Journal of Bone and Mineral Metabolism

, Volume 30, Issue 3, pp 326–337

Effects of teriparatide in Japanese and non-Japanese populations: bridging findings on pharmacokinetics and efficacy

Authors

    • Lilly Research Laboratories-Development Center of Excellence-Asia PacificEli Lilly Japan K.K.
  • Kazunori Uenaka
    • Lilly Research Laboratories-Development Center of Excellence-JapanEli Lilly Japan K.K.
  • Atsuko Iwata
    • Lilly Research Laboratories-Development Center of Excellence-Asia PacificEli Lilly Japan K.K.
  • Yoshihiro Higashiuchi
    • Lilly Research Laboratories-Development Center of Excellence-JapanEli Lilly Japan K.K.
  • Hideaki Sowa
    • Lilly Research Laboratories-Development Center of Excellence-JapanEli Lilly Japan K.K.
Original Article

DOI: 10.1007/s00774-011-0314-4

Cite this article as:
Tsujimoto, M., Uenaka, K., Iwata, A. et al. J Bone Miner Metab (2012) 30: 326. doi:10.1007/s00774-011-0314-4

Abstract

Teriparatide is an anabolic therapy for osteoporosis approved in the United States since 2002 and European Union since 2003; however, approval in Japan lagged significantly. This report describes analyses based on International Conference on Harmonisation (ICH) E-5 guidelines that support bridging between Japanese studies and the large Fracture Prevention Trial (FPT). We analyzed data from single teriparatide doses in healthy Japanese and Caucasian postmenopausal women (J-PK) and from studies of 6 months [Phase 2, dose ranging (J-Ph2)] and 12 months [Phase 3, efficacy and safety (J-Ph3)] of randomized, placebo-controlled, once-daily treatment in Japanese subjects with osteoporosis. In J-PK, apparent teriparatide area-under-the-curve (AUC) and peak concentration (Cmax) were up to 40% higher in Japanese versus Caucasian women; however, body weight-adjusted values were comparable between populations; these findings were supported by population pharmacokinetic analyses. Between the FPT and Japanese studies, baseline demographic characteristics were similar but bone mineral density (BMD) at lumbar spine (L1–L4) and body weight were lower for Japanese subjects. With teriparatide 20 μg/day, significant increases in BMD were observed compared to placebo at 12 months in both the FPT and J-Ph3 study, and percent change and actual change in BMD were comparable between studies. Dose response at 6 months was also comparable across populations. No novel safety signals were identified in Japanese subjects. These analyses show that teriparatide clinical data met ICH E-5 criteria for bridging. Findings from foreign trials such as the FPT can thus be extrapolated to Japanese subjects treated with teriparatide 20 μg/day.

Keywords

ICH E5 guidelineBridgingEthnic comparisonTeriparatideJapan

Introduction

Teriparatide (Forteo®) is the recombinant N-terminal fragment (residues 1–34) of human parathyroid hormone [rhPTH(1–34)]. Administered once-daily, teriparatide acts anabolically, leading to rapid increases in bone mineral density (BMD) and positive effects on bone microarchitecture that result in increased bone strength [15]. It is, therefore, particularly appropriate as a therapy for subjects with osteoporosis at high risk of fracture, where relatively rapid improvements in bone strength are urgently needed to reduce fracture risk. Teriparatide has shown potent activity in postmenopausal women [68], in men [9], and in men and women receiving glucocorticoid therapy [10], as evaluated using surrogate markers such as BMD and biological markers of bone metabolism. Teriparatide was also directly shown to reduce fracture risk in postmenopausal women in the large Fracture Prevention Trial (FPT) [7]. Thus, teriparatide is currently approved for use in over 80 countries and regions and has been in use in the United States since 2002 and in countries of the European Union since 2003.

Clinical development of drugs in Japan often lags behind that of other industrialized nations [11], and this has also been true for teriparatide, which received Japanese approval in July 2010. A strategy for addressing the lag in drug approval by bridging data from large foreign studies with smaller domestic studies is described in the International Conference on Harmonization (ICH) E-5 guidelines [12]. The E-5 guidelines address ethnic factors (i.e. extrinsic and intrinsic factors) in the acceptability of foreign clinical data and have supported the approval of numerous drugs in Japan [13]. In the case of osteoporosis treatments, extrinsic factors such as the definition and diagnostic criteria for osteoporosis, methods for the prevention and treatment for osteoporosis, and methods for evaluating the effectiveness of treatments for osteoporosis are all similar between Japanese and foreign populations. Thus, bridging strategies have already been used for raloxifene, alendronate, and risendronate [13]. Such strategies are particularly important for osteoporosis treatments because of the long duration of studies required to adequately address fracture risk, as well as the ethical dilemma raised by the need for long-term placebo treatment in subjects with osteoporosis at high risk of fracture in such studies [14]. Indeed, bridging strategies have been suggested both in the application of foreign data generally and for the extrapolation of fracture risk reduction data based on BMD response [15].

For teriparatide, bridging may have been particularly appropriate because intrinsic factors [i.e. the route of administration (subcutaneous injection), high bioavailability (~95%), and hormonal nature of the compound, where short periods of exposure (t1/2 of ~1 h) produce biological effects] would tend to reduce the likelihood of genetic/ethnic differences affecting drug availability [12]. This paper briefly explores pharmacokinetics (PK) in Japanese and Caucasian/non-Japanese subjects and then describes bridging of Japanese and foreign data based on comparisons of efficacy and safety data obtained in a large, international study conducted outside Japan with data obtained through up to 12 months of randomized, placebo-controlled, double-blind therapy in Japanese men and postmenopausal women with osteoporosis at high risk of fracture.

Materials and methods

Study designs and methods

Studies in subjects with osteoporosis

The FPT (ClinicalTrials.gov identifier NCT00670501) was a large, international placebo-controlled study that examined the safety and efficacy of teriparatide treatment [20 or 40 μg once daily (μg/day)] over a median of 19 months in postmenopausal women with osteoporosis [7, 16]. The study specifically assessed fracture risk, as well as the secondary endpoints of BMD and biochemical markers of bone metabolism. Two studies in Japanese subjects were prospectively planned to allow direct comparison with the FPT. In the J-Ph2 study (ClinicalTrials.gov identifier NCT00191867), the dose–response efficacy and safety of teriparatide 10, 20, and 40 μg/day in Japanese subjects were examined [17]. The results of the J-Ph2 study led to the selection of teriparatide dosed at 20 μg/day for development in Japanese subjects based on observations of dose-dependent increases in BMD, but increased rates of adverse events at 40 μg/day. The J-Ph3 study (ClinicalTrials.gov identifier NCT00433160) then assessed the long-term safety and efficacy of teriparatide 20 μg/day during 12 months of double-blind, placebo-controlled treatment with teriparatide, followed by up to 12 months of open-label teriparatide dosing [18]. The FPT included only postmenopausal women while the J-Ph3 study included postmenopausal women and a small proportion of men.

The overall study designs and the specific characteristics of the data that were analyzed for this report are summarized in Table 1. As no centers located in Japan participated in the FPT, the FPT study population is herein referred to using the general term “non-Japanese”; however, this term does not exclude the possibility of participants of Japanese extraction in the FPT, albeit at centers outside of Japan. The various dose groups in the FPT were 98–99% Caucasian [7].
Table 1

Study design and analysis characteristics for foreign and Japanese studies

Study

FPT (foreign Fracture Prevention Trial)

J-Ph3 (Japanese confirmation study)

J-Ph2 (Japanese dose–response study)

Design

Phase III, randomized, double-blind, placebo-controlled, comparative

Phase III, randomized, double-blind, placebo-controlled, comparative

Phase II, randomized, double-blind, placebo-controlled, comparative

Postmenopausal (>5 years) women with osteoporosis

Postmenopausal (>5 years) women and men with osteoporosis

Postmenopausal (>5 years) women with osteoporosis

99 Centers, 17 countries

28 Sites in Japan

19 Sites in Japan

Duration: median 19 months of treatment (36 months planned)

Duration: 12 months placebo-controlled; up to 12 months open-label

Duration: 6 months

Subjects and treatmentsa

Placebo (N = 544)

Placebo (N = 67)

Placebo (N = 38)

TPTD 20 μg/day (N = 541)

TPTD 20 μg/day (N = 136)

TPTD 10 μg/day (N = 38)

TPTD 40 μg/day (N = 552)

 

TPTD 20 μg/day (N = 39)

TPTD 40 μg/day (N = 39)

Calcium: 1,000 mg

Vitamin D: 400–1,200 IU

Calcium: 610 mg

Vitamin D: 400 IU

Calcium: 610 mg

Vitamin D: 400 IU

Primary efficacy endpoints

New vertebral fracture

Change in BMD at lumbar spine (L2–L4)

Change in BMD at lumbar spine (L2–L4)

Selected secondary efficacy endpoints

Change in:

Change in:

Change in:

Lumbar spine (L1–L4) BMD

Lumbar spine (L1–L4) BMD

Lumbar spine (L1–L4) BMD

Total hip BMD

Total hip BMD

Total hip BMD

Femoral neck BMD

Femoral neck BMD

Femoral neck BMD

Bone metabolism markers

Bone metabolism markers

Bone metabolism markers

BMD bone mineral density, CTX C-terminal telopeptide, ELISA enzyme-linked immunosorbent assay, NTX N-terminal telopeptide, PICP carboxy-terminal extension peptide of procollagen type 1, PINP procollagen I N-terminal propeptide, TPTD teriparatide

aN reflects number of patients randomized in the FPT and the number of patients in the full analysis populations for the J-Ph2 and J-Ph3 studies [17, 18]

Subjects in the FPT had osteoporosis with at least one moderate or at least two mild prevalent vertebral fragility fractures. In the J-Ph3 and J-Ph2 studies subjects had osteoporosis at high risk of fracture, which was defined as BMD at lumbar spine L2–L4 <80% of the young adult mean (YAM, based for all subjects in the J-Ph3 and J-Ph2 studies on the Japanese Normative Female Database [19]), with a minimum of one prevalent vertebral fragility fracture in the J-Ph3 study and one moderate or two mild prevalent vertebral fragility fractures in the J-Ph2 study; BMD at L2–L4 <70% of the YAM and age ≥65 years (in both the J-Ph3 and J-Ph2 study); or BMD at L2–L4 <65% of the YAM and age ≥55 years in the J-Ph3 study or BMD at L2–L4 <60% of the YAM in the J-Ph2 study. T scores represent the number of standard deviations from the YAM; for Japanese subjects, 80 and 70% of the YAM are approximately equivalent to T scores of −1.7 and −2.6, respectively [19]. The planned duration of study in the FPT was 36 months; however, the study was discontinued early by the sponsor [7] and the median duration of treatment was 19 months.

Study in healthy postmenopausal women

Following completion of the FPT and prior to the design of dose ranging and efficacy studies in Japanese subjects, a pharmacokinetic study (J-PK, not previously published) was conducted comparing teriparatide PK following acute dosing in healthy, postmenopausal Caucasian or first-generation Japanese subjects. Subjects received teriparatide once weekly at ascending doses of 10 (Japanese subjects only), 20, 40 (twice), and 60 μg, with an additional week for a placebo dose occurring at varying points in the dosing sequence. Pharmacokinetic parameters at all doses were used to assess dose proportionality using the power model within ethnicity (data not shown), while pharmacokinetic comparisons were based on the replicate 40 μg doses. The concentration of immunoreactive teriparatide in human serum samples was measured by a validated two-site immunoradiometric assay. Serum concentration–time data of teriparatide were analyzed individually by non-compartmental analysis using WinNonlin Professional Version 3.1 (Pharsight, Cary, NC, USA) to estimate pharmacokinetic parameters.

Comparisons of data from Japanese and foreign studies

Following the J-PK study in healthy Japanese and Caucasian postmenopausal women, the bridging strategy called for comparison of dose-ranging/pharmacodynamic results in Japanese and non-Japanese subjects based on the FPT and J-Ph2 study, and for comparison of long-term efficacy between the FPT and J-Ph3 study. Comparison of long-term efficacy was based primarily on BMD, which has been reported to account for 30–41% of the reduction in vertebral fracture risk seen during teriparatide therapy [20]. Although changes in BMD at lumbar spine (L2–L4) were the primary efficacy endpoint during the J-Ph3 study, changes at lumbar spine (L1–L4) were also assessed in order to allow comparison with foreign studies such as the FPT, where BMD at lumbar spine (L1–L4) is generally assessed. Analysis of the relationship between BMD at lumbar spine (L1–L4) versus (L2–L4) in the J-Ph3 study showed a strong correlation (Pearson correlation coefficient 0.970); it was thus judged non-problematic to use lumbar spine (L1–L4) BMD for comparison of the BMD response at lumbar spine between the Japanese and non-Japanese studies.

Comparisons for change in BMD were made for subjects receiving placebo and teriparatide 20 μg/day based on the scheduled visits performed at 12 months in the FPT and at the end of the placebo-controlled dosing period (12 months) in the J-Ph3 study. The primary comparison was for percent change in BMD, with actual change in BMD as a secondary comparison. The dose–response relationship was assessed through comparison of the results for actual and percent change in BMD at lumbar spine (L1–L4) in the teriparatide 20- and 40-μg dose groups in the FPT and J-Ph2 study through 6 months. The distribution of baseline BMD values and the relationship between baseline BMD and response to teriparatide were illustrated using scatter diagrams plotting the percent change or actual change in lumbar spine (L1–L4) BMD from baseline to 12 months versus baseline BMD in the FPT and J-Ph3 study; data were further analyzed through regression analyses. Similarly, the distribution of baseline BMD relative to population-specific means (T scores) versus BMD response was illustrated using scatter diagrams plotting the percent change or actual change in lumbar spine (L1–L4) BMD from baseline to 12 months versus baseline T score in the FPT and J-Ph3 study.

Safety data were assessed through a median of 19 months within the FPT study and through 12 months within the J-Ph3 study. For the sake of comparison to the FPT, data through 18 months from subjects in the J-Ph3 study originally randomized to teriparatide were also shown, despite the lack of a placebo-control arm after 12 months.

Laboratory methods and clinical measurements

Bone mineral density was measured by dual-energy X-ray absorptiometry as described previously [7, 17, 18], and central reading was provided by the same company for all three studies. Pharmacokinetic testing was performed by the same company for the FPT, J-PK, and J-Ph2 study. Other laboratory and clinical measurements have been described previously [7, 17, 18].

Statistical methods

Pharmacokinetics

Pharmacokinetic data were compared between healthy postmenopausal female Japanese and Caucasian subjects in the J-PK study by linear mixed-effects models. Following modeling, the effect of ethnicity was investigated for the 40-μg dose of teriparatide based on geometric means of the primary pharmacokinetic parameters maximum concentration (Cmax) and area-under-the-curve (AUC), and weight adjusted Cmax and AUC. The model included logarithm of pharmacokinetic parameters as dependent variables and ethnic group as a covariate with subject included as a random effect; results were presented as the ratio of geometric means for each ethnic group together with the 90% confidence interval.

In the J-Ph2 study, the PK of teriparatide in postmenopausal women with osteoporosis was evaluated using a population pharmacokinetic analysis. Blood samples were collected at predefined time points (from 5 to 120 min postdose) after 16 and 20 weeks of treatment. A one-compartment model parameterized in terms of apparent systemic clearance (CL/F), apparent volume of distribution (V/F), and absorption rate constant (Ka) was used as the base structural model based on previous analyses. A series of pharmacostatistical models was evaluated in order to identify the model that best described the data. The intersubject variability was assumed to be log-normally distributed and the first order conditional estimation with interaction method was used (NONMEM User’s Guide, Part VII). Pharmacokinetic results from the FPT and the J-Ph2 study were described for the distribution of individual clearance and volume of distribution estimates using fitted distributions.

Efficacy

The criteria for bridging of efficacy data were considered to be fulfilled if the lower limit of the 95% confidence interval for difference between the teriparatide 20-μg and placebo groups for percent change from baseline in lumbar spine (L1–L4) BMD at 12 months exceeded zero in both the FPT and the J-Ph3 study. Given that previous analyses of data from the FPT (unpublished observations) had suggested that subjects with lower baseline lumbar spine (L1–L4) BMD showed greater percent change in BMD during treatment, but that actual change from baseline in BMD was not impacted by baseline BMD, comparisons between studies were prospectively planned for both percent change and actual change in BMD. Changes in BMD were plotted as mean ± standard deviation. Actual change and percent change in BMD were plotted versus baseline BMD at lumbar spine and versus baseline T score for BMD at lumbar spine.

The relationship between baseline BMD and the percent or actual change by treatment group across the FPT and the J-Ph3 study was illustrated with scatter diagrams and further analyzed using multiple linear regression. Three sets of regression models were tested including (1) a simple regression model with pooled data; (2) two regression models with a common slope across studies; and (3) two individual regression models. Residuals of each model were compared to determine if a common model could be fitted across the studies.

Safety

Adverse events (AEs) were originally coded in the FPT using COSTART, and in the J-Ph3 study through 12 months using MedDRA Version 11.0 and MedDRA Version 12.0 for months 12 through 18. For the purposes of these analyses, all AEs were re-coded to reflect MedDRA Version 12.0. Adverse events were summarized by treatment and study; since extrinsic differences in medical care norms could result in different coding between populations, overall comparisons were made primarily between the teriparatide 20 μg and placebo groups within each study. For the sake of comparison of specific AEs to the FPT, AEs through 18 months from subjects in the J-Ph3 study originally randomized to teriparatide were also shown.

Results

Demographics and baseline characteristics

The J-PK study enrolled healthy postmenopausal Japanese (N = 18) and Caucasian (N = 15) subjects with a mean ± SD age of 60.4 ± 7.2 and 60.3 ± 4.7 years, respectively, and body weight of 52.0 ± 7.7 and 68.0 ± 7.6 kg, respectively. Subject demographics for the FPT and the J-Ph2 and J-Ph3 studies are shown in Table 2. Age and number of years post-menopause were closely comparable for the FPT and J-Ph3 study populations. Differences between the populations included height, body weight, and body mass index, which were all numerically greater in the non-Japanese population in the FPT compared to the Japanese subjects in the J-Ph3 study. The proportion of subjects reporting smoking and alcohol consumption were also higher in the FPT than in the J-Ph3 study.
Table 2

Baseline demographics for full analysis populations in the FPT and the J-Ph3 and J-Ph2 studies

Variable

FPT

J-Ph3

J-Ph2

Age (years)

 n

1637

203

154

 Mean ± SD

69.5 ± 7.0

69.6 ± 6.0

71.0 ± 5.1

 Median (min, max)

69.8 (42, 86)

70.0 (55, 87)

71.0 (58, 85)

Gender

 F

1637 (100.0)

189 (93.1)

154 (100.0)

 M

0 (0.0)

14 (6.9)

0 (0.0)

Years post-menopausal

 n

1448

189

154

 Mean ± SD

21.4 ± 8.5

20.4 ± 7.1

22.6 ± 7.8

 Median (min, max)

21.0 (5, 51)

20.0 (6, 41)

21.5 (6, 57)

Height (cm)

 n

1611

203

154

 Mean ± SD

157.23 ± 6.54

150.36 ± 5.89

149.49 ± 5.49

 Median (min, max)

157.40 (136.5, 180.3)

150.50 (134.1, 170.5)

150.00 (134.3, 161.7)

Weight (kg)

 n

1621

203

154

 Mean ± SD

65.80 ± 11.74

49.24 ± 7.29

48.18 ± 7.12

 Median (min, max)

65.00 (28.1, 138.0)

48.80 (31.0, 73.0)

48.20 (31.8, 75.6)

BMI (kg/m2)

 n

1609

203

154

 Mean ± SD

26.62 ± 4.41

21.80 ± 3.15

21.55 ± 2.94

 Median (min, max)

26.13 (11.7, 50.7)

21.74 (12.8, 33.4)

21.26 (14.0, 30.8)

Smoking

 Yes (%)

279 (17.0)

21 (10.3)

15 (9.7)

Alcohol use

 Yes (%)

610 (37.3)

53 (26.1)

19 (12.3)

Prior osteoporosis drug use

 Yes (%)

237 (14.5)

73 (36.0)

38 (24.7)

Number of vertebral fractures

 0

153 (9.3)

120 (59.1)

88 (57.1)

 1

472 (28.8)

51 (25.1)

37 (24.0)

 2

381 (23.3)

18 (8.9)

16 (10.4)

 3

223 (13.6)

6 (3.0)

7 (4.5)

 4 or more

336 (20.5)

8 (4.0)

6 (3.8)

 Unknown

72 (4.4)

0 (0.0)

0 (0.0)

BMD at total spine (g/cm2)

 n

1582

203

153

 Mean ± SD

0.8204 ± 0.1702

0.6143 ± 0.0726

0.6005 ± 0.0701

 Median (min, max)

0.8047 (0.327, 2.068)

0.6181 (0.372, 0.797)

0.6074 (0.364, 0.755)

T score

 n

1582

203

153

 Mean ± SD

−2.59 ± 1.44

−3.13 ± 0.61

−3.24 ± 0.62

 Median (min, max)

−2.72 (−6.8, 8.0)

−3.10 (−5.1, −1.7)

−3.16 (−5.3, −2.0)

For categorical data, n (%) is shown

BMD bone mineral density, BMI body mass index, FPT Fracture Prevention Trial, J-Ph2 Japanese dose–response study, J-Ph3 Japanese confirmation study, SD standard deviation

In terms of factors closely related to disease status, no vertebral fracture was present at baseline in 9.3% of subjects in the FPT versus 59.1% of subjects in the J-Ph3 study, reflecting the differences in inclusion criteria between these studies. The mean ± standard deviation of lumbar spine (L1–L4) BMD at baseline was 0.6143 ± 0.0726 g/cm2 in the J-Ph3 study and 0.8204 ± 0.1702 g/cm2 in the FPT; these values reflected average T scores of −3.13 and −2.59 relative to the population mean for each study, respectively.

Pharmacokinetics

Pharmacokinetics in healthy Japanese and Caucasian subjects (J-PK)

In healthy postmenopausal women, pharmacokinetic parameters were compared between Japanese (N = 18) and Caucasians (N = 15) at the 40 μg dose. The AUC and Cmax were approximately 30–40% higher in Japanese subjects than in Caucasian subjects [ratio (90% confidence interval) for Japanese/Caucasians subjects of 1.43 (1.19, 1.73) for AUC and 1.34 (1.09, 1.65) for Cmax]. When adjusted for body weight, however, the ratio between populations for AUC and for Cmax became close to 1 [ratio for Japanese/Caucasians of 1.09 (0.93, 1.27) for AUC and 1.02 (0.87, 1.19) for Cmax].

Pharmacokinetics in postmenopausal women with osteoporosis (FPT and J-Ph2)

The PK of teriparatide in postmenopausal women with osteoporosis was estimated using population pharmacokinetic analyses in the J-Ph2 study and for a subpopulation in the FPT. In the J-Ph2 study, body weight had a significant influence on the volume of distribution (V/F), and thus was retained in the final pharmacokinetic model (Table 3). The addition of body weight to the model reduced intersubject variability in V/F from 47.7 to 36.5% without significant changes in clearance (CL/F) or proportional residual error. Higher body weight was also associated with higher V/F in the FPT (data not shown); however, a significant effect of body weight on CL/F was not observed in either study. The distributions of individual estimates for CL/F and V/F from the two studies are described using fitted distributions in Fig. 1. Assessing the FPT and the J-Ph2 study, mean CL/F was comparable; however, inter-patient variability was larger in the FPT study. The V/F in the J-Ph2 study was slightly lower than in the FPT, possibly reflecting the effect of body weight on V/F. The overlap area of each graph, that is, the common portion under the two curves, represents the proportion of patients having CL/F and V/F in a specified interval common to the FPT and J-Ph2 studies. This area was 73% of the total area for CL/F and 85% for V/F.
Table 3

Pharmacokinetic and covariate parameters in final population pharmacokinetic model in Japanese patients (J-Ph2)

Parameter description

Population estimate (%SEE)

Interpatient variability (%SEE)

Rate of absorption

 Parameter for Ka (h−1)

17.5 (38.5)

 

Apparent clearance

 Parameter for CL/F (L/h)

54.3 (4.77)

28.9 (32.3)a

Apparent volume of distribution

 Parameter for V/F (l)

80.1 (5.57)

36.5% (35.0)a

 Effect of body weight on V/Fb

1.62 (16.6)

 

Interpatient variability interaction term (CL and V)

 

0.0606 (32.7)

Residual error (proportional)

25.9% (14.9)

SEE standard errors of estimations, Ka absorption rate constant, CL/F apparent clearance, V/F apparent volume of distribution

a%SEE for original variance estimate, not calculated % coefficient of variation

bV/F = 80.1*(body weight/47.4)1.62 where 47.4 = median population body weight

https://static-content.springer.com/image/art%3A10.1007%2Fs00774-011-0314-4/MediaObjects/774_2011_314_Fig1_HTML.gif
Fig. 1

Distribution of individual estimates of teriparatide clearance (a) and volume of distribution (b) in the FPT and the J-Ph2 study. The area of overlap between studies was 73% of the total area for clearance and 85% for volume of distribution. FPT, Fracture Prevention Trial; J-Ph3, Japanese confirmation study

Efficacy

Primary bridging criteria

The percent change in lumbar spine (L1–L4) BMD from baseline to 12 months of treatment is shown for the J-Ph3 study and the FPT in Table 4. In the J-Ph3 study, the percent change in BMD for the placebo and teriparatide 20 μg groups was 0.23 ± 4.44 and 10.43 ± 5.61%, respectively, compared to 0.84 ± 4.87 and 8.25 ± 6.10%, respectively, in the FPT. The difference [mean (95% confidence interval)] between the teriparatide 20-μg group and the placebo group in the percent change in lumbar spine (L1–L4) BMD from baseline to 12 months was 10.20% (8.57, 11.84%) in the J-Ph3 study and 7.41% (6.70, 8.12%) in the FPT, thereby confirming superiority of teriparatide 20 μg to placebo in both studies and fulfilling the primary bridging criterion. The mean actual change from baseline to 12 months of treatment was 0.0630 ± 0.0313 g/cm2 in the J-Ph3 study and 0.0632 ± 0.0426 g/cm2 in the FPT while the difference between teriparatide and placebo was 0.0623 ± 0.0296 in the J-Ph3 study and 0.0567 ± 0.0400 in the FPT.
Table 4

Actual and percent change in lumbar spine (L1–L4) BMD from baseline to 12 months in the FPT and the J-Ph3 study

 

FPT

J-Ph3

Placebo (n = 467)

TPTD 20 μg (n = 466)

TPTD-Placebo

Placebo (n = 60)

TPTD 20 μg (n = 121)

TPTD-placebo

Percent change (%)

 Mean

0.84

8.25

7.41

0.23

10.43

10.20

 SD

4.87

6.10

5.52

4.44

5.61

5.25

  95% CI

  

6.70, 8.12

  

8.57, 11.84

Actual change (g/cm2)

 Mean

0.0065

0.0632

0.0567

0.0008

0.0630

0.0623

 SD

0.0373

0.0426

0.0400

0.0256

0.0313

0.0296

  95% CI

  

0.0515, 0.0618

  

0.0530, 0.0715

BMD bone mineral density, CI confidence interval, FPT Fracture Prevention Trial, J-Ph3 Japanese confirmation study, SD standard deviation, TPTD teriparatide

Dose–response relationship (FPT and J-Ph2)

The comparison of the dose–response relationship for changes in lumbar spine (L1–L4) BMD through 6 months in the FPT and J-Ph2 studies is shown in Fig. 2. In both studies, the percent change in lumbar spine (L1–L4) BMD from baseline to 6 months increased with the dosage of teriparatide. The magnitude of percent change and actual change in BMD at 6 months was comparable across doses in the two studies.
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-011-0314-4/MediaObjects/774_2011_314_Fig2_HTML.gif
Fig. 2

Dose–response relationship in the FPT and J-Ph2 study. Data are mean ± standard deviation. BMD, bone mineral density; FPT, Fracture Prevention Trial; J-Ph2, Japanese dose–response study

Time course of changes in BMD at lumbar spine (L1–L4)

The time course of the percent change from baseline in lumbar spine (L1–L4) BMD in the placebo group and the teriparatide 20-μg group is shown for the FPT and J-Ph3 study in Fig. 3. The percent change in lumbar spine (L1–L4) BMD increased with a similar time course in both studies, while actual change in lumbar spine (L1–L4) BMD at the 3-, 6-, and 12-month time points was notably similar for the two studies.
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-011-0314-4/MediaObjects/774_2011_314_Fig3_HTML.gif
Fig. 3

Time course of percent change and actual change in BMD at lumbar spine (L1–L4) through 12 months in the FPT and J-Ph3 study. Data are mean ± standard deviation. BMD, bone mineral density; FPT, Fracture Prevention Trial; J-Ph3, Japanese confirmation study

Regression analyses of changes in BMD (FPT and J-Ph3)

As shown in Fig. 4a and b, the distribution of baseline BMD in the J-Ph3 study fell within that of the FPT, but the FPT included many patients with baseline BMD higher than the cutoff value corresponding to 80% of the YAM in Japanese patients that defined the upper limit allowable under the entry criteria for the Japanese study. Multiple linear regression analysis of baseline BMD at the lumbar spine versus the percent or actual change in BMD during the study assessed the relationship between response to teriparatide and baseline BMD in the FPT and J-Ph3 studies. The analyses demonstrated that those relationships were similar in the two studies, and that data were best represented by a common regression line. For teriparatide-treated subjects, the common linear regression model demonstrated a significant relationship between baseline BMD at the lumbar spine and both percent change in BMD (p < 0.0001) and actual change in BMD (p = 0.0008), while no relationship was observed for the placebo group (p = 0.372 for percent change and p = 0.835 for actual change).
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-011-0314-4/MediaObjects/774_2011_314_Fig4_HTML.gif
Fig. 4

Scatter diagrams for 12 month percent change (a) and actual change (b) in BMD at lumbar spine (L1–L4) versus BMD at baseline for the FPT and J-Ph3 study. Regression lines based on the common model including data from both the FPT and J-Ph3 study are shown for the Placebo (dashed line equations [a] y = 1.6 − 1.1*x and [b] y = 0.004 + 0.002*x) and Teriparatide 20-μg/day groups (solid line equations [a] y = 20.0 − 14.5*x and [b] y = 0.088 − 0.032*x). BMD, bone mineral density; FPT, Fracture Prevention Trial; J-Ph3, Japanese confirmation study; TPTD, teriparatide

Distribution of change in BMD relative to baseline T scores (FPT and J-Ph3)

In order to assess baseline BMD relative to population-specific means at baseline and also change in BMD during treatment relative to degree of osteoporosis, percent and actual change in BMD were plotted against baseline population-specific T scores (Japanese for J-Ph3; Caucasian for FPT) in Fig. 5a and b. A large degree of overlap was observed in the distribution of baseline T scores between the studies, although some subjects in the FPT had T scores above the BMD cutoffs that were part of the entry criteria for the J-Ph3 study. A rightward shift (horizontal axis) of the distribution of data from J-Ph3 relative to the FPT is seen in Fig. 5 relative to Fig. 4, reflecting the use of T scores in Fig. 5 which are based on the lower population mean observed for Japanese versus Caucasian subjects. Within the range of overlapping T scores for the 2 studies, distribution of BMD response to teriparatide 20 μg/day (vertical axis) appeared to be quite comparable, particularly for percent response (Fig. 5a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00774-011-0314-4/MediaObjects/774_2011_314_Fig5_HTML.gif
Fig. 5

Scatter diagrams for 12 month percent change (a) and actual change (b) in BMD at lumbar spine (L1–L4) versus population-specific T score at baseline for the FPT and J-Ph3 study. Note that the use of entry criteria based on prior fracture in FPT versus BMD loss (as a percentage of young adult mean) in J-Ph3 results in a maximum T score cutoff for the J-Ph3 but not the FPT population. BMD, bone mineral density; FPT, Fracture Prevention Trial; J-Ph3, Japanese confirmation study; TPTD, teriparatide

Safety

Adverse events in the FPT and J-Ph3 study are summarized in Table 5. No adverse events previously identified as characteristic of teriparatide use (such as headache or nausea) were seen at increased incidence in the J-Ph3 study for subjects receiving teriparatide compared to placebo.
Table 5

Treatment emergent adverse events in the FPT and J-Ph3 study (events occurring in >5% of teriparatide-treated subjects through 12 months in the J-Ph3 study)

 

FPT (19 months)

J-Ph3 (12 months)

J-Ph3 (18 months)a

Placebo (N = 544)

n (%)

TPTD 20 μg (N = 541)

n (%)

Placebo (N = 67)

n (%)

TPTD 20 μg (N = 136)

n (%)

TPTD 20 μg (N = 136)

n (%)

Nasopharyngitis

23 (4.2)

32 (5.9)

27 (40.3)

38 (27.9)

45 (33.1)

Back pain

115 (21.1)

84 (15.5)

10 (14.9)

17 (12.5)

22 (16.2)

Constipation

24 (4.4)

32 (5.9)

2 (3.0)

10 (7.4)

11 (8.1)

Fall

26 (4.8)

17 (3.1)

6 (9.0)

9 (6.6)

11 (8.1)

Headache

45 (8.3)

44 (8.1)

3 (4.5)

9 (6.6)

10 (7.4)

Osteoarthritis

29 (5.3)

16 (3.0)

4 (6.0)

9 (6.6)

10 (7.4)

Arthralgia

85 (15.6)

90 (16.6)

4 (6.0)

8 (5.9)

8 (5.9)

Contusion

21 (3.9)

9 (1.7)

4 (6.0)

8 (5.9)

11 (8.1)

Dizziness

33 (6.1)

47 (8.7)

3 (4.5)

8 (5.9)

11 (8.1)

Seasonal Allergy

4 (0.7)

1 (0.2)

4 (6.0)

8 (5.9)

8 (5.9)

Upper respiratory tract inflammation

22 (4.0)

11 (2.0)

3 (4.5)

8 (5.9)

9 (6.6)

Cystitis

26 (4.8)

17 (3.1)

2 (3.0)

7 (5.1)

7 (5.1)

FPT Fracture Prevention Trial, J-Ph3 Japanese confirmation study, TPTD teriparatide

aPlacebo treatment period ended at 12 months in the J-Ph3 study; to facilitate comparison with the FPT, data through 18 months are shown for subjects originally randomized to teriparatide in the J-Ph3 study

An adverse event of hypercalcemia was seen for no subjects in the placebo and teriparatide 20-μg groups in the FPT, and for no subjects receiving placebo and one subject receiving teriparatide 20 μg in the J-Ph3 study through 18 months. An adverse event of “blood uric acid increased” was seen in no subjects in the placebo and teriparatide 20 μg groups in the FPT, and for no subject receiving placebo and two subjects receiving teriparatide 20 μg in the J-Ph3 study through 18 months; notably, other AEs associated with increased blood uric acid, such as gout, were not observed in the 2 subjects with “blood uric acid increased” reported. Thus, the trend of common AEs in the Japanese study was similar to that in the overseas study, with no specific trend identified in Japanese subjects.

Discussion

Teriparatide is a drug for which efficacy and safety in humans were not expected to be substantially affected by extrinsic or intrinsic ethnic factors; thus, findings of well designed clinical trials would likely be relevant across different ethnic and cultural populations. Such results might include the reduction in risk of new vertebral fractures by 65% and of nonvertebral fractures by 35% compared with placebo in the FPT [7], which was shown to be accompanied by improvements in bone structural geometry (proximal femur) [21], cortical and cancellous bone structure [22], and trabecular microarchitecture [2] in analyses in subgroups of the FPT population. Further analyses of data from the FPT demonstrated that increases in lumbar spine BMD during teriparatide therapy accounted for 30–41% of the observed reduction in vertebral fracture risk [20]. In addition, a study comparing response to teriparatide and alendronate in patients with glucocorticoid-induced osteoporosis found greater improvements in BMD and reductions in incidence of vertebral fractures with teriparatide [10, 23]. Overall, the present analyses found that efficacy and safety was closely comparable in studies in Japanese subjects and in the FPT, strongly supporting the bridging of data between these studies.

Notably, successful bridging of study data, even absent a significant impact of ethnic differences, requires comparability in the design, conduct, and results of the bridging studies [13]. In order to facilitate comparisons, the dose-ranging J-Ph2 study and J-Ph3 confirmatory study in Japanese subjects were therefore designed to be easily comparable to the international FPT trial, using, for example, the same timing of sampling and the same central reader for BMD assessment. The largest differences in design (apart from ethnicity and nationalities) were the primary study endpoint of fracture prevention in the FPT, which also meant that the population was selected for prior vertebral fracture. In contrast, the J-Ph3 study selected subjects using BMD-based definitions of osteoporosis at high risk of fracture appropriate to labeling and treatment of osteoporosis in Japan. In addition, a small proportion of men were included in the J-Ph3 study.

In terms of intrinsic factors, the populations differed somewhat in body weight, which resulted in somewhat higher teriparatide exposure in Japanese subjects. The initial study (J-PK) comparing pharmacokinetic parameters between healthy Japanese and Caucasian postmenopausal women showed approximately 30–40% higher Cmax and AUC(0−tlast) in Japanese than in Caucasian subjects; however, the body weight-adjusted Cmax and AUC(0−tlast) were comparable between the two populations. The pharmacokinetic findings for Japanese patients from the J-Ph2 study appeared to be similar to those for the healthy Japanese subjects (J-PK study), and were also consistent with analyses of pharmacokinetic data from the FPT. Notably, in the primary comparison of safety results between the large FPT and the J-Ph3 study, the AE profile during placebo and teriparatide 20-μg/day dosing was comparable within each study, with no clear safety signals. This finding included AEs such as hypercalcemia, increased blood uric acid, and gout, which were rare in both studies and were not associated with other, potentially related AEs in either study; in addition, no novel AEs were identified in Japanese subjects. Finally, there was no increase in discontinuation rate for Japanese subjects receiving teriparatide 20 μg/day compared to placebo in either the J-Ph2 or J-Ph3 study [17, 18]. These findings indicate that the safety and tolerability profile of teriparatide 20-μg/day was similar in Japanese and non-Japanese subjects. Thus, safety and tolerability were not significantly affected by differences in body weight, indicating that no dosage adjustment for weight is needed for Japanese subjects on the basis of safety or tolerability when teriparatide is administered at 20 μg/day.

As expected based on previous reports for the Japanese population [24], the mean baseline BMD values in the J-Ph2 and J-Ph3 studies were substantially lower than in the FPT. However, when adjusted as T scores using population-specific means, the values were relatively comparable (−2.59 in the FPT and −3.13 in the J-Ph3 study). These differences were also illustrated in scatter plots of baseline BMD or T score versus response, where values for the J-Ph3 study clustered near the bottom of values from the FPT when plotted against baseline BMD but shifted toward the middle when plotted based on population-specific T scores, which adjust for the intrinsic population differences in BMD between Japanese and Caucasian subjects. Thus, the distributions showed that the baseline severity of osteoporosis in the J-Ph3 and FPT populations were similar, with the exception of data points from the FPT for subjects with BMD above the cutoff (maximum of 80% of the YAM) used for the Japanese studies.

In terms of bridging the efficacy results between the FPT and the J-Ph3 study, the primary bridging criterion was satisfied in as much as teriparatide 20 μg was superior to placebo at 12 months in both studies for percent change from baseline in lumbar spine (L1–L4) BMD. Mean percent changes in BMD were broadly comparable between the studies, but were numerically greater in the J-Ph3 study, while actual changes in BMD were quite similar in the two studies. However, when compared visually across the overlapping portion of the distribution of baseline population specific T scores, percent change in BMD appeared to be closely comparable between the studies. Thus, teriparatide 20 μg/day offered closely comparable percent response for subjects with the same degree of baseline osteoporosis across the Japanese and non-Japanese populations studied.

Notably, subjects originally randomized to teriparatide in the J-Ph3 study were able to continue receiving teriparatide through a total of 18 and 24 months. As reported previously, there were no notable safety findings at 18 and 24 months compared to those at 12 months, whereas percent change in BMD [lumbar spine (L2–L4) and femoral neck] continued to increase with further teriparatide treatment [18].

Previous reports for the J-Ph2 and J-Ph3 studies demonstrated that in addition to the similar BMD responses described here, procollagen I N-terminal propeptide (PINP), a bone metabolic marker, showed a rapid and sustained response in Japanese subjects [17, 18]. Although PINP was not originally assessed during the FPT study (the PINP assay was still under development), a retrospective analysis in a large subset of patients (placebo, n = 260; teriparatide 20 μg/day, n = 257) showed significant increases in PINP (p < 0.01) between baseline and 3 months [1]. In addition, in two later studies in largely Caucasian populations, increases in PINP remained significant throughout teriparatide dosing periods of 6 and 12 months, indicating that this pattern was not restricted to Japanese subjects [6, 25]. Response of markers of bone resorption during teriparatide 20-μg/day dosing in both the Japanese studies [17, 18] and the FPT [1] lagged behind markers of bone formation, but began to increase by 3 months and remained elevated throughout each of the studies. Thus, the overall pattern of changes in markers of bone metabolism, which reflects bone anabolic response after teriparatide treatment, was similar between the Japanese studies and the FPT.

Conclusions

Teriparatide is well suited to bridging studies because its mode of administration, mechanism of action, and high bioavailability make it relatively insensitive to many ethnic effects. Consistent with this notion, this analysis demonstrated that teriparatide met the ICH E5 criteria for bridging safety and efficacy results, based on studies of

Japanese and non-Japanese subjects. Thus, such findings as a reduced risk of fracture, improvements in bone microarchitecture, or increases in bone strength reported in the FPT [2, 3, 7, 21] and other studies [4, 5] can also be expected to occur in Japanese subjects receiving teriparatide. Given that safety and tolerability profiles were also comparable in the FPT and Japanese studies, these findings strongly support the use of teriparatide 20 μg/day in Japanese subjects with osteoporosis at high risk of fracture.

Acknowledgments

The studies described in this work were funded by Eli Lilly and Company. The authors would like to thank Thomas Melby of i3 Statprobe and Miho Hatano of Eli Lilly Japan K. K. for assistance in the preparation of the manuscript.

Conflict of interest

All authors are employees of Eli Lilly Japan, K.K.

Copyright information

© The Japanese Society for Bone and Mineral Research and Springer 2011