Osteoporosis International

, Volume 17, Issue 7, pp 1086–1095

Biochemical markers of bone turnover predict bone loss in perimenopausal women but not in postmenopausal women-the Japanese Population-based Osteoporosis (JPOS) Cohort Study

Authors

    • Department of Public HealthKinki University School of Medicine
  • A. Morita
    • Department of Public HealthKinki University School of Medicine
  • Y. Ikeda
    • Department of Public HealthKinki University School of Medicine
  • Y. Sato
    • Department of Nutrition, School of Nutrition and NursingTenshi College
  • T. Akiba
    • Department of Blood Purification and Internal Medicine, Kidney CenterTokyo Women’s Medical University
  • T. Matsumoto
    • Department of Medicine and Bioregulatory SciencesUniversity of Tokushima Graduate School of Health Biosciences
  • H. Nishino
    • Toyama Red Cross Blood Center
  • S. Kagamimori
    • Department of Welfare Promotion and EpidemiologyFaculty of Medicine, University of Toyama
  • Y. Kagawa
    • Kagawa Nutrition University
  • H. Yoneshima ·
    • Shuuwa General Hospital
  • For the JPOS Study Group
Original Article

DOI: 10.1007/s00198-005-0052-3

Cite this article as:
Iki, M., Morita, A., Ikeda, Y. et al. Osteoporos Int (2006) 17: 1086. doi:10.1007/s00198-005-0052-3

Abstract

Introduction

The predictive value of biochemical markers of bone turnover for subsequent change in bone density in a population sample of healthy women with a wide range of ages has not been fully established.

Methods

We followed 1,283 women aged 15–79 years at baseline selected randomly from the inhabitants of three areas in Japan for 6 years, and examined 1,130 subjects with no disease or administration of drugs affecting bone metabolism. The annual change in bone density at the spine, total hip, and distal one third of the radius was determined during the follow-up period by dual x-ray absorptiometry and was compared among the groups using different levels of biochemical markers at baseline, including serum osteocalcin (OC) and bone-specific alkaline phosphatase (bone ALP), free and total (tDPD) forms of immunoreactive deoxypyridinoline, and type I collagen crosslinked C-terminal telopeptide (CTX) in urine.

Results

Premenopausal women aged 45 years or older with elevated levels of OC, bone ALP, CTX, or tDPD showed significantly greater bone loss at most skeletal sites during the follow-up period than those with lower levels, after adjustment for the effects of age, height, weight, dietary calcium intake, regular exercise, and current smoking. The greatest coefficient of determination of the model was observed in the association between CTX and bone loss at the hip during the first 3 years of follow-up (42.8%). These subjects were pooled with perimenopausal women at baseline, and those who still menstruated at follow-up in this pooled group showed significant but more modest associations, whereas those who entered menopause during the follow-up period showed clear associations. However, early postmenopausal women with less than 5 or 10 years since menopause showed an association that was limited mostly to the distal radius, and other postmenopausal groups had virtually no association.

Conclusion

Biochemical markers of bone turnover may predict bone loss in women undergoing menopausal transition but may not predict bone loss in postmenopausal women.

Keywords

Bone mineral densityBone turnover markersJapanese womenLongitudinal studyPopulation-based epidemiologic studyRandom sample

Introduction

Preventing osteoporosis and osteoporotic fracture is of great importance in women’s health in an aging society. The strategy of this prevention can be summarized as two components: increasing peak bone mass and decreasing postmenopausal bone loss. The latter can be accomplished by giving potent preventive measures to individuals at risk for unacceptable bone loss around or after menopause. Biochemical markers of bone turnover are expected to predict which persons should be administered preventive measures.

Many studies have examined the association between levels of biochemical markers of bone turnover and subsequent change in bone mineral density (BMD) in postmenopausal women. Some earlier studies reported a strong relationship of the markers to bone loss at the forearm in early postmenopausal women, and they suggested great value of the markers for predicting the incidence of so-called fast bone losers [1, 2]. Later studies on this topic, however, reported a significant but more modest relationship at the forearm [37] and the spine [813], and some even failed to obtain a significant association at the hip [14, 15]. It has been concluded, according to the current knowledge available, that the levels of several biochemical markers of bone turnover are significantly associated with the change in BMD at the forearm in early postmenopausal women but are still not accurate enough to be used for predicting bone loss on an individual basis, and that the predictive value of the markers at other skeletal sites, such as the spine and hip, is more limited [16, 17].

Before drawing a pessimistic conclusion for this role of bone turnover markers, several points should be addressed. Most of the previous studies on this topic examined perimenopausal or postmenopausal women. Knowledge about younger or older people is still limited. In addition, those studies recruited subjects, mostly on a volunteer basis for screening purposes, from university workers, institutionalized people, and community inhabitants, so the results may not represent characteristics of the general population. This is particularly important regarding the prediction of BMD changes in perimenopausal or younger women because most of them dwell in the community without enjoying the benefits of screening. Furthermore, most of the investigations were performed on Caucasian women; knowledge regarding Asians, including Japanese, has been relatively limited.

To solve these problems, we launched the Japanese Population-based Osteoporosis (JPOS) Study [18], whereby Japanese women over a wide age range (15–79 years) were randomly selected on a large scale from the general population and were followed for 6 years to examine the association of levels of biochemical markers of bone turnover at baseline with subsequent changes in BMD at several skeletal sites.

Materials and methods

Subjects

This study was performed as part of a larger epidemiologic study, the JPOS Study, in which 50 women were randomly selected from each 5-year age group covering the ages 1579 years according to resident registrations in seven municipalities throughout Japan. Details of the study are found elsewhere [18]. The three municipalities were selected from the seven study areas for the cohort study: Hirara City in Okinawa, the southern island in the subtropical zone; Nishi-Aizu Town, located in a mountainous area of Honshuu Main Island; and Sangawa Town (which changed its name to Sanuki City in 2001) in Shikoku facing the Seto Inland Sea. Among 1,950 women selected, 1,651 (84.7%) completed the baseline study comprising interviews and measurements of BMD, body size, and biochemical markers of bone turnover. We excluded 147 subjects with a history or present case (as determined by the interviews or laboratory tests) of any disease affecting bone metabolism, such as amenorrhea, oligomenorrhea, bilateral oophorectomy, diseases of the parathyroid gland, rheumatoid arthritis, diabetes mellitus, cancers, and stroke; those who had ever taken drugs that affect bone turnover, such as calcium, estrogens, vitamin D, calcitonin, bisphosphonates, and glucocorticoids; and those who did not submit blood or urine samples. The remaining 1,504 (77.1% of the total samples) women were invited for two follow-up studies conducted 3 and 6 years after the baseline, which consisted of the same BMD and body size measurements as the baseline survey and similar interviews. Of these, 1,283 women (85.3% of the subjects eligible for follow-up) completed at least one of the follow-up surveys, and 125 women were further excluded from the study because 53 were found to have incident diseases affecting bone metabolism and 76 had begun taking drugs described above during the follow-up period (four subjects met both criteria).

The subjects who regularly menstruated at baseline according to the menstrual information obtained from the interviews were judged to be premenopausal, and those who had entered menopause at least 6 months prior to the survey were judged to be postmenopausal. The remaining women were classified as perimenopausal. Postmenopausal subjects whose age at menopause and years since menopause (YSM) could not be determined because of having had premenopausal hysterectomy or having lost exact memory were excluded from the analyses. Consequently, 1,130 women were analyzed as a representative sample of the Japanese female population free from apparent diseases affecting bone metabolism.

Written informed consent regarding all the study procedures was obtained from each subject in advance. The study protocol was approved by the ethics committee of the Kinki University School of Medicine.

Biochemical markers of bone turnover

Details of the methods for measuring biochemical markers of bone turnover have been described previously [19]. Briefly, we measured at baseline the levels of serum osteocalcin (OC), bone-specific alkaline phosphatase (bone ALP), type I collagen crosslinked C-terminal telopeptide (CTX), and free (fDPD) and total (tDPD) forms of immunoreactive deoxypyridinoline in urine using commercially available kits according to the manufacturers’ protocols.

Serum OC (nmol/l) was measured by a two-site immunoradiometric assay (IRMA; BGP IRMA kit, Mitsubishi Kagaku Iatron, Tokyo, Japan) with a sensitivity of 0.2 nmol/l [20]. Serum bone ALP (ng/ml) was measured by IRMA (Tandem-R Ostase, Hybritec, San Diego, CA, USA) with a sensitivity of 0.1 ng/ml [21]. CTX (μg/l) was measured by an enzyme-linked immunoassay (ELISA) (CrossLaps ELISA, Osteometer A/S, Rødovre, Denmark) with a detection limit of 0.5 μg/l [22]. Urine fDPD (nmol/l) was determined using ELISA (Pyrilinks-D, Metra Biosystems, Mountain View, CA, USA) with a minimal detection limit of 1.1 nmol/l [23]. Urine tDPD (nmol/l) was measured by the same method as for fDPD after hydrolysis of the urine specimen, with a minimal detection limit of 11 nmol/l. All results for urinary markers were corrected for creatinine (Cr) concentration (mmol Cr/l).

Intraassay precision represented by the coefficient of variation (CV) was 8.0%, 6.9%, 5.0%, 5.6%, and 6.8%, and interassay precision over the 9-month baseline study period was 14.2%, 11.9%, 14.5%, 12.6%, and 16.8% for OC, bone ALP, CTX, fDPD, and tDPD, respectively. Time trends of measured markers for the control specimens did not significantly deviate from zero throughout the study.

Other laboratory measurements

Serum intact parathyroid hormone (PTH) level (pg/ml) was measured by IRMA (Allegro intact PTH kit, Nichols Institute Diagnostics, San Clemente, CA, USA) [24] with a reference range for Japanese adults of 15–50 pg/ml, and serum 1,25 dihydroxy vitamin D (calcitriol) level (pmol/l) was determined by radioreceptor assay [1,25 (OH)2D kit SRL, Yamasa, Tokyo, Japan] [25] with a reference range of 48–144 pmol/l. Serum calcium and inorganic phosphorus levels were determined simultaneously.

Bone mineral density measurement

BMD was measured by dual x-ray absorptiometry (DXA) at the lumbar spine (L2–4) and right hip in a posteroanterior projection (QDR4500A, Hologic, Bedford, MA, USA) and the distal forearm of the nondominant side (pDXA, Norland/Stratec, Fort Atkinson/Pforzheim, USA/Germany) as previously described [18]. Subjects with a history or incident involvement of fractures or bone disease in the right hip or nondominant forearm were scanned on the other side. Morphometry of the spine was performed at the same time with QDR4500A to exclude subjects with fractures at the lumbar spine from the analysis of BMD. Densitometric data of the spine for subjects with vertebral fractures [26] or 4th-grade osteophytes using Nathan’s classification [27] or those with hip or forearm deformities in regions of interest were not used in the analysis. These procedures were performed in the baseline and two follow-up studies. All measurements of the spine and hip throughout the study were made with one scanner installed in a mobile test room in a large vehicle. BMD at the forearm was also measured with one scanner. BMD at the spine (LS), total hip (TH), and distal one third of the radius (DR) was used in the analysis. The short-term precision (CV) of the BMD measurements in vivo was 1.2%, 1.2%, and 1.2% for LS, TH, and DR, respectively. No remarkable drift in BMD for the spine or forearm phantom was observed throughout the 6 years of follow-up, and the CV in vitro was 0.48% or 0.34%, respectively.

Interviews

We gave the subjects a questionnaire regarding their menstrual history and its present status, the history and present involvement of gynecological and other diseases or medications that may have affected bone metabolism, and lifestyle factors such as smoking, drinking, exercise, and dietary habits, plus a validated food frequency questionnaire for estimating dietary calcium intake [28]. We asked the subjects to fill them out before the baseline visit. Detailed interviews were conducted during the baseline survey by trained nurses and dietitians according to these questionnaires, and similar questionnaires were given at two follow-up visits. We classified the subjects who had exercised with more than usual intensity by walking for at least 30 min at a regular time two or more times per week as regular exercisers, and the others as nonexercisers. The subjects who smoked at least 1 cigarette every day at the time of the survey were defined as current smokers.

Body size measurements

Height (cm) and weight (kg) were measured using an automatic scale (TK-11,868 h, Takei Kagaku, Tokyo, Japan) in every survey. The body mass index (BMI in kg/m2) was calculated as the weight (kg) divided by height (m) squared.

Statistics

The geographic mean and SD were used for PTH, calcitriol, OC, CTX, fDPD, and tDPD because they followed a logarithmic normal distribution. The mean percent change per annum in BMD during the first 3 years and entire 6 years of follow-up compared with BMD at baseline were used as indices of the change in BMD. Effects of confounding variables were adjusted for using an analysis of covariances (ANCOVA) when appropriate in comparisons of the mean values of change in BMD between groups. Tukey’s multiple comparison procedure was applied to adjust the statistical significance level to 5% when appropriate. All statistical analyses were performed using the SAS system for personal computers (release 6.12, SAS Institute, Cary, NC, USA).

Results

Basic characteristics of subjects

Table 1 shows the basic characteristics at baseline in premenopausal and postmenopausal subjects who were divided into three groups each by age or YSM, respectively, and in perimenopausal subjects. The premenopausal group aged 30–44 years was expected to have the peak bone mass and the lowest bone turnover. Roughly speaking, the younger the subjects, the taller and heavier they were, and they experienced menarche at a younger age and included less calcium in their diet.
Table 1

The basic characteristics of the subjects at baseline. Japanese Population-based Osteoporosis (JPOS) Cohort Study

 

N

Age (year)

Height (cm)

Weight (kg)

BMI (kg/m2)

Age at menarche

Age at menopause

YSM

Calcium intake (mg/day)

Regular exerciser

Current smoker

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Prevalence (%)

Prevalence (%)

All subjects

1130

47.7

16.8

153.0

6.5

53.8

8.0

23.0

3.4

14.0

2.1

49.6a

3.2a

14.8a

8.9a

629

262

19.9

5.2

Pre

15y≦Age<30y

194

23.3

4.2

157.8

5.8

52.5

8.2

21.0

2.9

12.3

1.2

514

206

22.5

9.4

30y≦Age<45y

312

37.5

4.2

156.3

4.9

54.2

7.7

22.2

3.1

12.9

1.2

585

224

11.0

7.7

 45y≦Age

131

48.3

2.5

154.0

4.9

56.6

8.4

23.9

3.5

14.1

1.2

597

236

23.1

3.8

Peri

34

51.0

3.4

151.1

5.1

53.0

7.1

23.2

2.7

14.3

1.7

659

268

20.6

8.8

Post

YSM<10

161

55.9

4.1

150.6

4.7

54.7

7.4

24.1

3.2

14.8

1.7

50.5

3.1

5.3

2.7

715

265

15.0

2.5

10≦YSM<20

157

64.5

4.1

148.5

5.0

54.3

7.9

24.6

3.4

15.8

1.8

49.8

3.1

14.7

2.8

744

305

27.1

0.6

20≦YSM

141

73.7

3.9

146.2

5.2

50.2

7.9

23.5

3.3

16.1

2.1

48.3

3.1

25.8

4.3

690

295

28.6

2.9

N: Number of subjects SD: Standard deviation BMI: Body mass index YSM: Years since menopause Pre: Premenopausal women Peri: Perimenopausal women Post: Postmenopausal women a: Calculated for 459 postmenopausal women

Table 2 shows the mean and SD values of biochemical indices relating bone metabolism in the same groupings. No difference was found in serum calcium or phosphorus levels among the groups. Increasing trends with aging were observed in the levels of PTH and calcitriol. Biochemical markers of bone turnover showed the lowest values in the middle age group of premenopausal women and increased markedly after menopause.
Table 2

The levels of the biochemical markers of bone turnover measured at baseline. Japanese Population-based Osteoporosis (JPOS) Cohort Study

 

N

Serum

Urine

Ca (mmol/l)

P (mmol/l)

PTH (pg/ml)

Calcitriol (pmol/l)

OC (mmol/l)

Bone ALP (ng/ml)

CTX (mg/mmol Cr)

fDPD (nmol/mmol Cr)

tDPD (nmol/mmol Cr)

Meana

SDa

Meana

SDa

Meanb

SDb

Meanb

SDb

Meanb

SDb

Meana

SDa

Meanb

SDb

Meanb

SDb

Meanb

SDb

All subjects

1130

2.2

0.1

1.1

0.2

35.6

1.5

91.9

1.4

1.1

1.5

12.8

6.1

229.9

1.9

5.4

1.4

9.7

1.5

Pre

15y≦Age<30y

194

2.3

0.1

1.2

0.2

31.2

1.5

86.7

1.4

1.2

1.5

12.7

4.9

221.4

1.7

5.3

1.4

9.7

1.5

30y≦Age<45y

312

2.2

0.1

1.1

0.2

34.5

1.4

86.2

1.4

0.8

1.4

9.2

3.6

142.8

1.8

4.6

1.3

7.6

1.3

 45y≦Age

131

2.2

0.1

1.1

0.2

35.1

1.4

93.8

1.4

0.9

1.5

10.2

4.2

179.3

2.0

5.0

1.4

8.7

1.5

Peri

34

2.2

0.1

1.2

0.1

36.8

1.4

91.7

1.4

1.2

1.5

14.0

6.6

305.8

1.8

5.8

1.3

11.3

1.5

Post

YSM<10

161

2.2

0.1

1.1

0.2

36.7

1.5

94.3

1.4

1.4

1.4

15.6

5.8

375.8

1.6

6.1

1.3

12.6

1.4

10≦YSM<20

157

2.3

0.1

1.2

0.2

39.0

1.4

99.1

1.4

1.4

1.4

16.3

6.6

328.9

1.7

6.2

1.4

11.5

1.4

 20≦YSM

141

2.3

0.1

1.1

0.2

40.5

1.5

101.5

1.4

1.4

1.5

16.0

7.4

311.9

1.7

5.9

1.3

11.0

1.4

N: Number of subjects SD: Standard deviation Pre: Premenopausal women Peri: Perimenopausal women Post: Postmenopausal women

YSM: Years since menopause Ca: Calcium P: Inorganic phosphorus PTH: Intact parathyroid hormone

Calcitriol: 1,25 dihydroxy vitamin D OC: Osteocalcin Bone ALP: Bone-specific alkaline phosphatase

CTX: Type I collagen cross-linked C-terminal telopeptide normalized by the creatinine concentration

fDPD: Free form of deoxypyridinoline normalized by the creatinine concentration

tDPD: Total deoxypyridinoline normalized by the creatinine concentration

a: Arithmetic mean and SD b: Geometric mean and SD

Table 3 shows BMD at baseline and its annual change over the first 3 years and the entire 6 years of follow-up. Subjects whose BMD was excluded from the analysis constituted less than 1% of the total subjects at TH or DR but 4.4% (N=48) at LS. Most of those excluded at LS were postmenopausal (N=45) and accounted for 9.8% of the postmenopausal subjects. BMD at LS and DR increased, and BMD at TH decreased during the follow-up period in the youngest group. BMD in the middle age group of premenopausal women did not change. Other groups showed significant decrease in BMD at every skeletal site except at LS in the oldest group of postmenopausal women.
Table 3

Bone mineral density (BMD) at baseline and annual BMD change (%) over the first 3 years and the entire 6 years of follow-up. Japanese Population-based Osteoporosis (JPOS) Cohort Study

 

BMD (g/cm2) at baseline

Annual BMD change (%) during the first 3 years of follow-up

Annual BMD change (%) during the entire 6 years of follow-up

Lumbar spine

Total hip

Distal 1/3 radius

Lumbar spine

Total hip

Distal 1/3 radius

Lumbar spine

Total hip

Distal 1/3 radius

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

N

Mean

SD

All subjects

1082

0.939

0.162

1123

0.827

0.132

1128

0.674

0.122

926

−0.5

1.4

982

−0.5

1.3

993

−0.6

1.6

830

−0.5

1.0

878

−0.5

0.9

888

−0.7

1.2

Pre

15y≦Age<30y

194

1.009

0.112

192

0.888

0.111

193

0.748

0.067

152

0.1

1.1

150

−0.4

1.1

152

0.3

1.1

134

0.2+

0.7

132

−0.3

0.7

135

0.4

0.5

 30y≦Age<45y

311

1.046

0.110

309

0.889

0.105

312

0.753

0.064

269

−0.1

1.0

266

0.1

1.0

271

0.0

1.0

248

−0.1

0.8

246

0.0

0.8

250

0.0

0.6

 45y≦Age

129

1.004

0.112

130

0.887

0.099

131

0.736

0.066

113

−1.3

1.6

115

−0.6

1.3

116

−0.8

1.5

120

−1.5

1.0

121

−0.9

0.8

123

−1.2

1.3

Peri

34

0.907

0.146

34

0.819

0.112

34

0.678

0.087

30

−1.2

1.9

30

−0.7

1.3

30

−1.6

1.9

30

−1.1

1.2

30

−0.8

0.9

30

−1.5

1.8

Post

YSM<10

153

0.877

0.134

160

0.806

0.107

161

0.648

0.082

136

−1.2

1.4

147

−0.8

1.1

148

−1.8

1.8

129

−0.9

0.9

139

−0.7

0.7

140

−1.5

1.2

 10≦YSM<20

139

0.806

0.120

157

0.747

0.104

157

0.569

0.094

118

−0.4

1.5

143

−0.5

1.2

145

−1.1

1.9

101

−0.3

0.9

124

−0.7

0.9

124

−1.2

1.1

 20≦YSM

122

0.726

0.108

141

0.665

0.100

140

0.486

0.089

108

−0.1

1.6

131

−1.0

1.7

131

−0.9

1.7

68

−0.2

1.1

86

−1.0

0.8

86

−1.1

1.1

N: Number of subjects SD: Standard deviation

Pre: Premenopausal women Peri: Perimenopausal women

Post: Postmenopausal women

YSM: Years since menopause BMD: Bone mineral density

+, , : Significant change at p<0.05, p<0.01 or p<0.001, respectively

Marker levels and BMD change in premenopausal women

Table 4 shows effects of biochemical markers of bone turnover at baseline on subsequent change in BMD during the follow-up period in the premenopausal groups. The subjects of each group were classified into three groups by tertile levels of each marker. The mean annual changes (%) in BMD at LS, TH, or DR during the first 3 years and the entire 6 years of follow-up were compared among these groups after being adjusted for the effects of age, height, weight, dietary calcium intake, regular exercise, and current smoking at baseline.
Table 4

The adjusted mean annual BMD change (%) during the first 3 years and during the entire 6 years of follow-up in premenopausal subjects classified according to the tertile level of each biochemical marker of bone turnover at baseline

https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0052-3/MediaObjects/198_2005_52_Tab4_HTML.gif

Significant associations in the younger two groups were found between the levels of some bone turnover markers and BMD change at LS or TH, where the subjects with the higher values of markers showed the greatest bone gain. The oldest group of premenopausal women showed a much clearer relationship, but the trends were all inverse to those in the younger groups. The subjects with the higher values of markers suffered from greater bone loss. This tendency was observed both in the first 3 years and in the entire 6 years of follow-up at LS and TH, with the association being stronger in the first 3 years. The corresponding association at DR was apparent only in the entire 6 years.

The greatest coefficient of determination R2 of the model incorporating age, height, weight, and one of the markers was observed in the association between CTX and BMD change at TH during the first 3 years in the subjects aged 45 or older (R2=42.8%), where CTX itself explained 21.6% of the total variance. The upper tertile value of CTX in this group was 240 mg/mmolCr. The second greatest R2 was found in the association between OC and BMD change at the same site in the same group (40.0% for the model, 15.0% for OC), where the upper tertile value of OC was 1.0 nmol/l.

Marker levels and BMD change in postmenopausal women

Table 5 shows the results in postmenopausal women from analyses similar to those shown in Table 4. Mean annual changes (%) in BMD were compared among the groups classified by tertile levels of each biochemical marker after being controlled for the effects of age at menopause, YSM, height, weight, dietary calcium intake, regular exercise, and current smoking at baseline. Significant associations of marker levels with BMD change were limited to DR in the subjects with less than 10 YSM. Some statistically significant trends were seen in other postmenopausal groups, but they were sporadic and inconsistent across either the markers or skeletal sites.
Table 5

The adjusted mean annual BMD change (%) during the first 3 years and during the entire 6 years of follow-up in postmenopausal subjects classified according to the tertile level of each biochemical marker of bone turnover at baseline

https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0052-3/MediaObjects/198_2005_52_Tab5_HTML.gif

Marker levels and BMD change in perimenopausal women

All subjects in the oldest group of premenopausal women menstruated regularly at baseline, but 74.4% of them entered menopause during the 6 years of follow-up. We pooled the subjects in this group with 34 women who were perimenopausal at baseline and divided the pooled group into two subgroups according to whether the subjects entered menopause during the study period or not. Table 6 shows the association between the marker levels at baseline and subsequent BMD change in these subgroups, together with the results for early postmenopausal subjects with less than 5 YSM at baseline.
Table 6

The adjusted mean annual BMD change (%) during the first 3 years and during the entire 6 years of follow-up in the subjects undergoing menopausal transition classified according to the level of each biochemical marker of bone turnover at baseline

https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0052-3/MediaObjects/198_2005_52_Tab6_HTML.gif

The strongest associations between the marker levels and BMD changes were observed in the subjects who entered menopause during the follow-up period. A similar significant trend was seen for OC and CTX in the subjects who still menstruated at follow-up. In the early postmenopausal group, however, most of the associations did not reach statistical significance.

Discussion

Fracture prevention is the ultimate aim of every sort of preventive activity against osteoporosis. Therefore, evaluation of fracture risk on an individual basis is an essential role of screening and diagnostic tests for this disease. Current evidence suggests that some biochemical markers of bone turnover are valid for predicting fracture risk in elderly people [2932]. But this is not confirmed for perimenopausal or early postmenopausal women because they rarely suffer from osteoporotic fractures at their current ages. Decrease in BMD may be used as a proxy outcome for fracture risk because those women currently face a high risk of rapid bone loss, which leads to an elevated risk of fracture later in life [33].

Christiansen et al. [1, 2] reported that a diagnostic model incorporating bone turnover markers predicted change in BMD at the forearm and the incidence of fast bone losers very accurately in early postmenopausal women. This study presented a possibility for predicting whether a woman would lose bone rapidly before she actually lost the bone. Many similar studies have been performed since then. However, no other authors have obtained such high accuracy in predicting BMD change at the spine [813] or hip [1315] or even at the forearm [37] in early or late postmenopausal women.

In the present study, we obtained some significant associations between the levels of bone turnover markers at baseline and subsequent bone loss at DR in postmenopausal women with YSM less than 5 or 10 years, but the association was not consistent across the skeletal sites. Furthermore, virtually no association was found in other postmenopausal groups. These findings confirm the conclusions drown by the reviews of Delmas et al. [16] and Stepan [17]. As shown in Tables 2 and 3, all the biochemical markers of bone turnover increased rapidly at the time of menopause, and BMD was synchronized to decrease. However, most markers remained elevated even at the late postmenopausal stage, while bone loss slowed down with the increase in YSM. Different mechanisms may be involved in the elevation of marker levels and the loss of bone in postmenopausal women with 10 YSM or more. It may be difficult, therefore, to predict change in BMD using the level of biochemical markers at baseline in women at this postmenopausal stage. However, we might have underestimated the real associations. We excluded 76 postmenopausal subjects from the analyses because they had started receiving drug therapy for osteoporosis after the baseline. We may thereby have excluded people at higher risk for rapid bone loss.

One of the interesting findings of the present study is a significant association between the levels of bone turnover markers and BMD change observed in premenopausal women aged 45 years or older. Only a few studies have evaluated such an association in premenopausal women, and no significant association was observed [6, 8, 13]. Similarly, in the present study no consistently significant association between the marker levels and BMD change was observed in premenopausal women less than 45 years of age. However, significant associations were found in the subjects aged 45 years or older, even when the subjects were limited to those who still menstruated at the end of the follow-up period. This tendency was more clearly observed in subjects who were premenopausal or perimenopausal at baseline and entered menopause during the follow-up period. Previous studies have also shown significant correlations between the level of bone turnover markers and BMD change when the study subjects included women who were perimenopausal at baseline [6, 13, 34, 35].

More interestingly, the association between the marker levels and BMD change was found to be more modest in women who were already in the early postmenopausal stage at baseline, and it was virtually absent in postmenopausal women with 10 or more YSM. Chaki et al. [12] reported that the urinary level of type I collagen crosslinked N-terminal telopeptide (NTX) had a relatively strong predictive value for subsequent bone loss up to 3 years from the baseline in Japanese women, even including late postmenopausal women; however, the researchers failed to obtain a significant association for other markers, including OC, bone ALP, CTX, and fDPD. It is not clear why only NTX showed a significant association with BMD change. The predictive value of NTX, however, was probably overestimated in the study of Chaki et al. [12], because the authors did not allow for the confounding effects of age, YSM, and body size in their analyses. It is probable, based on the present and related findings, that the biochemical markers of bone turnover measured for premenopausal women or perimenopausal women who will enter menopause within 6 years indicate the subsequent changes in BMD.

The value of this result, however, may be limited for practitioners of preventive medicine because it is difficult to determine whether a certain premenopausal woman will enter menopause within 6 years. Serum estradiol and gonadotropin levels may give some clues, but they can hardly predict accurately the time of menopause on an individual basis [36, 37]. Therefore, the result of the present study for premenopausal women aged 45 years or older may be important. The linear model for these women, incorporating age, height, weight, and CTX, explained 42.8% of the variance of BMD change at TH whether they would enter menopause or not. This coefficient of determination is the highest to date among the R2 values reported for this skeletal site, regardless of the age group. Therefore, we suggest that professionals in charge of osteoporosis prevention monitor BMD change carefully in premenopausal women aged 45 and older with relatively low BMD and elevated CTX levels.

The present study has several advantages over previous studies in study design, such as using a representative sample of the general population in a larger scale with an acceptable participation rate at baseline and follow-up. However, the study's limitations should also be addressed. The subjects were randomly selected from the inhabitants of the area, but the study areas were not randomly selected from all the municipalities in Japan. Although not perfect, our sampling method was a feasible way to conduct a nationwide survey. The number of perimenopausal women at baseline or the number of subjects undergoing menopausal transition during the follow-up period was small compared with the other groups. This led to a decrease in the statistical power of the analyses for this group and may have caused false negative results. Blood specimens for measuring the biochemical markers were not taken from subjects without strict control over sampling time or meals. This may have increased errors in the values of OC and bone ALP and may have blurred the results.

In conclusion, the present study provides relatively strong evidence for the predictive value of biochemical markers of bone turnover, including OC and CTX, for subsequent bone loss in premenopausal women aged 45 or older and for perimenopausal women, and little remarkable value for prediction in younger premenopausal women or in postmenopausal women. We may use the biochemical markers of bone turnover to predict bone loss in women at menopausal transition rather than for those after menopause.

Acknowledgements

This study was conducted by the JPOS Study Group, comprising Fumiaki Marumo (Chairman of the Study Group, Professor Emeritus, Tokyo Medical and Dental University), Toshihisa Matsuzaki (Co-chairman of the Study Group, Institute of Comprehensive Community Care), Etsuko Kajita (Nagoya University School of Health Sciences), Tomoharu Matsukura (Kanazawa University), and Takashi Yamagami (Hokuriku Health Service Association), along with the authors. Financial support for the baseline survey was provided by the Japan Milk Promotion Board and the Japan Dairy Council. The follow-up surveys were financially supported by a Grant-in-aid for Scientific Research (B #10470114, 1998–2000, and B #1437017, 2002–2003) from the Japanese Society for the Promotion of Science and a grant in 2000–2002 from the Research Society for Metabolic Bone Diseases, Japan. The authors wish to express special thanks to the personnel of the health departments of Hirara City, Sanuki City, and Nishi-Aizu Town for their excellent support of the study, and to those from SRL, Tokyo, Japan; Toyo Medic, Osaka, Japan; and Toyukai Medical Corporation, Tokyo, Japan, for their technical assistance with the surveys.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2006