Teriparatide in Bisphosphonate-Resistant Osteoporosis: Microarchitectural Changes and Clinical Results After 6 and 18 months
- First Online:
- Cite this article as:
- Jobke, B., Muche, B., Burghardt, A.J. et al. Calcif Tissue Int (2011) 89: 130. doi:10.1007/s00223-011-9500-6
- 375 Views
A number of osteoporotic patients under bisphosphonate treatment present persistent fragility fractures and bone loss despite good compliance. The objective of this 18-month prospective study was to investigate the effect of teriparatide [rhPTH(1–34)] in 25 female osteoporotics who were inadequate responders to oral bisphosphonates and to correlate microarchitectural changes in three consecutive iliac crest biopsies measured by micro-computed tomography (μCT) with bone mineral density (BMD) and bone serum markers. Scanned biopsies at baseline (M0), 6 months (M6), and 18 months (M18) demonstrated early significant (P < 0.01) increases in bone volume per tissue volume (+34%) and trabecular number (+14%) at M6 with only moderate changes in most μCT structural parameters between M6 and M18. μCT-measured bone tissue density was significantly decreased at M18, expressing an overall lower degree of tissue mineralization characteristic for new bone formation despite unchanged trabecular thickness due to increased intratrabecular tunneling at M18. μCT results were consistent with serum bone turnover markers, reaching maximal levels of bone alkaline phosphatase and serum β-crosslaps at M6, with subsequent decline until M18. BMD assessed by DXA demonstrated persistent increases at the lumbar spine until M12, whereas no significant change was observed at the hip. Type (alendronate/risedronate) and duration (3.5 ± 4 years) of prior bisphosphonate treatment did not influence outcome on μCT, BMD, or bone marker results. The overall results indicate a positive ceiling effect of teriparatide on bone microarchitecture and bone markers after 6 and 12 months for lumbar spine BMD, with no additional gain until M18 in bisphosphonate nonresponders.
Bone loss in male and female osteoporosis patients may be significantly decelerated, and associated fractures at the spine or hip can be reduced with different antiresorptive treatments documented by more than 10 years’ experience with bisphosphonate treatment in large cohorts [1, 2]. Preservation of bone microarchitecture and bone volume measured in bone samples from the iliac crest is possible by reducing or balancing bone turnover with antiresorptive medications [3–5]. Still, a number of individuals in clinical practice do not respond well to bisphosphonates despite correct intake of the drugs and good compliance . This may lead to persistent increases in fracture risk and continuous bone loss that require alternative treatment. In contrast to bisphosphonates, parathyroid hormone (PTH) analogues are partially able to restore bone microarchitecture by increasing bone turnover with the ability to enforce trabecular and cortical adaptability to mechanical loading [7–10]. An initial study on changing treatment regimes from antiresorptive to osteoanabolic therapy demonstrated data indicating a delayed PTH response on bone turnover markers and bone density , which were rebutted in a number of studies [12–17]. The current investigation presents data from an 18-month prospective teriparatide bone biopsy study in bisphosphonate nonresponders (BBB Study). Results of rapid microarchitectural changes in the context of the clinical response are presented and potentially new treatment regimes discussed. To the best of our knowledge this is the first human PTH study using a minimally invasive biopsy technique, not previously outlined, at three consecutive time points and investigating cancellous bone microarchitecture and tissue mineralization. The purpose of this study was to obtain additional information regarding the onset and duration of new bone formation reflected by architectural, densitometric, and volumetric changes under recombinant human PTH following inadequate response to long-term bisphosphonate therapy for osteoporosis.
Materials and Methods
Summary of mean (SD) baseline characteristics of 25 women enrolled in the BBB study
Total group (n = 25)
Alendronate (n = 12)
Risedronate (n = 13)
P (Student’s t-test)
Previous BP therapy duration (months)
Lumbar spine BMD (T score)
Total-hip BMD (T score)
Biochemical markers of bone turnover
Serum Ca (mmol/l)
Serum β-CTX (pg/ml)
Bone ALP (μg/l)
Clinical Tests at Baseline and Follow-Up Measures
BMD measurements at the lumbar spine (L1–L4) and total-hip scans were performed on the same DXA scanner (Prodigy; GE Lunar, Madison, WI) every 6 months by the same technician. Strict local quality measures were followed (standard quality-control and daily phantom scans).
Biochemical markers of bone turnover (bone markers; BTM), including calcium (Ca), bone-specific alkaline phosphatase (bone ALP; ACCESS Ostase Assay; Beckman Coulter, Krefeld, Germany), and C-terminal cross-linking telopeptide of type I collagen (S-CTX, Crosslaps; Roche Diagnostics, Mannheim, Germany), were assessed at months 0, 1, 3, 6, 12, and 18. Upper normal limits for CTX were set at 573 pg/ml and for bone ALP at 21.4 μg/l by the same laboratory that was used throughout the study period. Baseline values for BMD and bone markers were similar throughout the patient population and did not differ significantly between patients who had used either ALN or RIS.
Teriparatide (Forsteo®) was provided by Lilly Deutschland (Bad Homburg, Germany). All enrolled subjects were supplemented with a daily intake of 500 mg of elemental Ca and 400 IU of vitamin D3 . The women self-administered a once-daily subcutaneous injection of 20 μg teriparatide for 18 months. Due to the severity of the cases, PTH treatment was initiated immediately (within 1 month) following prior BPs without any washout phase. Elevations in serum Ca levels >2.55 mmol/l (ULN of the local lab, equivalent to 10.22 mg/dl) resulted in a discontinuation of Ca supplementation if necessary. Compliance was assessed by measuring the amount of medication in the applicator pens at all returning visits (M1, M3, M6, M12, and M18) using an Excel table provided by Lilly Deutschland.
Jamshidi bone biopsies were obtained from alternating sides of the dorsal iliac crest. The third biopsy was taken at the same side of the dorsal iliac crest as the first one, at a similar site with a different biopsy angle (craniocaudal direction).
No complications were reported following the biopsies. All biopsies (71 in total) were obtained by the same experienced clinician (B. M.) at one institution.
Biopsies were embedded undecalcified in methyl methacrylate. Following micro-computed tomographic (μCT) scanning, the biopsies were sectioned for histological investigation to exclude detrimental effects on mineralization, bone remodeling, and hematopoiesis. Semiquantitative histological bone turnover estimated by remodeling units per field of view was noted. A descriptive summary is presented.
All serious adverse events revealed by histology were reported. The number of samples needed at each time point was approximated based on reports from previous studies with PTH analogues [19–22], which showed statistically robust data with 8–12 biopsy specimens. To compensate for any loss of patients during the study (expected 20%) or improper, fragmented, or compressed bone samples (expected 10%) and to improve statistical power, it was calculated to include 25 patients having at least 15 biopsies at each time point available.
Samples were quantified 3-dimensionally using μCT (μCT 40; Scanco Medical, Brüttisellen, Switzerland) at 55 kVp, 145 μA. A 12-mm field of view was acquired, achieving a 12 μm isotropic nominal resolution. A fixed threshold of 400 (40% of the positive integer range) was chosen based on visual test, with different thresholds ranging 300–400. To avoid subcortical bone inhomogeneity, a 2-mm distance from the cortex was maintained. Also, the opposite spongeous end of the biopsy was excluded in order to avoid compression artifacts and microfractures. A minimum of 400 slices (5 mm, Ø 820 slices) of spongeous bone were required for μCT evaluation. Biopsies that did not meet these criteria due to biopsy-related microfractures were not included but were embedded for histological investigation. Cancellous bone volume fraction (BV/TV), structural indices (connectivity density [Conn.D], structure model index [SMI], trabecular number [Tb.N], trabecular thickness [Tb.Th], trabecular separation [Tb.Sp], and standard deviation of trabecular separation [Tb.Sp.SD]) were assessed in 19 paired biopsies [23–27]. We counted intratrabecular tunneling when resorption “lacunae” had at least twice the depth of their width and three or more intratrabecular tunnels per biopsy were observed. In addition to structural parameters, bone tissue density was determined from the original gray-scale images. The binary image map from the segmentation step was eroded by two voxels to remove partial volume components and then used to mask the gray-scale data. The mean linear attenuation was calculated as the sum of the masked image divided by the number of bone voxels in the eroded mask. This value was converted to mineral density (mgHA/ccm) based on a phantom-derived linear calibration [28, 29].
Clinical baseline values (mean ± SD) and absolute changes after 1, 3, 6, 12, and 18 months of teriparatide treatment
Baseline (n = 25)
1 month (n = 25)
3 months (n = 25)
6 months (n = 22)
12 months (n = 22)
18 months (n = 22)
Lumbar spine BMD (T score)
Total-hip BMD (T score)
Serum Ca (mmol/l)
Serum β-CTX (pg/ml)
Bone ALP (μg/l)
μCT baseline absolute values (mean ± SD) and after 6 and 18 months of teriparatide treatment in paired biopsies
Baseline (n = 19)
6 months (n = 22)
18 months (n = 17)
Tb. thickness (μm)
Tb. separation (μm)
Tb. number (1/mm)
Tb. separation SD (structural heterogeneity)
Intratrabecular tunnelling (%)
Tissue density (degree of mineralization, mgHA/ccm)
A total of 25 patients were enrolled between July 2005 and July 2006. Twenty patients finished the 18-month trial, and 71 biopsies totaling over three time points were acquired. μCT scans were evaluated for 19 (M0), 22 (M6), and 17 (M18) biopsies (58 total), which made paired follow-up comparison in 19 (M0–M6) and 17 (M6–M18) biopsies possible (Table 3). A number of patients had to discontinue the study. The dropout rate was 20%. Of the total 25 patients at study entry, six were excluded from μCT evaluations or further biopsies, including three who discontinued the clinical trial for the following causes: compliance problems at M3 (n = 1), evidence of mild primary hyperparathyroidism at M0 (n = 1), insufficient or fragmented biopsy material at M0–M18 (n = 3), and discovery of a recently unknown and asymptomatic chronic lymphatic leukemia in a 67-year-old female patient at M6. The patient was excluded from the study and assigned to a hematologist. One patient was found to have colon cancer incidentally during a regular colon cancer screening with colonoscopy at M12. Additionally, four patients developed intermittent episodes of mild asymptomatic hypercalcemia (Ca >2.55 mmol/l) during the first 3 months, which normalized quickly after reduction of calcium supplementation (Table 2). All adverse events were reported to the manufacturer. Compliance for teriparatide pen use was >90% over 18 months.
BMD and Biochemical Markers of Bone Turnover
Lumbar spine BMD steadily increased from baseline to M18 (Table 2). Mean BMD values after 18 months reached osteopenic T scores of −2.1 SD (P ≤ 0.001). Total-hip BMD insignificantly increased during the time of the study.
Bone formation markers (bone ALP) reached significant increases as early as 1 month after PTH treatment, with maximum levels at M6. Bone resorption markers (S-CTX) presented an equivalent but slightly protracted progression, reaching maximum levels at M6 (Table 2). Following M6 measurements, both biochemical markers of bone turnover steadily declined until M18 measurements and leveled off at M3 values.
Three patients did not show a response in biochemical markers of bone turnover, which was not correlated to a nonresponse in histological bone turnover and vice versa.
Intratrabecular Tunneling and Trabecular Heterogeneity
Mean tissue density, measured by μCT in milligrams of hydroxyapatite per bone tissue volume (mgHA/ccm), was significantly (P < 0.001) decreased at M18 (Ø −50 mgHA/ccm) compared with mean baseline values (Table 3), which were all within normal range following long-term BP use. The lowest values in recently mineralized remodeling sites ranged 750–850 mgHA/ccm. Despite a clear increase in bone remodeling units and extent of osteoid surface seams, no mineralization defects were observed in the histological investigation.
The new bone produced under PTH treatment had an increase in osteocyte density (Fig. 4) but normal lamellar structure and normal mineralization characteristics when observed histologically. Seven biopsies showed minor forms of local peritrabecular fibrosis but no signs of woven bone. The number of biopsies with microcallus formations was reduced from five (baseline) to two and three (M6 and M18), respectively.
The goal of our 18-month clinical trial was to investigate the effects of PTH on different biological levels in patients with osteoporosis previously treated with oral BPs and showing inadequate clinical response. There is no generally accepted definition for inadequate response to antiresorptive treatment, but unsatisfactory clinical outcome as measured by fragility fracture report or BMD decline >3% per year at the hip or lumbar spine have been previously used [11, 17, 30–33]. In accordance with previously published data, the results of our study indicate a quick anabolic bone marker response to teriparatide within 1 month for bone formation and within 3 months for bone resorption [11, 12, 17, 34]. Regardless of the type of BP used previously, bone markers reached a plateau at M6 and remained 100% (S-CTX) and ~30% (bone ALP) above baseline values. The fact that previous studies did not specifically include inadequate BP-treatment responders may explain our results that did not show statistically significant within-group differences between ALN and RIS users . Only 12% of the patients had no response in bone markers during the study period . Semiquantitative histological bone turnover presented a similar osteoanabolic trend, with 75% showing a strong increase of estimated bone remodeling units after 6 months, followed by a moderate decrease in 57% after 18 months.
Statistically significant differences with a positive development in most parameters were assessed at 6 months with the exception of spine BMD, which showed increases until 12 months whereas a plateau was reached in a majority of structural parameters or serum markers following 6 months of treatment [8, 15, 20, 35]. In contrast to the EUROFORS subgroup analysis , our study did not detect potentially deleterious effects on cortical bone as measured by stable total-hip BMD over the 18-month study period.
Although not particularly investigated in this trial due to the limited patient number, a positive correlative trend of BMD, BV/TV, and bone turnover markers could be established, as stated by others previously [11, 17, 36].
A trabecular reorganization from rod-like to more plate-like structures (significantly smaller SMI values) and increases in Tb.N with increases in trabecular tunneling could be observed . To comprehend the tunneling resorption and simultaneous decreasing SMI values, one must closely investigate the form and direction of resorption sites. In contrast to perforating resorptions that lead to loss of trabecular connectivity, tunneling resorption under PTH treatment, as observed here, moved along the long axis of plates and rods . A number of biopsies showed intense trabecular reorganization with a dense but heterogeneous structure expressed by Tb.Sp.SD (Fig. 5). Still, a deleterious effect of PTH on biomechanical stability was not observed in the small number of finite element analyses performed [37, 38].
Increases in BV/TV of >2% compared to baseline were observed in 68% (13/19) at M6 with a mean gain of 34%. At M18, the architectural parameter Conn.D increased once more significantly, indicating a higher trabecular connectivity, whereas the other parameters, including BV/TV, remained stable at levels above M6. Although not significant, 43% showed additional increases in bone volume (>2%) at M18 vs. M6 (Fig. 3).
For the first time, using a nondestructive μCT technique in a larger human biopsy study, we confirmed a significant decrease in mean tissue mineralization, most likely due to the newly formed bone matrix. Phantom validation studies have shown significant correlations with other destructive methods such as ash weight measurements [28, 29]. Thus far, mean tissue density, also referred to as bone mineralization density distribution, has generally been investigated by quantitative backscattered electron imaging (qBEI) [21, 39]. The advantage of μCT vs. qBEI is the possibility of a total biopsy scan, whereas the setup of qBEI only allows the investigation of a relatively small bone area.
There are several limitations to this study. Firstly, The absence of a control group has to be noted. However, the intention to treat with teriparatide was to investigate PTH clinical and microarchitectural responses in BP-resistant patients with ongoing fractures or BMD decline; thus, an untreated placebo group was no option in this cohort. Certainly, due to the size of the study, antifracture efficacy could not be investigated and beneficial effects could only be discussed based on macro- and microarchitectural changes. Secondly, the design of the study with consecutive bone biopsies always limits the number of patients available.
In contrast to a number of single-biopsy studies [7, 35], the paired-biopsy design allows use of baseline biopsies as a control. The minimally invasive type of biopsy used in this trial made the triple biopsy feasible and justifiable. Although the in-house experience with the usage of the Jamshidi technique for cellular and morphological diagnostics in metabolic bone diseases is large and used by others as well, data have not been documented in the literature until today. In experienced hands, Jamshidi biopsies are of excellent quality and succeed to the standard required for static histomorphometry . The total cylindrical volume of Jamshidi biopsies (Fig. 2) equates to the cancellous bone volume in bicortical Bordier biopsies without the limiting factor of bilateral subcortical bone heterogeneity . Still, intraindividual variability in structural parameters has long been known and must always be considered [41, 42]. The triple biopsy, as performed by us, also reduces bias based on biopsy location. The previously generously approximated number of at least 15 available biopsies at every time point was reached.
The influence of a vanishing BP effect on bone turnover markers after withdrawal and its potentiating effect on PTH efficacy has to be considered as well. Bone turnover markers have been found to remain reduced for several years after discontinuation of ALN, whereas others found bone turnover markers to rise within weeks and return to control group levels within 12 months of discontinuation of RIS but may remain at a level below normal turnover for months or years [1, 43–46]. Our study confirmed in a selected subclass of pretreated osteoporotic patients with insufficient response to BP treatment no significant delay in treatment response to PTH and may therefore be a valuable alternative treatment option. The mid-normal values of BTM at M0 furthermore underline inefficient BP effects in our cohort.
In conclusion, teriparatide therapy in patients with inadequate response to prior BP treatment produced quick increases in BMD and bone markers and improvement in bone microarchitecture within the first 6 months. The anabolic PTH effect on μCT parameters was only moderate or slightly reversed between 12 and 18 months of therapy. Furthermore, this is the first study using the Jamshidi biopsy technique at three time points that concluded changes in trabecular bone volume and trabecular structural parameters at the iliac crest comparable with the results acquired by conventional bicortical Bordier biopsy. Bone tissue mineralization as measured by μCT decreased continuously over the 18-month study period, remaining in the normal mineralization range.
Future studies may investigate the effects of shorter cyclic therapies (e.g., 6 months) of PTH followed by antiresorptive medications in severe osteoporotic patients  to optimize the therapeutic effects given the limited maximum treatment duration of PTH analogues and high costs of this medication.
We thank all of the patients who participated in this study. The study was supported by Lilly, Germany. We also greatly appreciate bone biopsy preparations by members of the former Institut of Bone Pathology, University of Hamburg-Eppendorf. B. J. received research funding for this study and travel support for ASBMR meetings from Lilly, Germany. However, the authors were fully responsible for all content and editorial decisions and received no financial support or other form of compensation related to the development of the report. The study sponsor was not involved in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit for publication.