Discussion

This study demonstrated that in patients receiving long-term oral GC and BP therapy, switching to ROMO for 12 mo resulted in greater increases in the BMD of the LS, FN, and TH compared to continuing BP. Even after adjusting for age, the ROMO group still exhibited a slight tendency toward a greater increase in the BMD of the LS and TH compared to the group that continued BP. These findings suggest that ROMO may be an effective sequential therapy for patients undergoing BP treatment in the context of GIOP.

To our knowledge, no previous studies have directly compared the efficacy of continuing BP therapy vs switching to ROMO for treating osteoporosis in patients receiving both GC and BP therapy. However, a RCT has reported the therapeutic effects of ROMO compared to denosumab (DMAb), an anti-RANKL antibody, in patients receiving long-term low-dose PSL (median dose 5.0 mg/d; treatment duration: mean ± SE, 122.3 ± 89 mo).11 In subgroup analysis, patients receiving ROMO with prior BP therapy showed a mean increase in LS BMD of 5.2% and FN BMD of 0.7% at 12 mo. In contrast, patients without prior BP therapy exhibited a 10% increase in LS BMD and a 2.8% increase in FN BMD, indicating a higher rate of improvement.

Similarly, in our previous retrospective study investigating the effectiveness of DMAb in patients with rheumatic diseases (68.5% using GCs; median PSL-equivalent dose 5.0 mg/d), patients with prior BP use exhibited a lower median percentage increase in LS BMD at 12 mo compared to those without prior BP use (3.6% vs 4.1%).12 Furthermore, patients receiving GCs at doses equivalent to 5 mg or more showed a lower median percentage increase in LS BMD at 12 mo compared to those not using GCs (2.8% vs 5.5%). Further research is needed to compare the efficacy of ROMO and DMAb in GIOP patients with prior BP use.

A case-control study comparing the effects of switching from BP therapy to ROMO, DMAb or teriparatide in patients with PMO reported a mean increases in LS BMD of 11.4%, FN BMD of 2.0% and TH BMD of 3.3% at 12 mo in the patients in the ROMO group.13 Our previous RomGo study evaluated the therapeutic effects of ROMO, DMAb, and BP in patients with rheumatic diseases who had no history of osteoporosis treatment and were initiating moderate to high doses of GCs (median 20.0 mg/d).10 After 12 mo, the median changes in BMD in the ROMO group were 8.6% of the LS, 0.5% of the FN and 2.1% of the TH. These results suggest that the BMD-increasing effect of ROMO may be diminished in patients who have previously received BP therapy, or long-term GC therapy even at low doses.

In this study, serum P1NP levels increased in the ROMO group at 3 mo and returned to baseline after 6 mo. Similar changes have been reported in the aforementioned RCT involving patients receiving low-dose GCs.11 In the ARCH study of postmenopausal women with osteoporosis, serum P1NP increased at 1 mo after ROMO administration, then gradually decreased, returning to baseline levels after 9 mo.8 In contrast, in the RomGo study, patients were initiated on moderate to high doses of GCs, and serum P1NP levels were below baseline from 3 mo.10 These results suggest that under long-term, low-dose GC treatment, the bone formation-promoting effect of ROMO outweighs the bone formation-suppressing effect of GCs. In this study, serum P1NP levels increased at 3 mo but decreased after 6 mo—earlier than in postmenopausal women—likely due to the effects of GCs. While serum P1NP is considered an early marker of bone formation, serum OC is regarded as a later marker because it is produced by mature osteoblasts.14 Although GCs are known to inhibit bone turnover and reduce OC levels, serum OC increased after ROMO administration in this study and, unlike P1NP, remained elevated for up to 6 mo. In the RomGo study, OC levels were also below baseline from 3 mo.10 The suppressive effect on bone formation appears to be stronger during the initiation of moderate to high doses of GCs compared to the use of long-term, low-dose GCs, under treated with ROMO.

The decrease in serum levels of TRACP-5b was greater in the ROMO group than in the BP group in this study. In the previously mentioned RCT involving patients on low-dose GC, serum C-terminal cross-linked telopeptide collagen, another marker of bone resorption, was measured, and it began to decrease 6 mo after the start of ROMO.11 Therefore, it was suggested that ROMO may exert a strong inhibitory effect on bone resorption, even in patients undergoing long-term GC administration and BP therapy.

Baseline levels of urinary pentosidine, which is thought to inversely reflect bone quality,14 were significantly higher in the ROMO group than in the BP group. This difference is likely attributable to the ROMO group consisting of elderly patients and a higher proportion of individuals with a history of fracture prior to the commencement of the study. After the start of the study, the urine levels decreased in both groups, but no significant differences were observed. Therefore, both ROMO and BP may slightly improve bone quality in GIOP.

Serum levels of sclerostin were markedly elevated in the ROMO group, consistent with our RomGo study.10 This increase is hypothesized to result from the prolonged elimination half-life of sclerostin due to its complex formation with ROMO. ROMO is a high-affinity antibody against sclerostin, and a similar mechanism is thought to apply here, as serum sclerostin levels increased in a dose-dependent manner following ROMO administration in the phase I study.15 Consequently, it is assumed that free sclerostin levels decreased in the ROMO group. Incidentally, GCs have been shown to suppress Wnt signaling and reduce bone formation by increasing the production of Wnt signaling inhibitors, such as sclerostin and Dkk-1, while decreasing ligands like Wnt3a.16,17 In the ROMO group, both free sclerostin and Dkk-1 levels decreased, indicating a reduction in Wnt signaling inhibitors, whereas the ligand Wnt3a increased. Furthermore, the Wnt3a/Dkk-1 ratio, representing the ligand-to-inhibitor ratio, increased, suggesting that Wnt signaling may be activated, thereby promoting bone formation. Conversely, in the BP group, although sclerostin and Dkk-1 levels increased, Wnt3a also increased. The Wnt3a/Dkk-1 ratio increased, but the Wnt3a/sclerostin ratio decreased, indicating that while Wnt signaling-mediated promotion of bone formation was observed, it was less pronounced than in the ROMO group.

No significant changes were observed in serum levels of RANKL and OPG in either group, but the rate of decrease in RANKL/OPG was greater in the ROMO group than in the BP group, suggesting that increased bone resorption due to promotion of osteoclast differentiation and maturation may have been more suppressed in the ROMO group.

Regarding adverse events, 1 patient in the ROMO group experienced a cerebral infarction. Because the patient was elderly and had vasculitis syndrome, the risk of cerebrovascular events was thought to be high, and the causal relationship with ROMO is unclear. Further studies are needed to establish the cerebrovascular and cardiovascular safety of ROMO in patients receiving long-term GCs.

This study has several limitations. First, it was an observational study, which may have led to selection bias. Second, due to the COVID-19 pandemic, some patients refused to receive ROMO, which requires monthly hospital visits. This hindered patient enrollment as planned, resulting in a small sample size. Third, some baseline characteristics differed between the ROMO and BP groups, including age, history of vertebral fractures, and BMD of the FN and TH. Due to the small number of patients analyzed, we were unable to adjust for variables other than age. These factors may have influenced the results. Furthermore, the limited number of patients and the short observation period make it difficult to conduct a comparative assessment of fracture prevention effects and to adequately evaluate safety issue. Therefore, future large-scale, long-term studies are needed to clarify the effects of ROMO administration on BMD increase and fracture prevention.

In conclusion, this study demonstrated that the sequential use of ROMO increased BMD of the LS, FN and TH compared to continuous administration of BP in patients with GIOP who had long-term BP exposure. However, the differences did not reach statistical significance. Future studies with longer durations and larger patient populations are warranted.