A novel dual radio- and stable-isotope method for measuring calcium absorption in humans: comparison with the whole-body radioisotope retention method

¹ From the Research Department of Human Nutrition, The Royal Veterinary and Agricultural University, Frederiksberg, Denmark (ABB, SB, CM, MH, and BS); the Department of Earth Sciences, Dartmouth College, Hanover, NH (SS); and the Department of Clinical Physiology and Nuclear Medicine, The Copenhagen University Hospital, Copenhagen (MJ and OWK).

² Reprints not available. Address correspondence to S Bügel, Research Department of Human Nutrition, The Royal Veterinary and Agricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark. E-mail: shb{at}kvl.dk.

ABSTRACT  
Background: Dietary calcium absorption can be determined only with the use of isotope techniques. Currently used isotope techniques require exclusive equipment or are not true tracer approaches.

Objective: The objective was to compare a dual-isotope method combining radioisotopes and stable isotopes with a whole-body radioisotope retention method for measuring calcium absorption.

Design: Seven healthy adults aged 21–27 y consumed a test meal containing 63 ± 14 ( Results: Mean (± SD) calcium absorption was 75 ± 9% with the dual-isotope method and was 74 ± 8% with the whole-body radioisotope retention method. There was a high degree of agreement between the methods.

Conclusion: The dual-isotope method is a valid approach for measuring calcium absorption from a single meal.

Key Words: Calcium absorption • radioisotopes • ⁴⁷Ca • stable isotopes • ⁴⁴Ca • whole-body counting • inductively coupled plasma mass spectrometry • ICPMS

INTRODUCTION  
Absorption of nutrients with a high intestinal reexcretion, such as calcium and zinc, can be measured with the use of radioactive or stable-isotope tracer techniques. The use of tracers is based on the fundamental assumption that an added isotope exchanges completely with the native element in the food and is absorbed and metabolized in the same proportions (1). Calcium absorption can be estimated from the direct measurement of the whole-body retention of a -emitting calcium radioisotope (2, 3). This method has a high measurement precision and is simple for the participating subjects. A potential disadvantage of this approach is that it requires an estimate of the endogenous losses during the time from the intake of the isotope until the time that the nonabsorbed isotope has been excreted. A power function was found suitable to describe the whole-body retention of radioisotope against time (4) and is used to extrapolate to the time for absorption (5, 6).

With simultaneous oral and intravenous administration of 2 different calcium isotopes, calcium absorption can be estimated from the relative enrichment or appearance of the 2 isotopes in urine or plasma. A double-radioisotope method based on this principle, and thus not requiring a human whole-body counter (7), has been used. However, this technique exposes the subject to relatively high radiation doses compared with the whole-body counting technique. As an alternative, a double stable-isotope method was developed (8, 9). This technique has the advantage that it can be used in studies of groups in whom radiation exposure should be restricted, such as children (10, 11) and pregnant and lactating women (12–16). Stable isotopes of calcium in urine can be measured with high precision and without time-consuming sample preparation by using high-resolution inductively coupled plasma mass spectrometry (ICPMS) (17). A potential disadvantage of the dual-isotope approach relative to the radiotracer technique is the amount of stable isotope that has to be administered orally to get a sufficient enrichment of the urine. In some studies, doses of up to 300 mg have been used (18). Improvements in mass spectrometric methods have made it possible to reduce the oral dose to 10 mg. However, in low-calcium meals this still represents a substantial increase in the total calcium content. In comparison, the radioisotope retention method uses nearly carrier-free ⁴⁷Ca solutions (< 0.1 mg/subject dose).

The aim of the present study was to evaluate the possibility of combining the advantages of the radioisotope and the stable-isotope methods into a novel dual-isotope technique for measuring the true absorption of calcium in humans. Absorption was estimated from the relation between orally administered radioisotope and intravenously administered stable isotope excreted in a urine sample. The results were compared with the results obtained with the whole-body radioisotope retention method.

SUBJECTS AND METHODS  
Subjects
Seven apparently healthy persons aged 21–27 y (3 women, 4 men) were recruited from local universities. To be eligible for inclusion, the subjects had to be nonsmokers, not be taking medicine daily, not be involved in sports on an athletic level, and not have received radioisotopes previously; the women had to be nonpregnant and nonlactating. The protocol was approved by the local Ethical Committee of Copenhagen and Frederiksberg [(KF) 01-238/98] and by the National Institute of Radiation Hygiene, Denmark. The volunteers were given oral and written information about the study protocol before they gave their written consent.

Study design
Absorption of calcium was measured from a single test meal in the morning after a 12-h fast. In addition to the meal, the subjects received 17 mg Ca labeled with ⁴⁷Ca (see below). The subjects were given 10–15 min to eat and alternated between eating and drinking. One hour after the test meal was finished, 18 mg ⁴⁴Ca was given intravenously. All feces and urine were collected from the day of isotope administration for 5 and 6 d, respectively. Whole-body retention of ⁴⁷Ca was measured 3 times weekly for 4 wk. The 24-h urine sample from day 2 was analyzed for excretion of ⁴⁷Ca and ⁴⁴Ca. To keep the subjects’ calcium intake constant during the study and at their habitual level, an experimental diet was provided for 1 wk: 2 d before and 5 d after ingestion of the test meal. During the following 3 wk, the subjects ate a self-selected diet but were instructed in how to keep their calcium intake constant by fixing the intake of dairy products at the habitual level indicated by a food-frequency questionnaire (see below). Compliance with these instructions was checked by interviews at each whole-body retention measurement.

Isotopes
The calcium solution provided as a drink with the test meal was extrinsically labeled with 0.11 MBq (per subject dose) ⁴⁷Ca (as ⁴⁷CaCl₂; Risø, Roskilde, Denmark), which resulted in an estimated whole-body equivalent radiation dose of 0.18 mSv. The labeling was done 1.5 h before administration to allow equilibration and isotope exchange. After the 200-mL Ca solution was consumed, the plastic mug was rinsed with 2 x 25 mL boiled deionized water, at a resistance of 18.2 cm (Millipore MilliQ water purification system; Millipore Corporation, Bedford, MA), which was also consumed.

Each subject received an intravenous dose of 18 mg ⁴⁴Ca in 10 mL isotonic saline. The dose was selected to enable evaluation of the minimum dose required for further similar studies. The stable isotope was purchased as ⁴⁴CaCO₃ (425.5 mg, purity 96.4%; Chemgas, Boulogne, France) and was dissolved in a few drops of a 12-mol HCl/L solution (analytic grade; Merck, Darmstadt, Germany). The pH was adjusted to 6 with a 1-mol NaOH/L solution (Bie & Berntsen A/S, Rødovre, Denmark), and the volume was adjusted to 100 mL with isotonic saline, giving a final concentration of 1.8 mg ⁴⁴Ca/mL. The solution was sterilized and packed in 10-mL vials for injection, with a final pH of 7.4. The exact amount of ⁴⁴Ca solution injected was determined by weighing the syringe before and after the injection on an electronic scale with a precision of 0.001 g. The infusion was made over 2 min. Before the infusion, passage of the intravenous cannula was secured with 5 mL isotonic saline. After injection of the stable isotope, an additional 5 mL was infused with the use of another syringe to ensure that the intravenous cannula was drained for ⁴⁴Ca and all the isotope had passed into the circulation. Blood pressure and heart rate were recorded for safety reasons.

Test meal
The test meal consisted of a bun with butter and raspberry marmalade, which was served with the radiolabeled calcium drink. The portion size of the bun was adjusted according to individual energy requirements. The calculated calcium content of the butter and the marmalade was 15 mg Ca/100 g (19). The calcium drink contained 17 mg Ca as calcium-L-lactate hydrate (Fluka Chemie, Buchs, Switzerland) dissolved in 200 mL boiled ultrapure water.

Experimental diet
Individual habitual calcium intake was measured by using a food-frequency questionnaire that was especially designed for estimating the intake of calcium in children and adolescents (20). A few questions were rephrased to match the older age group of the present study. The experimental diet was designed to match each individual’s average habitual calcium intake.

All meals were prepared in advance by the experimental kitchen and packed in portions matching each individual’s estimated energy requirement based on body weight and physical activity level (21). The lunch meal was eaten at the department, whereas breakfast and dinner (a hot meal) were eaten at home. Hot meals were stored at -20 °C and thawed and heated by the subjects themselves. Spring water (3.5 mg Ca/L; Harilds Kildevand, Rynkeby Food A/S, Ringe, Denmark) was the only drink allowed and was handed out for ad libitum use.

Urine and fecal collections
Six 24-h urine samples were collected in 2.5-L acid-washed bottles containing 10 mL of a 1-mol HNO₃/L solution (Suprapur; Merck, Darmstadt, Germany). Subjects were given p-aminobenzoic acid (PABA; 3 x 80 mg/d; Pharmacy of the Royal Veterinary and Agricultural University, Copenhagen) with the main meals over the 6-d collection period (22). Each 24-h urine sample was pooled, mixed, and weighed. Density was measured with a hand-held refractometer (Field; Bellingham & Stanley Ltd, Kent, United Kingdom).

Each single fecal excretion was collected in an acid-washed plastic container with a sealing lid. To monitor the collection and the transit of the radioisotope, radioopaque plastic markers in different shapes (Marker Capsules; Dunn Nutrition Center, Cambridge, United Kingdom) were given (23). One capsule containing 20 markers was taken at each of the 3 main meals for a total of 60 markers/d. Three different shapes were used during the study: one shape on each of the 3 d before ingestion of the test meal, a second shape on the day of the test meal, and a third shape on the following 4 d. Fecal samples were weighed and stored at -20 °C. After the excretion of ⁴⁷Ca was measured, the samples were freeze-dried and then X-rayed to count the excreted radioopaque plastic markers.

Measurement of ⁴⁷Ca
Whole-body retention of ⁴⁷Ca was measured with a whole-body counter at The Copenhagen University Hospital (Rigshospitalet, Copenhagen). Measurements were taken 12 times after intake of the test meal (nominally day 0) on days 1, 2, 4, 7, 9, 11, 15, 17, 18, 22, 24, and 25. The measurements were corrected for chamber background radiation and for the individual initial radionuclide burden in each subject, mainly due to naturally occurring ⁴⁰K, as measured by a background measurement made before administration of any activity. All measurements of ⁴⁷Ca were corrected for physical decay back to the time of administration. The whole-body counter used consisted of a lead-lined steel chamber (Nuclear Enterprises Ltd, Edinburgh) with 8 large-volume plastic scintillation detectors in standard "bed" configuration connected to conventional nuclear electronic modules and to a multichannel analyzer system. The system has almost 4 measuring geometry. Counting with plastic scintillators relies on the detection of the Compton events in the detector. The gain and window setting for each detector block were adjusted initially by using a 2-kBq ⁶⁰Co point source, thus establishing the uniform energy and efficiency response of each detector with a lower level cutoff just above the Compton edge for ^99mTc radiation (141 keV). The detector efficiency constancy (better than 1%) was checked on every measurement day by using the ⁶⁰Co standard.

Counting efficiency was established through measurements of water-filled phantoms with weights (55, 67, and 88 kg) and outlines approximating those of humans. In each instance, the phantoms were filled with an aliquot of radioisotope solution matching the dose given to the subjects. The detection efficiency ranged in a linear manner from 0.175 to 0.182 cps/Bq, going from the lightest to the heaviest phantom. To minimize contamination with atmospheric background activity, the subjects had a shower, including a hair wash, and were dressed in hospital clothing before each measurement. Acquisition time was always set to 600 s live time, thus providing automatic dead-line correction.

Excretion of ⁴⁷Ca in 24-h urine samples and in fecal samples was measured with a ring array of 4 NaI scintillation detectors (0.1 x 0.15 m, or 4 x 6 in) in a low-background shielding. Detection efficiency for the 24-h urine samples (topped up to 2.5 L with water) and the fecal samples was established by phantoms prepared from the same radioisotope solution used to prepare the test meals. The detection efficiency for urine and feces was 0.03 cps/Bq.

All whole-body, urine, and feces measurements were background subtracted, corrected for the physical decay of ⁴⁷Ca back to the day that the dose was administered, converted to activity by using the weight-matching detection efficiency, and expressed as a fraction of administered dose. Because of the high counting efficiency, the statistical errors of all whole-body counter measurements were < 0.5%, even the last measurement on a subject with a low uptake. Thus, the main contribution to measurement error in the whole-body counter must be attributed to variations in the whole-body distribution. From experiments using the abovementioned phantoms, but with nonuniform radioisotope distribution (as, for example, a limb-free distribution), the total error is known to be < 5%. This uncertainty is reflected directly into a worst-case measurement error of the whole-body retention of ⁴⁷Ca as 5% of the actual measurement, or 2.5% of the administered dose.

The relative statistical error in the urine count is more important (5%) because of the lower counting efficiency, which was chosen to minimize volumetric and matrix effects. The error in the urine measurements on days 1 and 2 adds directly to the error in the calculated absorption values because the urinary-excreted calcium is assumed to be part of the initially absorbed calcium. However, the total urinary excretion is low (1–2% of the daily dose), thus minimizing the influence of urine counting errors to < 1% of the administered dose.

In estimating the absorption with the use of the fecal recovery method, the errors in the feces counting have to be considered. For most of the subjects, most of the activity was excreted on days 1 and 2, giving a small relative error. For samples derived from the end of the experiment with low activity, and therefore quantitatively less important, the relative error could be as much as 10%.

Measurement of ⁴⁴Ca and total calcium in urine samples by ICPMS
The total calcium content and the enrichment of ⁴⁴Ca in the 24-h urine sample from day 2 were determined by using an Element1 sector field inductively coupled plasma mass spectrometer (Finnigan MAT GmbH, Bremen, Germany) equipped with a shielded torch in operation at the Dartmouth Trace Metal Analysis Core Facility (Dartmouth College, Hanover, NH). Throughout this study, the medium mass resolution setting was applied, which produces a mass resolution of 4000. All measurements were made with a standard sample introduction system that comprises a peristaltic pump (Perimax 12; Spetec, Erding, Germany), a concentric nebulizer (AR35–1-F04; Glas Expansion, Hawthorn, Australia), and a Scott-type spray chamber maintained at 6 °C. All samples and standards were prepared by dilution with 0.45 mol nitric acid/L. The nitric acid was prepared from 65% baseline HNO₃ (Seastar Chemicals, Sidney, Canada) diluted with ultrapure water (> 18.2 cm). Calcium standard solutions were prepared from a 1000-mg Ca/L standard (VHG Labs, Manchester, NH) by dilution with 0.45 mol HNO₃/L. A 1-mg/L standard solution was used for instrument optimization. The acidified human urine was diluted 40 times with 0.45 mol HNO₃/L and thereafter was aspirated directly into the ICPMS instrument. A detailed description of the method was recently published (24).

The percent atom enrichment (%XS⁴⁴Ca) of ⁴⁴Ca achieved in the urine sample from day 2 was calculated from the ⁴⁴Ca-⁴²Ca ratios measured in the urine sample and the baseline urine sample from day 2, collected before administration of ⁴⁴Ca. The measurement precision of the ⁴⁴Ca-⁴²Ca ratio is 0.05% relative SD (RSD) (24).

Total calcium was determined by isotope-dilution ICPMS spiking with enriched ⁴⁴Ca (97% ⁴⁴Ca) as an in-sample calibration to obtain more accurate results than possible with normal calibration applying external calibration or standard additional calibration. The enriched ⁴⁴Ca (as ⁴⁴CaCO₃) was obtained from Chemgas (Boulogne, France). Spiking the samples with 5 µL of enriched ⁴⁴Ca increased the ⁴⁴Ca-⁴²Ca ratio from 3.5–3.6 to 4.5–6.0 in the sample-spike mixture (24). The measurement precision was 3.0% RSD (n = 10), and the accuracy was verified by analyzing reference urine samples (Lyphocheck, 62081 and 62082; Bio-Rad Laboratories, ECS Division, Anaheim, CA). Values of 45.5 mg/L (certified range: 38–46 mg/L) and 121.3 mg/L (certified range: 98–128 mg/L) were obtained for the 2 samples that showed a high accuracy of the analytic method (24).

Measurement of total calcium in diet
The calcium content was analyzed in duplicate portions of the diets and the buns. The foods were homogenized, lyophilized, and microwave digested (MES-1000; CEM Corporation, Matthews, NC) with 65% HNO₃ (Suprapur) and 30% H₂O₂ (Suprapur). Calcium was measured by atomic absorption spectroscopy (SpectraAA-200; Varian Techtron Pty Ltd, Australia) after dilution with a lanthanium oxide solution (0.5% lanthanium oxide, 1% HNO₃, and water; Merck). Standards were prepared from a 1000-mg Ca/L standard (Tritisol; Merck) by dilution with the lanthanium oxide solution. A reference diet (Standard Reference Material 1548a; National Institute of Standards and Technology, Gaithersburg, MD) was analyzed in the same run. The analyzed calcium content was 1.94 ± 0.060 mg/g, and the certified value was 1.967 ± 0.113 mg/g. The CV was 3.1% (n = 4).

p-Aminobenzoic acid
The content of PABA in the 24-h urine samples was analyzed according to Bingham and Cummings (22) with a Gilford Staser Spectrophotometer (Gilford Instrument Laboratories, Inc, Orberlin, OH). Urine samples containing < 80% of the PABA dose (one collection from day 0 and one collection from day 4) were excluded from the estimates of total urinary calcium excretion but were used in the comparison of methods, because the same volume was used for both isotope measurements.

Calculation of calcium absorption with the whole-body retention method
Absorption of calcium was calculated according to Hansen et al (25, 26) on the basis of the power function given by Norris et al (4) and used by Nielsen et al (5). The measurements of retention, expressed as a percentage of the administered dose, were plotted against time in a log-log plot in which the power function is a straight line:

RESULTS  
The mean habitual calcium intake of the subjects was 895 ± 524 mg/d, which was well matched by the calcium content of the experimental diet (956 ± 483 mg/d). The total calcium content of the test meal was 52 ± 5 mg for women and 71 ± 13 mg for men. The intravenous dose of ⁴⁴Ca was 17.26 ± 0.995 mg.

Four days after administration, most of the fecal markers derived from the day of the test meal had been excreted (Table 1). The cumulative urinary and fecal excretion of ⁴⁷Ca on day 4 was 46.5 ± 7.7% of the dose. When differences in timing of the fecal samples and whole-body retention measurements were accounted for, the total recovery of the radioisotopes was 95.0 ± 2.7%. The losses of urinary radioisotope were highest in the first 24 h and decreased to < 1% of the dose after 48 h.

View this table:
TABLE 1 . Whole-body retention and fecal and urinary excretion of ⁴⁷Ca, excretion of fecal markers, and recovery of p-aminobenzoic acid (PABA) given each day after ingestion of the test meal¹  
The estimated calcium absorption from the test meal is given in Table 2. No significant difference between the absorption values determined by the whole-body-retention method and the dual-isotope method was observed, and there was good agreement between these 2 methods (mean difference: 0.43%; 95% CI: -8.9, -9.8%) (Figure 1). The fecal recovery method resulted in significantly lower absorption values when the recovery of 60 markers was used as the cutoff for the inclusion of fecal samples, whereas the recovery of 20 markers appeared to overestimate absorption (Table 2).

View this table:
TABLE 2 . Fractional absorption of calcium estimated with the whole-body radioisotope method, the double-isotope method, and the fecal recovery method  

View larger version (13K):
FIGURE 1. . A Bland and Altman plot of the difference in absorption between the dual-isotope method and the whole-body radioisotope retention method against the mean absorption of the 2 methods. The plot shows agreement between the 2 methods. The solid line represents the mean of the differences, and the dashed line represents the 95% CI of the differences ( ± two-tailed value for 6 df  

DISCUSSION  
This study showed that calcium absorption can be estimated with a simple method not requiring access to a human whole-body counter or extensive sample preparations. The intravenous dose of 17 mg ⁴⁴Ca resulted in an enrichment of ⁴⁴Ca in the 24-h urine sample from day 2 of 10.9 ± 2.1%; the minimum enrichment obtained was 8.2%. The applied dose is relatively high, but was used to ensure significant atomic enrichment in the day 2 urine sample and to allow evaluation of the minimum doses required in later experiments. In future experiments, a lower dose of ⁴⁴Ca can be used because it is customary to design nutritional experiments to result in an average shift in isotope ratio of 10 times the %RSD by which the isotope ratio involved can be measured (the quantification limit) (28). For the method presented here (%RSD = 0.05 for ⁴⁴Ca:⁴²Ca determinations), future nutritional experiments should be designed to produce a minimum %XS⁴⁴Ca of 0.5. Because the relation between dose and %XS does not follow any exact (known) equation but depends on factors such as age, sex, calcium status, and diet composition, we suggest that future experiments be designed to produce an %XS of 1–2% to ensure that no subject will have an enrichment < 0.5%. Given that future groups of subjects will have a calcium absorption similar to that in the subjects in the present study, the ⁴⁴Ca dose could be substantially reduced to 2–4 mg ⁴⁴Ca.

The main disadvantage of the suggested dual-isotope method is that it cannot be used in populations in whom radiation exposure should be restricted, such as infants, children, and pregnant and lactating women. Note, however, that because of the short half-life of ⁴⁷Ca, the radiation dose is low and under specific clinical conditions could be considered as an alternative.

A technical disadvantage of the method is that 3 separate analyses of the day 2 urine sample are required: the determination of ⁴⁴Ca:⁴²Ca by ICPMS, of total calcium by isotope-dilution ICPMS, and of ⁴⁷Ca by -ray detection. Therefore, the procedure is more time consuming than is the traditional double-stable-isotope procedure, in which only one isotope ratio measurement (eg, determination of ⁴⁴Ca:⁴³Ca and ⁴²Ca:⁴³Ca) is needed with ICPMS. The big advantage on the other hand is that the combined uncertainty of the estimated calcium absorption is improved because ICPMS analysis of the minor ⁴³Ca isotope is omitted with the dual-isotope procedure, ⁴³Ca has a low abundance and therefore the ICPMS signal is smaller than that for the other larger calcium isotopes and has a larger uncertainty attached to it, which is reflected in a higher overall combined uncertainty of the determined calcium absorption whenever ⁴³Ca is used in the measurements. The combined uncertainty (using the above setup) was estimated to be 6% RSD (24). Therefore, in our experience, more accurate estimates of calcium absorption can be obtained with the use of the dual-isotope procedure (ICPMS and -ray detection) or the whole-body radioisotope retention method than with the double-stable-isotope procedure (⁴⁴Ca, ⁴²Ca, and ICPMS detection) (24, 29).

The cost of the radioisotopes needed for the suggested dual-isotope technique is the same as that for the stable isotopes. The cost for the doses of ⁴⁷Ca used in this study were $216/person, whereas 10 mg ⁴⁴Ca and ⁴²Ca would cost $160/person and $800/person, respectively.

A similar dual-isotope procedure with ⁴⁷Ca as the oral tracer and ⁴⁸Ca as the intravenous tracer was published by Neer et al (30). They did not make a direct mass spectrometric determination of ⁴⁸Ca, but converted ⁴⁸Ca to ⁴⁹Ca by irradiation and measured ⁴⁹Ca by neutron activation. For the ICPMS method presented herein, ⁴⁴Ca was chosen as the intravenous tracer because ⁴⁴Ca can be measured free of interference with the use of high-resolution ICPMS and a mass resolution setting of 4000, which resolves the ⁴⁴Ca from known interferences. It is not possible to resolve the titanium interference on ⁴⁸Ca with the use of even the highest possible mass resolution (10 000).

To follow the ⁴⁷Ca excretion in urine for 5 d and to estimate the absorption based on whole-body retention measurements, total 24-h urinary collections were made. However, for the suggested dual-isotope method, incomplete collections could still be used because the volume is eliminated in the ratio calculations. It may even be possible to use collections over < 24 h. For the stable-isotope measurements, a spot urine sample collected after 24–30 h, for example, would be advantageous because it is assumed that most of the applied calcium spike will be excreted in the urine in the first 12–24 h after intake. Collecting and pooling the urine for 48 h, therefore, dilutes the applied calcium spike. Theoretically, the radioactivity could be measured with a conventional -counter, but this would require a long measuring time (3 h with the doses used in this study) to achieve sufficient counting statistics. Large-volume (100 mL) well counters, small-animal whole-body counters, or preconcentration of the radioisotopes in the sample by evaporation or precipitation could reduce counting time but would require more sample handling than would the procedure used in the present study.

The results of the present study clearly show that despite a high precision in the radioisotope measurements, the fecal monitoring technique we used was not suitable for calcium absorption measurements because of the high endogenous fecal calcium excretion and the difficulties involved in separating fecal excretion of unabsorbed ⁴⁷Ca from the test meal and reexcreted ⁴⁷Ca. Intestinal transit time, especially in the large bowel, is relatively long, and the mixing of the intake from several meals or of the intake over several days cannot be avoided. The use of radioopaque markers is simple from an analytic point of view because no preparation of the fecal samples is required, but it may not reflect the true transit of the unabsorbed isotope. An alternative approach with similar analytic simplicity would be the administration of a nonabsorbed radioisotope, eg, ⁵¹Cr (31, 32), which could be measured simultaneously with ⁴⁷Ca with the setup used in this study. Stable isotopes of rare earth metals (dysprosium, samarium, and ytterbium) have also been used in the fecal recovery method (33, 34). However, this requires more extensive fecal sample handling and preparation than does the direct measurement of radioisotopes.

A comparison of calcium absorption on day 4 estimated with the whole-body retention method or with the fecal recovery method suggests that 14–15% of the absorbed ⁴⁷Ca was reexcreted in the feces (Table 2). This is about twice the amount of calcium excreted in the urine, and if the same specific activity as in urine is assumed for the intestinal losses, endogenous calcium intestinal excretion would be 200–250 mg/d. However, if it can be shown that a constant calcium intake results in a fairly constant endogenous intestinal excretion in each subject, the fecal recovery method could be used to compare calcium absorption from different meals in the same subject. Theoretically, endogenous calcium excretion could be estimated from the excretion of the intravenously administered stable isotope; however, this would require extensive sample preparation. In conclusion, the dual-isotope method is a precise and valid technique for determining calcium absorption from single meals.

ACKNOWLEDGMENTS  
We thank the subjects who participated in this study for their cooperation, the technicians Vivian Anker and Hanne Lysdal Petersen for their excellent assistance with the biological sample collections and chemical analyses, and Susanne Svalling for her qualified work on the whole-body counting measurements.

REFERENCES  

1. Fairweather-Tait SJ, Fox TE. Intrinsic and extrinsic labeling of inorganic nutrients in food studies. In: Mellon FA, Sandström B, eds. Stable isotopes in human nutrition. London: Academic Press 1996:15–21.
2. Sjöberg HE, Reizenstein P, Arman E. Retention of orally administered ⁴⁷Ca in man measured in a whole-body counter. Scand J Clin Lab Invest 1970;26:67–71.
3. Shipp CC, Maletskos CJ, Dawson-Hughes B. Measurement of ⁴⁷calcium retention with a whole-body counter. Calcif Tissue Int 1987;41:307–12.
4. Norris WP, Tyler SA, Brues AM. Retention of radioactive bone-seekers. Science 1958;128:456–62.
5. Nielsen SP, Bärenholdt O, Munck O. Fractional intestinal absorption and retention of calcium measured by whole-body counting. Application of a power function model. Scand J Clin Lab Invest 1975;35:197–203.
6. Isaksson M, Fredlund K, Sandberg AS, Almgren A, Rossander-Hulthen L. Determination of the retention of ⁴⁷Ca by whole-body counting. Appl Radiat Isot 2000;52:1441–50.
7. DeGrazia JA, Ivanovich P, Fellows H, Rich C. A double isotope method for measurement of intestinal absorption of calcium in man. J Lab Clin Med 1965;66:822–9.
8. Smith DL, Atkin C, Westenfelder C. Stable isotopes of calcium as tracers: methodology. Clin Chim Acta 1985;146:97–101.
9. Yergey AL, Vieira NE, Covell DG. Direct measurement of dietary fractional absorption using calcium isotopic tracers. Biomed Environ Mass Spectrom 1987;14:603–7.
10. O’Brien KO, Abrams SA, Liang LK, Ellis KJ, Gagel RF. Increased efficiency of calcium absorption during short periods of inadequate calcium intake in girls. Am J Clin Nutr 1996;63:579–83.
11. Abrams SA, Copeland KC, Gunn SK, Stuff JE, Clarke LL, Ellis KJ. Calcium absorption and kinetics are similar in 7- and 8-y-old Mexican-American and Caucasian girls despite hormonal differences. J Nutr 1999;129:666–71.
12. Kent GN, Price RI, Gutteridge DH, et al. The efficiency of intestinal calcium absorption is increased in late pregnancy but not in established lactation. Calcif Tissue Int 1991;48:293–5.
13. Cross NA, Hillman LS, Allen SH, Krause GF, Vieira NE. Calcium homeostasis and bone metabolism during pregnancy, lactation, and postweaning: a longitudinal study. Am J Clin Nutr 1995;61:514–23.
14. Fairweather-Tait S, Prentice A, Heumann KG, et al. Effect of calcium supplements and stage of lactation on the calcium absorption efficiency of lactating women accustomed to low calcium intakes. Am J Clin Nutr 1995;62:1188–92.
15. Ritchie LD, Fung EB, Halloran BP, et al. A longitudinal study of calcium homeostasis during human pregnancy and lactation and after resumption of menses. Am J Clin Nutr 1998;67:693–701.
16. Moser-Veillon PB, Mangels AR, Vieira NE, et al. Calcium fractional absorption and metabolism assessed using stable isotopes differ between postpartum and never pregnant women. J Nutr 2001;131:2295–9.
17. Stürup S, Hansen M, Mølgaard C. Measurements of ⁴⁴Ca:⁴³Ca and ⁴²Ca:⁴³Ca isotopic ratios in urine using high resolution inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 1997;12:919–23.
18. Yergey A, Abrams S, Vieira NE, Aldroubi A, Marini J, Sidbury JB. Determination of fractional absorption of dietary calcium in humans. J Nutr 1994;124:674–82.
19. Møller A, Saxholt E. The composition of foods. Copenhagen: Gyldendals Forlagsekspedition, 1996.
20. Mølgaard C, Sandström B, Michaelsen KF. Evaluation of a food frequency questionnaire for assessing of calcium, protein and phosphorus intakes in children and adolescents. Scand J Nutr/Näringsforskning 1998;42:2–5.
21. WHO. Energy and protein requirements. Geneva: WHO, 1985.
22. Bingham S, Cummings JH. The use of 4-aminobenzoic acid as a marker to validate the completeness of 24 h urine collections in man. Clin Sci 1983;64:629–35.
23. Cummings JH, Jenkins DJA, Wiggins HS. Measurement of the mean transit time of dietary residue through the human gut. Gut 1976;17:210–8.
24. Stürup S. Uncertainty calculation on the determination of calcium absorption in nutritional experiments with enriched stable isotopes and detection by double focusing sector field ICP-MS with a shielded torch. J Anal Atom Spectrom 2002;17:1–7.
25. Hansen M, Sandström B, Jensen M, Sørensen SS. Casein phosphopeptides improve zinc and calcium absorption from rice-based but not whole-grain infant cereal. J Pediatr Gastroenterol Nutr 1997;24:56–62.
26. Hansen M, Sandström B, Jensen M, Sørensen SS. Effect of casein phosphopeptides on zinc and calcium absorption from bread meals. J Trace Elem Med Biol 1997;11:143–9.
27. Bland JM, Altman DG. Statistical methods for assessing agreement between 2 methods of clinical measurement. Lancet 1986;8:307–10.
28. Crews HM, Mellon FA. Stable isotopes in human nutrition—inorganic nutrient metabolism. London: Academic Press, 1996.
29. Hansen M, Stürup S, Mølgaard C, Jensen H, Sørensen SS, Sandström B. Methods for estimating calcium absorption: comparison of a stable isotope and a radioisotope method. Scand J Nutr 1999;2(suppl):S75.
30. Neer R, Tully G, Schatz P, Hnatowich DJ. Use of stable ⁴⁸Ca in the clinical measurement of intestinal calcium absorption. Calcif Tiss Res 1978;26:5–11.
31. Davidsson L, Cederblad Å, Lönnerdal B, Sandström B. Manganese retention in man: a method for estimating manganese absorption in man. Am J Clin Nutr 1989;49:170–9.
32. Flanagan PR, Cluett J, Chamberlain MJ, Valberg LS. Dual-isotope method for determination of human zinc absorption: the use of a test meal of turkey meat. J Nutr 1985;115:111–22.
33. Schuette SA, Janghorbani M, Young VR, Weaver CM. Dysprosium as a nonabsorbable marker for studies of mineral absorption with stable isotope tracers in human subjects. J Am Coll Nutr 1993;12:307–15.
34. Fairweather-Tait SJ, Minihane AM, Eagles J, Owen L, Crews HM. Rare earth elements as nonabsorbable fecal markers in studies of iron absorption. Am J Clin Nutr 1997;65:970–6.