The authors state that they have no conflicts of interest.
Differential Bone Metabolism Between Postmenopausal Women With Osteoarthritis and Osteoporosis†
Version of Record online: 26 NOV 2007
Copyright © 2008 ASBMR
Journal of Bone and Mineral Research
Volume 23, Issue 4, pages 475–483, April 2008
How to Cite
Jiang, L.-S., Zhang, Z.-M., Jiang, S.-D., Chen, W.-H. and Dai, L.-Y. (2008), Differential Bone Metabolism Between Postmenopausal Women With Osteoarthritis and Osteoporosis. J Bone Miner Res, 23: 475–483. doi: 10.1359/jbmr.071114
- Issue online: 4 DEC 2009
- Version of Record online: 26 NOV 2007
- Manuscript Accepted: 19 NOV 2007
- Manuscript Revised: 6 NOV 2007
- Manuscript Received: 7 JUL 2007
- bone metabolism;
A comparative study of bone metabolism between postmenopausal women with osteoarthritis and osteoporosis showed that differential levels of bone remodeling markers, leptin, free leptin index, and osteoprotegerin might partly contribute to the proposed inverse relationship in bone mass between postmenopausal women with osteoarthritis and osteoporosis.
Introduction: Osteoarthritis (OA) and osteoporosis (OP) are two common disorders affecting the quality of life in the elderly. The association between OA and OP has always been debated. The objective of this study was to compare bone metabolism between postmenopausal women with OA and OP.
Materials and Methods: A total of 120 postmenopausal women with OA and OP (n = 60, respectively) were included in this comparative study. Anthropometric parameters and BMD at the spine and the proximal femur were measured. Serum leptin, soluble leptin receptor (sLR), osteoprotegerin (OPG), and bone remodeling markers, including bone-specific alkaline phosphatase (BALP), osteocalcin (OC), deoxypyridinoline cross-links (DPD), and cross-linked N-telopeptides of type I collagen (NTX), were quantified with commercial ELISA or EIA kits. Free leptin index (FLI) was also calculated by the ratio between serum leptin and sLR levels.
Results: Postmenopausal women with OA had higher body weight, body mass index, fat mass, and percentage of fat than those suffered from OP. Compared with the patients in OP group, the patients in OA group had significantly higher BMD values at all sites measured. Higher serum leptin and FLI and lower OPG levels were shown in the OA group (leptin: 31.22 ± 6.4 versus 26.50 ± 9.27 ng/ml, p < 0.001; FLI: 3.20 ± 1.02 versus 2.50 ± 0.95, p < 0.05; OPG: 4.75 ± 1.97 versus 6.96 ± 2.75 pM, p < 0.001), whereas lower serum OC and higher urine DPD were noted in the OP group (OC: 16.45 ± 8.45 versus 13.06 ± 6.25 ng/ml, p < 0.05; DPD: 10.83 ± 7.12 versus 15.29 ± 6.65 nM BCE/mM Cr, p < 0.001). Serum OPG levels negatively correlated with BMD at all sites assessed. However, no correlation was found between leptin and BMD. Only in the OA group di positive correlations exist between FLI and Z-score at the femoral neck and Ward's triangle region. After stepwise regression analysis, it was found that differential factors were able to predict the variance of BMD at different sites to a certain extent.
Conclusions: Our study suggests that there are significant differences in bone metabolism between postmenopausal women with OA and OP and provides evidence for the inverse relationship between OA and OP. Differential levels of bone remodeling markers, leptin, FLI, and OPG may partly contribute to the proposed inverse relationship. Roles of leptin and its soluble receptor in bone metabolism regulation should be explored further.
Both osteoarthritis (OA) and osteoporosis (OP) are major public health problems affecting the quality of life in the elderly population. Although OA is a disease of articular cartilage degeneration, the subchondral bone has been suggested to play an important role in the pathogenesis of OA. The latest conception of OP is characterized by low bone mass, microarchitectural deterioration of bone tissue leading to enhanced bone fragility, and a consequent increase in fracture risk. Imbalance between bone formation and resorption, the coupled processes during bone metabolism, is considered to bring about the outcome of OP. Early studies showed that the femoral head excised from the patient with OP showed no progressive changes of coxarthritis, whereas bone loss was less likely to occur in the proximal femur taken from patients with OA. Subsequent studies showed that patients with OA had higher BMD at both axial and peripheral sites than both normal subjects and patients with OP.[5-11] When OA and OP cases are compared with normal controls matched for age, they are obviously different in anthropometry and geometry, biomechanics, biochemistry, and turnover of bone.[12-15] Also, these two diseases rarely coexist in the same patient.[3, 4, 15, 16] The clinical and research evidence suggest that there be an inverse relationship between OA and OP, although conflict and controversy exist.
The mechanism of the inverse relationship between OA and OP remains unclear. Studies of bone metabolism in patients with OA or OP will not only contribute to the understanding of the pathogenesis of these two age-related diseases, but can also provide biochemical evidence for the supposed relationship between OA and OP. Additionally, it has been reported that the incidence of OA after 50 yr of age increased more rapidly in women than in men.[18, 19] However, comparative study with regard to bone turnover between OA and OP is scarce, especially for age-matched postmenopausal women.
Anthropometric data have shown that postmenopausal women with OA are heavier and have more fat, even when controlled for age and body size. Conversely, most postmenopausal women with OP look thinner and smaller.[17, 20] Thus, obesity has been proposed to participate in the mechanism of these two diseases.[21, 22] Leptin, the soluble 16-kDa protein product of the obesity gene, is reported to regulate bone metabolism through a local or systemic pathway. Some researchers proposed that leptin decreased bone resorption through inhibiting the activity of osteoclasts and promoted bone formation by acting on marrow stromal cells to enhance their differentiation into osteoblasts.[23-25] On the contrary, with persuasive evidence from serial animal model–based experiments, others found that both the leptin-deficient (ob/ob) and -resistant (db/db) mice had a profound increase in bone mass, and intraventricular injection of leptin in both ob/ob and wildtype mice resulted in bone loss.[26, 27] Therefore, they concluded that leptin could inhibit bone formation through the hypothalamus pathway. Meanwhile, the soluble leptin receptor (sLR) and free leptin index (FLI) have been suggested to play an important role in the physiological function of leptin. It is essential to explain the exact role of leptin and leptin-related factors in the regulation of bone metabolism among postmenopausal population with OA and OP. As well, osteoprotegerin (OPG), a member of the TNF receptor family, has been confirmed to have positive effects on bone mass, especially in estrogen deficiency conditions.[29, 30] However, no study on the comparison of circulating OPG levels between OA and OP in postmenopausal women has been addressed yet.
Therefore, in this study, we attempted to explore the possible difference in the level of biochemical markers between postmenopausal women with OA and OP. The associations of leptin, leptin-related factors, and OPG with the regulation of bone mass were also studied. We hypothesized that differences in bone metabolism were associated with differences in phenotypes and that differences in the level of biochemical markers would partially explain the proposed inverse relationship between OA and OP.
MATERIALS AND METHODS
One hundred twenty postmenopausal women were selected for this study. Sixty patients were diagnosed with primary OA of the hip or knee according to the criteria of the American College of Rheumatology.[31, 32] Patients with any evidence of osteoporosis as shown on radiographs were excluded from the OA group using the Singh index in the proximal femur and the grading scale of spinal osteoporosis in the spine on plain radiographs as the exclusion criteria. Another 60 postmenopausal women with OP were defined according to the latest criteria of osteoporosis proposed by the World Health Organization (WHO). All patients in the OP group had one or more osteoporotic fracture histories at different body sites. In both groups, we excluded patients with diseases that might affect bone metabolism, such as hyperparathyroidism, hyperthyroidism, osteomalacia, renal dysfunction and diabetes mellitus, or drug supplementation.
All patients were informed and provided consent to participate in this research work. This study was approved by the Ethics Committee of the authors’ institution.
Body height and weight were measured, and body mass index (BMI) was calculated at the time of initial examination. The percentage of fat (%fat) and fat mass (FM) were measured by a body composition analyzer TBF-410 (Tanita Co., Tokyo, Japan). The percentage coefficients of the intra-analyzer variations (%CVs) for body composition were <2% in our institution.
Both lumbar spine (L2–L4) and the proximal femur, including the femoral neck, the greater trochanter, and Ward's triangle, were chosen as measurement sites for BMD. BMD was measured by DXA (Challenger; DMS, Pérols, France). For patients with osteoporotic fractures of the proximal femur, the unaffected side was used for measurements of BMD at the proximal femoral regions, whereas the affected side was measured for OA patients. The %CVs of the inter- and intra-analyzer variations for BMD were <3% at each site in our institution.
Blood and urine biochemical analyses
Blood samples were obtained from the median cubital vein at 8:00 a.m. in the morning after an overnight fasting. After at least 30 min, blood samples were centrifuged at 3000 rpm for 10 min at room temperature. The separated sera were stored at −80°C until analyzed. A second morning urine sample was obtained and stored at −80°C until analyzed. If the urine color was not clear before assay, centrifugation at the above-mentioned velocity was performed to remove sediment.
Serum leptin concentration was assayed by an ELISA kit (DSL-10–23100; Diagnostic System, Webster, TX, USA) and sLR by an EIA kit (Quantikine; R&D Systems, Minneapolis, MN, USA). The interassay CVs of leptin ranged from 3.3% to 5.3% and sLR from 5.3% to 8.6%. The intra-assay CVs of leptin ranged from 1.5% to 6.2% and sLR from 2.2% to 6.1%. FLI was calculated by the ratio between serum leptin and sLR levels.
Serum OPG level was measured by an ELISA kit (BI-20402; Biomedica Medizinprodukte, Wien, Austria). The minimum detection limit was 0.14 pmol/ml. The interassay CVs ranged from 7% to 8% and the intra-assay CVs from 4% to 10%.
Serum bone-specific alkaline phosphatase (BALP) and osteocalcin (OC) were measured as markers of bone formation. BALP was assessed using an EIA kit (Metra; Quidel, San Diego, CA, USA) and OC by an ELISA kit (DSL-10–7600; Diagnostic System). The interassay CVs of BALP ranged from 5.0% to 7.6% and OC from 3.7% to 8.0%. The intra-assay CVs of BALP ranged from 3.9% to 5.8% and OC from 4.6% to 6.4%.
Urine deoxypyridinoline cross-links (DPD), and cross-linked N-telopeptides of type I collagen (NTX) were measured as markers of bone resorption. Both of them were determined by EIA kits (DPD: Metra; Quidel; NTX: Osteomark; Wampole Laboratories, Princeton, NJ, USA). The interassay CVs of DPD ranged from 3.1% to 4.8% and NTX from 3% to 5%. The intra-assay CVs of DPD ranged from 4.3% to 8.4% and NTX from 5% to 8%.
All values were expressed as mean ± SD. The corresponding data from anthropometric, BMD, and biochemical markers measurements between OA and OP groups were compared statistically using the Mann-Whitney U test.
To determine the potential association of leptin, leptin-related factors, and OPG with other indices, such as BMD, anthropometric parameters, and bone remodeling markers, bivariate correlation analysis was used first to calculate Pearson's correlation coefficients. To avoid a mixed effect from age, partial correlation analysis was performed after adjustment of years since menopause.
Multiple linear regression analysis was used to estimate which parameter was the most important index to predict BMD. BMD at the lumbar spine and the proximal femoral regions, including the femoral neck, the greater trochanter, and Ward's triangle, were chosen as the dependent variable separately. A stepwise method was used to calculate the value of R2.
An SPSS 10.0 software package (SPSS, Chicago, IL, USA) was used for all statistical procedures. p < 0.05 was considered significant.
Anthropometric and biochemical measurements are presented in Table 1. The Mann-Whitney U test showed no significant difference between OA and OP patients in age, years since menopause, and body height. Patients with OA had higher weight, BMI, %fat, and FM than those with OP.
Higher serum leptin, FLI, and lower OPG levels were shown in the OA group, whereas lower serum OC and higher urine DPD were noted in the OP group. However, no significant difference was found in the levels of serum sLR between the OA and OP groups.
As shown in Fig. 1, at all body sites measured, including the lumbar spine, the femoral neck, the greater trochanter, and Ward's triangle, patients with OA had significantly higher BMD than those with OP (BMDLum: 0.79 ± 0.12 versus 0.72 ± 0.04 g/cm2, p < 0.001; BMDFN: 0.98 ± 0.13 versus 0.77 ± 0.06 g/cm2, p < 0.001; BMDGT: 0.90 ± 0.12 versus 0.66 ± 0.06 g/cm2, p < 0.001; BMDWT: 0.84 ± 0.15 versus 0.59 ± 0.07 g/cm2, p < 0.001). All T- and Z-scores were significantly lower in the OP group, except Z-score at the lumbar spine (OA versus OP: −0.30 ± 0.74 versus −0.30 ± 0.25, p = 0.983).
When all the patients were considered as a whole group, a positive correlation was found between leptin and FLI. Serum leptin levels and FLI both correlated positively with weight, BMI, %fat, and FM. A negative correlation was shown between serum leptin and urine DPD, between FLI and urine DPD, and between FLI and sLR. Also, serum leptin and FLI both positively correlated with BMD at the proximal femoral regions. After adjustment of age and years since menopause, the correlations between leptin and all the other above-mentioned parameters except body weight were no longer significant. However, all the correlations with regard to FLI still existed with statistical significance (Table 2). Serum OPG levels negatively correlated with BMI, FM, and BMD at all body sites measured (Table 3). Urine NTX positively correlated with serum BALP (r = 0.3779, p < 0.001) and OC levels (r = 0.3284, p = 0.001). Meanwhile, the level of serum BALP also positively correlated with serum OC concentration (r = 0.3131, p = 0.002). On stepwise regression analysis, years since menopause was the only significant factor that could predict BMD at the lumbar spine, which explained 0.4% of the variance in bone mineralization (r2 = 0.1944, p < 0.001). As for BMD at the femoral neck, age, BMI, and DPD contributed to 0.6%, 1.9%, and 0.4% of the variance in BMD (r2 = 0.5157, p < 0.001), respectively. These three independent variables also significantly predicted 0.7%, 2.5%, and 0.6% in BMD at the Ward's triangle region (r2 = 0.5268, p < 0.001), respectively. The %fat also significantly predicted 1.1% BMD at the greater trochanter (r2 = 0.5830, p < 0.001), in addition to the variables of years since menopause (0.5%), BMI (4.3%), and DPD (0.5%).
In the OA group, leptin levels and FLI positively correlated with each other (r = 0.5675, p < 0.001) and negatively correlated with sLR (r = −0.8098, p < 0.001). Also, FLI negatively correlated with serum BALP (r = −0.4085, p = 0.020) and urine DPD levels (r = −0.5539, p = 0.001). Meanwhile, there were positive correlations between FLI and Z-score at the femoral neck and Ward's triangle (FN: r = 0.4052, p = 0.021; WT: r = 0.3644, p = 0.040; Fig. 2). A positive correlation was observed between urine NTX and serum BALP (r = 0.3962, p = 0.025) and between serum BALP and OC levels (r = 0.3811, p = 0.031). All these relationships referred to above still maintained statistical significance after adjustment of age and years since menopause. No correlation was shown between serum OPG levels and the other parameters. Results from a stepwise regression model showed that years since menopause was the only independent variable that could predict 0.7%, 0.9%, 1.0%, and 0.6% of the variance in BMD at the lumbar spine, at the femoral neck, at Ward's triangle, and at the greater trochanter regions (BMDLum: r2 = 0.2010, p = 0.010; BMDFN: r2 = 0.3028, p = 0.001; BMDWT: r2 = 0.2910, p = 0.001; BMDGT: r2 = 0.1666, p = 0.020), respectively.
In the OP group, leptin levels positively correlated with weight and FLI, even after adjustment of years since menopause. FLI negatively correlated with sLR and positively correlated with body weight, BMI, FM, %fat, and BMD at Ward's triangle region. After adjustment of age and years since menopause, positive correlations between FLI and body weight, between FLI and %fat, and between FLI and BMD at Ward's triangle remained significant. Meanwhile, serum OPG level negatively correlated with BMD at the femoral neck and at the greater trochanter. Although the correlations between OPG levels and BMD at the lumbar spine and Ward's triangle were not significant, a negative trend could be noted (Fig. 3). Positive correlations between urine NTX and serum BALP (r = 0.3589, p = 0.004), between urine NTX and serum OC (r = 0.4156, P <0.001), and between serum BALP and OC levels (r = 0.3642, p = 0.003) also existed. On stepwise regression, years since menopause was the only significant factor that could predict BMD at the lumbar spine, which explained 0.1% of the variance in bone mineralization (r2 = 0.0727, p = 0.033). At the femoral neck, FM and years since menopause contributed to 0.7% and 0.2% of the variance in BMD (r2 = 0.2236, p < 0.001), respectively. Meanwhile %fat and years since menopause explained 0.8% and 0.2% of the variance in BMD at Ward's triangle region (r2 = 0.2820, p < 0.001), respectively. At the same time, years since menopause and BMI contributed to 0.2% and 1.6% of variance in BMD at the greater trochanter, respectively (r2 = 0.1947, p = 0.002).
A number of research papers have been published over the past years, debating the relationship between OA and OP.[3-5, 7, 13, 16] To our knowledge, there are few studies to compare the characteristics of bone metabolism between postmenopausal women with OA and OP based on the results of anthropometry, densitometry, and biochemistry. The main findings of our study were that serum leptin level and FLI were higher in OA patients, whereas serum OPG level was higher in OP patients. Instead of leptin, FLI seems to be more reasonable to explain the association between obesity and bone mineralization. Moreover, the higher serum OC and the lower urine DPD concentration in OA group implied that differential expression of biochemical markers of bone turnover should be related to the outcome in BMD between these two populations.
Our data indicated that postmenopausal women with OA had significantly higher BMD at all sites assessed by DXA. Hideaway et al. pointed out lower reproducibility and technical difficulty associated with DXA. The presence of osteophyte formation or aortic calcification could affect BMD measurement at local sites in patients with OA. However, Orwoll et al. found that the impact of osteophyte on BMD measurement mainly focused on the spine. Proximal femoral density was not different between those with and without osteophytes. Thus, we chose both the lumbar spine and the proximal femur region as the measurement sites assessed by DXA. Genetic, environmental, and hormonal factors have been suggested to contribute to the mechanism of bone mineralization.[20, 38-41] In postmenopausal population, acceleration of bone turnover caused by estrogen deficiency was confirmed to be one of the most important causes of bone loss.[42-45] However, in clinical settings, postmenopausal women with OA rarely suffer from OP.[3-5, 13, 16] In this study, we found that urine DPD, one of the bone resorption markers, was significantly higher in patients with OP, whereas serum OC, the marker of bone formation, was a little higher in patients with OA. More bone resorption and less bone formation is doomed to cause more bone loss in osteoporotic postmenopausal women. Stewart et al. thought that elderly women with OA possessed not only higher BMD, but also higher bone turnover. We believe that relatively lower bone turnover and higher BMD in OA population might be a more reasonable explanation. Differential status of bone turnover contributes to the difference in BMD between OA and OP.
Evidence from anthropometric studies showed that osteoporotic patients are smaller and thinner, whereas patients suffered from OA are heavier and have more fat. Thus, researchers believed that obesity should have a protective effect against OP.[21, 22] Increased weight bearing and aromatization of androgen to estrogen in adipose tissue has been proposed to be responsible for this bone-sparing effect.[46, 47] Our results also shown that higher body weight, BMI, fat mass, and percent of fat, were noted in OA group. Furthermore, according to the results of multiple linear regression analyses, BMI, FM, or %fat, could positively predict the variance of BMD more or less at different regions of the proximal femur in postmenopausal women. Some authors considered that lean mass positively correlated with bone mineralization status.[48, 49] However, Reid et al. believed that the effect of body weight on bone tissue should be contributed by both fat mass and lean mass, and the role of FM on BMD is consistently more important than that of the latter.
Leptin, known as the protein product of obesity gene, was found to regulate bone metabolism through local or systemic pathway.[23-27] With gene knocked-out animal models, Ducy and colleagues[26, 27] confirmed that leptin reduce bone mass through a central hypothalamic pathway. On the contrary, Thomas and colleagues[23, 24] showed that leptin prevented bone loss by acting on human marrow stromal cells to enhance their differentiation into osteoblasts and modulating bone resorption through OPG/RANKL pathway. Our data showed that there were significant differences in serum leptin level and FLI between OA and OP groups. Consistent with results from other clinical studies,[51, 52] we also attempted but failed to find a relationship between serum leptin levels and BMD. It is thought that the obese should have higher level of leptin than the nonobese. We noted that the mean BMI value did not reach the criteria of obesity (BMI > 30 kg/m2), even in patients with OA in our series (Table 1). Meanwhile, geographic or racial difference could influence the level of systemic leptin. However, it was reported that the ratio between leptin and sLR, namely FLI, determined the biologically active form of leptin.[54, 55] We also showed that FLI positively correlated with BMD at different sites measured to a certain extent in this study. These findings implied that FLI should be more reasonable to explain the potential association between obesity and bone mineralization in postmenopausal population.
In this study, a negative correlation was observed between serum OPG levels and BMD at all body sites measured. This kind of association turned to be insignificant when only the patients with OA were considered. OPG, a soluble decoy receptor to prevent the binding of RANKL to RANK, inhibits osteoclast differentiation and activation.[29, 56] Also, it regulates bone resorption by influencing the apoptosis of osteoclast. Misra et al. thought that OPG negatively correlated with lumbar bone density and markers of nutritional status, including BMI, %fat, and leptin. This has been explained to be a compensatory response to the bone loss. Higher mean OPG value noted in our OP group also supported this assumption in postmenopausal population.
In addition, sLR has been thought to determine the level of serum leptin with bioactivity. van Dielen et al. suggested that sLR level was significantly lower in morbidly obese subjects than that in lean individuals. Inconsistent with most studies about sLR,[28, 58, 59] we only identified a negative association between serum leptin, as well as FLI, and sLR level in OA group. When all the patients were considered, or just in OP group, only FLI negatively correlated with serum sLR levels. Inadequacy of the patients in our series to meet the criteria of obesity might be the reason for this inconsistent result. The limitation of our study may lie in our incapability to establish a “completely normal” group of population without OA and OP as a control. Certainly, an aging normal population as a control group will enhance the efficiency of comparative study. However, on the one hand, it is impossible to find age-matched postmenopausal women without any diseases, and on the other hand no animal model with generalized OA, up to now, has been developed for study yet. Thus, an alternative new experiment design should be set up to validate our conclusions.
In summary, we observed comparable differences in the level of biochemical markers between postmenopausal women with OA and OP. Differential levels of bone remodeling markers and OPG partially contributed to the proposed inverse relationship in bone mass between OA and OP. FLI seems to be more suitable to explain the association between obesity and bone mineralization. Roles of leptin and its soluble receptor in bone metabolism regulation should be explored further.
- 121996 The relationship between osteoarthritis and osteoporosis in the hip. J Orthop Rheumatol 9: 214–216.
- 311986 Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthrit Rheum 29: 1039–1049., , , , , , , , , , , , , , , , , , , , , , , ,
- 34World Health Organization Study Group 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. World Health Organ Tech Rep Ser 843: 1–129.
- 421998 A unitary model for involutional osteoporosis: Estrogen deficiency causes both type 1 and type 2 osteoporosis in postmenopausal women and contributes to bone loss in aging men. J Bone Miner Res 13: 763–773., ,