Vitamin D, the so-called sunshine vitamin, is in fact not really a vitamin but a hormone which the body can make using sunlight. Historically [1], vitamin-D deficiency was associated with the childhood disease of rickets characterized by severe growth retardation and the bending or bowing of the legs. Rickets was presumably a product of the Industrial Revolution with a high level of urbanization and child labor resulting in minimal exposure to the sun. Severe vitamin-D deficiency also caused some young women to have a deformed pelvis with the resultant difficulty in birthing, which incidentally gave rise to the practice of Cesarean sections. The suggestion that rickets was due to a lack of sunlight was advanced in 1822, but it was not until 1919 that a cure attributed to exposure to radiation from a mercury arc lamp gave strong support to this hypothesis. By the 1930s and 1940s the fortification of food with synthetically made vitamin D was popular. This was long before the photochemistry of the cutaneous (in the skin) production of vitamin D and the biochemistry and action of its metabolites were understood. With the almost complete disappearance of rickets, there was little interest in the possibility of residual or sub-clinical deficiency. Only recently has a serum marker for the vitamin D status been validated, and there has been renewed interest in the possibility of vitamin D deficiency and its implications which is quite recent and is in part due to the modern understanding of the multiplicity of biochemical actions of vitamin D metabolites. Today, research on the role of vitamin D metabolites in health and illness has gone well beyond their role in calcium homeostasis and bone health. They are implicated in cancer prevention, hypertension, rheumatoid arthritis, multiple sclerosis, and early-onset diabetes (type 1).

It is the nature of the human species that most of the vitamin D required is generated by the action of the sun. Natural food sources are very limited and provide only small amounts unless large quantities of oily fish are eaten. Humans are thought to have evolved in equatorial Africa and to have migrated from this area only about 80,000 years ago [2]. The dark skin of our ancestors is thought to have been a protective feature, reducing the destruction of folate by UV light, protecting the sweat glands from damage and increasing reflectivity of solar energy [3,4]. Even though dark skin reduces the efficiency of vitamin D production, our hunter-gatherer ancestors were exposed to very high levels of UV throughout most days. Migrations eventually took groups into the northern latitudes of Europe and Asia where the dark skin was a disadvantage because of reduced vitamin D production due to low winter levels of UV radiation. The lighter skin color associated with those living in northern latitudes that subsequently evolved [3,4] provided better utilization of solar UV and presumably allowed greater buildup of vitamin D stores during the summer months. The effect of severe vitamin D deficiency on both children and females of child-bearing age provides a plausible mechanism for selection and adaptation.

Compelling evidence-based arguments can now be made that many individuals have only marginal serum levels of the critical metabolite of this vitamin, and that deficiency is present in a significant fraction of the world's population [1,5,6]. This deficiency results from inadequate sun exposure and from the failure to eat large enough quantities of dark, oily fish and, where available, fortified foods. This review will examine a number of questions concerning vitamin D, its role in health, the dangers of deficiency, the need for supplements, and the currently expanding appreciation of its importance in biochemical processes other than those related to calcium homeostasis.

SOURCES AND METABOLISM OF VITAMIN D [1,5]

Humans acquire vitamin D from the action of sunlight and from food. The skin contains a cholesterol derivative, 7-dehydrocholesterol (provitamin D), which ultraviolet light (UVB, 290-315 nm) converts to vitamin D which is then either stored in body fat or converted in the liver to 25-hydroxyvitamin D, which we will denote as 25(OH)D. Vitamin D from dietary sources is also converted in the liver to 25(OH)D. Circulating 25(OH)D is converted, mostly in the kidney, to another derivative, 1,25(OH)2D, also called calcitriol, or vitamin D hormone, which regulates serum calcium and phosphorus levels by controlling the intestinal efficiency of absorption. Many tissues and cells in the body have receptors for vitamin D hormone, and it has been recognized for at least two decades that this hormone is a potent inhibitor of cellular proliferation and an inducer of cell maturation. This may have very important implications in connection with the incidence and progression of cancer. Vitamin D hormone receptors are known to exist, for example, in breast, prostate and colon tissue.

There are two forms of vitamin D, D2 and D3. Vitamin D3 is also called cholecalciferol, whereas vitamin D2 is called calciferol or ergocalciferol. The same conversion is used for both to convert from grams to International Units (IU), i.e. 100 IU = 2.5 micrograms (µg). However, these two forms are thought to have different biological activity, with D3 having between 1.7 and 2 times the conversion efficiency to 25(OH)D for approximately equivalent amounts [7]. However, this area remains uncertain and it is common practice not to differentiate between the two forms. Vitamin D2 is not a natural component of human biochemistry but can be manufactured, for example, by UV irradiation of a lipid extracted from yeast. Thus its existence in fortified food and therapeutic prescriptions is mainly for the sake of synthetic convenience. Supplements may contain either form, and sometimes this is not clear from the label. Typical supplement users probably consume 200-800 IU/d.

Fish are the primary natural food source of dietary vitamin D (the D3 form), with 100 grams of herring or salmon providing 1000 IU or 640 IU respectively. A teaspoon of cod liver oil provides about 400 IU, an egg only about 100 IU [5]. For those who consume only limited amounts of these foods, fortified foods and sunlight are the only sources. If one avoids fortified dairy or cereal products, and in addition minimizes exposure to the sun, deficiency becomes a real possibility. Babies who are nourished exclusively by nursing must get their vitamin D from the mother's milk or from sun. Breast milk is a very poor source of vitamin D and if sun exposure is limited, serious deficiencies can develop. A rebound is in fact being seen in the incidence of rickets [8], even in the US. In addition, the fear of skin cancer has promoted the extensive use of sunscreens which essentially eliminates any solar vitamin D generation. A sunscreen SPF of 8 reduces vitamin D3 production by about 98% [1]! To put these numbers in perspective, consider that an adult with white skin wearing a bathing suit generates about 10,000 IU of vitamin D3 in 15-30 minutes when exposed to the summer sun [9]. This is 25-50 times what is in the typical multivitamin. Lengthy sun exposure does not produce toxic levels because vitamin D is also photolabile and as it builds up it is converted (also by UVB) to compounds that do not lead to bioactive metabolites.

It may surprise some readers to learn that in the northern latitudes (>35°-40°N) the amount of UVB in sunlight is low to negligible in the winter months, except at higher altitudes, and contrary to popular belief, sunbathing in the winter in Boston or Edmonton does not generate significant Vitamin D [10]. The same is true in latitudes below about 35°S. Even the sunny French Rivera and Spain have low levels of UVB in the winter. The latitude effect is caused by increased light scattering and ozone absorption due to the tilt of the earth's axis. Thus there is a large and expected seasonal variation of vitamin D status in many populated regions. A number of correlations of latitude with disease incidence have been reported which my be due to vitamin D deficiency [1]

ESTABLISHING DEFICIENCY AS WELL AS HEALTHY LEVELS OF VITAMIN D

To establish daily requirements and the prevalence of deficiency, it is desirable to have a marker, ideally a blood marker. The concentration of vitamin D3 in the blood turns out to be uninformative. The consensus today is that the serum concentration of the metabolite 25(OH)D is the most informative measure of the vitamin status and should be used to define deficiency, sufficiency and perhaps toxicity [11]. Most labs offer this test. Given this consensus on a marker, the challenge is to establish a level below which deficiency exists and a level for optimum health, and to relate these levels to vitamin D intake, both orally and from sun exposure. A number of different approaches have been used.

• The level at which secondary hyperparathyroidism is evident. When 25(OH)D is low, there is a decrease in vitamin D hormone and thus a decrease in calcium absorption and a lower serum calcium concentration. This causes the parathyroid hormone (PTH) serum concentration to increase, and this in turn increases vitamin D hormone production. This keeps the vitamin D hormone concentration nearly constant at the expense of a higher PTH level. This is called secondary hyperparathyroidism (primary hyperparathyroidism generally involves parathyroid gland tumors). Serum calcium may still be within the reference range. The increased PTH level causes increase bone turnover and bone loss. Thus secondary hyperparathyroidism has been proposed as the principal mechanism connecting vitamin D deficiency with the pathogenesis of decreased bone mineral density and the risk of hip fracture in the elderly, and this disorder can also precipitate or exacerbate osteoporosis [6]. As the level of 25(OH)D continues to drop, PTH levels can double or triple. Numerous studies of the relationship between the levels of PTH and 25(OH)D in the blood reveal that at 25(OH)D levels below 50 to 100 nmol/L (nM), the PTH level begins to increase [12-14]. There is no consensus and those trained in the physical sciences would be alarmed at the scatter in some of the data, but it appears that the majority of researchers favor a range between 75 and 100 nM as the threshold below which secondary hyperparathyroidism begins, and this then establishes a threshold for deficiency. Rickets and osteomalacia are seen at levels below about 20 nM, so there is a big gap between the onset of hyperparathyroidism and the level just high enough to prevent what some call the reference disease.
• Keeping in mind our sun-drenched primitive ancestors in Equatorial Africa, some guidance regarding healthy levels of 25(OH)D can be gleaned from the following data based on studies of people living and working in sun-rich environments [14]. Farmers in Puerto Rico were found on average to have levels of 135 nM, whereas lifeguards in St Louis came in at 163 and lifeguards in Israel at 148 nM. Levels over 200 nM have been found in sun-exposed individuals. Those taking vitamin D supplements were excluded from these studies.
• Studies connecting calcium absorption with the serum levels of 25(OH)D in postmenopausal women suggest a level above about 80 nM is desirable [15].
• In two well accepted studies showing fracture prevention with vitamin D and calcium, mean 25(OH)D levels exceeded 100 nM [14].
• In a clinical report [16] on 15 patients with confirmed osteomalacia, the average 25(OH)D level was 13.5 nM with a range of 5-35 nM. PTH was also 3 to 10 times above the reference range.

Other studies could be quoted [1,5,6] but the point is clear that levels of 25(OH)D in the range of 75-100 nM can be justified as probably desirable for optimum health [6,14,17]. Note that there is some variation between laboratories as well as between various assay methods. The results obtained using two commonly used assay methods have been found to differ, on average, by about 30% [18].

THE PREVALENCE OF VITAMIN D DEFICIENCY[19]

There have been a large number of studies concerning the prevalence of low levels of 25(OH)D, some of which are summarized below to provide an indication of the widespread nature of the problem. Studies generally use vitamin D supplementation of 200 IU/d or more as grounds for exclusion, and frequently set a cut-off for deficiency at between 35 and 50 nM 25(OH)D with severe deficiency below 20 nM. There is no general agreement on nomenclature (deficiency, insufficiency, hypovitaminosis D, etc.) or precise cut-off values, but this does not change the picture that emerges.

• In a study [20] of hypovitaminosis D (low levels of vitamin D) in medical inpatients, 290 consecutive patients hospitalized in a general medical service at Massachusetts General Hospital were selected, 150 in March and 140 in September (when the marker would have been expected to be at its maximum). Using 37.5 nM as a cut- off, 57% were found to be vitamin D deficient and of these 22% had serum levels of 25(OH)D below 20 nM. No significant differences were found in the March and September groups. In a subgroup of 77 patients less than 65 years of age without risk factors for hypovitaminosis D, 43% were vitamin D deficient.
• Nesby-O'Dell et al [21] found that 42% of African American women in the US aged 15 to 49 had 25(OH)D levels below 35.7 nM and were described as having hypovitaminosis D.
• Tangpricha et al [22] reported that 32% of healthy young white men and women in Boston aged 18 to 29 were deficient at the end of the winter of 1999, with levels below 50nM.
• Centenarians living in Parma or Mantove, Italy (latitude 43°N) having no acute diseases were studied [23] and it was found that 99 out of 104 had 25(OH)D levels below the sensitivity of the test used (<5 nM).
• Plotnikoff and Quigley recently reported [24] a study of patients presenting with persistent musculoskeletal pain. Elderly patients refractory to the usual therapy had a high prevalence of vitamin D deficiency (<50nM). It is interesting that 90% of the 150 consecutive patients had been evaluated for their persistent muscuskeletal pain one year or more before the study and yet none were tested for vitamin D deficiency!
• Vieth et al [25] describe a study involving 796 young women (18-35 years) over one year in Toronto, Canada (latitude 43°N). During this period, the prevalence of low 25(OH)D of <40 nM was 25.6% for non-white, non-black subjects and 14.8% in white women. Of the 435 women studied during the winter half of the year, the prevalence of low 25(OH)D was independent of vitamin D intake up to 200 IU/d.

Many more studies could be listed[1,5,6], but the point is clear. Deficiency appears widespread in all age groups, but especially in the elderly. If a cut-off of 75 nM for 25(OH)D, one threshold suggested above, had been used in these and other studies, the prevalence of deficiency would have been much higher. As might be expected, black skinned individuals have the biggest problem followed by Hispanics [19]. In cultures where most of the skin is covered when the individual is outdoors, significant to severe vitamin D deficiency is routinely found. Finally, it is of considerable interest that 200 IU/d in general had an insignificant effect of serum 25(OH)D levels, and yet this dose is the currently recommended adequate intake for young persons.

VITAMIN D INTAKE TO ASSURE THE ABSENCE OF DEFICIENCY

One might think that the current recommendations for vitamin D intake were designed to ensure adequacy. To quote Reinhold Vieth and Donald Fraser of the University of Toronto [9], "In fact, the current recommendations for vitamin D are not designed to ensure anything. They are simply based on the old, default strategy for setting a nutritional guideline, which is to recommend an amount of nutrient similar to what healthy people are eating." The recommended daily allowance for vitamin D does not in fact as yet exist, and instead recommendations are referred to as "adequate intake" (AI). The AI for young adults was chosen to approximate twice the average vitamin intake reported by 52 young women in a study from Omaha, Nebraska in 1997. The use of the term AI is in fact an admission of the weak nature of the evidence used by the Food and Nutrition Board of the US Institute of Medicine. The current AI for young adults is 200 IU, for adults 400 IU and for the elderly, 600 IU/d. These recommendations assume some input from solar generated vitamin D, but as we have seen, this is highly variable.

There have been a number of studies concerning the relationship between vitamin D intake and serum 25(OH)D levels [14]. To keep the levels of this metabolite above 75-100 nM, a total daily intake of about 4000 IU from all sources is required [14]. This translates into adequate sun exposure in the summer months to maintain high summer levels and build up stores of vitamin D, plus supplements. To avoid undesirably low concentrations of serum 25(OH)D, all adults are encouraged by Vieth and Frazer [9] to take 1000 IU of vitamin D3 per day. In fact, at least 25 studies show that 800 IU/d results in an average 25(OH)D level of <80 nM [14], and Vieth et al [17] found that 1000 IU/d resulted in an average 25(OH)D level of about 70 nM. Increasing the dose to 4000 IU is predicted by Vieth et al [17] to yield an average of about 90 nM. For housebound elderly and others with almost no sun exposure at any season, 1000 IU/d would appear to be well below optimum, and 600 IU appears totally inadequate. In connection with the AI of 400 IU/d, Holick [26] found that even a dose of 600 IU/d was insufficient to maintain normal 25(OH)D levels for nuclear submariners submerged for 3 months. The view that the AIs are unrealistically low and that a daily oral intake of about 1000 IU/d is indicated has been put forward by others as well [6,27-30]. Obviously, the biggest problem for the concerned individual is to balance solar generation and supplementation. Fortunately, virtually unlimited solar generation appears safe, aside from skin cancer considerations.

TOXICITY

The maximum suggested dose currently is 2000 IU/d according to guidelines from the 1997 Food and Nutrition Board. Vieth argues in a reply to a letter by Munro [31] that this is unrealistically low. Toxicity has never been observed in cases where the high circulating 25(OH)D is derived from sunlight, and amounts can reach 235 nM, which is vastly more than 2000 IU could generate. If one sunbathes until the skin just shows a slight pink result, the estimated generation of vitamin D is equivalent to an oral intake of between 10,000 and 20,000 IU [14]. Therapeutic oral doses of 50,000 IU, generally D2, are available by prescription and are used to treat severe vitamin D deficiency. In a French study published in 2001 [32], three oral doses of 100,000 IU each of D3 were administered to male adolescents at the end of September, November and January, an intervention which maintained their March 25(OH)D levels at summer values of about 55 nM, as compared to controls that dropped to 20 nM. No side effects were observed. In a recently reported study [33], 2037 men and 649 women received an oral dose of 100,000 IU of D3 every four months for five years to test the hypothesis that there would be a beneficial effect on the incidence of fractures as well as mortality. Both were significantly reduced and no adverse effects were observed. In the study by Heaney et al [34], up to 10,000 IU/d resulted in no adverse effects, including hypercalcemia, and the subjects were carefully monitored because of the high doses used.

Toxic doses of vitamin D are described as producing vitamin D intoxication, which is generally accompanied by high or dangerous levels of serum calcium, i.e. hypercalcemia. There are only a few reports in the literature. The case of vitamin D poisoning reported in The Lancet in 2002 involved prolonged, accidental daily consumption by both a father and his son of >1,700,000 IU/d (this is not a misprint) from contaminated table sugar that occurred over a period of seven months [35]. In another case [36] the patient presented and was hospitalized with symptoms of hypercalcemia of a few weeks duration and was found to have a serum level of 25(OH)D of over 1200 nM! Analysis of the vitamin D supplement provided by the patient and an additional sample obtained from the company involved indicated a huge manufacturing error resulting in a daily dose of vitamin D of between 156,000 and 2,600,000 IU/d. It is not known how long this dose had been taken. Other cases [35] of toxicity have involved huge excesses of vitamin D added accidentally to milk, or where industrial concentrates of vitamin D were mistaken for cooking oil. Thus, it is impossible to make a case for toxicity even at levels well above 2000 IU/d. The reports of vitamin D intoxication have involved doses that were, by comparison, astronomical.

VITAMIN D AND PREVENTION OF CANCER

The reader is also referred to the review by Hans Larsen in the IHN Research Report Vitamin D and Cancer. Suspicion that there was a cancer-vitamin D connection was prompted by observations that the risk of some cancers varied with the latitude. As more became known about the metabolism of vitamin D and the actions of its metabolites, it was proposed that at lower latitudes there would be a higher circulating concentration of 25(OH)D and thus higher concentrations of 1,25-dihydroxyvitamin D (vitamin D hormone), which was known to be extremely potent in inhibiting cell proliferation. The trouble with this simple theory was that the serum concentration of vitamin D hormone is tightly regulated, and its concentration does not change significantly with UV exposure or oral supplementation. Only 25(OH)D changes dramatically. However, there is a way around this objection [37]. It has been found that a number of cell types have the enzyme capability to convert 25(OH)D to the vitamin D hormone, and thus serum and cellular levels of 25(OH)D could influence the intracellular concentration of vitamin D hormone, bypass the control mechanisms, and thus provide a connection between the cellular concentrations of these two metabolites. It is now know that a wide variety of normal tissues as well as various cancer cells, including breast, prostate and colon, can make vitamin D hormone from 25(OH)D. Epidemiologic studies designed to investigate the cancer-vitamin D hypotheses have had mixed success. We will very briefly examine positive results reported for colorectal, breast and prostate cancer, the only sites that have received significant attention.

COLORECTAL CANCER. By the mid 90s there was already considerable interest in the connection between vitamin D, calcium and colorectal cancer, but studies on humans had yielded inconsistent results [38]. Over the next eight years a number of intervention, case control and prospective studies were reported [39-48] with the majority providing evidence of an inverse relationship between vitamin D status or intake and either primary colon cancer or the recurrence of polyps in the colon. In addition, an inverse risk relationship has generally but not always been found with calcium intake. Grau et al [43] recently published the results of an important trial that was in part motivated by the reported association between 25(OH)D serum levels and the risk of colorectal cancer. They found in a randomized placebo-controlled clinical trial with both male and female subjects that vitamin D status strongly modified the effect of calcium supplementation on adenoma (polyp) recurrence. Calcium supplementation (1200 mg/d) lowered adenoma recurrence risk only among subjects with 25(OH)D levels above the median of the cohort, which was about 75 nM. What is noteworthy is the high median level of this vitamin D status marker. The highest value in the cohort was 91 nM. In another recently published study, Lieberman et al [48] examined a cohort of mostly men aged 50-75 who had completed a colonoscopy. Advanced neoplasia was found in 329 out of 1441 participants. When those with advanced neoplasia were compared to the total cohort as a function of vitamin D intake, an inverse association was found with apparent dose dependence and an odds ratio of 0.61 for intakes of greater than about 645 IU/d. These results are consistent with those reported by McCollough et al [47] on participants in the Cancer Prevention Study II Nutrition Cohort (60,886 men, 66,883 women). Vitamin D intake in this study was associated with reduced risk of colorectal cancer only in men, with an adjusted rate ratio of 0.58 for total vitamin D intake of greater than 525 IU/d and a highly significant trend for the rate ratio between this intake and <110 IU/d. They also found that calcium modestly reduced the risk of colorectal cancer. In a large prospective study based on two cohorts, one from the Nurses' Health Study, the other from the Health Professionals Follow-up Study, Wu et al [44] reported an inverse association between high total calcium intake (>700 mg/d) and distal (left sided) colon cancer, but only in participants with highest intake of vitamin D (median intake for the highest third the two cohorts was between 529 and 610 IU/d). Thus the picture emerges that calcium and vitamin D work together in this context, and while both are implicated in the incidence of colorectal cancer, the risk reduction is greater at high calcium intakes and a high vitamin D status, either determined from intake or by measuring 25(OH)D levels. Readers interested in the mechanistic and genetic aspects are referred to a review by Lamprecht and Lipkin [49].

BREAST CANCER. A possible connection between vitamin D and breast cancer was suggested over 10 years ago because of studies connecting incidence or mortality with the amount of solar radiation available. In the US, breast cancer rates in the Northeast sector are approximately twice those found in the southwest [50]. Dietary consumption factors are very similar in the four quadrants of the US [50]. The connection with sunlight was confirmed in a recent study by Freedman et al [51] where they found that mortality for breast cancer (and colon cancer) was negatively associated with both residential and occupational sunlight exposure. Janowski et al [52] found that after adjusting for confounding factors, the odds ratio for the lowest relative to the highest quartile of vitamin D hormone in a case control study (156 cancer cases, 184 controls) was 5.2. This is surprising, considering that serum vitamin D hormone is normally tightly controlled. Shin et al from Harvard in a large prospective study based on the Nurses' Health Study data base [53] have examined the connection between the incidence of breast cancer and calcium and vitamin D. It was found that both dietary calcium and vitamin D were inversely associated with breast cancer in premenopausal but not in postmenopausal women. It was, however, not possible to separate the effects of vitamin D and calcium. The strongest association with vitamin D was with the total intake including cutaneous production. This study is consistent with the NHANES I Epidemiologic Follow-up Study [54] where risk reductions between 0.35 and 0.75 were found for women who lived the US in regions of high solar radiation. This study took into account vitamin D intake from sunlight, diet and supplements. The failure of Shin et al to find a vitamin D effect in postmenopausal women is puzzling, since the other studies described included both pre and postmenopausal subjects.

PROSTATE CANCER [55]. The north-south gradient in prostate cancer mortality and the greater risk for prostate cancer among dark-skinned individuals are reminiscent of rickets and suggest that one of the causes of prostate cancer initiation or progression might be vitamin D deficiency. A large case control study reported in 2000 supports this hypothesis [56]. In a follow-up of 19,000 males, 149 prostate cancer cases were matched with 566 controls. Men with 25(OH)D levels below the median of 40 nM had an adjusted relative risk of l.7 compared to men with this marker above the median. Also, the prostate cancer risk was highest among men younger than 52 years of age (pre so-called andropause) with low serum 25(OH)D levels. They had a relative risk of 3.5! Another case control study [57] examined the relationship with sunlight exposure in some detail, and found significant odds ratios favoring the hypothesis that the protection from prostate cancer was exposure dependent. In recent studies where no connection with vitamin D was found, the majority of subjects had intakes below 600 IU/d. Chen and Holick, [55] in their recent review of vitamin D and prostate cancer prevention and treatment, suggest annual testing of 25(OH)D levels. In connection with the basic mechanism involved, Chen et al [58] have recently demonstrated that primary cultures of prostate cancer cells and prostate cancer cell lines show a marked decrease in activity of the enzyme that converts 25(OH)D to vitamin D hormone, with attendant loss of the growth-inhibitory activity of this hormone.

Since vitamin D and calcium are frequently taken together, the matter of the connection between prostate cancer and calcium arises. There have been a number of studies, both of supplement and dairy intake, with mixed results. Some indicated increased risk at high calcium intake. In a case control study [59], 605 diagnosed prostate cancer cases were compared with randomly selected controls. No effect of total calcium intake on incidence of localized prostate cancer was observed, with the highest quintile cut-off of >1163 mg/d. For cancer that had already spread outside the prostate or metastasized at the time of diagnosis, total calcium as a risk factor appeared above 518-850 mg/d. A large and very recent longitudinal (cohort) study [60] which was corrected for a number of confounding factors, found no connection up to a bit over 2000 mg/d. Over 96% of the cases were Stage B (organ confined). The Health Professionals Follow-up Study [61] found that an intake greater than 2000 mg/d was significantly associated with risk of advanced or metastatic cancer, but calcium intake near the recommended daily intake (1000 mg/d) was not significantly associated with total prostate cancer risk. To quote Patrick Walsh, a very well known urologist at Johns Hopkins Medical School, from an editorial in 2003 [62] concerning calcium supplements, "I have always found this association to be worrisome when advising patients about their dietary intake of calcium. I think now I can relax. Patients with moderate calcium intake (700 to 800 mg per day) are at no increased risk." Finally, it has recently been suggested that the positive connection between high levels of milk consumption and prostate cancer may in fact be partly due to its estrogen content rather than calcium [63].

VITAMIN D AND OTHER HEALTH PROBLEMS

BONE HEALTH. There have been a large number of studies relating vitamin D deficiency to bone health. For example, Mazquita-Raya et al [64] showed that deficiency (less than 40 nM 25(OH)D) in otherwise healthy postmenopausal women was a common risk factor for osteoporosis associated with increased bone remodeling and low bone mass. Dawson-Hughes et al [65] found that for both men and women over 65 years of age, supplementation with calcium (500 mg/d) and vitamin D (about 700 IU/d) moderately reduced bone loss over a three year period. Feskanich et al [66] in a study of 72,000 postmenopausal women (Nurses' Health Study cohort) found that adequate intake of vitamin D (highest risk reduction for greater than 500 IU/d) was associated with lower risk of osteoporotic hip fractures. They found that supplementation or dark (oily) fish consumption was the only satisfactory preventive measures, and that neither milk nor a high-calcium diet appeared to reduce risk. Nguyen et al [67] in a review titled Osteoporosis: Underrated, Underdiagnosed and Undertreated, examined the evidence that vitamin D and calcium supplementation can reduce hip fractures, particularly in institutionalized and housebound elderly and recommended supplements. Other studies could be quoted, but these, all very recent, make the point.

HYPERTENSION. Seasonal and geographic variations of blood pressure have been recognized for some time [68], leading to the hypothesis that variations in vitamin D photosynthesis results in diminished vitamin D levels and increased parathyroid hormone secretion which may result in higher blood pressure. In 1998 Krause et al [69] used full-body UV radiation (3 times a week over six weeks in February and March) on 18 patients with untreated mild essential hypertension randomized to UVB or UVA (longer wavelengths—the controls) to examine this question. Significant decreases in systolic and diastolic BP (average 6 mm Hg, range1- 14) were observed in the UVB but not the UVA group. In the UVB group, 25(OH)D increased by 160% from 58 to 151 nM and there was a 15% fall in PTH. Serum calcium, phosphorous and vitamin D hormone levels were unchanged. Both groups initially were deficient, with a substantial fraction having 25(OH)D levels below 50 nM.

What appears to be the first randomized, placebo-controlled double-blind trial investigating the effect of vitamin D and calcium supplementation on blood pressure was reported in 2001 [70]. The study involved elderly women. They received either 1200 mg calcium or 800 IU vitamin D3 per day or both for 8 weeks. Vitamin D plus calcium was found to be much more effective in lowering BP than calcium alone, with the former yielding a drop from 144 to 131 mm Hg on average for systolic and 85 to 78 for diastolic pressure. The average initial values of 25(OH)D were quite low (about 25 nM) and increased in the calcium plus vitamin D treatment to 65 nM. These studies are consistent with earlier work [71], including a study of the relationship between hypertension and bone-mineral loss in elderly women [72]. The authors cautiously conclude that inadequate vitamin D could play a contributory role in the pathogenesis and progression of hypertension and thus cardiovascular disease in elderly women.

DIABETES. Studies regarding Type 1 diabetes (insulin dependent diabetes) describing a latitude dependence and inverse dependence of incidence on mean monthly sunshine hours suggest that vitamin D might play a protective role, and a deficiency might favor the development of this form of diabetes. European studies [73-75] have provided some measure of confirmation of this hypothesis, but in only one was there any dose information. In that study [75] large daily doses (2000 IU/d) during the first year of life were found to confer protection against the later development of Type 1 diabetes over the ten year duration of the study. In another study, the use of cod liver oil during pregnancy was associated with lower risk of Type 1 diabetes in offspring, but no dose levels were given. It is interesting that suspected rickets early in life was found to yield a risk enhancement of about 3 times for developing Type 1 diabetes [75]. Most of these studies suffer from the failure to measure serum 25(OH)D levels, to characterize maternal vitamin D status or intake from all sources during the first year, or to examine dose dependence in general. This latter aspect is particularly critical since 2000 IU is ten times more than current US guidelines indicate as appropriate for children [76]. While there has been some work on possible mechanisms [77], most investigators suggest that vitamin D is acting as an immune system modulating agent which might inhibit autoimmune processes targeted against the beta cells of the pancreas. Further work, especially on the epidemiology, is clearly needed.

RHEUMATOID ARTHRITIS. Like Type 1 diabetes, rheumatoid arthritis (RA) can be considered an autoimmune disease. Vitamin D has been shown in animal models to have immune modulating effects, and this was part of the motivation for a study just reported that found vitamin D intake inversely associated with the risk of developing RA [78]. Almost 30,000 women aged 55-69 were followed for about 10 years in the Iowa Women's Health Study. An adjusted relative risk of developing RA was 0.66 for supplement users taking 400 IU/d or more of vitamin D. Unfortunately, the design of the study prevented clinical examination, the determination of serum 25(OH)D or sunlight exposure. Iowa is above 40°N and thus one would expect a fairly strong seasonal variation in vitamin D status in this age group.

PEDIATRIC ASPECTS IN GENERAL. The American Academy of Pediatrics (AAP) has recently revised their guidelines [76] regarding the prevention of rickets and vitamin D deficiency. Their recommendations are based on data indicating that 200 IU/d will prevent physical signs of vitamin D deficiency and maintain serum levels of 25(OH)D at or above 27.5 nM [76]. Since rickets has been seen [79] at 25(OH)D levels as high as 22 nM this is in keeping with the traditional philosophy of recommended levels that "just avoid the disease." The AAP recommends as an adequate intake, 200 IU/day for the following: (a) all breastfed infants unless they were weaned to at least 500 mL/d (about 2 cups) of vitamin D fortified formula or milk; (b) all non-breastfed infants who are ingesting less than 500 mL/d of vitamin D fortified formula or milk; (c) children and adolescents who do not get regular sunlight exposure, do not ingest at least 500 mL/d of vitamin D fortified milk, or do not take a daily multivitamin supplement containing at least 200 IU of vitamin D. Note that human breast milk contains about 25 IU/L, or less, and all milk and formulas sold in the US should have at least 400 IU/L. These guidelines do not caution against the low UVB in the winter sunlight at northern latitudes, nor do they focus on dark-skinned mothers whose children, if exclusively fed breast milk, account for the majority of cases of rickets currently being seen in the US [79]. However, the AAP in 1998 recommended 400 IU/d for deeply pigmented breastfed infants or those with inadequate exposure to sunlight [80]. Establishing "adequate" sunlight exposure is difficult for both parents and health care providers, and there is the strong recommendation from the AAP [81] that childhood exposure to sunlight be severely limited because of skin cancer concerns and that adequate sunscreens be used at all times, thus effectively eliminating generation of vitamin D by this normal route. A compromise solution, although not suggested by the AAP guidelines, might involve limiting sun exposure to that known to approximately provide adequate vitamin D levels and use sunscreen or protective clothing for all other exposure. It is estimated [80] that for infants and small children, 30 min of exposure per week in just a diaper, or 2 hours exposure per week if fully clothed with no hat is sufficient. African Americans possibly need six times as much exposure [82]. Presumably no one is going to recommend that all children have their 25(OH)D levels measured. More detailed studies are clearly needed. Also the target value for 25(OH)D that is known to be optimum for this age group appears to need investigating. It may very well be significantly above the AAP's 27.5 nM, especially if in adults it is 75-100 nM.

MULTIPLE SCLEROSIS. Munger et al have conducted the first large prospective study of vitamin D intake and the incidence of multiple sclerosis (MS), the results of which have just been published [83]. This study was based on cohorts from two Nurses' Health Studies totaling over 187,000 subjects. It was motivated by reports that the incidence of MS was latitude dependent and that lesion activity as judged by MRI studies was inversely correlated with vitamin D status. For those women who used supplemental vitamin D at levels equal to or greater than 400 IU/d, they observed a 40% lower risk of MS compared to women who did not use supplements. While they were unable to separate the effect of vitamin D from multivitamin use, an earlier study found that higher intakes of dietary carotenoids, vitamin C and vitamin E failed to reduce the risk of MS in women [84]. Thus, they favor the interpretation that involves vitamin D status.

CONCLUSIONS

It seems clear that anyone who is not paying attention to vitamin D status, either for themselves or for patients, is indeed ignoring the evidence. While much research remains to be done, and not all studies have provided positive results, the number of health issues that appear to relate to vitamin D status provides a strong incentive for being concerned. It should be clear that: (a) there is considerable evidence of rather widespread vitamin D deficiency; (b) numerous studies indicate the importance of maintaining high levels of serum 25(OH)D in order to optimize health; and (c) the importance of vitamin D transcends its role in bone health and calcium metabolism. The government mandated fortification of dairy products and cereals is indicative of a general awareness in public health circles of the importance of this vitamin, at least as regards to bone health. But because of the variation in eating patterns, geographical location of residence, sun exposure, and fear of skin cancer, becoming deficient may merely involve following the path of least resistance, since the alternative is to estimate intake from food and supplements, pay attention to levels of fortification, and estimate generation from sunlight, actions that take effort and some knowledge. Furthermore, there is a common opinion among health care professionals that since rickets is rare, there is no vitamin D problem. There is also the commonly held opinion that we get everything we need from food. Neither of these positions appears defensible.

Obviously, no one has ever taken a large group of presumably healthy subjects, kept their 25(OH)D levels above, say 75 nM for twenty years, and observed the results. Thus the vitamin D intake and 25(OH)D level for truly optimum long-term health is a matter of conjecture. The consensus among researchers as regards to the sensible level of supplementation appears to be about 1000 IU/d for adults, based mainly on keeping 25(OH)D levels high throughout the year. This is to be compared to the current recommendation of 400 IU/d with an increase to 600 IU/d for the elderly. From what is now known about toxicity, 1000 IU/d should not be a cause for concern. However, the intake should not be increased by increasing the number of multivitamin pills taken daily, since this may produce undesirable levels of, for example, vitamin A, which incidentally antagonizes calcium response to vitamin D [85]. It should also be clear that a high level of summer sunlight exposure builds up stored reserves, and the expected drop in the winter, especially in the northern (or southern) latitudes can be countered by supplementation. Concerns about skin cancer can be minimized by the practice of short but frequent exposure, e.g. 15-30 min in full summer sun, which can be very effective in producing a large amount of D3, followed by use of a sunscreen or protective clothing, although not everyone agrees that sunscreens are a good idea (see IHN Research Report Sunscreens: Do They Cause Skin Cancer, by Hans Larsen). Several reviewers and researchers [86,55] have suggested that 25(OH)D should be routinely measured annually, probably in mid-winter, in order that deficiency is not missed. This could be especially important for those with infrequent sun exposure, northern latitude residence, infrequent ingestion of fortified foods and no or low supplementation. Finally, it appears important to also pay attention to optimum calcium intake in the context of maximizing the benefits of adequate vitamin D.

REFERENCES