Women’s health initiative; PHS: Physicians’ health study; ATBC: Alpha-tocopherol and beta-carotene study cohort.

Forman et al[129] firstly reported an increased risk of incident hypertension in 1811 subjects selected from these two matched cohorts at 4-year follow-up (pooled RR = 3.18; 95%CI: 1.39-7.29, P < 0.05). In addition, the investigators extended this risk prediction, as a surrogate, to the overall study population including 38388 man from HPFS (adjusted RR = 2.31; 95%CI: 2.03-2.63, P < 0.05) and 77531 women from the NHS 2 (adjusted RR = 1.57; 95%CI: 1.44-1.72, P < 0.05). Afterwards, the same authors also designed a prospective nested case-control study including 1484 normotensive women from the NHS 2 that confirmed the previous results (inter-quartile OR = 1.66; 95%CI: 1.11-2.48, P value for trend = 0.01)[132]. Also the Intermountain Heart Collaborative Study Group provided similar results prospectively analyzing a large electronic medical database of a general healthcare population. In addition to recognize a wide prevalence of vitamin D deficiency, very low levels of 25(OH) vitamin D were directly associated with an increased risk of developing CV disease, including hypertension (HR = 1.62; 95%CI: 1.48-2.02, P < 0.01)[133]. Significant association between vitamin D deficiency and incidence of hypertension was also observed in a smaller subgroup analysis from both woman cohort of Michigan Bone Health and Metabolism Study (OR = 3.0; 95%CI: 1.01-8.7, P < 0.05)[134] and for male population of Physicians’ Health Study (HR = 0.69; 95%CI: 0.50-0.96, P < 0.05)[136]. On the other hand, other large sample size studies such as subgroup analyses from Ely study[131], Tromsø study (burdened with a 40% dropout rate)[104], Women’s Health Initiative[135] and Alpha-Tocopherol and Beta-Carotene study cohort[126], as well as cohort of general Copenhagen population[123] did not confirm any association between vitamin D levels and incidence of hypertension.

Randomized clinical trials

Table 3 summarizes randomized interventional clinical trials investigating the link between vitamin D and blood pressure.

Table 3.

Randomized clinical trial investigating the protective effect of vitamin D supplementation on blood pressure

Ref.	Year	Study design	Country (ethnicity) Age	Intervention	Findings
Lint et al[137]	1988	(sample size) Prospective randomized double-blind placebo-controlled trial (65 men with glucose intolerance of which 26 hypertensive)	Sweden (Caucasian) 61-65 yr	(follow-up) α-calcidol 0.75 μg (12 wk)	In hypertensive patients supplementation has addictive effect to concomitant antihypertensive therapy in reducing BP (P < 0.01). In the whole population there was only non-significant trend in BP lowering
Pan et al[138]	1993	Prospective randomized double-blind 2 × 2 interventional trial (58 institutionalized elderly persons)	Taiwan (Asian) not provided	calcium 800 mg/d or 1,25(OH)2 vitamin D 5 μg/d or calcium 800 mg/d + 1,25(OH)2 vitamin D 5 μg/d, or placebo (11 wk)	Any type of supplementation failed to reduce BP
Scragg et al[139]	1995	Prospective randomized double-blind placebo-controlled trial (189 elderly subjects)	United Kingdom (not provided) 63-76 yr	25(OH) vitamin D 2.5 μg/d or placebo (5 wk)	Although treatment was effective in increasing serum 1,25(OH)2 vitamin D (P < 0.01) and decreasing PTH (P < 0.01), there was not difference in BP change
Krause et al[140]	1998	Prospective randomized double-blind controlled trial (18 patients with untreated mild essential hypertension)	Germany (Caucasian) 26-66 yr	Full body UVB or UVA thrice weekly (6 wk)	In accordance with a 162% rise in plasmatic 25(OH) vitamin D (P < 0.01) and 15% fall in serum PTH (P < 0.01), the UVB group showed also a reduction in 24-h ambulatory systolic and diastolic BP (P < 0.01)
Pfeifer et al[141]	2001	Prospective randomized double-blind controlled trial (148 elderly subject with 25(OH)D < 50 nmol/L)	Germany (Caucasian) 70-86 yr	Calcium 600 mg × 2/d or calcium 600 mg + 25(OH) vitamin D 10 μg twice daily (8 wk)	In accordance with a 72% rise in plasmatic 25(OH) vitamin D (P < 0.01) and 17% fall in serum PTH (P < 0.05), combined supplementation significantly reduced systolic BP (P < 0.05)
Sudgen et al[142]	2008	Prospective randomized double-blind placebo-controlled trial (34 elderly type 2 diabetic patients with 25(OH)D < 50 nmol/L)	United Kingdom (not provided)	Loading dose ergocalciferol 2500 μg or placebo (8 wk)	Supplementation significantly rise plasmatic 25(OH) vitamin D (P < 0.01) and reduced systolic BP, whereas there was only a trend in diastolic BP decrease
Alborzi et al[143]	2008	Prospective randomized double-blind placebo-controlled trial (24 elderly type 2 diabetic patients with 25(OH)D < 50 nmol/L)	mean 64 years United States (Caucasian and African Americans) 56-80 yr	Paricalcitol 1 or 2 μg/d or placebo (4 wk)	Any dose of paricalcitol failed to reduce BP
Margolis et al[144]	2008	Prospective randomized double-blind controlled trial (36282 n post-menopausal women from WHI study)	United States (Caucasian, Asian, Hispanic, African American) 50-79 yr	Calcium 500 mg × 2/d or calcium 500 mg + 25(OH) vitamin D 5 μg twice daily (7 yr)	There was no significant difference in over time change of BP in the whole population. In addition, supplementation failed to reduce the risk of developing hypertension in non-hypertensive patients at baseline
Nagpal et al[145]	2008	Prospective randomized double-blind placebo-controlled trial (71 older overweight men )	India (Indian population) 36-54 yr	25(OH) vitamin D 3000 μg every 2 wk for 3 times or placebo (7 wk)	Supplementation failed to reduce BP
Daly et al[146]	2009	Prospective randomized double-blind controlled trial (124 community-dwelling men)	Australia (Caucasian) 55-69 yr	Milk fortified with calcium (500 mg) and 25(OH) vitamin D (10 μg) twice a day or standard milk (2 yr)	Supplementation failed to reduce BP
Hilpert et al[147]	2009	Prospective randomized double-blind controlled trial (23 hypertensive adults)	United States (not provided)	Dairy-rich, high fruits and vegetables diet or a high fruits and vegetables diet or an average Western diet (5 wk)	High fruits and vegetables diet dairy-rich or not significantly reduced BP (P < 0.05). Moreover, in dairy-rich, high fruits and vegetables diet there was a greater lowering of intracellular calcium (P < 0.01), strongly associated with fall in diastolic BP (P < 0.05)
Witham et al[148]	2010	Prospective randomized double-blind placebo-controlled trial (56 patients with history of stroke and baseline 25(OH)D < 75 nmol/L)	United Kingdom (not provided) 53-79 yr	Loading dose ergocalciferol 2500 μg or placebo (8 and 16 wk)	Supplementation significantly increased serum 25(OH) vitamin D to both controls (P < 0.01). However, treatment failed to reduced BP
Witham et al[149]	2010	Prospective randomized double-blind placebo-controlled trial (61 patients with type 2 diabetes and baseline 25(OH)D < 100 nmol/L)	United Kingdom (not provided) 55-76 yr	Loading dose ergocalciferol 2500 μg or 5000 μg or placebo (8 and 16 wk)	Supplementation significantly increased serum 25(OH) vitamin D to both controls (P < 0.01 for both). However, supplementation failed to reduced BP
Judd et al[150]	2010	Prospective randomized double-blind controlled trial (9 patients with baseline 25(OH)D within 25 and 75 nmol/L in addition to systolic BP between 130 and 150 mmHg)	United States (African American) mean 45 yr	loading dose ergocalciferol 2500 μg or placebo weekly for 3 wk or 25 (OH) vitamin D 0.5 μg twice a day for 1 wk (3 wk)	Only supplementation with 25(OH) vitamin D decrease by 9% mean systolic BP (P < 0.01) in accordance with rise of serum 25(OH) vitamin D (P < 0.05)
Scragg et al[151]	2011	Prospective randomized double-blind controlled trial (119 patients with baseline 25(OH)D < 50 nmol)	New Zealand (Pacific islander, Caucasian and Maori) 23-87 yr	24 whole body exposures of either UVB or ultraviolet A (6 and 12 wk)	In the UVB arm there was a significant increase in serum 25 (OH) vitamin D after both 6 and 12 wk (P < 0.01 for both). However, treatment failed to reduced BP
Salehpour et al[152]	2012	Prospective randomized double-blind placebo-controlled trial (77 pre-menopausal overweight and obese women)	Iran (Arabian) 30-46 yr	25 (OH) vitamin D 25 μg daily or placebo (12 wk)	Supplementation significantly rise plasmatic 25 (OH) vitamin D (P < 0.01) and fall PTH (P < 0.01). Moreover, although treatment improved lipid profile, there was no effect on BP
Gepner et al[153]	2012	Prospective randomized double-blind placebo-controlled trial (110 post-menopausal women with baseline 25(OH)D within 10 and 60 nmol/L)	United States (not provided) 60-67 yr	25 (OH) vitamin D 62.5 μg daily or placebo (16 wk)	Supplementation, although significantly raised serum 25(OH) vitamin D (P < 0.01), failed in improving BP control assessed by changes in FMD, PWV and Aix
Wood et al[154]	2012	Prospective randomized double-blind placebo-controlled trial (305 healthy post-menopausal women)	United Kingdom (not provided) 48-72 yr	25 (OH) vitamin D 10 μg or 25 μg/d or placebo (1 yr)	Supplementation failed in improving CV risk profile, including BP control
Larsen et al[155]	2012	Prospective randomized double-blind placebo-controlled trial (112 hypertensive patients)	Denmark (Caucasian) 48-72 yr	25 (OH) vitamin D 75 μg/d or placebo (20 wk)	Supplementation significantly rise plasmatic 25 (OH) vitamin D (P < 0.01) and fall PTH (P < 0.01) but failed in improving BP control. However, in a post-hoc subgroup analysis of patient with 25 (OH) vitamin D deficiency at baseline supplementation significantly decrease 24-h systolic and diastolic BP (P < 0.05)
Zhu et al[156]	2013	Prospective randomized double-blind placebo-controlled trial (43 healthy subjects)	China (Asian) 20-22 yr	Calcium 600 mg + 25 (OH) vitamin D 3.12 μg daily or placebo, in addition to 500 kcal/d of caloric deficit (7 yr)	Except a reduction in visceral fat mass, supplementation failed in improving CV risk profile, including BP control
Forman et al[157]	2013	Prospective randomized double-blind placebo-controlled trial (283 healthy black subjects)	United States (African American) mean 51 yr	25 (OH) vitamin D 25 μg or 50 or 100 μg/d or placebo (12 and 24 wk)	Supplementation significantly decrease BP consistent with increasing dose (P < 0.05). Moreover, there was linear correlation between systolic BP decrease and rise of serum 25 (OH) vitamin D (P < 0.05)
Witham et al[158]	2013	Prospective randomized double-blind placebo-controlled trial (159 with isolate systolic hypertension)	United States (not provided) mean 77 yr	Loading dose 25 (OH) vitamin D 2500 μg or placebo (12, 24 and 36 wk)	Supplementation significantly rise plasmatic 25 (OH) vitamin D (P < 0.01) but failed in improving BP control. Moreover, treatment failed to achieve secondary outcomes including 24-h blood pressure, arterial stiffness and endothelial function

α-calcidol: Synthetic analog of 1,25(OH)2D; BP: Blood pressure; 1,25(OH)2D: Calcitriol; UVB: 94.5% UVA and 3.5% UVB; UVA: 99.5% UVA and 0.05% UVB; 25(OH)D: Cholecalciferol; PTH: Parathyroid hormone; WHI: Women’s Health Initiative Calcium/vitamin D trial; HyD: 25(OH)D metabolite with hydrophilic properties and much shorter half-life; FMD: Brachial artery flow-mediated vasodilation; PWV: Carotid-femoral pulse wave velocity; Aix: Aortic augmentation index; CV: Cardiovascular.

Although most of the studies reported a significant serum 25(OH) vitamin D increase after supplementation, they are impeded by several limitations, mostly related to study design issues. The first one consists in the limited number of trials investigating blood pressure as a primary outcome. In addition, only few studies focused on vitamin D-deficient cohorts, more suitable for investigating the effectiveness of a supplementation with vitamin D. In this regard, a subgroup analysis of vitamin D-deficient subjects, from a sample of 112 Danish hypertensive patients randomized to high-dose 25(OH) vitamin D supplementation (75 μg/d) versus placebo, showed a significant decrease of 24-h systolic and diastolic blood pressure values (P < 0.05)[155]. These findings confirmed previous results from other small sample size cohorts of vitamin D-deficient patients[141,142,150]. For this reason, the recently results by Forman et al[157] from the largest published cohort of hypertensive patients (n = 283) randomized to vitamin D supplementation versus placebo appear of particular interest. The oral administration of 25(OH) vitamin D (25 to 100 μg/d) significantly decreased the blood pressure levels. Unfortunately, these studies present additional limitations, such as taking into account the different approaches used for vitamin D supplementation. Although sunlight exposition might be the more physiological way, the ultraviolet (UV)-B rays-induced skin synthesis of vitamin D is hard to quantify and thus poorly investigated[140,151]. Oral supplementation has been preferred because easier to manage (despite some variability in intestinal absorption may exist) if provided through diet regimen[147], nutritional supplements[146] or direct vitamin D administration (daily intake[137-139,141,143,144,152-157] or loading dose[142,148-150,158]). Finally, it should be reported the failure of Women’s Health Initiative study to prove changes in blood pressure in a very large sample size of post-menopausal women (n = 36282) randomized to receive calcium versus calcium plus 25(OH)D over 7-year follow-up[144].

Meta-analyses of clinical studies

Five meta-analyses were recently performed to quantify the prospective associations of vitamin D status with the risk of hypertension. Pittas et al[159] included the results of four observational longitudinal cohorts with 32181 subjects with a follow-up of 7 to 10 years. The pooled analysis showed an increased risk of developing hypertension in vitamin D-deficient subjects (RR = 1.76; 95%CI: 1.27-2.44, P < 0.05). Conversely, another meta-analysis of ten randomized clinical trials failed to prove the effectiveness of vitamin D supplementation in promoting blood pressure decrease[159]. Therefore, this mismatch between observational studies and randomized interventional clinical trials is retrieved in other meta-analyses The lack of relationship in interventional studies was reported by Witham et al[160] and Wu et al[161], while pooled analysis of observational studies showed a strong association between vitamin D status and blood pressure[162]. In particular, the meta-analysis of observational longitudinal studies by Kunutsor et al[163] recently reported that subjects in the higher tertiles of vitamin D levels have a 30% lower risk of developing hypertension as compared to those in the bottom tertiles (pooled RR = 0.70; 95%CI: 0.58-0.86, P < 0.05).

OPEN ISSUES AND PERSPECTIVES

Many questions recently emerged from efficacy and safety in interventional trials using vitamin D supplementation. In experimental mouse models, excessive intake of vitamin D induces vascular and soft-tissue calcifications. Thus, in human beings, caution has to be used on the pro-calcifying effects of exogenous vitamin D. In addition to derangement in calcium homeostasis, it should take into account the detrimental effects of vitamin D-induced phosphate overload involving also FGF23/klotho axis. On the other hand, the definition of the optimal vitamin D status from a CV point of view remains matter of debate and general consensus is still missing. “Bone health-driven” recommendations agree to define insufficient a 25(OH) vitamin D levels < 20 ng/mL, suggesting a target of 30 ng/mL. Similarly, reports from large cohorts (such as NHANES[164] and The Framingham offspring study[165]) showed a linear inverse association with CV outcome for 25(OH) vitamin D levels up to 30ng/mL. Considering hypertension, the results from the Vitamin D and Omega-3 Hypertension Trial (VITAL Hypertension) that is still enrolling patients[166] might clarify this point. Finally, the “local vitamin D system” is emerging as a pivotal topic that might explain the conflicting results between observational and interventional trials[167].

CONCLUSION

Neither the European society of Cardiology nor American Heart Association have published CV-focused algorithms regarding vitamin D deficiency and this is because the first results from randomized clinical trials have provided more questions than answers. Certainly, several factors involved in vitamin D biology are under-recognized or hard to assess, including physical activity, sunlight exposure, health status or dietary habits. Moreover, several confounding factors have not been considered in several studies, such as comorbidities, concomitant medications or differences in gender, age and race. In addition, also vitamin D compounds proposed were highly variable, ranging from native (cholecalciferol or ergocalciferol) or synthetic (α-calcidol) inactive vitamin D to active vitamin D (calcitriol) up to selective VDR activators (paricalcitol). However, it is likely that other unidentified factors are also involved in vitamin D biology, such as the possible relationship with other endocrine networks, emphasizing the need of pre-clinical studies.