FormalPara Key Summary Points

Why carry out this study?

There is no recommended treatment for polymyalgia rheumatica (PMR) other than glucocorticoids (GCs).

The effect of biological disease-modifying antirheumatic drugs (bDMARDs) for PMR under the real-world clinical setting was investigated.

What was learned from the study?

Interleukin-6 receptor inhibitor (IL-6Ri) reduced disease activity in PMR.

This study suggests that IL-6Ri may be effective for patients with treatment-resistant PMR or for patients who have difficulty using GC for PMR.

Introduction

Polymyalgia rheumatica (PMR) is an inflammatory disease of unknown etiology that causes proximal muscular pain, stiffness, and fever and commonly occurs in patients aged 50 years and older [1]. Although glucocorticoids (GCs) are the first-line treatment for PMR, patients often relapse with GC dose reduction. Approximately half of the patients relapse and flare, and as many as one-quarter of the patients require several years of GC treatment [2,3,4]. Long-term GC treatment in older individuals affects their prognosis as a result of organ damage caused by the treatment itself, such as susceptibility to infection, hyperglycemia, dyslipidemia, cardiovascular disease, GC-induced osteoporosis, Cushing’s syndrome, adrenal suppression, and psychiatric disturbances. In particular, the 2021 American College of Rheumatology (ACR) Guideline for the Treatment of Rheumatoid Arthritis describes long-term GC treatment for rheumatoid arthritis (RA) as a significant toxicity [5]. In the European League Against Rheumatism (EULAR) recommendations for the management of RA, it is repeatedly emphasized that GC should be used only for “short term” use and should be “discontinued” [6]. Therefore, it is necessary to establish a treatment that limits GC use to as short a period as possible.

Randomized controlled trials (RCTs) have been conducted to evaluate the efficacy of methotrexate (MTX) in PMR [2, 7, 8], and the EULAR/ACR guidelines for PMRs suggest the use of MTX with some conditions [1]. In older patients, PMR is prevalent, along with impaired renal and hepatic functions; however, the use of MTX may be contraindicated in this age group. In addition to MTX, biological disease-modifying antirheumatic drugs (bDMARDs) have also been investigated as a treatment for PMR. RCTs were initially conducted with tumor necrosis factor inhibitors (TNFi) but failed to demonstrate their efficacy [9, 10]. On the other hand, trials of tocilizumab (TCZ) and abatacept (ABT) for giant cell arteritis (GCA) [11, 12], which is often associated with PMR, suggested that they may be effective for GCA, suggesting that TCZ and ABT may also be effective for PMR [13, 14]. In particular, PMR is considered to be an interleukin-6 (IL-6) receptor-dependent disease, and IL-6 is considered to be an important cytokine involved in the pathogenesis of PMR after Weyand et al. showed that IL-6 is highly expressed even in PMR without GCA [15, 16]. Actually, two RCTs have suggested the efficacy of TCZ for PMR [17, 18]. The SAPHYR study (NCT03600818) was also conducted with sarilumab and showed that sarilumab was effective in achieving sustained remission in patients with relapsed PMR during glucocorticoid tapering [19]. As for ABT, the ALORS study (NCT03632187) was conducted but failed to show efficacy [20].

Although bDMARDs also cause AEs such as infections, as mentioned above long-term use of GCs is apparently toxic, so minimizing the use of GCs through the use of bDMARDs is an important strategy from both safety and efficacy perspectives. However, these trials are more rapid than usual for GC dose reduction protocols and diverge from real-world clinical practice. Although cases using TCZ for PMR have been reported, they are insufficient as real-world data as no control group was established [21,22,23,24], no comparisons were made before and after TCZ treatment [25], and the number of TCZ treatments was less than 10% of the total number of patients [26]. In other words, whether bDMARDs are effective for PMR and which of them is most effective remain unknown and are major clinical questions. In the study, we investigated the effects of bDMARDs on PMR which could be classified as RA in our department.

Methods

Study Design and Patients

This retrospective study observed a cohort of patients with PMR who were treated with bDMARDs. At present, since bDMARDs cannot be introduced to patients with PMR alone in Japan, patients with both PMR and RA were included in the study. Specifically, we selected patients with PMR diagnosed by the diagnostic criteria of Bird et al. or the 2012 EULAR/ACR provisional classification criteria for PMR among our hospitalized patients. Among patients with PMR, those who met the 1987 ACR RA classification criteria or the 2010 ACR/EULAR RA classification criteria were selected as patients with PMR and RA complications. Such patients with both PMR and RA who received bDMARDs were included in the study. Only cases that were definitively confirmed to strictly meet the diagnostic criteria were included. This confirmation occurred during a meeting that involved both the attending physician at the outpatient clinic and the attending physician when the patient was admitted to the hospital. At that meeting, the diagnosis of RA was confirmed by joint x-rays and other imaging studies with characteristic findings.

In this study, we also examined the number of items and scores for each diagnostic criterion to confirm that they were rigorously applied. The primary treatment for PMR was GCs, while conventional synthetic DMARDs (csDMARDs), mainly MTX, were used for RA. In cases where patients did not respond to GCs or MTX treatments or experienced adverse events preventing continued use of these medications, bDMARDs were administered. In other words, in this study, the target population included patients who met both the criteria for PMR and the classification criteria for RA, and who were introduced to bDMARDs because of poor disease control. The introduction of bDMARDs occurred between May 2002 and June 2022. The cohort continued to enroll patients who provided consent after May 2002, when infliximab became available in Japan. The bDMARDs used were TNFi (infliximab, etanercept, adalimumab, golimumab, and certolizumab pegol), IL-6Ri (tocilizumab and sarilumab), and cytotoxic T lymphocyte-associated antigen-4-Ig (CTLA4-Ig; abatacept). As rituximab is not approved for the treatment of RA in Japan, patients treated with rituximab were excluded from the study [27]. The other three types of bDMARDs are readily available in Japan and all have been included in the analysis. When bDMARDs were introduced, patient comorbidities, age, and disease characteristics were considered, and bDMARDs were selected for each patient on the basis of shared decision-making.

Assessment and Endpoints

The primary endpoints of the study were clinical Polymyalgia Rheumatica Activity Score (Clin-PMR-AS) and the factors associated with it at 26 weeks after bDMARD introduction. Essentially, Polymyalgia Rheumatica Activity Score (PMR-AS), the only validated composite measure, was used to assess disease activity in PMR. However, because we considered the possibility that administration of IL-6Ri could render laboratory values, such as C-reactive protein (CRP), negative regardless of disease activity, we decided to use Clin-PMR-AS instead [28]. As secondary endpoints, we evaluated continuation rates per drug class, PMR disease activity (PMR-AS, erythrocyte sedimentation rate (ESR)-PMR-AS, Clin-PMR-AS, and CRP imputed (CRP-imp) PMR-AS), and RA disease activity (Clinical Disease Activity Index (CDAI), Simplified Disease Activity Index (SDAI), and Disease Activity Score (DAS28-ESR)). The CRP-imp PMR-AS is a metric obtained by linearly transforming the Clin-PMR-AS, and it can be treated similarly to PMR-AS. We also assessed the proportion of patients with low disease activity (LDA), which has been previously defined in studies as a PMR-AS score ≤ 10 and a GC dose of ≤ 5 mg prednisolone (PSL) equivalent or a decrease of ≥ 10 mg from baseline. However, as a result of the inability to assess CRP levels in patients receiving IL-6Ri, we also considered LDA defined by CRP-imp PMR-AS score ≤ 10 and a GC dose of ≤ 5 mg PSL equivalent or a decrease of ≥ 10 mg from baseline [24]. Adverse events were defined as any new or worsening unanticipated medical event during the observation period, regardless of cause, based on the time before the start of bDMARDs. Moreover, adverse events were assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0.

Ethical Considerations

The study complied with the Declaration of Helsinki and the research protocol was approved by the Ethics Committee of the School of Medicine, University of Occupational and Environmental Health, Japan (UOEHCRB21-068). In addition, informed consent was obtained appropriately by the Ethical Guidelines for Medical and Health Research Involving Human Subjects published by the Ministry of Health, Labor, and Welfare in April 2015.

Statistical Methods

Patient characteristics were expressed as mean ± standard deviation, median (interquartile range [IQR]), or number (%) of patients. Wilcoxon rank sum and χ2 test, Fisher’s exact probability test, Kaplan–Meier curves, and generalized Wilcoxon tests were used to compare pre-and post-treatment results for each group. p values < 0.05 were considered statistically significant. In the test among the three groups, the q value was obtained using the Benjamini–Hochberg method and compared to the false discovery rate. The false discovery rate was set as 0.05 and was considered statistically significant when it was less than 0.05. When test values were not measured, they were supplemented with the multivariate imputation by chained equations (MICE) procedure. Statistical analyses were performed using JMP® 15 (SAS Institute Inc., Cary, NC, USA).

Propensity Score-Based Inverse Probability of Treatment Weighting (PS-Based IPTW)

Since each bDMARD can be used impartially, we anticipated a large selection bias due to shared decision-making. Therefore, we performed PS-based IPTW to adjust for the bias. Specifically, to adjust for baseline patient characteristics among groups, the calculated propensity score was used in IPTW as a weighting factor for stability with the “ratio of each formulation class group to all patients/propensity score” for each bDMARDs class group. Propensity scores were calculated using the following parameters: GC dose, MTX dose, previous use, PMR disease duration, MS, Pain Visual Analog Scale (VAS), DAS28-ESR, RA stage, Health Assessment Questionnaire Disability Index (HAQ-DI), body mass index (BMI), CRP, rheumatoid factor (RF), and anti-cyclic citrullinated peptide (anti-CCP) antibody. Paired-sample t test and Fisher’s exact probability test were used to compare treatment results for each group.

Results

Comparison of the Efficacy and Safety of TNFi, IL-6Ri, and CTLA4-Ig

A total of 203 patients with PMR which was resistant or intolerant to glucocorticoids and could be classified as RA were enrolled in the study. Out of these 203 patients, 142 had not previously received bDMARDs, while 61 had received some form of bDMARDs before this admission. To determine the features of patients with PMR which could be classified as RA, we compared them to patients who had introduced bDMARDs for RA. Cases in which bDMARDs were introduced for RA were extracted from our registry, the FIRST registry, a multicenter observational cohort of patients who met the 1987 ACR classification criteria for RA or the 2010 ACR/EULAR classification for RA and were enrolled in biologic/targeted synthetic DMARDs (b/tsDMARDs). Specifically, we identified 2669 patients who were administered their first b/tsDMARDs from the 4524 patients in the FIRST registry. A comparison of patients with PMR classifiable as RA and patients with RA revealed that patients with PMR classifiable as RA were older, had longer persistent morning stiffness, and had more impaired HAQ-DI than patients with RA (Table 1). In this study, we also evaluated how well each diagnostic criterion was met to ensure rigorous diagnosis. For the Bird criteria, 68% of patients met all or six items, with a mean score of 4.8 out of 6 points (5.4 out of 8 points), on the 2012 EULAR/ACR classification criteria. For the diagnosis of RA, 71% of patients diagnosed with the 1987 ACR classification criteria for RA met all or six of the criteria, and the 2010 ACR/EULAR classification for RA had a mean score of 7.6. Thus, only cases that strictly met the diagnostic criteria for both diseases were included. Approximately 14% of the data related to the primary endpoint were missing as a result of the discontinuation of bDMARDs. Since a complete case analysis may introduce bias, these cases were also included in the analysis by substituting missing data with the MICE procedure, regardless of the reason for censoring.

Table 1 Comparison of RA with PMR and RA without PMR in bDMARD-naïve cases in the study
Full size table

The 203 patients were divided into the TNFi, IL-6Ri, and CTLA4-Ig groups. Patient characteristics, such as age and duration of RA, differed significantly among the three groups (Table 2). Nearly half of the cases showed both myalgia symptoms associated with PMR and joint symptoms associated with RA concurrently. In most of the remaining cases, myalgia symptoms associated with PMR developed after the joint symptoms associated with RA had already manifested over time. While there were no significant differences among the three groups in terms of the diagnostic/classification criteria used, we observed a trend toward more cases in the IL-6Ri group meeting the 2012 EULAR/ACR provisional classification criteria for PMR. The concomitant use of MTX and GC was also significantly different between the groups. Factors involved in disease activity, such as evaluator’s global assessment and pain visual analogue scale, were also significantly lower in the CTLA4-Ig group than in the TNFi and IL-6Ri groups; DAS28 and SDAI, which indicate RA disease activity, tended to be lower in the CTLA4-Ig group. PMR disease activity was not significantly different among the three groups. The continuation rate of bDMARDs did not differ between bDMARDs classes (TNFi, IL-6Ri, and CTLA4-Ig = 82.2%, 85.7%, and 77.3%, respectively) (Fig. 1).

Table 2 Baseline characteristics of patients with PMR and RA in the study
Full size table
Fig. 1
figure 1

Retention rates up to 26 weeks after the introduction of biological disease-modifying antirheumatic drugs (bDMARDs). The tumor necrosis factor inhibitor (TNFi) is shown as a solid black line, interleukin-6 receptor inhibitor (IL-6Ri) as a solid gray line, and cytotoxic T lymphocyte-associated antigen-4-immunoglobulin (CTLA4-Ig) as a dashed line

Full size image

The trends in PMR activity after bDMARD initiation are displayed in Fig. 2a. Clin-PMR-AS after 26 weeks of bDMARD introduction was significantly lower in the IL-6Ri group than in the CTLA4-Ig group (IL-6Ri 8.6 + 9.6; CTLA4-Ig 15.7 + 12.7; p < 0.001, q < 0.001). The percentage of patients with low disease activity was significantly higher in the IL-6Ri group than in the TNFi and CTLA4-Ig groups (Fig. 2b, c). PMR-AS, ESR-PMR-AS, and CRP-imp PMR-AS demonstrated similar trends, with the IL-6Ri group exhibiting significantly higher levels than the TNFi and CTLA4-Ig groups (Supplementary Fig. S1). Supplementary Fig. S2 displays the trend in disease activity in patients with RA.

Fig. 2
figure 2

a Changes in clinical Polymyalgia Rheumatica Activity Score (Clin-PMR-AS) after the introduction of biological disease-modifying antirheumatic drugs (bDMARDs). b Percentage of patients reaching low disease activity (LDA) using C-reactive protein imputed Polymyalgia Rheumatica Activity Score (CRP-imp PMR-AS) after 26 weeks of the introduction of bDMARDs. c Percentage of patients reaching LDA using Polymyalgia Rheumatica Activity Score (PMR-AS) after 26 weeks of the introduction of bDMARDs. d Changes in glucocorticoids (GC) dose (prednisolone (PSL) equivalent) after the introduction of bDMARDs. The points denote the mean value and the bars indicate the standard deviation. TNFi tumor necrosis factor inhibitor, IL-6Ri interleukin-6 receptor inhibitor, CTLA4-Ig cytotoxic T lymphocyte-associated antigen-4-immunoglobulin

Full size image

CDAI was significantly lower in the IL-6Ri group than in both the TNFi (p = 0.003, q = 0.008) and CTLA4-Ig (p = 0.005, q = 0.008) groups 26 weeks after bDMARD introduction (Supplementary Fig. S2A), as well as in other disease activity measures.

The use of GCs 26 weeks after the introduction of bDMARDs did not differ among the three groups (Fig. 2d). The proportion of patients who were able to discontinue GC was also similar among the three groups (31.3%, 26.4%, and 36.0%, respectively). The adverse events during the observation period are displayed in Supplementary Table S1. There were no significant differences in overall or serious adverse events. On the other hand, severe infections were less common in the IL-6Ri group. One death due to cardiovascular disease occurred in the IL-6Ri group, but the patient had been on hemodialysis for 2 years before the introduction of IL-6Ri as a result of diabetic nephropathy and had also suffered an acute myocardial infarction 6 months before its introduction, suggesting that the combined use of GC was more influential than IL-6Ri.

Factors Associated with Clin-PMR-AS After 26 Weeks of bDMARD Introduction

As the patient characteristics varied widely, the factors involved in Clin-PMR-AS at 26 weeks were examined using single and multiple regression analyses (Table 3). In the single regression analysis, factors such as previous use of bDMARDs, tender joint count, morning stiffness, DAS28-ESR, PMR-AS, and BMI were identified as factors associated with Clin-PMR-AS at 26 weeks after bDMARD introduction. Multiple regression analysis, using explanatory variable factors (p < 0.05) in the single regression analysis and factors considered clinically important and excluding factors with multicollinearity, showed that the introduction of IL-6Ri was involved, along with the previous use of bDMARDs, BMI, and morning stiffness.

Table 3 Univariate and multivariate multiple regression analysis with Clin-PMR-AS after treatment as the objective variable
Full size table

Patient Characteristics After Adjustment by PS-Based IPTW

As IL-6Ri use was associated with Clin-PMR-AS 26 weeks after bDMARD introduction, we minimized the selection bias using PS-based IPTW and compared the results. Patient characteristics before PS-based IPTW adjustment were very different, as presented in Table 2, including the concomitant use of MTX and GC, BMI, and laboratory results (CRP and anti-CCP antibody). After the PS-based IPTW adjustment, as displayed in Table 2, there were no significant differences in any of the patient characteristics, and the standardized mean differences were all less than 0.1.

Comparison of Efficacy and Safety Among Three Groups After Adjustment by PS-Based IPTW

The persistence rate after bDMARD initiation, adjusted by PS-based IPTW, was not significantly different (TNFi, IL-6Ri, and CTLA4-Ig = 81.0%, 82.7%, and 70.0%, respectively) (Supplementary Fig. S3). In terms of PMR activity trends, the IL-6Ri group had significantly lower Clin-PMR-AS scores than the TNFi (p = 0.033, q = 0.042) and CTLA4-Ig (p = 0.001, q = 0.004) groups at 26 weeks after bDMARD introduction (Fig. 3a). The percentage of patients with low disease activity in the CRP-imp PMR-AS group was significantly different (33.1%, 60.0%, and 29.9%, respectively; p < 0.001), and in multiple testing of each group, the IL-6Ri group was significantly higher than that in the TNFi (p = 0.004, q = 0.006) and CTLA4-Ig (p = 0.003, q = 0. 010) (Fig. 3b). The residual analysis also demonstrated that the IL-6Ri group had a significantly higher proportion of patients with low disease activity (p = 0.002, q = 0.006) as compared to the other groups. The percentage of patients with low disease activity in the PMR-AS group was significantly higher in the IL-6Ri group compared to the TNFi and CTLA4-Ig groups. PMR-AS, ESR-PMR-AS, and CRP-imp PMR-AS at 26 weeks after bDMARD initiation were significantly lower in the IL-6Ri group than in the other two groups (Supplementary Fig. S4). Regarding RA disease activity, the CDAI, SDAI, and DAS28-ESR were significantly lower in the IL-6Ri group than in the other two groups after 26 weeks of bDMARD introduction (Supplementary Fig. S5). The overall trend of GC use was not significantly different between the two groups (Supplementary Fig. S6), but there was a significant difference in the percentage of patients who were able to reduce GC (58.5%, 77.7%, and 47.6%, respectively; p = 0.003); and multiple testing revealed that the IL-6Ri group was significantly lower than the TNFi (p = 0. 016, q = 0.025) and CTLA4-Ig group (p = 0.003, q = 0.009) (Fig. 3d). Supplementary Table S2 displays the adverse events during the observation period after adjusting for PS-based IPTW. There were no significant differences in overall or serious adverse events. Significantly fewer serious infections were observed in the IL-6Ri group than in the control group.

Fig. 3
figure 3

a Change in clinical Polymyalgia Rheumatica Activity Score (Clin-PMR-AS) after biological disease-modifying antirheumatic drugs (bDMARDs) introduction adjusted by propensity score-based inverse probability of treatment weighting (PS-based IPTW). b Percentage of patients reaching low disease activity (LDA) using C-reactive protein imputed Polymyalgia Rheumatica Activity Score (CRP-imp PMR-AS) after 26 weeks of the introduction of bDMARDs adjusted for PS-based IPTW. c Percentage of patients reaching LDA using PMR-AS after 26 weeks of the introduction of bDMARDs adjusted for PS-based IPTW. d Percentage of patients who were able to reduce the dose of glucocorticoids (GC) after 26 weeks of bDMARD introduction. The points denote the mean value and the bars indicate the standard deviation. TNFi tumor necrosis factor inhibitor, IL-6Ri interleukin-6 receptor inhibitor, CTLA4-Ig cytotoxic T lymphocyte-associated antigen-4-immunoglobulin

Full size image

Discussion

This is the first study to demonstrate in a real-world clinical setting that IL-6Ri was more effective than TNFi or CTLA4-Ig in PMR, classifiable as RA. In real-world clinical settings, patient characteristics usually differ widely, as efficacy and safety are considered on the basis of age, comorbidities, and concomitant medications, and bDMARDs are determined on the basis of joint decision-making with patients. The choice of any bDMARDs was not influenced by the temporal differences in the occurrence of myalgia symptoms associated with PMR and joint symptoms associated with RA, or by the diagnostic or classification criteria used. However, treatments such as MTX and GC, as well as CRP levels, did impact treatment selection. The selection of any bDMARDs was not influenced by the temporal differences in the occurrence of myalgia symptoms associated with PMR and joint symptoms associated with RA, or by the diagnostic/classification criteria used. However, treatments such as MTX and GCs, as well as CRP levels, did impact treatment selection. In the study, we used multiple regression analysis and the PS-based IPTW method to adjust for confounding factors as much as possible to make the IL-6Ri group comparable to the other two groups and demonstrated that the disease activity of PMR, including Clin-PMR-AS, was improved in the IL-6Ri group. In previous TCZ studies, the PMR-AS, which directly reflected CRP levels, was used for evaluation [17, 22, 24]. However, in this study, given that CRP could be used as an evaluation index due to IL-6Ri, the disease activity of PMR was evaluated using Clin-PMR-AS and CRP-imp PMR-AS, which was a fairer assessment of PMR disease activity than previous studies. The IL-6Ri group exhibited reduced disease activity in PMR compared to the other two groups, suggesting the efficacy of IL-6Ri in PMR. In addition, various cases with different disease durations were included in the study, ranging from early onset to relapse after treatment. Even after accounting for variations in the timing of myalgia symptoms related to PMR and joint symptoms associated with RA, the IL-6Ri still significantly reduced disease activity in PMR. These results suggest that the IL-6Ri was effective regardless of when joint and myalgia symptoms appeared, or which diagnostic criteria were used. Previous RCTs have focused on a uniform group of patients in the early stages of disease onset or at the time of relapse; however, we believe that the results of this case study reflect the real-world clinical setting, as a variety of patients were included in the study. In the CTLA4-Ig group, the efficacy of ABT for GCA has been previously reported [12], and even though an RCT for PMR is currently underway, its efficacy was not clear in this study. We believe that this may have been influenced by the small number of CTLA4-Ig groups and the short observation period.

Additionally, in the study, one patient in the IL-6Ri group died. The patient had been on hemodialysis for diabetic nephropathy for 2 years before the introduction of TCZ and had been hospitalized for acute myocardial infarction 6 months earlier. Data on bDMARDs for RA suggested that TNFi may exacerbate congestive heart failure [29], but there are no reports of IL-6Ri exacerbating congestive heart failure. We believe that the patient died because he was also being treated with GC and other drugs, which increased his risk of congestive heart failure. The IL-6Ri group demonstrated no increase in specific adverse events, and the results adjusted for PS-based IPTW revealed that, as in previous studies, the use of IL-6Ri increased the proportion of patients who were able to reduce GC use compared with other bDMARDs. This suggested that the use of IL-6Ri may reduce the number of GC-related adverse events. Therefore, the benefits of IL-6Ri use from the perspective of adverse events were significant.

The study has several limitations. First, this was a retrospective, observational study. Another limitation was that the study focused on patients with PMR classifiable as RA who required treatment with bDMARDs. Therefore, it was possible that patients with PMR in the study may have a higher rate of RF and anti-citrullinated protein antibodies (ACPA) positivity than patients with PMR in general, that PMR activity may be linked by RA disease activity, and that they may have a higher rate of organ complications, such as lung involvement [30]. However, previous findings have demonstrated that RF and ACPA were associated with treatment resistance and lung involvement was associated with adverse pulmonary infections; therefore, this study targeted patients with a prognosis worse than all PMR. Furthermore, it is considered that this study includes a heterogeneous population, which encompasses non-typical PMR and treatment-resistant PMR as indicated by the guidelines of the British Society for Rheumatology (BSR) and British Health Professionals in Rheumatology (BHPR) [31]. Moreover, this study subjects patients diagnosed with both RA and PMR as described in the exclusion criteria of the guidelines. However, as patients age, distinguishing between PMR and RA becomes more difficult, particularly in real-world clinical settings. The focus of this study is believed to reflect such real-world clinical settings. Third, the observation period was 26 weeks, and long-term observations regarding efficacy and safety were needed. Finally, in the study, we used multiple linear regression analysis and PS-based IPTW to adjust for confounders as much as possible to compare the efficacy and safety of bDMARDs; however, not all confounders were adjusted. It is possible that we did not adjust for unknown confounders, and a larger prospective study is required. It is necessary to investigate the efficacy of IL-6Ri in patients with PMR who do not have concurrent RA and explore the possibility of monotherapy with bDMARDs without using GCs or MTX.

Conclusions

This study compared the efficacy and safety of IL-6Ri and other bDMRADs in PMR which could be classified as RA by using multiple linear regression analysis and PS-based IPTW to adjust for confounding factors in real-world clinical settings. The results suggested that IL-6Ri was more effective than other bDMRADs in PMR which could be classified as RA and may further reduce GC.