Vitamin D supplementation, 25-hydroxyvitamin D concentrations, and safety1,2

1. Reinhold Vieth INTRODUCTIONMany arguments favoring higher intakes of calcium and other nutrients have been based on evidence about the diets of prehistoric humans (1). Likewise, the circulating 25-hydroxyvitamin D [25(OH)D; calcidiol] concentrations of early humans were surely far higher than what is now regarded as normal. Humans evolved as naked apes in tropical Africa. The full body surface of our ancestors was exposed to the sun almost daily. In contrast, we modern humans usually cover all except about 5% of our skin surface and it is rare for us to spend time in unshielded sunlight. Our evolution has effectively designed us to live in the presence of far more vitamin D (calciferol) than what most of us get now, yet there is no consensus about what vitamin D intakes are optimal or safe.

See corresponding editorial on page 825.

Unlike anything else used in the fortification of foods, the purpose of vitamin D is to correct for what is an environmental deficit (less ultraviolet exposure) and not to correct for lack due to classical nutritional reasons. With a few exceptions reviewed by Takeuchi et al (2), there is little or no vitamin D in the kind of foods that humans normally eat. Therefore, conclusions about the efficacy and safety of vitamin D must be in the context of the role of environmental factors.Before 1997, the recommended dietary allowance of vitamin D (RDA; 3) for infants and children was 10 μg (400 IU). In essence, the scientific basis for this dose was that it approximated what was in a teaspoon (5 mL) of cod-liver oil and had long been considered safe and effective in preventing rickets (4). The basis for adult vitamin D recommendations has been even more arbitrary. Thirty-six years ago, an expert committee on vitamin D could provide only anecdotal support for what it referred to as “the hypothesis of a small requirement” for vitamin D in adults and it recommended one-half the infant dose, just to ensure that adults obtain some from the diet (5). In England, an adult requirement of only 2.5 μg (100 IU)/d was substantiated on the basis of 7 adult women with severe nutritional osteomalacia whose bones showed a response when given this amount (6). The adult RDA of 5 μg (200 IU)/d was described as a “generous allowance” in the 1989 version of American recommended intakes (3)—but why was this “generous” and in relation to what? It is remarkable that despite the widespread intake of 5 μg (200 IU) vitamin D/d, there is still no published data showing that this dose has any effect on the serum 25(OH)D concentration in adults.The objective way to assess vitamin D nutritional status is through the circulating 25(OH)D concentration (7, 8). Concentrations <20–25 nmol/L indicate severe vitamin D deficiency (9, 10), which will lead to rickets and histologically evident osteomalacia (11). Concentrations between 25 and 40 nmol/L reflect marginal vitamin D deficiency (9, 10, 12), a situation that is common in countries north of the United States, where 40 nmol/L is a typical winter average in adults (13). Marginal concentrations of 25(OH)D are associated with mildly elevated parathyroid hormone (PTH) and diminished 1,25-dihydroxyvitamin D [1,25(OH)2D; calcitriol] concentrations (2, 9, 10, 14).According to Parfitt et al (7), “vitamin D deficiency implies the existence of an anatomic, physiological, or biochemical abnormality that can be corrected by the administration of vitamin D in nonpharmacological doses.” The regression of PTH versus 25(OH)D concentrations in the elderly shows that PTH concentrations become minimal when 25(OH)D concentrations exceed 100 nmol/L (12, 15). Therefore, concentrations of 25(OH)D below this may reflect deficiency in the elderly.

A report by Adams and Lee (16) has become cited as a classic example of vitamin intoxication by individuals taking nutritional supplements. They describe 4 subjects, 1 of whom claimed a vitamin D intake of 30 μg (1200 IU)/d, who presented with hypercalciuria that moderated after vitamin D intake was stopped. The paper was accompanied by an editorial by Marriott (17) of the NIH, which reemphasized concern about vitamin D intoxication. Although moderate vitamin D supplementation may help stave off osteoporosis when it is combined with calcium (18, 19), Marriott recommended care in initiating the approach (17).The issues surrounding vitamin D intake for adults have become confusing for both patients and physicians. For example, a 30 μg (1200-IU)/d intake in the report by Adams and Lee was a possible cause of vitamin D toxicity in one subject (16), although essentially the same intake has been reported in 2 studies to occur in some patients with distinct vitamin D deficiency (9, 20). How can 30 μg (1200 IU) vitamin D be insufficient for some and yet be implicated as mildly toxic for others? I will show that the confusion can be explained by environment because sunshine alone can bring 25(OH)D concentrations to 210 nmol/L in normal people and vitamin D intakes of 30 μg (1200 IU)/d contribute only a negligible fraction of this.A recent review by Hathcock (21) about the efficacy and safety of vitamins and minerals did not address vitamin D. The newly revised daily reference intake (DRI) for vitamin D is 3-fold higher than the previous RDA for people >70 y of age (19), but for most adults the DRI has remained unchanged from the 1960s RDA of 5 μg (200 IU)/d. The purpose of the present review is to assemble and interpret the substantial data now available about how vitamin D supply affects serum 25(OH)D concentrations and to relate these to issues of efficacy and safety of vitamin D.Previous Section Next Section

Commentary on Methods

More than for any other analyte in the endocrine test repertoire, circulating 25(OH)D concentrations vary by geography (latitude), culture, and legislation (food fortification laws). Technical issues also arise when results from different studies are compared because assays of 25(OH)D vary. In the survey reported in 1984 by Jongen et al (22) of 19 laboratories, the mean CV for plasma 25(OH)D was 35%, much of which was attributable to the purification procedures used. Over the past 6 y, an international proficiency survey of >30 laboratories has been available through the North West Thames, External Quality Assurance Survey (Charing Cross Hospital, London). In this ongoing quarterly survey, the one method most commonly used, the INCSTAR/Dia-Sorin radioimmunoassay (Stillwater, MN), exhibits a mean between-laboratory CV of 20–30%. The current between-laboratory CV remains essentially the same when all methods are combined (personal file of survey reports). The variability between laboratories today is only marginally better than that reported in 1984 for the various methods combined (22). For the purposes of this review, it must be assumed that differences in 25(OH)D concentration between group means in different publications reflect true differences. This assumption cannot be avoided. Any attempt to select preferential 25(OH)D methods for data inclusion would result in the omission of most of the information in the literature. Furthermore, the proficiency data available show that selection of any one method would result in only a marginal improvement in the CV of comparisons between laboratories.Excluded from this review are data on children, studies in which dose was not specified, and studies in which the duration of the vitamin D dose was <4 wk (0.93 mo). 25(OH)D values presented here are either published numerical values or they were redigitized from graphical data. All 25(OH)D values are presented as nmol/L (1 nmol/L = 1 ng/mL × 2.5). Amounts of vitamin D are given in μg, each being equivalent to 40 IU or 2.6 nmol vitamin D3. No attempt was made to

distinguish between vitamin D2 and D3 because there was not always a distinction in the literature. However, vitamin D2 is less effective at raising serum 25(OH)D concentrations than is vitamin D3 (13).Previous Section Next Section

THE EVIDENCE: EFFECTS OF VITAMIN D SOURCES ON SERUM 25(OH)D

Ultraviolet lightSerum 25(OH)D concentrations of people living or working in sun-rich environments are summarized in Table 1⇓. The highest individual serum 25(OH)D concentration obtained from sunshine was 225 nmol/L (23); in a farmer in Puerto Rico—there is no reason to think that he was taking a vitamin D supplement. In a report showing results for 391 subjects that excluded individuals taking calcium or vitamin D supplements, Dawson-Hughes et al (14) reported 3 subjects with serum 25(OH)D concentrations >200 nmol/L. From the data of Chapuy et al (26), as well as those of Dawson Hughes et al (12), the upper limits (+2 SDs, 97.5th percentile) were, respectively, 150 and 212 nmol/L in subjects taking only 20 μg (800 IU) vitamin D/d. Such concentrations are consistent with subgroups of individuals exposed to relatively more sunshine.The effects of artificial ultraviolet light treatment sessions on 25(OH)D concentrations are summarized in Table 2⇓. The highest individual 25(OH)D concentration attained was 274 nmol/L (38). The main problem in interpreting the data was that the exact dose of ultraviolet light was ambiguous because there is variability in the surface area of skin exposed and in the frequency and duration of exposure. Had the ultraviolet treatment sessions continued, one would expect that for those given full-body exposure, serum 25(OH)D concentrations would plateau at mean values comparable with those of the farmers and lifeguards shown in Table 1⇓.Two studies showed that in response to a given set of ultraviolet light treatment sessions, the absolute rise in serum 25(OH)D concentration was inversely related to the basal 25(OH)D concentration. In the study by Mawer et al (34), the increase in 25(OH)D in subjects with initial 25(OH)D concentrations <25 nmol/L was double the increase seen in subjects with initial concentrations >50 nmol/L. Snell et al (27) showed that in subjects with initial 25(OH)D concentrations <10 nmol/L, ultraviolet treatments increased 25(OH)D by 30 nmol/L, but in those with initial 25(OH)D concentrations approaching 50 nmol/L, the increase was negligible.At least 4 studies support the concept that one full-body exposure to sunlight can be equivalent to an oral vitamin D intake of 250 μg (10000 IU). Stamp (39) compared oral vitamin D to the effects of ultraviolet light treatment sessions and found that the rise in 25(OH)D was the same in subjects treated with ultraviolet light as in those given 250 μg (10000 IU) vitamin D/d. In a study of institutionalized elderly, Davie et al (28) exposed 600 cm2, ≈5% of skin surface, to ultraviolet light treatments over a 2–3-mo period and compared the resulting 25(OH)D concentrations with those achieved with oral vitamin D doses. They calculated a production of vitamin D in the skin equivalent to 0.045 nmol •d−1•cm−2 exposed skin. This is equivalent to 10.9 μg (435 IU) vitamin D/d for 5% of skin surface. An almost identical protocol was followed recently by Chel et al (30), confirming the relative effects of light and supplementation on 25(OH)D concentration (28). If these results for the elderly are extrapolated to total body surface area, it works out to 218 μg (8700 IU) vitamin D/d that can be acquired by the elderly. More recently, Holick (40) presented data that compared blood vitamin D concentrations in subjects taking vitamin D orally with those given ultraviolet light exposure. The ultraviolet treatment produced blood vitamin D concentrations comparable with an intake of 250–625 μg (10000–25000 IU) vitamin D/d orally.Ultraviolet exposure beyond the minimal erythemal dose does not increase vitamin D production further. The ultraviolet-induced production of vitamin D precursors is counterbalanced by

degradation of vitamin D and its precursors. The concentration of previtamin D in the skin reaches an equilibrium in white skin within 20 min of ultraviolet exposure (41). Although it can take 3–6 times longer for pigmented skin to reach the equilibrium concentration of dermal previtamin D, skin pigmentation does not affect the amount of vitamin D that can be obtained through sunshine exposure (42). However, aging does lower the amount of 7-dehydrocholesterol in the skin and lowers substantially the capacity for vitamin D production (43, 44).Effect of acute onset of ultraviolet light deprivationMost of what is known about ultraviolet light deprivation is summarized in Table 3⇓. The most recent data are for an American submarine crew (46), which show a percentage decline in 25(OH)D concentration comparable with that of a British crew reported 20 y earlier (45). The 30-nmol/L decline over 2 mo was despite “a standard US Navy diet which included milk and breakfast cereals fortified with vitamin D” (46). The initial mean 25(OH)D concentration was higher in the Americans than in the British, consistent with higher vitamin D supplies from diet and sun exposure in Americans.Effect of daily vitamin D supplementationStudies reporting a specified vitamin D intake and the resulting serum 25(OH)D concentration are summarized in Table 4⇓. It is important in this kind of comparison to know whether the treatments had achieved an equilibrium concentration in terms of 25(OH)D. The half-life of 25(OH)D in the circulation is reported as ≈1 mo in humans (68), the results for the submariners suggest a 2-mo half-life (Table 3⇓). Conventional pharmacology indicates it should take 4 half-lives before a drug's equilibrium is achieved. Unlike a conventional drug, 25(OH)D is a metabolite whose concentration can be altered through balance between its production and clearance so that an equilibrium can be achieved earlier than would be expected from the half-life. To permit some sort of comparison between the various studies in the literature, it was necessary to compromise on classical pharmacology. The data in Figure 1⇓ show that serum 25(OH)D is essentially at the plateau concentration by 1 mo (28, 64). Therefore, the results for 25(OH)D of studies in which the vitamin D supplementation was continued for a minimum of 4 wk (0.93 mo) are summarized in Table 4⇓. The eventual plateau of 25(OH)D concentration may be slightly underestimated in some cases.25(OH)D with pharmacologic doses of vitamin D and in cases of vitamin D toxicityReports in which pharmacologic doses of vitamin D were given for a prolonged time and in which the resulting serum 25(OH)D concentrations were provided are summarized in Table 5⇓. Eventually, as 25(OH)D concentrations rise, the classic situation of hypercalcemia becomes evident. Hypercalcemia due to vitamin D intoxication per se is always accompanied by serum 25(OH)D concentrations >220 nmol/L (70, 75, 77).Pettifor et al (78) measured the free 1,25(OH)2D concentration in 11 people intoxicated by the erroneous consumption of a vitamin D concentrate that they had used as cooking oil. The serum 25(OH)D concentrations in these patients covered the range (300–1000 nmol/L) that was required in the study of Heaney et al (79) to show a calcium-absorptive response to serum 25(OH)D itself. The mean serum free 1,25(OH)2D fraction in the study group of Pettifor et al was double the mean for normal individuals, implicating free 1,25(OH)2D as a contributor to toxicity when 25(OH)D concentrations match those used by Heaney et al. Not all of the subjects in Pettifor et al's study had elevated free 1,25(OH)2D, yet all subjects were hypercalcemic; therefore, the most likely agent causing vitamin D toxicity was a combination of inappropriate activity of both 25(OH)D and 1,25(OH)2D.The data of Table 5⇓ were incorporated into Figure 2⇓ to show the relation between vitamin D intake and the serum 25(OH)D concentrations achieved. Not all of the high-dose vitamin D intake

data involve vitamin D intoxication. There is one case of an individual with vitamin D toxicity for which the intake was 250 μg (10000 IU)/d.Depot injection or intermittent large oral dosing with vitamin DAnother way to augment vitamin D nutrition has been to administer vitamin D as a single large dose, either orally or through injection. Because vitamin D has a half-life >1 or 2 mo, a large dose should suffice for the better part of a year. The approach is often called stoss therapy (from the German for to bump) and it is most common in Europe. One Finnish study, by Heikinheimo et al (80), concluded that annual injection of vitamin D2 in the autumn (3750–7500 μg, or 150000–300000 IU) can lower the probability of osteoporotic fractures by ≈25%. Note that before treatment, the mean 25(OH)D concentration in the untreated control subjects was only 16 nmol/L, and thus, much of the fracture prevention appears to have been attributable to raising 25(OH)D concentrations out of the osteomalacic range.The effects of single large doses of vitamin D are summarized in Table 6⇓. Only a few studies monitored serum 25(OH)D concentrations during the days, weeks, or months after administration of large doses of vitamin D. Both Davie et al (28) and Weisman et al (87) showed that with oral dosing, there was a relatively rapid peak in serum 25(OH)D concentration, with concentrations falling progressively afterward. When doses were administered intramuscularly, there was a longer-lasting response in terms of serum 25(OH)D. However, in some cases it took ≈2 mo for the peak concentration to be achieved. In the study by Heikinheimo et al (80), serum 25(OH)D concentrations after injection remained higher than those of the control group for the entire year.Reliability of intermittent dosing was poor because of wide variability in the serum 25(OH)D concentrations achieved, both between individuals and in each individual over time. Consequently, the dose-response graph of Figure 2⇓ does not include results from groups given single large injections.Previous Section Next Section

ASSESSMENT OF EVIDENCE

Objectives achieved with vitamin D intake recommendationsIf the objective of vitamin D supplementation is simply to avoid severe osteomalacia, then the 2.5 μg (100 IU)/d requirement suggested by Dent and Smith (89) might be adequate for some individuals. Supplementation with 10 μg (400 IU) vitamin D/d raises 25(OH)D by ≈45 nmol/L (28, 30, 47, 56, 90). By extrapolating downward, it appears that the commonly consumed amount of 5 μg (200 IU)/d may be sufficient, as a sole source of vitamin D, to maintain average serum 25(OH)D concentrations at 25 nmol/L, but there are no data to verify this.There is a prevalent view that occasional exposure of the face and hands to sunlight is “sufficient” for vitamin D nutrition. Indeed, this exposure can provide 5–10 μg (200–400 IU) vitamin D during those months when the appropriate sunlight is available. However, a 5% skin exposure produces a mean 25(OH)D concentration of only 35 nmol/L (28), which would leave more than half of the population vitamin D insufficient. Gloth et al (20) found that in homebound, American elderly, the mean vitamin D intake was 12.9 μg (517 IU)/d, resulting in a mean 25(OH)D concentration of 40 nmol/L. Interestingly, for subjects with serum 25(OH)D concentrations <25 nmol/L, the mean vitamin D intake was 11.7 μg (467 IU)/d. Apparently, some people require more vitamin D than others to reach a given concentration of 25(OH)D in serum. More recent reports show that dietary vitamin D intake correlates poorly with 25(OH)D concentrations (2, 9) and that 25(OH)D concentrations can hover around what is considered to be marginal deficiency (38 nmol/L) despite consumption of the recommended amount of vitamin D (9). Ultraviolet exposure and time spent outdoors are better predictors of 25(OH)D concentration than is dietary vitamin D

intake (9). Obviously, the current, arbitrarily set vitamin D intakes play a minor role in the total economy of vitamin D for most adults.If the objective is to optimize the probability of good health, then it seems reasonable that the daily vitamin D supply should be >20 μg (800 IU)/d. In both of the well-accepted studies showing osteoporosis fracture prevention with vitamin D and calcium, mean 25(OH)D concentrations exceeded 100 nmol/L (12, 26). The osteoporosis studies of Chapuy et al (26) and Heikinheimo et al (86) have been interpreted as evidence that the value of the vitamin supplement is simply to raise 25(OH)D concentrations out of the osteomalacic or insufficient range. However, patients in the recent study of Dawson-Hughes et al (12) started off with perfectly acceptable basal 25(OH)D concentrations of 61 nmol/L and still derived benefit. Target 25(OH)D concentrations exceeding 100 nmol/L would also minimize PTH concentrations because the decrease in 25(OH)D with age has a greater effect on the secondary hyperparathyroidism of aging than does the decline in renal function (10).For most drugs and hormones, the relevant serum concentration attained is more important than is the actual intake of the agent. The recent increase in the DRI for vitamin D for adults >50 y of age was based on evidence of fracture prevention with vitamin D and calcium, as shown by Dawson-Hughes et al (12) and Chapuy et al (26). In both of those studies, mean serum 25(OH)D concentrations exceeded 100 nmol/L, which was unusual. Of the 21 groups summarized in Table 4⇓ to whom vitamin D was given at doses between 20 and 75 μg (800 and 3000 IU)/d, it was only the subjects in these 2 studies that attained average 25(OH)D concentrations >100 nmol/L (12, 26), except for subjects who lived near the equator (55). Calcium supplementation would not account for the unusual results because calcium by itself does not affect serum 25(OH)D in elderly women (91). A glance at Table 4⇓ will show that to ensure the desirable objective of mean 25(OH)D concentrations >100 nmol/L, the total vitamin D supply from dietary and environmental sources must be 100 μg (4000 IU)/d. This view is also based on the results of recent unpublished work (R Vieth, 1998) and on reference 67.The objective of vitamin D supplementation should be to compensate for insufficient ultraviolet light exposure. Studies in submariners offer a reasonable way to estimate the decrement in vitamin D. A low estimate for the vitamin D decrement of submariners is 15 μg (600 IU) vitamin D2/d. This was inferred by Holick on the basis of a small (n = 11) unpublished study done in 1984 that compared supplemented and unsupplemented sailors. Unfortunately, neither the actual 25(OH)D concentrations nor statistical significance were provided, and the summary includes the unreasonable observation that final 25(OH)D concentrations were 23% higher in subjects not taking any vitamin D (92). Studies with more complete data are summarized in Table 3⇓. One way to estimate the serum 25(OH)D that may have been lost in the submariners reported on by Dlugos et al (46) is to determine the level of vitamin D supplementation that would sustain their initial 25(OH)D concentration of 78 nmol/L. Data in Table 4⇓ show that an additional supplement of 10–25 μg (400–1000 IU) may be required to sustain the sailors' initial 25(OH)D concentrations. This is probably an underestimate because none of the vitamin D–supplemented subjects were totally sun deprived. Another approach to estimating the submariners' vitamin D decrement could be based on the amount of vitamin D required to produce the opposite effect on serum 25(OH)D within the same time period. In Table 4⇓, it can be seen that MacLennan and Hamilton (63) reported in 1977 that 50 μg (2000 IU) vitamin D/d increased 25(OH)D concentrations from basal values of 15 to final values of 81 nmol/L. This was verified by Himmelstein et al (64), who showed that 2 mo of 50 μg (2000 IU) vitamin D/d raised serum 25(OH)D concentrations from the initial 40 to 80 nmol/L (Figure 1⇓). On this basis, the sailors were deprived of ≈50 μg (2000 IU) environmental vitamin D/d during the time they were confined to the submarine.

Other reasons to consider increasing vitamin D nutritionThere are nonclassical and less well substantiated reasons to consider increasing vitamin D nutrition. As discussed above, the biochemical criterion of PTH suppression suggests that 25(OH)D concentrations >100 nmol/L are desirable in elderly subjects (12, 15). Epidemiologic studies show that higher serum 25(OH)D concentrations or environmental ultraviolet light exposure are associated with lower rates of breast, ovarian, prostate, and colorectal cancers (93–100). There is impressive circumstantial evidence that multiple sclerosis is more prevalent in populations having lower concentrations of vitamin D or ultraviolet exposure (98,101), and there are suggestions that vitamin D intake ranging from 32.5 to 95 μg (1300 to 3800 IU)/d helps prevent the disease (101). The probability that established osteoarthritis will progress to a more severe stage is reduced with better vitamin D nutritional status, based both on serum 25(OH)D concentrations and diet history. On the basis of these results, McAlindon et al (102) recommended that serum 25(OH)D should exceed 75 nmol/L in persons with osteoarthritis of the knee. The prevalence of hypertension in a population increases with distance north or south of the equator (103), and it was reported recently that hypertension becomes less severe in subjects whose 25(OH)D concentrations are increased to >100 nmol/L through ultraviolet exposure (38). Vitamin D deficiency impairs immune function in animals (104), and in children there is a strong association between pneumonia and nutritional rickets (105). Vitamin D nutrition probably affects major aspects of human health other than its classical role in mineral metabolism; however, the evidence is not conclusive enough yet to warrant considering these other potential health benefits as objectives in nutritional guidelines. If any of this evidence were taken into consideration it would require substantial upward revision of the current DRI.Some investigators found no statistically significant relations between serum 25(OH)D concentrations in archived tissue samples and eventual prostate cancer death (106, 107). Two points seem to have been missed in these studies. First, 25(OH)D concentrations vary with season and this will confound conventional approaches to cancer follow-up. Second, modern society in general is vitamin D–deprived compared with prehistoric humans. The concentrations of 25(OH)D observed today are arbitrary and based on contemporary cultural norms (clothing, sun avoidance, food choices, and legislation) and the range of vitamin D intakes being compared may not encompass what is natural or optimal for humans as a species.Dose-response curve of 25(OH)D concentration versus vitamin D intakeThe serum 25(OH)D concentration is maintained within a narrow range (Figure 2⇓), ≈75–220 nmol/L across vitamin D supplies from 20 μg (800 IU) to the physiologic limit of 250–500 μg (10000–20000 IU)/d. The most reasonable explanation for this kind of relation is that there are homeostatic control systems to regulate serum 25(OH)D and to buffer against variability in vitamin D supply. The metabolic points at which serum 25(OH)D can be regulated include the concentration of 25-hydroxylase in the liver (108), the catabolism of 25(OH)D by the liver into breakdown products excreted into bile (109), and the catabolism of 25(OH)D via the side-chain cleavage pathway initiated by 24-hydroxylase present in tissues throughout the body (110). Beyond the vitamin D supply limit, which is comparable with that attainable with sunshine, there is a classic rise in the dose-response curve. The sharp rise reflects the introduction of vitamin D and 25(OH)D at rates that exceed the capabilities of the various mechanisms to regulate 25(OH)D.One case illustrates the physiologic limit well. The patient, who had primary hypoparathyroidism, is indicated by the arrow in Figure 2⇓. This data point is an outlier because it is clearly the highest serum 25(OH)D concentration within the physiologic range of vitamin D intake. Although this subject did not take a remarkably large dose of

vitamin D daily, the effect was toxic because the dose was taken as a single monthly dose of 7500 μg (300000 IU) (77). This exceeded by 30-fold the physiologic range and resulted in the production of more 25(OH)D than the body had the capacity to clear from the system at one time—hence toxicity with one form of stoss therapy.Aside from the vitamin D supply itself, endocrine and dietary factors affect circulating 25(OH)D. In rats, both low (109, 111) and excessively high calcium intakes (111) can lower serum 25(OH)D concentrations. The calcium-raising hormones each tend to lower serum 25(OH)D. In hypoparathyroid humans, 1,25(OH)2D treatment speeds up the metabolic clearance of 25(OH)D from the circulation (68). In hyperparathyroid humans, there is accelerated clearance of 25(OH)D (112).The preceding discussion highlights the fact that 1,25(OH)2D is not the only metabolite that is regulated in the vitamin D endocrine system. The purpose of the mechanisms to regulate 25(OH)D concentrations may be to optimize the availability of 25(OH)D for tissues that require it, either for its direct action (79), or as the source of substrate for nonrenal, paracrine 1-hydroxylase (calcidiol 1-monooxygenase, EC 1.14.13.13) (113).Lowest dose causing harmThe recent paper by Adams and Lee (16) about mild vitamin D toxicity defined elevated 25(OH)D as anything >125 nmol/L, which was the upper limit of the reference range stated by their referral laboratory. Their subject with the highest concentration of urinary calcium had a serum 25(OH)D concentration of 140 nmol/L, and this was on only one occasion. When this subject's serum 25(OH)D fell to 102 and then to 75 nmol/L, the urinary ratio of calcium to creatinine remained unchanged at its highest value, suggesting some other metabolic cause of the hypercalciuria. The other cases in the report more closely resemble those given milk excessively supplemented with vitamin D reported by Jacobus et al (114) or those in the poisoned household reported by Pettifor et al (78).Except for the report by Adams and Lee (16), all instances of vitamin D toxicity have involved serum 25(OH)D concentrations in excess of 200 nmol/L (Table 5⇓). Adams and Lee came across their putative cases of vitamin D intoxication by checking urinary calcium concentrations in patients screened in an osteoporosis evaluation. Certainly, the first sign of vitamin D excess would involve an increase in urinary calcium, but whether this occurs with physiologic 25(OH)D concentrations in healthy individuals is by no means established by their study (115, 116). Furthermore, Adams later stated that the subjects had consumed amounts of vitamin D “at least one order of magnitude greater than” what was on the label (ie, ≥10 times 30 μg or 1200 IU/d) (117).Dawson-Hughes et al (14) found 25(OH)D concentrations comparable with those of Adams and Lee's subjects in ≈20 healthy elderly men and women who were not taking vitamin D supplements. Likewise, presumably healthy farmers in Puerto Rico and lifeguards also had such 25(OH)D concentrations (Table 1⇓). Although not strictly within the “normal” range for a clothed, sun-avoiding population, serum 25(OH)D concentrations ≤220 nmol/L are consistent with certain environments, are not unusual in the absence of vitamin D supplements, and should be regarded as being within the physiologic range for humans.The report of Adams and Lee (16), together with its accompanying editorial, suggest that serum 25(OH)D concentrations as low as 140 nmol/L are harmful. This is alarmist. Are we to start avoiding the sun for fear of raising urine calcium or increasing bone resorption? The question has never been addressed objectively. My view is that there is no harm in the 25(OH)D concentrations associated with sun exposure and that such concentrations are probably optimal for human health.Higher rates of osteoporosis and arteriosclerosis have been attributed to virtually any vitamin D intake (118). Furthermore, the US National Academy of Sciences (3) indicated in 1989 that the toxic

dose of vitamin D can be as low as “five times the RDA.” This view now seems to have been carried forward to the latest set of vitamin D recommendations, in which the tolerable upper intake level is indicated as 50 μg (2000 IU)/d (19).Throughout my preparation of this review, I was amazed at the lack of evidence supporting statements about the toxicity of moderate doses of vitamin D. Consistently, literature citations to support them have been either inappropriate or without substance. The statement in the 1989 US nutrition guidelines that 5 times the RDA for vitamin D may be harmful (3) relates back to a 1963 expert committee report (5), which then refers back to the primary reference, a 1938 report in which linear bone growth in infants was suppressed in those given 45–157.7 μg (1800–6300 IU) vitamin D/d (119). The citation is not related to adult nutrition and it does not form a scientific basis for a safe upper limit in adults. The same applies to the statement in the 1987 Council Report for the American Medical Association that “dosages of 10,000 IU/d for several months have resulted in marked disturbances in calcium metabolism…and, in some cases, death.” Two references were cited to substantiate this. One was a review article about vitamins in general, which gave no evidence for and cited no other reference to its claim of toxicity at vitamin D doses as low as 250 μg (10000 IU)/d (120). The other paper cited in the report dealt with 10 patients with vitamin D toxicity reported in 1948, for whom the vitamin D dose was actually 3750–15000 μg (150000–600000 IU)/d, and all patients recovered (121). If there is published evidence of toxicity in adults from an intake of 250 μg (10000 IU)/d, and that is verified by the 25(OH)D concentration, I have yet to find it.The issue of poorly substantiated claims of toxicity extends even to the most recent, 1997, revision for vitamin D intakes published by the National Academy of Sciences (19). The only study cited to address the question of critical endpoint doses for vitamin D (potential adverse effect levels) was an esoteric article by Narang et al (122). The basis for the current no observed adverse effect level (NOAEL) of 50 μg (2000 IU)/d is that Narang et al reported a mean serum calcium concentration >11 mg/dL(2.25 mmol/L) in the 6 normal subjects given 95 μg (3800 IU) vitamin D/d. That intake was then defined as the lowest observed adverse effect level (LOAEL). The next lowest test dose used by Narang et al, 60 μg (2400 IU)/d, with 20% less as the safety margin, became the NOAEL. Narang et al reported only serum electrolyte changes, the doses of vitamin D were not verified, and 25(OH)D concentrations were not reported. The National Academy committee missed evidence showing the safety of larger doses of vitamin D. Directly comparable with the protocol of Narang et al (122) is the study by Tjellesen et al (67), in which 19 normal subjects were given 95 μg (3800 IU) vitamin D2 or 110 μg (4400 IU) vitamin D3/d (these doses were validated by direct assay) (Table 4⇓). Serum calcium increased by a minute, but statistically significant, 0.05 mmol/L (0.12 mg/dL) with 110 μg (4400 IU) vitamin D3 and there was an increase in urinary calcium, as one should expect with improved intestinal calcium absorption. Another report indicative of the safety of larger doses is by Stern et al (84), who administered vitamin D2 at 2500 μg (100000 IU)/d for 4 d (Table 6⇓). This represented a far greater stress to the vitamin D system because it delivered in 4 d a total vitamin D dose equivalent to what Narang et al used over 3 mo. Nonetheless, Stern et al (84) detected no significant effects on serum calcium in the 24 normal adult subjects they treated in this way.The mechanism causing vitamin D toxicity involves the unbridled expression of 1,25(OH)2D-like activity. Whether 1,25(OH)2D or 25(OH)D is the main signaling molecule causing vitamin D toxicity remains a point of contention. There is a growing likelihood that 25(OH)D has biological activity in its own right. The concentration of 25(OH)D is often reported to be better correlated with absorption of calcium from the gut than is serum 1,25(OH)2D (123, 124). Recently, Heaney et al (79) showed that the circulating 25(OH)D concentration

affects intestinal calcium absorption. However, to show this, it was necessary to raise serum 25(OH)D concentrations into the range of 300–1000 nmol/L, and this is well beyond what could be considered physiologic.As vitamin D doses increase, the mechanisms to explain the toxic responses are 3-fold: 1) a possible conversion of vitamin D3 to 5,6-trans-vitamin D3, which contains a pseudo-1-α group; 2) a direct action of 25(OH)D at the 1,25(OH)2D receptor (79); or 3) an unbridled production of 1,25(OH)2D with inappropriate maintenance of its total concentration in the circulation despite its displacement from vitamin D binding protein to increase free 1,25(OH)2D (78, 125). The opinion that 95 μg (3800 IU) vitamin D/d is the LOAEL (19) is not consistent with these current theories of why vitamin D is toxic and is not consistent with the amounts of vitamin D needed to raise 25(OH)D concentrations to the hypercalcemic levels reported in all studies in which serum 25(OH)D is related to toxicity.Fraser (126) speculated that orally acquired vitamin D might be particularly toxic because it enters the body via an unnatural route for this nutrient. This hypothesis has never been put to the test. It does seem reasonable that the metabolism of vitamin D acquired through the skin might be more finely regulated than that of vitamin D obtained orally. Haddad et al (127) showed that the transport of vitamin D in the circulation is different for vitamin D acquired by dermal and oral routes. Oral vitamin D is primarily transported along with chylomicrons and lipoproteins until it is cleared by the liver within hours, whereas dermal vitamin D is transported on vitamin D binding protein and takes days to clear. We recently observed that the apparent self-regulation of 25-hydroxylase first observed for vitamin D generated in skin (27, 34) pertains equally well to vitamin D acquired orally. Oral vitamin D supplementation produced the greatest increase in 25(OH)D in those who initially had the lowest 25(OH)D concentrations (13). From what is known now, there is no practical difference whether vitamin D is acquired from ultraviolet exposed skin or through the diet.Hypersensitivity to vitamin DHypersensitivity to vitamin D can occur (128). Primary hyperparathyroidism is probably the most common example. It would be simplistic to avoid or minimize vitamin D intake because of this. Before the occurrence of hyperparathyroidism, vitamin D nutrition is preventive because it reduces parathyroid secretion and lowers the likelihood of parathyroid hyperplasia (129–131). Once primary hyperparathyroidism exists, production of 1,25(OH)2D is persistently up-regulated by the high PTH concentrations, and 1,25(OH)2D concentrations correlate directly with serum 25(OH)D (132). In hyperparathyroid individuals, vitamin D exaggerates hypercalcemia because of the connection between vitamin D nutrition and 1,25(OH)2D production. Vitamin D deficiency can mask primary hyperparathyroidism (132) and this could account for the occasional cases of hypercalcemia that occur when large groups of elderly people are given vita min D supplements (133). Some patients with sarcoidosis, tuberculosis, or lymphoma become hypercalcemic in response to any increase in vitamin D nutrition (122,134, 135). For these persons, it may be prudent to avoid any dietary or environmental sources of vitamin D.Efficacy and safetyAlthough the new DRI for vitamin D for most adults is 5 μg (200 IU)/d (19), the beneficial amount is more likely to be 10–12.5 μg (800–1000 IU)/d, on the basis of bone density measurements and fracture prevention in the elderly (12, 18, 26). For this reason, DRIs have been increased for those >70 y of age to 600 IU/d. This intake will also lessen the chance of vitamin D deficiency–induced secondary hyperparathyroidism and will bring serum 25(OH)D concentrations closer to those associated with other health benefits. Even if all adults consumed 12.5 μg (1000 IU)/d, it would be difficult to detect an increase in the number of individuals with 25(OH)D concentrations >140 nmol/L because ≥90% of the vitamin D

contributing to such concentrations would be from sunshine exposure, not oral intake.The assignment of a NOAEL, as defined by Hathcock (21), or allowable “tolerable upper intake level,” as defined by the Food and Nutrition Board, Institute of Medicine (19), is especially difficult for vitamin D. Because of environmental input, the concentration referred to must be that of 25(OH)D in the circulation and not simply dietary vitamin D. Another consideration is that rare individuals are hypersensitive to vitamin D. There is no simple answer here because the primary disease may be made more likely to occur by previous vitamin D deficiency (129). The following discussion disregards the possibility of vitamin D hypersensitivity. If it exists, hypersensitivity would appear to negate the value of any vitamin D intake or sunshine exposure.For vitamin D, a NOAEL could define the highest 25(OH)D concentration not suspected to cause hypercalciuria in healthy subjects. As discussed above, it could be difficult to prove that vitamin D is the cause of hypercalciuria because the condition is commonly caused by other things and it is a mild and nonclassicicl criterion for vitamin D intoxication. Nonetheless, because of the hypothesized predisposition to hypercalciuria of the Israeli lifeguards (25) and of the subjects in the study of Adams and Lee (16), 140 nmol 25(OH)D/L could be regarded as a very conservative limit for the NOAEL because the concentrations in those reports were higher than this. In the absence of sunshine, all available evidence indicates that this would require prolonged intake of ≈250 μg (10000 IU)/d to achieve. All of the reports of vitamin D toxicity showing the convincing evidence of hypercalcemia involve serum 25(OH)D concentrations well above 200 nmol/L (Table 5⇓), which requires a daily intake of ≥1000 μg (40000 IU), and which could thus be conservatively considered the LOAEL.

The current adult DRI for vitamin D approximates half the amount in the teaspoon of cod-liver oil that was a 19th-century folk remedy. Today, new drugs are passed through dose-finding studies before their efficacy is evaluated in clinical trials. This principle is not strictly applicable to nutrient recommendations because the bulk of what humans consume of them is from unfortified foods and this consumption is what recommended intakes tend to match. In contrast, vitamin D is a special case; the bulk of our dietary vitamin D intake is determined by legislation. I contend that this practice amounts to the dosing of populations with a drug, vitamin D, that is not present in the foods humans normally consume. If vitamin D is similar to a drug, then dose-finding studies are needed to use it properly, especially if nonclassical benefits are potentially relevant. Alternatively, if by analogy with other nutrients, vitamin D supplementation is intended to make up for what some people may not be getting from its natural source, in this case the sun, then the current adult DRI of 5 μg (200 IU)/d is woefully inadequate.

FIGURE 1.

Effect of the duration of vitamin D supplementation on the mean serum 25-hydroxyvitamin D [25(OH)D] concentration achieved. Vitamin D intakes for the groups were as follows: 10 μg (400 IU)/d (+; 28), 25 μg (1000 IU)/d (X; 28), 50 μg (2000 IU)/d (□; 64), 250 μg (10000 IU)/d (▪; 28).

FIGURE 2.Dose response for vitamin D intake versus final serum 25-hydroxyvitamin D [25(OH)D] concentration reported. Circles indicate group means from Tables 4 and 5⇑ ⇑ . Points indicated by “X” represent single values from Table 5⇑ for people reported as intoxicated with vitamin D. The arrow indicates the lowest dose reported as causing hypercalcemia, but which is an outlier because vitamin D was given as a single dose of 7500 μg (300000 IU)/mo, instead of 250 μg (10000 IU)/d (77). If authors reported 25(OH)D for several time points, only the final serum 25(OH)D is shown in the figure. The line representing the dose response passes through the points that represent group data

Relationship between Serum 25-Hydroxyvitamin D and Lower Extremity Arterial Disease in Type 2

Diabetes Mellitus Patients and the Analysis of the Intervention of Vitamin D

Wan Zhou and Shan-Dong Ye

Department of Endocrinology, Anhui Provincial Hospital Affiliated to Anhui Medical University, Hefei, Anhui 230001, China

Received 17 January 2015; Revised 6 March 2015; Accepted 11 March 2015

Academic Editor: Ronald G. Tilton

Copyright © 2015 Wan Zhou and Shan-Dong Ye. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The aim of this study was to explore the relationship between serum 25-hydroxyvitamin D [25(OH)D] concentrations and lower extremity arterial disease (LEAD) in type 2 diabetes mellitus (T2DM) patients and to investigate the intervention effect of vitamin D. 145 subjects were assigned to a control group (Group NC), T2DM group (Group DM1), and T2DM complicated with LEAD group (Group DM2); then Group DM2 were randomly divided into Group DM3 who received oral hypoglycemic agents and Group DM4 who received oral hypoglycemic drugs and vitamin D3 therapy. Compared to Group NC, 25(OH)D was significantly lower in Group DM2 and marginally lower in Group DM1. In contrast to baseline and Group DM3, 25(OH)D rose while low density lipoprotein (LDL), retinol binding protein 4 (RBP4), and HbA1c significantly lowered in Group DM4. Statistical analysis revealed that 25(OH)D had a negative correlation with RBP4, duration, HbA1c, homeostasis model assessment for insulin resistance (HOMA-IR), and fasting plasma glucose (FPG). LDL, systolic blood pressure (SBP), FPG, and smoking were risk factors of LEAD

while high density lipoprotein (HDL) and 25(OH)D were protective ones. Therefore, we deduced that low level of 25(OH)D is significantly associated with the occurrence of T2DM complicated with LEAD.

1. Introduction

Lower extremity arterial disease (LEAD) is a common peripheral arterial disease, which seriously affects the patient’s functional capacity and quality of life [1]. It is one of the factors that contribute to the progressive and the critical courses of foot ulceration and amputation in type 2 diabetes mellitus (T2DM) patients [2], and it is frequently associated with coronary, cerebral, and renal artery diseases [3]. Vitamin D is a secosteroid, which is obtained from exposure to sunlight and through dietary sources including food and supplements. It is hydroxylated in the liver to 25-hydroxyvitamin D [25(OH)D] and further hydroxylated in the kidney to form 1,25-dihydroxyvatamin D [1,25(OH)2D, calcitriol]. Although 1,25(OH)2D is considered to be the active form of vitamin D, its level in the serum does not correlate with overall vitamin D status, whereas the 25(OH)D level is a more clinically relevant marker [4]. In addition to the traditional involvement of vitamin D in bone metabolism, several lines of evidence suggest a role for vitamin D in glucose levels, insulin resistance (IR), and prevalence of T2DM [5–8], and there are more and more studies that have found that it participated in systemic inflammation, immune, and lipid metabolism to reduce the risk of cardiovascular diseases [9], but few investigations suggest the association between vitamin D and T2DM complicated with LEAD. Hence, demonstrating a relationship between 25(OH)D and T2DM complicated with LEAD is necessary. Retinol binding protein 4 (RBP4) is a new adipokine identified by Yang et al. [10] using gene chip technology, which involves the occurrence of IR, T2DM, and macrovascular complications. Several researches [11] have shown that vitamin D is correlated with adiponectin, leptin, and other adipokines, but there are few studies analyzing a relationship between 25(OH)D and RBP4 in T2DM patients complicated with LEAD. Therefore, we realize the importance of further investigation in this area. In this effort, we conducted a clinic-based case-control study to explore the relationship among 25(OH)D, RBP4, and T2DM complicated with LEAD and analyze the intervention effect of vitamin D for LEAD.

2. Materials and Methods

2.1. Subjects

107 patients (63 males and 44 females) who were recruited from outpatient department were in line with the T2DM diagnosis standard delivered by World Health Organization (WHO) in October 1999. They were hospitalized from October 2012 to January 2013 in the Department of Endocrinology in Anhui Provincial Hospital. This study was conducted in accordance with the tenets of the World Medical Association’s Declaration of Helsinki and had been approved by the China Ethics Committee. Written informed consent was obtained from all participants.

The subjects who met the following criteria were excluded: individuals who had acute complications of T2DM, retinopathy, nephropathy, and other chronic complications, cancer, vitamin A deficiency, liver and kidney dysfunction, osteoporosis, and other history of bone metabolic disorders; those taking drugs including vitamins, calcium, lipid lowering drugs, estrogen, and other drugs affecting bone metabolism. All the patients were assigned into 2 groups: the one that is without complications (Group DM1: 25 males, 20 females) and the other that is complicated with LEAD (Group DM2: 38 males, 24 females); then subjects in Group DM2

were further divided into Group DM3 (16 males, 15 females) and Group DM4 (22 males, 9 females). Patients in these 2 groups were both treated with hypoglycemic drugs, and additionally patients of Group DM4 were orally administered vitamin D3 1000 IU daily. In this period 2 subjects in Group DM3 were out of touch and 4 subjects in Group DM4 could not insist on the intervention therapy and quitted. The relevant characteristics were tested and ABI was reexamined in Groups DM3 and DM4 after 12 weeks of therapy. 38 age-sex-matched healthy people with no hypertension or impaired glucose tolerance were chosen at the same period as a control group (Group NC) (20 males and 18 females).

2.2. ABI Measurement

The Doppler instrument (Bidop ES-100-v3 produced in Japan) was adapted. All the ABI measurements were performed with the subjects in the supine position, a blood pressure cuff was placed on patients’ upper arms, and it was inflated until no brachial pulse was detected by the Doppler device. The cuff was then slowly deflated until the Doppler-detected pulse returned to measure brachial systolic blood pressure (BSBP). This maneuver was repeated on the leg, with the cuff being wrapped around the distal calf and the Doppler device being placed over the dorsalis pedis or the posterior tibial artery to measure ankle systolic blood pressure (ASBP) [12]. Ankle brachial index (ABI) = ASBP/BSBP and ABI < 0.9 indicated the presence of LEAD [13].2.3. Assays

All the subjects took balanced diets for 3 days and then fasted overnight for 12 hours. Clinical characteristics including sex, age, diabetes duration, height, weight, hip circumference, blood pressure, and history of smoking were collected. Body mass index (BMI) and waist-hip ratio (WHR) were calculated as weight divided by height squared (kg/m2) and waist circumference divided by hip circumference, respectively. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured using a digital automatic blood pressure monitor. Patients’ blood samples were kept frozen at −80°C until analysis. Fasting plasma glucose (FPG), total cholesterol (TC), triglycerides (TG), low density lipoprotein (LDL), and high density lipoprotein (HDL) were detected by Hitachi 7600 2020 automatic biochemical analyzer. HbA1C was conducted by high-pressure liquid chromatography (Bio-Rad variant turbo II analyzer). Fasting insulin (FINS) was tested by radioimmunoassay (Linco Research, St. Charles, USA). The homeostasis model assessment for insulin resistance (HOMA-IR) was calculated according to the formula HOMA-IR = FINS (mU/L) × FPG (mmol/L)/22.5. Serum 25(OH)D concentrations were detected via radioimmunoassay (Diasorin Stillwater, MN, USA). Sufficiency was indicated when 25(OH)D ≥ 30 ng/mL, insufficiency when 20 ng/mL ≤ 25(OH)D < 30 ng/mL, and deficiency when 25(OH)D < 20 ng/mL [14]. The level of RBP4 was measured by enzyme-linked immunoassay (AssayPro, MO, USA).2.4. Statistical Method

All data were calculated by SPSS17.0 statistical software (SPSS Inc., Chicago, IL, USA). Continuous variables are presented as mean ± standard deviation (), or median and interquartile range (25th and 75th percentile) in cases of skewed distributions, while categorical variables are presented as percentages. These groups were compared using two-sample -test, Mann Whitney test, or chi-square test. Paired sample -test was used to compare the differences before and after treatment. Pearson’s correlation coefficient or Spearman’s rank correlation coefficient was used to assess the relationship between 25(OH)D and other markers. Multivariable logistic regression analysis was used to calculate the ORs and 95% confidence intervals for LEAD; statistical significance was accepted at .

3. Results

3.1. Clinical and Biochemical Characteristics

The indexes of the subjects were listed in Table 1. There were no differences of sex between the 3 groups (). Participants in Group DM2 had a longer duration than in Group DM1; the differences were statistically significant . Serum 25(OH)D concentrations were  ng/mL,  ng/mL, and  ng/mL in Groups NC, DM1, and DM2, respectively; compared with Group NC, 25(OH)D decreased in Groups DM1 and DM2 and the decrease in Group DM2 was more significant . RBP4 level was lower in T2DM patients with LEAD compared to the patients without LEAD and healthy people . Raised FPG, FINS, HbA1C, TC, TG, LDL, HOMA-IR, SBP, and WHR and decreased HDL in Groups DM1 and DM2 were exhibited as compared with Group NC ; age, FPG, LDL, SBP, HbA1C, and the rate of smoking increased and HDL decreased in Group DM2 compared with the other 2 groups .Table 1: The comparison of basic characteristics in Groups NC, DM1, and DM2.

3.2. Cases with Different Levels of 25(OH)D among Three Groups

As depicted in Table 2, approximately 18.4% of subjects were classified as 25(OH)D sufficient and 65.8% were classified as 25(OH)D insufficient; only 15.8% of subjects were classified as 25(OH)D deficient in Group NC. In Group DM1, only 8.9% of subjects had normal 25(OH)D level and 26.7% had insufficiency 25(OH)D level while 64.4% were 25(OH)D deficient. The percentages of 25(OH)D deficiency, insufficiency, and sufficiency were 77.4%, 14.5%, and 8.1%, respectively, in Group DM2. The percentages of sufficient 25(OH)D level in Groups DM1 and DM2 were both lower than the percentage in Group NC, and the difference was statistically significant ().Table 2: Cases with different levels of 25(OH)D in Groups NC, DM1, and DM2.

3.3. The Levels of Clinical Characteristics before and after the Treatment in Diabetic Groups Complicated with LEAD

There was no significant difference between Groups DM3 and DM4 at baseline. After 12 weeks of vitamin D supplementation in Group DM4, 25(OH)D level increased and LDL, RBP4, and HbA1C decreased significantly when compared with baseline , while, following 12-week basic therapy of hypoglycemic drugs in Group DM3, only RBP4 and HbA1C decreased (). Supplementation with vitamin D for 12 weeks in diabetics with LEAD significantly raised 25(OH)D level and lowered LDL, RBP4, and HbA1c relative to baseline and unsupplemented diabetics (). Reexamination of ABI by Doppler ultrasonography for subjects in Group DM3 and Group DM4 after 12 weeks showed that there were, respectively, 5 cases in Group DM3 and 10 cases in Group DM4 whose values increased to ≥0.9, and the difference was not statistically significant () (Table 3).Table 3: The comparison of characteristics before and after treatment in Groups DM3 and DM4.

3.4. Correlations between 25(OH)D and Other Indexes

As depicted in Table 4 the correlation analysis results suggested that the level of 25(OH)D was inversely associated with duration (, ), HbA1c (, ), RBP4 (, ), FPG (), and HOMA-IR (, ), while there was no significant correlation between 25(OH)D and the remaining indicators (). 25(OH)D had a positive correlation trend with HDL, but the difference was not statistically significant ().Table 4: The analysis of the correlation between 25(OH)D and the indicators of patients in Group T2DM.

3.5. Independent Factors for the Presence of LEAD

As shown in Table 5, the presence of LEAD in T2DM patients was used as the dependent variable, while 25(OH)D, TC, TG, HDL, LDL, BMI, RBP4, HOMA-IR, SBP, FINS, FPG, WHR, age, sex, duration, and smoking were used as the independent variables. Multivariable logistic regression analysis showed that smoking (OR = 5.565, 95% CI = 1.379–22.458), FPG (OR = 1.818, 95% CI = 1.027–3.217), SBP (OR = 1.167, 95% CI = 1.081–1.260), and LDL (OR = 2.746, 95% CI = 1.122–6.721) were significant negative predictors. HDL (OR = 2.746, 95% CI = 0–0.432) and 25(OH)D (OR = 0.892, 95% CI = 0.813–0.978) were positive predictors. The equation is as follows:Table 5: Multivariable logistic regression analysis for the predictors of LEAD.

4. Discussion

RBP4 is a molecule found in the circulation, thought to be secreted mainly by adipose tissue and the liver. Increased serum RBP4 levels have been reported in subjects with obesity, IR, and T2DM and in other insulin-resistant states, such as metabolic syndrome and vascular complications of DM [15]. Cabré et al. [16] pointed out that RBP4 would increase in patients of T2DM complicated with coronary heart disease, and when the level of RBP4 increased by 1/4, the risk of cardiovascular disease would increase by 2.5 times. In this study, the level of RBP4 was higher in patients of T2DM with LEAD than in patients of T2DM and healthy people. Therefore, RBP4 is involved in the pathogenesis of T2DM with LEAD. Metheniti et al. [17] reported that 25(OH)D was low in ultra obese young females and it was significantly associated with RBP4 and neutrophil gelatinase-associated lipocalin (NGAL). We found that there was a negative correlation between 25(OH)D and RBP4; the concentrations of RBP4 decreased after vitamin D supplementation in the study. Vitamin D may reduce peripheral IR by inhibiting the expression of peroxisome proliferator-activated receptor-γ (PPAR-γ) and the differentiation from preadipocytes to mature adipocytes [18]. This result suggests that one of the reasons for IR led by vitamin D deficiency in T2DM patients is the regulation of RBP4.The third national health and nutrition examination survey (NHANES-III) found that, compared with residents with normal concentrations of 25(OH)D, residents with low concentrations of 25(OH)D had a higher incidence of peripheral arterial disease [19]. In this study, the level of 25(OH)D in T2DM group was lower than the one in control group, and it was the lowest in T2DM with LEAD group; 25(OH)D was a factor to protect DM patients from the pathogenesis of LEAD. Our findings were consistent with the research of Fahrleitner-Pammer et al. [20] about the relationship between 25(OH)D and T2DM with LEAD. The possible mechanisms of the involvement for 25(OH)D include the following. Vitamin D can promote pancreatic β cells to secrete insulin by adjusting vitamin D receptor (VDR) and vitamin-D-dependent calcium-binding protein (DBP) in pancreatic tissue. It can also reduce IR in peripheral tissues by regulating inflammatory cytokines and inhibiting the expression of PPAR-γ [18]. Besides, it can affect insulin sensitivity by stimulating (PPARδ) and the gene expression of insulin receptor [21]. Calcitriol can regulate rennin angiotensin-aldosterone system (RAAS), inhibit the expression of renin, and adjust the synthesis and secretion of atrial natriuretic peptide. Vitamin D can upregulate the expression of related protein of delaying arterial calcification including vascular endothelial growth factor and matrix metalloproteinase-9. Vitamin D can reduce the occurrence of arterial calcification and atherosclerosis by inhibiting angiogenesis, smoothing muscle cells proliferation, and playing a role in immune regulation with VDR mediation of immune cells. (5) A lack of calcitriol results in an increase in the serum parathyroid hormone (PTH) levels. Excess PTH levels may at least in part promote cardiovascular disease by increasing the cardiac contractility and

myocardial calcification [22]. (6) Vitamin D can improve the inflammation state [23, 24] and reduce chronic inflammatory reaction of the arterial wall [25] by inhibiting the secretion of inflammatory cytokines, such as IL-6, TNF-α, and C-reactive protein (CRP) [26]. (7) Calcitriol can inhibit foam cells formation in vascular wall by reducing acetylation and reduce the level of oxidized LDL in T2DM patients.A 20-year retrospective study has found that [27] the incidence of DM complicated with cardiovascular diseases reduced by 33% in population with 800 IU vitamin D and 1200 mg calcium of daily intake compared to those with 400 IU vitamin D and 600 mg calcium of daily intake, which suggests that vitamin D supplementation might become an effective measurement to prevent the occurrences of T2DM complicated with macrovascular diseases. Major et al. [28] suggested that oral supplementation of vitamin D could regulate blood lipid. In this study, the levels of LDL and RBP4 reduced in T2DM group after the intervention of vitamin D, and, simultaneously, the incidence rate of LEAD in T2DM group also decreased after the intervention of vitamin D. Therefore, vitamin D supplementation can protect T2DM patients from being complicated with LEAD ultimately by improving IR and lipid metabolism and inhibiting inflammation.

5. Conclusion

There were several limitations in our study. The sample size was relatively small and we did not consider the influence of outdoor activities and seasonal variation on the level of vitamin D. However, the study was conducted in winter so that the impact of sun exposure was minimized. In conclusion, the phenomenon of vitamin D deficiency amongst T2DM patients was more and more prevalent, and, with the advantages of low price and validated long-term safety, the supplementation of vitamin D would be one of the intervening measurements for the presentation of T2DM and its vascular complications.

Vitamin D status and peripheral arterial disease: evidence so far

GT Chua, YC Chan, and SW Cheng

Abstract

Introduction

It is now more than 75 years since the discovery of calciferol, known commonly as vitamin D, and of its ability to cure rickets in animals and children.1–3 Recent epidemiologic and clinical studies have indicated that vitamin D may have far more roles than just in the prevention of rickets. Diseases such as osteoporosis, muscle weakness, several types of cancer, diabetes, hypertension, and cardiovascular disease may all result from subtle and chronic vitamin D deficiency.4

Vitamin D is a generic name for a group of fat-soluble steroids of which the two major forms are vitamin D2 (ergocalciferol) and vitamin D3 (cholecalciferol).5 Vitamin D, or calciferol, without a subscript, refers to either D2, D3, or both. In 1919, a British doctor, Edward Mellanby, noticed that dogs fed cod liver oil did not develop rickets and concluded that cod liver oil could prevent the disease.6 In 1921, Elmer McCollum tested modified cod liver oil in which the vitamin A had been destroyed, which also cured sick dogs, so McCollum concluded that the factor in cod liver oil which cured rickets was distinct from vitamin A.7–9 It was called “vitamin D” because it was the fourth vitamin to be named, following an

alphabetical code. The molecular geometry of vitamin D2 was determined in 193210 and that of vitamin D3 in 1936 as a result of ultraviolet irradiation of 7-dehydrocholesterol.11

Go to:

Measurements of serum vitamin D levels

The concentration of 25-hydroxyvitamin D (25-OHD) is commonly used for measurement of serum vitamin D levels, despite the active form being 1,25-OHD. This is because parathyroid hormone will increase renal conversion when there is an insufficiency, causing the active form to be normal or elevated. Therefore, vitamin D status is better reflected by 25-OHD. However, measurement of 25-OHD levels may not be accurate in patients with severe renal diseases (creatinine clearance less than 15 mL/min or undergoing dialysis), as renal conversion of 25-OHD level is reduced. Hence, parathyroid hormone levels should also be taken into account when identifying vitamin D deficiency.12

There is no consensus regarding the cut-off level for vitamin D deficiency. It has been suggested previously that vitamin D deficiency in adults should be defined as a serum 25-OHD level below 50 nmol/L (20 ng/mL).12,13 However, further research suggests that optimal serum 25-OHD status should be maintained above 80 nmol/L (32 ng/mL), because the risk for various medical conditions, including reduced bone density, periodontal disease, colon cancer, hypertension, and impaired lung function, is lowest above this level. The risk of undiagnosed diabetes and impaired glucose tolerance was also found to be lowest when the serum 25-OHD level was above 83 nmol/L; while that of myocardial infarction was lowest when serum 25-OHD was above 107 nmol/L.13

Technically, there are still some uncertainties about the optimal method for measuring vitamin D levels. Multiple methodologies for 25-OHD measurement (radioimmunoassay, high-performance liquid chromatography, and liquid chromatography tandem mass spectroscopy) are available and subject to variability, which causes difficulties in defining the cut-off value for vitamin D insufficiency or deficiency. Therefore, aiming at a higher 25-OHD level seems to be clinically safe and reasonable, as it improves vitamin D status with minimal risks.14 Most laboratories have increased the lower limit of the reference range to 75 nmol/L (30 ng/mL). In addition, 25-OHD level varies with seasons, exposure to sunlight, and dietary intake.15 Therefore, unless issues such as choice of assay, timing of measurement, and reference range are clearly addressed, the optimal method of measuring vitamin D level remains uncertain.

Go to:

Literature search method and results

A literature review using standard databases including PubMed, Medline, and Embase was performed in March 2011, searching for studies on peripheral arterial disease and vitamin D and its analogs performed since 2000. The search was conducted using the keywords “peripheral arterial disease” (PAD) and “vitamin D.” Inclusion criteria were: systematic reviews and meta-analysis, journal reviews, cohort, case-control studies and population studies, and English-language articles where the full text was available. Commentaries and letters to the editor were excluded.

Twenty-three articles were initially identified and 13 of these excluded. Five articles had no full text available, five articles were irrelevant to the topic, two articles were letters to the editor, and one was commentary. The remaining ten articles were included for review. These included journal reviews, longitudinal, and cross-

sectional studies, written in English with full text available. Publications on the relationships between vitamin D and cardiovascular risk factors, cardiovascular diseases, and peripheral arterial disease were carefully reviewed. Other relevant articles were also cross-referenced.

Go to:

Vitamin D deficiency and cardiovascular risk factors

Vitamin D deficiency has been shown in various studies to be associated with an increase in the prevalence of various cardiovascular risk factors.16–19 Vitamin D deficiency is inversely correlated with body mass index, hypertension, diabetes mellitus, and hyperlipidemia (total cholesterol and triglyceride).16 When subjects were categorized into normal (>30 ng/mL or 75 nmol/L), low (16–30 ng/mL or 40–75 nmol/L) or very low (≤15 ng/mL or 37.5 nmol/L) vitamin D level groups, it was also found that there were highly significant graded inverse associations with hypertension, hyperlipidemia, diabetes mellitus, and PAD. In addition, there was an increase in hazard ratios for these cardiovascular risk factors between individuals with very low and normal vitamin D levels, which remained statistically significant after adjustment.17

In a randomized placebo-controlled double-blind trial, Witham et al found that one single dose of high-dose vitamin D2 improves endothelial function, as assessed by flow-mediated dilatation of the brachial artery in response to 5 minutes of forearm occlusion.18 Values were significantly higher in the intervention group at 8 weeks’ follow-up, despite the blood pressures of subjects not being significantly affected during the short-term follow-up period of 16 weeks.18,19

Vitamin D inadequacy has also been shown to be inversely correlated with the proliferation of vascular smooth muscle cells,20,21 inflammatory effects upon arterial endothelial cells, and with increased carotid intima-media thickness.22,23 Vitamin D deficiency is linked to adverse morphological changes in large arteries, even in children.24 Low vitamin D has been shown to affect survival rate independently of vascular calcification and stiffness.25 Calcification can be induced by 1,25-OHD in cultured arterial smooth muscle cells.26 Arterial calcification is common in patients with diabetes and end-stage renal failure with peripheral arterial disease, and paricalcitol, a vitamin D receptor agonist, has been shown to improve survival for renal failure patients in comparison to traditional calcitriol therapy.27 Reik et al found that 1,25-OHD affects macrophages by lowering acetylated and oxidized low-density lipoprotein uptake in type 2 diabetic patients, thus reducing foam-cell formation and subsequently atherosclerosis.28 In addition, 1,25-OHD can induce cell-cycle arrest and apoptosis, an important process in the development of atherosclerosis and in some normal and malignant cell types.29 More recently, vitamin D deficiency has been associated with peripheral vascular disease,30 and may be an important contributory factor in the development of peripheral arterial disease in different ethnicities.

Gaddipati et al reviewed the possible mechanisms and role of vitamin D in vascular health.30 Vitamin D decreases blood pressure through the renin-angiotensin system.31 Vitamin D deficiency is inversely related to vascular smooth muscle proliferation and increase in carotid intima-media thickness. However, excessive vitamin D may do more harm than good, as excessive dietary consumption is potentially angiotoxic.32 Such toxicity can lead to vascular calcification and arterial stiffness. Norman and Powell showed that 1,25-OHD (which is the active form of vitamin D) interacts with vitamin D receptors and directly or indirectly regulates the expression of a number of proteins relevant to the arterial wall, such

as vascular endothelial growth factor, matrix metalloproteinase type 9, myosin, and elastin, and type I collagen. Cultured arterial smooth muscles can also undergo calcification when treated with 1,25-OHD, through a mechanism relying on suppression of parathyroid hormone–related peptide, which is an endogenous inhibitor of calcification, and parathyroid hormone–receptor signaling. Directly or indirectly, vitamin D exposure may downregulate the paracrine mechanism that normally protects the vasculature from calcification.33

Go to:

Vitamin D deficiency and cardiovascular diseases

Cardiovascular diseases and adverse events are significantly higher in subjects with very low vitamin D levels. These conditions include coronary artery disease, heart failure, atrial fibrillation, PAD, stroke, myocardial infarction; the rate of death is also significantly higher in subjects with very low vitamin D levels.17,34,35 In another study, Wang et al also analyzed the relationship between vitamin D and risk of cardiovascular disease, and found that participants with 25-OHD levels < 15 ng/mL (37.5 nmol/L) had a higher incidence of cardiovascular events compared with those with 25-OHD > 15 ng/mL, with a hazard ratio (HR) of 1.62. Similar observations were made in participants with hypertension (HR = 2.13, 95% confidence interval [CI] 1.30–3.48), but there were no significant differences between participants without hypertension (HR = 1.04, 95% CI 0.55–1.96). A graded inverse relationship between cardiovascular risk and 25-OHD level was also observed in this study.36 These results suggest that vitamin D deficiency may potentially become a future risk factor for cardiovascular diseases.

However, these results may not be conclusive. A recent systematic review and meta-analysis was unable to demonstrate a statistically significant reduction in mortality and cardiovascular risks, including stroke and myocardial infarction, but the quality of the available evidence was low to moderate at best.37 However, another systematic review by Grandi et al showed significant results.38 Therefore, clinical association between vitamin D and cardiovascular risks and events should be further investigated and supported by high-quality evidence.

Go to:

Vitamin D deficiency and peripheral arterial disease

Recent studies have suggested that vitamin D may have a direct relationship with PAD.16,39–40 Reis et al performed a cross-sectional study based on the National Health and Nutrition Examination Survey 2001–2004, investigating racial differences in vitamin D levels and the incidence of PAD between black and white populations.39 The authors concluded that Afro-Caribbean populations were at higher risk of vitamin D deficiency, possibly due to greater cutaneous melanin content, lower dietary intake, and racial differences in vitamin D metabolism. They also showed that vitamin D concentrations were significantly lower in blacks that in whites (P < 0.001) and PAD was significantly more prevalent among black adults than white (8.5% and 5.3% respectively, P < 0.05), with the odds ratio being 2.11. After adjustment for demographic factors and cardiovascular risk factors (diabetes, hypertension, hypercholesterolemia, C-reactive protein, estimated glomerulofiltration rate, and history of cardiovascular disease), the odds ratio remained significant at 1.67. In contrast, adjustment for serum 25-OHD concentration yielded insignificant results.

Serum vitamin D level has been shown to have an inverse relationship with the prevalence of PAD in a graded manner.

Melamed et al studied the association between 25-OHD levels and prevalence of PAD, and found that for each 10 ng/mL decrease in 25-OHD, the prevalence ratio of PAD was 1.35 (95% CI 1.15–1.59) after multivariable adjustment. Furthermore, vitamin D deficiency increases the amputation rate in patients with PAD.40 Gaddipati et al showed that individuals with vitamin D deficiency (<20 ng/mL) had a significantly higher amputation rate compared with those who were not defined as vitamin D deficient (6.7% and 4.2%, P = 0.029).16

Go to:

Experimental evidence

Emerging laboratory evidence supports the idea of a direct relationship between vitamin D and PAD. Oral supplementation of vitamin D3 has been shown to suppress atherosclerosis in a mouse model. Takeda et al tested apolipoprotein E-knock-out mice with calcitriol and demonstrated that higher doses of calcitriol reduced the formation of atherosclerotic lesions by 39.1%, which was significantly more than those in the control group. It was shown in the atherosclerotic lesions that more regulatory T-cells (Tregs) and less mature dendritic cells are present compared with the control group.41 Studies have shown that Tregs suppress the inflammatory responses of effector T-cells, which are suggested to be potentially proatherogenic, while dendritic cells stimulate the differentiation of naive T-cells to effector T-cells.21,42,43 In addition, Takeda et al has also demonstrated that suppression of Tregs has resulted in an increase in atherosclerotic lesions. This showed that oral vitamin D3 reduces atherosclerosis by the suppression of inflammatory cells in mice, which warrants confirmation of such benefits in humans.41

All the above findings imply that vitamin D deficiency is an important direct contributor to PAD in a dose-related manner and proper supplementation could possibly reverse the disease. Vitamin D deficiency could also affect subsequent adverse events.

Go to:

Limitations of the current review

There are several limitations to this review. Firstly, most of the studies included were population-based, cross-sectional, or longitudinal studies. Only associations, not causal relationships, could be established between vitamin D and PAD. Although most studies concluded that vitamin D deficiency was a potential risk factor of PAD, it was also sensible to think that vitamin D deficiency could be a complication of PAD as a result of impaired mobility and subsequent lack of sunlight exposure. Secondly, there was a lack of information about certain aspects in these studies, including the geographical location of residence, seasonal variations during which blood samples were obtained, and measurement of parathyroid hormone, which were significant confounding variables. Thirdly, these studies were targeted to Western populations. The associations between vitamin D and PAD in Asians or other ethnic populations, such as those of Arabic descent, remain unclear. Currently, there is still no level-one evidence or any randomized-controlled trials studying whether vitamin D supplementation is useful in reducing the rate or the risk of developing PAD, so a systematic review or meta-analysis of randomized-controlled trials is not yet possible.

Go to:

Conclusion

Emerging studies are showing that there is an association between low vitamin D levels and PAD, with animal models suggesting possible mechanisms. Future studies should focus on various ethnic

populations, and investigate whether vitamin D should be recommended for prevention of PAD, and if so, at what dosage. Only randomized controlled trials will be able to provide the answers.

Serum Vitamin D, Parathyroid Hormone Levels, and Carotid Atherosclerosis

Jared P. Reis,*,a Denise von Mühlen,b Erin D. Michos,c Edgar R. Miller, III,a Lawrence J. Appel,a Maria R. Araneta,band Elizabeth Barrett-Connorb

1. Introduction

Vitamin D and parathyroid hormone (PTH) are widely recognized for their important roles in maintaining extracellular calcium and phosphorous homeostasis and in regulating bone formation and bone resporption. Vitamin D, largely obtained from cutaneous exposure to ultraviolet B radiation (290–315 nm) and to a lesser extent from dietary and supplemental sources, is metabolized by the liver to 25-hydroxyvitamin D [25(OH)D], the primary circulating storage form of vitamin D. 1,25-dihydroxyvitamin D [1,25(OH)2D], produced largely in the kidney following a second hydroxylation, is the tightly regulated activated molecule of vitamin D. PTH, on the other hand, is a peptide hormone secreted by the parathyroid gland in response to low circulating calcium and phosphorus concentrations. PTH stimulates the reabsorption of calcium in the kidney, the resporption of calcium from the skeleton, and enhances the renal production of 1,25(OH)2D.

Evidence from several lines of scientific inquiry suggests vitamin D and PTH may also play a role in the pathogenesis of cardiovascular disease. Hypovitaminosis D has been independently associated with increased rates of hypertension [1], diabetes [2], peripheral arterial disease [3], myocardial infarction [4,5], and related mortality [6,7]. Numerous studies have shown patients with hyperparathyroidism experience an excess risk of mortality from cardiovascular disease [8,9]. High PTH levels have also been independently associated with higher rates of cardiovascular disease in the general population [10]. Vitamin D and PTH may influence cardiovascular risk through a shared association with atherosclerotic plaque formation and progression. However, both vitamin D and PTH have been inconsistently found to be associated with early signs of atherosclerosis, such as increased carotid intima-media wall thickness (IMT) determined with B-mode ultrasound [11–13]. Existing studies have been limited by small clinic-based samples and the failure to examine the combined influence of both vitamin D metabolites and PTH. Therefore, in the current study we aimed to determine the individual and combined associations of 25(OH)D, 1,25(OH)2D, and PTH levels with the extent of carotid artery IMT in a population-based cohort of community-dwelling older adults from the Rancho Bernardo Heart and Chronic Disease Study.

Go to:

2. Methods

2.1 Study population

The Rancho Bernardo Study has been described in detail elsewhere [14]. Briefly, between 1972 and 1974, 82% of all adults living in the southern California community of Rancho Bernardo were enrolled in a study of heart disease risk factors as part of the Lipid Research Clinics Prevalence Study. Nearly all were white and middle to upper-middle class, as assessed by Hollingshead’s index [15]. Between 1997 and 1999, 89% of the local surviving community-dwelling members attended a clinic visit when a medical evaluation was performed and a blood sample was obtained. Of these, approximately 70% agreed to

participate in the current study involving the measurement of carotid artery IMT. Sub-study participants were generally older and more likely to be male than non-participants; however, no differences were observed in 25(OH)D, 1,25(OH)2D, or PTH levels between participants and non-participants.

Because this analysis focused on the role of vitamin D and PTH early in the atherosclerotic process, those with a history of clinical cardiovascular disease, including coronary heart disease, coronary revascularization, or stroke were excluded (n = 103). After further excluding 1 participant who did not have a blood sample adequate for measurement of vitamin D and PTH levels, there remained 654 participants for these analyses. All participants had serum calcium concentrations within the reference range and no participant reported a history of thyroid or parathyroid surgery. The University of California, San Diego Human Subjects Protections Program approved this study and all participants provided written informed consent prior to participation.

2.2 Clinical evaluation

Participants were evaluated at the University of California, San Diego Rancho Bernardo Research Clinic. During the clinic visit, standardized questionnaires were used to obtain demographic information, medical history, and lifestyle behavior information. Data on alcoholic beverage consumption (g/wk during the previous 2 weeks) [16], cigarette smoking (current, former, never), and participation in strenuous exercise three or more days per week (yes/no) were obtained. Current use of diabetes and blood pressure medications as well as vitamin D and calcium supplement use was validated by the examination of pills and prescriptions brought to the clinic for that purpose.

Two morning blood pressure readings were recorded and averaged in seated subjects after a five-minute rest using a mercury sphygmomanometer according to the Hypertension Detection and Follow-up Program protocol [17]. Hypertension was defined by meeting any of the following 4 criteria: systolic blood pressure ≥140 mmHg, diastolic pressure ≥90 mmHg, previous physician’s diagnosis, or medication use. Height and weight were measured with a standard physician’s scale and stadiometer while participants were wearing light clothing and no shoes. Waist circumference was measured at the bending point (the natural indentation when bending side-wards) and at the narrowest circumference. The correlation between these two measures was 98% and the bending point measure was used in these analyses. Hip circumference was measured at the widest circumference below the waist.

2.3 Laboratory methods

Blood was obtained by venipuncture after an overnight fast and placed into tubes that were protected from sunlight. Serum was separated and stored at −70°C within 30 minutes of collection. Serum 25(OH)D [25(OH)D2 + 25(OH)D3] was measured in the research laboratory of M. Holick, using vitamin D competitive binding protein recognition and chemiluminescence detection (Nichols Institute Diagnostics, San Clemente, CA) as described by Chen et al. [18] The rat serum vitamin D binding protein has high affinity for 25(OH)D. The intra- and inter-assay coefficients of variation were 8% and 10%, respectively. The limit of detection was 5.0 ng/mL and the reference range was 10–55 ng/mL. Serum 1,25(OH)2D was measured in the same laboratory using a radioreceptor assay (Nichols Institute Diagnostics, San Clemente, CA) as described previously [19]; intra- and inter-assay coefficients of variation were 10% and 15%, respectively, with a reference range of 20–45 ng/L. The PTH assay was performed in the same laboratory using a chemiluminescence

assay for the measurement of intact PTH (Nichols Institute Diagnostics, San Clemente, CA); intra- and inter-assay coefficients of variation were both 6% with a reference range of 10–65 ng/L.

Plasma total and HDL-cholesterol were measured in a Centers for Disease Control-certified Lipid Research Clinic laboratory. Total cholesterol was measured by enzymatic techniques using an ABA-200 biochromatic analyzer (Abbott Laboratories, Irving, TX). HDL-cholesterol was measured after precipitation of the other lipoproteins with heparin and manganese chloride according to standardized procedures of the Lipid Research Clinics manual [20]. Fasting plasma glucose was measured by the glucose oxidase method. Serum creatinine levels were measured by Smith Kline Beecham clinical laboratories. Glomerular filtration rate, a measure of kidney filtration function, was estimated with the abbreviated Modification of Diet in Renal Disease Study equation, calculated as 175 × (serum creatinine, mg/dL)−1.154 × (age, years)−0.203 × 0.742 (if female) [21]. Diabetes was defined by history, use of diabetes medications, or a fasting blood glucose level ≥126 mg/dl.

2.4 Carotid artery intima-media thickness (IMT)

B-mode ultrasonography of the left and right common carotid arteries and the internal carotid artery was performed, and IMT was determined, as described by O’Leary et al. [22] Four standardized images were obtained bilaterally for each subject: 1 of the common carotid artery, 1 cm proximal to the dilation of the carotid bulb at a lateral angle and 3 of the internal carotid artery at the site of maximal thickness in 3 angles (anterior, posterior, and lateral). Ultrasound measurements were sent to a central reading facility (DH O’Leary, Principal Investigator) where data were processed in a blinded fashion. The common carotid artery IMT score was calculated as the mean of the left and right measurements of the common carotid artery IMT. The internal carotid artery IMT score was determined as the mean of the 6 internal carotid artery IMT measurements.

2.5 Statistical analyses

All analyses were conducted with SAS version 9.1 (SAS Institute, Cary, NC). Carotid IMT and PTH concentrations were positively skewed and were therefore log-transformed or categorized as appropriate, depending upon the type of analysis conducted; geometric means and 95% confidence intervals were calculated. Analysis of covariance was used to examine the association of 25(OH)D, 1,25(OH)2D, and PTH with log-transformed common carotid and internal carotid IMT. Models adjusted for potential confounding factors based upon a review of the relevant literature, including age, sex, smoking status, alcohol intake, waist-to-hip ratio, exercise, and season of blood draw. Additional models adjusted further for potential intermediate factors, including diabetes and hypertension. Tests for a linear trend were performed by entering the quintile categories of 25(OH)D, 1,25(OH)2D, and PTH into the model as an ordinal term. We also examined the association of 25(OH)D, 1,25(OH)2D, and PTH with log-transformed common carotid and internal carotid IMT according to high or low levels of the other vitamin D metabolite and PTH (two categories, defined by the sample median); factors that are thought to influence the concentration of 25(OH)D, 1,25(OH)2D, or PTH; and diabetes or hypertension status. We tested for the presence of interaction by including a product term with the respective stratification variable and the continuous exposure variable in the model. Statistical significance was defined at p<0.05 using two-sided tests.

Go to:

3. Results

The demographic, anthropometric, and clinical characteristics of the study population are presented in Table 1. On average, the 654 adults included in this analysis were aged 75.5±8.5 (range 55–96) years, 64% were women, and the mean body mass index was 25.6±4.1 kg/m2. Most reported a healthy lifestyle, with current smoking in only 4.2%, regular exercise in 74.6%, calcium supplement use in 46.1%, and vitamin D supplement use in 23.6%. Mean 25(OH)D, 1,25(OH)2D, and geometric mean (95% CI) PTH levels were 41.5±14.1 ng/mL, 32.4±17.3 ng/L, and 44.7 (44.2, 47.3) ng/L, respectively. The prevalence of 25(OH)D levels >20 or >30 ng/mL was 96.8% and 83.3%, respectively.

Table 1

Descriptive characteristics of the 654 participants: Rancho Bernardo, CA, 1997–1999.

Table 2 displays the demographic, anthropometric, and clinical predictors of 25(OH)D, 1,25(OH)2D, and PTH levels. Individuals with the highest 25(OH)D levels were younger, had lower systolic blood pressure and body mass index levels, and were more likely to be male, physically active, drink alcoholic beverages more frequently, and use calcium and vitamin D supplements. Those with the highest 1,25(OH)2D levels were also younger and drank alcohol more frequently, but were more likely to be female, have a lower waist-to-hip ratio and total cholesterol:HDL ratio, and higher kidney function. On the other hand, participants with the lowest PTH levels had a lower body mass index and total cholesterol:HDL ratio, and were more likely to exercise regularly as well as use calcium and vitamin D supplements.

Table 2

Characteristics of participants according to quintiles (Q) 1, 3, and 5 of 25-hydroxyvitamin D [25(OH)D], 1,25-dihydroxyvitamin D [l,25(OH)2D], and parathyroid hormone concentration: Rancho Bernardo, CA, 1997–1999.

Table 3 displays the adjusted geometric mean common carotid IMT and internal carotid IMT according to quintiles of 25(OH)D, 1,25(OH)2D, and PTH. Lower geometric mean internal carotid IMT was observed across increasing quintiles of 25(OH)D. These results were independent of age, sex, smoking status, alcohol use, waist-to-hip ratio, physical exercise, and season of blood draw. Further

adjustment for potential mediators of this association, including diabetes and hypertension, did not explain these findings. Adjustment for body mass index as opposed to waist-to-hip ratio or systolic blood pressure rather than hypertension status had little influence on these results. No association was observed between 25(OH)D levels and common carotid IMT, or between 1,25(OH)2D or PTH and the common carotid or internal carotid IMT.

Table 3

Adjusted geometric mean (95% CI) common carotid artery IMT and internal carotid artery IMT according to quintiles (Q) of 25-hydroxyvitamin D [25(OH)D], 1,25-dihydroxyvitamin D [l,25(OH)2D], and parathyroid hormone concentration: Rancho Bernardo, CA, 1997–1999. ...

There was no evidence the inverse association between 25(OH)D and internal carotid IMT differed significantly by age, sex, adiposity, exercise, kidney function, hypertension, diabetes, 1,25(OH)2D, or PTH levels (p for interaction > 0.1, for all). An inverse association between 1,25(OH)2D and internal carotid IMT was observed only among those with hypertension (p for interaction 0.036); for each 1-SD increase in the concentration of 1,25(OH)2D, multivariable-adjusted log-transformed carotid IMT decreased by 0.050 (0.005, 0.095). PTH showed no association with the internal or the common carotid IMT in subgroups defined by age, sex, adiposity, exercise, kidney function, hypertension, diabetes, 25(OH)D, or 1,25(OH)2D (p for interaction > 0.1, for all). In stratified subgroup analyses, 25(OH)D and 1,25(OH)2D showed no association with common carotid IMT.

Go to:

4. Discussion

In this community-based cohort, we found a significant inverse stepwise association between serum 25(OH)D levels and internal carotid IMT, but no association with common carotid IMT. The association with internal carotid IMT was independent of age, sex, and multiple other potential confounding factors. There was no evidence of an association between 1,25(OH)2D, the activated metabolite of vitamin D, or PTH with either common carotid or internal carotid IMT. In only one of several subgroup analyses, an inverse association was found between 1,25(OH)2D and internal carotid IMT among those with hypertension.

The results of previous studies seeking an association between vitamin D and atherosclerosis have been inconsistent. Targher et al. [13] reported an inverse association between 25(OH)D and IMT among 390 diabetic patients, but no association was observed in another study of 109 postmenopausal women [23]. Arad et al. [11] found no evidence of an association between 1,25(OH)2D levels and the extent of coronary artery calcification among 50 patients who had recently undergone coronary angiography for the evaluation of coronary artery disease, valvular heart disease, or both. On the other hand, Watson et al. [12] observed a weak but statistically significant inverse association between 1,25(OH)2D concentrations and coronary artery calcium mass (r = −0.18) among 153 subjects at moderate risk

for developing coronary heart disease according to Framingham Study estimates. Doherty et al. [24] later confirmed this finding among 283 asymptomatic adults at high risk of coronary disease.

Differences in the selection criteria and demography of the study populations may be responsible for the contradictory findings. Factors that contribute to carotid IMT may also be different from coronary artery calcification. The inconsistent findings of studies measuring 1,25(OH)2D may also be due to its tight physiologic control, reported biological half-life of only 4–6 hours, and concentration within the normal range even in the presence of vitamin D deficiency [25]. On the other hand, 25(OH)D has been widely recommended as the best indicator of vitamin D status in individuals without chronic kidney disease because it has a reported half-life of 12–19 days and its levels are closely associated with the amount of sunlight to which the epidermis is exposed as well as dietary intake of vitamin D [25].

Some studies have suggested that 25(OH)D and 1,25(OH)2D are only weakly correlated and therefore might provide unique prognostic information regarding cardiovascular risk. Dobnig et al. [7] recently showed that low levels of 25(OH)D and 1,25(OH)2D were each associated with higher rates of all-cause mortality, even after adjustment for the other vitamin D metabolite and PTH. Furthermore, the mortality risk was greatest among those in the bottom quartile of 25(OH)D and 1,25(OH)2D, suggesting a possible synergistic effect for those deficient in both vitamin D metabolites. We found that the inverse association of 25(OH)D on internal carotid IMT did not differ between those with high or low levels of 1,25(OH)2D or PTH.

We observed an independent association between serum 25(OH)D levels and internal carotid, but not common carotid IMT. The IMT of both the internal and common carotid arteries is strongly associated with incident atherosclerotic cardiovascular disease events even after adjustment for major cardiovascular risk factors [22,26]. However, recent evidence suggests a higher internal carotid IMT may reflect greater atherosclerotic plaque burden, whereas a higher common carotid IMT may be the result of vascular changes in response to shear stresses [27–29]. Thus, the overwhelming effects of endothelial shear stress may mask any potential role that vitamin D may play in regulating the IMT of the common carotid artery.

In subgroup analyses, higher 1,25(OH)2D levels were associated with lower internal carotid IMT values only among those with high blood pressure. We are unsure as to why this association appeared limited to persons with hypertension. Furthermore, we could not confirm a modifying influence of hypertension on the association of the circulating metabolite of vitamin D [25(OH)D] with either internal carotid or common carotid IMT. Admittedly, we cannot rule out that this finding may be due to chance, since we performed numerous subgroup analyses. As discussed earlier, the measurement of the activated metabolite of vitamin D and its interpretation in clinical research studies among apparently healthy persons is complicated by many its physiologic parameters, namely its inability to provide an indication of usual vitamin D status.

There is accumulating evidence that vitamin D may influence vascular function and the development or progression of atherosclerosis. Vitamin D receptors have a broad tissue distribution which includes vascular smooth muscle cells [30,31], macrophages [32], and lymphocytes [33]. Vascular smooth muscle and endothelial cells possess the enzyme 25(OH)D-1 α-hydroxylase, which is responsible for the conversion of 25(OH)D to 1,25(OH)2D [34–36]. Vitamin D induces prostacyclin in vascular smooth muscle cells, which prevents thrombus formation, cell adhesion, and smooth muscle cell proliferation [37]. Furthermore, vitamin D regulates the expression of a number of other proteins relevant to the arterial wall, including

vascular endothelial growth factor, matrix metalloproteinase type 9, myosin, elastin, type I collagen, and γ-carboxyglutamic acid, a protein that protects against arterial calcification [38,39]. Vitamin D suppresses pro-inflammatory cytokines, including interleukin-6 and tumor necrosis factor- α in vitro and in vivo [40]. In mice, vitamin D is an inhibitor of the renin-angiotensin system [41], and vitamin D receptor knock-out mice develop cardiac hypertrophy [42]. Furthermore, a transgenic rat model which constitutively expresses vitamin D-24-hydroxylase, an enzyme responsible for breaking down an activated form of vitamin D thereby causing deficiency, develops aortic atherosclerosis and hyperlipidemia [43].

We found no evidence of an association between PTH concentrations and internal carotid or common carotid IMT in either the entire cohort or in subpopulations. A role for PTH in the development of cardiovascular disease has been suggested by several studies [8–10]. In a recent report, Choi et al. [23] observed a direct association between PTH and carotid IMT independent of established cardiovascular risk factors among postmenopausal women attending an endocrine clinic. Patients with primary hyperparathyroidism have been reported to have increased arterial stiffness, altered vascular reactivity, left ventricular hypertrophy, and valvular calcification [8]. Furthermore, receptors for PTH and PTH-related peptide have been located in the arterial endothelium as well as vascular smooth muscle cells [30].

The high prevalence of sufficient 25(OH)D levels observed in the current study sample are likely due to latitude and the southern California climate which includes an abundance of year-round sun exposure. However, the high 25(OH)D levels may also be due to the competitive binding protein assay used to measure circulating vitamin D levels. At the time the blood samples were measured, high-performance liquid chromatography was more costly and labor intensive. Recent laboratory comparison studies have shown that competitive binding protein assays may produce higher 25(OH)D levels than other available assay methods [44,45]; however, routine assays accurately rank individuals across the range of 25(OH)D levels [45]. Thus, the use of the competitive binding protein assay should not have influenced the association between 25(OH)D and carotid IMT reported here. The magnitude of the association may have been even greater if studied in a population that was predominantly vitamin D deficient, since epidemiologic studies have shown a non-linear association between 25(OH)D and cardiovascular risk and mortality, with the greatest risk conferred by 25(OH)D levels <15 ng/ml [6]. In addition, a single measurement of vitamin D or PTH may not be reflective of lifetime status and subclinical atherosclerosis progresses over many years. We acknowledge the cross-sectional design of our study limits causal inference. Furthermore, the external validity of our findings may be limited to apparently healthy middle- to upper-middle class white adults and may or may not be generalizable to other groups or other geographic areas. While the homogeneity of our sample may limit the external validity of our findings, it minimizes the potential for residual confounding by unmeasured characteristics.

In conclusion, in this population-based cohort of older adults, we found an inverse dose-response association between serum 25(OH)D levels, but not 1,25(OH)2D or PTH levels, with internal carotid IMT, independent of traditional cardiovascular disease risk factors. These findings provide evidence that vitamin D may play a role in the development of atherosclerosis. Our results support additional investigation into whether vitamin D may influence the progression of atherosclerosis as well as the effects of supplementation on the atherosclerotic process.

Vitamin D, Shedding Light on the Development of Disease in Peripheral Arteries

Abstract

Vitamin D is generally associated with calcium metabolism, especially in the context of uptake in the intestine and the formation and maintenance of bone. However, vitamin D influences a wide range of metabolic systems through both genomic and nongenomic pathways that have an impact on the properties of peripheral arteries. The genomic effects have wide importance for angiogenesis, elastogenesis, and immunomodulation; the nongenomic effects have mainly been observed in the presence of hypertension. Although some vitamin D is essential for cardiovascular health, excess may have detrimental effects, particularly on elastogenesis and inflammation of the arterial wall. Vitamin D is likely to have a role in the paradoxical association between arterial calcification and osteoporosis. This review explores the relationship between vitamin D and a range of physiological and pathological processes relevant to peripheral arteries.

The primary role of vitamin D and its active metabolites is maintaining calcium homeostasis by increasing intestinal calcium absorption and, depending on circulating calcium levels, influencing the balance between bone resorption and formation. The physiological role of vitamin D in skeletal and cellular health has been reviewed elsewhere.1 Vitamin D also has effects on the microendocrine systems of the vasculature, some of which have only been appreciated recently. This review reflects on the possible influence of vitamin D on peripheral arterial disease, which includes diseases (both occlusive and dilating) of the abdominal aorta and the distal arteries supplying the lower limb. Atherosclerosis is a major contributor to peripheral arterial disease, but the risk factors are subtly different from those for coronary artery disease: smoking is the dominating risk factor for peripheral arterial disease. Does vitamin D have any influence on the disease process?

Previous Section Next Section

Vitamin D Metabolites and Analogues

Cholecalciferol is a prohormone that is synthesized in the skin by photochemical conversion of 7-dehydrocholesterol (Figure 1). It is subsequently hydroxylated to 25-hydroxycholecalciferol [25(OH)D3] in the liver and finally to the active metabolite, 1α,25 didydroxcholecalciferol [1α,25(OH)2D3] in the kidney.2 Some dietary sources provide a cholecalciferol derivative, with a double carbon-carbon at position 22,23 known as vitamin D2 that also undergoes 1,25 hydroxylation. The active vitamin D metabolites are transported in the circulation by vitamin D-binding protein which has additional effects, including the binding of globular actin and fatty acids and immunomodulation. Because little vitamin D is available in most diets, photochemical synthesis of vitamin D is paramount for its homeostasis, and it is for this reason that rickets became a problem in industrialized cities. 1α,25(OH)2D3 is conformationally flexible and is able to generate a large array of different ligand shapes with numerous effects.3 In addition, it is estimated that >1000 vitamin D analogues with variable biologic profiles have been synthesized.3

Figure 1. The biosynthetic pathway of vitamin D.

Actions of Vitamin D in Target Cells and Tissues

Genomic Effects1α,25(OH)2D3 is a steroid hormone shown to regulate >60 genes.2,4 This is accomplished by the translocation of 1α,25(OH)2D3 into cells where it binds with high affinity to the vitamin D receptor (VDR), which is a member of the nuclear receptor superfamily. This complex then interacts with the vitamin D response elements in the promoter region of target genes thereby altering rates of gene expression (Figure 2).5 By this pathway, 1α,25(OH)2D3 influences a number of genes relevant to the arterial wall, including vascular endothelial growth factor, matrix metalloproteinase type 9, myosin, and structural proteins, such as elastin and type I collagen.4–8 In addition, there is evidence of an alternative pathway for 1α,25(OH)2D3 altering gene transactivation through intracellular vitamin D-binding proteins.9 The newly recognized use of 1α,25(OH)2D3 analogues as immunomodulatory agents is based on the ability of these analogues to influence gene expression in cells of the immune system and cytokine expression by other cells (reviewed by Mathiew and Adorini10).

Figure 2. Genomic and nongenomic responses to1α,25 dihydroxycholecalciferol (genomic responses based on Barsony et al and Brown et al5,13).

Nongenomic EffectsIn common with other steroid hormones, 1α,25(OH)2D3 induces a range of effects which occur too rapidly to involve gene expression. These include increases in intracellular calcium and cGMP levels, activation of protein kinase C and changes in phosphinositide metabolism11 (Figure 2). The effects are known to be mediated by 1 or more plasma membrane receptors, but their role is unclear in most cell types.2,11 An example relevant to the arterial wall includes

the stimulation of vascular smooth muscle cell (VSMC) migration through activation of phosphatidylinositol 3-kinase.12

Target Cells and Tissues for 1α,25(OH)2D3

Classically, the action of 1α,25(OH)2D3 is to maintain calcium and phosphate homeostasis, with the intestine and bone being key targets. It also acts on a wide range of nonclassical target tissues, including the heart and arterial wall.13 The influence of 1α,25(OH)2D3 on these tissues could have important implications for vascular function and disease. 1α,25(OH)2D3 appears to cause cell cycle arrest and inhibit proliferation of most cell types, including lymphocytes.13 VDRs are present in VSMCs. There is controversy concerning the action of vitamin D on VSMCs, with some studies reporting stimulation of proliferation and others reporting inhibition of proliferation and a synthetic phenotype being induced by 1α,25(OH)2D3.14,15The direction of the effect may depend on the culture conditions, with effects in vivo being uncertain.16 VDRs have been identified in dermal capillaries, and cultured endothelial cells appear to express a 1α-hydroxylase enzyme, indicating the possibility of a vitamin D microendocrine system in endothelial cells.171α,25(OH)2D3 causes apoptosis in tumor endothelial cells, interferes with vascular endothelial growth factor signaling, and suppresses angiogenesis.18 Macrophages and lymphocytes are other important target cells for vitamin D in the diseased artery wall.19 The vitamin D microendocrine systems of the diseased arterial wall are shown in Figure 3.

View larger version: In this page

In a new window

Figure 3. VDR-expressing cells of the diseased artery wall and their known responses to vitamin D. Cells with blue cytoplasm are known to express VDR, with the density of the blue corresponding approximately to the probable density of VDR expression; T lymphocytes have the highest density of VDR. Lymphocytes are shown as T or B. IEL indicates internal elastic lamina; EEL, external elastic lamina.

Previous Section Next Section

Dietary Requirements and Human Consumption of Vitamin D

After the discovery that cod liver oil prevented rickets in 1919, overall consumption of vitamin D in Europe and North America was probably excessive during the first half of the 20th century.20 It was not unusual for infants to consume 100 μg per day and an “epidemic” of infantile hypercalcemia and supravalvular aortic stenosis coincided with the increased vitamin D supplementation of milk in the 1940s. It was not until 1963 that infant food supplementation was reduced in the United States.21 Sensitivity to vitamin D appears to vary because not all infants developed hypercalcemia, and a reduced susceptibility to rickets may be associated with an increased susceptibility to vitamin D toxicity.21 Cod liver oil may have disappeared from infant food supplements, but a significant proportion of adults in Western countries continue to use fish oil supplements. A recent dietary study in southwest England indicated that ≈20% of healthy adults were using fish oil supplements.22

Dietary recommendations for vitamin D consumption vary among advisory bodies, reflecting uncertainty about vitamin D status and health (particularly osteoporosis).23 Recommended daily intake

ranges from 0 in those at low risk of osteoporosis to 15 μg per day for those at high risk (>65years, dark skin, restricted exposure to sunlight).23 Some authorities feel that fear of toxicity has tended to keep the recommended daily intake at “woefully inadequate” levels.24

Previous Section Next Section

Epidemiological Studies of Vitamin D and Arterial Disease

Despite a number of articles over the last 50 years or more arguing that vitamin D is a risk factor in atherosclerosis,25,26 the published data, at least with respect to ischemic heart disease, remain conflicting. Linden initially reported that patients with myocardial infarction appeared to have a higher intake of vitamin D than controls.27 This has been followed by reports that serum levels of 25(OH)D3 were not elevated in patients with myocardial infarction,28,29 and the mortality from ischemic heart disease (in postmenopausal women) was not associated with intake of vitamin D.30 Scragg et al even reported an inverse relationship between levels of 25(OH)D3 and myocardial infarction.31 Ironically, vitamin D deficiency [serum levels of 25(OH)D3 <9 ng/mL] because of immobility causing lack of sunlight exposure has been reported in patients with peripheral arterial disease.32

Similar uncertainty surrounds the relevance of vitamin D in hypertension (an established risk factor for peripheral arterial disease33). Despite evidence from various animal models that vitamin D may be important in blood pressure control and hypertension,34 its clinical importance remains unclear. There is an inverse relationship between exposure to sunlight (needed to synthesize cholecalciferol) and both blood pressure and prevalence of hypertension.35 Some studies have found a similar inverse relationship between levels of 1α,25(OH)2D3 and blood pressure and plasma renin activity,36,37 although others report a positive association.38 Recent results from a large population-based study failed to find any convincing association between serum vitamin D concentrations and blood pressure.39 Vitamin D supplementation appears to lower blood pressure in those with preexisting hypertension.40 Li et al have recently shown this to be mediated through the direct actions of 1α,25(OH)2D3 in suppressing renin expression.41Critical assessment of the role of vitamin D in blood pressure control is beyond the scope of this review. However, in the context of blood pressure control, vitamin D may provide some protection against peripheral arterial disease.

Previous Section Next Section

Vitamin D and Peripheral Arterial Calcification

Arterial calcification is important because it has been shown to be a forerunner of cardiovascular events. The molecular biology of arterial calcification will not be described in detail as it has been the subject of recent reviews.42,43 However, the possible role of vitamin D in arterial calcification will be discussed.There are 2 distinct patterns of arterial calcification: calcification of the media (Monckeberg’s sclerosis, seen in aging, chronic renal failure, and diabetes) and calcification of the intima (seen in atherosclerosis). Intimal calcification has attracted considerable attention, particularly in the context of the prognostic significance of coronary artery and aortic arch calcification.44,45 However, most of the increase in arterial calcium with age is concentrated in the medial layer.46 This medial calcification is not usually occlusive or associated with atherosclerotic plaque but is nevertheless a predictor of lower limb amputation and cardiovascular mortality.47,48 Medial calcification also results in incompressible arteries and difficulty in measuring true ankle pressures, which can complicate the noninvasive diagnosis of peripheral arterial disease. An inverse relationship between serum 1α,25(OH)2D3 levels and total (intimal and medial) coronary artery

calcification has been reported.49,50 This association may depend on medial calcification, whereas another study failed to demonstrate an association between intimal calcification and serum 1α,25(OH)2D3 levels.51 The significance of an inverse relationship between levels of 1α,25(OH)2D3 and coronary calcification is uncertain, particularly as levels of 25(OH)D3 are a better indicator of vitamin D status.24

There is increasing evidence of a paradoxical association between osteoporosis and vascular calcification.52–54 The mechanisms underlying this association are beginning to be unraveled42,55–57 and may account for the inverse association between coronary artery calcification and serum levels of 1α,25(OH)2D3.49 Various inhibitors of bone resorption, including biphosphonates (alendronate and ibandronate), osteoprotegrin, and an inhibitor of osteoclastic V-H+-ATPase (SB 242784) have been shown to inhibit calcification of the arterial media in animal models.55,56,58 Because none of these agents are known to act directly on the arterial smooth muscle cells, it has been proposed that arterial calcification is directly linked to bone resorption in this model.56 1α,25(OH)2D3 may, however, act directly on smooth muscle cells or osteoclast-like cells within the arterial wall.59,60

Medial calcification is common in diabetes and end-stage renal failure, both conditions being associated with peripheral arterial disease. In renal failure, 1α,25(OH)2D3 analogues are used to prevent secondary and tertiary hyperparathyroidism, and treatment with vitamin D has been implicated in calcification of soft tissues, including the arterial wall.61 Recently a large observational study has indicated that a selective VDR antagonist (paricalcitol) improves survival for renal failure patients, by 16% to 25%, in comparison to traditional calcitriol therapy.62 In support of this finding, in vitro experiments show that VSMCs undergo calcification when treated with 1α,25(OH)2D3 through a mechanism dependent on suppression of an endogenous inhibitor of calcification (PTH-related peptide) and PTH receptor signaling.59,63 Hence, vitamin D exposure may downregulate the paracrine mechanisms that, under normal circumstances, protect the vasculature from calcification. This has led to recent speculation that inflammation of the vascular adventitia with local synthesis of 1α,25(OH)2D3 by macrophages could lead to medial calcification.64

1α,25(OH)2D3, through its interaction with VDR, can induce the calcification of cultured arterial smooth muscle cells.59 Apoptosis, which is well-documented in atherosclerosis and after arterial injury, may provide an initial stimulus for calcification within the arterial wall.65,66 1α,25(OH)2D3 can induce cell cycle arrest and apoptosis in some normal and malignant cell types.18,67 Although 1α,25(OH)2D3 has been shown to inhibit angiogenesis by induction of apoptosis, there is no direct evidence linking vitamin D to peripheral arterial calcification through this mechanism.Although adequate vitamin D nutrition is essential for optimal vascular function,68both exogenous and endogenous 1α,25(OH)2D3 are possible axes for the association. In contrast to endogenous vitamin D, which is carried by circulating vitamin D-binding protein, exogenous vitamin D may be carried by lipoproteins.69This may facilitate accumulation of vitamin D within atherosclerotic plaque and alter macrophage gene expression.19,70,71

Previous Section Next Section

Arterial Disease and the Genomic Effects of Vitamin D: Clinical

Studies

The relationship between common polymorphisms of the VDR gene and osteoporosis has attracted considerable research (and controversy) over the last decade.72,73 The link between osteoporosis and arterial calcification has led to the hypothesis that functional polymorphisms in the VDR gene are associated with the prevalence and severity of coronary artery disease. This has been tested in clinical populations, and although a possible association between the

VDR BB genotype (resulting in decreased levels of vitamin D) and severity of coronary artery stenosis has been reported, the relationship was not statistically significant.74 The importance of such studies depends on whether they are large enough to determine effects of the magnitude attributed to the polymorphism(s) or haplotypes in vitro. There are contradictory results reported as to whether the BsmI polymorphism in intron 8 or the FokI polymorphism in the 5 regulatory region of the VDR gene has ′functional effects in vitro. Therefore, even the results of large studies focusing on a single polymorphism must be treated with skepticism. Although Ortlepp et al75 report that the BsmI polymorphism has no influence on the incidence of coronary artery disease in 3441 patients, such evidence does not refute the possibility that genetic variation in response to vitamin D has an impact on the development of arterial disease. There have been no studies of the role of the genomic effects of 1α,25(OH)2D3 in the development of peripheral arterial disease. Genomic associations with the more relevant clinical phenotype of arterial calcification have not been evaluated.

Previous Section Next Section

Nongenomic Effects of Vitamin D on Peripheral Vascular Resistance

Although the evidence concerning vitamin D exposure and hypertension is controversial, it is possible that vitamin D influences vascular tone. The single study that has investigated the short-term effect of vitamin D on cardiovascular hemodynamics showed that vitamin D caused rapid changes in patients with essential hypertension but not in controls.76 In patients with essential hypertension, the cardiac output decreased by ≈15% within 2 hours of intravenous administration of 1α,25(OH)2D3 (0.2 μ/kg). This was associated with transient smaller increases in pulse rate and mean blood pressure, suggesting that 1α,25(OH)2D3 had a nongenomic effect to increase peripheral resistance. This would be supported by an earlier report that indicated a correlation between calf vascular resistance, both before and during reactive hyperemia, with serum concentrations of 25(OH)D3 in hypertensive patients.77 The ability of 1α,25(OH)2D3to increase vascular resistance is supported by animal studies.78 1α,25(OH)2D3increases the sensitivity of resistance arteries to norepinephrine in hypertensive but not normotensive rats79 and rapidly enhances arterial force generation by modulation of intracellular calcium concentration.34,80 Taken together these studies suggest that hypertension induces or sensitizes putative plasma membrane VDRs which modulate intracellular calcium concentrations and hence resistance artery force generation. The identity of these plasma membrane receptors is obscure, as are the downstream kinase signaling cascades. In the absence of hypertension, these fast nongenomic responses have not been observed.Even in the presence of hypertension, in the longer term, the fast nongenomic responses may be counteracted by the genomic effects of 1α,25(OH)2D3 which may function to decrease vascular resistance. Likely mechanisms include altered expression of myosin isoforms in resistance vessels.81 This would be consistent with the reported decrease of systolic blood pressure after 8 weeks of oral calcium and vitamin D supplementation in the late winter.40

Previous Section Next Section

Vitamin D, Elastin, and Inflammatory Vascular Disease

Several years ago the hypothesis was elaborated that impaired synthesis of elastin in the walls of the aorta and other elastic or conduit arteries during fetal development was an initiating event in the pathogenesis of hypertension.82Because tropoelastin synthesis is known to be downregulated by vitamin D, through a posttranscriptional mechanism,6 excess vitamin D during fetal development could cause the impaired synthesis of elastin discussed

by Martyn and Greenwald. Loss of medial elastin is a pathological hallmark of abdominal aortic aneurysm (AAA). This led to Norman elaborating a further hypothesis, that excess vitamin D consumption in early life led to the development of AAA in later life.83 The transportation of vitamin D across the placenta is specifically enhanced during the last one third of pregnancy (a period of maximal aortic elastin deposition).84,85 The neonate is dependent on stored vitamin D, because mammalian milk contains minimal vitamin D.86 This suggests that if maternal intake is excessive then fetal and neonatal exposure will also be excessive. Although there are no clinical studies to support an association between increased maternal vitamin D intake and impaired aortic elastogenesis, an experimental animal study demonstrated that exposure to increased vitamin D in early life was associated with a reduction in elastin content and elastic lamellae number in the abdominal aorta.87 The studies were not extended to investigate the possibility of aneurysm formation in aging rats.Another pathological hallmark of AAA is inflammation.88 Interestingly, there are recent scientific observations to support an important role for vitamin D and its analogues on the immune system, particularly macrophages and T lymphocytes expressing VDR, which could have important implications for both AAA and other peripheral arterial disease. The highest expression of VDR is in CD8 lymphocytes, with less expression in CD4 lymphocytes and macrophages, whereas B cells do not express VDR. Moreover, VDR expression in CD8 cells increases in response to 1α,25(OH)2D3.89 Therefore, vitamin D can regulate cytokine expression in diseased arteries (Figure 3). Laboratory data suggest that 1α,25(OH)2D3 influences cytokine production by both CD4+ and CD8+ subsets, preferentially inhibiting cytokine production (interleukin [IL]-2 and interferon-γ) from Th1 cells and hence favoring Th2 responses with the production of IL-4, IL-5, and IL-10, as well as IL-6.90,91 It is these findings that might translate to a link between vitamin D and AAA, a condition where Th2 responses predominate.92

Clinical studies of the effects of vitamin D supplementation are limited. In postmenopausal women, vitamin D supplementation (2 μg per day) increased CD3 and CD8+ subsets of lymphocytes.93 Therefore, vitamin D may influence T cell activity and inflammation of the artery wall through several different pathways.Wjst and Dold94 have hypothesized that deficiency of vitamin D in early life leads to allergic diseases in later life. Allergy, autoimmune deficiency, and transplant rejection are all controlled by Th1 responses, inflammatory vascular disease being important in transplant rejection. Although there is abundant experimental evidence (eg, Raisanen-Sokolowski et al95) to indicate that vitamin D supplementation reduces transplant rejection, there are no clinical studies or trials of this phenomenon. Instead vitamin D supplementation is used to target posttransplantation osteoporosis.Activated macrophages express 1α-hydroxylase and produce 1α,25(OH)2D3.96This may have a role in limiting the extent of local inflammation64,97 but also has the potential to alter smooth muscle cell migration and proliferation in the vessel wall. Moreover, 1α,25(OH)2D3 influences the function of macrophages, with subsequent effects on the production of alkaline phosphatase by cultured smooth muscle cells, promoting calcification.19 These findings, together with the observation that vitamin D supplementation increases serum transforming growth factor type β levels,98 provide other mechanistic possibilities for a more widespread influence of vitamin D on peripheral arterial disease.

Previous Section Next Section

Vitamin D and Animal Models of Atherosclerotic Arterial Disease

The harmful effect of excess vitamin D on arteries has been studied in many animal models over the last 40 years or more.99,100 Numerous dosage regimens have been described although the majority have

used short courses of potentially toxic doses of vitamin D resulting in acute hypercalcemia. Chronic less toxic treatment also results in metastatic calcification and deteriorating renal function.101 In general, vitamin D results in arterial wall calcification and a variety of other “arteriosclerotic” changes. Loss of collagen and disruption of elastic lamellae are additional features.102 These latter changes are usually associated with increased aortic stiffness,102 although one study reported a paradoxical reduction in stiffness.103 Vitamin D also exacerbates the intimal hyperplasia seen in balloon-injured rat carotid arteries.104 This is probably caused by the stimulation of migration12 and proliferation of smooth muscle cells.79,105,106 Recently, the combination of vitamin D2 and cholesterol has been used to induce both peripheral atherosclerosis and aortic valve stenosis in a rabbit model.107 In this model, the addition of vitamin D2 to the high cholesterol diet resulted in significantly higher levels of circulating cholesterol.107 A combination of vitamin D and nicotine, causing hypercalcemia, results in stiffer rat conductance arteries and exacerbates the atherosclerotic effects of cholesterol feeding.102,107 The relevance of any of these models to the development of cardiovascular disease in humans is uncertain.Another possible link between vitamin D and peripheral arterial disease has been found in a study of transgenic rats that constitutively express vitamin D-24-hydroxylase [which catalyzes the conversion of 25(OH)D3 to 24,25(OH)2D3].108These rats had low levels of plasma 24,25(OH)2D3) and developed hyperlipidemia and aortic atherosclerosis.108 This is an example of the complexity of the relationship between differing vitamin D metabolites and the arterial wall.

Previous Section Next Section

Summary

In addition to its role in calcium and phosphate homeostasis, vitamin D is important in many physiological and pathological processes relevant to peripheral arterial disease. Vitamin D is essential for the development and maintenance of a healthy arterial tree. 1α,25(OH)2D3 influences the migration, proliferation, gene expression of VSMCs, elastogenesis, and immunomodulation, all processes which are involved in the pathogenesis of atherosclerotic and aneurysmal arterial disease. 1α,25(OH)2D3 has additional, poorly understood, nongenomic effects on vessel contractility in essential hypertension and is likely to have a pivotal role in the paradoxical association between osteoporosis and vascular calcification. New clinical studies are required to fully understand the contribution of vitamin D to the health of peripheral arteries.