Biochemical effects of calcium supplementation in postmenopausal women: influence of dietary calcium intake 1 3

Transcription

1 Biochemical effects of calcium supplementation in postmenopausal women: influence of dietary calcium intake 1 3 Patrice Fardellone, Michel Brazier, Saïd Kamel, Jean Guéris, Anne-Marie Graulet, Jacques Liénard, and Jean-Luc Sebert ABSTRACT We studied the biochemical effects of calcium supplementation during a 2-mo course in postmenopausal women (x ± SD: 64 ± 5 y of age and 14.5 ± 6.7 y since menopause). The effects on calcium homeostasis and bone remodeling were assessed after 1 and 2 mo of daily administration of either calcium carbonate (1200 mg elemental Ca/d, n = 60) or a placebo (n = 56). The daily dietary calcium intake assessed before the beginning of calcium supplementation was 786 mg/d. We found a significant inverse relation between baseline intact parathyroid hormone (ipth) and dietary calcium intake before supplementation (r = 0.48, P = ). A significant increase in urinary excretion of pyridinoline was observed when the dietary calcium intake was lower than the median value. Calcium supplementation resulted in a significant increase in 24-h urinary calcium (39%, P < 0.02) and a significant reduction of bone alkaline phosphatase at 2 mo and of all bone-resorption markers (hydroxyproline, pyridinoline, and deoxypyridinoline) at 1 and 2 mo without significant changes in PTH fragments or ipth concentrations. When the dietary calcium intake was low (mean ± SD: 576 ± 142 mg/d), calcium supplementation was responsible for a greater increase in urinary calcium excretion and a greater decrease in markers of bone turnover. The greatest variations were observed for deoxypyridinoline at 1 and 2 mo ( 18.5%, P < 0.05) and for pyridinoline at 1 mo ( 16.3%, P < 0.01). Two months of calcium supplementation in postmenopausal women was efficient in reducing markers of bone turnover, with a greater effect in women with a low dietary calcium intake. Am J Clin Nutr 1998;67: KEY WORDS Calcium supplementation, postmenopausal women, bone-remodeling markers, parathyroid hormone, 1,25- dihydroxycholecalciferol, 25-hydroxycholecalciferol, dietary calcium intake, calcidiol, calcitriol, hydroxyproline, pyrodinoline, deoxypyridinoline, bone alkaline phosphatase INTRODUCTION Bone loss in women starts at the time of menopause and persists at an accelerated rate for up to 10 y. Afterward, continuous bone loss at a slower rate is observed, defined as age-related bone loss (1). It is agreed that postmenopausal bone loss is due to estrogen deficiency, which is the major etiologic factor of postmenopausal osteoporosis. Several reports suggest that increased parathyroid hormone (PTH) concentrations are involved in the development of age-related bone loss (2, 3). Serum PTH concentrations increase with age and this mainly results from a reduction of both 1,25-dihydroxycholecalciferol (calcitriol) production and gastrointestinal calcium absorption (4). Vitamin D and calcium supplementation in the elderly can reduce secondary hyperparathyroidism (5, 6), leading to a reduction in bone turnover (7) and prevention of hip fractures (6). After menopause, bone loss is also considered to be the result of a negative external calcium balance (8). It was suggested that the previous recommended dietary allowance of 800 mg Ca is not sufficient to prevent this negative calcium balance, which led to the proposal of an optimal calcium intake of 1500 mg/d after menopause (9). The benefits of high dietary calcium intake and calcium supplementation to prevent both postmenopausal bone loss and osteoporotic fractures are still being debated (10, 11). Most studies favor a positive role of calcium, particularly in women long past menopause (10, 12, 13) and when the usual dietary calcium intake is low (14). The effects of calcium supplementation on biochemical markers of bone remodeling are not yet well documented. The present study was designed to evaluate the effects of a 2-mo course of calcium supplementation in healthy postmenopausal women aged 65 y by assessing the biochemical indexes of calcium homeostasis, particularly intact PTH (ipth) and fragments of parathyroid hormone ([44 68]PTH), and by measuring the biochemical markers of bone formation, such as osteocalcin (b-gp), total (t-ap) and bone (b-ap) alkaline phosphatase, and of bone resorption such as urinary excretion of total hydroxyproline and hydroxypyridinium derivatives, pyridinoline and deoxypyridinoline. Pyridinoline and deoxypyridinoline are cross-links found in mature collagen (15) and their urinary excretion is a sensitive index of bone breakdown in the course of various metabolic disorders (16). SUBJECTS AND METHODS Subjects 1 From the Departments of Rheumatology, Gerontology, and Biochemistry, CHU, Amiens, France; the Laboratory of Clinical Pharmacy, Faculty of Pharmacy, Amiens, France; Laboratory of Biochemistry, Lariboisière Hospital, Paris. 2 Supported by Lederle Laboratories, Rungis, France. 3 Address reprint requests to P Fardellone, service de Rhumatologie, CHU NORD Amiens Cedex 1, France. michel.brazier@sa. upicardie.fr. Received February 10, Accepted for publication November 12, Am J Clin Nutr 1998;67: Printed in USA American Society for Clinical Nutrition 1273

2 1274 FARDELLONE ET AL We studied 116 healthy women <75 y of age (x ± SD: 64 ± 5 y of age) and >5 y postmenopause (time since menopause: 14.5 ± 6.7 y). Women with severe vitamin D insufficiency, defined by 25-hydroxycholecalciferol (calcidiol) concentrations <15 nmol/l (< 6 ng/ml) (17), were excluded from this study. None had any known systemic disease or was taking any medication (eg, calcium, vitamin D, phosphate, fluoride, bisphosphonate, or estrogens) able to influence calcium homeostasis or bone status. In addition, the subjects answered a food-frequency questionnaire before the beginning of the study to assess their daily calcium intake (18). Protocol The subjects were randomly assigned to receive either calcium carbonate in a solid oral form (Caltrate; Lederle Laboratories, Paris; n = 60) or similar-looking placebo tablets (n = 56) for 2 mo. Active treatment (1200 mg elemental Ca/d) and placebo consisted of two doses to be taken every day with the morning and evening meals. Specimen collection Blood and urine samples were collected at baseline and at 1 and 2 mo of treatment. Blood and 2-h second-void urine samples were taken in the morning before the calcium or placebo administration. The subjects were studied while consuming their usual diet without specific recommendations but were asked to keep fasting from 2100 the day before sampling. Blood samples were drawn at 0800, serum was removed, then frozen at 20 C. The 2-h fasting urine samples collected between 0800 and 1000 and the 24-h urine samples of the previous day were stored at 20 C. The study protocol was approved by the independent local ethics committee of Picardie (Amiens, France). Written informed consent was obtained from all the patients and the trial was conducted in accordance with good clinical practices. Biochemical measurements In serum Measurements of calcidiol were performed by using a competitive protein binding assay after ethanol extraction followed by chromatographic purification (19). ipth was measured by an immunoluminometric assay (Magic Lite intact PTH; Ciba-Corning, Cergy-Pontoise, France), [44-68]PTH by a radioimmunoassay method (20), intact osteolcalcin by chemoluminometric diagnostic kits (Lumitest osteocalcin; Brahms Diagnostica, Berlin) and b-ap by immunoenzymatic kits (Alkphase-B; Metra-Biosystems, Mountain View, CA). Serum t-ap and calcium were measured by colorimetric methods using an Ektachem 500 autoanalyzer (Johnson and Johnson Company, Rochester, NY). Interassay reproducibility, expressed as CVs, was 10.8% for calcidiol, 3.4% for ipth, 9.2% for [44 68]PTH, 12% for osteolcalcin, 7.8% for b-ap, and 1.4% for calcium. In urine Measurements were performed on the 2-h fasting urine samples and expressed as a ratio with creatinine. Total hydroxyproline was assessed by using HPLC (21) after acid hydrolysis of urine samples (at 102 C for 18 h) to release the amino acids from their peptidyl forms (hydroxyproline kit; Bio-Rad Laboratory, Ivry sur Seine, France). The interassay CV of the chromatographic assay was 3.9%. Total pyridinoline and deoxy-pyridinoline were measured, after hydrolysis of samples (at 125 C for 3 h) to release the three hydroxypyridinium derivatives from their conjugated forms, by extraction on cellulose and HPLC separation (22). The interassay CVs of this assay, including the extraction step, were 5.4% for pyridinoline and 8.2% for deoxypyridinoline. Urinary calcium and creatinine were determined by colorimetric methods using an automated system (Ektachem 500). Urinary calcium was measured in 2-h fasting urine samples (expressed as a ratio with creatinine) and also in 24-h urine collections. The interassay reproducibility of urinary calcium and creatinine, expressed as CVs, was 2.6% and 5.3%, respectively. Statistical analysis All data are expressed as means ± SDs. Homogeneity of the two treated groups was assessed on baseline characteristics by using Student s t test. The level of significance was established at 5%. Comparisons between placebo and calcium-treated groups were performed by using two-way analysis of variance for repeated measures. In case of significant interaction between time and treatment, a complementary analysis was done by comparing the two groups at each time according to Winer s procedure (23). Moreover, Dunnett s tests were performed within groups to assess changes from baseline. The relation between dietary calcium intake and ipth was tested by using linear regression analysis. Because ipth and dietary calcium intake values were not normally distributed, logarithmic transformations were performed. The computer program SAS PC (v 6.12; SAS Institute, Cary, NC) was used for the analyses. RESULTS Baseline characteristics of the subjects The two groups were not significantly different in age: 64 ± 5.3 compared with 63.4 ± 5.6 y; time since menopause: 172 ± 69 compared with 175 ± 92 mo; or calcidiol concentrations: 29.3 ± 9.3 compared with 28.3 ± 11 nmol/l (11.7 ± 3.7 compared with 11.3 ± 4.4 ng/ml), respectively, and the other biochemical measurements did not differ significantly before treatment (Table 1). According to dietary calcium intake The mean daily dietary calcium intake, assessed before the beginning of calcium supplementation, was 811 ± 330 mg/d (median: 786 mg/d, range: mg/d). We used this median value to classify the subjects by their dietary calcium intake as normal or low when it was, respectively, > or < 786 mg/d. In the total population (n = 116) we therefore derived two subpopulations, a normal one, ie, a reference subpopulation with a mean (± SD) dietary calcium intake of 1062 ± 278 mg/d and a low one with a mean (± SD) intake of 566 ± 131 mg/d. These two subpopulations were not significantly different in age (63.1 ± 5.6 compared with 64.0 ± 5.3 y), in time since menopause (172 ± 89 compared with 175 ± 72 mo), and in calcidiol concentrations (27.8 ± 8 compared with 30.0 ± 12 nmol/l or 11.1 ± 3.2 compared with 12.0 ± 4.8 ng/ml). In these two subpopulations, the biochemical measurements did not significantly differ except for pyridinoline, which was significantly higher in the subpopulation with a low calcium intake (56.2 ± 16.3 compared with 50.8 ± 13.4 nmol/mmol creatinine, P = 0.05). Moreover, a highly significant negative correlation between ipth values and dietary calcium intake was observed only in the subpopulation with a low calcium intake (r = 0.48, P = ) and not in the sub-

3 CALCIUM SUPPLEMENTATION IN POSTMENOPAUSAL WOMEN 1275 TABLE 1 Biochemical measurements in calcium-supplemented and placebo groups at baseline and at 1 and 2 mo 1 Calcium-supplemented group Placebo group Biochemical (n = 60) (n = 56) measurements Baseline 1 mo 2 mo Baseline 1 mo 2 mo Calcium (mmol/l) 2.36 ± ± ± ± ± ± h urinary calcium (mmol/d) 3.8 ± ± ± 2.6 2,3 3.8 ± ± ± h urinary ratio of 0.33 ± ± ± ± ± ± 0.22 calcium to creatinine (mmol/mmol) ipth (ng/l) 45.6 ± ± ± ± ± ± 24.2 [44 68]PTH (ng/l) ± ± ± ± ± ± 60.7 Osteocalcin ( g/l) 9.5 ± ± ± ± ± ± 3.6 Alkaline phosphatase Total (U/L) 71.5 ± ± ± ± ± ± 21.8 Bone (U/L) 14.0 ± ± ± ± ± ± 6.4 Hyp:creatinine ( mol/mmol) 17.9 ± ± ± ± ± ± 5.9 Pyd:creatinine (nmol/mmol) 53.2 ± ± ± ± ± ± 15.2 Dpd:creatinine (nmol/mmol) 9.5 ± ± ± ± ± ± x ± SD. ipth, intact parathyroid hormone; [44 68]PTH, PTH fragments 44 68; Hyp, hydroxyproline; Pyd, pyridinoline; Dpd, deoxypyridinoline. 2,4,5 Significant effect of time and treatment compared with baseline (Dunnett s multiple comparison test): 2 P 0.001, 4 P 0.05, 5 P Significant interaction between treatment and time, P 0.05 (Winer s analysis). population with a normal calcium intake (Figure 1). In the calcium-treated group (n = 60), we also derived two subgroups for calcium intake: a normal one with a mean (± SD) dietary calcium intake of ± mg/d (n = 29) and a low one with a mean calcium intake of ± 142 mg/d (n = 31). These two subgroups were not significantly different in age (63.5 ± 5.5 compared with 64.3 ± 5.1 y), time since menopause (167 ± 72 compared with 177 ± 66 mo), or calcidiol concentrations (27.5 ± 8 compared with 31.0 ± 10.3 nmol/l, or 11.0 ± 3.2 compared with 12.4 ± 4.1 ng/ml). There were slight differences between these two subgroups in their biochemical indexes, but they were not significant (Table 2). Effects of calcium supplementation on calcium homeostasis During the course of calcium supplementation in the calciumtreated group a significant increase in 24-h urinary calcium (39%) and in 2-h urinary calcium:creatinine (30%) excretion was observed whereas urinary calcium excretion did not change in the placebo group. At 2 mo, the 24-h urinary calcium differed significantly between the calcium and placebo groups. Simultaneously, [44 68]PTH and ipth were not significantly modified (Table 1). These effects observed in the two calcium-treated subgroups were FIGURE 1. Relation between intact parathyroid hormone (ipth) concentrations and dietary calcium intake at baseline in postmenopausal women (n = 116). The correlations were separately calculated in the subpopulations with low (R = 0.48, P = , n = 58) or normal (R = 0.02, NS, n = 58) dietary calcium intakes defined from the median value of the dietary calcium intake assessed before the beginning of the study.

4 1276 FARDELLONE ET AL TABLE 2 Biochemical measurements during the course of calcium supplementation according to the dietary calcium intake assessed before the beginning of the study 1 Normal dietary calcium intake Low dietary calcium intake Biochemical (n = 29) (n = 31) measurements Baseline 1 mo 2 mo Baseline 1 mo 2 mo Calcium (mmol/l) 2.37 ± ± ± ± ± ± h urinary calcium (mmol/d) 3.7 ± ± ± ± ± ± h urinary ratio of 0.37 ± ± ± ± ± ± calcium to creatinine (mmol/mmol) ipth (ng/l) 46.1 ± ± ± ± ± ± 14.1 [44 68]PTH (ng/l) ± ± ± ± ± ± Osteocalcin ( g/l) 9.7 ± ± ± ± ± ± 3.7 Alkaline phosphate Total (U/L) 71.4 ± ± ± ± ± ± 19.4 Bone (U/L) 13.9 ± ± ± ± ± ± Hyp:creatinine ( mol/mmol) 18.9 ± ± ± ± ± ± Pyd:creatinine (nmol/mmol) 49.9 ± ± ± ± ± ± Dpd:creatinine (nmol/mmol) 9.3 ± ± ± ± ± ± x ± SD. ipth, intact parathyroid hormone; [44 68]PTH, PTH fragments 44 68; Hyp, hydroxyproline; Pyd, pyridinoline; Dpd, deoxypyridinoline. 2 4 Significant differences among time periods (Dunnett s multiple comparison test): 2 P 0.001, 3 P 0.05, 4 P enhanced in the low dietary calcium intake subgroup, with a significant increase of 24-h urinary calcium (44%) and 2-h urinary calcium:creatinine (48%), and with a significant decrease of [44 68]PTH observed only at 2 mo. When we compared the two subgroups, however, there was no significant difference between them (Table 2). Effects of calcium supplementation on bone-remodeling markers When we compared the values of bone-remodeling markers between the calcium and placebo groups, no significant differences were observed (Table 1). In the calcium-treated group, as compared with baseline, bone-formation marker values were not significantly modified, except b-ap at 1 (P = 0.05) and 2 (P = ) mo. All the bone-resorption marker values decreased significantly, principally deoxypyridinoline at 2 mo ( 16%, P < 0.01). When considering the influence of dietary calcium intake assessed before the beginning of the study, the two subgroups did not differ significantly during the course of calcium supplementation. However, a significant effect was observed in the low calcium intake subgroup: the main indexes of bone remodeling (b-ap, hydroxyproline, pyridinoline, and deoxypyridinoline) decreased significantly at 1 and 2 mo after calcium supplementation. In contrast, no significant difference was observed in the subgroup with normal dietary calcium intake after calcium supplementation (Table 2). In the low calcium intake subgroup the most important variations were observed for deoxypyridinoline at 1 and 2 mo ( 18.5%, P < 0.05) whereas the other significant variations did not exceed 12%. Moreover, in this subgroup, when the results were expressed as z scores derived from the baseline values of the reference subgroup (Figure 2), we observed that among remodeling markers, deoxypyridinoline and hydroxyproline were particularly modified at 1 and 2 mo during the course of calcium supplementation. DISCUSSION Most epidemiologic studies have shown that the usual dietary calcium intake of postmenopausal women is lower than the optimal daily calcium requirements, estimated at 1500 mg/d (9). The benefits of calcium supplementation have been debated (11); however, several studies have proven its positive effects in elderly women in the prevention of bone loss and fractures (13, 14, 24, 25). In women just past menopause, the effect of calcium supplementation on bone loss is moderate and weaker than that obtained with hormone substitution therapy (26). In the elderly, vitamin D and calcium insufficiencies lead to secondary hyperparathyroidism, which contributes to age-related bone loss. During the course of calcium therapy in our study, we measured PTH concentrations and biochemical markers of bone turnover simultaneously in a homogeneous population of postmenopausal women aged on average 65 y. We excluded women with a calcidiol concentration < 15 nmol/l (6 ng/ml), to avoid the possibility of severe vitamin D insufficiency (17). The mean calcidiol concentration of our studied population was 28.8 nmol/l (11.5 ng/ml), a value very similar to the threshold proposed for vitamin D insufficiency when using an assay with a chromatographic purification step (27, 28). Therefore, part of our population is probably vitamin D insufficient, but not severely so. Because the effects of calcium supplementation on bone density may vary according to dietary calcium intake (14), the latter was assessed by using a food-frequency questionnaire at baseline before the beginning of calcium supplementation. We used the median value of dietary calcium intake (786 mg/d) to separate the subjects as having a normal or low dietary calcium intake. At baseline in the subjects with low dietary calcium intake, only urinary pyridinoline excretion was greater than for subjects with normal dietary calcium intake. Moreover, when calcium intake was low, an inverse relation between ipth values and dietary calcium intake was observed; this relation was no longer present when dietary calcium intake was normal. This suggests that a low dietary calcium intake is a determining factor in PTH concentrations in late postmenopausal women. Calcium administration for 2 mo had no significant effect on PTH concentrations measured by using two different assays. However, we do not exclude an acute and transient decrease in ipth concentrations after the oral calcium load, as shown after a single calcium administration in elderly men and women (29) and also in healthy adults (30, 31). These acute changes, which have been shown to be maximal in the first hours after a calcium load, may have been missed because blood sampling was performed in the morning long after the last calcium dose. In contrast with our

5 CALCIUM SUPPLEMENTATION IN POSTMENOPAUSAL WOMEN 1277 FIGURE 2. z Scores of bone-remodeling markers in the low dietary calcium intake subgroup (n = 31) at 1 and 2 mo of dietary calcium supplementation (z scores were calculated by reference to the normal subgroup and expressed as box plots, with 10th, 25th, 50th, 75th, and 90th percentiles) for urinary pyridinoline (Pyd), deoxypyridinoline (Dpd), hydroxyproline (Hyp), and serum bone alkaline phosphatase (b-ap). results, other authors have shown an effect of chronic calcium supplementation on PTH concentrations. A 20% reduction in ipth concentrations was shown by Kochersberger et al (5) in elderly subjects after 4 wk of treatment (1000 mg Ca given as calcium carbonate). McKane et al (32) showed that elderly women (mean age 69 y) treated with calcium citrate (1600 mg elemental Ca) for 3 y had significantly lower mean PTH concentrations ( 40%) and parathyroid gland responsiveness ( 47%) than untreated women with a mean dietary calcium intake of 815 mg. The 2-mo period of calcium supplementation affected mostly the bone-resorption markers and to a lesser extent the bone-formation markers. This discrepancy may be due to the relatively short duration of our study, because it is expected that a reduction in bone resorption will be followed by a parallel reduction in bone formation when a new steady state has been reached. The decrease in bone resorption was reflected by sensitive assays such as the urinary excretion of pyridinoline cross-links but also by hydroxyproline, usually considered to be a less sensitive index. Only b-ap, measured by a new immunologic assay, showed a significant variation. The changes in bone-remodeling markers were more marked in subjects with a low dietary calcium intake, but were moderate overall. When expressed as z scores, the most responsive assays were deoxypyridinoline and hydroxyproline. The greatest decrease was in deoxypyridinoline excretion, which was reduced by 20%. This effect is weaker than that reported previously by McKane et al (32), who observed a 35% difference in urinary deoxypyridinoline between the calcium- and placebo-treated groups, but this result could be explained by the use of a larger dose of calcium and a longer period of treatment. When using similar urinary markers, the effects of calcium also appear to be lower than those observed with other treatments like estrogens and bisphosphonates (26, 33). The decrease in bone-resorption markers after hormone replacement therapy was at least twice as high as that observed after calcium supplementation in early postmenopausal women (19% compared with 39%, respectively, for urinary collagen cross-links and 22% compared with 52% for urinary hydroxyproline) (27). In postmenopausal women with low bone mass, the decrease in urinary deoxypyridinoline excretion was 47% after 3 mo of treatment with alendronate (33). Our results suggest that calcium supplementation is more efficient in subjects with a low dietary calcium intake as shown previously by others for the prevention of age-related bone loss (14). Recker et al (25) showed recently that calcium supplementation is capable of reducing the incidence of new vertebral fractures in postmenopausal women with severe osteoporosis and a low calcium intake (mean value 450 mg/d) before therapy. In conclusion, our results show that calcium supplementation is able to decrease the concentrations of biochemical markers of bone turnover in patients with a low dietary calcium intake. This agrees with studies showing a preferential effect of calcium on the prevention of bone loss (14) and vertebral fractures (25) in postmenopausal women when dietary calcium intake is low. Grateful thanks to A Morel and C Diot for their technical assistance. REFERENCES 1. Nordin BEC, Need AG, Chatterton BE, Horowitz M, Morris HA. The relative contribution of age and years since menopause to postmenopausal bone loss. J Clin Endocrinol Metab 1990;70: Orwoll E, Meier D. Alterations in calcium, vitamin D and parathyroid hormone physiology in normal men with aging: relationship to the development of senile osteoporosis. J Clin Endocrinol Metab 1986;63: Forero MS, Klein RF, Nissenson RA, et al. Effects of age on circulating immunoreactive and bioactive parathyroid hormone levels in women. J Bone Miner Res 1987;2: Prince R, Dick I, Devine A, et al. The effects of menopause and age on calcitropic hormones: a cross-sectional study of 655 healthy women aged 35 to 90. J Bone Miner Res 1995;10: Kochersberger G, Bales C, Lobaugh B, Lyles KW. Calcium supplementation lowers serum parathyroid hormone levels in elderly subjects. J Gerontol 1990;45: Chapuy MC, Arlot ME, Dubœuf F, et al. Vitamin D3 and calcium

6 1278 FARDELLONE ET AL prevent hip fracture in elderly women. N Engl J Med 1992; 327: Brazier M, Kamel S, Maamer M, et al. The markers of bone remodeling in the elderly subject: effects of vitamin D insufficiency and its correction. J Bone Miner Res 1995;10: Heaney RP. Editorial: calcium, parathyroid function, bone, and aging. J Clin Endocrinol Metab 1996;81: NIH Consensus Conference. Optimal calcium intake. JAMA 1994;272: Cummings RG. Calcium intake and bone mass: a quantitative review of the evidence. Calcif Tissue Int 1990;47: Kanis JA, Passmore R. Calcium supplementation of the diet. Br Med J 1989;289: Horowitz M, Need AG, Philcox JC, Nordin BEC. Effect of calcium supplementation on urinary hydroxyproline in postmenopausal osteoporosis. Eur J Clin Nutr 1988;42: Reid IR, Ames RW, Evans MC, Gamble GD, Sharpe SD. Effect of calcium supplementation on bone loss in postmenopausal women. N Engl J Med 1993;328: Dawson-Hughes B, Dallal GE, Krall EA, Sadowski L, Sahyoun N, Tannenbaum S. A controlled trial of the effect of calcium supplementation on bone density in postmenopausal women. N Engl J Med 1990;323: Eyre DR, Paz MA, Gallop PM. Cross-linking in collagen and elastin. Annu Rev Biochem 1984;53: Delmas P. Biochemical markers of bone turnover. J Bone Miner Res 1993;8 (suppl):s Nordin BEC, Heyburn PJ, Peacock M, et al. Osteoporosis and osteomalacia. J Clin Endocrinol Metab 1980;9: Fardellone P, Sebert DL, Bouraya M, et al. Evaluation de la teneur en calcium du régime alimentaire par autoquestionnaire fréquentiel. (Evaluation of the calcium content of the diet by a food-frequency questionnaire.) Rev Rhum Mal Osteoartic 1991;58: (in French). 19. Zeghoud F, Jardel A, Guillozo H, N Guyen TM, Garabedian M. 25- Hydroxyvitamine D: mise au point d un microdosage par radiocompétition. Intérêt en pédiatrie. (25-Hydroxyvitamin D: development of a microassay by radiocompetition. Pediatric interest.) Immunoanal Biol Spec 1991;27:29 33 (in French). 20. Gueris J. Role physiologique et quantification de la parathormone humaine dans les liquides biologiques: difficultés méthodologiques et interprétation. (Physiological influence and quantification of human parathyroid hormone in the biological fluids: methodological difficulties and interpretation.) Immunoanal Biol Spec 1993; 8: (in French). 21. Castelain S, Kamel S, Picard C, Desmet G, Sebert JL, Brazier M. A simple and automated HPLC method for determination of total hydroxyproline in urine. Comparison with excretion of pyridinolines. Clin Chim Acta 1995;235: Kamel S, Brazier M, Desmet G, Picard C, Mennecier I, Sebert JL. High-performance liquid chromatographic determination of 3- hydroxypyridinium derivatives as new markers of bone resorption. J Chromatogr 1992;574: Winer BJ. Statistical principles in experimental design. New York: McGraw-Hill Book Company, Chevalley T, Rizzoli R, Nydegger V, et al. Effects of calcium supplements on femoral bone mineral density and vertebral fracture rate in vitamin-d-replete elderly patients. Osteoporos Int 1994: 4: Recker RR, Hinders S, Davies KM, et al. Correcting calcium nutritional deficiency prevents spine fractures in elderly women. J Bone Miner Res 1996;11: Aloia JF, Vaswani A, Yeh JK, Ross PL, Flaster E, Dilmanian A. Calcium supplementation with or without hormone replacement therapy to prevent postmenopausal bone loss. Ann Intern Med 1994;120: McKenna MJ. Differences in vitamin D status between countries in young adults and the elderly. Am J Med 1992;93: Ooms ME, Lips P, Roos JC, et al. Vitamin D status and sex hormone binding globulin: determinants of bone turnover and bone mineral density in elderly women. J Bone Miner Res 1995;10: Goddard M, Young G, Marcus R. Short-term effects of calcium carbonate, lactate, and gluconate on the calcium parathyroid axis in normal elderly men and women. Am J Clin Nutr 1986;44: Gonelli S, Cepollaro C, Camporeale A, Nardi P, Rossi S, Gennari C. Acute biochemical variations induced by two different calcium salts in healthy perimenopausal women. Calcif Tissue Int 1995; 57: Reginster JY, Denis D, Bartsch V, Deroisi R, Zegels B, Franchimont P. Acute biochemical variations induced by two different calcium salts in healthy male volunteers. Osteoporos Int 1993;3: McKane WR, Khosla S, Egan KS, Robins SP, Burritt MF, Riggs BL. Role of calcium intake in modulating age-related increases in parathyroid function and bone resorption. J Clin Endocrinol Metab 1996;81: Chesnut CH, McClung MR, Ensrud KE, et al. Alendronate treatment of the postmenopausal osteoporotic woman: effect of multiple dosages on bone mass and bone remodeling. Am J Med 1995;99: