In women who have not been on prior osteoporosis

Transcription

1 CLINICAL TRIALS JBMR Hip and Spine Strength Effects of Adding Versus Switching to Teriparatide in Postmenopausal Women With Osteoporosis Treated With Prior Alendronate or Raloxifene Felicia Cosman, 1,2 Tony M Keaveny, 3,4 David Kopperdahl, 4 Robert A Wermers, 5 Xiaohai Wan, 6 Kelly D Krohn, 6 and John H Krege 6 1 Helen Hayes Hospital, West Haverstraw, NY, USA 2 Department of Medicine, Columbia University College of Physicians and Surgeons, New York, NY, USA 3 Departments of Mechanical Engineering and Bioengineering, University of California, Berkeley, CA, USA 4 O.N. Diagnostics, LLC, Berkeley, CA, USA 5 Mayo College of Medicine, Mayo Clinic, Rochester, MN, USA 6 Eli Lilly and Company and/or one of its subsidiaries, Indianapolis, IN, USA ABSTRACT Many postmenopausal women treated with teriparatide for osteoporosis have previously received antiresorptive therapy. In women treated with alendronate (ALN) or raloxifene (RLX), adding versus switching to teriparatide produced different responses in areal bone mineral density (abmd) and biochemistry; the effects of these approaches on volumetric BMD (vbmd) and bone strength are unknown. In this study, postmenopausal women with osteoporosis receiving ALN 70 mg/week (n ¼ 91) or RLX 60 mg/day (n ¼ 77) for 18 months were randomly assigned to add or switch to teriparatide 20 mg/day. Quantitative computed tomography scans were performed at baseline, 6 months, and 18 months to assess changes in vbmd; strength was estimated by nonlinear finite element analysis. A statistical plan specifying analyses was approved before assessments were completed. At the spine, median vbmd and strength increased from baseline in all groups (13.2% to 17.5%, p < 0.01); there were no significant differences between the Add and Switch groups. In the RLX stratum, hip vbmd and strength increased at 6 and 18 months in the Add group but only at 18 months in the Switch group (Strength, Month 18: 2.7% Add group, p < 0.01 and 3.4% Switch group, p < 0.05). In the ALN stratum, hip vbmd increased in the Add but not in the Switch group (0.9% versus 0.5% at 6 months and 2.2% versus 0.0% at 18 months, both p group difference). At 18 months, hip strength increased in the Add group (2.7%, p < 0.01) but not in the Switch group (0%); however, the difference between groups was not significant (p ¼ 0.076). Adding or switching to teriparatide conferred similar benefits on spine strength in postmenopausal women with osteoporosis pretreated with ALN or RLX. Increases in hip strength were more variable. In RLX-treated women, strength increased more quickly in the Add group; in ALN-treated women, a significant increase in strength compared with baseline was seen only in the Add group. ß 2013 American Society for Bone and Mineral Research. KEY WORDS: TERIPARATIDE; RALOXIFENE; ALENDRONATE; POSTMENOPAUSAL OSTEOPOROSIS; FINITE ELEMENT ANALYSIS Introduction In women who have not been on prior osteoporosis medications, the effect of teriparatide to increase bone mineral density (BMD) and to reduce fracture risk in postmenopausal women at both vertebral and nonvertebral skeletal sites is well documented. (1 4) However, a significant proportion of teriparatide use is in patients who have been treated with antiresorptive drugs. In these patients, there is less information regarding fracture efficacy. (5,6) In a study of drug-naïve and antiresorptive-pretreated patients given teriparatide for 24 months, (7) the BMD increase from baseline was statistically significant at all sites; however, the increment in spine BMD was significantly greater in the drugnaïve patients. The time course for increase in hip BMD was different, with early increases in hip BMD being greater in the Received in original form September 19, 2012; revised form December 3, 2012; accepted December 10, Accepted manuscript online December 21, Address correspondence to: Felicia Cosman, MD, Helen Hayes Hospital, Route 9W, West Haverstraw, NY 10993, USA. cosmanf@helenhayeshosp.org or John H Krege, MD, Eli Lilly and Company, Lilly Corporate Center, Drop Code 4109, Indianapolis, IN 46285, USA. kregejh@lilly.com Journal of Bone and Mineral Research, Vol. 28, No. 6, June 2013, pp DOI: /jbmr.1853 ß 2013 American Society for Bone and Mineral Research 1328

2 drug-naïve than in the antiresorptive-pretreated groups. However, the 24-month increments in BMD at the hip in the antiresorptive-pretreated groups were approximately twice the increments seen after 18 months of therapy. Some studies have suggested that previous treatment with less potent antiresorptive drugs has less effect on the BMD response to teriparatide than more potent antiresorptive drugs, (8,9) whereas other studies have suggested no difference. (10) In patients previously treated with antiresorptive drugs, two approaches are available when initiating teriparatide: stopping the antiresorptive agent when teriparatide is initiated (the Switch approach), (7 10) or continuing the antiresorptive agent when teriparatide is initiated (the Add approach). (11,12) To compare these approaches in a formal head-to-head study, we randomized women on antiresorptive therapy, either prior alendronate (ALN) or prior raloxifene (RLX), to either stop their antiresorptive therapy and switch to teriparatide (the Switch groups) or to stay on their antiresorptive therapy and add teriparatide (the Add groups). From this trial, we previously reported that no clinically meaningful increases in mean predose serum calcium were observed in the Add or the Switch groups. (13) In addition, we have reported that greater bone turnover increases were achieved in the Switch groups, whereas greater areal BMD (abmd) increases were achieved in the Add groups after 18 months. (14) The differences were more pronounced in the ALN- than in the RLX-pretreated patients. Additionally, fewer adverse events were reported in the Add than in the Switch groups. (14) Here, we report data from quantitative computed tomography (QCT) scans obtained in this same study. Under some circumstances, changes in dual-energy X-ray absorptiometry (DXA) abmd measurements, which average cortical and trabecular bone responses, may not accurately reflect changes in overall whole-bone strength, particularly at the hip. (15,16) Accordingly, the current study was conducted using QCT scans acquired from our previous clinical study to determine estimated bone strength responses to teriparatide in women previously treated with RLX and ALN. Our prespecified objectives related to spine and hip after 6 and 18 months of treatment with teriparatide in postmenopausal women previously treated with RLX or ALN were: 1) to determine if volumetric BMD (vbmd) and strength increased from baseline in the Add and Switch groups, and 2) to compare vbmd and strength changes in the Add versus Switch groups. Materials and Methods Design overview Methods for this open-label study, which was conducted in the United States to assess the effects of adding versus switching to teriparatide 20 mg/day in patients pretreated with ALN (total of 70 mg/week) or RLX 60 mg/day for at least 18 months, have been reported. (13,14) The primary objective was to examine the change in mean serum calcium values during the first 6 months of treatment with teriparatide in the Add and Switch groups, and those data have been published. (13) The treatment phase was followed by a 12-month extension that was prespecified in the protocol to compare the effects of 18 months of teriparatide therapy on abmd, DXA, and vbmd at the hip and spine in the Add and Switch groups. Data on abmd have been published. (14) Here, we report on vbmd data as assessed by QCT, which provides 3-dimensional images to observe changes in volumetric density for both trabecular and cortical bone; we also report on estimates of whole-bone strength from finite element analysis (FEA) of those QCT scans. All scans were analyzed by O.N. Diagnostics (Berkeley, CA, USA) while blinded to treatment assignment (Add/Switch) and prior antiresorptive therapy but not to the order of the scans. Patients The patient population has been previously described. (13,14) In brief, women had to be at least 50 years old and to have a prior diagnosis of osteoporosis based on fracture history and/or BMD. Women were excluded if they had a history of hypercalcemia (except for surgically corrected hyperparathyroidism) or metabolic bone diseases other than osteoporosis; secondary causes of osteoporosis or malignant neoplasms within the past 5 years; active urolithiasis within the past 2 years or at high risk for urolithiasis in the opinion of the investigator; active liver disease; substantially impaired renal function; history of excessive alcohol consumption; or treatment with other bone-active drugs. Women with evaluable baseline and at least one post-baseline QCT scan were included in the present analysis. All women provided written informed consent, and an institutional review board approved the protocol at each study center. The study was conducted in accordance with the ethical standards of the Declaration of Helsinki. Treatments Women were entered into either a prior ALN stratum or a prior RLX stratum based on their previous antiresorptive therapy (Fig. 1). They continued their antiresorptive regimen during a 2- month antiresorptive phase and were then randomized (1:1) by computer program to either stop or continue antiresorptive treatment. All women initiated teriparatide 20 mg/day by subcutaneous injection. Patients were required to have been taking stable calcium supplementation (at least 500 mg/day elemental calcium) for at least 1 month to enter the study. Patients were required to maintain their existing stable calcium regimen (at least 500 mg/day of elemental calcium) and vitamin D supplementation for the duration of the study. Assessments QCT scans of the lumbar spine and hip were performed 1 month before randomization and then again after 6 and 18 months of therapy with teriparatide. A standard anthropomorphic spine phantom was used to correct for scanner drift at each study site. Variables of interest were: 1) vbmd measured through QCT, and 2) hip and spine strength estimated by FEA of QCT scans. We also provided measures of both QCT-density and strength for the trabecular and peripheral bone. The protocol specified that investigative sites would not change CT scanners during the conduct of the study. Even so, some study sites did change CT scanners, leading the central reader (Bio-Imaging Technologies, Inc., Newtown, PA, USA) to conclude that the QCT data were not reliable before database Journal of Bone and Mineral Research TERIPARATIDE IN WOMEN ON PRIOR ANTIRESORPTIVES 1329

3 Fig. 1. Study schematic. QCT ¼ quantitative computed tomography. lock. (14) However, analysis by O.N. Diagnostics of the CT scans from study sites that changed CT scanners revealed that most patients completed both the baseline and follow-up visits either before or after the machine was changed, and only 14 patients had a machine change between baseline and follow-up visits. O.N. Diagnostics implemented machine-specific field uniformity correction factors for all CT machines in the study, and analyses including and excluding the 14 patients with machine changes showed similar results. O.N. Diagnostics excluded scans that had metal artifacts, appreciable imaging artifacts (patient motion, obvious streaking), and all scans for patients who lacked an evaluable baseline CT scan. Concentrations of serum procollagen type 1 N-terminal propeptide (Total P1NP Assay, Roche Diagnostics, Mannheim, Germany), a biochemical marker of bone formation, were measured at baseline, month 1, and month 3. FEA specification Finite element analysis has shown improved statistical power to detect treatment effects compared with both DXA and QCT, and results of FEA have been shown to be statistically similar to measurements obtained directly from biomechanical testing of cadaver bones. (17) L 1 vertebral strength for a simulated compression overload and femoral strength for a simulated sideways fall were estimated for each patient, at each time point, using nonlinear 3-dimensional FEA of the QCT scans. All FEAs were performed by O.N. Diagnostics as described previously for longitudinal studies. (15,16,18,19) In living patients, these estimates of strength have been shown to be predictive of new clinical spine and hip fractures, prevalent radiographic spine, and all clinical fractures. (20 25) Volumetric BMD and strength measures for the trabecular and peripheral bone were also determined. For the hip, we defined a peripheral compartment as all the cortical bone (calibrated apparent density >1.0 g/cm 3 ) plus any trabecular bone within 3 mm of the periosteal surface; the trabecular compartment was defined as all remaining trabecular bone. (15) Hip trabecular strength was computed by turning all the bone in the peripheral compartment at each time point into plastic (a cortical plastic and an endosteal trabecular plastic) and then reloading to failure. Hip peripheral strength was computed by turning all the bone in the trabecular compartment into plastic (an internal trabecular plastic) and then reloading to failure. (15) For the spine, the trabecular compartment was defined as all the bone less the outer 2 mm of bone (considered peripheral bone). Spine trabecular strength was computed by performing a new FEA on only the trabecular compartment; (18) the spine peripheral strength was then calculated as the overall vertebral strength minus this trabecular strength. These trabecular and peripheral strengths are additive for the spine but not for the hip. This added complexity for the hip is because of the complex nonparallel load-transfer paths in the hip between the trabecular and peripheral compartments, and thus it is not reasonable biomechanically to simply remove the peripheral bone when computing a trabecular strength for the femur and vice versa. Statistical analyses A statistical analysis plan was approved before the FEA assessment of the QCT scans. No values were imputed for missing QCT data. The statistical analyses population was defined to include patients with baseline and at least one postbaseline follow-up QCT. Summary statistics were calculated for 1330 COSMAN ET AL. Journal of Bone and Mineral Research

4 change and percent change from baseline at 6 and 18 months in BMD, strength, and a strength/density ratio at both the spine and hip. For each treatment group, Wilcoxon signed-rank tests were applied to compare baseline to post-baseline for all continuous outcomes. Within each pretreatment stratum, Wilcoxon ranksum tests were used to compare continuous outcomes between the Add and Switch groups. The relationship between change from baseline at month 1 and month 3 in bone biomarker P1NP and percent change from baseline at 18 months in spine strength in each of the four treatment groups was assessed using Pearson correlation coefficients and associated significance tests. All analyses used SAS Drug Development software (SAS Institute, Cary, NC, USA). Results Baseline characteristics of the groups were similar (Table 1). Serum P1NP, a biochemical marker of bone turnover, was similar between Add and Switch groups within each stratum but was lower in the ALN- compared with the RLX-pretreated women, which is consistent with the bisphosphonate drug ALN being a more potent antiresorptive agent than the selective estrogen receptor modulator RLX. Additional baseline characteristics are provided in earlier publications from this study. (13,14) Volumetric BMD and strength in the spine (Table 2 and Fig. 2A, B) At the spine, integral vbmd and vbmd of individual compartments of trabecular and peripheral bone increased in all groups at the 6- and 18-month time points. Increases at 18 months were greater than those at 6 months in all groups. At 18 months, integral vbmd increased 7.5% and 7.9% in the ALN Add and Switch Groups, respectively, and 10.1% and 10.0% in the RLX Add and Switch groups, respectively (all p < 0.01 versus respective baselines). In the RLX stratum at 6 months, the increase in peripheral vbmd was greater in the Add versus the Switch group, but this difference was no longer significant at 18 months. There were no other significant differences between adding and switching to teriparatide in either the RLX or ALN stratum. Increases in trabecular vbmd appeared to exceed those of peripheral vbmd at both 6- and 18-month time points in all groups. Strength changes in the spine followed a pattern similar to the changes in vbmd. In all four groups, significant increases in strength were seen, with no persistent differences between the Add and Switch groups by 18 months for the integral, trabecular, or peripheral compartments. Increases in spine strength were greater at 18 versus 6 months for both groups but were particularly marked in the ALN groups. At 18 months, spine strength increased 13.2% and 15.6% in the ALN Add and Switch groups, respectively, and 17.5% and 15.7% in the RLX Add and Switch groups, respectively (all p < 0.01 versus respective baselines, no group differences). Volumetric BMD and strength in the hip (Table 3 and Fig. 2C, D) In the RLX stratum, integral hip vbmd increased at 6 months in the Add but not in the Switch group, though integral vbmd increases were similar in the Add and Switch groups by 18 months. At 6 months, increases in trabecular vbmd were seen in both Add and Switch groups, whereas peripheral vbmd declined (Switch group) or did not change significantly (Add group). At 18 months, both trabecular and peripheral vbmd increased, though increases in trabecular vbmd were consistently greater than those in peripheral vbmd. Integral vbmd increased 3.2% and 2.4% in the Add and Switch groups, respectively, at 18 months (both p < 0.01 versus respective baselines, no group difference). In the ALN stratum at 6 months, integral hip vbmd increased in the Add but not in the Switch group. Trabecular vbmd increased Table 1. Baseline Characteristics of Postmenopausal Women With Osteoporosis With Baseline and Post-Baseline QCT Scans RLX-pretreated stratum ALN-pretreated stratum Add group Switch group Add group Switch group Variable (n ¼ 43) (n ¼ 44) (n ¼ 46) (n ¼ 49) Age (years) Body mass index (kg/m 2 ) Duration of prior treatment (months) T-score Lumbar spine Femoral neck Total hip Spine vbmd (mg/cm 3 ) Spine strength (newtons) Hip vbmd (mg/cm 3 ) Hip strength (newtons) P1NP (ng/ml) [ ] ALN ¼ alendronate; P1NP ¼ serum procollagen type 1 N-terminal propeptide; QCT ¼ quantitative computed tomography; RLX ¼ raloxifene; vbmd ¼ volumetric bone mineral density. Continuous variables are shown as least squares mean SE, except for P1NP, which is shown as median values. Categorical variables are shown as number (%). Reference range for P1NP measurements is given in brackets, from Cosman and colleagues. (14) Journal of Bone and Mineral Research TERIPARATIDE IN WOMEN ON PRIOR ANTIRESORPTIVES 1331

5 Table 2. Median Change (%) From Baseline (25th, 75th IQ Range) for vbmd and Strength of the Spine in Postmenopausal Women With Osteoporosis Previously Treated With RLX or ALN Then Randomized to Add or Switch to TPTD Outcome RLX þ TPTD (Add group) RLX! TPTD ALN þ TPTD (Switch group) p Value a (Add group) ALN! TPTD (Switch group) p Value a Volumetric BMD Month 6 n ¼ 38 n ¼ 39 n ¼ 45 n ¼ 46 Integral 7.0 (3.8, 9.7) 5.7 (1.3, 7.8) (2.1, 6.8) 3.0 (0.8, 5.7) 0.18 Trabecular 8.7 (4.8, 14.3) 6.6 (1.3, 10.8) (0.9, 7.0) 4.0 (1.1, 7.7) 0.56 Peripheral 5.2 (2.4, 7.5) 3.7 (0.6, 6.2) (1.3, 6.0) 1.8 y ( 0.9, 5.6) 0.42 Month 18 n ¼ 29 n ¼ 25 n ¼ 40 n ¼ 40 Integral 10.1 (6.7, 16.0) 10.0 (6.3, 13.0) (3.7, 12.7) 7.9 (5.3, 10.7) 0.91 Trabecular 12.0 (6.7, 20.3) 12.4 (6.3, 16.0) (5.0, 15.9) 10.3 (5.9, 14.7) 0.57 Peripheral 8.6 (6.1, 10.7) 7.2 (4.7, 11.8) (2.5, 9.0) 5.2 (1.7, 8.1) 0.36 Strength Month 6 n ¼ 38 n ¼ 39 n ¼ 45 n ¼ 46 Integral 11.4 (6.4, 17.4) 9.7 (2.3, 13.6) (1.7, 12.4) 7.1 (1.8, 12.6) 0.95 Trabecular 11.5 (6.9, 21.9) 11.2 (0.9, 16.7) (2.1, 11.7) 7.0 (1.9, 14.9) 0.70 Peripheral 9.5 (6.3, 13.9) 7.8 (2.7, 12.1) (2.4, 11.2) 5.7 (1.2, 9.4) 0.49 Month 18 n ¼ 29 n ¼ 25 n ¼ 40 n ¼ 40 Integral 17.5 (11.8, 27.2) 15.7 (10.5, 25.6) (7.2, 21.4) 15.6 (10.1, 20.7) 0.33 Trabecular 18.2 (8.1, 34.2) 18.0 (12.1, 29.3) (6.2, 23.2) 17.3 (9.9, 26.2) 0.15 Peripheral 16.8 (12.3, 25.2) 13.9 (9.2, 21.1) (6.8, 17.9) 13.6 (7.7, 19.0) 0.74 Strength/density ratio b Month (2.3, 7.4) 3.1 (0.7, 5.5) (1.0, 5.4) 3.3 (1.1, 5.8) 0.21 Month (4.5, 9.8) 5.2 (2.6, 11.0) (2.3, 7.7) 6.5 (4.2, 10.3) ALN ¼ alendronate; RLX ¼ raloxifene; TPTD ¼ teriparatide; vbmd ¼ volumetric bone mineral density. Wilcoxon signed-rank test p < 0.01 percentage change from baseline; y p < 0.05 percentage change from baseline. a Between-group p value, Wilcoxon rank-sum test (bolded values are statistically significant). b Strength per unit measure of vbmd. slightly in both the Add and Switch groups, whereas peripheral vbmd declined in the Switch group and remained unchanged in the Add group. Between 6 and 18 months, integral vbmd increased in the Add but not in the Switch group. The difference in integral vbmd at 18 months between the Add and Switch groups was largely the result of the persistent decline in cortical vbmd in the Switch group compared with an increase in cortical vbmd in the Add group. At 18 months, integral vbmd increased by 2.2% in the Add group (p < 0.01 versus baseline) and 0% in the Switch group (p ¼ between-group difference). Strength changes in the hip followed a similar pattern to vbmd changes. In the RLX groups, integral strength increased at 6 months in the Add but not in the Switch group. However, by 18 months, both Add and Switch groups showed a similar increase in strength (2.7% and 3.4%, no group difference). In the ALN stratum at 6 months, there were no significant changes in either Add or Switch groups in integral strength, though in trabecular bone, strength declined significantly in the Switch group only. In the ALN stratum, by 18 months, integral strength increased by 2.7% (p < 0.01 versus baseline) in the Add Group and 0% in the Switch group (p ¼ between-group difference). Increments in hip strength were also noted in both trabecular and peripheral compartments in the Add but not in the Switch group at 18 months. Correlations between increments in serum PINP and spine strength Absolute changes from baseline in serum P1NP level at 1 and 3 months were correlated with percent change in spine strength at 18 months in all four groups (r ¼ 0.48 to 0.71, all p 0.01), Table 4. Discussion In postmenopausal women with osteoporosis treated with antiresorptive drugs, the clinical need may sometimes arise to initiate treatment with teriparatide. To our knowledge, this is the only clinical study of postmenopausal women with osteoporosis previously treated with RLX or ALN who were randomized to add teriparatide to their antiresorptive therapy or to switch to teriparatide. Here, we report findings on vbmd from QCT and estimated strength from FEA. For the spine, in patients previously treated with ALN or RLX, there were no differences in the observed increases in either integral vbmd or strength between the Add and Switch groups. For the RLX stratum, in the analysis of the hip, adding teriparatide to RLX resulted in an earlier increase in hip strength, but by 18 months there was no significant difference in the improvement in integral hip vbmd or 1332 COSMAN ET AL. Journal of Bone and Mineral Research

6 Fig. 2. Effects of adding versus switching to teriparatide in postmenopausal women with osteoporosis pretreated with raloxifene or alendronate, including (A) spine vbmd at 6 and 18 months, (B) spine strength at 6 and 18 months, (C) hip volumetric BMD at 6 and 18 months, and (D) hip strength at 6 and 18 months. Note that statistically significant changes from baseline are indicated by footnote symbols, and p values from statistical comparisons between the Add and Switch groups in each stratum are shown above each pair of bars. strength in those who continued versus those who stopped RLX. In contrast, in the ALN stratum, hip vbmd increased from baseline at both 6 and 18 months only in the Add group and hip strength increased only in the Add group. With switching from ALN to teriparatide, the lack of an increase in integral hip vbmd was a result of a statistically significant decrease in peripheral vbmd, which may have been the result of an increase in cortical remodeling (porosity). (26) In general, trends for changes were similar for measures of vbmd and mass (data not presented), suggesting that there were no volume changes associated with either treatment arm. Long-term ALN pretreatment may have resulted in bone with low remodeling space and relatively high mineralization. Activation of new bone metabolic units by teriparatide reopens the remodeling space and causes bone with higher mineralization to be removed and replaced with new, less mineralized bone. (8) Higher levels of cortical remodeling are seen in those women who switch versus those who add teriparatide to ongoing ALN, consistent with our observation of significantly greater bone turnover when switching to teriparatide. (14) Higher bone turnover associated with switching to versus adding teriparatide may result in relatively greater replacement of older bone with newer bone, a potential mechanism of improved bone quality, which was not captured by the FEA BMD-assessed bone strength determinations. Spine strength improvements with teriparatide were correlated to absolute changes in the bone formation marker P1NP at both 1 and 3 months in all groups. This information demonstrates that increases in serum P1NP with teriparatide treatment predict later increases in strength in patients on prior antiresorptive therapy. This relationship has also been demonstrated in osteoporosis drug-naïve patients. (27) As limitations, the study duration was 18 rather than 24 months. This is important because studies have shown that Journal of Bone and Mineral Research TERIPARATIDE IN WOMEN ON PRIOR ANTIRESORPTIVES 1333

7 Table 3. Median Change (%) From Baseline (25th, 75th IQ Range) for vbmd and Strength of the Hip in Postmenopausal Women With Osteoporosis Previously Treated With RLX or ALN Then Randomized to Add or Switch to TPTD Outcome RLX þ TPTD (Add group) RLX! TPTD ALN þ TPTD (Switch group) p Value a (Add group) ALN! TPTD (Switch group) p Value a Volumetric BMD Month 6 n ¼ 40 n ¼ 36 n ¼ 42 n ¼ 46 Integral 1.3 ( 0.5, 3.0) 0.4 ( 1.0, 1.9) (0.0, 1.7) 0.5 ( 2.1, 0.9) Trabecular 2.9 (1.0, 4.9) 1.7 ( 0.3, 4.1) (0.0, 2.5) 1.0 y ( 0.8, 3.5) 0.77 Peripheral 0.3 ( 2.6, 2.1) 0.9 y ( 3.2, 0.9) ( 1.3, 1.7) 1.8 ( 3.9, 0.6) Month 18 n ¼ 30 n ¼ 24 n ¼ 36 n ¼ 39 Integral 3.2 ( 0.5, 5.7) 2.4 (1.0, 4.0) (0.7, 4.7) 0.0 ( 1.7, 2.6) Trabecular 4.9 (0.0, 7.8) 3.5 (1.2, 6.5) (1.1, 5.6) 2.8 (0.0, 4.5) 0.44 Peripheral 1.2 ( 2.1, 3.7) 1.0 ( 1.2, 3.1) (0.5, 3.0) 2.2 ( 4.1, 0.3) < Strength Month 6 n ¼ 40 n ¼ 36 n ¼ 42 n ¼ 46 Integral 1.7 ( 0.5, 3.1) 1.1 ( 2.9, 3.4) ( 2.4, 1.8) 1.6 ( 3.9, 1.9) 0.10 Trabecular 1.3 ( 0.6, 3.4) 0.3 ( 2.3, 2.0) ( 3.2, 1.6) 1.8 y ( 4.6, 1.9) 0.14 Peripheral 0.6 ( 1.0, 2.1) 1.1 ( 2.2, 0.5) ( 1.1, 1.3) 1.1 ( 2.5, 1.4) Month 18 n ¼ 30 n ¼ 24 n ¼ 36 n ¼ 39 Integral 2.7 ( 1.1, 6.2) 3.4 y ( 1.5, 5.8) ( 0.5, 6.1) 0.0 ( 3.4, 4.7) Trabecular 0.7 ( 0.9, 4.1) 2.6 ( 0.9, 4.5) (0.7, 5.6) 0.1 ( 2.8, 4.4) Peripheral 2.0 (0.0, 3.8) 1.9 y ( 0.9, 3.2) (0.0, 3.2) 0.2 ( 1.8, 2.6) Strength/density ratio b Month ( 1.6, 2.4) 0.8 ( 2.6, 1.9) y ( 2.0, 0.6) 0.8 ( 2.5, 0.9) 0.79 Month ( 1.2, 1.8) 0.9 ( 2.8, 2.6) ( 1.7, 2.1) 0.4 ( 2.1, 2.7) 0.84 ALN ¼ alendronate; RLX ¼ raloxifene; TPTD ¼ teriparatide; vbmd ¼ volumetric bone mineral density. Wilcoxon signed-rank test p < 0.01 percentage change from baseline; y p < 0.05 percentage change from baseline. a Between-group p value, Wilcoxon rank-sum test (bolded values are statistically significant). b Strength per unit measure of vbmd. patients switched from antiresorptive treatment to teriparatide may have relatively large increases in BMD and bone strength between 18 and 24 months of treatment, specifically at cortical bone-rich sites, probably reflecting the filling of the cortical remodeling spaces. (7,28) Thus, if the current study had continued to 24 months, there may have been significant catch-up in hip vbmd and estimated strength in the ALN Switch group. This possibility suggests that a full 24-month course of treatment may be especially important in patients who switch from antiresorptive treatment to teriparatide. Furthermore, this study did not include follow-up after cessation of teriparatide treatment, during subsequent antiresorptive treatment; additional changes in vbmd and bone strength might occur owing to closure of the remodeling space. Additionally, FEA did not capture the effects Table 4. Pearson Correlations Between Absolute Change in P1NP and Percent Change in Spine Strength Variable RLX þ TPTD (Add) RLX! TPTD (Switch) ALN þ TPTD (Add) ALN! TPTD (Switch) Month 1 Absolute change P1NP at 1 month (SD) (30.286) (40.695) (36.161) (37.991) Mean % change spine strength at 18 months (SD) (14.389) (12.175) (11.954) (12.849) Pearson r Pearson p value Month 3 Absolute change P1NP at 3 months (SD) (61.568) (52.549) (41.036) (76.877) Mean % change spine strength at 18 months (SD) (14.389) (12.175) (11.954) (12.849) Pearson r Pearson p value ALN ¼ alendronate; P1NP ¼ serum procollagen type 1 N-terminal propeptide; RLX ¼ raloxifene; SD ¼ standard deviation; TPTD ¼ teriparatide COSMAN ET AL. Journal of Bone and Mineral Research

8 of prior antiresorptives on bone tissue material properties, and the study lacked power to compare fracture outcomes. Finally, our results for the ALN stratum might not be applicable to other bisphosphonates. In conclusion, in RLX-treated women with osteoporosis, continuing RLX at the time teriparatide was started provided an earlier but ultimately similar increase in hip strength by 18 months. For women previously treated with ALN, adding and switching to teriparatide conferred similar increases in spine vbmd and estimated strength. Therefore, for patients in whom the major risk of fracture appears to be the spine, adding versus switching to teriparatide would be anticipated to result in a roughly similar outcome. However, this study suggests there is uncertainty about the best approach for optimizing hip outcome. In the ALN stratum at 18 months, hip vbmd and strength increased significantly in the Add but not in the Switch group, with the between-group difference being statistically significant for vbmd but not statistically significant for strength. In the ALN Switch group, the lack of improvement in these parameters appears to be the result of greater cortical remodeling, a phenomenon that might ultimately improve cortical bone material properties (by replacing more old bone with new bone). Additionally, vbmd and strength of the hip might improve between 18 and 24 months in patients who are switched from ALN to teriparatide, highlighting the importance of a full treatment course in this clinical setting. However, for women who have been treated with prior ALN (or perhaps other oral bisphosphonates) who appear to be at particularly high risk of hip fracture, such as those who have hip fractures while on ALN or those who have very low hip BMD, adding teriparatide to ongoing bisphosphonate therapy is a therapeutic option that our data suggest will improve hip strength. Disclosures FC has served on the advisory board at Eli Lilly & Co., Merck, and Amgen Inc.; has been a consultant for Eli Lilly & Co., Merck, Amgen, Novartis, Unigene Laboratories, Inc., and Tarsa Therapeutics; has received research support from Eli Lilly & Co., Novartis, and Amgen; and has served on the Speakers Bureau for Eli Lilly & Co., Novartis, and Amgen. TMK has received consulting fees from O.N. Diagnostics, Merck, and Wright Medical Technology; has received grants and/ or contracts from Amgen, GSK, J&J, Lilly, Merck, Novartis, Pfizer, and Wright Medical Technology; owns stock in O.N. Diagnostics; and has ownership interests related to work reported in this article. DK is an employee of O.N. Diagnostics. RAW has no conflicts of interest. XW, KDK, and JHK are employees of Lilly and own stock in the company. Acknowledgments This work was sponsored by Eli Lilly and/or one of its subsidiaries. The authors thank Eileen R Gallagher, a full-time employee of PharmaNet/i3, a part of the inventiv Health Company, for help with the writing of the manuscript. Authors roles: Study design: FC, XW, KDK, and JHK. Study conduct: DK and RW. Data collection: FC, DK, and RW. Data analysis: TMK, XW, and JHK. Data interpretation: FC, TMK, DK, RW, XW, KDK, and JHK. Drafting the manuscript: FC, TMK, XW, and JHK. Revising manuscript content: FC, TMK, DK, RW, XW, KDK, and JHK. Approving final version of manuscript: FC, TMK, DK, RW, XW, KDK, and JHK. XW takes responsibility for the integrity of the data analysis. References 1. Neer RM, Arnaud CD, Zanchetta JR, Prince R, Gaich GA, Reginster J-Y, Hodsman AB, Eriksen EF, Ish-Shalom S, Genant HK, Wang O, Mitlak BH. Effect of parathyroid hormone (1 34) on fractures and bone mineral density in postmenopausal women with osteoporosis. N Engl J Med May; 344(19): Prevrhal S, Krege JH, Chen P, Genant H, Black DM. Teriparatide vertebral fracture risk reduction determined by quantitative and qualitative radiographic assessment. Curr Med Res Opin Apr; 25(4): Krege JH, Wan X. Teriparatide and the risk of nonvertebral fractures in women with postmenopausal osteoporosis. Bone Jan; 50(1): McClung MR, San Martin J, Miller PD, Civitelli R, Bandeira F, Omizo M, Donley DW, Dalsky GP, Eriksen EF. Opposite bone remodeling effects of teriparatide and alendronate in increasing bone mass. Arch Intern Med Aug; 165(15): Jakob F, Oertel H, Langdahl B, Barrett A, Karras K, Walsh JB, Fahrleitner-Pammer A, Rajzbaum G, Barker C, Lems WF, Marin F. Effects of teriparatide in postmenopausal women with osteoporosis pre-treated with bisphosphonates: 36-month results from the European Forsteo Observational Study. Eur J Endocrinol Jan; 166(1): Hadji P, Zanchetta JR, Russo L, Recknow CP, Saag KG, McKiernan FE, Silverman SL, Alam J, Burge RT, Krege JH, Lakshmanan MC, Masica DN, Mitlak BH, Stock JL. The effect of teriparatide compared with risedronate on reduction of back pain in postmenopausal women with osteoporotic vertebral fractures. Osteoporos Int Aug; 23(8): Obermayer-Pietsch BM, Marin F, McCloskey EV, Hadji P, Farrerons J, Boonen S, Audran M, Barker C, Anastasilakis AD, Fraser WD, Nickelsen T. Effects of two years of daily teriparatide treatment on BMD in postmenopausal women with severe osteoporosis with and without prior antiresorptive treatment. J Bone Miner Res Oct; 23(10): Ettinger B, San Martin J, Crans G, Pavo I. Differential effects of teriparatide on BMD after treatment with raloxifene or alendronate. J Bone Miner Res May; 19(5): Miller PD, Delmas PD, Lindsay R, Watts NB, Luckey M, Adachi J, Saag K, Greenspan SL, Seeman E, Boonen S, Meeves S, Lang TF, Bilezikian JP. Early responsiveness of women with osteoporosis to teriparatide after therapy with alendronate or risedronate. J Clin Endocrinol Metab Oct; 93(10): Boonen S, Marin F, Obermayer-Pietsch B, Simões ME, Barker C, Glass EV, Hadji P, Lyritis G, Oertel H, Nickelsen T, McCloskey EV. Effects of previous antiresorptive therapy on the bone mineral density response to two years of teriparatide treatment in postmenopausal women with osteoporosis. J Clin Endocrinol Metab Mar; 93(3): Journal of Bone and Mineral Research TERIPARATIDE IN WOMEN ON PRIOR ANTIRESORPTIVES 1335

9 11. Cosman F, Nieves J, Zion M, Woelfert L, Luckey M, Lindsay R. Daily and cyclic parathyroid hormone in women receiving alendronate. N Engl J Med Aug; 353(6): Cosman F, Nieves JW, Zion M, Barbuto N, Lindsay R. Effect of prior and ongoing raloxifene therapy on response to PTH therapy. Osteoporos Int Apr; 19(4): Wermers RA, Recknor CP, Cosman F, Xie L, Glass EV, Krege JH. Effects of teriparatide on serum calcium in postmenopausal women with osteoporosis previously treated with raloxifene or alendronate. Osteoporos Int Jul; 19(7): Cosman F, Wermers RA, Recknor C, Mauck KF, Xie L, Glass EV, Krege JH. Effects of teriparatide in postmenopausal women with osteoporosis on prior alendronate or raloxifene: differences between stopping and continuing the antiresorptive agent. J Clin Endocrinol Metab Oct; 94(10): Keaveny TM, Hoffmann PF, Singh M, Palermo LK, Bilezikian JP, Greenspan SL, Black DM. Femoral bone strength and its relation to cortical and trabecular changes after treatment with PTH, alendronate, and their combination as assessed by finite element analysis of quantitative CT scans. J Bone Miner Res Dec; 23(12): Keaveny TM, McClung WR, Wan X, Kopperdahl DL, Mitlak BH, Krohn K. Femoral strength in osteoporotic women treated with teriparatide or alendronate. Bone Jan; 50(1): Keaveny TM. Biomechanical computed tomography non-invasive bone strength analysis using clinical CT scans. Ann NY Acad Sci Mar; 1192: Keaveny TM, Donley DW, Hoffman PF, Mitlak BH, Glass EV, San Martin JA. Effects of teriparatide and alendronate on vertebral strength as assessed by finite element modeling of QCT scans in women with osteoporosis. J Bone Miner Res Jan; 22(1): Lewiecki M, Keaveny TM, Kopperdahl D, Genant HK, Engelke K, Fuerst T, Kivitz A, Davies RY, Fitzpatrick LA. Once-monthly oral ibandraonate improves biomechanical determinants of bone strength in women with postmenopausal osteoporosis. J Clin Endocrinol Metab Jan; 94(1): Wang X, Sanyal A, Cawthon PM, Palermo L, Jekir M, Christensen J, Ensrud KE, Cummings SR, Orwoll E, Black DM. Osteoporotic Fractions in Men (MrOS) Research Group, Keaveny TM. Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans. J Bone Miner Res Apr; 27(4): Orwoll ES, Marshall LM, Nielson CM, Cummings SR, Lapidus J, Cauley JA, Ensrud K, Lane N, Hoffmann PR, Kopperdahl DL, Keaveny TM. Finite element analysis of the proximal femur and hip fracture risk in older men. J Bone Miner Res Mar; 24(3): Kopperdahl DL, Hoffmann P, Sigurdsson S, Aspelund T, Siggeirsdottir K, Eiriksdottir G, Harris T, Gudnason VG, Keaveny TM. Enhancement of hip fracture prediction using finite element analysis of CT scans. J Bone Miner Res. 2010;25(Suppl S1):S114. Abstract. 23. Melton LJ 3rd, Riggs BL, Keaveny TM, Achenbach SJ, Hoffman PF, Camp JJ, Rouleau PA, Bouxsein ML, Amin S, Atkinson EJ, Robb RA, Khosla S. Structural determinants of vertebral fracture risk. J Bone Miner Res Dec; 22(12): Melton LJ 3rd, Riggs BL, Keaveny TM, Achenbach SJ, Kopperdahl D, Camp JJ, Rouleau PA, Amin S, Atkinson EJ, Robb RA, Therneau TM, Khosla S. Relation of vertebral deformities to bone density, structure, and strength. J Bone Miner Res Sep; 25(9): Amin S, Kopperdahl DL, Melton LJ 3rd, Achenback SJ, Therneau TM, Riggs BL, Keaveny TM, Khosla S. Association of hip strength estimates by finite element analysis with fractures in women and men. J Bone Miner Res Jul; 26(7): Sato M, Westmore M, Ma YL, Schmidt A, Zeng QQ, Glass EV, Vahle J, Brommage R, Jerome CP, Turner CH. Teriparatide [PTH(1 34)] strengthens the proximal femur of ovariectomized nonhuman primates despite increasing porosity. J Bone Miner Res Apr; 19(4): Chen P, Glass EV, Krege JH. Early changes in bone turnover markers (BTMs) predict vertebral strength changes in teriparatide- or alendronate-treated postmenopausal women with osteoporosis [abstract]. In: ENDO 2007 Program & Abstracts. Toronto: Endocrine Society; June Borggrefe J, Graeff C, Nickelsen TN, Marin F, Gluer CC. Quantitative computed tomographic assessment of the effects of 24 months of teriparatide treatment on 3D femoral neck bone distribution, geometry, and bone strength: results from the EUROFORS Study. J Bone Miner Res Mar; 25(3): COSMAN ET AL. Journal of Bone and Mineral Research