Bone Mineral Density Response With Denosumab in Combination With Standard or High-Dose Teriparatide: The DATA-HD RCT

Abstract

Context

In the Denosumab and High-Dose Teriparatide Administration (DATA-HD) study, we reported that 15 months of combined high-dose (HD) teriparatide and denosumab increased mean areal bone mineral density (aBMD) at the hip and spine more than combined denosumab and standard-dose (SD) teriparatide.

Objective

In the current analysis, we compare the individual rates of aBMD response between the treatment groups.

Design

Single-site, open-label, randomized controlled trial in which postmenopausal women received either teriparatide 20-μg daily (SD) or 40-μg daily (HD) given months 0 through 9, overlapped with denosumab 60 mg, given months 3 through 15 (15 months’ total duration). The proportion of participants in the SD and HD groups experiencing total hip, femoral neck, and lumbar spine aBMD gains of >3%, >6%, and >9% were compared.

Participants

Postmenopausal women with osteoporosis completing all study visits (n = 60).

Main outcome measure(s)

aBMD (dual x-ray absorptiometry).

Results

At the end of the 15-month treatment period, a higher proportion of women in the HD group had aBMD increases >3% (83% vs. 58%, P = .037) and >6% (45% vs. 19%, P = .034) at the total hip, and >3% at the femoral neck (86% vs. 63%, P = .044). At the lumbar spine, >3% response rates were similar, whereas the >6% and >9% response rates were greater in the HD group (100% vs. 79%, P = .012 and 93% vs. 59%, P = .003, respectively).

Conclusion

Compared with the SD regimen, more women treated with the HD regimen achieved clinically meaningful and rapid gains in hip and spine aBMD. These results suggest that this approach may provide unique benefits in the treatment of postmenopausal osteoporosis.

Areal bone mineral density (aBMD) derived from dual x-ray absorptiometry (DXA) is commonly used to monitor an individual’s response to osteoporosis therapy (1). In patients with osteoporosis or increased fracture risk, improvements in aBMD during treatment predict the observed reduction in fracture risk, though to varying degrees (2–8). Additionally, in a recent meta-regression of 38 placebo-controlled trials of 19 osteoporosis agents, larger increases in aBMD were strongly associated with greater reductions in vertebral and hip fractures, further supporting the use of treatment-related increases in aBMD as a surrogate endpoint for fracture (9).

Conventional osteoporosis therapies produce modest gains in aBMD (10–15), so there is an ongoing need to develop therapeutic regimens that will more rapidly restore skeletal integrity. We have previously reported that the combination of teriparatide, an osteoanabolic agent, and denosumab, a potent antiresorptive agent, produced large increases in aBMD by fully inhibiting teriparatide-induced bone resorption but not modeling-based bone formation (16,17). In the subsequent Denosumab and High-Dose Teriparatide Administration (DATA-HD) study, we reported that 15 months of combined high-dose (HD) teriparatide (40 μg) and denosumab produced even greater increases in hip and spine aBMD than the standard-dose (SD, 20 μg) combination (18). In the current analysis, we compare the relative response rates, as measured by aBMD gains at prespecified thresholds, among postmenopausal osteoporotic women treated with either combined HD teriparatide and denosumab or SD teriparatide and denosumab.

Methods

Study design and participants

Details of the DATA-HD study design and participants have been described previously (18). In brief, the DATA-HD study is a 15-month open-label, single-site, randomized controlled trial conducted from October 2014 to June 2016. Seventy-six postmenopausal women aged 45 years or older were enrolled in the study. Women were eligible for inclusion if they were at least 36 months from last menses (or since hysterectomy if FSH concentration was >40 U/L) and at high fracture risk. High fracture risk was defined as an aBMD T-score of -2.5 or less at the lumbar spine, total hip, or femoral neck; T-score of -2.0 or less with at least 1 aBMD-independent risk factor (fragility fracture after age 50, parental history of hip fracture after age 50, prior hyperthyroidism, inability to rise from a chair with arms elevated, or current smoking) (19); or T-score of -1.0 or less with a history of a fragility fracture. Women were excluded if they had evidence of hyperparathyroidism; 25-hydroxy vitamin D deficiency (<20 ng/mL); other congenital or acquired bone disease; history of malignancy (with the exception of nonmelanoma skin cancer); history of ionizing radiation therapy; significant cardiopulmonary, liver, or renal disease; major psychiatric disease; or excessive alcohol intake. Women were also excluded if they had any prior use of parenteral bisphosphonates or strontium; use within 6 months of oral bisphosphonates, denosumab, or glucocorticoids; or use within 3 months of estrogens, selective estrogen receptor modulators, or calcitonin. The study was approved by the Partners Human Research Committee and is registered with ClinicalTrials.gov, number NCT02176382. All participants provided written informed consent before inclusion in the study.

Study participants were stratified by age (<65 or ≥65 years) and previous bisphosphonate use (yes or no) and then randomized in a 1:1 ratio to receive either teriparatide 20 μg daily (SD) or 40 μg daily (HD) by subcutaneous injection, given months 0 through 9, overlapped with denosumab 60 mg every 6 months by subcutaneous injection, given months 3 through 15, for a total study duration of 15 months (Fig. 1). Additionally, women were given calcium carbonate and vitamin D supplements, if needed, to achieve a total daily intake of 1200 mg elemental calcium and to maintain serum 25-hydroxy vitamin D concentrations of at least 20 ng/mL. Adherence to teriparatide was assessed by subject diary.

Figure 1.

Flow diagram of treatment schedule over the15-month study period.

Bone mineral density measurements

The aBMD of the posteroanterior lumbar spine, total hip, and femoral neck was assessed by DXA using a Hologic QDR 4500A densitometer (Waltham, MA, USA) at baseline, 3-, 9-, and 15-month time points. All scans of an individual subject were performed on the same densitometer. Quality control measurements were performed daily with a Hologic anthropomorphic spine phantom. The SDs of in vivo reproducibility were 0.005, 0.006, and 0.009 g/cm² for posteroanterior lumbar spine, total hip, and femoral neck aBMD measurements, respectively. Vertebrae were excluded from analysis if there were obvious deformities or focal sclerosis detected. Physicians interpreting aBMD assessments were blinded to treatment allocation.

Statistical analysis

Participants who completed all study visits during the 15-month study period were included in the responder analysis. If an aBMD result was not available for a particular anatomic site (total hip, femoral neck, or lumbar spine), participants were excluded from that analysis and no imputation of missing data was performed. Bone mineral density increases of >3%, >6%, and >9% at each anatomic site as well as the combination of all 3 sites were used as thresholds of response. The 3% threshold was predetermined based on the widely accepted least significant change in DXA aBMD measurements, which assumes a DXA precision error of 1% (20). More robust aBMD increase of >6% and >9% were also analyzed. Between-group differences in the proportion of responders for each threshold were analyzed using 2-sided chi-square tests. The Fisher’s exact test was used if the number of responders was fewer than 5 in either treatment group. P values were not adjusted for multiplicity. Baseline characteristics were compared between treatment groups using independent 2-sample t tests for continuous variables and 2-sided chi-square tests for discrete variables.

Role of the funding source

The funding sources and drug supply sources did not have any role in the design, conduct, or analysis of this study.

Results

The baseline demographics and characteristics of the 60 participants who completed all study visits (79% of the enrolled cohort) are shown in Table 1. Baseline characteristics did not differ between treatment groups and were comparable to those in the overall study population (data not shown) (18). Briefly, the mean age was 66.2 years, 95% were non-Hispanic white, 52% had a prior fragility fracture, and 55% had previously been treated with a bisphosphonate.

Table 1.

Demographics and Baseline Characteristics of DATA-HD Study Participants Included in the Responder Analysis^a

Characteristic	SD TPTD (20 μg) Plus Denosumab (n = 31)	HD TPTD (40 μg) Plus Denosumab (n = 29)	P Value
Age (y)	65.8 (7.4)	66.6 (7.2)	.68
Body mass index (kg/m2)	22.8 (2.7)	22.5 (3.7)	.73
White, non-Hispanic	29 (94%)	28 (97%)	.59
History of adult fragility fracture	13 (42%)	18 (62%)	.12
Previous bisphosphonate use	18 (58%)	15 (52%)	.62
Duration of use (mo)	44.2 (32.1)	71.9 (51.9)	.07
Time since discontinuation (mo)	73.0 (64.2)	76.8 (60.0)	.86
Serum P1NP (μg/L)	48.0 (16.8)	51.5 (16.0)	.42
Serum CTX (ng/L)	426.2 (166.1)	464.7 (164.7)	.37
Baseline areal bone mineral density by DXA (g/cm 2)
Total hip	0.748 (0.105)	0.740 (0.073)	.71
Femoral neck	0.652 (0.097)	0.629 (0.071)	.30
Posteroanterior lumbar spine	0.836 (0.109)	0.800 (0.111)	.37

Values are presented as mean (SD) or number (%).

Abbreviations: CTX, serum C-telopeptide of type I collagen; DXA, dual x-ray absorptiometry; HD, high dose; P1NP, serum procollagen type I N-terminal propeptide; SD, standard dose; TPTD, teriparatide.

^aIncludes 60 participants who had areal bone mineral density determinations at all study visits (baseline, month 9, and month 15, 79% of the Denosumab and High-Dose Teriparatide Administration population at enrollment).

The mean percent changes in aBMD at 9 and 15 months are shown in Table 2. Bone mineral density measurements were not available for 1 participant in the SD group for femoral neck at 9 and 15 months and 2 participants in the SD group and 2 participants in HD group for lumbar spine at 9 and 15 months. Note that by study end (15 months), there was a greater increase in aBMD at each site in the HD group than the SD group: 6.1% vs 3.9% (mean difference, 2.2%; 95% CI, 0.6-3.8; P = .009) at the total hip; 6.8% vs 4.3% (mean difference, 2.5%; 95% CI, 0.5-4.5; P = .017) at the femoral neck; and 17.5% vs 9.5% (mean difference, 8.1%; 95% CI, 5.5-10.6; P < .0001) at the lumbar spine.

Table 2.

Percent Change in Bone Mineral Density From Baseline to 9 and 15 Months in DATA-HD Study Participants Included in the Responder Analysis^a

	SD TPTD (20 μg) Plus Denosumab (n = 31)	HD TPTD (40 μg) Plus Denosumab (n = 29)	P Value
Change in areal bone mineral density by DXA from baseline to 9 months (%)
Total hip	3.8 (2.9)	5.2 (3.3)	.076
Femoral neck	4.0 (3.5)	5.4 (3.4)	.138
Posteroanterior lumbar spine	8.6 (3.8)	14.1 (5.4)	<.0001
Change in areal bone mineral density by DXA from baseline to 15 months (%)
Total hip	3.9 (2.9)	6.1 (3.4)	.009
Femoral neck	4.3 (3.7)	6.8 (4.1)	.017
Posteroanterior lumbar spine	9.5 (3.2)	17.5 (6.0)	<.0001

Values are presented as mean (SD).

Abbreviations: DXA, dual x-ray absorptiometry; HD, high dose; SD, standard dose; TPTD, teriparatide.

^aIncludes 60 participants who had areal bone mineral density (aBMD) determinations at all study visits (baseline, month 9, and month 15, 79% of the Denosumab and High-Dose Teriparatide Administration population at enrollment). Data missing for 1 participant in the SD group for femoral neck aBMD at 9 and 15 months and 2 participants in the SD group and 2 participants in the HD group for lumbar spine aBMD at 9 and 15 months.

The cumulative relative frequency distributions of percent change in aBMD from baseline to 9 and 15 months in the SD and HD groups are shown in Fig. 2. Across all 3 sites, at both 9 and 15 months, a greater proportion of participants in the HD group achieved a higher percent change in aBMD compared with participants in the SD group. The maximum percent change in aBMD achieved at all sites, at both 9 and 15 months, was also greater in the HD group compared with the SD group. At 15 months, bone loss occurred in 1 participant at both the total hip and femoral neck and in 1 participant at the femoral neck, with no bone loss observed at the lumbar spine. Both of these participants were in the SD group. No participant randomized to the HD group experienced bone loss at the hip or spine.

Figure 2.

Cumulative relative frequency distribution plots of percent change in bone mineral density from baseline to 9 and 15 months. ^aBMD, bone mineral density; HD, high-dose teriparatide (40 μg) plus denosumab; SD, standard-dose teriparatide (20 μg) plus denosumab.

^aIncludes 60 participants who had areal bone mineral density (aBMD) determinations at all study visits (baseline, month 9, and month 15, 79% of the DATA-HD population at enrollment). Data missing for 1 participant in the SD group for femoral neck aBMD at 9 and 15 months and 2 participants in the SD group and 2 participants in the HD group for lumbar spine aBMD at 9 and 15 months.

As shown in Fig. 3, at 15 months, there was a higher proportion of participants in the HD group with total hip aBMD increases of >3% (83% vs 58%, P = .037) and >6% (45% vs 19%, P = .034), with no significant differences in the >9% (17% vs 6%, P = .193) response rates between the 2 groups. At the femoral neck, there was a higher proportion of participants in the HD group with increases >3% (86% vs 63%, P = .044), as well as similar nonsignificant trends in the >6% (55% vs 33%, P = .091) and >9% (24% vs 7%, P = .062) response rates. At the lumbar spine, all but 1 patient in the SD group achieved an aBMD gain of at least 3%, with no significant differences between the groups (HD = 100% vs SD = 97%, P = .330). The >6% and >9% response rates were higher in the HD group than the SD group (100% vs 79%, P = .012, and 93% vs 59%, P = .003, respectively). At 9 months, response rates at all anatomic sites were similar between the 2 groups, except at the lumbar spine, where the >6% and >9% response rates were higher in the HD group (96% vs 76%, P = .029, and 93% vs 59%, P = .003, respectively). Based on an exploratory subgroup analysis, there were no significant differences in bone density response rates at each anatomic site between women with or without prior oral bisphosphonate exposure (data not shown).

Figure 3.

Proportion of participants achieving bone mineral density increases of >3%, >6%, and >9% at the total hip, femoral neck, and lumbar spine from baseline to 9 and 15 months.^a BL, baseline; HD, high-dose teriparatide (40-μg) plus denosumab; M, month; SD, standard-dose teriparatide (20 μg) plus denosumab.

*P < .05.

**P < .01.

^aIncludes 60 participants who had areal bone mineral density (aBMD) determinations at all study visits (baseline, month 9, and month 15, 79% of the Denosumab and High-Dose Teriparatide Administration population at enrollment). Data missing for 1 participant in the SD group for femoral neck aBMD at 9 and 15 months; and 2 participants in the SD group and 2 participants in the HD group for lumbar spine aBMD at 9 and 15 months.

Fifty-five participants had aBMD assessments at all sites (total hip, femoral neck, and lumbar spine) and at all time points. As shown in Fig. 4, at 15 months, there was a greater proportion of participants in the HD group achieving >3% and >6% gains at all three anatomic sites (78% vs 43%, P = .008, and 41% vs. 14%, P = .028, respectively). Few patients achieved >9% gains at all sites with no significant difference between groups (11% for HD vs 4% for SD, P = .282).

Figure 4.

Proportion of participants achieving bone mineral density increases of >3% at all 3 sites from baseline to 9 and 15 months.^b BL, baseline; FN, femoral neck; HD, high-dose teriparatide (40 μg) plus denosumab; LS, lumbar spine; M, month; SD, standard-dose teriparatide (20 μg) plus denosumab; TH, total hip.

*P < .05.

**P < .01. ^b

Includes 55 participants who had areal bone mineral density determinations at all study visits (baseline, month 9, and month 15) and at all sites (total hip, femoral neck, and lumbar spine).

Discussion

In the present study, we have demonstrated that more women treated with 9 months of HD teriparatide (40 μg) partially overlapped with 12 months of denosumab achieved clinically relevant aBMD gains than women receiving a similar treatment regimen with SD teriparatide (20 μg), although the majority of women in both groups had meaningful gains in aBMD at the hip and spine. Notably, the differences in response rates at the total hip and femoral neck were more pronounced after 15 months of treatment than 9 months even though both groups were receiving only denosumab from months 9 through 15. This observation may potentially be explained by denosumab-induced secondary mineralization of a greater volume of newly formed undermineralized bone during months 9 through 15 (21).

The response rates achieved with combined HD teriparatide and denosumab are also greater and more rapid when compared with prior studies of osteoporosis medications. For example, using the same 3% threshold, the total hip aBMD response rates have been reported to be 37% and 22% after 12 months of alendronate and risedronate treatment, respectively, and 53% after 24 months of denosumab (22,23). The few studies that have evaluated total hip aBMD response rates with teriparatide 20 μg have reported >3% response rates of 47% (n = 93) and 57% (n = 39) after 18 months and 36% after 24 months (n = 28) (23,24). In the present study, the total hip aBMD response rates among women in the HD group were significantly higher at 76% after just 9 months of treatment. Additionally, in a recent study that reported aBMD response rates of abaloparatide, the >6% response rate at the total hip after 18 months of treatment was 26.3% compared with an almost twofold higher rate of 45% among women in the HD group after just 15 months of treatment (25).

In addition to an overall greater response rate achieved in women treated with the HD combination, the response rates were also more consistent among anatomic sites. In the subgroup of women (n = 55) who had aBMD measurements at all 3 sites (total hip, femoral neck, and lumbar spine) and at all timepoints, nearly 4 of 5 women in the HD group achieved >3% aBMD increases at all 3 sites compared with approximately 2 of 5 women in the SD group at 15 months. Perhaps most strikingly, by the end of the study period, no participant randomized to the HD group experienced bone loss at the hip or spine and only 2 women in the SD group experienced bone loss at the hip.

The precise mechanisms underlying the larger effects on aBMD observed with HD compared with SD teriparatide when combined with denosumab are not fully understood. Teriparatide, an osteoanabolic agent, stimulates bone formation through both remodeling- and modeling-based bone formation (26–28). We have previously reported in the DATA-HD study that combining teriparatide (20 μg) with denosumab, a potent inhibitor of receptor activator of nuclear factor kappa-B ligand, led to profound suppression of bone remodeling, which was similar to that seen with the use of denosumab monotherapy, while allowing for continued bone formation (16,17). Hence, we hypothesized that use of a larger anabolic stimulus (i.e., HD teriparatide [40 μg]), would lead to greater separation between bone resorption and formation, favoring overall more bone formation. Notably, this hypothesis was supported by the changes in bone turnover markers observed in DATA-HD (18). Future studies that include histomorphometric analysis of bone biopsy specimens would help elucidate these mechanisms more fully.

This study has several potential limitations. First, the study is limited by the absence of fracture data. That said, there is increasing evidence to support observations that larger gains in aBMD during treatment are associated with greater reductions in fracture risk (2–9). Additionally, there is little reason to hypothesize that combining 2 medications, which both reduce fracture risk individually, would not do so when combined. In a phase 3 trial that compared 20 and 40 μg of teriparatide with placebo, both doses resulted in numerically similar decreases in vertebral and nonvertebral fracture rates despite greater increases in aBMD with the higher dose. Although this study was not powered to detect differences in incident fracture rates between teriparatide 20 and 40 μg, it has been postulated that the lack of greater antifracture efficacy observed with the higher dose may have been due to higher resorption of cortical bone and increased cortical porosity compared with the 20-µg dose (29–33). This does not apply to the present study, in which bone resorption was potently inhibited by combined treatment with denosumab.

Second, although there were no differences in the rates of hypercalcemia or participant withdrawal resulting from adverse effects, formal safety comparisons between the 2 therapeutic regimens are not possible because of the limited study size. Finally, the relatively small cohort size, the post hoc nature of the analysis, and the lack of correction for multiple comparisons may be considered limitations and restrict subgroup analysis, for example, baseline aBMD or women with or without prior bisphosphonate exposure.

In summary, the majority of postmenopausal women treated with combined HD teriparatide (40 μg) and denosumab for 15 months experienced meaningful increases in hip and spine aBMD within just 9 to 15 months, with more than 3 of 4 women experiencing >3% increases at all 3 measured sites (total hip, femoral neck, and lumbar spine). The superiority of this HD combination on overall aBMD accrual and the consistency of the response to treatment provide a compelling rationale to conduct larger studies to investigate the antifracture efficacy and safety of this treatment regimen because it may provide significant benefits to postmenopausal women at highest risk of fragility fracture.

Glossary

Abbreviations

aBMD

areal bone mineral density

DATA-HD

Denosumab and High-Dose Teriparatide Administration

DXA

dual x-ray absorptiometry

HD

high dose

SD

standard dose

Additional Information

Disclosure Summary: J.N.T. has a financial interest in Amgen; her interests were reviewed and are managed by Massachusetts General Hospital and Partners HealthCare in accordance with their conflict of interest policies. B.Z.L. served on an advisory board for Amgen. All other authors have nothing to declare.

Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.