Dr Hodsman is a member of the advisory board for Eli Lilly Canada and NPS Pharmaceuticals. Drs Lindsay and Cosman have received speaker fees from Eli Lilly & Co. Dr Dempster has served as a consultant for and received speaker fees from Eli Lilly & Co. All other authors state that they have no conflicts of interest.
Effects Of a One-Month Treatment With PTH(1–34) on Bone Formation on Cancellous, Endocortical, and Periosteal Surfaces of the Human Ilium†
Article first published online: 8 JAN 2007
Copyright © 2007 ASBMR
Journal of Bone and Mineral Research
Volume 22, Issue 4, pages 495–502, April 2007
How to Cite
Lindsay, R., Zhou, H., Cosman, F., Nieves, J., Dempster, D. W. and Hodsman, A. B. (2007), Effects Of a One-Month Treatment With PTH(1–34) on Bone Formation on Cancellous, Endocortical, and Periosteal Surfaces of the Human Ilium. J Bone Miner Res, 22: 495–502. doi: 10.1359/jbmr.070104
- Issue published online: 4 DEC 2009
- Article first published online: 8 JAN 2007
- Manuscript Accepted: 3 JAN 2007
- Manuscript Revised: 11 NOV 2006
- Manuscript Received: 14 SEP 2006
- bone formation;
- bone histomorphometry;
Using bone histomorphometry, we found that a 1-month treatment with PTH(1–34) [hPTH(1–34)] stimulated new bone formation on cancellous, endocortical, and periosteal bone surfaces. Enhanced bone formation was associated with an increase in osteoblast apoptosis.
Introduction: The precise mechanisms by which hPTH(1–34) increases bone mass and improves bone structure are unclear. Using bone histomorphometry, we studied the early effects of treating postmenopausal women with osteoporosis with hPTH(1–34).
Materials and Methods: Tetracycline-labeled iliac crest bone biopsies were obtained from 27 postmenopausal women with osteoporosis who were treated for 1 month with hPTH(1–34), 50 μg daily subcutaneously. The results were compared with tetracycline-labeled biopsies from a representative control group of 13 postmenopausal women with osteoporosis.
Results: The bone formation rate on the cancellous and endocortical surfaces was higher in hPTH(1–34)–treated women than in control women by factors of 4.5 and 5.0, respectively. We also showed a 4-fold increase in bone formation rate on the periosteal surface, suggesting that hPTH(1–34) has the potential to increase bone diameter in humans. On the cancellous and endocortical surfaces, the increased bone formation rate was primarily caused by stimulation of formation in ongoing remodeling units, with a modest amount of increased formation on previously quiescent surfaces. hPTH(1–34)–stimulated bone formation was associated with an increase in osteoblast apoptosis, which may reflect enhanced turnover of the osteoblast population and may contribute to the anabolic action of hPTH(1–34).
Conclusions: These findings provide new insight into the cellular basis by which hPTH(1–34) improves cancellous and cortical bone architecture and geometry in patients with osteoporosis.
The effects of pth on the skeleton are complex and differ between states of elevated endogenous PTH and exogenous administration of PTH. When given as treatment for osteoporosis, a daily subcutaneous injection of hPTH(1–34) produces a rapid increase in markers of bone formation, followed by a more delayed increase in markers of bone resorption.(1) These effects peak after ∼6–12 months of administration and decline thereafter, despite continuing treatment, providing a rationale for limiting hPTH(1–34) treatment to a course of 2 years or less.(2) Bone mass increases more markedly at sites of predominantly cancellous bone compared with sites of predominantly cortical bone, where some declines, such as at the radius, have been recorded.(3) However, the more important outcome is a decrease in the risk of both vertebral and nonvertebral fractures with a median treatment duration of 19 months.(3)
The biochemical data seen on initiation of hPTH(1–34) treatment(1) suggest that the effects of hPTH(1–34) on bone are dependent on two separate processes: early stimulation of osteoblast activity, recruitment, or lifespan, followed by a more gradual activation of remodeling. The early increase in markers of bone formation are assumed to reflect increased formation of new bone tissue by osteoblasts, although whether this is caused by enhanced bone formation by osteoblasts already synthesizing osteoid when hPTH(1–34) treatment is begun or is related to activation of cells on inactive surfaces of bone is not clear.
Hodsman et al.,(4) as part of a longer-term study of hPTH(1–34) effects on bone mass and structure,(5,6) showed an increase in bone formation rate in cancellous bone after 1 month of hPTH(1–34) treatment and concluded that this must be because of activation of osteoblasts on resting surfaces in cancellous bone. Bone formation at the endocortical junction and the periosteal surface was not evaluated in that study. To date, there is no definitive evidence of stimulation of bone formation by hPTH(1–34) on the periosteal surface of humans. To examine the mechanism of early hPTH(1–34) actions on bone in more detail, we have re-evaluated the biopsies(5,6) to pursue a more in-depth study of the early changes in bone formation, not only in cancellous bone but also at the endocortical junction and on the periosteal surface. One previously suggested mechanism for the anabolic action of hPTH(1–34) is that it reduces osteoblast apoptosis and thereby increases osteoblast life span and bone formation.(7) This hypothesis is based on observations in mice and has not yet been tested in humans. We therefore also assessed osteoblast apoptosis in this study.
MATERIALS AND METHODS
All patients signed informed consent, and the study was approved by the Institutional Review Board for Research Involving Human Subjects at the University of Western Ontario. Standard 8-mm transiliac crest biopsies were obtained from 27 patients with postmenopausal osteoporosis treated with hPTH(1–34) (50 μg/day subcutaneously) for 28 days, as part of a clinical trial of cyclical hPTH(1–34) and sequential calcitonin therapy, as previously described.(4) The results of this study have previously been published in full.(5,6) Dual tetracycline labels were administered following a standardized 2-week interlabeling protocol, with the first label administered for 3 days at 2 weeks after initiation of hPTH(1–34). The biopsies were performed 2–4 days after completion of the second label. All patients continued to receive hPTH(1–34) throughout the labeling period. For logistical reasons, it was not possible to perform pretreatment biopsies in these patients. Therefore, the histomorphometric data reported here are compared with those obtained from 13 patients with severe postmenopausal osteoporosis. These control patients were recruited to an earlier study, exploring the use of cyclical hPTH(1–34) in the treatment of osteoporosis, and the biopsies were obtained as part of the baseline evaluation on entry to this study.(8) Because the inclusion criteria for both studies were similar, the two groups of patients were well matched for age, weight, dietary calcium intake, prevalent vertebral fractures, lumbar spine BMD, and biochemical markers of bone turnover (Table 1).(4,8) Neither the hPTH(1–34)–treated nor the control patients had been started on therapy for osteoporosis before initiating these protocols. The transiliac biopsies were performed under the same labeling conditions, by the same operator (ABH), and processed using the same methacrylate embedding technique. Biopsy cores were stored in the dark, but for the purposes of the study reported here, all sections were recut from the original blocks and batch-processed by one of us (HZ). Therefore, the interpretation of tetracycline labels was performed under strictly standardized conditions.
All biopsies were prepared in a standard fashion for histomorphometric analysis.(4,6,8) Serial sections from each patient were examined under UV light for evaluation of tetracycline labels and under polarized light after staining with methylene blue for evaluation of reversal and cement lines. There were no tetracycline labels in 2 of 27 PTH-treated patients and 3 of 13 controls. Two adjacent sections from each of three levels ∼100 μm apart were analyzed from each biopsy. Histomorphometric analysis was performed as previously described.(8) Where only single labels were observed in a biopsy, the mineral apposition rate was set at 0.3 μm/day; where there were no tetracycline labels, it was set at zero.
In addition to estimation of bone formation from the extent and width of the double labels, we identified different types of bone formation, based on pre-established criteria.(9) Where we identified osteons in which formation was occurring over a crenated reversal line, we designated this as regular, remodeling-based formation (RBF). We also identified osteons in which bone formation was occurring over smooth cement lines and designated this as modeling-based formation (MBF). Data were expressed as the number of each type of forming osteon per section. Remodeling- and modeling-based bone formation was further subdivided into those osteons that displayed single tetracycline labels (sL) and those that displayed double tetracycline labels (dL).
Osteoblast apoptosis was assessed as previously described.(7) Briefly, sections of transiliac bone biopsies and sections of positive controls from mammary tissue of a weaned rat were deplasticized and incubated in 10 mM citrate buffer, pH 6.0, in a microwave oven at 48°C for 5 minutes. They were placed in 0.5% pepsin in 0.1N HCl for 20 minutes at 37°C, rinsed with Tris-buffered saline, and reincubated in 30% H2O2 in methanol for 5 minutes. DNA fragmentation was detected by TUNEL using the Klenow-FragEL DNA Fragmentation kit (Oncogene Research Products, Boston, MA, USA). The application of Klenow terminal deoxynucleotidyl transferase was omitted for the negative control section. To improve the sensitivity of the reaction, sections were subsequently incubated for 2 minutes in 15% copper sulfate in saline. The sections were counterstained with 3% methyl green, dehydrated, and mounted. The cells whose nuclei were clearly dark brown from the horseradish peroxidase (HRP)-biotinylated nucleotide conjugate, instead of blue green from the methyl green, were interpreted as being positive for apoptosis. Quantitative analysis was performed on the TUNEL-stained sections using the OsteoMeasure digitizing image/analysis system (OsteoMetrics, Atlanta, GA, USA). Apoptotic osteoblasts were expressed as the number of TUNEL+ osteoblasts over the total osteoblast-covered perimeter (N.Apo.Ob/Ob.Pm, no./mm) on cancellous (Cn), endocortical (En), or intracortical (Ct) envelopes. The individual reading the biopsies (HZ) was blinded to the treatment assignment of the subjects.
Statistical analysis was performed using the statistical analysis programs NCSS 2004 and PASS (NCSS, Kaysville, UT, USA). To compare differences in all variables between PTH-treated and control biopsies, the distribution of data was first checked for normality. A two-sample t-test was performed for normally distributed variables, whereas a Mann-Whitney U test/Wilcoxon rank sum test was performed for data that were not normally distributed. Linear regression and correlation analysis was performed between variables of apoptotic osteoblasts and mineralizing perimeter, bone formation rate, and osteoblast perimeter. p < 0.05 was considered to be significant.
The hPTH(1–34)–treated and control groups were not significantly different from each other with respect to a number of demographic variables, including age, sex, height, weight, daily calcium intake, BMD by DXA, and number of vertebral fractures, as described previously (Table 1).(4–6,8) To establish similarity between laboratories, we remeasured bone formation rate in cancellous bone in all biopsies and obtained a value for BFR in hPTH(1–34)–treated individuals very similar to that in the original publication, and a value of BFR in controls slightly (although not significantly) lower than that originally published (Table 2). At all skeletal sites, including the periosteal surface, BFR was significantly higher in hPTH(1–34)–treated individuals than controls (Table 2). This was associated with a marked and significant increase in osteoid perimeter in cancellous bone and at the endocortex, without any change in eroded perimeter in cancellous bone and a slight decrease at the endocortex (Table 3).
Characteristics of early bone formation in cancellous bone
In cancellous bone, the bone formation rate was 4.5-fold higher in the hPTH(1–34)–treated subjects than in the control subjects (Table 2). This was the result of a significantly higher mineralized perimeter and a significantly higher mineral apposition rate (Table 3). The higher bone formation rate with hPTH(1–34) treatment was accompanied by a 60% higher number of osteons with remodeling-based formation (RBF), which was mainly caused by a significantly higher number of single-labeled osteons with RBF and a trend toward a higher number of osteons with modeling-based formation (MBF; Table 4). MBF with double labels was seen in the hPTH(1–34) –treated group but not in the controls (Table 4). Apart from this difference, the relative proportions of MBF to RBF were unchanged by hPTH(1–34) treatment (Fig. 1A). The average absolute length of mineralized perimeter per osteon was 38% higher in osteons with RBF and 271% higher in osteons with MBF in the hPTH(1–34)–treated group than in controls (Fig. 2A). On cancellous and endocortical surfaces, osteoblast apoptosis, evaluated by TUNEL assay, was significantly increased in the hPTH(1–34)–treated group, relative to the control group (Table 5). However, this did not result in compromised bone formation. On the contrary, the number of apoptotic osteoblasts per osteoblast perimeter was positively correlated with mineralized perimeter (r = 0.51, p = 0.02), mineral apposition rate (r = 0.39, p = 0.08), and bone formation rate (r = 0.54, p = 0.01; Fig. 3).
Characteristics of early bone formation on endocortical bone
As for cancellous bone, the bone formation rate on the endocortical envelope was 5-fold higher in the hPTH(1–34)–treated group (Table 2). This was the result of a higher mineralized perimeter and a slightly, but nonsignificantly, higher mineral apposition rate (Table 3). The higher bone formation rate was accompanied by a 106% increase in the number of osteons with RBF but, unlike cancellous bone, this was mainly caused by an increase in the number of double-labeled osteons with RBF. There was no change in the number of osteons with MBF (Table 4). This resulted in an increase in the proportion of osteons with RBF (97% in PTH versus 74% in control, p < 0.01) and a reciprocal decrease in the proportion of osteons with MBF (3% in PTH versus 26% in control, p < 0.01; Fig. 1B). The average absolute length of mineralized perimeter per osteon was 58% higher in osteons with RBF and 771% higher in osteons with MBF in the hPTH(1–34)–treated group than in controls (Fig. 2B). As for cancellous bone, the number of apoptotic osteoblasts was higher in the hPTH(1–34)–treated subjects than in the controls (Table 5).
Characteristics of early bone formation on the periosteal surface
Relative to the control group, hPTH(1–34) treatment increased mineralized perimeter, mineral apposition rate, and bone formation rate by 659%, 143%, and 300%, respectively, on the periosteal surface (Table 6). Single labels were found in 12 of 27 hPTH(1–34)–treated subjects but in only 3 of 13 controls. Double labels were seen in 2 of 27 hPTH(1–34)–treated patients, but none were observed in the controls (Table 6). The two types of forming osteon (RBF versus MBF) could not be reliably distinguished on the periosteal surface.
We previously reported that a 28-day course of hPTH(1–34) (50 μg/day) dramatically stimulated bone formation as assessed both biochemically and histomorphometrically.(4) In that study, the first tetracycline label was given just 14 days after initiation of treatment; therefore, we postulated that the increased bone formation occurred on previously quiescent bone surfaces rather than as a “coupled” consequence of a completed resorption cycle. To understand better the early effects of hPTH(1–34) treatment on bone formation, we subjected the same biopsies to a more detailed analysis in which we have distinguished between bone formation occurring on previously resorbed surfaces, which we term here RBF or on previously quiescent surfaces, termed MBF. We also extended the analysis to both the endocortical and periosteal surfaces of the cortex. The results are consistent with those obtained more recently(9) after treatment for 2 months with a lower dose of hPTH(1–34) (25 μg/day) and using a novel, quadruple tetracycline labeling schedule.
Compared with the control biopsies, the short period of treatment with hPTH(1–34) increased bone formation on all three envelopes: cancellous, endocortical, and periosteal. In cancellous bone, the bone formation rate in hPTH(1–34)–treated women was higher than in control women by a factor of 3.5, with an increase in both mineralized perimeter and mineral apposition rate. As we reported previously,(9) the primary mechanism for increased formation with hPTH(1–34) was an increase in remodeling-based formation. There was a strong trend toward an increase in modeling-based formation, but this was not statistically significant. However, double-labeled, modeling-based formation was observed in the hPTH(1–34)–treated biopsies, but not in the controls, suggesting that this phenomenon was occurring with greater frequency and/or with longer duration with hPTH(1–34) treatment. Furthermore, there was a marked increase in the absolute length of mineralized perimeter in osteons with MBF in the hPTH(1–34)–treated group, which is consistent with our previous suggestion that an important action of hPTH(1–34) is to extend formation beyond its original boundaries.(9) Nevertheless, only 5% of the new bone formation after hPTH(1–34) represented MBF, and this proportion was not different from controls. We should note that the increase in the number of osteons with RBF does not necessarily indicate that the corresponding resorption cavities were also started after the beginning of treatment. These osteons could well represent remodeling units that were in the resorption or reversal phases of the cycle at the start of treatment. Prolongation of the formation period in remodeling units would be another possible explanation for the increase in the number of osteons with RBF in the hPTH(1–34)–treated group.
Recently, Ma et al.(10) have studied the effects of longer-term (12–24 months) treatment with 20 and 40 μg/day of teriparatide on bone formation in postmenopausal women. Using similar methods to those used in this study, these authors also distinguished between modeling and remodeling osteons and included an additional category of mixed remodeling–modeling osteons. Similar to our current observations, the majority (>90%) of the forming osteons in the teriparatide-treated patients were classified as remodeling osteons with RBF. An increase in osteons with MBF was seen at the 40-μg/day dose, and an increase in mixed remodeling–modeling osteons was seen at both doses. These findings are consistent with our observations at earlier time-points in this study and in our previous study.(9) Also consistent with our finding of increased periosteal bone formation with teriparatide treatment, Ma et al.(10) reported enhanced IGF-II expression on the periosteal surfaces in the teriparatide-treated groups. The authors speculate that IGF-II may play a key role in creating a positive bone balance on the periosteum.
On the endocortical surface, hPTH(1–34)–stimulated bone formation was primarily the result of an increase in the number of osteons with RBF. There was not even a trend toward a change in the number of osteons with MBF. Indeed, the proportion of osteons with MBF was reduced in the hPTH(1–34)–treated subjects relative to controls. However, again, there was a dramatic increase in the absolute length of labeled perimeter per osteon with MBF, suggesting that the area of bone undergoing formation at these sites was extended by hPTH(1–34) treatment. In our earlier report on these biopsies,(4) we concluded that hPTH(1–34) treatment must have stimulated new bone formation on previously quiescent surfaces. This conclusion was solely based on the fact that the biopsies were removed 1 month after the beginning of treatment and that there was insufficient time to see a wave of new remodeling. This analysis supports the view that there is some stimulation of bone formation on previously quiescent surfaces on both cancellous and endocortical surfaces but that the bulk of the enhanced formation during early treatment with hPTH(1–34) occurs in remodeling units that were already underway at the time that treatment was initiated.
Our finding that an average of 26% of the osteons on the endocortical surface of the control biopsies were classified as displaying MBF with smooth cement lines was unexpected. This seems high when expressed as a proportion of the total number of forming osteons. We should note, however, that only 3 of the 13 control subjects showed such modeling-based formation on the endocortex. Some evidence for modeling in adult trabecular bone has previously been presented. Takahashi et al.(11) reported that 3% of cement lines in cancellous bone from normal individuals were smooth, representing MBF, leading Frost to speculate that a small amount of modeling can occur in adult cancellous bone.(12,13) More recently, Kobayashi et al.(14) obtained evidence of trabecular modeling, which they called minimodeling, in the iliac crest of subjects at the time of total hip replacement, but in whom there was no indication that the iliac crest was abnormal. Indeed, in that study, 62% of 27 postmenopausal women displayed formation sites in which fluorochrome labeling occurred over smooth cement lines. In that study, labeled surfaces of minimodeling sites made up as much as one fourth to one half of the entire labeled surface. Further work is needed to determine the extent to which modeling-based formation occurs in normal bone, particularly on the endocortical surface, which has not been previously characterized in this regard.
One postulated mechanism for the enhanced bone formation after daily PTH administration is a reduction in osteoblast apoptosis and a consequent increase in osteoblast life span.(7) To date, this has only been shown in mice. Our present observations do not support such a mechanism in humans. Indeed, the number of apoptotic osteoblasts per unit osteoblast perimeter was significantly higher in the hPTH(1–34)–treated subjects than in controls. One might have expected this to be associated with reduced formation. However, this did not seem to be the case because there was a positive correlation between apoptotic osteoblast number and bone formation rate. One possible explanation for this observation is that increased osteoblast number and or activity, stimulated by hPTH(1–34), is associated with a higher rate of osteoblast death. In support of this hypothesis, Chen et al.(15) showed that hPTH(1–34) enhanced apoptosis in postconfluent MC3T3-E1 osteoblastic cells, whereas in preconfluent cells, hPTH(1–34) protected against dexamethasone-induced apoptosis. The authors suggested that such dual effects of PTH act to accelerate osteoblast turnover by promoting viability of osteoblasts early and enhancing their departure from the differentiation scheme at a later stage of development. They further speculated that promotion of turnover of more mature cells may help to clear the way for less differentiated osteoblasts to produce bone matrix and thereby contribute to PTH's anabolic action. It is well recognized that cell proliferation and apoptosis are inevitably associated.(16) Hock and colleagues (16,17) used bromodeoxyuridine to label proliferating cells in young rats and showed that 3–5 days of hPTH(1–34) treatment increased osteoblast proliferation and that over the same time interval osteoblast apoptosis increased transiently, reaching a peak at 5–7 days of treatment. However, increased osteoblast proliferation with PTH treatment has not been a universal finding.(7,18) Whether intermittent hPTH(1–34) treatment decreases or increases osteoblast apoptosis or affects osteoblast proliferation is an important question that needs to be explored further in humans.
It is well established that PTH treatment is able to stimulate periosteal bone formation in animals.(19–23) One study has provided indirect evidence to support this in the distal radius of humans,(24) although that study was limited by its cross-sectional design. Moreover, a follow-up, longitudinal study(25) failed to confirm this finding in the proximal femur, using hip structural analysis (HSA), a technique with a number of limitations and assumptions.(25,26) Indeed, although there were improvements in femoral neck cross-sectional area with teriparatide treatment in that study, the outer diameter at the femoral neck and intertrochanteric regions were lower in the teriparatide-treated group than in placebo.(25) Our results, on the other hand, provide direct evidence that hPTH(1–34) is able to stimulate periosteal bone formation in humans. This did not occur in all individuals, but 45% of the hPTH(1–34)–treated patients displayed single tetracycline labels on the periosteal surface compared with only 23% in the controls. Furthermore, double labels were observed in three of the treated patients, whereas there was none in the controls. Comparing the mean values for the two groups, the hPTH(1–34)–treated group had a 7.6-fold higher mineralized perimeter, a 2.4-fold higher mineral apposition rate, and a 4-fold higher bone formation rate on the periosteal surface. These mean differences are in fact comparable with the increases in bone formation rate that occurred on the cancellous and endocortical surfaces. This finding lends support to the concept that daily treatment with hPTH(1–34) is capable of stimulating new bone formation on the periosteal surface in humans, providing a mechanism for increased cortical diameter and, as a consequence, improved cross-sectional moment of inertia. We also confirmed our finding that hPTH(1–34) stimulates endosteal bone formation, contributing to the observed increases in endosteal wall thickness and cortical thickness.(27,28)
This study has a number of important limitations. First, the sample size in each group is small, and a paired biopsy design was not used. Second, subjects were not randomized to treatment or control groups. Rather, the subjects in the control group were recruited as part of a previous study (see the Materials and Methods section). They were, however, not significantly different from the treated group with respect to age, sex, height, weight, daily calcium intake, number of vertebral fractures, or BMD.(4,8) Third, the dose (50 μg/day) of hPTH(1–34) was higher than that approved for clinical use, limiting the ability to extrapolate from this study to the clinical experience. However, it is worth noting that in the pivotal fracture trial,(3) the 40-μg dose was no more effective than the 20-μg dose in lowering vertebral and nonvertebral fracture risk.
In conclusion, we showed that a 1-month treatment of postmenopausal women with osteoporosis with hPTH(1–34) dramatically stimulates the bone formation rate on cancellous, endocortical, and periosteal surfaces of the iliac crest. On the cancellous and endocortical surfaces, the increased bone formation rate was primarily caused by stimulation of formation in ongoing remodeling units, with a modest amount of increased formation on previously quiescent surfaces. Demonstration of enhanced periosteal bone formation indicates that hPTH(1–34) has the potential to increase bone diameter and, thereby, cross-sectional moment of inertia in humans. hPTH(1–34)–stimulated bone formation was associated with an increase in osteoblast apoptosis, which may reflect enhanced turnover of the osteoblast population, leading to an overall increase in bone-forming activity. These findings provide further insight into the cellular basis by which hPTH(1–34) treatment improves cancellous and cortical bone architecture and geometry and lowers fracture risk in osteoporosis.
This study was supported by NIH Grant AR051454, an unrestricted research grant from Bristol Myers Squibb, and the Helen Hayes Foundation.
- 51997 A randomized controlled trial to compare the efficacy of cyclical parathyroid hormone versus cyclical parathyroid hormone and sequential calcitonin to improve bone mass in postmenopausal women with osteoporosis. J Clin Endocrinol Metab 82: 620–628., , , , , , ,
- 92006 A novel tetracycline labeling schedule for longitudinal evaluation of the short-term effects of anabolic therapy with a single iliac crest bone biopsy: Early actions of teriparatide. J Bone Miner Res 21: 366–373., , , , , , ,
- 111964 Evidence that bone resorption precedes formation at the cellular level. Henry Ford Hosp Med Bull 12: 359–364., , ,
- 131995 Introduction to a New Skeletal Physiology, vol 1. Bone and Bones. Pajaro Group, Pueblo, CO, USA.
- 222001 Intermittently administered human parathyroid hormone(1–34) treatment increases intracortical bone turnover and porosity without reducing bone strength in the humerus of ovariectomized cynomolgus monkeys. J Bone Miner Res 16: 157–165., , , , ,
- 262000 Structural trends in the aging femoral neck and proximal shaft: Analysis of the Third National Health and Nutrition Examination Survey dual-energy X-ray absorptiometry data. J Bone Miner Res 15: 2297–2304., , , ,