Factors influencing therapeutic efficacy of denosumab against osteoporosis in systemic lupus erythematosus

Abstract

Objective

Osteoporosis is a common comorbidity in patients with SLE, and bone loss in patients with SLE has a multifactorial aetiology. This study aimed to evaluate the therapeutic efficacy of denosumab in patients with SLE with osteoporosis and to analyse the factors influencing therapeutic efficacy.

Methods

A total of 166 patients with SLE with osteoporosis who initiated denosumab between January 2016 and December 2023 were included. Changes in the T-score and areal bone mineral density (BMD) at the lumbar spine, total hip and femur neck from denosumab initiation to 12 months were measured. Correlation analysis was performed between the degree of BMD improvement and covariates including SLE-specific factors such as SLE duration, SLE Disease Activity Index 2000 (SLEDAI-2K) score, glucocorticoid dose and hydroxychloroquine use. Multiple linear regression analysis was conducted to identify predictors of the therapeutic efficacy of denosumab.

Results

Denosumab significantly increased BMD and decreased bone turnover markers at 12 months compared with baseline. The degree of BMD improvement revealed a significant negative correlation with SLEDAI-2K score, hydroxychloroquine use, prior osteoporosis treatment and baseline BMD values. In contrast, body mass index and c-telopeptide of collagen type 1 levels were positively correlated with the degree of BMD improvement. Higher baseline BMD values, SLEDAI-2K scores and hydroxychloroquine use were significant predictors of attenuated BMD improvement.

Conclusions

Our study suggests that denosumab is an effective treatment option for osteoporosis in patients with SLE. The therapeutic efficacy of denosumab can be predicted by baseline BMD values, SLEDAI-2K scores and hydroxychloroquine use.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Patients with SLE have a higher risk and prevalence of osteoporosis compared with the general population due to multifactorial aetiology.

• Information on the therapeutic efficacy of denosumab in patients with SLE and potential factors that may attenuate its efficacy is lacking.

WHAT THIS STUDY ADDS

• Denosumab demonstrates therapeutic effectiveness against osteoporosis in patients with SLE; however, its efficacy may be attenuated by higher baseline bone mineral density (BMD) values, SLE Disease Activity Index 2000 (SLEDAI-2K) scores and hydroxychloroquine use.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• In the treatment of osteoporosis in patients with SLE, various factors beyond glucocorticoid use and vitamin D levels need to be considered.

Introduction

Advancements in diagnosis and treatment have improved the survival of patients with SLE; however, long-term comorbidities have become more prevalent.¹ Osteoporosis, which results in increased functional impairment and decreased quality of life by raising the risk of fragility fractures,² is one of the common comorbidities associated with SLE.¹ Studies have shown reduced bone mineral density (BMD) in patients with SLE compared with the general population.3,6 Reported incidence of osteoporosis in patients with SLE in cohort studies ranges from 1.4% to 68%.⁷ A meta-analysis reported an osteoporosis prevalence of 16% among patients with SLE, with a 2.03 times increased risk of developing osteoporosis compared with those without SLE.⁸ Systemic chronic inflammation, glucocorticoid (GC) use and low vitamin D levels are considered to contribute to the higher prevalence of osteoporosis in patients with SLE than in the general population.¹

Denosumab, a promising option for osteoporosis treatment, is a fully human monoclonal antibody that targets the receptor activator of NFκB ligand (RANKL).⁹ By preventing RANKL from interacting with RANK on osteoclasts, denosumab inhibits the activity and survival of osteoclasts.⁹ This antiresorptive mechanism leads to increased bone density and has proven effective in treating both postmenopausal10,12 and GC-induced osteoporosis.^{13 14} Regarding autoimmune diseases, denosumab has been shown to improve bone metabolism and inhibit joint erosion in patients with rheumatoid arthritis (RA).^{15 16}

However, studies evaluating the therapeutic efficacy of denosumab specifically in patients with SLE with osteoporosis are extremely limited. Sawamura et al reported that denosumab effectively prevents bone resorption and BMD loss in patients with inflammatory diseases, including SLE.¹³ Larger sample sizes focusing solely on patients with SLE are lacking. Furthermore, many reports have addressed the risk factors for osteoporosis and fragility fractures in patients with SLE,517,20 and the aetiology of bone loss in SLE is known to be multifactorial.^{21 22} Based on this, the therapeutic efficacy of denosumab in patients with SLE may also be multifactorial; however, no studies to date have analysed the factors associated with denosumab therapeutic efficacy.

This study aimed to evaluate the therapeutic efficacy of denosumab in increasing BMD in patients with SLE with osteoporosis, as well as to identify factors associated with the degree of BMD improvement. Additionally, we sought to determine the predictors influencing the therapeutic efficacy of denosumab using multiple linear regression analysis.

Methods

Study population

We retrospectively obtained medical records of patients with SLE who visited Seoul St. Mary’s Hospital, a university-affiliated hospital and tertiary referral centre in Korea. The diagnosis of SLE was based on the 2019 EULAR and American College of Rheumatology classification criteria.²³ Initially, data from 206 patients with SLE who initiated and received at least two doses of denosumab (60 mg via subcutaneous injection) for osteoporosis between January 2016 and December 2023 were screened. Patients who started or continued bisphosphonates or other osteoporosis treatments during this period were not included. Osteoporosis was diagnosed according to the WHO criteria.²⁴ We excluded 40 patients who had intervals longer than 6 months between denosumab administration or who did not undergo BMD measurement. Ultimately, 166 patients with SLE who initiated and received denosumab at 6-month intervals and underwent BMD measurements at the denosumab initiation and after 12 months were included in the study.

Covariates

Data on age, sex, menopausal status, smoking history, body mass index (BMI), SLE disease duration, SLE Disease Activity Index 2000 (SLEDAI-2K) score,²⁵ complement 3 and 4 (C3 and C4) levels, anti-double stranded DNA antibody levels, calcium levels, 25-hydroxyvitamin D levels, cumulative GC dose, duration of GC treatment, concurrent use of hydroxychloroquine, prior osteoporosis medications and bone turnover markers at the time of denosumab initiation were collected as baseline data. The cumulative GC dose refers to the total amount of GCs administered prior to denosumab initiation. The duration of GC treatment indicates the period of GC use before the initiation of denosumab. Additionally, data on concurrent GC doses administered from the time of denosumab initiation to 12 months were collected. Cumulative and concurrent GC doses were calculated by multiplying the daily dose recorded at each clinic visit by the treatment duration for each dose. All GC doses are expressed as prednisolone equivalents.

Bone mineral density and bone turnover markers measurements

BMD values at baseline and after 12 months were collected from the medical records. T-scores and areal BMD (aBMD) of the lumbar spine, total hip and femur neck were assessed using dual-energy X-ray absorptiometry (DEXA) devices (Horizon W, Hologic). The DEXA assessment of the lumbar spine considered lumbar levels 1–4, excluding any vertebral bodies affected by compression fractures. For patients with avascular bone necrosis of the hip or those who had undergone joint replacement, the BMD of the opposite hip was used. The T-score difference was defined as the value obtained by subtracting the baseline T-score from the T-score at 12 months. The change in aBMD was defined as the value obtained by subtracting the baseline aBMD from the aBMD at 12 months, divided by the baseline aBMD and multiplying by 100.

Serum bone turnover markers, including the C-telopeptide of collagen type 1 (CTx) for bone resorption and procollagen 1 N-terminal propeptide (P1NP) for bone formation, were measured using an electrochemiluminescence immunoassay at baseline and 12 months.

Statistical analysis

The normality of the distribution of continuous variables was assessed using the Kolmogorov-Smirnov test. Continuous variables are expressed as mean and SD. Categorical variables are expressed as frequencies and percentages. A paired t-test was used to assess the significance of changes from baseline to 12 months. Pearson’s correlation analysis was conducted to assess the relationship between the covariates and the therapeutic efficacy of denosumab. Multiple linear stepwise regression was used to identify the predictors influencing the therapeutic efficacy of denosumab. All statistical analyses were performed using the SPSS software (V.24.0; International Business Machine Corporation, Chicago, Illinois, USA). Statistical significance was set at a p value of <0.05.

Results

Characteristics of the study population

The baseline characteristics of the patients with SLE are presented in table 1. Of the 166 patients included in this study, the mean age was 54.0±11.0 years. The majority of patients (97.0%) were female, and 71.4% were postmenopausal. Approximately 2.4% of patients had a history of smoking. The mean BMI was 21.65±3.24 kg/m². The average duration of SLE was 16.2±8.5 years, with a mean SLEDAI-2K score of 3.1±3.2. The SLEDAI-2K scores ranged from 0 to 12, with 33.1% of patients in remission at baseline. The mean levels of C3, C4 and anti-DNA antibodies were 87.8±20.4 mg/dL, 19.68±9.03 mg/dL and 12.657±25.627 IU/mL, respectively. The mean calcium and 25-hydroxyvitamin D levels were 9.11±0.46 mg/dL and 30.159±14.572 ng/mL, respectively. The mean baseline T-scores for the lumbar spine, total hip and femur neck were −2.32±0.91, –1.9±0.78 and −2.34±0.71, with the proportion of patients having a T-score ≤−2.5 being 46.4%, 17.5% and 39.4%, respectively. The baseline mean aBMD values for the lumbar spine, total hip and femur neck were 0.7998±0.0973 g/cm², 0.7192±0.0880 g/cm² and 0.6079±0.0730 g/cm², respectively. The mean cumulative GC dose administered prior to denosumab initiation was 19159.28±17 349.58 mg. The average GC treatment duration was 197.6±99.2 months. Since studies have reported that the highest rate of bone loss occurred within the first 3–6 months of GC treatment,²⁶ we investigated the proportion of patients with a GC treatment duration of 6 months or less. Only 2.5% of the patients were within 6 months of GC treatment. The mean GC dose administered during the 12 months following denosumab initiation was 1480.56±1475.81 mg, and 59.6% of patients were taking hydroxychloroquine. The mean dosage of hydroxychloroquine was 201.2±50.5 mg/day. Regarding previous osteoporosis treatment, 72.9% were treatment-naïve, 19.3% had previously received bisphosphonates and 7.8% had been treated with selective oestrogen receptor modulators. The mean bone turnover markers were 0.38939±0.23692 ng/mL for CTx and 48.70±35.04 ng/mL for P1NP.

Table 1. Baseline clinical and laboratory variables of included patients with SLE (n=166).

Variables	Values
Age, years	54.0 (11.0)
Female	161 (97.0)
Postmenopause	115/161 (71.4)
Smoking history	4 (2.4)
BMI, kg/m2	21.65 (3.24)
Duration of SLE, years	16.2 (8.5)
SLEDAI-2K score	3.1 (3.2)
C3, mg/dL	87.8 (20.4)
C4, mg/dL	19.68 (9.03)
Anti-DNA antibody, IU/mL	12.657 (25.627)
Calcium, mg/dL	9.11 (0.46)
25-Hydroxyvitamin D, ng/mL	30.159 (14.572)
T-score
Lumbar spine	−2.32 (0.91)
≤−2.5	77 (46.4)
Total hip	−1.90 (0.78)
≤−2.5	28/160 (17.5)
Femur neck	−2.34 (0.71)
≤−2.5	41/104 (39.4)
Areal BMD, g/cm2
Lumbar spine	0.7998 (0.0973)
Total hip	0.7192 (0.0880)
Femur neck	0.6079 (0.0730)
Cumulative GC, mg	19 159.28 (17 349.58)
Duration of GC treatment, months	197.6 (99.2)
≤6 months	4/161 (2.5)
Concurrent GC, mg	1480.56 (1475.81)
Hydroxychloroquine use	99 (59.6)
Prior osteoporosis treatment
None	121 (72.9)
Bisphosphonates	32 (19.3)
SERMs	13 (7.8)
Bone turnover markers
CTx, ng/mL	0.38939 (0.23692)
P1NP, ng/mL	48.70 (35.04)

Data are shown as mean (standard deviationSD) or n (%).

BMDbone mineral densityBMIbody mass indexCTxC-telopeptide of collagen type 1GCglucocorticoidP1NPprocollagen 1 N-terminal propeptideSERMselective oestrogen receptor modulatorSLEDAI-2KSLE Disease Activity Index 2000

BMD value changes from baseline to 12 months

Figure 1 presents the changes in the T-scores, aBMD and bone turnover markers from baseline to 12 months in the included patients with SLE. The mean T-score differences from baseline to 12 months were 0.112 (95% CI 0.061 to 0.162) for the lumbar spine, 0.117 (95% CI 0.084 to 0.150) for the total hip and 0.106 (95% CI 0.080 to 0.132) for the femur neck (figure 1A). The mean aBMD changes from baseline to 12 months were 1.7226% (95% CI 1.0316 to 2.4136) in the lumbar spine, 2.0604% (95% CI 1.4897 to 2.6310) in the total hip and 2.1312% (95% CI 1.6566 to 2.6058) in the femur neck (figure 1B). Compared with baseline, there were statistically significant increases in all parameters at 12 months (all p<0.01). Serum CTx levels at baseline and 12 months were 0.38939±0.23692 ng/mL and 0.18108±0.13572 ng/mL, respectively (figure 1C). Serum P1NP levels at baseline and 12 months were 48.70±35.04 ng/mL and 23.36±9.95 ng/mL, respectively (figure 1D). Both serum CTx (p<0.01) and P1NP (p=0.043) levels significantly decreased at 12 months compared with baseline.

Figure 1. (A) T-score differences (∆T-score) and (B) areal bone mineral density (aBMD) changes from baseline to 12 months. Bars indicate the mean±95% CI. Bone turnover markers, (C) c-telopeptide of collagen type 1 (CTx) and (D) procollagen 1 N-terminal propeptide (P1NP), at baseline and 12 months. Bars indicate the mean±SD. *P<0.05, **p<0.01 vs baseline.

Factors associated with the therapeutic efficacy of denosumab

To evaluate the factors associated with the therapeutic efficacy of denosumab, a correlation analysis was performed between covariates and the differences in T-scores and changes in aBMD for the lumbar spine, total hip and femur neck (table 2). The T-score difference and aBMD change in the lumbar spine revealed a statistically significant negative correlation with the SLEDAI-2K score (T-score difference, p=0.003; aBMD change, p=0.002), prior osteoporosis treatment (T-score difference, p=0.007; aBMD change, p=0.007) and baseline BMD value (T-score difference, p=0.000; aBMD change, p=0.000), and a positive correlation with the serum CTx level (T-score difference, p=0.017; aBMD change, p=0.010). The T-score difference and aBMD change in the total hip revealed a significant negative correlation with the baseline BMD value (T-score difference, p=0.008; aBMD change, p=0.001). The aBMD change in the total hip was positively correlated with the CTx level (p=0.047). The T-score difference and aBMD change in the femur neck exhibited a statistically significant negative correlation with the SLEDAI-2K score (T-score difference, p=0.039; aBMD change, p=0.009) and hydroxychloroquine use (T-score difference, p=0.032; aBMD change, p=0.025). The change in aBMD of the femur neck was positively correlated with BMI (p=0.041). Concurrent GC doses tended to negatively correlate with T-score differences and aBMD changes at all sites; however, no statistical significance was observed. The cumulative GC dose and 25-hydroxyvitamin D levels did not significantly correlate with the degree of BMD improvement at any site. Males, patients with a smoking history and those with a GC treatment duration of <6 months were not analysed as covariates due to their minor proportions (<5%).

Table 2. Correlation analysis of covariates and therapeutic efficacy of denosumab.

Covariates	Lumbar spine				Total hip				Femur neck
	T-score difference		aBMD change (%)		T-score difference		aBMD change (%)		T-score difference		aBMD change (%)
	r	P value	r	P value	r	P value	r	P value	r	P value	r	P value
Age, years	0.072	0.360	0.079	0.322	0.127	0.110	0.116	0.143	0.031	0.751	0.025	0.804
Postmenopause	0.003	0.970	0.022	0.784	0.107	0.184	0.096	0.233	0.015	0.885	0.010	0.920
BMI, kg/m2	0.029	0.730	0.018	0.830	0.013	0.880	−0.051	0.540	0.190	0.064	0.210	0.041*
Duration of SLE, years	−0.028	0.725	−0.010	0.901	0.047	0.551	0.020	0.804	0.016	0.872	0.048	0.629
SLEDAI-2K score	−0.228	0.003**	−0.240	0.002**	−0.129	0.105	−0.068	0.394	−0.202	0.039*	−0.255	0.009**
Calcium, mg/dL	0.122	0.136	0.139	0.090	0.100	0.223	0.074	0.369	0.124	0.224	0.160	0.117
25-Hydroxyvitamin D, ng/mL	−0.169	0.057	−0.169	0.059	−0.008	0.933	0.014	0.878	0.006	0.955	0.009	0.935
Cumulative GC, mg	−0.096	0.224	−0.096	0.227	0.086	0.280	0.049	0.538	0.065	0.508	0.040	0.686
Concurrent GC, mg	−0.143	0.069	−0.154	0.051	−0.089	0.263	−0.040	0.614	−0.079	0.421	−0.116	0.241
Hydroxychloroquine use	0.021	0.791	0.010	0.897	−0.010	0.903	0.047	0.556	−0.209	0.032*	−0.220	0.025*
Prior osteoporosis treatment	−0.211	0.007**	−0.213	0.007**	−0.029	0.720	−0.004	0.962	−0.072	0.463	−0.077	0.438
CTx, ng/mL	0.329	0.017*	0.356	0.010*	0.260	0.065	0.279	0.047*	0.040	0.831	0.034	0.857
P1NP, ng/mL	0.409	0.362	0.363	0.424	0.257	0.578	−0.290	0.528	0.676	0.095	0.262	0.571
Baseline BMD value	−0.320	0.000**	−0.319	0.000**	−0.210	0.008**	−0.266	0.001**	−0.079	0.423	−0.131	0.184

Baseline BMD value: T-score or aBMD value at denosumab initiation for the respective site.

*P<0.05, **P<0.01.

aBMDareal bone mineral densityBMDbone mineral densityBMIbody mass indexCTxc-telopeptide of collagen type 1GCglucocorticoidP1NPprocollagen 1 N-terminal propeptideSLEDAI-2KSLE Disease Activity Index 2000

Multiple linear regression models of predicting therapeutic efficacy of denosumab in SLE

Multiple linear stepwise regression analysis was conducted on factors that showed a significant association with the therapeutic efficacy of denosumab (table 3). The T-score difference and aBMD change in the lumbar spine were significantly predicted by baseline BMD value (T-score difference: β=−0.170, p=0.001; aBMD change: β=−21.484, p=0.001). Similarly, the T-score difference and aBMD change in the total hip were significantly predicted by baseline BMD value (T-score difference: β=−0.058, p=0.008; aBMD change: β=−19.413, p=0.002). The T-score difference and aBMD change in the femur neck were significantly predicted by the baseline SLEDAI-2K score (T-score difference: β=−0.010, p=0.037; aBMD change: β=−0.238, p=0.012) and hydroxychloroquine use (T-score difference: β=−0.074, p=0.031; aBMD change: β=−1.341, p=0.042).

Table 3. Multiple linear regression models of factors influencing therapeutic efficacy of denosumab in patients with SLE.

Predictors	Lumbar spine				Total hip				Femur neck
	T-score difference		aBMD change (%)		T-score difference		aBMD change (%)		T-score difference		aBMD change (%)
	β	P value	β	P value	β	P value	β	P value	β	P value	β	P value
BMI, kg/m2	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	0.160	0.110
SLEDAI-2K score	−0.070	0.582	−0.047	0.718	N/A	N/A	N/A	N/A	−0.010	0.037*	−0.238	0.012*
Hydroxychloroquine use	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A	−0.074	0.031*	−1.341	0.042*
Prior osteoporosis treatment	−0.070	0.585	−0.040	0.759	N/A	N/A	N/A	N/A	N/A	N/A	N/A	N/A
CTx, ng/mL	0.197	0.141	0.240	0.071	N/A	N/A	0.130	0.360	N/A	N/A	N/A	N/A
Baseline BMD value	−0.170	0.001**	−21.484	0.001**	−0.058	0.008**	−19.413	0.002**	N/A	N/A	N/A	N/A
R2	0.211		0.202		0.044		0.182		0.084		0.106
Adjusted R2	0.195		0.186		0.038		0.165		0.066		0.087

N/A: Not applicable. Baseline BMD value: T-score or aBMD value at denosumab initiation for the respective site.

^*P<0.05, **P<0.01.

aBMDareal BMDBMDbone mineral densityBMIbody mass indexCTxC-telopeptide of collagen type 1N/Anot applicableSLEDAI-2KSLE Disease Activity Index 2000

Overall, the results indicated that baseline BMD values are possible predictors of T-score differences and aBMD changes in the lumbar spine and total hip. Baseline SLEDAI-2K scores and hydroxychloroquine use are possible predictors of T-score differences and aBMD changes in the femur neck.

Discussion

Our study showed that denosumab significantly increased T-scores and aBMD, and decreased bone turnover markers at 12 months compared with baseline in patients with SLE with osteoporosis (figure 1). The degree of lumbar spine BMD improvement was significantly associated with SLEDAI-2K score, prior osteoporosis treatment, CTx level and baseline BMD value. Improvement in total hip BMD was significantly associated with CTx and baseline BMD values. The degree of improvement in femur neck BMD was significantly associated with BMI, SLEDAI-2K score and hydroxychloroquine use (table 2). Multiple linear regression analysis of associated factors revealed that baseline BMD values were significant predictors of the therapeutic efficacy of denosumab in the lumbar spine and total hip, while baseline SLEDAI-2K score and hydroxychloroquine use were significant predictors of the therapeutic efficacy of denosumab in the femur neck (table 3).

In this study, denosumab significantly reduced bone turnover markers and improved BMD values in the lumbar spine, total hip and femur neck after 12 months in patients with SLE with osteoporosis. These findings are consistent with those of a previous study that evaluated the efficacy of denosumab after 12 months in patients treated with long-term GCs for inflammatory diseases, primarily SLE.¹³ The tendency of relatively smaller aBMD increases in the lumbar spine in our study may be attributable to its higher baseline aBMD values compared with the hips. The significance of our study lies in its larger sample size, the exclusive focus on patients with SLE and the additional evaluation of total hip BMD values.

The SLEDAI-2K score, hydroxychloroquine use, prior osteoporosis treatment and baseline BMD values revealed a significant negative correlation with the therapeutic efficacy of denosumab in patients with SLE. Since systemic chronic inflammation can lead to bone loss, numerous previous studies have established that SLE inherently increases the risk of osteoporosis.^{1 21 22 27} In terms of fractures, Nakajima et al reported that vertebral fractures were associated with the Systemic Lupus International Collaborating Clinics and American College of Rheumatology Damage Index score,¹⁷ and Carli et al identified disease duration as one of the risk factors for osteoporosis and fragility fractures in patients with SLE.⁵ Our findings suggest that a higher SLEDAI-2K score at the initiation of denosumab treatment may attenuate therapeutic efficacy. The increased expression of RANKL and accelerated osteoclastogenesis observed in patients with SLE¹ may be more pronounced in those with high disease activity, potentially reducing the degree of BMD improvement achieved with denosumab, a monoclonal antibody targeting RANKL. The effects of hydroxychloroquine on bone resorption reported in in vitro and in vivo studies are inconsistent.^{28 29} Reports on the association between hydroxychloroquine use and BMD in patients with SLE are lacking,^{17 30 31} with most studies focusing primarily on lumbar spine BMD. The findings of our study suggest that patients with SLE using hydroxychloroquine showed a lower degree of femur neck BMD improvement with denosumab than those not taking hydroxychloroquine. Further studies are needed to investigate the possible mechanisms of interaction between hydroxychloroquine and osteoporosis therapeutic agents, including denosumab, while also considering the compositional differences between spinal and femoral bones. In this study, patients with prior osteoporosis treatment, including bisphosphonates, revealed lower therapeutic efficacy of denosumab in the lumbar spine than the treatment-naïve group. This finding is consistent with a previous study reporting that denosumab treatment was less effective in the trabecular bone, which predominantly comprises the lumbar spine, in the prior bisphosphonate group than in the naïve group.³² This reduced efficacy is thought to be due to the higher concentrations and accumulated effects of bisphosphonates in the trabecular bone than in the cortical bone.^{33 34} Nevertheless, the relatively weaker association of osteoporosis treatment history with BMD improvements, compared with baseline BMD values, may have contributed to its exclusion from the multiple regression model. Additionally, our results indicated that lower baseline BMD in the lumbar spine and total hip leads to a greater increase in BMD in response to denosumab. This finding is supported by previous osteoporosis drug trials, which have shown that treatment is more effective in individuals with low BMD.^{35 36}

In this study, BMI and CTx levels were significantly positively correlated with the therapeutic efficacy of denosumab. A low BMI has been reported as an independent risk factor for increased bone loss²² and is associated with lower lumbar and femoral BMD¹⁷ in patients with SLE. However, further studies are needed to elucidate the mechanisms underlying the attenuation of the therapeutic efficacy of denosumab in the femur neck. CTx is a marker of bone resorption, and it is well-known that denosumab acts as an antiresorptive agent by decreasing bone resorption and inhibiting bone turnover.^{9 13} Eastell et al reported significant correlations between reduced CTx and increased BMD in patients with postmenopausal osteoporosis.³⁷ In this study, the association between high baseline CTx levels and greater therapeutic efficacy of denosumab is thought to be attributable to its mechanism of action.

Interestingly, the therapeutic efficacy of denosumab revealed no significant association with factors such as the GC dose or vitamin D level, both of which are known to contribute to bone loss in patients with SLE.¹ In patients with RA, systematic reviews have shown that GC use does not affect the therapeutic efficacy of denosumab.^{15 16} Although studies on the therapeutic efficacy of denosumab in patients with SLE are lacking, reports have suggested that GC dose is not associated with an increased risk of fractures.^{17 18} The findings regarding the 25-hydroxyvitamin D level align with that of a previous study reporting that baseline vitamin D sufficiency status does not affect denosumab efficacy related to increasing BMD.³⁸

Our study has certain limitations. First, it was conducted at a single centre, which could have introduced selection bias. Second, sex, smoking history and GC treatment duration of <6 months could not be analysed as covariates due to the proportion imbalance arising from the retrospective study design. Therefore, future prospective cohort studies exploring the relationship between these factors and the therapeutic efficacy of denosumab are needed. Moreover, our analysis was restricted to 12 months following the initial administration of denosumab. Further longitudinal studies are required to determine the long-term efficacy of denosumab.

Despite these limitations, our findings suggest that denosumab significantly improves BMD in both the lumbar spine and femur in patients with SLE with osteoporosis. However, factors such as low BMI, high SLEDAI-2K scores, hydroxychloroquine use, prior osteoporosis treatment, low CTx levels and high BMD values at denosumab initiation negatively correlated with the degree of BMD improvement. The findings of multiple linear regression revealed that higher baseline BMD values were predictors of lower therapeutic efficacy of denosumab at the lumbar spine and total hip, while higher SLEDAI-2K scores and hydroxychloroquine use were predictors of lower efficacy at the femur neck.

In conclusion, our study shows that denosumab is a promising treatment option for osteoporosis in patients with SLE. The therapeutic efficacy of denosumab in patients with SLE could be predicted by baseline BMD values, SLEDAI-2K scores and hydroxychloroquine use, rather than by GC dose or vitamin D levels.

Data availability statement

Data are available on reasonable request.