Overview, Vol 14, Issue 2

Only doubt is certain and disbelief worth believing.
Without this courage there can be no learning.
Believe nothing.

"The quarterly journal Progress in Osteoporosis began in October 1993 as Advances in Osteoporosis. Its purpose was to provide readers without easy access to the literature with summaries of the most important literature. We now inhabit a virtual world. Information is instantaneously accessible to all at the tap of a keyboard; understanding is not. In the spirit captured by the anonymous author*, the purpose of this publication is to provide critical evaluation of the most important literature and so to provoke discussion. It is our intention to promote dialogue which examines the quality of information published and so its credibility. The opinions expressed are my own and do not necessarily reflect those of the International Osteoporosis Foundation."

We invite readers to comment on and discuss this journal entry at the bottom of the page.

PDF Version

FREE DOWNLOAD of PIO Volume 14-2 Figures


Denosumab is the most powerful remodeling suppressant available. Remodeling removes old or damaged bone and so it is beneficial provided that each remodeling transaction carried out by each basic multicellular unit (BMU) removes and replaces the same volume of bone. However, when BMU balance becomes negative, as occurs around midlife, each remodeling event produces structural decay. So, it can be argued that any remodeling is bad for the structure of bone. However, the price that may be paid for remodeling suppression is the potential for compromising the material composition of bone, especially in persons with low baseline remodeling who may assemble a bone with a mean bone matrix mineral density at the upper part of the normal distribution for this trait.

In the presence of little or no remodeling, microdamage is unlikely to be removed. As more and more of the unremodeled matrix undergoes secondary mineralization, a process that takes months to years to reach completion, adjacent osteons become similarly and homogeneously mineralized. The increased homogeneity, i.e., reduced heterogeneity, in matrix mineralization density and increase pentosidine crosslinking of collagen are likely to reduced ductility of bone matrix – its ability to deform without cracking. This leads to microdamage as the matrix becomes stiff so that when a load is imposed, instead of matrix deforming, changing length and absorbing energy in that way, the energy is released as a microcrack. If these accumulate, or one major crack progresses in length, catastrophic structural failure may occur.

The question is what is worse; to allow some remodeling and sacrifice structure while maintaining material composition of the slowly decaying bone, or to suppress remodeling as much as possible and preserve structure but allow deterioration in material composition. There is an answer to this question. There needs to be a measure pretreatment material composition and structure and monitoring of the relative effects of treatment on the material composition, structure and whole bone strength. From this, it may be possible to administer and titrate therapies accordingly.

Four papers published recently address the effects of denosumab on bone histomorphetry, structure, comparisons with bisphosphonates and the effects on intrauterine development. Brown et al evaluated 41 subjects (13 crossover from placebo and 28 long-term denosumab groups representing up to 5 years denosumab treatment) (1). The mean (SD) duration from the last denosumab dose to the first dose of tetracycline was 5.7 (0.5) months. Qualitative histology showed normal mineralized lamellar bone. Bone remodeling decreased as reflected by eroded surface in crossover and long-term subjects. 11/13 (85%) crossover subjects and 20/28 (71%) long-term subjects had double or single tetracycline label in trabecular and/or cortical compartments; specimens from 5 crossover subjects and 10 long-term subjects were evaluable for dynamic trabecular bone parameters; remodeling indices were low, consistent with reduced remodeling. Information on the material composition of bone and its consequences in the longer term on bone strength were not studied. I find this disappointing because we need information regarding the material composition of bone with longer term denosumab therapy given its very efficient remodeling suppressant effects, we need to know if there is a risk of increased tissue mineralization and glycosylation and so, an increased risk for atypical fractures.

Brown et al report that postmenopausal women ≥55 years suboptimally adherent to bisphosphonate (BP) were randomized to denosumab 60 mg subcutaneously every 6 months (N=852) or oral 150 mg monthly ibandronate or risedronate (N=851) for 12 months (2). Denosumab produced greater gains in BMD at 12 months than monthly (p<0.0001 all). In higher risk subjects, denosumab led to greater gains in BMD than oral BPs at the total hip (2.2 vs. 0.8 %), femoral neck (1.8 vs. 0.3 %), and lumbar spine (3.7 vs. 1.4 %). Denosumab also led to greater decreases in sCTX-1 in the whole group and higher risk subjects (all p<0.0001). There are no surprises here given how powerfully denosumab reduces remodeling. What is of interest is where is the remodeling continuing in the skeleton. It is likely that this is intracortical remodeling that continues with the BPs. This needs formal study but has been reported. Zebaze et al report greater suppression of remodeling and a greater reduction in porosity with denosumab than alendronate (3). If denosumab suppresses intracortical remodeling more than alendronate, could this drug have greater antifracture efficacy than alendronate? We don’t know. No comparator studies have been done, but it is clearly one of the great challenges for the new leaders in the field. A 20% reduction in nonvertebral fracture efficacy reported in most trials is not good enough.

Poole et al used 3D cortical bone mapping of the proximal femur in 80 postmenopausal women with osteoporosis to determine the timing and location of effects of denosumab (4). Cortical 3D bone thickness and surface density maps of both hips were created from CT scans at baseline, 12, 24 and 36 months. After registration of each bone to an average femur shape model followed by statistical parametric mapping to identify differences between denosumab and control. Denosumab increased cortical mass surface density and thickness by 12 months and by 5.4% over 3 years, and by up to 12% relative to placebo at locations such as the lateral femoral trochanter. The authors report one third of the increase came from increasing cortical density, and two thirds from increasing cortical thickness relative to placebo.

Figure 1. Summary of the steps involved in 3D cortical bone mapping and analysis of changes over time using statistical parametric mapping (SPM). (1) Triangulated 3D surface of a left proximal femur (one individual) and corresponding coronal CT plane after segmentation with Stradwin software. (2) An axial CT slice to demonstrate sampling the CT data along an 18mm cyan trans-cortical line at the surface location indicated by the red circular dot. (3) Parametric fit (red curve) and original CT data (cyan curve), superimposed on a graph of distance (x axis) vs. Hounsfield units (y axis) along the 18 mm transcortical line. The cortical thickness is the distance between the red arrows. (4) Many thousands of cortical thickness estimates painted by reference to a colour scale onto the femoral surface of same individual. (5) Registration (morphing) of the individual’s femoral surface shape to the canonical femur shape, CFS. (6) Resultant map displaying cortical thickness for the individual now on the CFS after left and right femurs are averaged to give a single map per participant at each visit. (7) Subtracting the baseline cortical thickness map from the 36 month map in an individual to create a difference map. (8) After unblinding, there are two groups; cortical thickness difference maps from the placebo (upper), and denosumab groups (lower). (9) Typical results from one arm of the study (in this case, the denosumab group) using the statistical parametric mapping process to find and quantify where the 36 month maps differ from baseline maps. The left image shows a yellow p-value map displayed on the CFS. The yellow foci are vertices where cortical thickness changed significantly between baseline and 36 months, ranging from grey (non-significant) via orange to increasingly bright yellow as the p value decreases. The right image shows the amount of change as a colour scale on the CFS. In this hypothetical case we see increasing thickness with denosumab. (10) Final result showing the magnitude of increase in cortical thickness on the CFS with only areas of statistically significantly greater cortical thickness retained, the rest greyed out. Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2325 with permission of the American Society of Bone and Mineral Research.

Figure 2. Percentage changes in overall* cortical mass surface density (left), overall* cortical thickness (middle) and overall* cortical density (right) during 3 years treatment with denosumab or placebo. These results are averaged across the entire femoral cortical surface. *By ‘overall’, we denote the mean value for all femurs, each averaged over the entire canonical femur surface; i.e., the average of approximately 6000 points per femur from 78 femurs (every femur). The values are shown for each of 4 visits (baseline, 12 months, 24 months, 36 months). Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2325 with permission of the American Society of Bone and Mineral Research.

As discussed by the authors, it is possible that cortical thickening is produced by infilling of porosity with subendocortical cortex. It is difficult to explain this by any other means. Antiresorptives are not anabolic, they do not increase bone mass by depositing bone upon the endocortical surface to thicken the cortex. At best, they allow partial refilling of endocortical resorption cavities excavated before treatment – these excavated sites partly refill while newly excavated sites are virtually abolished, so at best, if a hemiosteon is dug on the endocortical surface it may be about 100 microns in depth and then refills to about 97% of its starting value – this cannot build bone unless the resorption cavity dug was reduced by denosumab; this is possible but remains unproven. Another mechanism of cortical thickening may be continued bone modeling that becomes detectable because it is no longer obscured by high remodeling which is now suppressed (5).

Boyce et al report subcutaneous 50 mg/kg/month denosumab was given to pregnant cynomolgus monkeys from gestation day (GD) 20 to parturition (6). For up to 6 months postpartum (birth day [BD] 180/181), remodeling markers decreased at birth day. Spontaneous long bone fractures were detected in 4 denosumab-exposed infants at BD28 and BD60. BD1 infants exposed to denosumab in utero had decreased bone length; increased radio-opacity of the axial and appendicular skeleton and base of the skull with decreased marrow cavities, widened growth plates, flared metaphyses, altered jaw/skull shape, reduced jaw length; delayed secondary ossification centers, increased trabecular BMD, decreased cortical BMD. Bones with endochondral ossification consisted of dense primary spongiosa with reduced marrow space. Retained woven bone was observed in bones formed by intramembranous ossification. Reductions in toughness at the femur diaphysis and lack of correlation between energy and bone mass at the vertebra were observed. Tooth eruption was unimpaired, the reduced growth and increased bone density of the mandible resulted in tooth malalignment and dental dysplasia. Radiographic changes at BD1 persisted at BD28, with evidence of resumption of resorption and remodeling in most infants at BD60 and/or BD90. There was recovery from bone-related changes in infants necropsied at BD181 where exposure to denosumab had been below limits of quantitation for 3 months. The phenotype resembles infants with osteoclast-poor osteopetrosis due to inactivating mutations of RANK or RANKL.

Cathepsin K Inhibition

Cathepsin K (CatK) inhibition prevents bone resorption by preventing collagen degradation (7). There are many fascinating aspects concerning the mechanisms of action of this class of drug. The most consistent and credible effect is the reduction in the size of the resorption cavities excavated. Antiresorptives like the BPs and denosumab probably do this too but it is well documented experimentally with CatK inhibitors and other models perturbing collagen degradation (8).

After this, the data becomes challenging to interpret. BPs and denosumab reduce the rate of bone remodeling – they reduce the number of resorption sites appearing per unit bone volume per unit time and this reduction in activation frequency is clearly identifiable in the reduction in circulating remodeling markers. CatK inhibition also reduces remodeling rate but this appears to be site-, surface- and species-specific. The trabecular and intracortical surface extent of remodeling, as measured by the mineralizing surface per unit bone surface is reduced but not necessarily upon the endocortical surface, at least in monkeys (9).

If remodeling removes a larger volume of bone than is subsequently deposited upon the endocortical surface, this will erode and thin the cortex. If remodeling is slowed, as it appears to be upon intracortical and trabecular surfaces, then structural decay will still occur at these locations but more slowly and will produce cortical porosity and trabecular thinning respectively, provided there is a negative BMU balance.

If the depth of resorption produced by each BMU is lessened using a CatK inhibitor, and the volume of bone deposited is now equal to the volume resorbed, then there will be no net bone loss despite continued remodeling. If the volume of bone deposited exceeds the volume resorbed then this may produce a positive BMU balance. Under these circumstances, high remodeling rate is an advantage because each remodeling event will add bone. This desirable scenario remains unproven experimentally.

Several studies have been published recently, but the data need to be examined carefully. Ochi et al report that vehicle, ONO-5334 (3, 10 or 30 mg/kg) or alendronate (0.5 mg/kg) were administered for 8 months to sham and OVX monkeys (10). Alendronate prevented OVX-induced increase in remodeling rate, but it did not appear to reduce remodeling markers; this is surprising as virtually all studies in animals and humans demonstrate a convincing and easily visualized reduction in remodeling with alendronate. ONO-5334 at 30 mg/kg did not suppress remodeling, again this is quite different to studies done in human subjects. In this study, remodeling upon periosteal, osteonal and endocortical surfaces continued as determined by histomorphometry. If remodeling continued, then bone structure should deteriorate if BMU balance is negative. However, the higher dose maintained urinary CTX near zero and kept serum osteocalcin around the level of the sham animals. So, the drug prevented a rise in remodeling markers, but it did not prevent remodeling at the surface of bone. ONO-5334 reversed the effect of OVX on vertebral BMD with improvement in strength. Both ONO-5334 and alendronate prevented OVX-induced changes in vertebral microstructure. Femoral neck pQCT showed that ONO-5334 increased total and cortical BMD and strength. I don’t understand this work.

Figure 3. Changes in urinary CTX (A) and serum osteocalcin (B) in ovariectomized cynomolgus monkeys. Monkeys orally received ONO-5334 (0.3, 3 and 30 mg/kg), alendronate (0.5 mg/kg) or vehicle for 8 months. Serum was collected 4 h after administration on each sampling day for determination of osteocalcin level. 24-h cumulated urine was collected on each sampling day for determination of CTX level. Data are presented as % baseline (mean+SE, n=6–8). #, ##: p<0.05, 0.01 vs. sham group, respectively (Student t test). **: p<0.01 vs. OVX group (Dunnett test). $: p<0.05 vs. OVX group (Student t test). Reproduced from Bone, 65:1-8, Copyright (2014), with permission from Elsevier.





Figure 4. Bone mechanical strength, maximum load (A), stiffness (B) and energy absorption (C) in the 4th lumbar vertebra. Monkeys orally received ONO-5334 (0.3, 3 and 30 mg/kg), alendronate (0.5 mg/kg) or vehicle for 8 months. Compression test of the isolated 4th lumbar vertebra was conducted after necropsy. Data are presented as mean+SE (n=6-8). #, ##: p<0.05, 0.01 vs. sham group, respectively (Student t test). *, **: p<0.05, 0.01 vs. OVX group, respectively (Dunnett test). $: p<0.05 vs. OVX group (Student t test). Reproduced from Bone, 65:1-8, Copyright (2014), with permission from Elsevier.


Engelke et al report a double-blind, placebo-controlled, 2-year trial involving postmenopausal women randomized to odanacatib (ODN) 50 mg weekly or placebo; hip QCT scans were available for 158 women (11). The ODN minus placebo effects were significant for total hip integral (5.4%), trabecular (12.2%) and cortical vBMD (2.5%), not integral bone volume. A small but statistically significant increase in cortical volume (1.0-1.3%) and thickness (1.4-1.9%) was reported. The proportions of total bone mineral content (BMC) gain attributed to cortical gain ranged from 40-52% depending on the location. ODN improved integral, trabecular and cortical vBMD and BMC relative to placebo. Cortical volume and thickness increased in all regions except the femoral neck. The increase in cortical volume and BMC paralleled the increase in cortical vBMD, demonstrating a consistent effect of ODN on cortical bone. These data are difficult to interpret. One possibility is cortical volume within the periosteal and endocortical surfaces increased, but there was no report of either an increase in periosteal perimeter or a decrease in endocortical perimeter. The other possibility is that there is infilling of pores with matrix which then mineralizes. A third possibility is that there is no increase in matrix volume at all, but the existing matrix undergoes more complete secondary mineralization producing this increase in cortical vBMD. Whether these methods have the resolution to accurately quantify small changes in volume and to distinguish these alternatives is not clear.

Figure 5. Total hip least squares (LS) mean percentage changes of bone mineral density over 24 month in the ODN and PBO groups. Left: aBMD measured by DXA; right: integral vBMD measured by QCT. Error bars indicate standard error. OW: once weekly. Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2292 with permission of the American Society of Bone and Mineral Research.

Figure 6. Least squares (LS) mean percentage change from baseline (BL) after 24 months for cortical, subcortical and trabecular vBMD. ODN-PBO differences were significant for all VOIs and compartments shown in the graph. The numbers indicate percentage change of the corresponding parameter versus baseline. Error bars indicate standard error. Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2292 with permission of the American Society of Bone and Mineral Research.

Cheung et al randomized 214 postmenopausal women (mean age 64.0±6.8 years and baseline lumbar spine T-score -1.81±0.83) to oral ODN 50 mg or placebo, weekly for 2 years (12). Increases from baseline in total vBMD occurred at the distal radius and tibia. At both sites, differences from placebo were also found in trabecular vBMD, cortical vBMD, cortical thickness, cortical area, and strength estimated using finite element analysis (treatment differences at radius and tibia = 2.64% and 2.66%). At the distal radius, ODN improved trabecular thickness and bone volume/total volume (BV/TV) vs. placebo. At a more proximal radial site, ODN attenuated the increase in cortical porosity found with placebo (treatment difference = -7.7%, p=0.066). At the distal tibia, odanacatib improved trabecular number, separation, and BV/TV vs. placebo.

Figure 7. Two-year percent changes from baseline in other HR-pQCT parameters at (A) radius and (B) tibia. The distal radius and tibia (to the left of the vertical line) were scanned for trabecular number, trabecular separation, trabecular thickness, trabecular bone volume/total volume (BV/TV), cortical area, and cortical porosity. The more proximal regions (to the right of the vertical line) were scanned for assessment of cortical porosity. LS: least squares. Reproduced from J Bone Miner Res 2014;29:1786-94 with permission of the American Society of Bone and Mineral Research.







Figure 8. Percent changes from baseline in estimated strength at (A) the distal radius and (B) the distal tibia, based on finite element analysis of HR-pQCT scans. LS: least squares. Reproduced from J Bone Miner Res 2014;29:1786-94 with permission of the American Society of Bone and Mineral Research.











Pennypacker et al report that adult OVX rhesus monkeys were treated with vehicle or ODN (6 or 30 mg/kg, once per day [q.d., p.o.]) for 21 months. Calcein and tetracycline double-labeling were given at 12 and 21 months, and the femoral cross-sections were subjected to dynamic histomorphometric and cement line analyses (13). ODN increased periosteal and endocortical bone formation (BFR/BS) with increased endocortical mineralizing surface (102%, p<0.01) with the 6 mg/kg dose. Both doses reduced remodeling hemiosteon numbers by 51% and 66% (p<0.05), respectively, and ODN 30 mg/kg numerically reduced activation frequency without affecting wall thickness. On the same endocortical surface, ODN increased all modeling-based parameters, while reducing intracortical remodeling, consistent with the observed no treatment effects on cortical porosity. ODN 30 mg/kg increased cortical thickness (CtTh, p<0.001), reduced marrow area (p<0.01) and increased femoral structural strength (p<0.001). Peak load correlated with the increases in BMC (r2=0.9057, p<0.0001) and CtTh (r2=0.6866, p<0.0001). The authors claim that treatment reduced cortical remodeling and stimulating modeling-based bone formation, improved cortical dimension and strength in OVX monkeys. These are interesting observations, but they have not been reproduced or reported yet in the bone biopsy data presented at ASBMR last week (see next issue of PIO).

Discontinuing BPs after 5 years is regarded as a viable option to avert adverse events of long term remodeling suppression. However, remodeling rate eventually increases and will erode the skeleton if there is a negative remodeling balance. Bauer et al tested methods of predicting fracture risk among women discontinuing alendronate after 4-5 years in the prospective Fracture Intervention Trial Long-term Extension (FLEX) study, which randomized women aged 61-86 years previously treated with 4-5 years of alendronate to 5 more years of alendronate or placebo (14). The analysis included only the placebo group. Hip and spine DXA were measured when placebo was begun (FLEX baseline) and after 1-3 years of follow-up. Urinary type I collagen crosslinked N-telopeptide (NTX) and serum bone-specific alkaline phosphatase (BAP), were measured at FLEX baseline and after 1 and 3 years. During 5 years of placebo, 94 of 437 women (22%) had 1 or more symptomatic fractures; 82 had fractures after 1 year. One-year changes in hip DXA, NTX, and BAP were not related to fracture, but older age and lower hip DXA at time of discontinuation were related to increased fracture risk (lowest tertile of baseline femoral neck DXA vs. other 2 tertiles relative hazard ratio: 2.17 [1.38-3.41]; total hip DXA relative hazard ratio: 1.87 [1.20-2.92]). The authors infer that age and hip BMD at discontinuation predict clinical fractures during the next 5 years.

The authors’ inference from this work is “Follow-up measurements of DXA 1 year after discontinuation and of BAP or NTX 1-2 years after discontinuation are not associated with fracture risk and cannot be recommended.” Women with BMD<-3.5 SD, those with BMD below the baseline in the fracture intervention (FIT) trial, those losing bone rapidly, and 12 women sustaining fractures before repeat measurements of BMD and bone remodeling markers, were excluded. Bone remodeling markers were measured in only ~90 women, and 83 women discontinuing the placebo “took bone-active medications”. It is possible that that these features truncated the extremes of these traits producing limited power to detect an association with fracture.

Three questions leave uncertainties as to ‘how long’ to treat. First, does 10 years treatment confer continued antifracture efficacy? There was no randomized untreated control group in the FLEX trial, nor in the 10-year follow-up of the study by Bone et al (15), so this question cannot be answered. Second, if treatment is stopped, do fracture rates increase? It is implied that fractures in the second 5 years of FLEX when placebo is given is due to stopping alendronate, but that inference depends on whether covariates were equally distributed at rerandomization of the 1099 women into placebo or one of two alendronate doses, and whether retention and compliance during the further 5 years were high; ~20% of participants discontinued the allotted treatment (16). Third, is any benefit from 5 years treatment sustained? It is also implied (but not stated) that the absence of fractures during the second 5 years is, in part, due to prior alendronate, but this inference requires demonstration of a lower fracture rate than in a control group. So, the data in FLEX trial is difficult to interpret.

Thus, even though the evidence base is limited, stopping treatment or poor compliance increase remodeling rate, bone loss, and fractures (17). We know this. More ‘typical’ fractures result from stopping than atypical fractures produced by continuing treatment (18). We know this too. Treatment guidelines pay no attention to remodeling rate, material composition or structure. Perhaps there are factors (low remodeling, high tissue mineralization density, high collagen crosslinking, severe microstructural decay) that signal risk for atypical femoral fractures and identify those at risk. We won’t know until we look.

Jaw and PTH

Kim et al report that in 24 cases of osteonecrosis of the jaw associated with BP use (BRONJ), 15 subjects were assigned to 20 µg teripartide (TPTD) for 6 months and 9 were assigned to the none (19). All continued calcium and vitamin D. While 60% of the non-TPTD group showed one stage of improvement in BRONJ, 40.0% did not show any improvement. In the TPTD group, 62.5% showed one stage of improvement and 37.5% demonstrated a marked improvement, including two stages of improvement or complete healing. All TPTD cases improved. Patients with higher baseline serum 25(OH)D levels showed better clinical therapeutic outcomes with TPTD.

Vitamin D May Not Increase Calcium Absorption

Gallagher et al report that 198 white and African American women aged 25-45 years with serum 25(OH)D <20 ng/mL were randomized double-blind to vitamin D3 400, 800, 1600, 2400 IU, or placebo plus calcium supplement (20). Calcium absorption was measured at baseline and 12 months using radiocalcium-45 and 100 mg of calcium. Mean baseline serum 25OHD was 13.4 ng/mL (33.5 nmol/L) and increased to 40 ng/mL (100 nmol/L) with 2400 IU without an increase in calcium absorption. There was no relationship between 12-month calcium absorption and final serum 25(OH)D. There was no evidence of a threshold suggesting that active transport of calcium is saturated at 25(OH)D levels <5 ng/mL. There is no need to recommend vitamin D for increasing calcium absorption in normal subjects.

Milk Supplements

Sahni et al report 830 men and women completed a food frequency questionnaire in 1988-89. Energy adjusted intakes of milk, yogurt, cheese, cream and milk+yogurt (servings/wk) were calculated (21). The mean age at baseline was 77 y (68-96). 97 hip fractures occurred over 11.6 y (0.04-21.9). OR for medium (>1 and <7 servings/wk) or higher (≥7 servings/wk) milk intake vs. low intake (≤1 serving/wk) intake was 0.58 (0.31-1.06), P=0.078. OR for medium vs. low intake: 0.61 (0.36-1.08), P=0.071; P trend: 0.178]. Hip fracture risk was 40% lower among those with medium/high milk intake, compared to those with low intake (P=0.061). A similar threshold was observed for milk+yogurt intake (P=0.104). Greater intakes of milk and milk+yogurt may lower risk for hip fracture. While of borderline significance, this study does support the possibility that calcium supplementation by dietary means may reduce the risk of hip fracture. It is a case-control study and so it is hypothesis generating, not hypothesis testing – what is needed are randomized controlled trials with both fracture outcomes and cardiac outcomes measured.

Radavelli-Bagatini et al evaluated the association between dairy food and bone in 564 elderly women aged 80-92 y (mean 84.7) who attended the 10-year follow-up to tertiles of dairy intake: first tertile (≤1.5 servings/d), second tertile (1.5-2.2 servings/d) and third tertile (≥2.2 servings/d) (22). pQCT at the 15% tibia showed that compared with those in the first tertile of dairy intake, women in the third tertile had 5.7% greater total bone mass (p=0.005), principally because of higher cortical and subcortical bone mass (5.9%, p=0.050), resulting in a 6.2% higher total vBMD (p=0.013). Trabecular, but not cortical and subcortical, vBMD was also higher (7.8%, p=0.044). DXA assessment showed that women in the third tertile of dairy intake had greater appendicular bone mass (7.1%, p=0.007) and skeletal muscle mass (3.3%, p=0.014) compared with tertile 1. The associations with bone measures were dependent on dairy protein and calcium intakes, whereas the association with appendicular muscle mass was not totally dependent on dairy protein intake.

Figure 9. pQCT assessments at 15% tibia according to dairy intake. (A) Total bone mass. (B) Total vBMD. (C) Cortical and subcortical mass. (D) Trabecular vBMD. Values are expected mean and SEM. Means without a common letter differ, p<0.05, multivariate adjusted for age, BMI, physical activity, alcohol consumption, smoking habit, calcium supplementation during the intervention phase, current calcium supplementation, current vitamin D supplementation, and current osteoporosis medication (ANCOVA with Bonferroni post hoc test). Reproduced from J Bone Miner Res 2014;29:1691-700 with permission of the American Society of Bone and Mineral Research.

Figure 10. Appendicular bone and muscle mass according to dairy intake. (A) Appendicular bone mass. (B) Appendicular bone area. (C) Appendicular skeletal muscle mass. Values are expected mean and SEM. Means without a common letter differ, p<0.05; A and B were multivariate adjusted for age, BMI, physical activity, alcohol consumption, smoking habit, calcium supplementation during the intervention phase, current calcium supplementation, current vitamin D supplementation, and current osteoporosis medication; C was adjusted for age, BMI, physical activity, smoking habit, calcium supplementation during the intervention phase, and current vitamin D supplementation (ANCOVA with Bonferroni post hoc test). Reproduced from J Bone Miner Res 2014;29:1691-700 with permission of the American Society of Bone and Mineral Research.





Calcium Supplements During Growth

Zhang et al randomly assigned to 40 g of milk powder with 300 mg (Low-Ca group), 600 mg (Mid-Ca group) or 900 mg of calcium (High-Ca group) for 2 years in 111 girls and 109 boys (aged 12-14 y) enrolled, 91 girls and 91 boys completed the trial (23). The girls in the High-Ca group (1110 mg/d) had 2.3%, 2.7% and 2.6% greater BMD accretion at the total hip, femoral neck and shaft (P<0.05) than those in the Low-Ca group (655 mg/d). A significant effect of higher intake was also observed for percentage change of size-adjusted BMC at femur neck (P=0.047). No differences in the percentage changes in BMD, BMC or size-adjusted BMC between the Mid- and Low-Ca groups and between the High- and Mid-Ca groups. Extra calcium had no observable additional effect in the boys.

Calcium Supplements and Cardiovascular Risk
Much ado about nothing?

Lewis et al report an ancillary study of 1103 women, 75.2±2.7 y, assessed common carotid artery intimal medial thickness (CCA-IMT) and carotid atherosclerosis at year 3 of CAIFOS (24). By intent to treat, women randomized to calcium had no higher mean CCA-IMT (calcium 0.778±0.006 mm, placebo 0.783±0.006 mm, p=0.491) and maximum CCA-IMT (calcium 0.921±0.007 mm, placebo 0.929±0.006 mm, p=0.404). Women randomized to calcium did not have increased carotid atherosclerosis (calcium 47.2%, placebo 52.7%, p=0.066). Women taking >80% supplements had reduced carotid atherosclerosis in unadjusted but not in multivariate adjusted models (p=0.033 and p=0.064, respectively). Participants in the highest tertile of total calcium (diet and supplements) had reduced carotid atherosclerosis in unadjusted and multivariable adjusted analyses compared with participants in the lowest tertile (OR=0.67, 0.50-0.90, p=0.008, and OR=0.70, 0.51-0.96, p=0.028, respectively). These findings do not support that calcium supplementation increases carotid artery intimal medial thickness or carotid atherosclerosis, and high calcium intake may reduce this surrogate cardiovascular risk factor.

Paik et al found no independent associations between supplemental calcium intake and risk of incident coronary heart disease (CHD) and stroke in a prospective cohort study of 74,245 women in the Nurses' Health Study with 24 years of follow-up (25). During 24 years of follow-up, 4565 cardiovascular events occurred (2709 CHD and 1856 strokes). At baseline, women who took calcium supplements had higher levels of physical activity, smoked less, and had lower trans fat intake compared with those who did not take calcium supplements. After multivariable adjustment, the relative risk of cardiovascular disease for women taking >1000 mg/d of calcium supplements compared with none was 0.82 (95% CI 0.74-0.92; p for trend <0.001), the multivariable adjusted relative risk for CHD was 0.71 (0.61-0.83; p for trend <0.001) and for stroke was 1.03 (0.87-1.21; p for trend = 0.61).

Lewis et al undertook a meta-analysis of randomized controlled trials with placebo or no-treatment control groups to determine if these supplements increase myocardial infarction (MI), angina pectoris and acute coronary syndrome, and chronic CHD. 18 studies met the inclusion criteria and contributed 63,563 participants with 3390 CHD events and 4157 deaths (26). Five trials contributed CHD events with pooled relative RR of 1.02 (0.96-1.09; P=0.51). Seventeen trials contributed all-cause mortality data with pooled RR of 0.96 (0.91-1.02; P=0.18). The RR for MI was 1.08 (0.92-1.26; P=0.32), angina pectoris and acute coronary syndrome 1.09 (0.95-1.24; P=0.22) and chronic CHD 0.92 (0.73-1.15; P=0.46). The authors infer that current evidence does not support the hypothesis that calcium supplementation with or without vitamin D increase coronary heart disease or all-cause mortality risk in elderly women. However, a glance at Figure 11 shows that the duration of most of the studies was 1-3 years. If calcium supplementation has an adverse effect, unless acute, it is unlikely to be detected in brief studies such as this. How many lung cancers will be observed in smokers of 1 pack of cigarettes per year?

Figure 11. Random effects estimates of calcium supplementation with or without vitamin D for a) myocardial infarction, b) angina pectoris and acute coronary syndrome and c) chronic coronary heart disease compared to no calcium. For Grant et al (2005), events were reported in those who received calcium cf. placebo (Ca) and calcium plus vitamin D cf. vitamin D only (CaD). M-H: Mantel-Haenszel, this method estimates the amount of between-study variation by comparing each study’s result with a Mantel-Haenszel fixed-effect meta-analysis result. Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2311 with permission of the American Society of Bone and Mineral Research.


Figure 12. Sensitivity analyses based on type of supplementation. *Post hoc subgroup analysis of the Women’s Health Initiative (WHI) in participants with no personal supplements at baseline (NPS) using the trial investigators internal dataset (28). M-H: Mantel-Haenszel, this method estimates the amount of between-study variation by comparing each study’s result with a Mantel-Haenszel fixed-effect meta-analysis result. Reproduced from J Bone Miner Res 2014;doi:10.1002/jbmr.2311 with permission of the American Society of Bone and Mineral Research.

Sclerostin Deficiency and Stronger Bone

The prospect of a new anabolic agent in antisclerostin antibodies requires demonstration that the new bone will be healthy bone. Hassler et al studied cortical bone of Sost-knockout (KO) mice (n=9, 16-week-old) and sclerosteosis patients (young [4-14 y], n=4 and adults [24 and 43 y], n=2) (27). In Sost-KO mice, endocortical bone had reduced matrix mineralization -1.9%, p<0.0001 (younger tissue age) and -1.5%, p<0.05 (older tissue age), and relative proteoglycan content was increased. Bone matrix mineralization density distribution was shifted towards lower matrix mineralization in samples of compact bone of sclerosteosis patients associated with increased mineralization heterogeneity in the young population. The relative proteoglycan content was increased. The altered bone composition may contribute to increased strength.

Sclerostin Antibody

Roschger et al assessed once-weekly intravenous Sost-ab injections for 4 weeks in male Col1a1Jrt/+mice, a model of severe dominant osteogenesis imperfecta (OI), starting either at 4 weeks (growing mice) or at 20 weeks (adult mice) of age. Sost-ab had no effect on weight or femur length (28). In OI mice, no treatment-associated differences in remodeling markers. µCT analyses at the femur showed that Sost-ab was associated with higher trabecular bone volume and higher cortical thickness in wildtype mice at both ages and in growing OI mice, but not in adult OI mice. Three-point bending tests of the femur showed that in wildtype but not in OI mice, Sost-ab was associated with higher ultimate load and work to failure. Quantitative backscattered electron imaging of the femur did not show any effect of Sost-ab on CaPeak regardless of genotype, age or measurement location. Thus, Sost-ab had a larger effect in wildtype than in Col1a1Jrt/+mice. The data suggest that Sost-ab is less effective in severe OI mouse model.

Ross et al report that in rats and nonhuman primates treated with vehicle or sclerostin antibody (Scl-Ab), despite up to 54% increases in the bone volume after Scl-Ab, the mean global mineralization of trabecular and cortical bone was unaffected (29). However, in BMDD in the primate trabecular bone had an increase in the number of pixels with a low mineralization value and a decrease in the standard deviation of the distribution. Tissue age-specific measurements in the primate model showed that Scl-Ab did not affect the mineral-to-matrix ratio, crystallinity, or collagen crosslinking in the endocortical, intracortical, or trabecular compartments. Scl-Ab was associated with a nonsignificant trend toward accelerated mineralization intracortically and a nearly 10% increase in carbonate substitution for tissue older than 2 weeks in the trabecular compartment (p<0.001). Scl-Ab does not negatively impact matrix quality.

Bone formation may be remodeling-based (RBF) or modeling-based (MBF), the former coupled to bone resorption and the latter occurring directly on quiescent surfaces. Scl-Ab increases bone formation while decreasing bone resorption. Ominsky et al tested the hypothesis that Scl-Ab produces a modeling based anabolic response by examining bones from OVX rats and male cynomolgus monkeys (cynos) (30). Histomorphometry was performed to quantify and characterize bone surfaces in OVX rats administered vehicle or Scl-Ab (25 mg/kg) subcutaneously (sc) twice/week for 5 weeks and in adolescent cynos administered vehicle or Scl-Ab (30 mg/kg) sc every 2 weeks for 10 weeks. Fluorochrome-labeled surfaces in L2 vertebra and femur endocortex (cynos only) were considered to be MBF or RBF based on characteristics of their associated cement lines. In OVX rats, Scl-Ab increased MBF by 8-fold (from 7% to 63% of bone surface, compared to vehicle). In cynos, Scl-Ab increased MBF on trabecular (from 0.6% to 34%) and endocortical surfaces (from 7% to 77%) relative to vehicle. Scl-Ab did not affect RBF in rats or cynos despite decreased resorption surface in both species. In cynos, Scl-Ab resulted in a greater proportion of RBF and MBF containing sequential labels from week 2, indicating an increase in the lifespan of the formative site. This extended formation period was associated with robust increases in the percent of new bone volume formed. Scl-Ab increased bone volume by increasing MBF and prolonged the formation period at both modeling and remodeling sites while reducing bone resorption.

Figure 13. Effects of Scl-Ab on modeling bone formation on trabecular surfaces in OVX rats. Histomorphometry was performed on LV2 from OVX rats treated twice weekly with sc Vehicle or Scl-Ab (25 mg/kg) for 5 weeks. (A) For each group, images are shown of the whole sagittal section stained with Goldner's trichrome (top panel) and with magnified epifluorescent images (middle panel) showing the fluorochrome labels administered on day 0 (xylenol orange, red), days 22–23 (calcein, green), and days 32–33 (tetracycline, orange). Red arrows indicate the locations of faint xylenol orange labels. The corresponding polarized light micrographs (bottom panel) show the collagen orientation, reflecting changes in the cement lines used for assessment of modeling (white dashed lines) and remodeling (orange dotted lines), and showing the lamellar architecture of the newly formed bone. (B) Bone surfaces were characterized as modeling-based formation (MBF), remodeling-based formation (RBF), quiescent (QS), or osteoclastic (OcS), and are expressed as a percentage of the total surface. Trabecular bone volume/total volume (BV/TV) was also measured. Data are expressed as mean±SEM; *p<0.05 vs. vehicle by two-tailed t test. Reproduced from J Bone Miner Res 2014;29:1424-30 with permission of the American Society of Bone and Mineral Research.

Figure 14. Effects of Scl-Ab on modeling bone formation on trabecular surfaces in cynomolgus monkeys. Histomorphometry was performed on LV2 from male cynos treated sc every 2 weeks with vehicle or Scl-Ab (30 mg/kg) for 10 weeks. (A) For each group, epifluorescent images (top panel) reflect a pair of tetracycline labels (orange) injected on days 14 and 24 and calcein labels injected on days 56 and 66 (green). The corresponding polarized light micrographs (bottom panel) show the collagen orientation, reflecting changes in the cement lines used for assessment of modeling (white dashed lines) and remodeling (orange dotted lines), and showing the lamellar architecture of the newly formed bone. (B) Bone surfaces were characterized as modeling-based formation (MBF), remodeling-based formation (RBF), quiescent (QS), or eroded (ES), and are expressed as a percentage of the total surface. Trabecular bone volume/total volume (BV/TV) was also measured. Data are expressed as mean±SEM; *p<0.05 vs. vehicle by two-tailed t test. Reproduced from J Bone Miner Res 2014;29:1424-30 with permission of the American Society of Bone and Mineral Research.


1. Brown JP, Reid IR, Wagman RB, et al. Effects of up to 5 years of denosumab treatment on bone histology and histomorphometry: The FREEDOM Study Extension. J Bone Miner Res 2014;29:2051.

2. Brown JP, Roux C, Ho PR, et al. Denosumab significantly increases bone mineral density and reduces bone turnover compared with monthly oral ibandronate and risedronate in postmenopausal women who remained at higher risk for fracture despite previous suboptimal treatment with an oral bisphosphonate. Osteoporos Int 2014;25:1953.

3. Zebaze R, Ghasem-Zadeh A, Bohte A, et al. Intracortical remodeling and porosity: Rational targets for fracture prevention. Lancet 2010;375:1729.

4. Poole KES, Treece GM, Gee AH, et al. Denosumab rapidly increases cortical bone in key locations of the femur: A 3D bone mapping study in women with osteoporosis. JBMR 2014;doi:10.1002/jbmr.2325.

5. Ominsky M, Libanati C, Boyce RW, et al. Continuous modeling-based bone formation: a novel mechanism that could explain the sustained increases in hip bone mineral density (BMD) with denosumab treatment. J Bone Miner Res 2013;28 (Suppl 1). Available at http://www.asbmr.org/education/AbstractDetail?aid=89fd1f4a-ff73-4a62-acc.... Accessed September 26, 2014.

6. Boyce RW, Varela A, Chouinard L, et al. Infant cynomolgus monkeys exposed to denosumab in utero exhibit an osteoclast-poor osteopetrotic-like skeletal phenotype at birth and in the early postnatal period. Bone 2014;64:314.

7. Delaissé JM, Eeckhout Y, Vaes G. In vivo and in vitro evidence for the involvement of cysteine proteinases in bone resorption. Biochem Biophys Res Commun 1984; 14;125:441.

8. Rodan SB, Duong LT. Cathepsin K – A new molecular target for osteoporosis. IBMS BoneKEy 2008;5:16.

9. Cusick T, Chen CM, Pennypacker BL, et al. Odanacatib treatment increases hip bone mass and cortical thickness by preserving endocortical bone formation and stimulating periosteal bone formation in the ovariectomized adult rhesus monkey. J Bone Miner Res 2012;27:524.

10. Ochi Y, Yamada H, Mori H, et al. Effects of eight-month treatment with ONO-5334, a cathepsin K inhibitor, on bone metabolism, strength and microstructure in ovariectomized cynomolgus monkeys. Bone 2014;65:1.

11. Engelke K, Fuerst T, Dardzinski B, et al. Odanacatib treatment affects trabecular and cortical bone in the femur of postmenopausal women – Results of a 2-year [lacebo-controlled trial. J Bone Miner Res 2014;doi:10.1002/jbmr.2292.

12. Cheung AM, Majumdar S, Brixen K, et al. Effects of odanacatib on the radius and tibia of postmenopausal women: Improvements in bone geometry, microarchitecture, and estimated bone strength. J Bone Miner Res 2014;29:1786.

13. Pennypacker BL, Chen CM, Zheng H, et al. Inhibition of cathepsin K increases modeling-based bone formation, and improves cortical dimension and strength in adult ovariectomized monkeys. J Bone Miner Res 2014;29:1847.

14. Bauer DC, Schwartz A, Palermo L, et al. Fracture prediction after discontinuation of 4 to 5 years of alendronate therapy: The FLEX study. JAMA Intern Med 2014;174:1126.

15. Bone HG, Hosking D, Devogelaer JP, et al. Ten years’ experience with alendronate for osteoporosis in postmenopausal women. N Engl J Med. 2004;350:1189.

16. Black DM, Schwartz AV, Ensrud KE, et al.; FLEX Research Group. Effects of continuing or stopping alendronate after 5 years of treatment: The Fracture Intervention Trial Long-term Extension (FLEX): A randomized trial. JAMA 2006;296:2927.

17. Olsen KR, Hansen C, Abrahamsen B. Association between refill compliance to oral bisphosphonate treatment, incident fractures, and health care costs ‒ An analysis using national health databases. Osteoporos Int 2013;24:2639.

18. Shane E, Burr D, Abrahamsen B, et al. Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force of the American Society for Bone and Mineral Research. J Bone Miner Res 2014;29:1.

19. Kim KM, Park W, Oh SY, et al. Distinctive role of 6-month teriparatide treatment on intractable bisphosphonate-related osteonecrosis of the jaw. Osteoporos Int 2014;25:1625.

20. Gallagher JC, Jindal PS, Smith LM. Vitamin D does not increase calcium absorption in young women: a randomized clinical trial. J Bone Miner Res 2014;29:1081.

21. Sahni S, Mangano KM, Tucker KL, et al. Protective association of milk intake on the risk of hip fracture: Results from the Framingham Original Cohort. J Bone Miner Res 2014;29:1756.

22. Radavelli-Bagatini S, Zhu K, Lewis JR, Prince RL. Dairy food intake, peripheral bone structure, and muscle mass in elderly ambulatory women. J Bone Miner Res 2014;29:1691.

23. Zhang ZQ, Ma XM, Huang ZW, et al. Effects of milk salt supplementation on bone mineral gain in pubertal Chinese adolescents: A 2-year randomized, double-blind, controlled, dose-response trial. Bone 2014;65:69.

24. Lewis JR, Zhu K, Thompson PL, Prince RL. The effects of 3 years of calcium supplementation on common carotid artery intimal medial thickness and carotid atherosclerosis in older women: An ancillary study of the CAIFOS randomized controlled trial. J Bone Miner Res 2014;29:534.

25. Paik JM, Curhan GC, Sun Q, et al. Calcium supplement intake and risk of cardiovascular disease in women. Osteoporos Int 2014;25:2047.

26. Lewis JR, Radavelli-Bagatini S, Rejnmark L, et al. The effects of calcium supplementation on verified coronary heart disease hospitalization and death in postmenopausal women: A collaborative meta-analysis of randomized controlled trials. J Bone Miner Res 2014;doi:10.1002/jbmr.2311.

27. Hassler N, Roschger A, Gamsjaeger S, et al. Sclerostin deficiency is linked to altered bone composition. J Bone Miner Res 2014;doi:10.1002/jbmr.2259.

28. Roschger A, Roschger P, Keplinger P, et al. Effect of sclerostin antibody treatment in a mouse model of severe osteogenesis imperfecta. Bone 2014;66:182.

29. Ross RD, Edwards LH, Acerbo AS, et al. Bone matrix quality after sclerostin antibody treatment. J Bone Miner Res 2014;29:1597.

30. Ominsky MS, Niu Q-T, Li C, Li X, Ke HZ. Tissue‐level mechanisms responsible for the increase in bone formation and bone volume by sclerostin antibody. J Bone Miner Res 2014;29:1424.



Much has been said that milk consumption increases fracture risk in postmenopausal women. May results shown in this publication be taken as a solid argument against this idea? Are there other publications showing the opposite as far as you know? Best regards CDT Raúl Pineda Aquino / Mexico city.
Ego Seeman's picture

The data are controversial. I would like to see the paper you refer to. There is a need for a randomised trial of milk and milk products on fracture risk and cardiovascular health. Very difficult studies to design and execute.

What is the verdict? Will 250mg BID dosage be safer than 500mg OD dosage? Is there a potentially unsafe dose with regards to CV risk? Dharmanand/India
Ego Seeman's picture

We do not know. The data has still to be properly adjudicated regrading safety. MSD is doing this.

Professor, Thank you for your fabulous research!!! My wife with seven fractures in her spine has been told to take Forteo which is quite scary for her. What is the current thinking as it relates to this drug?? With appreciation!! Brian
Ego Seeman's picture

Sorry, but IOF cannot give personal medical advice. 

Supported by an unrestricted educational grant from