HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Heaney, R. P. Articles by Bendich, A. Search for Related Content PubMed PubMed Citation Articles by Heaney, R. P. Articles by Bendich, A.

Absorbability and Cost Effectiveness in Calcium Supplementation

Creighton University, Osteoporosis Research Center, Omaha, Nebraska (R.P.H., M.S.D., J.B.), GlaxoSmithKline, Parsippany, New Jersey (C.A.H., A.B.)
Creighton University, Osteoporosis Research Center, GlaxoSmithKline, Parsippany, New Jersey (C.A.H., A.B.)

Address reprint requests to: Robert P. Heaney, MD, Creighton University, Osteoporosis Research Center, 601 North 30^th Street—Suite 4841, Omaha, NE 68131. E-mail: rheaney{at}creighton.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Background: Cost-effectiveness of calcium supplementation depends not only on the cost of the product but on the efficiency of its absorption. Published cost-benefit analyses assume equal bioavailability for all calcium sources. Some published studies have suggested that there are differences in both the bioavailability and cost of the major calcium supplements.

Design: Randomized four period, three-way cross-over comparing single doses of off-the-shelf commercial calcium supplements containing either calcium carbonate or calcium citrate compared with a no-load blank and with encapsulated calcium carbonate devoid of other ingredients; subjects rendered fully vitamin D-replete with 10 µg/day 25(OH)D by mouth, starting one week prior to the first test.

Subjects: 24 postmenopausal women

Methods: Pharmacokinetic analysis of the increment in serum total and ionized calcium and the decrement in serum iPTH induced by an oral calcium load, based upon multiple blood samples over a 24-hour period; measurement of the rise in urine calcium excretion. Data analyzed by repeated measures ANOVA. Cost calculations based on average retail prices of marketed products used in this study from April through October, 2000.

Results: All three calcium sources (marketed calcium carbonate, encapsulated calcium carbonate and marketed calcium citrate) produced identical 24-hour time courses for the increment in total serum calcium. Thus, these were equally absorbed and had equivalent bioavailability. Urine calcium rose slightly more with the citrate than with the carbonate preparations, but the difference was not significant. Serum iPTH showed the expected depression accompanying the rise in serum calcium, and there were no significant differences between products.

Conclusion: Given the equivalent bioavailability of the two marketed products, the cost benefit analysis favors the less expensive carbonate product.

Key words: calcium absorption, calcium carbonate, calcium citrate, bioavailability, cost-effectiveness

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
There is general acceptance of the importance of achieving adequate calcium intakes throughout life, and in most adults effort in that regard means taking some form of calcium supplement. Over half the women enrolled in the Women’s Health Initiative reported using supplements, and that figure rose to nearly 60% in women over age 70 [1]. While calcium supplementation has generally been considered a cost effective intervention [2,4], much depends upon the cost of the preparation. Thus Torgerson and Kanis, in the UK, calculated that calcium was not cost effective for a preparation they priced at $0.50/g in current dollars [5]. Lowering that cost modestly produced a more favorable relationship. Bendich et al. [4] found that calcium supplementation at 1200 mg/day and a cost of $0.10–0.12/g was cost effective for all US women 75 years of age or older when calculated against the costs of care associated only with hip fracture. If the endpoint was increase in bone mineral density and its associated lower total fracture risk, then calcium supplementation was cost effective even with universal supplementation of all US men and women 65 years of age or older.

An additional consideration, given virtually no attention to date, involves factoring in bioavailability of the calcium source. Most, if not all, analyses to date have assumed equivalent bioavailability for different salts and different consumer formulations. Recent publications by Heller et al. [6,7] suggested that this might not be the case. The authors reported absorbability for a calcium citrate supplement superior to that of a commercially marketed calcium carbonate product. Since the two salts, in pure form, had been shown in several studies to be absorbed equally well [8–10], a question arose as to whether differences in pharmaceutical preparation of marketed products might have interfered with or enhanced the absorbability of one or the other preparation. Such absorptive effects, if they exist, would alter cost effectiveness calculations, once calcium actually delivered into the blood stream becomes the basis for the computation.

Accordingly we set out to compare two commercial supplements, using standard pharmacokinetic methods, both with one another and with non-pharmaceutical calcium carbonate ingested without excipients. This communication describes the results of this investigation. Additionally, we then used the bioavailability data to calculate the costs associated with providing the two commercially available calcium salts to the US population at greatest risk of hip fracture.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Subjects
Subjects were 24 postmenopausal women aged 56.1 ± 7.1 years and in good general health. Their BMI was 29.3 ± 5.2 kg/m². Thirteen subjects were receiving estrogen replacement therapy, and the remaining 11 were not. One was African-American; the others were Caucasian. Subjects taking calcium supplements were asked to abstain throughout the course of the study, starting at least one week in advance of the first test. Additionally, subjects were counseled by our research dietitian to hold calcium from dietary sources to under 400 mg/day, mainly by avoiding all dairy products. Also, they were instructed to avoid high sodium foods (such as commercial fast foods and canned soups or soup mixes) starting two days prior to and including each test day. To eliminate any variability in absorptive performance due to vitamin D insufficiency or to seasonal change in vitamin D status, all subjects were given 10 µg 25(OH)D₃ (Calderol®, Organon, West Orange, NJ)/day starting one week before the first test and continuing throughout the study. This dose is approximately equivalent to 1000 IU (25 µg) of cholecalciferol, but produces a rapid elevation of serum 25(OH)D, in contrast with the five month time-to-equilibrium required when using cholecalciferol. Further, this dose is the amount required, at Omaha’s latitude, to bring serum 25(OH)D concentration up to 32 ng/mL (80 nmol/L), a level widely considered to be the lower limit of physiological normal. The study was approved by the Creighton University Institutional Review Board, and each subject gave written consent.

Design
The study was a four-period, three-way randomized cross-over, within-subject design, with each individual receiving Os-Cal® (a product manufactured by GlaxoSmithKline and consisting of calcium carbonate derived from oyster shell), Citracal® (a product manufactured by Mission Pharmacal and consisting of calcium citrate), a gelatin capsule containing precipitated calcium carbonate or an empty gelatin capsule (the blank). The test source was ingested midway through a standard light breakfast containing two pieces of Italian-style white bread (Center-baked from unenriched flour), toasted and buttered, together with a cup of coffee, tea or water (with artificial sweetener if desired), plus additional water to ensure adequate urine volume. Blood samples were taken at 0, 1, 3, 5, 7, 9, 12, and 24 hours for measurement of total and ionized calcium and parathyroid hormone (PTH). Urine was collected in two pools, from 0 to 5 hours, and from 5 to 24 hours, and was analyzed for calcium, creatinine and sodium. Calcium sources were given only on the test day and only at the breakfast meal. The noon meal was provided by Center staff between the 5 and 7 hour blood draws and was designed to be low in both calcium and sodium. The evening meal was ingested between the 9 and 12 hour blood draws. Tests were separated typically by seven days; in this way the entire suite of studies was completed for most subjects within a 22-day period so as to minimize temporal variability in absorptive performance.

Test Sources
For the two commercial products (Os-Cal® and Citracal®), the sources were purchased from a retail pharmacy. The labeled content of elemental calcium for the Os-Cal® was 500 mg, plus 200 I.U. of vitamin D (Control No. 9K2228; exp. date 11/01). In order to approximate the load size of the Os-Cal®, the Citracal® dose required a combination of two different formulations, one labeled to contain 200 mg elemental calcium (Lot 9D12; exp. date 4/02) and the other 315 mg plus 200 I.U. vitamin D (Lot 9E86; exp. date 5/01). Precipitated calcium carbonate was prepared in the Center’s laboratory by dissolving reagent grade calcium chloride in distilled water, heating to 80°C with stirring and adding a slight stoichiometric excess of a heated aqueous solution of sodium carbonate, timed so that the reaction was completed within one minute. The resulting precipitate was collected on a fritted glass filter, washed with deionized water to remove adsorbed sodium chloride, dried at 90°C overnight, ground in a mortar and packed loosely into tared gelatin capsules in sufficient quantity to provide a 500 mg calcium load per dose. All preparations were chemically analyzed; actual ingested loads of calcium were as follows: for Os-Cal®, 503 mg; for Citracal®, 516 mg, and for precipitated calcium carbonate, 497 mg.

Analytical Methods
Calcium in serum, urine and the ingested sources was analyzed by atomic absorption spectrophotometry (AAnalyst 100, Perkin-Elmer, Norwalk, CT), creatinine in urine by an auto analyzer method (Chiron Express Plus, Ciba Corning Diagnostics, Medfield, MA) and sodium in urine by an ion selective electrode method (Cobas Integra, Roche Diagnostics, Basel, Switzerland). Serum ionized calcium was analyzed under standardized test conditions by an ion selective electrode method (Nova Nucleus, Nova Biomedical, Waltham, MA). Serum immunoreactive parathyroid hormone (iPTH) was measured as the intact molecule by IRMA (Nichols, San Juan Capistrano, CA).

Data Handling and Statistical Analysis
The primary outcome measures were the usual pharmacokinetic variables, area under the curve (AUC), both at five hours and at 24 hours (for both total and ionized serum calcium), as well as the time of maximum serum concentration (Tmax) and the magnitude of the elevation (Cmax). AUC was calculated by the trapezoidal method, and Cmax and Tmax were analyzed both by taking the observed values for concentration and time and by fitting the means of the timed serum increments for each source, using a first-order, two-compartment model with an absorptive delay of 0.5 hours (PKAnalyst; Micro-Math Scientific Software, Salt Lake City, Utah). The curves were plotted, and the pharmacokinetic parameters were calculated, both as the absolute values and as the increment above baseline. Secondary variables were serum iPTH and urine calcium, the latter with and without adjustment for urine sodium. AUC for iPTH was calculated using the same approach as for serum calcium. The sodium adjustment was made in two ways, using a slope factor of either 0.004 mg Ca/mEq sodium or 0.010 mg Ca/mEq Na. In each case adjustment was to the mean sodium excretion value for a given calcium source. The first factor is in the middle of the range reported in the literature for the relationship of urine calcium and sodium [11,12]. The second factor was derived from the slope of urine calcium to urine sodium observed with the blank meal in the subjects of this investigation. For the test calcium sources, urine calcium values are reported as the increment above the calcium content of the corresponding collections obtained on the test day with the blank load.

A standard bioequivalence analysis [13] was performed both on serum total and serum ionized calcium, using AUC from 0 to 5 and 0 to 24 hours, as well as Cmax and Tmax. AUC for serum PTH was also compared. Only the data from the first three periods were used in these bioequivalence analyses, since the treatment in the fourth period (non-pharmaceutical calcium carbonate) was not in random order. A general linear model was fit with the natural logarithm of the variate as the dependent variable, test source, sequence, period and subject nested in sequence as factors and the pre-dose value of the parameter as a covariate. The test sources in this equivalence analysis were Os-Cal®, Citracal® and blank. The sequence (or order) effect was tested using the subject in sequence mean square as the error term. The adjusted mean difference between the carbonate and citrate sources was computed and its 90% and 99% confidence intervals were constructed. The difference and the bounds of the confidence interval were exponentiated to obtain the ratio of the carbonate source mean to the citrate source mean and its confidence interval. As set forth in the applicable FDA Guidelines [13], if the confidence interval for the ratio fell in the range from 0.80 to 1.25, bioequivalence was considered to have been demonstrated.

Cmax and Tmax were compared between treatment groups using paired t tests. Pharmacokinetic parameters for Os-Cal® and Citracal® were each compared to blank using linear contrasts in the general linear model described above. Pharmacokinetic parameters for Os-Cal® and Citracal® were each compared to CaCO₃ using paired t tests. Changes from pre-dose serum concentrations of total and ionized calcium were compared among treatment groups at each time point using doubly-repeated measures ANOVA. Each pairwise comparison among test sources was tested and type I error was controlled at the 5% level using Holm’s step-down method. Urine calcium and sodium-adjusted urine calcium were compared between calcium sources using paired t tests. AUC values for incremental calcium and PTH ratio to baseline were correlated by standard Pearsonian regression. All of these analyses used within-subject differences to make inferences concerning the pharmacokinetic parameters, and in this way full adjustment was made for between-subject differences in absorptive efficiency.

To determine the cost of these supplements, we used the average price at all US outlets and also calculated separately the mass market costs/g of elemental calcium for Os-Cal® and for Citracal® between April and September 30, 2000. The data are provided by AC Nielson. The savings associated with hip fracture reduction were based on a previous analysis of this issue [4] for calcium supplements generally, which in turn used the average 1995 cost per discharged patient with a hip fracture, the size of the age cohort concerned and the fractional reduction in risk derived from published trials of calcium supplementation.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Table 1 presents the pharmacokinetic parameters for both total and ionized serum calcium for the four test sources, and Fig. 1 and 2 show the time courses of total and ionized calcium, respectively. The AUC values for the three calcium sources were all highly significantly different from the blank (p < 0.001), but there was no significant difference between the three calcium-containing sources for either of the AUC values or any of the other pharmacokinetic parameters. Also, as Fig. 1 shows graphically, the three sources produced virtually identical total serum calcium time courses, whether expressed as absolute values (Fig. 1A) or as increment above baseline (Fig. 1B). Serum calcium values differed significantly from the corresponding values following the blank load at all time points from 3 to 12 hours for Os-Cal® and from 1 to 9 hours for Citracal®, but there were no significant differences between the calcium sources at any time point. Fig. 2B shows that the incremental elevation of serum ionized calcium for the citrate source was somewhat greater from 5 to 12 hours compared to Os-Cal® and from 5 to 9 hours compared to the plain calcium carbonate. Consistent with this difference, the AUC₂₄ for ionized calcium (Table 1) was greater for the citrate than for the carbonate preparations. However, given the dispersion of the individual AUC data, none of these differences was statistically significant. There was no effect of the order of the test substance on any of the outcome variables. Similarly, age and estrogen status were also tested and were without effect on the relative absorbabilities of the test calcium sources.

View larger version (21K): [in this window] [in a new window]   Fig. 1. Time course of the total serum calcium, both as absolute values (A) and as increment above baseline (B), for the three calcium sources and for the blank load. Error bars are 1 SEM. (Copyright Robert P. Heaney, 2000. Used with permission.)

View larger version (22K): [in this window] [in a new window]   Fig. 2. Time course of the ionized serum calcium increment above baseline for the three calcium sources and for the blank load, both as absolute values (A) and as increment above baseline (B), for the three calcium sources and for the blank load. Error bars are 1 SEM. (Copyright Robert P. Heaney, 2000. Used with permission.)

View larger version (20K): [in this window] [in a new window]   Fig. 3. Time course of serum iPTH following ingestion of the three calcium sources and for the blank load, both as absolute values (A) and as fractional values relative to baseline (B). Error bars are 1 SEM. (Copyright Robert P. Heaney, 2000. Used with permission.)

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Shangraw [14] had previously shown marked differences in dissolution of calcium supplement preparations, due solely to pharmaceutical formulation differences, and unpublished experience of one of us (RPH) has demonstrated that not all preparations of the same salt exhibit equivalent absorbability. Finally, Heller et al. [7] explicitly raised this question in their recent paper. It is reassuring, therefore, to note that, in this study, Os-Cal® and the non-pharmaceutic, precipitated calcium carbonate exhibited identical bioavailability values. Thus for at least one marketed calcium carbonate product, pharmaceutical formulation does not alter the intrinsic bioavailability of its calcium salt. The same conclusion is probably applicable to the marketed citrate product as well. This is because it did not differ from non-pharmaceutic calcium carbonate in this study and because we had previously shown that the bioavailability values of the pure carbonate and citrate salts were identical [8].

Interestingly, however, and not previously described, several small differences were noted in pattern of response between the citrate and carbonate sources. None was statistically significant in isolation, but taken together, their mutual consistency suggests underlying differences in metabolic response to the two salts. These effects were i) although the rise in total calcium was the same, slightly less of the increment in serum calcium following the carbonate products was carried as the ionized form and slightly more as the bound form, relative to the citrate salt; ii) PTH suppression was slightly greater for the Citracal® than for the Os-Cal®, and the difference approximately coincided with the time points at which the ionized calcium differences were most prominent; and iii) urine calcium excretion in the 5 to 24 hour pool was higher for the Citracal® than for Os-Cal®. The relative depression is shown most clearly in Fig. 4, which plots ionized calcium as a percent of total calcium and shows slightly lower values for the Os-Cal® from 5 to 9 hours. This relative depression may reflect a very slight degree of alkalosis due to exhalation of CO₂ from the carbonate anion, but the reason for the delay after ingestion is unclear. Physiologically, these changes are mutually consistent, since a higher ionized calcium would be expected to lead to a greater depression of PTH release, to an increased filtered calcium load at the kidney and, through lowered PTH, to decreased tubular reabsorption of calcium. Although the greater rise in urine calcium with calcium citrate was not statistically significant in this study, it is worth noting that Heller et al. [7] reported a significant loss of calcium in urine following supplementation with calcium citrate (Citracal®) which was not seen with an equivalent dose of calcium carbonate (Os-Cal®).

View larger version (18K): [in this window] [in a new window]   Fig. 4. Time course of serum ionized calcium expressed as a percent of total serum calcium for the three calcium sources. Error bars are 1 SEM. (Copyright Robert P. Heaney, 2000. Used with permission.)

On a methodologic note, it may be worth mentioning that the increments in urine calcium were substantially more variable than the increments in serum calcium. The coefficients of variation (CVs) of the serum and urine calcium increments at their peak values (3 and 5 hours for serum and 0 to 5 hours for urine), for all calcium sources, were 38% to 60% for serum and 77% to 99% for urine. This roughly twofold greater variability underscores, as we have noted previously [8], the relative weakness of using the rise in urine calcium to estimate absorptive performance, particularly for loads as small as 500 mg.

For this study, the retail cost per 1000 mg of ingested calcium was between $0.16 and $0.20 for the marketed calcium carbonate product and between $0.24 and $0.38 for the marketed calcium citrate product. Since both sources exhibited equivalent bioavailability, it is clear that the carbonate source was the less expensive of the two per unit of absorbed calcium and would therefore exhibit a more favorable cost-benefit relationship in a cost-effectiveness analyses such as set forth in Table 4. Additionally, although not usually considered in cost benefit analysis, the greater calcium density of carbonate-based products means that fewer pills are needed to achieve a desired supplement intake, a factor known to influence patient compliance [15].

In this study we used 25(OH)D as a rapid and efficient means of ensuring approximately equivalent vitamin D status in all subjects. Such treatment would not be a part of population-level supplementation, and its costs are, accordingly, not a part of our calculations. Vitamin D is contained in both of the supplements tested here, and its cost is, therefore, already factored into the analysis summarized in Table 4.

While we tested only two commercially available products in this analysis, our purpose was not so much to contrast these two specifically as to use them as examples for a type of calculation and analysis that should be performed for all marketed calcium supplement products. It was beyond the scope of this project to undertake an exhaustive survey of different pharmaceutical formulations, although we believe this should be done. It is a matter of commonplace experience that there are many other calcium products available, at least some of which explicitly meet the USP disintegration and dissolution standards for calcium supplements (and therefore can be presumed to have a bioavailability comparable to what we found here). Their prices range from as low as $0.09 per 1000 mg to as much as $0.53. Lacking bioavailability data for most of these products, it is uncertain whether any of them would exhibit an advantage over the products tested here.

In conclusion, based upon bioavailability, cost and clinical efficacy, calcium carbonate, in the form of Os-Cal®, would appear to be a good choice for calcium supplementation in a US population at risk for both low bone mineral density and hip fracture.

	ACKNOWLEDGMENTS

 
Study supported both by Creighton University research funds and by a contract from GlaxoSmithKline. The assistance of Kurt Balhorn for the ionized serum calcium measurements is gratefully acknowledged.

Received December 11, 2000. Revised March 26, 2001. Accepted March 26, 2001.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Jackson RD, LaCroix A, Cauley J, McGowan J: WHI calcium and vitamin D trial baseline monograph. Annals Epidemiol (submitted) 2000.
2. Chrischilles EA: Public health implications of interventions to promote calcium intake: cost-benefit considerations. Paper presented to the NIH Consensus Development Conference on Optimal Calcium Intake, June 1994, Washington, DC.
3. Eddy DM, Johnson Jr CC, Cummings SR, Dawson-Hughes B, Lindsay R, Melton III LJ, Slemenda CW: Osteoporosis: Review of the evidence for prevention, diagnosis and treatment and cost-effectiveness analysis. Osteoporos Int 8(Suppl 4): S1–S88, 1998.
4. Bendich A, Leader S, Muhuri P: Supplemental calcium for the prevention of hip fracture: potential health-economic benefits. Clin Ther 21: 1058–1072, 1999.[Medline]
5. Torgerson DJ, Kanis JA: Cost-effectiveness of preventing hip fractures in the elderly population using vitamin D and calcium. Q J Med 88: 135–139, 1995.
6. Heller HJ, Stewart A, Haynes S, Pak CYC: Pharmacokinetics of calcium absorption from two commercial calcium supplements. J Clin Pharmacol 39: 1151–1154, 1999.[Abstract]
7. Heller HJ, Greer LG, Haynes SD, Poindexter JR, Pak CYC: Pharmacokinetic and pharmacodynamic comparison of two calcium supplements in postmenopausal women. J Clin Pharmacol 40: 1237–1244, 2000.[Abstract]
8. Heaney RP, Dowell MS, Barger-Lux MJ: Absorption of calcium as the carbonate and citrate salts, with some observations on method. Osteoporos Int 9: 19–23, 1999.[Medline]
9. Sheikh MS, Santa Ana CA, Nicar MJ, Schiller LR, Fordtran JS: Gastrointestinal absorption of calcium from milk and calcium salts. N Engl J Med 317: 532–536, 1987.[Abstract]
10. Recker RR: Calcium absorption and achlorhydria. N Engl J Med 313: 70–73, 1985.[Abstract]
11. Itoh R, Suyama Y: Sodium excretion in relation to calcium and hydroxyproline excretion in a healthy Japanese population. Am J Clin Nutr 63: 735–740, 1996.[Abstract/Free Full Text]
12. Nordin BEC, Need AG, Morris HA, Horowitz M: The nature and significance of the relationship between urinary sodium and urinary calcium in women. J Nutr 123: 1615–1622, 1993.
13. Center for Drug Evaluation and Research, Food and Drug Administration, US Dept of Health and Human Services: Statistical procedures for bioequivalence studies using a standard two-treatment crossover design 1992. http://www.fda.gov/cder/guidance/index/htm
14. Shangraw RF: Factors to consider in the selection of a calcium supplement. In "Proceedings of the 1987 Special Topic Conference on Osteoporosis." Public Health Rep S104: 46–49, 1989.
15. Eisen SA, Miller DK, Woodward RS, Spitznagel E, Przybeck TR: The effect of prescribed daily dose frequency on patient medication compliance. Arch Intern Med 150: 1881–1884, 1990.[Abstract]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Heaney, R. P. Articles by Bendich, A. Search for Related Content PubMed PubMed Citation Articles by Heaney, R. P. Articles by Bendich, A.