|Home | About | Journals | Submit | Contact Us | Français|
Vitamin D deficiency, which classically manifests as bone disease (either rickets or osteomalacia), is characterized by impaired bone mineralization. More recently, the term vitamin D insufficiency has been used to describe low levels of serum 25-hydroxyvitamin D that may be associated with other disease outcomes. Reliance on a single cutoff value to define vitamin D deficiency or insufficiency is problematic because of the wide individual variability of the functional effects of vitamin D and interaction with calcium intakes. In adults, vitamin D supplementation reduces the risk of fractures and falls. The evidence for other purported beneficial effects of vitamin D is primarily based on observational studies. We selected studies with the strongest level of evidence for clinical decision making related to vitamin D and health outcomes from our personal libraries of the vitamin D literature and from a search of the PubMed database using the term vitamin D in combination with the following terms related to the potential nonskeletal benefits of vitamin D: mortality, cardiovascular, diabetes mellitus, cancer, multiple sclerosis, allergy, asthma, infection, depression, psychiatric, and pain. Conclusive demonstration of these benefits awaits the outcome of controlled clinical trials.
BMD = bone mineral density; CI = confidence interval; 1,25(OH)2D = 1,25-dihydroxyvitamin D; HR = hazard ratio; HRT = hormone replacement therapy; NHANES = National Health and Nutrition Examination Survey; 25(OH)D = 25-hydroxyvitamin D; OR = odds ratio; PTH = parathyroid hormone; RCT = randomized controlled trial; RR = relative risk; WHI = Women's Health Initiative
The past decade has seen renewed interest in the sunshine vitamin, vitamin D, because new data suggest that its benefits extend beyond healthy bones. Accompanying this renewed interest has been a proliferation of published studies related to the effects of vitamin D in many varying clinical conditions. This article discusses the definition of vitamin D insufficiency, identifies the sources of variation in vitamin D status, reviews the evidence for the clinical benefits of vitamin D, and recognizes indications for vitamin D testing.
Representative studies were selected to highlight some of the limitations of current knowledge related to vitamin D insufficiency and the clinical benefits of vitamin D. We selected studies with the strongest level of evidence for clinical decision making related to vitamin D and health outcomes from our personal libraries of the vitamin D literature and from a search of the PubMed database using the term vitamin D in combination with the following terms related to the potential nonskeletal benefits of vitamin D: mortality, cardiovascular, diabetes mellitus, cancer, multiple sclerosis, allergy, asthma, infection, depression, psychiatric, and pain. The level of evidence was assessed with the following hierarchy: meta-analyses of randomized controlled trials (RCTs), RCTs, nonrandomized intervention studies, meta-analyses of observational studies (cohort and case-control studies), and observational studies.1
The road to the discovery of vitamin D began with recognition of the childhood bone disease of rickets. The first formal medical treatise on rickets was published by Francis Glisson in 1650, when it was identified as a new disease that was more frequent in the rich than in the poor. During the industrial revolution of the 1800s, the prevalence of rickets increased dramatically, ranging from 40% to 60% among children in crowded and polluted urban areas. In 1822, Sniadecki was the first to recognize and report the association of rickets with a lack of sunlight exposure. By the mid-1800s, cod liver oil had been established as an effective treatment for rickets. The work of Mellanby and McCollum led to the discovery of vitamin D as the agent in cod liver oil that had antirachitic properties. This discovery eventually led to the fortification of milk and other foods with vitamin D in the 1930s, and as a result rickets all but disappeared in North America and Europe.
The terminology related to the biochemistry of vitamin D can be confusing. Vitamin D has 2 forms and several metabolites. The 2 forms are vitamin D2 and vitamin D3, also called ergocalciferol and cholecalciferol, respectively. Vitamin D3 is produced in the skin in response to ultraviolet B radiation from sunlight or can be obtained from the diet (ie, animal sources such as deep sea fatty fish, egg yolks, or liver) or from supplements. Few foods naturally have substantial vitamin D content, and dietary vitamin D is obtained primarily through fortified foods or supplements. Vitamin D2, which is found in some plants in the diet and is produced commercially by irradiation of yeast, is used for fortification and supplementation. Both vitamin D2 and vitamin D3 can be used for supplementation.
Both forms of vitamin D undergo identical metabolism (Figure). Some evidence indicates that vitamin D2 may be metabolized more rapidly than vitamin D3,2,3 but with regular daily intake they can be considered bioequivalent.4,5 Both forms of vitamin D are converted to 25-hydroxyvitamin [25(OH)D] in the liver, and the serum level of 25(OH) D is measured to determine the adequacy of vitamin D status. In the kidney, 25(OH)D is hydroxylated to 1,25-dihydroxyvitamin D [1,25(OH)2D], which is the only biologically active form of vitamin D. Acting principally on the duodenum, 1,25(OH)2D increases calcium absorption. It also acts on bone cells, both osteoblasts and osteoclasts, to mobilize calcium.
The characteristics of 1,25(OH)2D are those of a hormone, and consequently vitamin D is a prohormone rather than a true vitamin. The structure of 1,25(OH)2D is similar to that of other steroid hormones. As long as sunlight exposure is adequate, 1,25(OH)2D can be produced by the body without the requirement for ingestion in the diet. Like other hormones, 1,25(OH)2D circulates at picogram concentrations that are 1000 times less than those of the precursor 25(OH)D. Based on the need for increased calcium absorption, the synthesis of 1,25(OH)2D is tightly regulated and stimulated primarily by serum parathyroid hormone (PTH), as well as low serum calcium or phosphorus levels, and inhibited by circulating FGF23 produced by osteocytes.6 Although produced in the kidney, 1,25(OH)2D acts at a distance in the intestinal cell to increase calcium absorption or in the bone to stimulate differentiation and activation of osteoblasts and osteoclasts.7
Determination of vitamin D status is not based on measurement of serum 1,25(OH)2D concentrations. Vitamin D status is assessed by measuring the prohormone 25(OH) D, which is an indicator of supply rather than function. The most stable and plentiful metabolite of vitamin D in human serum, 25(OH)D has a half-life of about 3 weeks, making it the most suitable indicator of vitamin D status. In the past, vitamin D deficiency was identified by the presence of bone disease, either rickets or osteomalacia. Bone disease caused by vitamin D deficiency is associated with serum 25(OH)D values below 10 ng/mL (to convert to nmol/L, multiply by 2.496). More recently, the term vitamin D insufficiency has been used to describe suboptimal levels of serum 25(OH)D that may be associated with other disease outcomes. Precisely defining vitamin D deficiency or insufficiency on the basis of 25(OH)D values is still a matter of much debate. A useful but rather simplistic classification of vitamin D status is shown in the Table. A cutoff value of 30 ng/mL is sometimes used for optimal vitamin status. On the basis of measured concentrations of 25(OH)D, many patients are given a diagnosis of vitamin D deficiency or insufficiency when most have no evidence of disease.
As discussed in detail in recent reviews,8,9 investigators have considered various functional measures to assess the adequacy of vitamin D status. One functional definition of optimal vitamin D status is the 25(OH)D level that maximally suppresses PTH secretion, because the major stimulus for PTH secretion is a low level of serum ionized calcium. In adults, multiple cross-sectional examinations of the relationship between serum PTH and 25(OH)D levels demonstrate a plateau in suppression of PTH when the 25(OH)D level reaches approximately 30 ng/mL.10 This is the rationale for selecting 30 ng/mL as the cutoff value for defining optimal vitamin D status. However, this definition represents an average value at a population level but does not account for the wide variation in the 25(OH) D level that represents adequacy at an individual level. Many patients have very low 25(OH)D values without evidence of increased production of PTH, and conversely, 25(OH)D levels greater than 30 ng/mL do not guarantee PTH suppression.10 Another limitation of this definition is that, in children, an elevated PTH level does not indicate inadequate vitamin D status and has been associated with increased calcium absorption.11 In puberty, the PTH concentration increases, which may stimulate increased periosteal bone formation and increased bone accrual. In fact, preliminary evidence suggests that, with adequate calcium intake, a high-normal PTH level and low-normal 25(OH)D level may result in greater bone size and mass during puberty.12
Another method used in some research studies for defining optimal vitamin D status is the 25(OH)D level at which there is no incremental increase in 1,25(OH)2D levels after administration of vitamin D, because the level of 1,25(OH)2D is adequate to meet demand.13-15 Similar to the findings related to PTH in adults, an incremental increase in the level of 1,25(OH)2D was observed after administration of vitamin D in children when values of 25(OH)D were less than 25 to 30 ng/mL.5 In situations of very low calcium intakes, some evidence suggests that the demand for 1,25(OH)2D may be greater.14,16,17 Thus, vitamin D requirements may vary based on customary calcium intake.
Another functional measure of vitamin D status is the 25(OH)D level that results in maximal intestinal calcium absorption. By combining the results of 3 studies in adults, Heaney18 concluded that optimal calcium absorption occurred at 25(OH)D levels of 32 ng/mL or greater. In contrast, another study found no association between 25(OH)D levels and calcium absorption in healthy women.19 Fractional calcium absorption was high (>50%) in Nigerian children with presumed dietary calcium deficiency rickets and low dietary calcium intakes despite low normal serum 25(OH)D concentrations.5,20 After vitamin D administration and a marked increase in 25(OH) D and 1,25(OH)2D concentrations, fractional calcium absorption did not increase any further.5 In these studies in children, fractional calcium absorption was not related to serum 1,25(OH)2D levels either before or after vitamin D administration. In a study of adults attending an osteoporosis clinic, concentrations of 1,25(OH)2D and intestinal calcium absorption did not appear to decline until 25(OH) D concentrations fell to 4 ng/mL or less, a level that is generally considered to be indicative of severe vitamin D deficiency.21
More recently, the criterion for optimal vitamin D status has moved away from being defined as the 25(OH) D concentration needed to achieve skeletal health to that which demonstrates optimal benefits on nonskeletal health outcomes. The evidence related to these outcomes will be considered later in this review.
Factors known to influence 25(OH)D levels include race, vitamin D intake, sun exposure, adiposity, age, and physical activity. Even when all the factors known to influence 25(OH)D concentrations are taken into account, most of the individual variation of 25(OH)D values is difficult to explain. Consequently, it is difficult to assess the risk of clinical or biochemical consequences of vitamin D insufficiency in a patient on the basis of concentrations of 25(OH) D alone. The duration of vitamin D insufficiency, the responsiveness of the vitamin D receptor, dietary calcium intake, and individual calcium requirements likely modify the clinical consequences of vitamin D deficiency or insufficiency based on levels of 25(OH)D.
A single exposure to summer sun in a bathing suit for 20 minutes produces the equivalent of 15,000 to 20,000 IU of vitamin D3. In a study of Hawaiian surfers with sun exposure of at least 15 hours per week for the preceding 3 months, 25(OH)D levels ranged from 11 up to 71 ng/mL, demonstrating wide individual variation.22 Outdoor sun exposure and time spent outdoors are better predictors of serum 25(OH)D values than dietary vitamin D intake.23
The 25(OH)D level achieved with the same oral dose of vitamin D varies widely by individual.23,24 The level of 25(OH)D that results in clinical consequences probably varies with calcium intake, race, age, body fat, and individual genetic factors, all of which may influence calcium homeostasis. Genetic variation represented by polymorphisms of certain genes in the vitamin D metabolic pathway explains some of the interindividual variability of 25(OH)D concentrations, particularly polymorphisms of the enzyme 7-dehydrocholesterol reductase in the skin, cytochrome P450 25-hydroxylase in the liver, and vitamin D–binding protein in the circulation.25 The functional effect of a particular level of 25(OH)D depends on the uptake of 25(OH)D by target cells and the efficiency of 1α-hydroxylation to produce 1,25(OH)2D.
Some controversy exists regarding the best method for measuring 25(OH)D levels. Radioimmunoassay has been the most common method reported in the literature and was the method used in some of the large-scale population studies of vitamin D, such as the National Health and Nutrition Examination Survey (NHANES) and the Women's Health Initiative (WHI).
The accuracy of measurement varies widely between individual laboratories and between different assay methods. In one study, identical serum samples were provided to 6 different laboratories, and the chemiluminescent assay tended to return higher values for 25(OH)D.26 Competitive protein-binding assays are also known to generally yield higher 25(OH)D values. When serum samples were spiked with an additional 20 ng/mL of 25(OH)D, the increment in 25(OH)D level was less than 20 ng/mL in all the laboratories, except the one using high-performance liquid chromatography. Antibodies used in some radioimmunoassays do not detect both 25(OH)D2 and 25(OH)D3. The use of a standard cutoff value for adequate vitamin D status is problematic if applied to all laboratories and all methods. A single serum sample could be assessed as showing adequate vitamin D status in one laboratory and an insufficient level in another, with differences of up to 17 ng/mL.26,27
More recently, large medical laboratories have begun using liquid chromatography–tandem mass spectrometry, which identifies the 25-hydroxylated forms of both vitamin D2 and D3.28 The total 25(OH)D, which is the sum of 25(OH)D2 and 25(OH)D3, is used to evaluate vitamin D status. Since 2003, there has been more than a 15-fold increase in the volume of 25(OH)D measurements at Mayo Clinic in Rochester, MN (Singh R., personal communication), reflecting the increasing attention clinicians are giving to vitamin D status.
The classical manifestation of vitamin D deficiency is nutritional rickets, which results from inadequate mineralization of growing bone. Consequently, rickets is a disease of children. Far from being eradicated, nutritional rickets continues to occur throughout the world, with reports from at least 60 countries in the past 20 years.29 In a review of published cases of rickets in the United States, most occurred in children younger than 30 months.30 The vast majority of cases in the United States occurred in African American infants who were fed with breast milk rather than formula. Florid rickets manifests with leg deformities; enlargement of the growth plates of the wrists, ankles, and costochondral junctions; and rib cage deformities. Subtle symptoms that should raise the clinical suspicion of rickets in children include bone pain in the legs, delayed age of standing or walking, frequent falling, and delayed growth. Hypocalcemic seizures in the first year of life may be the initial manifestation of rickets.
Radiography of the long bones at the knees and the wrists is necessary to confirm the diagnosis of rickets. Radiography demonstrates impaired mineralization of the growth plates, evident by widening of the growth plate and fraying of the margin of the metaphyses.31 Biochemical features most consistently include hypophosphatemia and an elevated alkaline phosphatase level. As a result of vitamin D deficiency, serum concentrations of 25(OH)D are very low in patients with rickets, usually less than 5 ng/mL. However, concentrations of 25(OH)D may not be markedly reduced if rickets results from calcium deficiency or if the child has recently received vitamin D or sun exposure. In some tropical countries, where sun exposure is plentiful, calcium deficiency is more important than vitamin D deficiency as a cause of rickets.32,33 However, even in the United States, only 22% of children with nutritional rickets had deficient levels of 25(OH)D, indicating that calcium deficiency as a cause of rickets needs to be considered domestically as well.34
Osteomalacia refers to the failure of organic osteoid formed by osteoblasts to become mineralized with calcium and phosphorus. Although histological osteomalacia is characteristic of rickets, the term osteomalacia is generally used to describe the bone disease caused by vitamin D deficiency in adults, who no longer have growing bones. The clinical manifestations of these 2 conditions are different.
Bone pain is a characteristic feature of osteomalacia, and it can be confused with arthritis or fibromyalgia. Bone pain due to osteomalacia primarily affects the bones between the joints, whereas arthritis usually causes predominantly joint pain, and fibromyalgia causes more diffuse muscle and soft tissue pain; however, it can be difficult to distinguish between these disorders. Proximal muscle weakness and gait instability are often present. Because the growth plates have closed in adults, the radiographic features differ from those typical of rickets. Radiography may reveal pseudofractures of the pelvis, femurs, metatarsals, or lateral margins of the scapulae. The biochemical features of osteomalacia are similar to those of rickets, with increased serum alkaline phosphatase and PTH values, and low calcium, phosphorus, and 25(OH)D values in most cases. A review of all the archived cases of bone biopsy–proven osteomalacia seen by the Bone Histomorphometry Laboratory at Mayo Clinic concluded that radiographic examination as well as serum calcium, phosphorus, and alkaline phosphatase assays are adequate screening tests in patients who have a clinical presentation suggestive of osteomalacia, but that 25(OH)D values may be normal.35
In a cross-sectional study of iliac bone biopsy specimens obtained at autopsy, an excess accumulation of osteoid, which corresponds with histological osteomalacia, was found only in patients with 25(OH)D values less than 25 ng/mL.36 However, even patients with very low values of 25(OH)D did not consistently have evidence of osteomalacia.
Apart from the deficiency diseases of rickets and osteomalacia, recent evidence suggests other skeletal and nonskeletal benefits of vitamin D. In evaluating the evidence, it is important to recognize the limitations inherent in the study design and methodology. Important issues that apply to vitamin D research include the following:
In the following section, representative studies of the available evidence related to the skeletal and nonskeletal effects of vitamin D are reviewed.
Bone Density. In addition to the treatment and prevention of vitamin D–deficiency rickets in children,37 vitamin D has been associated with other beneficial skeletal effects. A retrospective cohort study of pubertal girls demonstrated increased bone mineral density (BMD) of the femoral neck, but not of the spine or radius, among those who received supplemental vitamin D in infancy.38 Evidence of a positive association between BMD and serum 25(OH)D concentrations in adolescents is fair, but the evidence for a positive association in infants is inconsistent.39 Serum 25(OH)D concentration was related to hip BMD in community-dwelling women and men aged at least 20 years who participated in the US NHANES III survey.40 Higher calcium intake was significantly associated with higher BMD only for women with 25(OH)D values less than 20 ng/mL. One of the limitations of a cross-sectional study like the NHANES survey is that it can demonstrate only associations, not cause-and-effect relationships. Another confounding factor may be associated with low vitamin D intake and low bone density. For example, healthier people who exercise more outside in the sun may have greater bone density because of their exercise and higher 25(OH)D levels because of sun exposure. The WHI calcium and vitamin D supplementation trial showed that hip bone density was 1.06% higher in women receiving calcium and vitamin D supplementation vs placebo at 9 years, but that their lumbar spine and total body bone density did not differ significantly from those receiving placebo during this interval.41
Fractures and Falls. On the basis of RCTs, the strongest evidence for the benefit of vitamin D relates to the prevention of fractures and falls. In a meta-analysis of 12 RCTs, a reduced nonvertebral fracture risk was demonstrated only for doses of vitamin D greater than 400 IU/d (relative risk [RR], 0.80; 95% confidence level [CI], 0.72-0.89).42 Similarly, a meta-analysis of 8 RCTs demonstrated that vitamin D reduced the risk of falls (RR, 0.78; 95% CI, 0.64-0.94), but only if the dose was 700 IU/d or greater and the 25(OH)D concentration was at least 25 ng/mL.43 The benefit of vitamin D could have been limited to those with unrecognized osteomalacia, which is associated with proximal muscle weakness and gait instability. These high-quality studies provide clear evidence that a minimum dose of 800 IU/d of vitamin D will reduce the risk of falls and fractures in older adults. However, a recent RCT of a 500,000 IU annual dose of vitamin D in women of advanced age increased the median 25(OH)D concentration from 20 ng/mL to 48 ng/mL one month later but resulted in an increased risk of falls and fractures in the group receiving this regimen.44
Interest in the nonskeletal effects of vitamin D has been increasing since the discovery of vitamin D receptors and the 1α-hydroxylase enzyme in multiple tissues, including cells of the pancreas, immune system, macrophages, vascular endothelium, stomach, epidermis, colon, and placenta.6 In these tissues, 25(OH)D can be converted to 1,25(OH)2D locally, without altering serum 1,25(OH)2D concentrations. Through these paracrine effects, 1,25(OH)2D influences the expression of genes in local tissues. However, the evidence for the nonskeletal benefits of vitamin D is not as strong as the evidence for the skeletal effects.
Lower Mortality Rate. In a prospective observational study of adults older than 65 years participating in NHANES III, the risk of death was 45% lower in those with 25(OH)D values greater than 40 ng/mL compared with those with values less than 10 ng/mL (hazard ratio [HR], 0.55; 95% CI, 0.34-0.88).45 However, this may simply reflect the fact that people with underlying illness or immobility (who are more likely to die) tend to have lower 25(OH)D levels, in part as a result of having spent less time outdoors or of having less adequate nutrition. Because vitamin D is sequestered in adipose tissue, obesity is also associated with lower 25(OH)D levels. However, observational studies cannot prove whether low 25(OH)D status is the cause of greater mortality or just a marker of other underlying risk factors.
In contrast, a meta-analysis of 18 RCTs of vitamin D supplementation in postmenopausal women of advanced age, with dosages ranging from 300 to 2000 IU/d, reported a 7% lower risk of death in those receiving a vitamin D supplement (RR, 0.93; 95% CI, 0.87-0.99).46 This highlights the difference often found between RCTs and observational studies. The effect sizes found in observational studies are often attenuated or absent in RCTs.47 The situation with vitamin D is analogous to that of hormone replacement therapy (HRT) in postmenopausal women. The beneficial effects of HRT were demonstrated for multiple health outcomes in observational studies,48-50 but the WHI RCT in older postmenopausal women failed to confirm the beneficial effects of HRT on dementia and cardiovascular disease.51,52 In the observational trials, healthier women were more likely to use estrogen replacement and had fewer adverse health outcomes, indicating a “healthy user” bias. Only an RCT definitively demonstrated that the risks of first-time use of HRT outweighed the benefits in women older than 60 years.
Despite the slight reduction in mortality associated with vitamin D supplementation, the primary aim of the RCTs included in the meta-analysis was not to assess mortality. Not all trials of vitamin D reported mortality outcomes, so those trials could not be included in the meta-analysis. Trials that showed a mortality effect would be more likely to report this outcome, leading to a high likelihood of reporting bias that could render the slight mortality reduction statistically insignificant.
Lower Cardiovascular Mortality. The reduced mortality in the aforementioned observational study mirrored in large part the reduced cardiovascular mortality in those with 25(OH)D values greater than 40 ng/mL compared with those with values less than 10 ng/mL (HR, 0.42; 95% CI, 0.21-0.85).45 In another observational cohort study, patients who had angiography and 25(OH)D measurements were followed up for 8 years. Those from the highest 25(OH)D quartile (median, 28 ng/mL) had a lower mortality (HR, 0.45; 95% CI, 0.32-0.64) than those from the lowest quartile (median, 8 ng/mL).53 Although these observational studies do not demonstrate that low 25(OH)D values accelerate cardiovascular mortality, low 25(OH)D concentrations were associated with serum markers of inflammation that are indicators of cardiac risk.
Recently, concern has been expressed that vitamin D could potentially accelerate vascular disease. In a study of African Americans with type 2 diabetes mellitus, 25(OH)D levels correlated with increased calcified plaque in the aorta and carotids, but not in the coronary arteries.54 Vascular disease associated with chronic kidney disease, especially that associated with very low bone turnover, may also be accelerated with supplementation with standard doses of vitamin D. Furthermore, concern has been raised recently that other disorders characterized by vascular inflammation, such as diabetes mellitus, rheumatoid arthritis, or systemic lupus erythematosus, may not benefit from standard recommended doses of vitamin D supplementation.
Vitamin D may affect other cardiovascular and metabolic disease risks. In an observational study of adolescents in NHANES III, those with the lowest 25(OH)D values (<15 ng/mL) had more than a 2-fold greater odds of having an elevated blood pressure compared with the group of adolescents with higher 25(OH)D levels (>26 ng/mL) (odds ratio [OR], 2.4; 95% CI, 1.3-4.2).55
The NHANES III data in adults indicated that those with 25(OH)D levels of less than 21 ng/mL had an increased risk of hypertension, diabetes, obesity, and high triglyceride levels—all metabolic manifestations associated with increased cardiovascular mortality.56 Although obesity is associated with lower serum 25(OH)D levels because of the sequestration of vitamin D in adipose tissue, it is likely not the consequence of low 25(OH)D levels. Additionally, the 25(OH)D level may be a marker of other factors associated with obesity, such as physical inactivity and reduced outdoor sun exposure.
Reduced Risk of Diabetes Mellitus. A meta-analysis of 5 observational studies of vitamin D supplementation in childhood reported a nearly 30% reduction in the risk of type 1 diabetes in children who had ever received vitamin D supplements (OR, 0.71; 95% CI, 0.60-0.84).57 Unfortunately, most studies had no information about vitamin D dosage or adherence. Because these were observational studies, and vitamin D was not randomly assigned to children, it is possible that characteristics of families who provided supplemental vitamin D to their children contributed to the decreased risk of type 1 diabetes in children receiving supplements.
Vitamin D receptors are present in pancreatic β cells, and vitamin D may augment insulin secretion and insulin sensitivity. Adolescents in NHANES III with serum 25(OH)D levels of less than 15 ng/mL were more likely to have elevated blood glucose levels than those with the highest 25(OH)D values (>26 ng/mL) (OR, 2.5; 95% CI, 1.0-6.4).55 The observational Nurses Health Study found that vitamin D supplementation and calcium supplementation were both associated with a reduction in risk of type 2 diabetes.58 Current data related to vitamin D and the risk of type 2 diabetes are limited by inadequate adjustment for confounding variables, post hoc analyses, and inability to identify the separate effects of calcium and vitamin D.59 Because milk is the major source of both vitamin D and calcium in the diet, it is difficult to identify the independent effects of dietary calcium and vitamin D on the basis of intake or 25(OH)D levels. Skim milk intake is also inversely associated with obesity, which could account for an association between the intake of dietary calcium and vitamin D and a reduced risk of type 2 diabetes.
Reduced Risk of Cancer. Vitamin D is known to promote cellular differentiation, inhibit cellular proliferation, and reduce the growth of certain tumors in laboratory animals. A meta-analysis of case-control studies of those with and without colon cancer found that, for each 20 ng/mL increase in serum 25(OH)D levels, the odds of colon cancer were reduced by more than 40% (OR, 0.57; 95% CI, 0.43-0.76).60 Other studies have shown that dietary calcium intake is also associated with reduced colon cancer risk and adenoma formation.61,62 Because milk intake is a major determinant of serum 25(OH)D levels, it is difficult to separate the effect of vitamin D from that of calcium intake.
In the case of colon cancer, one large RCT was performed to evaluate the effect of combined supplementation with calcium and vitamin D on the risk of colon cancer. In the WHI trial, supplementation with calcium and vitamin D had no significant effect on the risk of colorectal cancer during 8 years of follow-up.63 Several limitations of this study may have contributed to this lack of effect. Colorectal cancer is a long latency disease, and 8 years may not have been sufficient time to observe the effect of calcium and vitamin D. Another criticism is that the relatively low dose of 400 IU of vitamin D may have not been protective or sufficient to increase serum 25(OH)D levels adequately. Concentrations of 25(OH)D were measured at baseline but not during follow-up. Declining adherence over time would have further reduced the effective doses of calcium and vitamin D.
Breast cancer has also been associated with vitamin D insufficiency. A meta-analysis combining 7 observational studies reported a lower risk of breast cancer among women in the highest compared with the lowest quartile of 25(OH) D values (OR, 0.55; 95% CI, 0.38-0.80).64 As with colon cancer, calcium intake was also associated with a reduced risk of breast cancer. Because obesity is associated with an increased risk of breast cancer and low 25(OH)D levels, it is a confounding factor in the association between breast cancer risk and vitamin D.
As with colon cancer, the WHI RCT of a combined regimen of calcium and vitamin D showed no benefit of supplementation on the risk of breast cancer, again highlighting the different conclusions of observational studies and RCTs.65 The limitations of the breast cancer study are similar to those of the study focused on colon cancer. This study demonstrated the potential confounding effects of physical activity and obesity. Baseline 25(OH)D levels were greater among women with lower body mass index and more recreational physical activity. When controlling for body mass index and physical activity, serum 25(OH)D concentration was not associated with breast cancer risk.
In a meta-analysis of 11 observational studies, prostate cancer was not associated with serum 25(OH)D levels.66 The evidence regarding an association between pancreatic cancer and 25(OH)D levels is conflicting.67 A multinational cohort study found no protective association between greater 25(OH)D values and gastric, esophageal, endometrial, ovarian, kidney, non-Hodgkin lymphoma, and pancreatic cancers.68 Drake et al69 recently showed that event-free survival and overall survival were reduced in vitamin D–insufficient patients newly diagnosed as having diffuse large B-cell lymphoma and T-cell lymphoma during 34.8 months of follow-up.
To date, studies have not shown impressive effects of vitamin D treatment on malignancies.70
Reduced Risk of Multiple Sclerosis. The incidence of multiple sclerosis increases with increasing latitude, corresponding with reduced ultraviolet B sun exposure and lower serum levels of 25(OH)D. A case-control study demonstrated that the odds of having multiple sclerosis were lower in the group with the highest 25(OH)D levels.71 However, the association was found only in white patients [OR, 0.59; 95% CI, 0.36-0.97 for a 20 ng/mL increase in 25(OH)D], not in African American patients. It is difficult to exclude the possibility that other confounding exposures associated with increasing latitude and greater indoor activity during winter months contribute to the risk of multiple sclerosis. Little evidence supports a therapeutic role for vitamin D in the treatment of multiple sclerosis.72
Reduced Risk of Allergy and Asthma. Several lines of evidence demonstrate the effects of vitamin D on proinflammatory cytokines, regulatory T cells, and immune responses, with conflicting interpretation of the effects of vitamin D on allergic diseases.73,74 In a cross-sectional study of Costa Rican children, low 25(OH)D levels were associated with elevated IgE and eosinophil counts, as well as with increased asthma-related hospitalizations and use of anti-inflammatory medication.75 However, an association does not prove causation, and alternative explanations can account for this association. For example, children with more severe asthma may spend more time indoors and have less sun exposure.
Low maternal vitamin D intake in pregnancy has been associated with an increased likelihood of childhood wheezing at ages 3 and 5 years.76,77 In contrast, maternal 25(OH)D levels of greater than 30 ng/mL in pregnancy have been associated with childhood eczema at age 9 months and asthma at age 9 years.78 Vitamin D supplementation in infancy has been associated with increased atopy and allergic rhinitis in adulthood.79 Increasing 25(OH)D levels were associated with increasing risk of allergic rhinitis among adults in NHANES III.80 The conflicting data indicate the need for RCTs to demonstrate the effect of vitamin D on the prevention and control of allergic diseases.
Reduced Risk of Infection. Vitamin D is required for the expression of cathelicidin by macrophages, which is involved in bacterial killing.81 A meta-analysis of 7 observational studies noted a reduced risk of active tuberculosis in those with the highest vs the lowest values of 25(OH)D (OR, 0.68; 95% CI, 0.43-0.93).82 However, an RCT in a West African population with baseline mean 25(OH)D values of 31 ng/mL showed no effect of 100,000 IU of supplemental vitamin D given at the beginning and at 3 and 8 months of tuberculosis treatment on the rate of sputum conversion or resolution of markers of clinical severity.83 However, this dose of vitamin D may have been insufficient because the increase in 25(OH)D concentration during treatment did not differ between the supplement and placebo groups.
In observational data from NHANES III, persons with 25(OH)D values lower than 10 ng/mL were more likely to have had a recent upper respiratory tract infection than those with higher 25(OH)D values in all 4 seasons of the year.84 This association was even stronger in those with asthma or chronic obstructive pulmonary disease. Whether this association is explained by the fact that people who remain indoors are more likely to catch colds remains unclear.
A case-control study reported that mean 25(OH)D values were lower in children with bronchiolitis or pneumonia admitted to the pediatric intensive care unit than in healthy control children or in children with pneumonia admitted to the general pediatric ward.85
Reduced Risk of Mental Illness. A cohort of Finnish children who received supplemental vitamin D in their first year of life had a lower risk of developing schizophrenia.86 However, the significance of this association is unclear because it was unrelated to adherence to vitamin D supplementation, was only evident in males, and was not found with any other mental illness.
To examine the effect of vitamin D on depression, overweight and obese patients were randomized to receive 20,000 or 40,000 IU of vitamin D or placebo weekly for 1 year.87 At baseline, those with 25(OH)D concentrations lower than 16 ng/mL had greater Beck Depression Inventory scores, indicating that they were more depressed, than those with higher 25(OH)D levels. The 2 groups receiving vitamin D supplementation had significant improvement in their scores, whereas the placebo group did not.
Less Musculoskeletal Pain. A small descriptive study reported that most patients (93%) with persistent musculoskeletal pain had 25(OH)D values of 20 ng/mL or less.88 In one RCT, patients with diffuse musculoskeletal pain or osteoarthritis and 25(OH)D values lower than 20 ng/mL were randomized to receive vitamin D or placebo for 3 months.89 Those given vitamin D had no improvement in their pain compared with baseline or compared with placebo-treated patients. In another double-blind RCT, primary care patients with 25(OH)D levels of 10 to 25 ng/mL were randomized to receive 50,000 IU of vitamin D or placebo weekly for 8 weeks.90 The treated group showed significantly greater improvement in fibromyalgia assessment scores than the placebo group. Patients with 25(OH)D values lower than 10 ng/mL were treated in an unblinded fashion with 50,000 IU of vitamin D weekly for 8 weeks but had no symptom improvement.
Reduced Risk of Renal Disease. In a subgroup analysis of the NHANES III data set, low 25(OH)D values were associated with a greater risk of kidney failure in African American but not in white participants.91 However, the opposite trend was observed in whites.
Measurement of serum 25(OH)D levels is indicated in select circumstances. If clinical symptoms of rickets in children or osteomalacia in adults are present, measurement of 25(OH)D levels will confirm vitamin D deficiency. Such testing would be appropriate in adults or children with bone pain, elevated serum alkaline phosphatase or PTH levels, and low serum calcium or phosphorus levels. Persons of advanced age, those with osteoporosis, or those at increased risk of falls or fractures may also benefit from measurement of 25(OH)D levels. However, one could argue that providing at-risk groups with routine supplementation of adequate doses of vitamin D may make testing for vitamin D insufficiency unnecessary. When to test and how to treat adults with vitamin D deficiency have recently been reviewed in this journal.8 No evidence shows benefit for screening 25(OH)D levels in the asymptomatic population.
Critically evaluating the evidence regarding the purported benefit of vitamin D on a multitude of health outcomes is difficult. The bulk of current data is based on observational, epidemiological studies, which are useful for generating hypotheses but not for proving causality. It is particularly difficult to tease out the effects of confounding variables that relate both to health outcomes and to vitamin D status, such as physical activity, milk intake, and adiposity. Few of the observational associations have been confirmed by RCTs, and many of the interventional studies of vitamin D also included calcium supplementation. Future clinical trials, including a National Institutes of Health–funded 5-year 20,000-participant prospective RCT comparing the effect of supplementation with 2000 IU/d of vitamin D3 or placebo, will help clarify the benefits and risks of vitamin D supplementation in many of the disorders discussed in this review.92
On the basis of the current data, it seems prudent for persons older than 60 years to take a vitamin D supplement of 800 to 2000 IU/d to reduce the risk of falls and fractures. These recommendations are consistent with the recently released report of the Institute of Medicine, which recommended that healthy adults take 600 IU/d to maintain skeletal health and also concluded that information about the health benefits beyond bone health could not be considered reliable.93 Dark-skinned infants who are exclusively breast-fed are at greater risk of rickets and should receive 400 IU/d of supplemental vitamin D. Vitamin D supplementation in these ranges is likely to prevent bone loss, may improve bone density, may reduce fractures, and appears to reduce falls. Although vitamin D intoxication has been associated only with intakes of 50,000 to 1 million IU/d over the course of months or years,7 the potential risks of kidney stones, vascular disease, and fractures with high-dose vitamin D regimens are unclear. Until more data from RCTs are available, a healthy dose of skepticism should be maintained regarding the other health claims for vitamin D.