SEARCH

SEARCH BY CITATION

Keywords:

  • osteoporosis;
  • bone markers;
  • estrogens;
  • androgens;
  • parathyroid hormone

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The mechanisms leading to increased bone loss and skeletal fragility in women with postmenopausal osteoporosis are still poorly understood. Increased bone resorption, low serum estradiol and high serum sex-hormone-binding globulin (SHBG) recently have been reported as predictors of vertebral and hip fractures in elderly women. In a cohort of healthy untreated younger postmenopausal women aged 50–89 years (mean, 64 years), we compared baseline levels of bone markers and endogenous hormones in 55 women who subsequently had a fracture (20 vertebral and 35 peripheral fractures) with levels in the 380 women who did not fracture during a mean 5 years of follow-up. Women with levels in the highest quartile of four bone resorption markers including urinary-free deoxypyridinoline (D-Pyr), urinary type I collagen N-telopeptides (NTX), and urinary and serum type I collagen C-telopeptides (CTX) had about a 2-fold increased risk of fractures compared with women with levels in the three lowest quartiles with relative risk (RR) and 95% CI of 1.8 (1.0-3.4) for free D-Pyr, 1.7 (0.9-3.2) for urinary NTX, 2.3 (1.3-4.1) for urinary CTX, and 2.1 (1.2-3.8) for serum CTX. Serum levels of bone alkaline phosphatase (BAP) in the highest quartile were associated with an RR of fracture of 2.4 (1.3-4.2). Women with serum levels of estradiol and dehydroepiandrosterone (DHEA) sulfate in the lowest quartile had an RR of fracture of 2.2 (1.2-4.0) and 2.1 (1.2-3.8), respectively. Increased levels of SHBG and intact parathyroid hormone (PTH) were moderately associated with an increased risk of fracture. Similar results were obtained when the analysis was restricted to symptomatic vertebral and nonvertebral fractures. Adjustment of biochemical markers by hormone levels did not significantly alter the results. Women with both high bone resorption markers and low estradiol (or low DHEA sulfate) had a higher risk of fracture with RRs of 3.0-3.3 (p < 0.001). After adjustment for bone mineral density (BMD) of the hip, spine, radius, or total body, bone markers and hormones were still predictive of fracture risk with similar RRs. We conclude that high levels of some biochemical markers of bone turnover, low serum estradiol, low DHEA sulfate, high SHBG, and high PTH are associated with increased risk of osteoporotic fracture in postmenopausal women, independently of each other and of BMD. The mechanism by which some postmenopausal women have an increased rate of bone turnover leading to an increased risk of fracture remains to be elucidated.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

THE MECHANISMS responsible for accelerated bone loss and increased fracture risk in some—but not all—postmenopausal women are still poorly understood. Indeed, although the importance of estrogen deficiency in the rapid bone loss around the time of menopause has been recognized for several years,(1,2) circulating estradiol levels explains only a small proportion of the interindividual variance of bone mineral density (BMD) and bone loss.(3) Lower androgen concentrations and increased sex-hormone-binding globulin (SHBG) also have been reported to be weakly associated with lower BMD and rapid bone loss at some but not all skeletal sites.(3) However, in elderly women, three recent reports of the Osteoporotic Fracture Study suggest that undetectable levels (<5 pg/ml) of estradiol, which occurred in about one-third of the population, are strongly associated with low BMD,(4) rapid bone loss,(5) and thus increased fracture risk.(6) In the same study, increased levels of SHBG and low levels of 1,25-dihydroxyvitamin D3[1,25(OH)2D3] but not 25-hydroxyvitamin D [25(OH)D] also were associated with fractures.(6)

Estrogen deprivation at the time of menopause is associated with an increased bone turnover(7–10) that persists in the elderly(10) and that can be reduced to premenopausal levels with estrogen replacement therapy in young women(8,9,11,12) but also in late postmenopausal women.(13) In elderly women, increased levels of biochemical markers of bone turnover, more specifically of bone resorption, have been shown to be associated with an increased risk of hip fracture independently of BMD,(14,15) possibly because of trabeculae perforation that decreases bone strength.(16) Thus, the relationships between low estradiol and increased fracture risk could be mediated through an increase of bone turnover. To our knowledge, no study has investigated, in the same population, the relation between bone turnover, hormonal levels, and fracture risk. In addition, most prospective studies looking at the determinants of fractures have been performed in elderly women (because of the high incidence of osteoporotic fractures) but the role of the residual secretion of steroids is more likely to play a role in younger postmenopausal women.

To test the hypothesis that decreased estrogen levels leading to increased bone turnover could play an important role in skeletal fragility in postmenopausal women, we compared baseline concentrations of markers of bone turnover and steroids and calciotropic hormones in women who subsequently sustained a fracture with women who did not fracture during the follow-up. All women belonged to a population-based cohort of postmenopausal women with a wide age range of 50–89 years.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Subjects

Four hundred and thirty-five healthy female volunteers who had been enrolled in a study of the determinants of bone loss (OFELY study) were included. The cohort of this study comprises 1039 women, 31–89 years of age, randomly selected from the regional section of a health insurance company (Mutuelle Générale de l'Education Nationale). Among the 672 postmenopausal women (menopause was defined as an absence of menses for at least 12 months) recruited at baseline, 237 women were subsequently excluded because of the presence of treatment or diseases that could influence levels of markers of bone turnover and/or hormones. These included 162 women who had received hormone replacement therapy in the past 12 months, 7 patients on tamoxifen, 11 patients taking fluoride, 16 women receiving bisphosphonates, 4 women treated with calcitonin, 24 women receiving thyroid hormones, 5 women taking corticosteroid, and 8 women who had a disease known to affect bone metabolism (Paget's disease of bone, primary hyperparathyroidism, hyperthyroidism, cirrhosis, or diabetes mellitus). The OFELY cohort has been described elsewhere.(10,17)

Identification of fractures

At baseline, women were submitted to a detailed questionnaire including history of fragility fractures, that is, resulting from a low trauma, since the age of 30 years; physical activity; calcium intake; and smoking habits. Past and present physical activity was registered by questionnaire. The monthly hours of physical exercise, the daily walking distance and number of stairs climbed, the weekly hours of home work, and the physical demands of professional work (setting, light, medium, and heavy) were recorded and used to calculate an individual physical activity score. Dietary calcium intake was assessed by a sequential self-questionnaire. Subsequently, each woman was seen every year for a maximum of 6 years. At each visit, the occurrence of fractures within the previous year was registered. For women who did not continue, mail was sent every year to identify the occurrence of fractures. All peripheral fractures were confirmed by radiographs.

Lateral X-ray films of the thoracic and lumbar spine were obtained at baseline for all women and at follow-up for 79% of them after an average of 3.8 years (2.8-5.3 years). All vertebral prevalent and incident fractures were identified by quantitative morphometry by two individuals who were unaware of the baseline assay results. A vertebra was classified as having a prevalent fracture on the baseline radiograph if any vertical height (anterior, middle, and/or posterior) was reduced by more than 20%. A new fracture was defined by a decrease of 20% or more and of at least 4 mm in any vertebral height of one or more thoracic or lumbar vertebrae between follow-up and baseline X-ray films.(18,19)

Biochemical analyses

For each woman, fasting blood samples were collected at baseline between 7:30 a.m. and 9:30 a.m. for measurements of hormones and serum bone turnover markers. Total 24-h urinary excretions also were collected at baseline without any preservative for measuring urinary resorption markers. Serum and urine samples were stored frozen at −80°C until assayed.

Markers of bone turnover

Bone formation markers. Serum osteocalcin was measured with a human immunoradiometric assay (IRMA), which uses two monoclonal antibodies recognizing, respectively, the 5–13 and 43–49 sequence of the molecule and purified intact human bone osteocalcin as a standard (ELSA-OST-NAT; Cis Biointernational, Gif/Yvette, France). The intra- and interassay CVs are typically below 10%, and the sensitivity is 0.3 ng/ml.(20) Serum bone-specific alkaline phosphatase (BAP) was measured with a human specific IRMA using two monoclonal antibodies directed against the human bone isoenzyme and BAP purified from human SAOS-2 osteosarcoma cells as a standard (Ostase; Hybritech, Inc., San Diego, CA, U.S.A.). The intra- and interassay CVs are less than 10%.(21) Serum C-terminal propeptide of type I collagen (PICP) was measured by a two-site ELISA (Prolagen-C; MetraBiosystems, Palo Alto, CA, U.S.A.). The intra- and interassay CVs are below 8%. Serum intact N-terminal propeptide of type I collagen (PINP) was measured by a radioimmunoassay (Intact PINP; Farmos Diagnostica, Upsalla, Finland).(22) The intra- and interassay CVs of PINP immunoassay are less than 5% and 8%, respectively.

Bone resorption marker.

Urinary free deoxypyridinoline (D-Pyr) was measured by an ELISA that uses a monoclonal antibody, which has less than 1% cross-reactivity with free Pyr (Pyrilinks-D;, MetraBiosystem) and no significant interaction with cross-linked peptides.(23) Urinary type I collagen N-telopeptides (NTX) breakdown products were measured by ELISA (Osteomark; Ostex International, Inc., Seattle, WA, U.S.A.). The intra- and interassay CVs are less than 10%.(24) Urinary type I collagen C-telopeptide (CTX) breakdown products were measured by an ELISA (CrossLaps ELISA; CIS Biointernational) based on an immobilized synthetic peptide with an amino acid sequence specific for a part of the C-telopeptide of the α1-chain of type I collagen (Glu-Lys-Ala-His-βAsp-Gly-Gly-Arg; CrossLaps antigen).(25) The intra- and interassay CVs are less than 5% and 8%, respectively. Serum CTX breakdown products were measured by a two-site ELISA (Serum Crosslaps one step; Osteometer Biotech, Ballerup, Denmark) using monoclonal antibodies raised against an amino acid sequence specific for a part of the C-telopeptide of the α1-chain of type I collagen (Glu-Lys-Ala-His-β Asp-Gly-Gly-Arg).(26) Intra- and interassay CVs are lower than 5% and 8%, respectively.

Urinary-free D-Pyr, NTX, and CTX data were corrected by the urinary creatinine (Cr) concentration measured by a standard colorimetric method.

Hormones

Serum total estradiol, estrone, testosterone, and dehydroepiandrosterone (DHEA) sulfate were measured by tritiated radioimmunoassays after diethylether extraction or dilution (for DHEA sulfate).(27) The total assay precision is 10% for estradiol, 10% for estrone, 9% for testosterone, and 10% for DHEA sulfate. The limit of detection is 3 pg/ml for estradiol and estrone and 60 pg/ml for testosterone. SHBG was measured by IRMA (125I SBP coatria; Bio-Mérieux, Marcy l'Etoile, France) with an interassay variation of 7%. Serum intact parathyroid hormone (PTH) was measured with a two-site immunochemiluminometric assay (Magic Lite PTH; Ciba-Corning, East Walpole, MA, U.S.A.). The intra- and interassay variations are lower than 6% and 11%, respectively, and the limit of detection is 1.4 pg/ml. 25(OH)D level was determined by a radiobinding assay kit (Bühlmann Laboratories AG, Basel, Switzerland) with intra- and interassay variations below 7% and 14%, respectively, and a sensitivity of 4 ng/ml.

BMD

We measured BMD of the spine (L1-L4), femoral neck, distal radius, and whole body by dual-energy X-ray absorptiometry (DXA) on a QDR 2000 device (Hologic, Bedford, MA, U.S.A.). The short-term in vivo precision of DXA was 0.9, 1.0, 0.6, and 1.0% for lumbar spine, total hip, distal radius, and whole body, respectively.(17)

Statistical analyses

Comparison of baseline characteristics between fractures and controls was assessed by unpaired Student's t-tests, after logarithmic transformation of biochemical marker data. The relationships between baseline levels of bone turnover markers, hormones, and the risk of fracture were analyzed by logistic-regression analysis after adjustment for potential confounding variables such as age, body mass index (BMI), prevalent fractures, and physical activity. The mean and SD values used to calculated T scores for biochemical markers and BMD were obtained from the premenopausal women of the same OFELY cohort.(10,17)

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

During the mean 5.0 ± 1.3 years of follow-up, 21 vertebral fractures and 37 peripheral osteoporotic fractures (12 wrist, 5 hip, 5 rib, 5 ankle, 3 patella, 3 humerus, 1 sacrum, 1 pelvis, and 2 metatarsi) were recorded in 55 women. The time of follow-up was different for vertebral fractures, which were identified on X-ray films of the spine that were obtained an average of 3.8 years after the baseline visit, and the peripheral fractures, which occurred after an average of 3.0 years follow-up. Ten out of the 20 women with vertebral fractures suffered a symptomatic fracture. Women who had osteoporotic fractures during the study were older; had slightly lower physical activity; and had a lower whole body, hip, spine, and radius BMD but did not differ for weight, height, BMI, or thiazide use (Table 1). As expected, the proportion of women with prevalent osteoporotic fractures was also higher among women who fractured during the study. Baseline levels of biochemical markers of bone turnover and hormones are shown in Table 2.

Table Table 1.. Baseline Characteristics of Women with Incident Osteoporotic Fractures and Controls
Thumbnail image of
Table Table 2.. Baseline Biochemical Markers of Women with Incident Osteoporotic Fractures and Controls
Thumbnail image of

Markers of bone turnover

After adjustment for age, prevalent osteoporotic fractures, and physical activity, women with levels of markers of bone turnover in the highest quartile or above the upper limit of the premenopausal range had a greater risk of osteoporotic fractures than women who had lower values, with an increased risk of 1.3-2.4 according to the marker. The relative risks (RRs) were significant for serum BAP and urinary and serum CTX (Table 3). As shown on Fig. 1, the number of incident fractures is higher in those women with baseline urinary CTX levels in the highest quartile compared with those women with levels in the three lowest quartiles, irrespective of the time of follow-up. When women were categorized according to quartiles of baseline bone marker levels, there was no proportional increase in the risk of osteoporotic fractures, but women with levels in quartile 4 had a 2-fold increased risk of fracture compared with women with levels in quartiles 1, 2, and 3 (Fig. 2). When results were additionally adjusted for body weight and height, increased levels of serum BAP (RR [95% CI], 2.3 [1.3-4.1]), urinary CTX (2.2 [1.2-3.9]), and serum (2.0 [1.1-3.6]) CTX remained significantly associated with an increased risk of fracture. When women with prevalent fractures were excluded from the analyses, increased levels of serum BAP, urinary CTX, and serum CTX were associated with increased fracture risk, although not significantly for serum BAP (Table 4).

Table Table 3.. Adjusteda Association Between Increased Bone Turnover and the Risk of Osteoporotic Fracture in Postmenopausal Women
Thumbnail image of
Table Table 4.. Bone Turnover and Hormonal Predictors of Osteoporotic Fracture in Postmenopausal Women Without Prevalent Fractures
Thumbnail image of
thumbnail image

Figure Figure 1. Cumulative number of fractures according to time of follow-up in women with baseline levels in the highest quartile of urinary CTX (high risk group) and in women with levels in the three lowest quartiles (low risk group).

Download figure to PowerPoint

thumbnail image

Figure Figure 2. Serum levels of BAP and urinary levels of CTX at baseline and age-adjusted risk of subsequent fracture in postmenopausal women. The reference group consisted of the women with levels of bone markers in the lowest quartile.

Download figure to PowerPoint

Serum hormones

Steroid hormones

Women with serum levels of estradiol in the lowest quartile of the population, corresponding to values below 11 pg/ml, had an RR of fracture of 2.2 compared with women with levels greater than 11 pg/ml (Table 5; Fig. 3). Decreased levels of DHEA sulfate were associated with increased risk of fractures (Table 5). Women with levels in the lowest quartile had an RR of 2.1 compared with women with levels in quartiles 2, 3, and 4 (Fig. 3). Each increase of 1 μg/dl in the concentration of SHBG was associated with an RR of fracture of 1.5 (95% CI, 1.1-2.1) and women with levels in the highest quartile had an RR of fracture of 1.6 (95% CI, 0.9-2.9) compared with women whose concentrations are in the three lowest quartiles. There was no significant association between increased or decreased serum levels of estrone or testosterone and the risk of osteoporotic fractures (Table 5). After adjustment for BMI, RRs of fracture were not significantly altered and remained significant for estradiol (RR [95% CI], 2.1 [1.2-3.9]) and DHEA sulfate (2.0 [1.1-3.7]). However, BMI-adjusted RR was NS for SHBG (1.4 [0.9-2.0] per each increase of 1 μg/dl). Adjustment by body weight and height gave results similar to those observed after adjustment by BMI (data not shown). Adjustment for thiazide use did not influence the relationship between hormone levels and the risk of fracture with the same RRs than unadjusted values. After exclusion of women with prevalent fractures, decreased levels of estradiol and DHEA sulfate and increased levels of SHBG remained associated with a higher risk of fracture with similar a odds ratio, which, however, did not reach significance for estradiol (RR, 1.9 [0.97-3.8]; p = 0.06; Table 4).

Table Table 5.. Adjusteda Association Between Serum Hormone Levels and the Risk of Osteoporotic Fracture in Postmenopausal Women
Thumbnail image of
thumbnail image

Figure Figure 3. Serum levels of endogenous hormones at baseline and age-adjusted risk of subsequent fracture in postmenopausal women. The reference group consisted of women with levels of hormones in the highest quartile for serum estradiol and DHEA sulfate and women with levels in the lowest quartile for serum SHBG and intact PTH.

Download figure to PowerPoint

Serum PTH and 25(OH)D

Women whose serum concentrations of intact PTH were in the highest quartile (>51 pg/ml) had an increased risk of fractures (RR, 1.8; 95% CI, 1.01-3.4; Table 5 and Fig. 3. Adjustment for BMI did not significantly alter the results (RR [95% CI], 1.8 [1.01-3.2]). Exclusion of women with prevalent fractures slightly improved the association between increased levels of serum PTH and higher fracture risk (Table 4). Women with both levels of intact PTH in the highest quartile and serum levels of estradiol below 11 pg/ml, which accounted for 6% of the population, had an RR of fracture of 3.7 (95% CI, 1.6-8.4) compared with all the other women. There was no statistically significant association between serum 25(OH)D and the risk of fractures (Table 5).

Markers of bone turnover, endogenous hormones, and the risk of symptomatic vertebral and nonvertebral fractures

When we restricted the analysis to women who sustained either a symptomatic vertebral fracture (n = 10) or a nonvertebral fracture (n = 35) or to those with only nonvertebral fractures, we found that increased levels of serum BAP, urinary CTX, and serum CTX are significantly associated with an increased risk of fracture with an RR of about 2.3-2.5 (Table 6). Decreased levels of serum estradiol, DHEAS sulfate, and increased levels of PTH also were associated with increased risk of symptomatic vertebral and nonvertebral fractures (Table 6). However, when the analysis was restricted to nonvertebral fractures only, the RRs did not reach significance for serum estradiol and PTH.

Table Table 6.. Bone Turnover and Hormonal Predictors of Symptomatic Osteoporotic Fractures in Postmenopausal Women
Thumbnail image of

Markers of bone turnover and endogenous hormones

After adjustment for age, serum BAP (r = −0.13; p = 0.006), urinary CTX (r = −0.12; p = 0.01), and serum CTX (−0.13; p = 0.008) were negatively correlated with serum estradiol. Bone turnover markers were weakly correlated with serum intact PTH (r = 0.17 and p = 0.0005, r = 0.12 and p = 0.02, and r = 0.15 and p = 0.02 for serum BAP, urinary CTX, and serum CTX, respectively) and serum SHBG (r = 0.20 and p < 0.0001 and r = 0.15 and p = 0.002 for urinary and serum CTX, respectively). As shown on Table 7, after adjustment of bone turnover markers by serum levels of estradiol, DHEA sulfate, SHBG, and PTH, serum BAP, urinary CTX, and serum CTX in the highest quartile remained associated with increased risk of fracture with RRs similar to nonadjusted values, although not significantly for serum CTX (Tables 3 and 7). Adjustment of bone turnover markers by the ratio estradiol/SHBG, an indirect index of free estradiol, provided data similar to that obtained after adjustment for total estradiol (data not shown). Additional adjustment by BMI, did not alter the results presented on Table 7 with similar odds ratios and significance levels (data not shown), except for SHBG in multivariable adjusted models (RR [95%CI], 1.4 [0.9-2.0]). Women with both levels of serum CTX in the highest quartile and serum estradiol below 11 pg/ml (8% of the population) had an RR of fracture of 3.0 (95% CI, 1.3-6.7). Similarly, women with both serum levels of CTX in the highest quartile and serum levels of DHEA sulfate in the lowest quartile (8% of the population) had an RR of fracture of 3.3 (95% CI, 1.6-6.9).

Table Table 7.. Bone Turnover and Hormonal Predictors of Osteoporotic Fracture in Postmenopausal Women According to Multivariate Modelsa
Thumbnail image of

Adjustment for BMD

Adjustment for femoral neck BMD slightly weakened the association between increased levels of serum BAP, urinary CTX, serum CTX and the risk of fracture (Table 7), which remained significant for serum BAP and urinary CTX. Adjustment for hip BMD slightly decreased the association between hormones and the risk of fracture. After adjustment for BMD, the association between SHBG, intact PTH, and fracture risk was no longer significant (Table 7). After adjusting for BMD as a dichotomous variable with a T score of −2.5 to separate low and high BMD, similar results were obtained. For example, levels of urinary CTX in the highest quartile and levels of estradiol in the lowest quartile were associated with a higher risk of fracture with an odds ratio of 2.0 (1.1-3.6) and 2.2 (1.2-4.1), respectively. Adjustment for BMD of other skeletal sites including spine, radius, and total body gave similar results (data not shown). Women with both low hip BMD (T score ≤ −2.5) and levels of urinary CTX in the highest quartile were at a higher risk of fracture than women with either low BMD or high urinary CTX (Fig. 4). Women with both prevalent fractures and increased urinary CTX had a, RR of fracture of 3.8. Women with low BMD, prevalent fractures, and increased urinary CTX had a higher risk of fracture with an odds ratio of 5.8 (Fig. 4).

thumbnail image

Figure Figure 4. Combination of different independent predictors to identify women with the highest risk of fracture. Low hip BMD was defined as values at 2.5 SD or below the mean of young adults. High urinary CTX corresponds to values in the highest quartile. RRs are adjusted for age and physical activity.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

In this study we found that increased levels of some sensitive biochemical markers of bone turnover are associated with increased risk of osteoporotic fractures in a large prospective cohort of healthy untreated postmenopausal women from 50 to 89 years of age. Low serum levels of estradiol and DHEA sulfate and increased levels of SHBG and intact PTH were associated with an increased risk of fracture.

We found that women with levels of BAP above the premenopausal range were at higher risk for fracture with an RR of 2.4. Increased levels of the other markers including serum osteocalcin and propeptides of type I collagen also were associated with increased fracture risk, although not significantly. The reasons for the discrepancy between the predictive value of the different markers of bone formation are unclear, but different markers reflect different aspects of bone metabolism. BAP is an enzymatic activity of the osteoblastic cells unlike serum osteocalcin and type I collagen propeptides. In addition, serum BAP is likely to reflect a different stage of osteoblastic differentiation compared with osteocalcin and type I collagen. Different behaviors of serum BAP and serum osteocalcin have been reported in other clinical situations such as Paget's disease of bone, chronic renal failure, hyperthyroidism, osteolytic metastasis, and after glucocorticoid treatment,(28–30) although in most cases the reasons for these differences remain unclear. Our results contrast with the absence of a significant relationship between markers of bone formation including serum BAP and the risk of hip fracture that we previously reported in the Epidémiologie des Ostéoporoses (EPIDOS) cohort.(14) Several differences between the two studies may account for this discrepancy. In the present study we analyzed the predictive value for all types of osteoporotic fractures including vertebral and not only hip fractures, although when the analysis was restricted to peripheral fractures increased levels of BAP remained associated with higher risk of fractures. Women were on an average 19 years younger (64 years vs. 83 years). The duration of follow-up in the EPIDOS study was only of 22 months compared with 5 years in this study. Recently, we have shown that increased levels of these bone formation markers are associated with a significantly greater rate of bone loss in postmenopausal women.(31) Thus, if the increased risk of fracture is mediated through a more rapid rate of bone loss,(32) a follow-up of several years may be necessary to detect it. Although increased levels of all four resorption markers investigated were associated with a higher risk of fracture, the predictive value of free D-Pyr and NTX was slightly lower and NS. The reasons for this discrepancy between resorption markers are unclear. Potential explanations include a lower bone specificity and responsiveness(33,34) of the free fraction of pyridinium cross-links, which does not seem to be generated directly from osteoclastic bone resorption,(35) but is mainly released by the degradation of peptide bound cross-links though renal metabolism.(36) Urinary NTX is likely to reflect thedegradation of both isomerized and not isomerized type I collagen(37) in contrast to serum and urinary CTX, which specifically recognize the β-isomerized form.(38) β-Isomerization, a spontaneous post-translational modification, is believed to reflect aging of type I collagen(37,38) and has shown to be more prominent in bone than in soft tissues.(39) That pattern might explain the lower sensitivity of urinary NTX, which was also not a significant predictor of the risk of hip fracture in the EPIDOS study.(14) Thus, measuring the degradation of β-isomerized type I collagen by serum or urinary CTX may reflect, more specifically, resorption of mature bone matrix than urinary NTX. Despite the difference in the predictive value of the various biochemical bone markers, it is of interest to note that in both EPIDOS and OFELY studies, the RR of fracture associated with increased bone turnover markers is consistently around 2, that the nonproportional pattern of relationship between the levels of bone markers and the risk of fracture is similar, and that bone resorption markers predict osteoporotic fractures independently of BMD measured at several sites. Thus, combination of the assessment of bone turnover and BMD may improve the detection of women with the highest risk of fracture not only in elderly but also in younger postmenopausal women.

Women who had serum estradiol levels in the lowest quartile, that is, with levels below 11 pg/ml, had an increased risk of subsequent osteoporotic fractures. These data extend the results, recently reported by Cummings et al.,(6) in elderly women to younger postmenopausal women. We used a level of 11 pg/ml to identify patients at high risk of fracture compared with 5 pg/ml in the study of Cummings et al.,(6) because only 6% of the women in our study had levels below 5 pg/ml compared with 33% in the Cummings's study. The change in the cut-off could be related to differences in the standardization of the assays or more probably because of the younger age of our population. Increased levels of SHBG also were associated with an increased risk of osteoporotic fractures, probably because of decreased bioavailable estradiol although other unidentified mechanisms could be involved. We found for the first time that decreased levels of serum DHEA sulfate are associated with incident osteoporotic fractures. Previous small retrospective case-control studies have either reported decreased levels of DHEA(40) in patients with vertebral fractures or no significant differences.(41) Although DHEA sulfate, which is interconvertible with DHEA, is not androgenic by itself and does not bind to androgen receptors, it is metabolized to androgens such as testosterone and 5α-dihydrotestosterone and also may be metabolized to 17β-estradiol and the weakly estrogenic D5-androstene-3β. Thus, decreased DHEA sulfate may reflect decreased levels of both androgens and estrogens. We also found that osteoporotic fractures were more common in women with high levels of PTH. These results contrast with the absence of significant association between PTH levels and increased vertebral or hip fractures reported in two prospective studies performed in elderly women(6,14) for unclear reasons. Thus, this association needs to be confirmed in other studies. Interestingly, increased PTH and decreased estradiol predicted osteoporotic fractures independently and the combination of both estradiol levels below 11 pg/ml and PTH levels above 51 pg/ml indicated a high risk of fracture.

We found low correlation between hormones and bone turnover marker levels with hormone levels explaining less than 7% of the interindividual variance of bone turnover markers. After single or multiple adjustments for hormones, increased levels of biochemical markers of bone turnover still predict osteoporotic fractures with very similar RRs. In addition, women with both increased bone resorption and decreased estradiol or decreased DHEA sulfate were at higher risk of fracture than women with either one of these two predictors. These findings suggest that the association between high bone turnover and increased risk of fracture may not be explained by the association between decreased hormone levels and increased fracture risk. Low estradiol and indirectly low DHEA sulfate may be associated with higher osteoporotic fracture risk by a non bone turnover mechanism such as increased osteocyte death.(42,43) An alternative hypothesis could be that because we did not measure bioavailable estradiol but did measure total estradiol we could not detect such a relationship. However, the ratio of estradiol/SHBG, an indirect index of bioavailable estradiol, provided data similar to total estradiol.

Our study has some limitations. We may have not identified all vertebral fractures, because X-ray evaluation was performed only once after 4 years. However, all symptomatic vertebral fractures were recorded and the potential presence of few nonsymptomatic vertebral fractures in the control group would have resulted in an underestimation of the association between increased bone turnover (or decreased hormone levels) and fracture risk. The follow-up was different for vertebral and nonvertebral fractures, although when the analysis was restricted to nonvertebral fractures, increased levels of bone turnover markers also were associated with increased fracture risk. The number of fractures investigated is relatively low and thus the CIs on the RRs are quite large. We measured 25(OH)D and not 1,25(OH)2 D3, which has been shown to be associated with increased risk of hip fracture.(6) Our results are based on a single measurement of markers of bone turnover and hormones, which may not accurately reflect the bone turnover and hormone status because of the day-to-day variability in the measurements. However, this variability is likely to underestimate the true association that we have found between markers, hormones, and fracture risk.

In conclusion, we found in a large cohort of postmenopausal women from 50 to 89 years of age that increased bone turnover, especially bone resorption, is associated with an increased risk of osteoporotic fracture independently of BMD. Women with low levels of estradiol or DHEA sulfate and increased levels of SHBG or PTH were also at increased risk of fracture but these hormones did not account for the increase in bone turnover. Further studies should try to elucidate the mechanisms by which a fraction of postmenopausal women have increased rate of bone turnover leading to increased risk of fractures.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We thank O. Borel, A. Bourgeaud, C. Boulu-Chataigner, S. Jaisse, C. Valverde, and B. Vey-Marty for excellent technical assistance. We thank Dr. M.-C. Chapuy and S. Arnaud for measurements of serum intact PTH and 25(OH)D and Françoise Munoz for statistical advice. This work was supported by a contract INSERM-MSD-Chibret (OFELY Study).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Albright F, Smith PH, Richardson AM 1941 Postmenopausal osteoporosis—its clinical features. JAMA 116:24652473.
  • 2
    Lindsay R, Hart DM, Forrest C, Baird C 1980 Prevention of spinal osteoporosis in oophorectomized women. Lancet 2:11511154.
  • 3
    Slemenda C, Longcope C, Peacock M, Hui S, Johnston CC 1996 Sex steroids, bone mass, and bone loss. A prospective study of pre, peri- and postmenopausal women. J Clin Invest 97:1421.
  • 4
    Ettinger B, Pressman A, Sklarin P, Bauer D, Cauley JA, Cummings SR 1998 Associations between low levels of serum estradiol, bone density, and fractures among elderly women: The study of osteoporotic fractures. J Clin Endocrinol Metab 83:22392243.
  • 5
    Stone K, Bauer DC, Black DM, Sklarin P, Ensrud KE, Cummings SR 1998 Hormonal predictors of bone loss in elderly women: A prospective study. J Bone Miner Res 13:11671174.
  • 6
    Cummings SR, Browner WS, Bauer DB, Stone K, Ensrud K, Jamal S, Ettinger B 1998 Endogeneous hormones and the risk of hip and vertebral fractures among older women. N Engl J Med 339:733738.
  • 7
    Stepan JJ, Pospichal J, Presl J, Pacovsky V 1987 Bone loss and biochemical indices of bone remodeling in surgically induced postmenopausal women. Bone 8:279284.
  • 8
    Seibel M, Cosman F, Shen V, Gordon S, Dempster DW, Ratcliffe A, Lindsay R 1993 Urinary hydroxypyridinium crosslinks of collagen and markers of bone resorption and estrogen efficacy in postmenopausal women. J Bone Miner Res 8:881889.
  • 9
    Uebelhart D, Schlemmer A, Johansen J, Gineyts E, Christiansen C, Delmas PD 1991 Effect of menopause and hormone replacement therapy on the urinary excretion of pyridinium crosslinks. J Clin Endocrinol Metab 72:367373.
  • 10
    Garnero P, Sornay-Rendu E, Chapuy M-C, Delmas PD 1996 Increased bone turnover in late postmenopausal women is a major determinant of osteoporosis. J Bone Miner Res 11:337349.
  • 11
    Field C, Ory SJ, Wahner HW, Herrmann RR, Judd HL, Riggs BL 1993 Preventive effects of trandermal 17 beta estradiol on osteoporotic changes after surgical menopause: A two-year placebo-controlled trial. Am J Obstet Gynecol 168:114121.
  • 12
    Delmas PD, Pornel B, Felsenberg D, Garnero P, Hardy P, Pilate, Dain M-P 1999 A dose-ranging trial of a matrix transdermal 17 beta estradiol for the prevention of bone loss in early postmenopausal women. Bone 24:517523.
  • 13
    Prestwood KM, Pilbeam CC, Burleson JA, Woodiel FN, Delmas PD, Deftos LJ, Raisz LG 1994 The short-term effects of conjugated estrogen on bone turnover in older women. J Clin Endocrinol Metab 79:366371.
  • 14
    Garnero P, Hausherr E, Chapuy M-C, Marcelli C, Grandjean H, Muller C, Cormier C, Bréart G, Meunier PJ, Delmas PD 1996 Markers of bone resorption predict hip fracture risk in elderly women: The EPIDOS Prospective Study. J Bone Miner Res 11:15311538.
  • 15
    van Daele PLA, Seibel MJ, Burger H, Hofman A, Grobbee DE, van Leewen JPTM, Birkenhager JC, Pols HA 1996 Case-control analysis of bone-resorption markers, disability, and hip fracture: The Rotterdam study. BMJ 312:482483.
  • 16
    Parfitt AM, Mathews CHE, Villanueva AR, Kleerekoper M, Frame B, Rao DS 1983 Relationship between surface, volume and thickness of iliac trabecular bone in aging and in osteoporosis. J Clin Invest 72:1396.
  • 17
    Arlot ME, Sornay-Rendu E, Garnero P, Vey-Marty B, Delmas PD 1997 Apparent pre and postmenopausal bone loss evaluated by DXA at different skeletal sites in women: The Ofely cohort. J Bone Miner Res 12:683690.
  • 18
    National Osteoporosis Foundation Working Group on Vertebral Fractures 1995 Assessing vertebral fractures: A report by the National Osteoporosis Working Group on vertebral Fractures. J Bone Miner Res 10:518523.
  • 19
    Black DM, Palermo L, Nevitt MC, Genant HK, Christensen L, Cummings SR 1999 Defining incident vertebral deformity: A prospective comparison of several approaches. J Bone Miner Res 14:90101.
  • 20
    Garnero P, Grimaux M, Demiaux B, Préaudat C, Seguin P, Delmas PD 1992 Measurement of serum osteocalcin with a human-specific two-site radioimmunoassay. J Bone Miner Res 7:13891397.
  • 21
    Garnero P, Delmas PD 1993 Assessment of the serum levels of bone akaline phosphatase with a new immunoradiometric assay in patients with metabolic bone disease. J Clin Endocrinol Metab 77:10461053.
  • 22
    Melkko J, Kauppila S, Niemi S, Risteli L, Haukipuro K, Jukkola A, Risteli J 1996 Immunoassay for intact amino-terminal propeptide of human type I procollagen. Clin Chem 46:947954.
  • 23
    Robins SP, Woitge H, Hesley R, Ju J, Seyedin S, Seibel MJ 1994 Direct, enzyme-linked immunoassay for urinary deoxypyridinoline as a specific marker for measuring bone resorption. J Bone Miner Res 9:16431649.
  • 24
    Hanson DA, Weiss MAE, Bollen AM, Maslan SL, Singer FR, Eyre DR 1992 A specific immunoassay for monitoring human bone resorption: Quantitation of type I collagen cross-linked N-telopeptides in urine. J Bone Miner Res 7:12511258.
  • 25
    Bonde M, Qvist P, Fledelius C, Riis BJ, Christiansen C 1994 Immunoassay for quantifying type I collagen degradation products in urine evaluated. Clin Chem 40:20222025.
  • 26
    Rosenquist C, Fledelius C, Christgau S, Pedersen BJ, Bonde M, Qvist P, Christiansen C 1998 Serum Crosslaps one step Elisa. First application of monoclonal antibodies for measurement in serum of bone-related degradation products from C-terminal telopeptides of type I collagen. Clin Chem 44:22812289.
  • 27
    Bremond AG, Claustrat B, Rudigoz RC, Seffert P, Corniau J 1982 Estradiol, androstenedione, and dehydroepiandrosterone sulfate in the ovarian and peripheral blood of postmenopausal patients with and without endometrial cancer. Gynecol Oncol 14:119124.
  • 28
    Duda RJ, O' Brien JF, Katzmann JA, Peterson JM, Mann KG, Riggs BL 1988 Concurrent assays of circulating bone-gla-protein and bone alkaline phosphatase: Effects of sex, age, and metabolic bone disease. J Clin Endocrinol Metab 66:951957.
  • 29
    Delmas PD, Demiaux B, Malaval L, Chapuy M-C, Meunier PJ 1986 Serum bone-GLA-protein is not a sensitive marker of bone turnover in Paget's disease of bone. Calcif Tissue Int 38:6061.
  • 30
    Garnero P, Vassy V, Bertholin A, Riou JP, Delmas PD 1994 Markers of bone turnover in hyperthyroidism: Effects of treatment. J Clin Endocrinol Metab 78:955959.
  • 31
    Garnero P, Sornay-Rendu E, Duboeuf F, Delmas PD 1999 Markers of bone turnover predict forearm bone loss: A 4-yr follow-up study. J Bone Miner Res 14:16141621.
  • 32
    Riis BJ, Hansen MA, Jensen AM, Overgaard K, Christiansen C 1996 Low bone mass and fast rate of bone loss at menopause: Equal risk factors for future fracture: A 15 year follow-up study. Bone 19:912.
  • 33
    Garnero P, Gineyts E, Arbault P, Christiansen C, Delmas PD 1995 Different effects of bisphosphonate and estrogen therapy on free and peptide-bound bone crosslinks excretion. J Bone Miner Res 10:641649.
  • 34
    Randall AG, Kent GN, Garcia-Webb P, Bhagat CI, Pearce DJ, Gutteridge DH, Prince RL, Stewart G, Stuckey B, Will RK, Retallack RW, Price RI, Ward L 1996 Comparison of biochemical markers of bone turnover in Paget disease treated with pamidronate and a proposed model for the relationships between measurements of the different forms of pyridinoline crosslinks. J Bone Miner Res 11:11761184.
  • 35
    Apone S, Lee MY, Eyre DR 1997 Osteoclasts generate crosslinked collagen N-telopeptides (NTx) but not free pyridinolines when cultured on human bone. Bone 21:129136.
  • 36
    Colwell A, Eastell R 1996 The renal clearance of free and conjugated pyridinium cross-links of collagen. J Bone Miner Res 11:19761980.
  • 37
    Garnero P, Gineyts E, Schaffer AV, Seaman J, Delmas PD 1998 Measurement of urinary excretion of nonisomerized and β-isomerized forms of type I collagen breakdown products to monitor the effects of the bisphosphonate zoledronate in Paget's disease. Arthritis Rheum 41:354360.
  • 38
    Fledelius C, Johnsen AH, Closs PC, Bonde M, Qvist P 1997 Characterization of urinary degradation products derived from type I collagen. Identification of a beta-isomerized Asp-Gly sequence within the C-terminal telopeptide (alpha 1) region. J Biol Chem 272:97559763.
  • 39
    Gineyts E, Cloos PAC, Borel O, Grimaud L, Delmas PD, Garnero P 2000 Racemization and isomerization of type I collagen C-telopeptides in human bone and soft tissues: assessment of tissue turnover. Biochem J 345:481485.
  • 40
    Nordin BEC, Robertson A, Seamark RF, Bridges A, Philcox JC, Need AG, Horowitz M, Morris HA, Deam S 1985 The relation between calcium absorption, serum dehydroepiandrosterone, and vertebral mineral density in postmenopausal women. J Clin Endocrinol Metab 60:651656.
  • 41
    Davidson BJ, Riggs BL, Wahner HW, Judd HL 1983 Endogeneous cortisol and sex steroids in patients with osteoporotic spinal fractures. Obstet Gynecol 61:275.
  • 42
    Tomkinson A, Reeve J, Shaw RW, Noble BS 1997 The death of osteocytes via apoptosis accompanies estrogen withdrawal in human bone. J Clin Endocrinol Metab 82:128135.
  • 43
    Dunstan CR, Evans RA, Hills E, Wong SYP, Higgs RJED 1990 Bone death in hip fracture in the elderly. Calcif Tissue Int 47:270275.