Kind of an interesting study of the relationship to Vita D levels and calcium/etc... absorption.
Calcium Absorption Varies within the Reference Range for Serum 25-Hydroxyvitamin D
Robert P. Heaney, MD, FACN, M. Susan Dowell, PhD, Cecilia A. Hale, PhD and Adrianne Bendich, PhD, FACN
Creighton University, Omaha, Nebraska (R.P.H., M.S.D.)
GlaxoSmithKline, Parsipanny, New Jersey (C.A.H., A.B.)
Address reprint requests to: Robert P. Heaney, M.D., Creighton University, 601 N. 30th St., Suite 4841, Omaha, NE 68131. Email: firstname.lastname@example.org
Background: Calcium absorption is generally considered to be impaired under conditions of vitamin D deficiency, but the vitamin D status that fully normalizes absorption is not known for humans.
Objective: To quantify calcium absorption at two levels of vitamin D repletion, using pharmacokinetic methods and commercially marketed calcium supplements.
Design: Two experiments performed in the spring of the year, one year apart. In the first, in which participants were pretreated with 25-hydroxyvitamin D (25OHD), mean serum 25OHD concentration was 86.5 nmol/L; and in the other, with no pretreatment, mean serum concentration was 50.2 nmol/L. Participants received 500 mg oral calcium loads as a part of a standard low calcium breakfast. A low calcium lunch was provided at mid-day. Blood was obtained fasting and at frequent intervals for 10 to 12 hours thereafter.
Methods: Relative calcium absorption at the two 25OHD concentrations was estimated from the area under the curve (AUC) for the load-induced increment in serum total calcium.
Results: AUC9 (± SEM), was 3.63 mg hr/dL ± 0.234 in participants pretreated with 25OHD and 2.20 ± 0.240 in those not pretreated (P < 0.001). In brief, absorption was 65% higher at serum 25OHD levels averaging 86.5 nmol/L than at levels averaging 50 nmol/L (both values within the nominal reference range for this analyte).
Conclusions: Despite the fact that the mean serum 25OHD level in the experiment without supplementation was within the current reference ranges, calcium absorptive performance at 50 nmol/L was significantly reduced relative to that at a mean 25OHD level of 86 nmol/L. Thus, individuals with serum 25-hydroxyvitamin D levels at the low end of the current reference ranges may not be getting the full benefit from their calcium intake. We conclude that the lower end of the current reference range is set too low.
Key words: vitamin D, calcium absorption, 25-hydroxyvitamin D, vitamin D status, parathyroid hormone, vitamin D requirement
In 1997 the Food and Nutrition Board of the Institute of Medicine accepted serum 25-hydroxyvitamin D concentration (25OHD) as the functional indicator of vitamin D status , but data were insufficient at that time to characterize fully the physiological normal range for this indicator. Most reference laboratories cite lower limits varying from 37.5 to 50 nmol/L (15 to 20 ng/mL), but this lower limit is empirically based on measurements in ostensibly "normal" individuals. Since there is a growing consensus that vitamin D insufficiency is more common than previously thought, such empiric estimates may well be circular; furthermore, individuals who would benefit from higher vitamin D status may be inappropriately classified as "normal" if their serum 25OHD concentration falls within the reference range.
Studies that have evaluated serum parathyroid hormone (PTH) concentration as a function of 25OHD level have generally found that PTH is higher at low 25OHD levels than at higher 25OHD values. The curve flattens out above 25OHD values that range, in various reports, from 75 to 110 nmol/L [2–4]. The tendency for higher PTH levels below such inflection points is usually interpreted to indicate a physiological adaptation to reduced calcium entry into the body, and there are arguments on both sides of the question whether such physiological adaptation is conducive to or indicative of optimal health [5–6]. What is lacking in the data available to date is evidence of quantitative variation in function at 25OHD levels in the disputed zone between the lower end of the nominal reference range and the level where PTH becomes constant.
In an attempt to provide such data, we present in this paper the results of paired studies of calcium absorption in healthy postmenopausal women performed under different conditions of vitamin D repletion.
Two studies were conducted in Omaha, Nebraska at 41.3° N. latitude, approximately one year apart, in the spring of the year at the time of the seasonal nadir for serum 25OHD. Each was a randomized, cross-over study, designed to test the relative absorbabilities of two calcium supplement sources, ingested at single, 500 mg loads taken as part of a standard low calcium breakfast. Results from the first study, comparing the two calcium sources, have been published previously . The protocol for the first study included pre-dosing with 25OHD (CalderolTM, Organon, West Orange, NJ), at a dose of 20 µg given on alternate days for an average of three weeks prior to the absorption measurements, a stratagem designed to bring all participants rapidly into a state of vitamin D sufficiency. By contrast, in the second study, there was no pretreatment with vitamin D or 25OHD. The test substances in both studies were commercially marketed preparations of Os-CalTM and CitracalTM in doses providing 500 mg of calcium (as calcium carbonate for Os-Cal) and 515 mg of calcium (as calcium citrate for Citracal). Both sources also provided 200 IU of vitamin D, which the participants received one time only, on the day of the test. Subjects were fed a low calcium lunch five hours after the test breakfast.
In both studies the two marketed products tested were found to be bioequivalent, with nearly identical (and nonsignificantly different) indices of absorption. Hence, for purposes of this analysis, within-subject averages of calcemia for the two calcium sources in each study were taken as the best estimates of each participant’s absorptive performance under the then prevailing vitamin D status.
Participants were 34 postmenopausal women, 14 of whom took part in both studies. Average age at time of study was 56 ± 7 years in the first study and 64 ± 9 years in the second. Body mass index (kg/m2) was 29.2 ± 5.2 for the study with 25OHD supplementation and 28.8 ± 3.8 for the study without. Twelve of the 24 women in the first study were receiving estrogen replacement therapy, and 10 in the second study. Individuals with digestive disorders, antibiotic use within five days or unstable medical conditions of any sort were excluded. The study was approved by the Creighton University Institutional Review Board, and each subject gave written consent.
Relative absorption was estimated from the area under the curve over intervals ranging from 9 to 12 hours after dosing (AUCt). Blood was drawn immediately prior to the test breakfast (time zero) and at frequent intervals thereafter out to 24 hours in the first study and to 12 hours in the second. AUC was calculated for 9, 10 and 12 hours by the trapezoidal method using the increment above each individual’s baseline serum calcium value, and the individual participant AUCt values were aggregated across each study. Comparative absorption was expressed as the ratio of the means of the two AUCt values. Since the timing of the blood samples was not identical for the two studies, AUCt was calculated for each using values confected at 9 or 10 hours (as the case may be) by linear interpolation between measurements on either side of the desired time point. As the values were close to baseline by nine hours, this assumption of linearity can have introduced at most only a trivial error. Because AUC9 has been shown elsewhere  to be better correlated with true absorption than AUC values at earlier or later times, the primary comparisons we report will be based on AUCt calculated over nine hours. Fractional absorption was calculated from the AUC9 values using the following formula :
Serum calcium was measured by atomic absorption spectrophotometry (AAnalyst 100, Perkin-Elmer, Norwalk, CT). Serum 25OHD was measured once at baseline on each subject in each study (Nichols Institute Diagnostics, Catalog No. 40-2135, San Juan Capistrano, CA). Serum immunoreactive parathyroid hormone (iPTH) was measured as the intact molecule by IRMA (Nichols, San Juan Capistrano, CA).
Data were analyzed in two ways. Since approximately half of the subjects in each group were not common to the two studies, the two sets of values were analyzed as independent samples, using ANOVA and testing for period and treatment effects. For the 14 women common to each study, a repeated measures ANOVA was performed, also testing for treatment and period effects. In both instances we used SAS (SAS Institute, Cary, North Carolina), as well as the various descriptive statistics provided by Excel (Microsoft, Redmond, WA).
Table 1 sets forth the numerical values for the principal findings in this study. Serum 25OHD at baseline was 36 nmol/L higher in the study in which participants had been pretreated with 25OHD. The fact of this difference is not surprising, since that was the intent of the pretreatment. The size of the difference is of greater interest. The mean within-individual difference in the 14 participants common to both studies was 34.3 nmol/L ± 3.7 (SEM), virtually the same as the difference between the two group means. The primary outcome was the incremental AUC9 for serum calcium, which was 65% higher in the 25OHD treated study (D+) than in the untreated study (D-). This difference was highly significant.
View this table:
[in this window]
[in a new window]
Table 1. Serum Ca AUC and Related Variables*
Fig. 1 shows the calcemia time course for each calcium source in the two studies (with and without vitamin D repletion). The figure shows graphically not only that baseline vitamin D status significantly influenced calcium absorption but that the effect of D-status was equivalent for the two calcium salts.
View larger version (25K):
[in this window]
[in a new window]
Fig. 1. Time course of the incremental calcemia following ingestion of 12.5 mmol Ca loads from two different sources and with and without pretreatment with 25OHD. The top two lines are the calcemia curves observed in participants pretreated with 25OHD, and the bottom two, those measured without vitamin D. Error bars are 1 SEM. ( & represent CitracalTM and & represent OscalTM.) (Copyright Robert P. Heaney, 2002. Used with permission.)
Fig. 2 presents the full time course of the mean serum calcium increments for the two salts combined over the 10 to 12 hour periods following the test calcium loads. Not only was the AUCt different, but at most of the time points from 3 to 10 hours, the serum calcium increment was significantly greater in the D+ study than in the untreated. Additionally, as Table 1 shows, even the baseline serum calcium values, while all within the reference normal range (8.8 to 10.2 mg/dL), were significantly higher in the D+ participants (by 0.33 mg/dL; P < 0.002). While basal serum PTH was higher in the D- group, the difference was not statistically significant, largely because, unlike calcium, which is tightly regulated, PTH values exhibited broad dispersion with a coefficient of variation more than 10x greater. Similarly, the AUC for PTH showed less of a drop in the study without supplemental vitamin D, but the difference was not statistically significant. Finally, there was no difference in AUC9 in either study between those receiving estrogen replacement therapy and those not so treated (data not shown).
View larger version (24K):
[in this window]
[in a new window]
Fig. 2. Time course of the mean increment in serum total calcium in two studies, in one of which vitamin D status was elevated ("D+"), and in the other, it was not ("D-"). Error bars are 1 SEM. (Copyright Robert P. Heaney, 2002. Used with permission.)
Even within each of the two treatment groups there was a broad range of 25OHD values; in fact, 25OHD values were continuously distributed across both groups. Hence we pooled the two studies and regressed AUC9 on 25OHD concentration (Fig. 3). As expected there was a significant positive association (P < 0.05). The coefficient of correlation improved substantially if the regression were confined to 25OHD values below 90 nmol/L (above which level the correlation was effectively zero).
View larger version (16K):
[in this window]
[in a new window]
Fig. 3. Correlation of AUC9 with serum 25OHD concentration in 48 measurements of calcium absorption in 34 postmenopausal women. The lines represent the least squares regression line through the data and its 95% confidence limits. (Copyright Robert P. Heaney, 2002. Used with permission.)
Subjects in the second study, without supplemental 25OHD, were significantly older than those in the first study. Because of this age difference and the possibility that some or all of the absorptive difference was due to age rather than to vitamin D status, we separately evaluated the within-individual differences in AUC9 in the 14 participants common to both studies. In these women, the mean difference in AUC9 was 1.132 mg hr/dL ± 0.339 (SEM; P < 0.01), a value 45% greater than without supplemental 25OHD and, in absolute magnitude, nearly as large as the intergroup difference. Using the formula for conversion of AUC9 to true fractional absorption values (Eq. 1), mean calcium absorption under vitamin D repletion was 35.3% of load (± 11.
and, without supplemental vitamin D, 22.5% of load (± 12.0).
Many nutrients exhibit what has been termed "threshold behavior," that is, the values for the physiological response change directly with intake up to some threshold value, above which the response does not change with further increases in intake. Calcium, iron and ascorbic acid are well recognized examples. Vitamin D is usually considered to exhibit similar behavior, and the limited evidence available is consistent with that interpretation. Threshold behavior in this instance would mean that, in a state of vitamin D sufficiency, variations in vitamin D intake within the physiological range would not alter calcium absorption efficiency, while absorption would vary with intake at subthreshold values for vitamin D status.
As noted earlier, the Food and Nutrition Board in its 1997 recommendations  was not able to specify the serum 25OHD value at which vitamin D status reaches the response threshold. Mortensen and Charles  had shown an improvement in absorption in healthy Danish subjects given short-term pretreatment with vitamin D, but unfortunately their study provides no data on serum 25OHD concentrations. Scotti et al.  more recently have shown very similar results in Italian male subjects, in this case with accompanying serum 25OHD values. Specifically, the index of calcium absorption used was higher at a serum 25OHD level of 67.5 nmol/L than at 55.5 nmol/L. Unfortunately both studies used, as their measure of calcium absorption, the induced rise in urine calcium excretion, which, while convenient, is a very insensitive indicator.
In this study, using classical pharmacokinetic methods, we found that healthy, postmenopausal women with serum 25OHD levels averaging 86.5 nmol/L had calcium absorptive efficiencies from 45% to 65% greater than those with mean 25OHD levels of 50.1 nmol/L. Aside from a general congruence with the findings of Scotti et al. , this is the first such evidence of which we are aware. Presumably it is precisely this reduced calcium absorption at 25OHD levels below 80 nmol/L which is the stimulus to the higher PTH secretion reported by others [2–4] at 25OHD values below 80 nmol/L.
We have recently shown that a drop in serum 25OHD from 122 to 74 nmol/L did not produce a significant difference in calcium absorption , in contrast to the findings in this study where a drop from 86.5 to 50.1 nmol/L produced a large drop in absorption. In the context of the threshold model, that means that a serum 25OHD of 50 nmol/L is on the ascending limb of the response curve, while 86 nmol/L is close to or above the threshold. Consistent with placement of the threshold near 80–90 nmol/L is not only the PTH behavior alluded to above [2–4], but our finding in this study that the correlation between 25OHD and AUC9 was tightest for 25OHD values below 90 nmol/L, and effectively non-existent above that level.
A limitation of this study is that the data were obtained solely in postmenopausal women. Additional work needs to be done in men and in women at other ages. A further limitation is the pharmacokinetic method itself. While more sensitive and specific than urine calcium-based methods, the pharmacokinetic approach presents formidable analytic challenges. The degree of peak calcemia produced by a single 500 mg load, especially in the face of lower vitamin D status, is, as Figs 1 and 2 show, on the order of only 5% above basal values. Of necessity, some individuals will have below average peak elevations, and earlier and later values will be lower still. The error in quantifying the calcemic rise in individual subjects is, therefore, unavoidably large. While sample size can help overcome this limitation, it is of little help in evaluating associations, in which individual values must be employed. Thus, while we found a significant positive correlation between AUC9 and serum 25OHD concentration, the dispersion of the data around the regression line was large, mainly, we believe, because of the unavoidably poor precision of the individual AUC estimates for the load-induced calcemia.
Since calcium absorption is critical to the ability to maintain calcium balance, it follows that reduced absorptive performance at 25OHD levels between 50 and 80 nmol/L must be considered suboptimal, and, accordingly, 25OHD values in that range ought to be considered subnormal. While the precise location of the threshold remains uncertain, the evidence presented here points to a value closer to 80 or 90 nmol/L, consistent with the studies of PTH concentration [2–4]. In any event, it seems more certain now that the lower boundaries of the reference ranges (i.e., 37.5 to 50 nmol/L) are incorrect, i.e., such levels of serum 25 hydroxyvitamin D are associated with suboptimal calcium absorption, thereby exacerbating the negative effects of the low calcium intakes that are today found across most population segments.
Note Added to Proof
A very recent publication corroborates our conclusion about the inadequacy of 25OHD values in the lower half of the reference range. Trivedi et al. , in a population-based, randomized controlled trial of 2,686 individuals, aged 65 to 85, showed that treatment with vitamin D3 in a dose sufficient to raise serum 25OHD from 53 to 74 nmol/L, decreased fracture risk at hip, forearm, or spine by 33% (P = 0.02). This effect occurred over almost exactly the 25OHD range for which we here report a substantial difference in calcium absorption efficiency.
Work supported in part by contracts with GlaxoSmithKline.
Received November 1, 2002. Accepted January 27, 2003.
Food and Nutrition Board, Institute of Medicine:"Dietary Reference Intakes for Calcium, Magnesium, Phosphorus, Vitamin D, and Fluoride." Washington, DC: National Academy Press,1997 .
Chapuy M-C, Preziosi P, Maamer M, Arnaud S, Galan P, Hercberg S, Meunier PJ: Prevalence of vitamin D insufficiency in an adult normal population.Osteoporos Int7 :439 –443,1997 .[Medline]
Thomas MK, Lloyd-Jones DM, Thadhani RI, Shaw AC, Deraska DJ, Kitch BT, Vamvakas EC, Dick IM, Prince RL, Finkelstein JS: Hypovitaminosis D in medical inpatients.N Engl J Med338 :777 –783,1998 .[Abstract/Free Full Text]
Dawson-Hughes B, Harris SS, Dallal GE: Plasma calcidiol, season, and serum parathyroid hormone concentrations in healthy elderly men and women.Am J Clin Nutr65 :67 –71,1997 .[Abstract]
Burckhardt P: Calcium and vitamin D in osteoporosis: supplementation or treatment?Calcif Tissue Int70 :74 –77,2002 .[Medline]
Heaney RP: The importance of calcium intake for lifelong skeletal health.Calcif Tissue Int70 :70 –73,2002 .[Medline]
Heaney RP, Dowell MS, Bierman J, Hale CA, Bendich A: Absorbability and cost effectiveness in calcium supplementation.J Am Coll Nutr20 :239 –246,2001 .[Abstract/Free Full Text]
Heaney RP: Measuring calcium absorption by pharmacokinetic methods [Abstract].J Am Coll Nutr2 :478 ,2002 .
Mortensen L, Charles P: Bioavailability of calcium supplements and the effect of vitamin D: comparisons between milk, calcium carbonate, and calcium carbonate plus vitamin D.Am J Clin Nutr63 :354 –357,1996 .[Abstract]
Barger-Lux MJ, Heaney RP: Effects of above average summer sun exposure on serum 25-hydroxyvitamin D and calcium absorption fraction.J Clin Endocrinol Metab87 :4952 –4956,2002 .[Abstract/Free Full Text]
Scotti A, Bianchini C, Gianalfredo A, Marzo A: Absorption of calcium administered alone or in fixed combination with vitamin D to healthy volunteers.Arzneim-Forsch/Drug Res51 :493 –500,2001 .
Trivedi DP, Doll R, Khaw KT: Effect of four monthly oral vitamin D3 (cholecalciferol) supplementation on fractures and mortality in men and women living in the community: randomised double blind controlled trial.BMJ326 :469 –474,2003 .[Abstract/Free Full Text]