Denosumab versus romosozumab for postmenopausal osteoporosis treatment

Denosumab and romosozumab, a recently approved new drug, are effective and widely known molecular-targeted drugs for postmenopausal osteoporosis treatment. However, no studies have directly compared their therapeutic effects or safety in postmenopausal osteoporosis. This retrospective observational registry study compared the efficacy of 12-month denosumab or romosozumab treatment in postmenopausal osteoporosis patients. The primary outcome was the change in bone mineral density (BMD) at the lumbar spine. Secondary outcomes included BMD changes at the total hip and femoral neck, changes in bone turnover markers, and adverse events. Propensity score matching was employed to assemble patient groups with similar baseline characteristics. Sixty-nine patients each received either denosumab or romosozumab for 12 months. The mean 12-month percentage change from baseline in lumbar spine BMD was 7.2% in the denosumab group and 12.5% in the romosozumab group, indicating a significant difference between the groups. The percentage changes in BMD at both the total hip and femoral neck were also significantly higher at 12 months in the romosozumab group than in the denosumab group. In denosumab patients, bone formation and bone resorption markers were significantly decreased at 6 and 12 months from baseline. In the romosozumab group, the bone formation marker was significantly increased at 6 months and then returned to baseline, while the bone resorption marker was significantly decreased at both time points. Adverse events were few and predominantly minor in both groups, with no remarkable difference in the incidence of new vertebral fractures. Romosozumab showed a higher potential for improving BMD than denosumab in this clinical study of postmenopausal osteoporosis patient treatment.

Primary and secondary outcomes of interest. The primary outcome of interest was the percentage change from baseline in areal BMD by DXA at the lumbar spine during 12 months of treatment (mean values at 6 and 12 months). The secondary outcomes were the percentage changes in total hip and femoral neck BMD at 6 and 12 months as well as the percentage changes in the serum bone turnover markers procollagen type 1 N-terminal propeptide 1 (P1NP) and tartrate-resistant acid phosphatase isoform 5b (TRACP-5b). A previous report demonstrated that TRACP-5b levels were useful bone resorption markers that demonstrated higher clinical sensitivity and signal-to-noise ratio as compared with serum CTX levels 22 . P1NP and TRACP-5b were measured by the enzyme immunoassay and chemiluminescent enzyme immunoassay methods, respectively, at the time of treatment introduction (baseline) and at 6 and 12 months afterwards. Statistical analysis. Patient background parameters are expressed as the mean ± standard deviation. P1NP and TRACP-5b are expressed as the median. Percentage changes from baseline to the 6-and 12-month time points for BMD, P1NP, and TRACP-5b were assessed using the Wilcoxon signed-rank test. The Wilcoxon ranksum test was employed to evaluate the differences between the groups with regards to the percentage changes from baseline for the primary and secondary outcomes. Differences between the study groups were determined by ANOVA or Fisher's exact test. A two-tailed P-value of < 0.05 was considered statistically significant for all analyses. All statistical testing was conducted using R version 3.6.0 (R Core Team, 2019; http:// www.R-proje ct. org/).

Results
Study proportions. Between April 2015 and August 2020, 571 osteoporosis patients received denosumab or romosozumab treatment for 12 months (Fig. 1). Seventy-seven male osteoporosis patients and 229 secondary osteoporosis patients were excluded, leaving 265 patients who met the inclusion criteria of this study. Of them, 131 received denosumab and 134 received romosozumab. Before propensity score matching, there were significant differences between the treatment groups for several baseline variables (Table 1), including a lower T-score and higher incidence of prior fragility fracture in the romosozumab group. These factors confirmed that romosozumab was used in patients with a higher risk of further fractures. After propensity score matching, 69 patients each had received denosumab or romosozumab. No remarkable differences in patient background were detected between the groups (Table 2). Mean ± standard deviation age was 74.2 ± 11.3 years in the denosumab group and 75.8 ± 9.70 years in the romosozumab group. Respective mean T-scores for the denosumab group and romosozumab group were − 2.50 ± 1.13 and − 2.62 ± 1.25 for the lumbar spine, − 2.55 ± 0.73 and − 2.57 ± 0.84 for the total hip, and − 3.12 ± 0.62 and − 3.12 ± 0.82 for the femoral neck. Twenty-six (37.7%) subjects had a prevalent vertebral fracture in the denosumab group, as compared with 25 (36.2%) subjects in the romosozumab group. Twelve (17.4%) and 13 (18.8%) subjects had a history of prior non-vertebral fracture in the denosumab group and romosozumab group, respectively. Twenty-seven (39.1%) denosumab patients and 20 (29.0%) romosozumab patients had some kind of osteoporosis treatment history and had been switched to  Propensity score matching was employed to extract subjects. Table 1. Demographic and clinical characteristics of subjects at baseline before extraction by propensity score matching. Data are expressed as the mean ± standard deviation (SD) or the number (%) of all patients who completed 12 months of denosumab or romosozumab treatment. P1NP and TRACP-5b are expressed as median values. BMI body mass index, P1NP procollagen type 1 N-terminal propeptide, TRACP-5b tartrateresistant acid phosphatase isoform 5b, eGFR estimated glomerular filtration rate, 25OHD 25-hydroxyvitamin D; Differences between the groups were determined by ANOVA or Fisher's exact test.
Primary outcome. Sixty-nine patients each in the denosumab group and romosozumab group were included in the primary outcome analysis. The respective percentage changes from baseline (mean ± 95%CI) in areal BMD tested by DXA at the lumbar spine at 6 and 12 months were 6.0% ± 4.1 (P < 0.001 versus baseline) and 7.2% ± 4.3 (P < 0.001 versus baseline) in the denosumab group and 7.4% ± 1.7 (P < 0.001 versus baseline) and 12.5% ± 2.4 (P < 0.001 versus baseline) in the romosozumab group (Fig. 2a). The percentage change in lumbar spine BMD was significantly higher in the romosozumab group than in the denosumab group at 6 (P < 0.01) and 12 months (P < 0.001).
Secondary outcomes. The respective percentage changes in total hip BMD from baseline at 6 and 12 months were 2.4% and 3.6% in the denosumab group and 3.4% and 6.0% in the romosozumab group (Fig. 2b). Similar results of 2.0% and 2.6% in the denosumab group and 3.0% and 5.5% in the romosozumab group were observed for BMD at the femoral neck (Fig. 2c). All increases were significant (P < 0.01) versus baseline values for both drugs. There was no remarkable difference in percent increases between the denosumab group and romosozumab group at 6 months (total hip: P = 0.394, femoral neck: P = 0.331), although significant differences were noted at 12 months (total hip: P < 0.05, femoral neck: P < 0.01). Those data supported a possible advantage of romosozumab for elevating bone density over denosumab. Next, as the other important factors for osteoporosis treatment, the changes in major serum bone turnover markers, P1NP and TRACP by the action of these treatment drugs were focused. Serum P1NP level was significantly decreased at 6 months (− 63.1%; P < 0.001) and 12 months (− 68.2%; P < 0.001) compared with baseline in the denosumab group. In the romosozumab group, P1NP was significantly higher at 6 months (5.9%; P < 0.01), and then normalized at 12 months (− 5.6%; P = 0.705) (Fig. 3a). There were significant differences between the groups at 6 months (P < 0.001) and 12 months (P < 0.001). Serum TRACP-5b level in the denosumab group was significantly decreased at 6 months (− 56.0%; P < 0.001) and 12 months (− 60.5%; P < 0.001) versus baseline values (Fig. 3b). The romosozumab group displayed a similar trend at 6 months (− 32.1%; P < 0.001) and 12 months (− 42.9%; P < 0.001). A significant difference was observed between the groups both time points (both P < 0.001).
Adverse events and new fractures. The adverse events recorded during the 12 months of treatment are listed in Table 3. Although injection site reactions occurred more frequently in the romosozumab group, they did not lead to drug discontinuation. Injection site reactions often occurred at the time of first administration. The presence of new vertebral fractures during the 12 months of treatment was evaluated by regular X-ray photographs. Two patients in each group (both 2.9%) suffered a new fracture, an incidence that was statistically comparable. Table 2. Demographic and clinical characteristics of subjects at baseline after extraction by propensity score matching. Data are expressed as the mean ± standard deviation (SD) or the number (%) of subjects. P1NP and TRACP-5b are expressed as median values. BMI body mass index, P1NP procollagen type 1 N-terminal propeptide, TRACP-5b tartrate-resistant acid phosphatase isoform 5b, eGFR estimated glomerular filtration rate, 25OHD 25-hydroxyvitamin D; Differences between the groups were determined by ANOVA or Fisher's exact test. www.nature.com/scientificreports/

Discussion
Using a propensity score-matching cohort design, the present study found that the increasing rates of lumbar spine, total hip, and femoral neck BMD were significantly higher with romosozumab than with denosumab after a treatment period of 12 months, with few serious adverse effects for either drug. Denosumab is a fully human monoclonal antibody to the receptor activator of nuclear factor-kappa B ligand (RANKL) that blocks its binding to RANK, thus inhibiting the development and activity of osteoclasts, decreasing bone resorption, and increasing bone density 23 . On the other hand, romosozumab is a bone-forming agent that inhibits sclerostin to promote bone formation and suppress bone resorption through a so-called "dual-effect" 24,25 . These molecular-targeted drugs are prominent in the field of osteoporosis treatment.
Three factors are involved in the increase of BMD: (1) initial closure of the bone remodeling space, (2) a subsequent increase in mineralization, and (3) the steady contribution of modeling-based bone formation 26,27 . Especially in bone remodeling, the transition of bone metabolism markers affects the size of the anabolic window due to the difference between the levels of bone formation markers and bone resorption markers 28 . Denosumab strongly suppresses bone resorption, which in turn inhibits bone formation as well. Bone remodeling is considered to proceed under these conditions. In contrast, romosozumab promotes bone formation and suppresses bone resorption, resulting in a larger anabolic window and presumably a greater effect on increasing bone density. In this study, denosumab decreased both the bone formation marker and the bone resorption marker, while the bone formation marker did not decrease throughout 12 months and only the bone resorption marker was decreased for . Bars indicate the median ± first or third quartile. **P < 0.01 and ***P < 0.001 versus baseline (Wilcoxon's signed-rank test). † † † P < 0.001 versus denosumab (Wilcoxon's rank-sum test). www.nature.com/scientificreports/ romosozumab. Accordingly, we considered that a larger anabolic window was created. Romosozumab also had a greater effect on bone modeling than on bone remodeling in a recent report 29 . Taken together, the considerable effects of romosozumab on bone remodeling and modeling appear more effective to increase bone density levels. Both our primary and secondary clinical results support this theory. As a notable point, the low levels of vitamin D in the cohort are not unique circumstances in Japan. In the real-world setting, approximately 90% of Japanese patients suffer from a vitamin D deficiency or insufficiency as a complication of osteoporosis 30 . We routinely advise the intake of active vitamin D or calcium preparations for patients with lower 25OHD; however, some patients find it undesirable to take additional medicine, and ultimately reject the prescription. For those individuals, we suggest other strategies, such as the intake of supplements and insolation, and provide suitable daily life guidance as well as further medical treatment. In addition, romosozumab is usually reserved for patients with severe osteoporosis in actual clinical practice, and we believe that immediate intervention is required without losing time waiting for 25OHD elevation to a sufficient level to prevent further fractures.
As limitations of this study, the following factors require further consideration: (1) there was no discussion on treatment-naïve vs. switch (non-naïve) patients because we focused on the standardization of patient background characteristics by propensity score matching, (2) as the observation period of this study was short at 1 year, longer follow-up for adverse events and new fractures is needed, and (3) the data on adverse events during the 12 months of treatment were obtained from the clinical records of patients, which was a retrospective process.
In conclusion, this investigation used propensity score matching to directly compare the clinical effects of denosumab and romosozumab in patients with postmenopausal osteoporosis. In terms of BMD of the lumbar spine, total hip, and femoral neck, the 12-month gains in the romosozumab group were all significantly higher than those in the denosumab group, indicating a potential therapeutic advantage that warrants further validation.  www.nature.com/scientificreports/