Variability in urinary excretion of bone resorption markers: limitations of a single determination in clinical practice

Clinical Biochemistry 34 (2001) 571–577

Variability in urinary excretion of bone resorption markers: limitations of a single determination in clinical practice Didier Borderiea, Christian Rouxb, Benedicte Toussainta, Maxime Dougadosb, Ohvanesse G. Ekindjiana, Brigitte Cherruaua,* b

a Laboratoire de Biochimie A, Hoˆpital Cochin, 27 rue du Faubourg Saint Jacques 75014 Paris, France Service de Rhumatologie, Centre d’Evaluation des Maladies osseuses Hoˆpital Cochin, Universite´ Rene´ Descartes, 27 rue du Faubourg Saint Jacques, 75 679 Paris Cedex 14, France

Received 11 July 2001; received in revised form 9 October 2001; accepted 9 October 2001

Abstract In this study we assessed the within and between-subject variability of the concentrations of two urinary markers, free deoxypyridinoline (DPD) and C telopeptide (CTX-I), in healthy patients with the aim of setting reliable thresholds to enable physicians to take decisions about individual patients with confidence. Between-subject variability for the women was 25.4% for DPD and 38.2% for CTX-I, and for the men was 12.9% for DPD and 23.8% for CTX-I. The coefficients of variation were similar for daily, weekly and monthly determinations, giving means of 13.8 and 28.1% for DPD and CTX-I respectively. Critical difference (CD) was lower for DPD than for CTX-I (about 44 and 80% respectively). The number of samples required to determine the true mean with a CD at the 5% level was 29 for DPD and more than 113 for CTX-I. DPD was the least biologically variable. One determination was not sufficient to determine bone resorption status and a 44% decrease in DPD levels and an 80% decrease in CTX-I levels were required to demonstrate the efficacy of antiresorptive therapy in individual patients. © 2001 The Canadian Society of Clinical Chemists. All rights reserved. Keywords: Deoxypyridinoline; Collagen telopeptides; Bone resorption; Biologic variability

1. Introduction Bone remodeling is a dynamic phenomenon that occurs via the coupled processes of resorption and formation. These two different cell activities are assessed by different markers. These markers include deoxypyridinoline (DPD), a cross-linker of collagen synthesized posttranscriptionally and CTX-I (Crosslaps™), a type I collagen peptide corresponding to the C telopeptide of the ␣1 chain. These molecules are released into the blood stream during resorption and are subsequently excreted in urine [1]. Biochemical bone markers have been used to identify patients with a variety of metabolic bone diseases [2,3], to predict bone changes [4] and future fracture risk [5] and to assess the efficiency of antiresorptive treatments such as hormone replacement therapy, raloxifene, and bisphospho* Corresponding author. Tel.: ⫹33-1-58-41-15-91; fax: ⫹33-1-58-4115-85. E-mail address: [email protected] (B. Cherruau).

nates [6 –9]. The effects of treatment can also be confirmed by measuring bone mineral density, but extended follow-up periods are generally necessary to observe a significant effect of the treatment and the precision error for BMD is estimated to be 5% [10]. Biochemical markers have the advantage of being a noninvasive method for detecting and following changes in bone turnover [11]. However, in all of these studies, significant changes were observed for groups of patients and results were expressed as mean values. Patients were included according to selective criteria and followed according to rigorous protocols. In practice, the clinical value of biochemical markers is less clear in individual patients than in these studies. In studies, knowledge of the variability of biochemical markers is required to interpret the value of a single measurement in an individual patient. The variability observed may be of several different types, including analytical variability and, particularly, biologic variability, reflecting exogenous and endogenous perturbations of biologic systems [12]. Biologic variability seems to be the major contributor to total intra-individual

0009-9120/01/$ – see front matter © 2001 The Canadian Society of Clinical Chemists. All rights reserved. PII: S 0 0 0 9 - 9 1 2 0 ( 0 1 ) 0 0 2 6 9 - 7

572

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577

variability in many bone markers [13]. The analytical variability of biochemical markers of bone resorption is well documented [14] but only a few studies have evaluated the biologic variability of bone markers in healthy patients; none have analyzed this variability over a long period. The aim of this study was, to assess and to compare daily, weekly and monthly variabilities in healthy patients in the absence of circadian and seasonal variations, and to evaluate the limitations and usefulness of single measurements in individuals.

with an amino acid sequence specific for part of the Ctelopeptide of the ␣1 chain of type I collagen, according to the recommendations of the manufacturers. The concentrations of these markers are expressed relative to those of urinary creatinine concentration determined on a Hitachi 917 analyzer using the kinetic Jaffe method (Roche Diagnostics). All samples from an individual volunteer were analyzed in the same assay in duplicate. The analytical variability of the bone markers was determined from at least 10 repeat measurements in the same run and 10 repeat measurements in different runs.

2. Methods

2.4. Analytical evaluation

2.1. Subjects

Analytical performances were evaluated by determining intraassay (CVA) and interassay (CVB) coefficients of variation. Intraassay precision was assessed by analysis of 2 urine samples, with 10 replicates for each. Both DPD and CTX-I had intraassay CVs ⬍ 10% (DPD nmol/L: U1 30 ⫾ 2.3, CV 7.9%; U2 90 ⫾ 8.2, CV 7.4%; CTX-I ␮g/L: U1 255 ⫾ 21.7, CV 8.5%, U2 1455 ⫾ 119.3, CV 8.2%). Interassay precision was assessed by analysis of 2 urine samples and kit controls 5 times over a 2- week period. Interassay: DPD nmol/L: U1 16.5 ⫾ 1.4, CV 8.5%; U2 90.6 ⫾ 7.2, CV 7.9%; CTX-I ␮g/L: U1 263 ⫾ 27.9, CV 10.6%, U2 1475 ⫾ 141.6, CV 9.6%. CVs were higher for CTX-I than for DPD, especially for the highest concentrations.

Forty-two healthy premenopausal women aged from 21 to 49 yr (mean, 28.7 ⫾ 6.1 yr) and 21 men aged from 23 to 37 yr (mean, 25.7 ⫾ 7.3) were recruited in winter. None of the volunteers had medical disorders known to affect bone metabolism such as renal diseases, hyperparathyroidism, hyperthyroidism, recent fracture, prolonged immobilization or joint diseases. None of the subjects used drugs that could affect bone metabolism such as corticosteroids, nonsteroidal antiinflammatory drugs (NSAID) or anticonvulsants. Twenty-four of the women used estroprogestative contraception; the others did not use hormonal contraception. The interindividual variability of the bone resorption markers was determined for these 42 women and 21 men. A subset of subjects provided additional specimens to determine daily, weekly and monthly intraindividual variability. This subgroup consisted of 20 healthy premenopausal women aged from 24 to 36 yr (mean, 28.6 ⫾ 3.6 yr) and 6 men aged from 26 to 31 yr (mean, 28.7 ⫾ 3.3). 2.2. Urine samples Second-void fasting morning urine specimens were collected between 0800 and 1000 h. For each subject, five samples were collected during the first week, three more samples were collected during the next three weeks and one sample per month was collected in each of the next two months, giving a total of 10 samples per subject. All urine specimens were collected during the same three months, to eliminate the possibility of seasonal variation. Urine samples were frozen within one hour of collection and stored at ⫺80°C until analysis. 2.3. Biochemical measurements Urinary free deoxypyridinoline (DPD) was determined with an ELISA method using a monoclonal antibody that does not interact with cross-linked peptides (Pyrilinks-D Metra Biosystems). Urinary type I collagen C-telopeptide breakdown products (CTX-I) were determined by ELISA (Crosslaps, Cis Bio) using an immobilized synthetic peptide

2.5. Statistical analysis Data are presented as means ⫾ SD. Individual measures of within-subject variability are reported as coefficients of variation (CVs; SD/mean x100) and were calculated by nested ANOVA from replicate analyses [15]. The weekly intrasubject CV’s were calculated from the values for the samples taken on the Mondays of the four consecutive weeks for each subject. Monthly intrasubject CV’s were calculated from the values for the samples taken on the Monday of the first week in each of the three consecutive months. The index of individuality (I) was calculated for each analysis as the ratio CV1/CV2, where CV1 is the biologic within-subject coefficient of variation and CV2 is the between-subject coefficient of variation. This index indicates the degree to which a single measurement is able to differentiate between normal and abnormal values in a subject in relation to a reference interval [15]. The critical difference (CD) is the minimum difference between two serial results required for statistical significance. This critical difference at the 5% level was calculated as [16]: CD ⫽ 1.96 ⫻ 公2 ⫻ CVT, where CVT is the total ‘‘apparent’’ within-subject coefficient of variation. It can be partitioned into the coefficient of within-subject biologic variation (CV1) and the coefficient of within-subject analytical variation (CVA): CVT ⫽ (CV12 ⫹ CVA2)1/2.

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577

Fig. 1. Distribution of DPD by age for the group of women. (A) Individual values for the very first sample taken from each female subject are plotted as a function of age. (B) Each point represents the mean of all ten results of for a given subject included in the intrasubject variability study and is plotted as a function of age. The solid line is the best fit to the data using a locally weighted regression plot (LOWESS).

The number of samples (Ns) required for determination of the true mean for an individual subject to within 5%, with 95% confidence, was estimated using the formula of Fraser and Harris [15]. Spearman’s rank correlation coefficients were calculated to determine the degree of association between two continuous parameters. Mann-Whitney U-tests were used to compare values of the biochemical markers for categorical data. The relationship between age and urinary markers was analyzed with a nonparametric locally weighted regression plot (LOWESS).

3. Results 3.1. Demographic characteristics and bone resorption markers Figs. 1 and 2 show the distribution of DPD and CTX-I values as a function of age. The fitted line suggests a relationship between age and DPD/creatinine before the age of 27 yr, with no association thereafter. We observed no difference between the scatterplot of resorption markers

573

Fig. 2. Distribution of CTX-I by age for the group of women. (A) Individual values for the very first sample taken from each female subject are plotted as a function of age. (B) Each point corresponds to the mean of all ten results for a given subject included in the intrasubject variability study and is plotted as a function of age. The solid line is the best fit to the data using a locally weighted regression plot (LOWESS).

based on the very first sample taken from each subject and the scatterplot obtained if each point plotted corresponded to the mean of all ten results for a given subject. Regression analysis confirmed that there was a negative association between DPD and age for the 20 to 27 yr age group and no association after the age of 27 yr. Creatinine-corrected DPD values were significantly higher for the group of women under the age of 27 yr than for women over 27 yr of age (p ⬍ 0.03). The mean ⫾ SD urinary DPD/creatinine value for women aged 21 to 27 yr (n ⫽ 14) was 7.46 ⫾ 2.45 nmol/mmol Cr (range: 3.4 –12.16). In the group of women aged over 27 yr (n ⫽ 28), the mean value of DPD/creatinine was 5.43 ⫾ 1.56 nmol/mmol Cr (range: 3.2–10.2). CTX-I, unlike DPD, was not associated with age. CTX-I values were not significantly different in the two subgroups of women: in women aged 21 to 27 yr, the mean CTX-I concentration was 198.5 ⫾ 79.5 ␮g/mmol, whereas that in women over the age of 27 yr was 163.1 ⫾ 61.9 ␮g/mmol. Mean CTX-I concentration for the subgroup of women using estroprogestative contraception was significantly lower (171.1 ⫾ 12.7 ␮g/mmol Cr) than that for the subgroup not using estroprogestative contraception (213.1 ⫾ 8 ␮g/mmol Cr) (p ⬍ 0.02). DPD concentration did not differ significantly between these two subgroups.

574

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577

Table 1. Mean, standard deviation and between-subject CVs of bone resorption markers in healthy men and premenopausal women Women (n ⫽ 42)

Mean Standard deviation Median Range CV (%)

Men (n ⫽ 21)

DPD (nmol/mmol Cr)

CTX-I (␮g/mmol Cr)

DPD (nmol/mmol Cr)

CTX-I (␮g/mmol Cr)

6.7 1.7 5.5 3.3–10.2 25.4

192.7 73.6 194.5 59.3–405.0 38.2

4.5* 0.6 4.6 3.6–5.0 12.9

174.8 41.6 161.2 143.2–233.6 23.8

* p ⬍ 0.05 versus the group of women.

For the men, no relationship was observed between age and the bone markers. 3.2. Population variation Mean DPD levels were lower for the men than for the women (p ⬍ 0.05), but there was no significant difference for CTX-I (Table 1). Between-subject CVs were higher for the women than for the men. Between-subject CVs were also higher for CTX-I than for DPD. Tables 2 and 3 show the within-subject CVs for DPD and CTX-I. There was no statistically significant difference in daily, monthly or weekly within-subject variability, whatever the marker. However, a correlation was found between daily and weekly CVs (r ⫽ 0.41, p ⬍ 0.01) and between daily and monthly CVs (r ⫽ 0.32, p ⬍ 0.01) variability. To illustrate the variability in DPD and CTX-I levels, we plotted daily, weekly and monthly, individual values for all participating subjects (Fig. 3 and 4). Mean CV values were similar in men and women. We therefore calculated mean CV values for all subjects: 13.8 ⫾ 1.6% and 28.1 ⫾ 1% for DPD and CTX-I respectively. Index of individuality values were lowest for women (Table 4). The critical difference, which reflects both biologic and analytical variability, was lower for DPD than for CTX-I (close to 44% and 80% respectively). The number of samples required to determine the true mean, with a critical difference at the 5% level, was 29 for DPD and more than 113 for CTX-I.

4. Discussion The value of biochemical markers of bone resorption has been evaluated essentially for groups of patients, with values pooled and represented as a mean value for the population studied. We therefore evaluated the analytical and biologic variability of these markers in healthy individual patients in the absence of circadian and seasonal variation, to facilitate the correct interpretation of these assays in clinical practice. We found that mean urinary DPD excretion was highest in women aged 20 to 27 years. No such difference was observed for CTX-I. The high level of DPD excretion between the ages of 20 and 27 suggests a slow but persistent increase in bone mass during this period. This indirectly reflects the bone remodeling balance with a probable increase in bone formation in women aged 20 to 27 yr. This finding is important because this skeletal phase, corresponding to peak bone mass, serves as a reference for evaluating normal bone remodeling, and must be taken into account in defining the normal range of values. Orwoll et al. have reported similar result, using NTX-I as the index of bone resorption in a group of healthy men [17]. We observed a narrower range of DPD and CTX-I values for men than for women. Free cross-links are considerably less variable than peptide-bound crosslinks. Other studies have also reported that DPD exhibits lower day-to-day variability than CTX-I [14]. An explanation was put forward by Colwell and Eastell, who suggested that peptide-bound

Table 2. Distribution of within-subject CVs (%) of urinary DPD/creatinine: daily, weekly and monthly variability Men (n ⫽ 6)

Mean SD Median Range

Women (n ⫽ 20)

All (n ⫽ 26)

Daily

Weekly

Monthly

Daily

Weekly

Monthly

Daily

Weekly

Monthly

13.6 1.9 16.7 9.2–11.7

15.3 10.5 15.7 14.1–34.6

12.4 8.5 16.0 6.5–27.2

12.9 6.7 13.6 4.2–26.9

15.8 8.8 15.1 5.4–34.1

12.8 8.7 11.2 2.6–37.8

12.4 6.0 10.7 4.2–26.9

15.1 6.3 15.1 5.4–34.6

13.5 8.5 12.4 2.6–27.2

weekly within-subject CV’s: Monday values for the four consecutive weeks for each subject monthly within-subject CV’s: Monday values for the first week of each of the three consecutive months

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577

575

Table 3. Distribution of within-subject CVs (%) of urinary CTX-I/creatinine : daily, weekly and monthly variability Men (n ⫽ 6)

Mean SD Median Range

Women (n ⫽ 20)

All (n ⫽ 26)

Daily

Weekly

Monthly

Daily

Weekly

Monthly

Daily

Weekly

Monthly

28.7 13.9 29.2 11.6–43.2

29.7 27.8 30.6 8.5–63.2

29.0 18.0 30.9 13.2–53.0

26.8 14.7 27.7 8.4–57.1

28.7 23.8 30.9 8.4–6.2

27.9 15.0 29.4 2.2–67.2

29.4 14.3 29.1 4.8–57.1

26.9 23.9 27.4 8.4–68.2

25.0 16.0 31.2 2.2–67.2

weekly within-subject CV’s: Monday values for the four consecutive weeks for each subject monthly within-subject CV’s: Monday values for the first week of each of the three consecutive months

crosslinks may be metabolized by a saturable enzyme in the kidney [18]. The mean CVs for daily, weekly and monthly values were similar for DPD and CTX-I. There was also no significant difference in mean CV between men and women. We observed no significant change in bone resorption markers over the menstrual cycle. Our findings for DPD, are consistent with those of Schlemmer et al. [19] and Chiu et al. [20], but differ from those of Gorai et al., [21,22]. Gorai et al. reported differences in DPD and CTX-I concentrations between subphases of the cycle, but the patterns of change differed in the two published studies. Various sources of biologic variation have been shown to affect bone turnover, including marked circadian rhythm [23] and seasonal variations [24]. In our study, these factors

Fig. 3. Variations in excretion of creatinine-corrected DPD in the subgroup of women participating in the within-subject variation study. (A) Day-today variation, (B) Weekly variation, (C) Monthly variation. Individual subject values are plotted as a function of the time of specimen collection. Solid lines connect the individual values for each subject.

should not affect the results because all specimens were collected between 8 and 10 AM over consecutive days, weeks or months in winter. Woitge et al. [24] reported higher concentrations of DPD in women in winter than in summer, with no seasonal variation in men. Variability, as expressed by CV %, was similar across the range of values in our study. However differences may occur in postmenopausal women, a population in which higher values may be observed. The index of individuality was used to assess the performance of the tests for distinguishing normal from unusual results in a subject. This ratio was lower for women (0.48 for DPD and 0.70 for CTX-I) than for men group (1.1 for DPD and 1.30 for CTX-I), indicating that the day-to-day variation in individuals was markedly greater than the underlying variation between individuals for men. It would therefore not be wise to attempt to distinguish between

Fig. 4. Variations in excretion of creatinine-corrected CTX-I excretion values in the subgroup of women that participated in the within-subject variation study. (A) Day-to-day variation, (B) Weekly variation, (C) Monthly variation. Individual subject values are plotted as a function of the time of specimen collection. Solid lines connect the individual values for each subject.

576

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577

Table 4. Individuality indices, critical difference values, reference change values and numbers of samples required to estimate the true mean DPD

Iindex CD (%) Number of samples

CTX-I

Men

Women

Men

Women

1.15 43.8 29

0.48 44.1 29

1.30 84.0 130

0.70 78.8 113

Iindex: Index of individuality, CD: critical difference at the 5% level.

individuals in the male population on the basis of a singleday measurement; it would instead be preferable to use the mean of repeat measurements on different days. Beck Jensen et al. reported a similar index for DPD (0.35) for a group of women [25]. The critical difference is defined as the smallest difference in sequential measurements that gives a significant result for an individual at the level chosen. Its reflects the clinical utility of the markers in individual patients. For the women in this study, a 44% change in DPD or a 79% change in CTX-I was required to obtain a difference that was significant and unlikely to be due to variability alone. These values are higher than the CD value obtained for NTx by Orwoll et al. [17] in a group of healthy men (35%), but are lower than the CD value obtained by Hannon et al. [26] for DPD and CTX-I (55% and 132% respectively) for a group of healthy postmenopausal women. However, these values suggest that DPD would be useful for following the efficacy of antiresoptive therapy as Fradinger et al. [27] found that this marker decreased by 46 ⫾ 14% after six months of alendronate treatment. The number of repeat samples, taken on different days, required to determine the true mean to within ⫾ 5% with 95% confidence intervals was much higher for CTX-I than for DPD. A single measurement is clearly insufficient for estimation of the true value of these markers with 95% confidence but is sufficient for the estimation of DPD with 75% confidence. Several strategies can be developed to minimize the variability. For example, the mean level of duplicate measurements can be used, as proposed by Hannon et al. [26]. Measuring the serum concentrations of these markers could also circumvent some of the limitations of using urinary indices. A single determination is sufficient to demonstrate the efficacy of antiresorptive therapy if the decrease obtained after three months is over 44% for DPD and 79% for CTX-I. In conclusion, this study shows that intra- and intersubject variabilities must be considered when interpreting quantitative urinary bone marker data for the assessment of bone loss in individual patients, and particularly for monitoring antiresorptive therapy. The two assays have similar analytical variabilities, but differ in biologic variability. In our study, DPD had the lowest biologic variability, both between and within individuals.

References [1] Delmas PD, Schlemmer A, Johansen JS, Gineyts E, Riis B, Christiansen C. Urinary excretion of pyridinoline cross-links correlates with bone turnover measured on iliac crest biopsy in patients with vertebral osteoporosis. J Bone Miner Res 1991;6:639 – 44. [2] Garnero P, Gineyts E, Riou JP, Delmas PD. Assessment of bone resorption with a new marker of collagen degradation in patients with metabolic bone disease. J Clin Endocrinol Metab 1994;79:780 –5. [3] Taylor AK, Kueken SA, Libanati C, Baylink DJ. Biochemical markers of bone turnover for the clinical assessment of bone metabolism. Rheum Dis Clin North Am 1994;20:589 – 607. [4] Borderie D, Cherruau B, Dougados M, Ekindjian OG, Roux C. Biochemical markers as predictors of bone mineral density changes after GnRH agonist treatment. Calcif Tissue Int 1998;62:21–5. [5] Garnero P, Hausherr E, Chapuy M-C, et al. Markers of bone resorption predict hip fracture in elderly women: the EPIDOS prospective study. J Bone Miner Res 1996;11:1531– 8. [6] Garnero P, Weichnung JS, Gineyts E, Karpf D, Delmas PD. Comparison of new biochemical markers of bone turnover in late postmenopausal osteoporotic women in response to alendronate treatment. J Clin Endocrinol Metab 1994;79:1693–700. [7] Riis BJ, Overgaard K, Christiansen C. Biochemical markers of bone turnover to monitor the bone response to postmenopausal hormone replacement therapy. Osteoporosis Int 1995;5:276 – 80. [8] Bonde M, Qvist P, Fledeelius C, Riis BJ, Christiansen C. Applications of an enzyme immunoassay for a new marker of bone resorption (CrossLaps): follow-up of hormone replacement therapy and osteoporosis risk assessment. J Clin Endocrinol Metab 1995;80:864 – 8. [9] Ravn P, Clemmensen B, Riis BJ, Overgaard K, Christiansen C. The effect of bone mass and bone markers of different doses of ibandronate: a new bisphosphonate for prevention and treatment of postmenopausal osteoporosis: a 1 year, randomized, double-blind, placebo controlled, dose-finding study. Bone 1996;19:527–33. [10] Genant HK, Jergas M, Palermo L, et al. Comparison of semi-quantitative visual and quantitative morphometric assessment of prevalent and incident vertebral fractures in osteoporosis. The Study of Osteoporotic Fractures Research Group. J Bone Miner Res 1996;11:984–96. [11] Watts NB. Clinical utility of biochemical markers of bone remodeling. Clin Chem 1999;8(Suppl 2):1359 – 68. [12] Widjaja A, Morris RJ, Levy JC, Frayn KN, Manley SE, Turner RC. Within-, and between- subject variation in commonly measured anthropometric, and biochemical variables. Clin Chem 1999;45:561– 6. [13] Ju HSJ, Leung S, Brown B, et al. Comparison of analytical performance and biological variability of three bone resorption assays. Clin Chem 1997;43:1570 – 6. [14] Popp-Snijder C, Lips P, Netelenbos JC. Intra-individual variation in bone resorption markers in urine. Ann Clin Biochem 1996;33:347– 8. [15] Fraser CG, Harris EK. Generation and application of data on biological variation in clinical chemistry. Crit Rev Clin Lab Sci 1989;27:409–37. [16] Magid E, Petersen PH, Christensen M. A note on the theory of references changes. Scand J Clin Lab Invest 1992;52(Suppl 208):95–101. [17] Orwoll ES, Bell NH, Nanes MS, et al. Collagen N-telopeptide excretion in men: the effects of age and intrasubject variability. J Clin Endocrinol Metab 1998;83:3930 –5. [18] Colwell A, Eastell R. The renal clearance of free and conjugated pyridinium crosslinks of collagen. J Bone Miner Res 1996;11:1976– 80. [19] Schlemmer A, Hassager C, Risteli L, Jensen SB, Christiansen C. Possible variation in bone resorption during the normal menstrual cycle. Acta Endocrinol 1993;129:388 –92. [20] Chiu KM, Mayes D, Bacchetti P, Weitz S, Arnaud CD. Changes in bone resorption during the menstrual cycle. J Bone Miner Res 1999; 14:609 –15. [21] Gorai I, Chaki O, Nakayama M, Minaguchi H. Urinary biochemical markers for bone resorption during the menstrual cycle. Calcif Tissue Int 1995;57:100 – 4.

D. Borderie et al. / Clinical Biochemistry 34 (2001) 571–577 [22] Gorai I, Taguchi Y, Chaki O, et al. Serum soluble interleukin-6 receptor and biochemical markers of bone metabolism show significant variations during the menstrual cycle. J Clin Endocrinol Metab 1998;83:326 –32. [23] Aoshima H, Kushida K, Takahashi M, et al. Circadian variation of urinary type I collagen crosslinked C-telopeptide and free and peptide-bound forms of pyridinium crosslinks. Bone 1998;22:73– 8. [24] Woitge HW, Scheidt-Nave C, Kissling C, et al. Seasonal variation of biochemical indexes of bone turnover: results of a population-based study. J Clin Endocrinol Metab 1998;83:68 –75.

577

[25] Beck Jensen JE, Kollerup G, Sorensen HA, Pors Nielsen S, Sorensen OH. A single measurement of biochemical markers of bone turnover has limited utility in the individual person. Scand J Clin Lab Invest 1997;57:351– 60. [26] Hannon R, Blumsohn A, Naylor K, Eastell R. Response of biochemical markers of bone turnover to hormone replacement therapy: impact of biological variability. J Bone Miner Res 1998;13:1224–33. [27] Fradinger EE, Rodriguez G, Bogado C, Zanchetta JR. Effect of antiresorptive therapy on day to day variation of urinary free deoxypyridinoline excretion. Clin Chem 1998;44:2224 –5.