Three-drug combination therapy to prevent glucocorticoid-associated osteonecrosis in patients with systemic lupus erythematosus: a proof-of-concept study

Abstract

Background:

Systemic lupus erythematosus (SLE) is a major underlying disease of glucocorticoid-associated osteonecrosis of the femoral head (ONFH). Despite its clinical significance, no prophylactic treatment has been established to prevent ONFH in patients receiving systemic glucocorticoid therapy.

Objectives:

To investigate the efficacy and safety of a three-drug combination therapy consisting of clopidogrel sulfate, pitavastatin calcium hydrate, and tocopherol acetate, administered concurrently with initial glucocorticoid therapy to prevent ONFH in patients with SLE.

Design:

A multicenter, single-arm, interventional clinical trial conducted as an advanced medical treatment approved by the Ministry of Health, Labor and Welfare of Japan.

Methods:

This study was conducted at 12 sites in Japan between August 2014 and March 2024. Patients with SLE who required initial glucocorticoid therapy (⩾0.5 mg/kg/day of prednisolone) received the three study drugs concurrently with glucocorticoids for 90 days. Magnetic resonance imaging of both hip joints was performed 180 days after initiation of glucocorticoid therapy to determine ONFH occurrence. The primary endpoint was ONFH incidence, and safety and potential risk factors were also evaluated using logistic regression analysis.

Results:

Of the 50 enrolled patients, 43 completed the 90-day regimen. ONFH was identified in 8 of 43 patients (18.6%), which was below the threshold incidence of 25% based on the historical control, suggesting a potential signal of reduced incidence (p = 0.1664). Treatment-emergent adverse events were observed in 19 patients; the only severe adverse event was a drug eruption in one patient. The exploratory analysis identified the period of drinking as a significant risk factor for ONFH occurrence.

Conclusion:

The findings demonstrate the feasibility and acceptable safety of this three-drug combination therapy. Although the results should be regarded as preliminary and hypothesis-generating, they suggest a potential signal that warrants further investigation in adequately powered randomized controlled trials.

Trial registration:

Clinical trial for the control of osteonecrosis of the femoral head secondary to the initial corticosteroid treatment in patients with systemic lupus erythematosus (UMIN000008230; https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000009636).

Introduction

Osteonecrosis of the femoral head (ONFH) is a refractory disease that causes irreversible hip joint dysfunction due to femoral head collapse and is commonly associated with systemic glucocorticoid therapy.^1,2 Furthermore, ONFH development is influenced by multiple factors, including lifestyle-related factors, such as alcohol consumption and smoking, highlighting the multifactorial nature of this condition. ³ Despite the established association between glucocorticoid therapy and ONFH,^4,5 the precise mechanism by which glucocorticoids induce osteonecrosis remains unclear, and no established interventions are available to prevent its occurrence.

Systemic lupus erythematosus (SLE) is a systemic autoimmune disease characterized by diverse organ manifestations caused by the production of various autoantibodies. ⁶ SLE, which often requires high-dose glucocorticoid treatment, is the most common underlying disease of glucocorticoid-associated ONFH, ¹ accounting for approximately 30% of ONFH cases. ⁴ While the reported incidence of ONFH in patients with SLE varies,^7,8 prospective studies that followed patients from initial glucocorticoid therapy reported relatively high ONFH incidences of 25.0%–44.4%.^9–11 Therefore, developing a prophylaxis to prevent ONFH in patients with SLE is crucial. However, accumulating evidence suggests that, in addition to glucocorticoid exposure, SLE-related factors—particularly active disease itself^12,13 and the presence of antiphospholipid antibodies^14,15—may also contribute to the development of ONFH, indicating that establishing effective prophylactic strategies is not straightforward in clinical practice.

Human and animal studies have suggested several possible factors that may be involved in the pathogenesis of glucocorticoid-associated ONFH, including coagulation abnormalities,^16,17 dyslipidemia,^16,18 and oxidative stress.^19,20 Based on these findings, several animal-model studies have shown the ONFH preventive effects of an antiplatelet drug (clopidogrel sulfate), ²¹ a 3-hydroxymethyl-3-glutaryl-CoA reductase inhibitor (pitavastatin calcium hydrate), ²² and vitamin E (α-tocopherol).^23,24 Additionally, the combination of two drugs with different mechanisms (anticoagulant and lipid-lowering agents) has demonstrated greater inhibitory effects on the occurrence of glucocorticoid-associated osteonecrosis than monotherapy. ²⁵ These findings suggest that the pathophysiology of ONFH induced by glucocorticoids is multifactorial and that a combination of multiple drugs may be effective in controlling such abnormalities.

To date, two clinical studies have attempted to prevent the occurrence of ONFH by concurrently administering drugs with initial glucocorticoid therapy for SLE.^11,26 Both studies examined the effect of warfarin potassium, an anticoagulant, or atorvastatin calcium hydrate alone on the prevention of ONFH; however, neither showed a difference from the control group that did not receive any preventive drugs,^11,26 suggesting the difficulty of preventing ONFH with a single drug in humans.

Therefore, we aimed to conduct a proof-of-concept study to investigate the efficacy and safety of a three-drug combination therapy consisting of clopidogrel sulfate, pitavastatin calcium hydrate, and tocopherol acetate administered concurrently with initial glucocorticoid therapy to prevent ONFH in patients with SLE.

Methods

Study design

The multicenter, single-arm, interventional clinical trial (UMIN000008230) was conducted at 12 sites in Japan between August 1, 2014, and March 31, 2024, as an advanced medical treatment with the approval of the Ministry of Health, Labor and Welfare of Japan. The three study drugs were administered concurrently with the initial glucocorticoid therapy for 90 days. Magnetic resonance imaging (MRI) of the bilateral hip joint was performed 180 days later to confirm ONFH. The trial was conducted in accordance with the Declaration of Helsinki and the Clinical Trials Act (https://www.mhlw.go.jp/file/06-Seisakujouhou-10800000-Iseikyoku/0000213334.pdf). The study protocol, informed consent forms, and other study-related documents were reviewed and approved by the local or central Institutional Review Boards at all study sites. The Independent Data Monitoring Committee, composed of clinical trial and orthopedic experts, periodically reviewed and evaluated the trial to ensure participant safety and monitor their conduct and progress. Written informed consent was obtained from all patients before enrollment.

Interim analysis and sample size calculation

An interim analysis was set up for this trial owing to ethical considerations, aiming to minimize the number of patients who participated in this trial if prophylactic treatment was ineffective. Based on previous reports indicating an ONFH incidence of 25.0%–44.4% in patients newly diagnosed with SLE,^9–11 the threshold ONFH incidence rate was set at 0.25, and expected ONFH incidence rate was 0.19 under the null hypothesis, with a one-sided type I error of 20% and a power of 80%.

The ratio of the amount of information at the time of the interim analysis to that at the time of final analysis was set at 0.33:1. Critical values were identified using Pocock’s method to accept the null hypothesis for the lower test on proportions. Consequently, the required number of enrolled cases in the interim and final analyses was estimated to be 48.87 and 146.62, respectively; therefore, the number of enrolled cases was set to 50 and 150, respectively.

The trial was considered invalid and discontinued if ONFH occurred in more than 25% of patients in the interim analysis. A report on the interim analysis, prepared by a statistician from the Independent Data Monitoring Committee, was reviewed by the committee to determine whether to continue the trial.

Patients

The inclusion criteria were as follows: patients (1) with a diagnosis of SLE based on the diagnostic criteria of the American College of Rheumatology, ²⁷ (2) who required initial glucocorticoid therapy for SLE at a dose of ⩾0.5 mg/kg/day of prednisolone, (3) who were hospitalized at the start of the study drug administration, (4) who were at least 20 years of age at the time of consent, (5) who provided written consent to participate in the trial, and (6) whose latest laboratory values within 14 days before enrollment met all of the following criteria—a platelet count ⩾70,000/mL, aspartate aminotransferase and alanine aminotransferase levels ⩽100 U/L, and a creatinine clearance of >50 mL/min. Patients were excluded from the trial if they met any one of the following criteria: (1) a history of systemic glucocorticoid administration; (2) a history of or suffering from ONFH; (3) active hemorrhagic lesions; (4) active infection; (5) contraindicated to any of the study drugs; (6) a history of hypersensitivity to any of the study drugs; (7) pregnant, lactating, or possibly pregnant patients; (8) use of any of the following drugs contraindicated with glucocorticoids or study drugs: (i) antiplatelet agents other than clopidogrel sulfate, (ii) anticoagulants, (iii) fibrate drug, and (iv) cyclosporine; (9) contraindication to MRI examination; and (10) those deemed inappropriate for the study by an investigator or a sub-investigator at each site. Patients could participate in this trial even if they had been taking all or any of the three study drugs immediately before starting the study drug administration or had taken any of the three drugs in the past. Patients could take any immunosuppressive agent other than cyclosporine, which is a contraindication for pitavastatin calcium hydrate. Patients on statins other than pitavastatin calcium hydrate could participate in the trial as long as they discontinued the drug at least 1 day before study drug administration.

Procedure

A centralized web-based registration system was used for case enrollment. An investigator or sub-investigator at each site entered the basic information of eligible patients who provided written consent. Basic patient information included age, sex, height, weight, and drinking and smoking history. MRI of the bilateral hip joint was performed within 28 days before starting initial glucocorticoid therapy to confirm the absence of ONFH. Conversely, an MRI was performed within 1 week after starting initial glucocorticoid therapy if it was difficult to obtain MRI findings before the procedure, such as when glucocorticoid therapy had to be started immediately.

Eligible patients were hospitalized and started on initial glucocorticoid therapy. The timing of discharge was determined by an investigator or a sub-investigator at each site based on the patient’s medical condition. All glucocorticoids and concomitant medications administered for SLE during the study period were recorded in a case report form. The three study drugs were administered simultaneously or immediately before starting initial glucocorticoid therapy (on the same day) and continued for 90 days in principle: (1) clopidogrel sulfate, 300 mg orally once on the first day only, then 75 mg orally once daily from the second day onward; (2) pitavastatin calcium hydrate, 2 mg orally once a day; and (3) tocopherol acetate, 100 mg orally three times a day (300 mg per day). Medication compliance was verified using a medication diary maintained by the patient. Single-medication noncompliance was defined as failure to take any one of the three study drugs and was evaluated three times per day. The patients were monitored for medication noncompliance every 30 days during the 90-day treatment period. Patients were discontinued from the study if they had more than 19 noncompliance events in 30 days to achieve a medication compliance rate of at least 80%. In addition, patients were discontinued from the study if they had ⩾10 noncompliance events within the first 15 days of study drug administration.

Routine laboratory tests were performed 14 days before the start to the day prior to the start and 14, 30, 60, 90, and 180 days after starting the study (up to 1 week before and after was allowed). SLE disease activity was assessed using the SELENA-SLEDAI SCORE ²⁸ 14 days before the start to the day prior to the start and at 30, 90, and 180 days after the start (up to 1 week before and after was allowed). MRI of the bilateral hip joint was performed if hip pain was observed during the study period to determine whether ONFH had occurred. Otherwise, MRI was performed 180 days later (up to 1 week before and after was allowed) if no ONFH occurrence was observed on the MRI performed during the study period or no hip pain was observed throughout the study period, which would complete the study protocol. The data obtained in this study were collected and managed at the Center for Clinical and Translational Research.

To minimize selection bias, patient eligibility criteria were predefined, and all patients were prospectively enrolled through a centralized web-based registration system before the initiation of the study drugs. Enrollment was based on clinical disease activity requiring initiation of high-dose glucocorticoid therapy, regardless of prior SLE treatment history or disease course. Participating centers were instructed to consecutively enroll all eligible patients who met the inclusion criteria during the study period.

Endpoints

The primary endpoint of the study was ONFH incidence. The diagnostic criterion for ONFH was a band-like low-signal region within the femoral head recognized on T1-weighted images. ²⁹ The presence or absence of ONFH was determined based on the results of an imaging review by an Independent Central Review Committee comprising three orthopedic experts. When ONFH was confirmed, the stage and type were classified according to the Japanese Investigative Committee classification. ²⁹ The secondary endpoint of the study was the safety of the combined administration of the three study drugs. Treatment-emergent adverse events (TEAEs) were defined as all unwanted or unintended injuries, illnesses, or signs (including abnormal laboratory values) that occurred in participants after starting the study treatment, regardless of whether they were causally related to the study drugs. TEAEs were recorded using MedDRA/J (Medical Dictionary for Regulatory Activities, Japanese version) version 18.1 System Organ Class and Preferred Term. ³⁰ Another secondary endpoint was the evaluation of known risk factors for ONFH and the identification of unknown risk factors. Patient background, total glucocorticoid dosage (up to 30, 90, and 180 days after initiation), maximum daily glucocorticoid dosage, presence of glucocorticoid pulse therapy, SLE disease activity, and laboratory data obtained before starting initial glucocorticoid therapy were also evaluated.

Statistical analysis

The incidence of ONFH was prospectively assumed to be <25% and analyzed using a normal approximation of the binomial distribution test within an exploratory, signal-detection framework, applying a one-sided significance level of 20% with a 60% Wald confidence interval. This statistical approach was predefined to allow detection of a potential preventive signal in this proof-of-concept study, rather than establish definitive efficacy. Logistic regression was used to explore the risk factors for ONFH. Univariate logistic regression was performed using patient background factors, glucocorticoid dosage, presence of glucocorticoid pulse therapy, SLE disease activity before initial glucocorticoid treatment, and laboratory values before initial glucocorticoid treatment as covariates to calculate odds ratios for the occurrence of ONFH, their Wald confidence intervals, and p-values. The two-sided significance level was set at 5%, with a 95% confidence level. Multivariate logistic regression analysis was performed with the total dose of glucocorticoids fixed as a covariate (three patterns: up to 30, 90, and 180 days), and the variables selected by the best subset selection were added. The same analysis was performed when the maximum daily glucocorticoid dose was fixed as a covariate. The two-sided significance level was set at 5%, with a 95% confidence level. Missing data were not imputed; analyses were conducted using available data only. SAS software (version 9.4; SAS Institute Inc., Cary, NC, USA) was used for statistical analysis.

We used the CONSORT reporting guideline ³¹ to draft this manuscript, and the CONSORT reporting checklist ³² when editing, included in Supplemental Documents.

Results

Trial progress

Overall, 50 patients were enrolled between January 2015 and March 2023 in this multicenter, single-arm, intervention clinical trial, and an interim analysis was conducted according to the protocol. Although the results of the interim analysis did not meet the criteria for discontinuation of this trial, the Independent Data Monitoring Committee decided to discontinue the trial because continuation would lead to a loss of reliability due to changes in the medical background with a further extended trial duration. Therefore, the principal investigator closed the trial, and the data collected were analyzed in the final analysis.

Patient characteristics

Among the 50 patients enrolled in this trial, 4 were discontinued before starting the study drug administration owing to protocol deviation (including violation of inclusion or exclusion criteria). Among the 46 patients for whom study drug administration was started, 3 discontinued treatment within 1 month: one withdrew consent, one deviated from the protocol, and one discontinued treatment due to the appearance of a drug eruption. The remaining 43 patients completed 90 days of the study drug administration (Figure 1). Medication compliance rates for the three study drugs were 97.8% at 30 days and 100% at 60 and 90 days after starting administration. The baseline characteristics of the 43 patients are shown in Table 1.

Figure 1.

Flow diagram of patient enrollment and study progression.

Table 1.

Patient characteristics.

Background factors	N = 43
Age (years) a	40.2 ± 14.6
Sex (female/male)	35/8
Body weight (kg) a	54.7 ± 8.5
Height (cm) a	161.5 ± 7.3
Smoking (yes/no)	14/29
Drinking (yes/no)	18/25
Period (years) b	23.3 ± 11.8
Frequency (days/week) b	3.4 ± 3.2
Average daily drinking (mL)b,c	379.7 ± 317.4
SELENA-SLEDAI score a	15.7 ± 8.3
ACR classification	(yes/no)
Malar rash	19/24
Discoid rash	2/41
Photosensitivity	4/39
Oral ulcers	4/39
Arthritis	30/13
Serositis	12/31
Renal disorder	18/25
Neurologic disorder	5/38
Hematologic disorder	36/7
Immunologic disorder	41/2
Antinuclear antibody	42/1
Anticardiolipin antibody (positive/negative/not examined)	20/22/1
Presence of ONFH before treatment	0

^aValues are presented as means ± standard deviation.

^bValues are given as means ± standard deviation of 18 cases with a history of drinking.

^cCalculated based on the alcohol content of each beverage.

ACR, The American College of Rheumatology; ONFH, osteonecrosis of the femoral head; SELENA-SLEDAI, Safety of Estrogens in Lupus Erythematosus National Assessment – Systemic Lupus Erythematosus Disease Activity Index.

All 43 patients received an initial glucocorticoid therapy with ⩾0.5 mg/kg/day of prednisolone and continued glucocorticoids until the end of the study period. Additionally, glucocorticoid pulse therapy was administered to 14 (32.6%) patients. The total and maximum daily glucocorticoid dosages are shown in Table 2. In addition to glucocorticoid therapy, various other treatments were concomitantly used for SLE. Hydroxychloroquine was administered to 33 (76.7%) patients. Immunosuppressive drugs were used in 29 (67.4%) patients: mycophenolate mofetil in 18, tacrolimus in 11, cyclophosphamide in 7, azathioprine in 2, and methotrexate in 2 patients. Biologic treatment was administered to three (7.0%) patients, with belimumab used in two and anifrolumab in one. Eight patients had received treatment other than glucocorticoids for SLE prior to study entry (hydroxychloroquine in six, hydroxychloroquine plus belimumab in one, and mycophenolate mofetil in one).

Table 2.

Dosage of glucocorticoids.

Dose metric	Duration after the start of initial glucocorticoid therapy
	30 days	90 days	180 days
Total dosage (mg)	2041.6 ± 1546.1	3572.4 ± 1817.3	4933.4 ± 2036.5
Maximum daily dosage (mg)	284.7 ± 380.3	284.7 ± 380.3	284.7 ± 380.3

Values are given as means ± standard deviation.

Efficacy

ONFH was identified in 8 of the 43 patients (Table 3). The incidence of ONFH was 18.6%, which was below the threshold incidence of 25% set from the historical control,^9–11 suggesting a potential signal of reduced incidence that may warrant further study (60% confidence interval, 13.6–23.6; p = 0.1664). All patients had asymptomatic ONFH at the time of diagnosis; five had bilateral ONFH, while three had unilateral ONFH.

Table 3.

List of patients with the occurrence of ONFH.

Case	Year of registration	Age	Sex	ONFH type a		ONFH stage b
				Right	Left	Right	Left
Case 1	2016	42	Female	B	C2	1	1
Case 2	2017	38	Male	C2	C2	1	1
Case 3	2018	48	Female	C1	C1	1	1
Case 4	2018	59	Male	C2	C2	1	1
Case 5	2019	44	Female	–	B	–	1
Case 6	2020	33	Male	–	A	–	1
Case 7	2020	42	Female	B	–	2	–
Case 8	2022	38	Female	A	A	1	1

^aThe classification system of the Japanese Investigation Committee.

^bThe staging system of the Japanese Investigation Committee.

ONFH, osteonecrosis of the femoral head.

The clinical characteristics of patients with and without ONFH under the prophylactic regimen are summarized in Supplemental Table 1. In addition, individual clinical profiles of all patients who developed ONFH, including disease activity, glucocorticoid pulse therapy, and concomitant immunosuppressive treatments, are presented in Supplemental Table 2.

Safety

The safety analysis set comprised 46 patients who received at least 1 dose of the 3 study drugs. Table 4 summarizes the TEAEs. During the study period, TEAEs occurred in 19 (41.3%) patients. The most common TEAE was shingles, which occurred in 2 (4.3%) patients (Supplemental Table 3). A drug eruption was the only severe TEAE that occurred in one patient. The study drugs were discontinued in this patient because it was difficult to identify the causative agent. No fatal TEAEs occurred.

Table 4.

Overview of adverse events during the study period.

Adverse event category	N = 46 n (%)
Overall treatment-emergent adverse events	19 (41.3)
Serious adverse events	2 (4.3)
Adverse events leading to discontinuation of treatment	1 (2.2)
Fatal adverse events	0 (0.0)

Values are presented as numbers (%) of patients.

However, two TEAEs were considered as serious adverse events because of hospitalization for their treatment. One patient was hospitalized due to convulsions. The patient had a history of central nervous system lupus and temporal lobe epilepsy. After admission, the patient’s condition resolved with increasing doses of antiepileptic drugs. The other patient was diagnosed with shingles and was hospitalized. As the patient was treated for SLE with glucocorticoids and mycophenolate mofetil, the immunosuppression associated with these drugs was considered to be related to shingles. Nevertheless, in both cases, a causal relationship with the study drugs was deemed to be negative by an Independent Data Monitoring Committee.

Risk factors for the occurrence of ONFH

Based on the univariate logistic regression analysis (Table 5), neither the total or maximum daily glucocorticoid dosages nor glucocorticoid pulse therapy was a significant risk factor for the occurrence of ONFH. Additionally, no significant differences were observed in SLE disease activity or laboratory data before starting initial glucocorticoid therapy. However, significant differences were found in the history of drinking, period of drinking, and average daily drinking. Multivariate logistic regression analysis of the total dosage (30, 90, and 180 days) or maximum daily dosage of glucocorticoids as covariates and variables selected by best subset selection showed that the period of drinking was the only significant variable in all four analyses (Table 6).

Table 5.

Results of univariate logistic regression analysis.

Explanatory variables	Odds ratio	95% Confidence intervals	p-Value
Total dosage of glucocorticoids (30 days) (100 mg)	1.00	0.95–1.05	0.9586
Total dosage of glucocorticoids (90 days) (100 mg)	1.01	0.97–1.05	0.7788
Total dosage of glucocorticoids (180 days) (100 mg)	1.01	0.97–1.04	0.7837
Maximum daily dosage of glucocorticoids (10 mg)	0.99	0.97–1.02	0.6283
Glucocorticoid pulse therapy	0.64	0.11–3.66	0.6152
History of drinking	5.75	1.00–32.94	0.0496
Period of drinking (year)	1.07	1.01–1.14	0.0244
Frequency of drinking (days/week)	1.23	0.95–1.59	0.1233
Average daily drinking (100 mL)	1.30	1.01–1.69	0.0455
Smoking	1.31	0.26–6.48	0.7413
SELENA-SLEDAI score a	0.96	0.86–1.06	0.4025
White blood cell count (/mL) a	1.00	1.00–1.00	0.8545
Hemoglobin (g/dL) a	0.90	0.52–1.54	0.6978
Platelet count (104/mL) a	0.99	0.90–1.08	0.7921
Aspartate aminotransferase (U/L) a	1.02	0.95–1.09	0.6076
Alanine aminotransferase (U/L) a	1.02	0.97–1.07	0.4007
Blood urea nitrogen (mg/dL) a	0.96	0.84–1.11	0.6093
Creatinine (mg/dL) a	1.27	0.07–23.67	0.8706
Total cholesterol (mg/dL) a	0.98	0.96–1.01	0.2565
Triglycerides (mg/dL) a	0.99	0.97–1.01	0.2045
C3 (mg/dL) a	1.02	0.99–1.05	0.1655
C4 (mg/dL) a	1.04	0.97–1.11	0.3157
CH50 (U/mL) a	1.02	0.98–1.07	0.2847
Anti-double stranded DNA antibody a	0.29	0.05–1.60	0.1547
Anti-cardiolipin antibody a	0.11	0.01–1.02	0.0521

^aData before the start of initial glucocorticoid therapy.

SELENA-SLEDAI, Safety of Estrogens in Lupus Erythematosus National Assessment – Systemic Lupus Erythematosus Disease Activity Index.

Table 6.

Results of multivariate logistic regression analysis.

Explanatory variables	Odds ratio	95% Confidence intervals	p-Value
[Analysis 1]
Total dosage of glucocorticoids (30 days) (100 mg)	1.01	0.95–1.07	0.8008
Period of drinking (year)	1.07	1.01–1.14	0.0240
[Analysis 2]
Total dosage of glucocorticoids (90 days) (100 mg)	1.01	0.97–1.06	0.5719
Period of drinking (year)	1.07	1.01–1.14	0.0212
[Analysis 3]
Total dosage of glucocorticoids (180 days) (100 mg)	1.01	0.97–1.06	0.5928
Period of drinking (year)	1.07	1.01–1.14	0.0214
[Analysis 4]
Maximum daily dosage of glucocorticoids (10 mg)	1.00	0.97–1.02	0.7859
Period of drinking (year)	1.07	1.01–1.14	0.0281

Discussion

In this study, we examined the preventive effect of a three-drug combination consisting of clopidogrel sulfate, pitavastatin calcium hydrate, and tocopherol acetate. We confirmed that the three-drug combination was safe and tended to reduce the incidence rate of ONFH to <25% of the historical control.

However, the results also suggested a limitation of the three-drug combination therapy. A major difference from animal studies that have successfully prevented glucocorticoid-associated osteonecrosis is the timing of the prophylactic drug administration.^21–25 Since MRI has been used to diagnose ONFH, it has become evident that ONFH develops early after the initiation of glucocorticoid therapy,^9,10 as early as within 3 weeks. ³³ Therefore, prophylactic treatment of glucocorticoid-associated ONFH should ideally be started at least 1 week before starting the initial glucocorticoid therapy, as in animal studies.^21–25 However, it is ethically inappropriate to start prophylaxis for ONFH in humans by delaying the glucocorticoid therapy necessary for SLE treatment; therefore, we had to start prophylaxis simultaneously with glucocorticoid therapy, as in previous clinical studies.^11,26

Osteonecrosis in SLE is widely recognized as a multifactorial condition influenced by a broad range of disease-related clinical and serological factors.^7,8,12–15 Consideration of these factors is essential when interpreting the risk profile of osteonecrosis in SLE. However, in the present study, all patients were uniformly treated with a three-drug combination therapy administered concomitantly with the initiation of glucocorticoid treatment. Accordingly, the findings should be interpreted within the context of this uniform pharmacological intervention, rather than as reflecting the natural risk architecture of osteonecrosis in SLE. Although disease-related factors may still modulate individual susceptibility under combination therapy, a comprehensive assessment of all such factors was beyond the scope of this proof-of-concept study.

Although high-dose glucocorticoid therapy has long been reported as a risk factor for ONFH in patients with SLE,^7,8 this study found no association between glucocorticoid dosage and ONFH occurrence. One possible reason is that the inclusion criteria of this study required an initial glucocorticoid dose of at least 0.5 mg/kg/day of prednisolone, which excluded patients with low glucocorticoid doses from the study. In addition, the administration of the three study drugs for 90 days may have modified the effects of glucocorticoid dosage on the occurrence of ONFH.

Since the absence of ONFH on MRI was confirmed before the initiation of glucocorticoid therapy, the ONFH that occurred in this study could theoretically be glucocorticoid-associated ONFH. In contrast, in this study, the period of drinking was identified as an important factor related to the occurrence of ONFH, indicating that longer alcohol consumption increased the risk of ONFH in patients with SLE concurrently treated with the three study drugs at the start of initial glucocorticoid therapy. This is consistent with previous studies, which report that habitual alcohol consumption is the most common factor associated with ONFH, along with the systemic use of glucocorticoids,^1,4,34 and it is common for patients to have both triggers. ³⁵ Although a previous study investigating the effect of the interaction between systemic glucocorticoid use and alcohol intake on the occurrence of ONFH suggested that the additive effect of alcohol may be trivial, ⁵ to the best of our knowledge, whether high-dose glucocorticoid therapy has additional effects on the occurrence of ONFH in patients with long-term alcohol consumption remains unknown. Consequently, further studies are required to elucidate the significance of the drinking period in the occurrence of glucocorticoid-associated ONFH.

In this study, the combination of the three study drugs was shown to be safe. However, patients with SLE who required urgent renal biopsy could not be enrolled in this study due to concerns regarding bleeding with clopidogrel sulfate administration. Thus, antiplatelet or anticoagulant agents may be indispensable prophylactic agents for ONFH because thrombus formation and hypercoagulable states have been suggested to be involved in the pathogenesis of glucocorticoid-associated ONFH.^16,17 Therefore, in the future, the prevention of ONFH in patients at risk of bleeding should be explored.

Limitations

Nonetheless, the present study has some limitations. First, as a proof-of-concept study, it adopted a single-arm design with historical controls. The reference ONFH incidence of 25% was derived primarily from a prospective randomized controlled trial that enrolled steroid-naïve patients with SLE who received initial glucocorticoid therapy at ⩾0.5 mg/kg/day of prednisolone. ¹¹ These criteria are consistent with those of the present study. Although complete equivalence of background factors across studies cannot be assumed, key shared design elements—namely, disease entity, steroid-naïve status, initial glucocorticoid dose, and systematic MRI-based detection of ONFH—were carefully considered when defining this reference incidence. Nevertheless, even with these considerations, the single-arm design with historical controls inherently restricts causal inference and makes it challenging to conclusively determine the efficacy of the three-drug regimen. Second, our statistical design used a one-sided significance level of 20% with a confidence interval of 60%. Although this approach was prospectively selected to detect a preliminary signal, it is unconventional compared with standard randomized trials and weakens the strength of statistical inference. Third, the number of patients analyzed (n = 43) was substantially lower than the planned enrollment (n = 150). Recruitment was challenging because of narrow eligibility criteria, requiring patients with SLE who had not been treated with glucocorticoids previously and changes in treatment practice that reduced initial glucocorticoid doses through concomitant use of immunosuppressants.^36,37 As a result, enrollment progressed more slowly than anticipated. Despite a prespecified sample size calculation with an interim analysis design, the reduced sample size may have limited the statistical power of the final analysis. Fourth, the study spanned nearly a decade, during which treatment strategies for SLE, including immunosuppressants and biologics, evolved substantially. ³⁶ This likely introduced heterogeneity in concomitant therapies that was not accounted for in the study design, adding complexity to the interpretation of results. In addition, a small subset of patients had received SLE treatment prior to study enrollment, which may have further contributed to baseline heterogeneity. However, all enrolled patients, including those who received prior treatment, required initiation of high-dose glucocorticoid therapy because of active disease at baseline, indicating that the major risk factor for osteonecrosis—exposure to high-dose glucocorticoids—was present across the entire cohort. Nevertheless, the limited number of previously treated patients precluded reliable evaluation of the independent impact of prior treatment on osteonecrosis risk. Despite these limitations, this study has a unique value. To the best of our knowledge, this is the first human trial to examine a prophylactic regimen for glucocorticoid-associated ONFH using a standardized multicenter protocol.

Conclusion

The present study demonstrated feasibility and acceptable safety of a three-drug combination therapy. Given the multifactorial nature of ONFH and limited sample size, the findings should be regarded as preliminary and hypothesis-generating. Nevertheless, we hope that these results will encourage and guide the future design of well-powered randomized controlled trials; such trials are essential for establishing effective strategies to prevent glucocorticoid-associated ONFH in patients with SLE.