Parts of this manuscript were presented at the International Osteoporosis Foundation, Toronto, Canada, June 2–6, 2006 and at the Sixth European Congress on Clinical and Economic Aspects of Osteoporosis and Osteoarthritis, Vienna, Austria, March 15–18, 2006. Dr Miller serves as a consultant and receives honoraria from Eli Lilly and Company, Merck, Novartis, Procter & Gamble, Amgen, and Roche. Dr Recker serves as a consultant and receives an honorarium from Eli Lilly and Company, Merck, Procter & Gamble, Novartis, Roche, GSK, Amgen, NPS, and Wyeth. Dr Resch has no conflict of interest. Drs Chen, Rana, Pavo, and Sipos are employees of Eli Lilly and Company.
Increases in BMD are correlated with improvements in 2D and 3D trabecular microarchitecture indices with teriparatide treatment. Therefore, improvements in trabecular bone microarchitecture may be one of the mechanisms to explain how BMD increases improve bone strength during teriparatide treatment.
Introduction: Bone strength is determined by BMD and other elements of bone quality, including bone microarchitecture. Teriparatide treatment increases BMD and improves both cortical and trabecular bone microarchitecture. Increases in lumbar spine (LS) BMD account for ∼30–41% of the vertebral fracture risk reduction with teriparatide treatment. The relationship between increases in BMD and improvements in cortical and trabecular microarchitecture has not yet been studied.
Materials and Methods: The relationship between increases in BMD and improvements in cortical and trabecular microarchitecture after teriparatide treatment was assessed using data from a subset of patients who had areal BMD measurements and structural parameters from transiliac bone biopsies in the Fracture Prevention Trial. 2D histomorphometric and 3D μCT parameters were measured at baseline and 12 (n = 21) or 22 (n = 36) mo. LS BMD was assessed at baseline and 12 and 18 mo, and femoral neck (FN) BMD was measured at baseline and 12 mo. Pearson correlation was performed to assess the relationship between actual changes in BMD and actual changes in microarchitectural parameters.
Results: Changes in LS BMD at 12 mo were significantly correlated with improvements in trabecular bone structure at 22 mo: 2D bone volume (r = 0.45, p = 0.02), 2D mean wall thickness (r = 0.41, p = 0.03), 3D bone volume (r = 0.48, p = 0.006), 3D trabecular thickness (r = 0.44, p = 0.01), 3D trabecular separation (r = −0.37, p = 0.04), 3D structural model index (r = −0.54, p = 0.001), and 3D connectivity density (r = 0.41, p = 0.02). Changes in LS BMD at 18 mo had similar correlations with improvements in bone structure at 22 mo. Changes in FN BMD at 12 mo were significantly correlated with changes in 2D mean wall thickness (r = 0.56, p = 0.002), 3D bone volume (r = 0.51, p = 0.004), 3D trabecular thickness (r = 0.44, p = 0.01), 3D trabecular separation (r = −0.46, p = 0.01), and 3D structural model index (r = −0.55, p = 0.001).
Conclusions: Increases in BMD are correlated with improvements in trabecular microarchitecture in iliac crest of patients with teriparatide treatment. Therefore, improvements in trabecular bone microarchitecture may be one of the mechanisms to explain how BMD increases improve bone strength during teriparatide treatment.
Bone strength depends on two groups of interdependent properties representing structural and material features.(1,2) Structural properties include macro- (size and shape) and microarchitecture, whereas characteristics of mineral and collagen contents determine the material properties.(3) In clinical practice, changes in BMD reflect mainly changes in mineral content (material properties) and, to a lesser extent, changes in bone structure. BMD measured by DXA offers sufficient precision to follow osteoporosis disease progression.(4) In case of antiresorptive therapies, although BMD increases account for a small portion of the fracture risk reduction,(5–8) BMD measurement is currently the best validated technology for evaluating the clinical response to these type of therapies.(9) On the other hand, histomorphometry is a useful tool for assessing the quality of the bone.
Teriparatide (rDNA origin) injection [recombinant human PTH(1–34)] is a bone-forming agent for the treatment of osteoporosis. In the Fracture Prevention Trial (FPT), daily self-injections of teriparatide (20 and 40 μg) reduced the risk of new vertebral and nonvertebral fractures by 65% and 53%, respectively, in postmenopausal women with osteoporosis.(10) Teriparatide significantly increases BMD and improves both trabecular and cortical bone microarchitecture(10,11) through a mechanism of action on bone remodeling that is essentially opposite from that of bisphosphonates.(12) It has been recently shown that increases in BMD account for a larger portion of the vertebral fracture risk reduction with teriparatide treatment than with antiresorptive treatments.(5–8,13) Consequently, BMD increases during teriparatide treatment may reflect different structural and material changes than during antiresorptive therapies. The improvement in different elements of bone architecture has been suggested in patients with PTH(1–34) treatment.(14–18) It is of interest to understand whether increases in BMD as measured by DXA could capture improvements in bone microarchitecture with teriparatide treatment. The objective of this study, therefore, was to investigate whether changes in BMD correlate with bone structural improvements observed in patients treated with teriparatide treatment.
MATERIALS AND METHODS
These posthoc analyses were conducted using data from a subset of patients who had areal BMD measurements and structural parameters from transiliac bone biopsies in the FPT. Methods and results of this trial have been previously published.(10) Briefly, 1637 ambulatory, postmenopausal women ranging in age from 42 to 86 yr were randomized in a double-blinded manner to receive daily, self-administered, subcutaneous injections of placebo (n = 544), teriparatide 20 μg/d (n = 541), or teriparatide 40 μg/d (n = 552). The primary endpoint of this study was the number of women who experienced a new vertebral fracture. Secondary endpoints were nonvertebral fracture and BMD, assessed by DXA. The design and conduct of this trial have been previously reported.(10)
A total of 102 patients from 11 sites and 5 countries participated in the biopsy substudy of the FPT. Specimens of 15 patients at baseline were not eligible for analysis because of the lack of one or two cortices. Thus, baseline 2D histomorphometry analyses were performed in biopsies of the remaining 87 patients. Correlations between specific 2D histomorphometry parameters and corresponding individual BMD values were reported in patients where both were available. 3D measurements were performed in paired biopsies. Thirty-six of these paired biopsies (13 in placebo-treated and 23 in teriparatide-treated patients) were collected at baseline and at study end (mean: 22 ± 2 mo; range, 19–25 mo) and were the subject of this study. Quantitative and qualitative analyses of the baseline characteristics and the efficacy response confirmed that the substudy cohort was reflective of the entire FPT cohort.(10,11) Comparisons between the teriparatide-treated groups and the placebo group from the 2D histomorphometry and 3D μCT analyses have been previously reported.(11)
Patients signed informed consent to the treatment and investigation protocol, which was approved by the Institutional Review Board for Research Involving Human Subjects, at each participating center.
BMD was assessed by DXA using Hologic (Hologic, Bedford, MA, USA), Norland (Cooper-Surgical, Trumbull, CT, USA), and Lunar equipment (GE Medical Systems, Madison, WI, USA). Lumbar spine BMD (LS BMD) measurements were performed at baseline, 12 mo, 18 mo, and study endpoint. To eliminate differences attributable to densitometer manufacturer, spine BMD values were converted to standardized units (expressed as milligrams per square centimeter).(19) Vertebrae whose fracture status changed during the trial as assessed by lateral spine radiography were removed from the calculation of both baseline and follow-up LS BMD. At least two evaluable vertebrae were required for assessment of spine BMD; this was calculated from the sum of the BMCs and the sum of the areas of all evaluable spine vertebrae. Femoral neck BMD (FN BMD) measurements were performed at baseline, 12 mo, and study endpoint.
Histomorphometry and μCT
The transiliac crest bone biopsies were carried out using a Bordier-Meunier 9-mm needle trephine system, after in vivo double labeling with tetracycline fluorochromes given orally in a 3:12:3-day sequence as published earlier.(11) The methods for biopsy specimen evaluations have been published previously.(11,20,21) Biopsies were obtained from patients at baseline, before receiving treatment. Patients were randomized to have a follow-up biopsy from the contralateral iliac crest after either 12 mo or at study end. The 12-mo biopsies (mean: 12 ± 1 mo; range, 11–15 mo) were performed in 8 placebo patients and 13 teriparatide-treated patients. The biopsy at study end (mean: 22 ± 2 mo; range, 19–25 mo) was performed in 13 placebo patients and 23 teriparatide-treated patients.
Static and dynamic 2D parameters were measured under light microscopy on 5-μm-thick biopsy sections stained with Goldner's trichrome. The μCT image acquisition and analysis were performed on specimens already sectioned for histomorphometry as described previously.(11,20,21) Briefly, nondecalcified biopsy specimens embedded in methylmethacrylate were trimmed to fit the specimen hold and examined using a compact fan-beam-type μCT 20 (Scanco Medical AG, Bassersdorf, Switzerland). A scout view scan was obtained for selection of the examination volume of the specimens, by automatic positioning, measurement, and offline reconstruction. For each sample, a total of −600 microtomographic sections were acquired with a slice increment of 17 μm. The field of view was 17 × 17 mm2, and the matrix size was 1024 × 1024. Images with istotropic resolution of 17 μm3 were obtained. μCT 3D parameters were directly measured without stereological model assumption. All measured and derived variables were expressed according to the standard nomenclature recommended by the American Society of Bone and Mineral Research nomenclature committee.(22)
Structural changes during teriparatide treatment were assessed using both 2D and 3D indices. The static 2D histomorphometry parameters included in this analysis were total bone volume per tissue volume (trabecular bone volume [BV/TV]), mean wall thickness (W.Th, μm) of completed osteons, marrow star volume (Ma.St.V, mm3), trabecular diameter (Tb.Dm, μm), trabecular number (Tb.N, mm−1), trabecular separation (Tb.Sp, μm), and cortical porosity. Cortical thickness (Ct.Th, μm), expressed as the average of evenly spaced measurements across each of the inner and outer cortices, was also measured. The 3D structural parameters included trabecular bone volume per total volume (BV/TV), trabecular thickness (Tb.Th, μm), trabecular separation (Tb.Sp, μm), trabecular number (Tb.N, mm−1), structural model index (SMI), connectivity density (CD), degree of anisotropy (DA), cortical porosity (Ct.Po), and cortical thickness (Ct.Th, mm).(11)
Pearson correlation was performed to assess the relationship between the baseline BMD and baseline static 2D and 3D indices. To further assess the role of baseline BMD in determining later structural changes during teriparatide treatment, Pearson correlation coefficients between baseline BMD and change from baseline in structural indices were also calculated. The relationship between changes in BMD and improvements in bone structure with teriparatide therapy was assessed by correlating change from baseline in BMD to change from baseline in structural indices. To establish the true relationship between BMD and structural responses, subjects from the placebo and treated groups were analyzed as a single group. Combining these groups reflects the poor adherence to therapy that can be encountered in clinical practice,(23–26) yielding practical information for the clinician. The α level for significance was 0.05.
Of 1637 randomized patients from the FPT, the baseline characteristics of 21 placebo-treated patients and 36 teriparatide-treated patients who had paired iliac crest biopsies specimens, with at least one 2D or 3D index suitable for analysis, has been described elsewhere.(27) Briefly, there were no statistically significant differences in the baseline characteristics including baseline BMD and 2D histomorphometry and 3D μCT indices between the teriparatide and placebo groups. However, the correlation coefficients in this analysis were based on all biopsies.
Correlations between baseline BMD and 2D histomorphometry and 3D μCT indices
Baseline LS BMD was positively correlated with baseline 2D BV/TV, 2D Tb.Dm, 2D Ct.Th, 3D Tb.Th, and 3D CD. LS BMD was inversely correlated with 2D Ma.St.V and 3D Tb.Sp. There were no significant correlations between LS BMD and the other structural parameters at baseline (Table 1).
Table Table 1.. Correlation Coefficients Between Baseline BMD and Baseline 2D and 3D Structural Parameters
There were more correlations between baseline FN BMD and baseline structural parameters. Baseline FN BMD was positively correlated with baseline 2D BV/TV, 2D W.Th, 2D Tb.Dm, 2D Ct.Th, 3D BV/TV, 3D Tb.N, and 3D Tb.Th. FN BMD was inversely correlated with 2D Ma.St.V, 3D Tb.Sp, and 3D DA. There were no significant correlations between baseline FN BMD and the other structural parameters (Table 1). Baseline 2D structural parameters (BV/TV, Tb.N, and Tb.Sp) were significantly correlated with those of the 3D analyses (r = 0.57, 0.43, and 0.47, respectively, all p < 0.001).
Correlations between baseline BMD and change in 2D histomorphometry and 3D μCT indices
In general, the correlations between baseline BMD and changes in structural parameters between the start and end of teriparatide treatment were weak. Baseline LS BMD was inversely correlated with changes in 3D Tb.Sp (r = −0.35, p = 0.05) and SMI (r = −0.36, p = 0.05) and was positively correlated with changes in 3D Tb.N (r = 0.35, p = 0.05) at 22 mo. There were no significant correlations between baseline LS BMD and the other structural parameters. Baseline FN BMD was not significantly correlated with changes in any of the structural parameters.
Correlations between change in LS BMD and change in 2D histomorphometry and 3D μCT indices
Changes in LS BMD at 12 mo were significantly positively correlated with changes in 2D BV/TV, 2D W.Th, 3D BV/TV, 3D Th.Th, and 3D CD at 22 mo. Changes in LS BMD at 12 mo were significantly negatively correlated with changes in 3D SMI and Tb.Sp at 22 mo (Table 2). Changes in LS BMD at 18 mo were significantly correlated with the same parameters as changes in LS BMD at 12 mo, with the exception of 2D W.Th and 3D Tb.Sp (Table 2; Fig. 1).
Table Table 2.. Correlation Coefficients Between Changes in BMD and Changes in 2D and 3D Structural Parameters at Study End (Median of 22 Mo of Treatment)
Correlations between change in FN BMD and change in 2D histomorphometry and 3D μCT indices
Changes in FN BMD at 12 mo were significantly positively correlated with changes in 2D W.Th, 3D BV/TV, and 3D Tb.Th at 22 mo. Changes in FN BMD at 12 mo were also significantly negatively correlated with changes in 3D SMI and 3D Tb.Sp (Table 2; Fig. 2).
We found that low BMD was correlated with the worsening of most trabecular histologic indices in patients with established postmenopausal osteoporosis. These correlations suggest the presence of an impaired bone microstructure in patients with low BMD. These findings are in agreement with earlier observations about the significant correlation between trabecular bone volume of the iliac crest and spinal density measured by dual photon absorptiometry in patients with osteoporosis.(28) The deterioration in trabecular microarchitecture, thinning of trabeculae in paired iliac crest biopsies, was also shown in women during the 5-yr menopausal transition period, generally characterized by fast bone loss.(29) These data indicate that BMD, to some extent, reflects trabecular bone microstructure. A significant portion of variance for structural histomorphometry indices, however, could not be explained by BMD but may be related to other bone quality aspects. For example, in the same cohort reported here, after adjustment to age and BMD, histomorphometric bone volume was lower in women with moderate and severe fractures than in women with no fractures.(30)
Numerous data have shown that deteriorated bone microarchitecture is associated with fractures.(30–35) Parfitt et al.(24) were the first to report that the lower cancellous bone volume and loss of structural elements that occur with aging were more severe in men and women with vertebral or hip fractures compared with patients without fractures. In other studies, trabecular thickness, number, and separation were correspondingly worse in the patients with fractures.(32,35) Similarly, spatial arrangement of bone, assessed by fractal analysis or histomorphometry, is compromised in women with vertebral fractures, and described changes in trabecular complexity that are associated with a biomechanically compromised structure.(31) Moreover, the increase in vertebral fracture severity was associated with lower cancellous bone volume, impaired trabecular connectivity, and a transformation of trabeculae from platelike to more rodlike morphology, suggesting reduced biomechanical competence after adjusting for age and spine bone density.(30) Our data suggest an interrelationship between BMD and microarchitecture independent of fractures and that low BMD may explain 20–30% of the deterioration in bone microarchitecture in an osteoporotic patient population with prevalent fractures.
We showed that increases in both LS BMD and FN BMD were associated with improvements in indices of trabecular microarchitecture after long-term teriparatide treatment. This is the first observation suggesting that microarchitecture improvement induced by teriparatide correlates with the increase of BMD. Teriparatide treatment provides a good opportunity to study this relationship because teriparatide increases BMD and improves both cancellous and cortical bone structure.(10,11) Similar, but less consistent microarchitectural improvements were described in earlier studies with PTH(1–34).(14–18) These studies were different from those with teriparatide in one or more of the following: shorter treatment duration, smaller sample size, concomitant use of antiresorptive medications, different sexes, or cyclical treatment regimens. In contrast to teriparatide, therapies with bisphosphonates increase BMD without significantly affecting bone microstructure. For example, risedronate treatment resulted in increased BMD but largely unchanged bone volume and architectural parameters. The preservation of microarchitecture on risedronate compared with the deterioration observed in the placebo group may contribute to the fracture risk reduction observed in risedronate therapy.(36) Our observations are the first to suggest that the fracture risk reduction with an anabolic agent may also be related to improvements in bone microarchitecture and that this microarchitecture improvement is also correlated to the improvement in BMD. The findings of this study may also explain the fact that teriparatide-mediated increases in BMD account for a higher percentage of the fracture risk reduction compared with those observed in antiresorptive agents.(5–8,13)
The present data indicate that the structural improvements during teriparatide treatment are largely independent of baseline BMD status. These results are consistent with the findings that BMD increases and vertebral fracture risk reduction with teriparatide treatment was also independent of baseline BMD.(37) Consequently, the observed trabecular microarchitecture improvement with teriparatide(11) may occur regardless of baseline bone mass in patients with established osteoporosis.
In contrast to trabecular bone, changes in BMD did not correlate significantly with structural changes in cortical bone, although a significant increase in 3D cortical thickness had been observed in these patients treated with teriparatide.(11) These results are consistent with the findings that the early changes of bone formation markers did not correlate significantly with structural changes in cortical bone.(27) This lack of correlation may be caused by the relatively low sample size or can be explained by a different relationship between dynamic changes in bone turnover and bone formation within cortical and trabecular bone compartments.(38,39)
In our analyses, compared with traditional 2D structural histomorphometry, the correlations between BMD and 3D microstructure indices were generally stronger. This is in agreement with the growing evidences that the more stereologically correct indices like μCT-based 3D parameters could be more accurate for use of the analysis of small samples, such as bone biopsies of the iliac crest.(11,29) However, our data are insufficient to conclude which modality (2D or 3D) could have more value in clinical practice. The small sample size in this cohort precluded the analysis on the relationship between microarchitecture and the fracture risk.
Our results showed a correlation of BMD measured in the spine or femoral neck with microarchitecture investigated in the iliac crest. The iliac crest is the skeletal site used for the invasive bone sampling, and studies are ongoing to determine to which extent iliac crest microarchitecture reflects the structure of more clinically relevant sites such as vertebra.(40–42) A growing number of evidence suggests that iliac crest microarchitecture reflects the structure and strength of more clinically relevant sites such as vertebra. For example, postmenopausal women with increasing severity of vertebral fractures had progressively worse microstructure integrity, evaluated from iliac crest histomorphometry indices.(30) In addition, Dempster et al.(41) found histomorphometric BV/TV and Tb.Sp of the iliac crest correlated well with vertebral ultimate compressive stress. Our study showed that trabecular bone architecture of the iliac crest correlated well with vertebral and femoral neck BMD and may reflect certain bone quality and quantity elsewhere in the skeleton. Moreover, the power to detect structural abnormalities and treatment effects by transiliac biopsy is surprisingly strong given that the biopsy is such a small sample of the skeleton, from a site that usually does not fracture.
Areal BMD measurement is an accepted clinical surrogate for bone mass, but the technique has several limitations. For example, only a 2D assessment of the skeleton is obtained, spatial resolution is low, and the geometry of most bones cannot be evaluated with routine analysis.(43) The increase of spine areal BMD is markedly less than volumetric BMD increase in patients treated with teriparatide.(30) Nevertheless, the increase in BMD in the lumbar spine significantly correlated with the volumetric increase (data not published). This correlation suggests that areal BMD to some extent could also reflect microstructural changes in patients on teriparatide.
In conclusion, this study showed that increases in BMD are correlated with improvements in 2D and 3D trabecular microarchitecture indices with teriparatide treatment. Therefore, improvements in trabecular bone microarchitecture may be one of the mechanisms how BMD increases explain the improvements in bone strength during teriparatide treatment.
The authors thank the investigators of the Fracture Prevention Trial. This study was supported by Eli Lilly and Company. Data were analyzed at Lilly Research Laboratories, Eli Lilly and Company.