Menopause contributes to significant risk for postmenopausal women in developing osteoporosis (23
) and atherosclerotic CVD (24
). Many risk factors including elevated Hcy (7
) and CRP are more likely to emerge after rather than prior to menopause. To our knowledge, this is the only study examining the relationship between total Hcy and CRP versus trabecular bone assessed via pQCT in healthy postmenopausal women. A few studies have related high total Hcy to low BMD (2
). However, no studies have examined the relationship between total Hcy and trabecular bone.
Contrary to previous findings (2
), we did not demonstrate a relationship between circulating total Hcy and areal bone via DXA. Although we hypothesized that total Hcy would be inversely associated with trabecular bone, correlation analysis did not reveal a significant association between total Hcy and trabecular BMC (r=−0.041, p
=0.58) or BMD (r=−0.060, p
=0.42). Nonetheless, once other factors were taken into account in the regression model, the presence of an inverse relationship (albeit not significant) suggested the potential role of Hcy in osteoporotic risk. The vast majority of women in our study had normal total Hcy, whereas previous epidemiological studies have demonstrated an increased risk of osteoporotic fracture only in subjects with high Hcy (>15 μmol/L). For example, in the Hordaland Homocysteine Study (3
), the multiple adjusted odds ratio for low BMD was 1.96 in those with high (≥15 μmol/L) compared with low (<9 μmol/L) Hcy. Similarly, the Framingham Study (4
) showed that subjects in the highest quartile (18.6 ± 6.4 μmol/L) had a greater risk of hip fracture than those in the lowest quartile (7.6 ± 1.0 μmol/L) of total Hcy. Gerdhem et al. (2
) reported a decrease in femoral neck BMD in elderly women in the highest Hcy quartile (17.4–109.0 μmol/L) compared to those in the lower three quartiles (6.2–17.3 μmol/L). However, this relationship was not observed for lumbar spine BMD. Plasma total Hcy concentration was slightly higher (7.7 ± 2.2 μmol/L) in our subjects than that recently reported for women (6.8 ± 0.1 μmol/L) from the 2001–2002 NHANES data (6
), but was comparable to the lowest quartile of total Hcy for women (7.6 ± 1.0 μmol/L) in the Framingham Study (3
). Only four women in our study had high (≥15 μmol/L) plasma total Hcy, possibly explaining why we did not detect a relationship between Hcy and bone in these healthy postmenopausal women.
Nonetheless, we identified not only known but also less recognized factors that contributed to the variability in trabecular bone. Hemoglobin emerged as a significant contributor to trabecular bone. This may be because trabecular bone is highly vascularized, well innervated, and responsible for hematopoiesis and mineral homeostasis. Subjects with a higher hemoglobin concentration may have better oxygen delivery to all tissues, including bone, thereby increasing trabecular bone. It is important to note that hemoglobin is fairly stable in postmenopausal women who no longer have monthly blood losses. Reporting similar results, Cesari et al. (26
) concluded that hemoglobin was significantly (positive direction) associated with total (at the calf; p
=0.03), as well as both trabecular (p
=0.02) and cortical (p
=0.03) bone density, using pQCT in older women. One may speculate that a low hemoglobin concentration due to iron deficiency anemia may impact bone by reducing trabecular bone density. To illustrate, the lumbar vertebrae of female rats either on an iron restricted (12 mg Fe/kg/day) but calcium adequate (5.2 g Ca/kg diet) diet (27
) or an iron deficient diet (< 8 mg Fe/kg/day; [28
]) had a lower trabecular number and greater trabecular separation using micro-computed tomography. The activity of prolyl hydroxylase, essential to hydroxyproline cross-linking in bone, is iron-dependent (29
), providing a possible explanation for the role of iron in collagen and hence bone. Perhaps this is one explanation as to why we noted the consistent association of hemoglobin with trabecular bone in these women, suggesting that preventing anemia may be important for reducing risk of osteoporosis. Although serum ferritin is a better indicator of iron status than hemoglobin, surprisingly, we did not find any relationship between serum ferritin and trabecular bone (data not shown) in this sample of healthy postmenopausal women.
Although women in our study had normal kidney function, we observed a negative association between circulating uric acid and trabecular BMC, perhaps illustrating the relationship of kidney function to bone. Pietschmann et al. (30
) reported that patients with absorptive hypercalciuria and lower vertebral BMD exhibited a trend toward increased uric acid excretion. In our study, we did not measure urinary uric acid, but rather serum uric acid. The relationship between serum uric acid and BMD should be further investigated. Additionally, weight contributed to trabecular BMC, similar to what has been reported in previous studies (31
A few studies have demonstrated an association between CRP concentration and BMD or biochemical markers of bone turnover in immune (32
) and inflammatory (33
) disease states. However, there is a paucity of data examining this relation in healthy postmenopausal subjects without inflammatory conditions, similar to our subjects. Our median CRP value was less than what was reported by Bassuk et al. (10
) for American women not on hormone therapy. Contrary to our hypothesis, CRP was not associated with distal tibia trabecular BMC or BMD either with correlation or multiple regression analyses.
The median CRP concentration (1.1 mg/L) in our study was similar to that reported by Koh et al. (22
) in Korean women (1.0 mg/L), but was slightly lower than the concentration (1.5 mg/L) in apparently healthy American women (35
). Our results indicating no correlation between CRP and trabecular bone do not agree with Koh et al. (22
), who demonstrated that higher circulating CRP was associated with lower femoral neck BMD in postmenopausal Korean women. However, women from their study were late postmenopausal (7.7 ± 5.7 years) compared to our women (3.5 ± 2.0 years since menopause) and we focused on trabecular bone. More importantly, osteopenic and osteoporotic women in their study had a significantly greater (p
≤0.014) CRP concentration compared with healthy postmenopausal women. Since we excluded women with a BMD T-score below -1.5, it is possible that we were unable to detect a relationship between CRP and bone because our women were more uniform with respect to BMD. To further explore this relationship, we examined women with CRP values above and below 1.8 mg/L, based on the association of CRP above this value with osteopenia and/or osteoporosis as noted in the Korean study (22
). In our women with elevated CRP, we found no association between CRP and any trabecular bone measurement. Contrary to results from Koh et al. (22
), high CRP (>1.8 mg/L) did not appear to confer elevated risk of osteoporosis in this group of healthy postmenopausal women.
In conclusion, total Hcy and CRP were not significantly related to trabecular bone in our healthy, non-osteoporotic postmenopausal women. Yet, once various factors were taken into account, our regression results suggested a weak (albeit not significant) but negative relationship between Hcy and trabecular bone. However, the majority of our participants had normal Hcy and CRP concentrations, with only a few moderately elevated values, suggesting that low or normal values do not impact trabecular bone in healthy postmenopausal women. We acknowledge that peripheral measurements of trabecular bone using pQCT do not perfectly reflect the axial skeleton, but this technology is nevertheless quite useful. Further investigation is needed to examine the relationship of hyperhomocysteinemia and elevated CRP to trabecular bone loss in early postmenopausal women and the response of trabecular bone to dietary intervention.