The author states that he has no conflicts of interest.
Quantitative aspects of vitamin D3 endocrine physiology are briefly reviewed, together with the supporting evidence. Net calcium absorption of at least 200 mg/d is required to offset typical obligatory losses and thereby to protect the skeleton. The ability of the body to regulate intestinal calcium absorption is optimal at serum 25-hydroxyvitamin D3 concentrations >80 nM. Vitamin D3 inputs from all sources required to sustain such a level amount to 3600–4200 IU/d. Daily oral intakes as high as 10,000 IU are safe.
This brief review presumes familiarity with the canonical scheme whereby vitamin D is synthesized in the skin, 25-hydroxylated in the liver, and subsequently, 1-α-hydroxylated to produce 1,25(OH)2D (calcitriol). In the endocrine portion of the vitamin D system, the last step takes place in the kidney and is responsive to controls such as PTH and fibroblast growth factor 23 (FGF23). 25(OH)D serum concentration is not effectively limiting in this reaction, because even low 25(OH)D levels will support renal 1-α-hydroxylation. In contrast, the autocrine/paracrine aspect of the system involves 1-α-hydroxylation in a wide variety of tissues as a part of the cell signaling system important for various aspects of the immune response and for cell differentiation, replication, and apoptosis. Calcitriol synthesis is controlled locally through tissue-specific, stimulus-evoked expression of the 1-α-hydroxylase gene. Tissue functional response depends on the amount of calcitriol produced locally, and because 1-α-hydroxylase operates well below its km, calcitriol synthesis is limited by serum 25(OH)D concentration. The autocrine/paracrine portion of the system accounts for at least 80%, and possibly as much as 95%, of the metabolic consumption of the vitamin each day.
My focus, in this brief review, is on the endocrine portion of the system, and in doing so, I shall concentrate primarily on the quantitative aspects of the operation of vitamin D's control of calcium absorption and on the various downstream consequences thereof.
Vitamin D has long been recognized as important for calcium absorption, but the quantitative relationships between intake, vitamin D status, and absorption are less well documented. Figure 1 quantifies these relationships with a series of contour lines representing varying degrees of vitamin D–mediated active calcium absorption, plotted as a function of the ingested calcium load. As Fig. 1 shows, in the absence of active transport (the zero-contour line), calcium intake would have to rise above 1100 mg/d for net calcium absorption to be zero or above and to ∼2500 mg/d for net absorption to equal 200 mg/d. The 200-mg net absorption figure is important here because 200 mg/d is the typical amount of obligatory calcium loss in adults that must be offset by absorbed calcium if bone is to be spared. In contrast, at an active transport figure of 16%, net absorption reaches 200 mg at a calcium intake between 1000 and 1200 mg/d (i.e., the current recommendations for calcium intake for adults).
Only two studies have been performed to date assessing absorption efficiency as a function of vitamin D status. Both showed clinically important improvement in absorption efficiency as serum 25(OH)D rose from ∼50 to 66 nM in one and to 83 nM in the other. Furthermore, the rate of rise per unit change in serum 25(OH)D was identical for both studies. In a third study, only a very small difference in absorption efficiency was noted in individuals studied at 25(OH)D levels of ∼120 and ∼78 nM. Taken together, these studies suggest that absorption efficiency rises up to serum 25(OH)D levels of ∼80 nM, with a plateau occurring at higher levels. What is unclear is how serum 25(OH)D itself is influencing absorption fraction in this range of vitamin D status values. Not only is absorption fraction positively correlated with 25(OH)D in cross-sectional studies, but exogenous administration of 25(OH)D significantly increases absorption fraction without altering calcitriol levels.
Two corollary observations suggest that the observed improvement in absorption as serum 25(OH)D rises to 80 nM is biologically meaningful. One is the reduction in osteoporotic fractures reported for vitamin D supplementation producing change in serum 25(OH)D over precisely the same range of values as the absorption studies above, and the second is the positive correlation of hip BMD with serum 25(OH)D observed in NHANES III.
PARATHYROID RESPONSE TO VITAMIN D INSUFFICIENCY
One of the expected concomitants of low 25(OH)D status is a rise in PTH secretion, evoked by reduced calcium absorption. This relationship has been noted in essentially all studies performed to date. Above some 25(OH)D concentration, the relation to PTH is flat, but below that concentration, PTH and 25(OH)D vary inversely. The inflection point of this biphasic relationship has been used to define the lower level of vitamin D adequacy, and in various studies that value ranges from 50 to 80 nM. This basis for estimating adequacy is not robust, because the inflection point itself varies inversely with calcium intake. What is more interesting is the fact that not all individuals with low serum 25(OH)D exhibit the expected rise in serum PTH. It is not clear what the clinical significance of that failure may be. Sahota et al. have recently shown, using a magnesium tolerance test, that patients with low vitamin D status and low-to-normal PTH levels have subclinical magnesium deficiency and respond to magnesium supplementation by an elevation in PTH. It would seem from these data that perhaps one half of all patients with vitamin D deficiency are also to some extent magnesium deficient. Whether this is caused by the vitamin D deficiency or is an independent condition is unclear. Similarly, it is not known whether magnesium supplementation, despite normalizing PTH response, alters clinical status to a meaningful extent.
QUANTIFYING INPUTS AND METABOLIC UTILIZATION
Controlled oral dosing experiments have defined the rise in serum 25(OH)D that can be produced by any given input of cholecalciferol. The several studies reporting such data exhibit a high degree of consistency, with most showing that serum 25(OH)D rises from 0.6 to 1.2 nM/μg cholecalciferol/d. Taking values toward the middle of that range, this means that, at 80 nM, the metabolic consumption of vitamin D is on the order of 3600–4200 IU/d. In contrast, typical documentable intakes from food sources (natural and fortified) are in the range of 150–200 IU/d, and including multivitamin preparations, perhaps up to 600 IU/d (in individuals who use supplements). Unless there are unrecognized food sources, most of the daily consumption of vitamin D must be offset by cutaneous inputs. It is important, therefore, to define those inputs quantitatively.
It is well known, for example, that vitamin D synthesis is stimulated through a photoconversion of 7-dehydrocholesterol to previtamin D by UVB wave lengths, that this conversion is greater in light skinned individuals than in dark, and that it is greater in younger individuals than in older. What is needed, however, is quantification of these ordinal relationships. Armas et al. have recently shown that the 25(OH)D response to controlled doses of UVB can be captured by a simple equation incorporating skin lightness and UVB dose. For example, in individuals of Northern European extraction, a three times weekly dose of 30 mJ/cm2 over 90% of body surface for 4 wk raises 25(OH)D by 24 nM. The same dose in blacks raises 25(OH)D by 16 nM and in Sub-Saharan Africans by 10 nM.
The rate of metabolic consumption of 25(OH)D is, to a certain extent, independent of input and may explain some of the variability between individuals of the same skin type and apparently similar vitamin D inputs. One such factor is the circulating level of calcitriol, which in turn is induced by PTH and ultimately by absorbed calcium intake. For reasons that are teleologically unclear, high calcitriol levels shorten the half-time of 25(OH)D substantially. This seems to explain the generally poor vitamin D status of patients with primary hyperparathyroidism and the spontaneous improvement in vitamin D status after removal of a parathyroid adenoma. Even in individuals without primary hyperparathyroidism, large variations in calcium intake would be expected to produce corresponding variations in PTH, and with it, variations in average 24-h circulating calcitriol concentrations. Consistent with this expectation, calcium supplements have been shown to raise 25(OH)D levels and lower calcitriol levels by clinically meaningful amounts.
A thorough evaluation of the literature, mostly published since the Food and Nutrition Board promulgated the tolerable upper intake level (TUIL or UL) for vitamin D in 1997, supports the need to increase the UL for vitamin D. Briefly, there are no documented cases of vitamin D intoxication at serum 25(OH)D levels <500 nM. It would take continuous oral intakes in excess of 25,000 IU/d to produce such a level. Evaluating a large body of literature, Hathcock et al. argued for a UL of 10,000 IU/d. Whereas there are no physiological reasons for oral intakes as high as 10,000 IU/d, nevertheless, it is worth noting that a daily dose that high can be produced by sun exposure and that typical 25(OH)D levels produced at that intake are on the order of 220 nM, far below the level at which toxicity manifestations might occur.
This attempt to develop a more quantitative description of vitamin D physiology reveals some interesting and potentially important questions.
Currently measured cutaneous and oral inputs seem insufficient to account fully for observed serum 25(OH)D levels in many individuals. Where are we getting our vitamin D? Are there are unrecognized food sources? How much of the interindividual variability in 25(OH)D concentration can be explained by variation in calcium intake?
25(OH)D has an effect on calcium absorption that is at least an order of magnitude greater than its binding to the VDR can explain. How is 25(OH)D acting in this regard? Why is absorption low in osteomalacia despite “normal” or sometimes high levels of calcitriol [but always low values for 25(OH)D]?
What is the clinical significance of the subclinical magnesium deficiency that is apparently common in vitamin D–deficient patients?