Discontinuation of concomitant methotrexate in Japanese patients with rheumatoid arthritis treated with tocilizumab: An interventional study

Abstract

Objectives: To evaluate the efficacy and safety of methotrexate (MTX) discontinuation in Japanese rheumatoid arthritis (RA) patients with sustained low disease activity undergoing combination therapy with tocilizumab (TCZ) plus MTX.

Methods: This multicenter, open-label, uncontrolled, prospective study included RA patients maintaining low disease activity (Clinical Disease Activity Index (CDAI) ≤10) for ≥12 weeks with TCZ plus MTX. Methotrexate was discontinued following 12 weeks of biweekly administration while continuing TCZ therapy. The primary endpoint was the proportion of patients maintaining low disease activity with no flare at week 36.

Results: A total of 49 patients completed 36 weeks of therapy. The proportion of patients maintaining low disease activity at week 36 was 75.5%. The lower limit of the 95% confidence interval exceeded the assumed threshold response rate of 60%, demonstrating the clinical feasibility of MTX discontinuation. The prevalence of gastroesophageal reflux disease, defined as a Frequency Scale for Symptoms of Gastroesophageal reflux disease score ≥8, significantly decreased from week 0 to 12 (27.1–18.4%; p= .025).

Conclusion: Discontinuation of concomitant MTX is clinically feasible for maintaining low disease activity, and may be beneficial from the perspective of reducing gastrointestinal symptoms in Japanese RA patients treated with TCZ.

Trial registration number: UMIN000021247.

Keywords:

• Rheumatoid arthritis
• tocilizumab
• methotrexate
• gastroesophageal reflux
• clinical trial

Introduction

Treatment targets in rheumatoid arthritis (RA) are remission, especially in patients naïve to disease-modifying antirheumatic drugs (DMARDs), and low disease activity for those who have failed previous therapies [1,2]. Treatment strategies for RA have greatly advanced over the last few decades, and aggressive therapies that combine the use of biologics and methotrexate (MTX) have enabled more patients to achieve treatment goals.

MTX is an important anchor drug for RA patients and is used alone (monotherapy) or in combination with biologics. However, some patients discontinue MTX due to toxicity [3,4], including gastrointestinal (GI) disorders [4–6] and liver damage [4,5]. In addition, the use of MTX may lead to the development of adverse events (AEs) such as pneumonia and cytopenia. According to real-world cohort studies, approximately one-third of RA patients who underwent biologic treatment received biologics as monotherapy, often due to MTX intolerance [7–9]. Thus, de-escalation of MTX while maintaining a favorable disease activity state – a challenge in RA clinical research – may be beneficial from the perspective of reducing AEs.

Tocilizumab (TCZ), a humanized anti-human interleukin-6 receptor monoclonal antibody, has proven effective in RA patients who have failed to respond to MTX. Most biologics (primarily tumor necrosis factor inhibitors) have been shown to improve outcomes to a greater extent when used in combination with MTX, compared to monotherapy [10,11]. The efficacy of TCZ has been demonstrated in monotherapy as well as with concomitant MTX [12,13], opening up the possibility of MTX discontinuation in these patients if disease control can be maintained.

The dose of MTX approved by the Ministry of Health, Labour, and Welfare of Japan had an upper limit of 8 mg/week until January 2011. This was subsequently increased to 16 mg/week in February 2011 for patients with RA, which is still lower than doses recommended for the treatment of RA in Western countries [14]. Previous studies in Japan have reported that biological therapy for RA with concomitant 8 mg/week MTX has additive effects and more AEs than biological therapy alone [15–18]. Given this unique situation in Japan, the possibility of discontinuing MTX in Japanese RA patients warrants consideration.

This study aimed to evaluate the efficacy and safety of MTX discontinuation in Japanese RA patients with sustained low disease activity undergoing combination therapy with TCZ plus MTX.

Participants and methods

Study design and participants

The present study, the TCZ treatment with Reducing and stopping methotreXate in patients with RA in stable low disease activity-state study (hereafter, the T-ReX study) is a multicenter, open-label, uncontrolled, prospective 64-week study. Participants were RA patients aged ≥20 years with sustained low disease activity (Clinical Disease Activity Index (CDAI) ≤10) for ≥12 weeks while undergoing combination therapy with TCZ plus MTX. Patients fulfilled the 1987 American College of Rheumatology (ACR) classification criteria or the new ACR/European League Against Rheumatism (EULAR) diagnostic criteria for RA, and had a serum C-reactive protein (CRP) level of ≥0.6 mg/dL at the time of initiating TCZ therapy. Patients had to be receiving MTX orally at a stable dose of ≥6 mg/week, and TCZ at a stable dosage regimen irrespective of the route of administration, for ≥12 weeks prior to obtaining informed consent. Patients receiving conventional synthetic DMARDs (csDMARDs) other than MTX or glucocorticoids were eligible, but doses had to be stable for ≥12 weeks prior to obtaining informed consent.

This report analyses data from the first 36 weeks of the 64-week study. The protocol was reviewed and approved by the Ethics Committee of the Nagoya University Graduate School of Medicine (2016-0086) and ethics committees of participating centers, and was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN000021247). This study was conducted in accordance with the International Conference on Harmonization Good Clinical Practice guidelines and Japanese ethical guidelines, and complied with the Declaration of Helsinki. Written informed consent was obtained from all patients.

Study treatment

The study treatment scheme is shown in Figure 1(a). At week 0, the dosing frequency of MTX was decreased from weekly to biweekly without a change in dose, regardless of the initial dose. At week 12, MTX was discontinued if low disease activity was maintained. TCZ and csDMARDs other than MTX were continued at a stable dose and interval throughout the course of the study. Glucocorticoids were continued at a stable dose up to week 36, and allowed to taper after week 36. The use of oral analgesics (non-steroidal anti-inflammatory drugs, acetaminophen, pregabalin, and tramadol) was not prohibited during the study period. One or more of the following rescue treatments were performed if the CDAI score was >10 and at the discretion of the investigator and/or upon patient request: changing the dosing frequency back to weekly administration, restarting, or increasing doses of MTX; increasing doses of or adding csDMARDs other than MTX or glucocorticoids; and administering an intraarticular injection of corticosteroids, hyaluronic acid, or lidocaine.

Figure 1. Study treatment and patient disposition. (a) Study treatment scheme. At week 0, the dosing frequency of methotrexate was decreased from weekly to biweekly without a change in dose regardless of the initial dose. At week 12, methotrexate was discontinued if low disease activity was maintained. (b) Patient disposition. csDMARDs: conventional synthetic disease-modifying antirheumatic drugs.

Data collection and assessments

Data collected at baseline (week 0) included patient demographics and disease characteristics. All eligible patients were assessed at weeks 0, 12, 24, 36, and 64, regardless of disease status. Assessments included the following parameters: tender joint count, swollen joint count, patient global assessment using a visual analogue scale (VAS), evaluator global assessment using a VAS, serum CRP level, serum matrix metalloproteinase-3 (MMP-3) level, Health Assessment Questionnaire Disability Index (HAQ-DI) [19], EuroQol-5 dimension (EQ-5D), and Frequency Scale for Symptoms of Gastroesophageal reflux disease (FSSG). The FSSG, a Japanese scale for gastroesophageal reflux disease (GERD) symptoms, is a self-administered questionnaire comprising 12 items rated on a five-point scale ranging from 0 (never) to 4 (always) [20]. The FSSG has been shown to correlate strongly with upper GI endoscopic findings. A cut-off score of 8 is used to diagnose GERD. At each visit, patients were monitored for laboratory evaluations. Patients could contact the study site whenever necessary.

Outcome measures

The primary endpoint was the proportion of patients maintaining low disease activity without a flare at week 36 (24 weeks after MTX discontinuation). Disease flare was defined as a CDAI score >10 or intervention with the above-mentioned rescue treatments for any reason. Secondary endpoints included the following parameters from week 0 to 64: disease activity (CDAI, Simple Disease Activity Index (SDAI), and Disease Activity Score with 28 joint counts with CRP (DAS28-CRP)), serum MMP-3 level, physical function (HAQ-DI), quality of life (EQ-5D), and GI symptoms (FSSG; prevalence of GERD was defined as a FSSG score ≥8). Safety was evaluated throughout the study period by observing the occurrence of AEs and monitoring laboratory evaluations.

Statistical analyses

The primary endpoint was assessed by estimating the proportion and two-sided 95% confidence interval (CI) of patients who maintained low disease activity without a flare at week 36 using binomial distribution. Assuming that 80% of patients would maintain low disease activity at week 36, 43 patients were calculated as necessary to prove the clinical feasibility of discontinuing MTX at a power of ≥80% with a threshold response rate of 60%. A two-sided statistical test of no difference at 5% significance level was used. On the assumption that 15% of patients would not complete the study, we planned to enroll 51 patients. The cumulative flare-free rate was estimated using the Kaplan–Meier curves. Results for CDAI, SDAI, DAS28-CRP, serum MMP-3 level, HAQ-DI, EQ-5D, and FSSG were evaluated using a linear mixed model. Estimates for differences between mean values at week 0 and weeks 12, 24, and 36 were compared using the Dunnett–Hsu correction for multiple comparisons. Prevalence of GERD was compared by McNemar’s test between week 0 and weeks 12, 24, and 36.

Efficacy analyses were performed in the full analysis set, excluding those who were found ineligible. Safety analyses were performed in all treated patients. All statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC) and Stata software version 13.1 (StataCorp, College Station, TX). p < .05 was considered statistically significant.

Results

Patient characteristics

Of the 53 enrolled patients, four did not meet the inclusion criteria. Consequently, data from 49 patients who completed the therapy up to week 36 were subjected to efficacy analysis, whereas data from 53 patients were subjected to safety analysis (Figure 1(b)). Table 1 shows the baseline (week 0) characteristics of patients included in the efficacy analyses. Data for weight, rheumatoid factor positivity, and FSSG score were available for 47, 48, and 48 patients, respectively, and other data items were available for all patients. Of the 49 patients, six received TCZ at a longer dosing interval than that described in the drug label (i.e. intravenously (IV) every 4 weeks, or subcutaneously (SC) every 2 weeks). Thirty-two patients received TCZ IV; doses were higher than that described in the drug label (8 mg/kg) for three patients and lower for three patients.

Table 1. Patient characteristics at baseline (week 0).

	Total, n = 49
Age, years	62 ± 10
Female, %	84
Weight, kg^a	55 ± 11
Disease duration, years	11 ± 8
Steinbrocker stage, I/II/III/IV, %	14/35/25/27
Steinbrocker class, I/II/III/IV, %	33/63/4/0
Rheumatoid factor positive, %^b	79
Anticyclic citrullinated peptide positive, %	90
Route of tocilizumab, intravenous/subcutaneous, %	65/35
Methotrexate dose, mg/week	8.2 ± 2.3
Use of glucocorticoids, %	29
Glucocorticoid dose, mg/day^c	2.4 ± 1.4
Use of csDMARDs other than methotrexate, %	16
Previous biologic use, %	63
CDAI	2.7 ± 2.5
CDAI ≤2.8, %	67
SDAI	2.8 ± 2.5
DAS28-CRP	1.5 ± 0.4
CRP, mg/dL	0.04 ± 0.06
Matrix metalloproteinase-3, ng/mL	64.0 ± 28.8
HAQ-DI^d	0.472 ± 0.613
EuroQol-5 dimension	0.822 ± 0.170
FSSG^b	5.0 ± 5.4
FSSG ≥8, %^b^,^e	27

csDMARDs: conventional synthetic disease-modifying antirheumatic drugs; CDAI: Clinical Disease Activity Index; SDAI: Simple Disease Activity Index; DAS28: Disease Activity Score with 28 joint counts; CRP: C-reactive protein; HAQ-DI: Health Assessment Questionnaire Disability Index; FSSG: Frequency Scale for Symptoms of Gastroesophageal reflux disease.

Data are shown as mean ± SD or percentage.

^a Data were obtained from 47 patients.

^b Data were obtained from 48 patients.

^c Prednisolone equivalent mg/day. Mean among patients receiving the drug.

^d HAQ-DI scores ranged from 0 to 3, with higher scores indicating greater disability.

^e A cut-off score of 8 was used to diagnose gastroesophageal reflux disease.

Efficacy

A total of 12 patients had a disease flare by week 36: of these, three patients with CDAI scores >10, and six patients who maintained CDAI ≤10, received rescue treatment, whereas three patients with CDAI scores >10 requested not to (and did not) undergo rescue treatment. The proportions (95% CI) of patients who maintained low disease activity without a flare at weeks 12, 24, and 36 were 87.8% (75.2–95.4%), 81.6% (68.0–91.2%), and 75.5% (61.1–86.7%), respectively (Figure 2(a)). The lower limit of the 95% CI at week 36 exceeded the assumed threshold response rate of 60%. According to the Kaplan–Meier estimates, the cumulative flare-free rate was 87.8% at week 12 (MTX reduction phase) and 74.7% at week 36 (Figure 2(b)).

Figure 2. Maintenance of low disease activity based on CDAI score. (a) Proportion of patients maintaining low disease activity without a flare. (b) The Kaplan–Meier estimates of cumulative flare-free rate. CDAI: Clinical Disease Activity Index.

To assess the potential effects of administration route of TCZ and glucocorticoid use, patients were divided into the following groups: TCZ IV and SC groups, and glucocorticoid (+) and (–) groups. Proportions of patients who maintained low disease activity at weeks 36 were 72% and 82% in TCZ IV and SC groups, respectively, and 64% and 80% in glucocorticoid (+) and (–) groups, respectively. Univariate logistic regression analysis revealed that administration route and glucocorticoid use had no significant impact on maintaining low disease activity at weeks 36.

Mean CDAI score at week 36 was modestly but significantly higher than that at week 0 (3.9 vs. 2.7, p = .026) (Figure 3(a)), consistent with the results for SDAI score (4.2 vs. 2.8, p = .013) and DAS28-CRP (1.7 vs. 1.5, p = .001) (Figure 3(b,c)). All three patients with CDAI scores >10 who received rescue treatment (one patient changed dosing frequency back to weekly, and two patients restarted MTX) regained low disease activity by week 36 (Figure 4). No significant changes were observed in serum MMP-3 level, HAQ-DI, and EQ-5D from week 0 to 36 (Figure 3(d–f)).

Figure 3. Secondary efficacy endpoints. Changes in (a) Clinical Disease Activity Index (CDAI), (b) Simple Disease Activity Index (SDAI), (c) Disease Activity Score with 28 joint counts with C-reactive protein (DAS28-CRP), (d) matrix metalloproteinase-3 (MMP-3), (e) Health Assessment Questionnaire Disability Index (HAQ-DI), and (f) EuroQol-5 dimension (EQ-5D) from week 0 to 36. Values represent mean (standard deviation). *p<.05 vs. week 0 by the Dunnett–Hsu test.

Figure 4. Changes in CDAI score in patients who received rescue treatment due to CDAI score >10. CDAI: Clinical Disease Activity Index. *Changing the dosing frequency back to weekly. ^†Restarting MTX.

FSSG scores (mean ± standard deviation (SD)) at weeks 0, 12, 24, and 36 were 5.0 ± 5.4, 4.0 ± 4.5, 3.6 ± 4.5, and 4.6 ± 5.6, respectively (Figure 5(a)). The FSSG score at week 24 was significantly lower than that at week 0 (p = .005). The prevalence of GERD was 27.1% at week 0, 18.4% at week 12, 18.8% at week 24, and 20.0% at week 36 (Figure 5(b)). The prevalence of GERD at week 12 was significantly lower than that at week 0 (p = .025). These results were consistent even after replacing missing data by using the last observation carried forward method.

Figure 5. Gastrointestinal symptoms. (a) Changes in Frequency Scale for Symptoms of Gastroesophageal reflux disease (FSSG) score from week 0 to 36. Values represent mean (standard deviation). (b) Prevalence of gastroesophageal reflux disease defined as an FSSG score ≥8. *p<.05 vs. week 0 by the Dunnett–Hsu test. ^†p<.05 vs. week 0 by McNemar’s test.

Safety

AEs reported during the period from week 0 to 36 are summarized in Table 2. Laboratory test abnormalities were defined as grade ≥1 abnormalities according to the Common Terminology Criteria for Adverse Events version 5.0. Overall, 31 (58%) patients reported 54 AEs, and five (9%) patients reported five serious AEs. Safety analyses revealed no unexpected TCZ safety issues.

Table 2. Adverse events.

	n = 53
	AE	SAE
Total patients with ≥1 AE or ≥1 SAE	31 (58%)	5 (9%)
Blood and lymphatic system disorders	5 (9%)	0 (0%)
Eye disorders	1 (2%)	1 (2%)
Gastrointestinal disorders	2 (4%)	1 (2%)
Infections and infestations
Bronchial infection	1 (2%)	1 (2%)
Skin infection	1 (2%)	1 (2%)
Upper respiratory infection	1 (2%)	0 (0%)
Laboratory test abnormalities
Increased aspartate aminotransferase	10 (19%)	0 (0%)
Increased alanine aminotransferase	6 (11%)	0 (0%)
Increased creatinine	6 (11%)	0 (0%)
Decreased platelet count	6 (11%)	0 (0%)
Decreased white blood cell count	5 (9%)	0 (0%)
Decreased lymphocyte count	9 (17%)	0 (0%)
Neoplasms (benign, malignant, and unspecified)	1 (2%)	1 (2%)

AE: adverse events; SAE: serious adverse events.

Data are shown as number and percentage.

Discussion

This study examined the possibility of discontinuing MTX in RA patients who maintained low disease activity with TCZ plus MTX combination therapy. Discontinuation of MTX was associated with modest but statistically significant increases in disease activity; however, more than 75% of patients maintained low disease activity at 24 weeks after MTX discontinuation. The lower limit of the 95% CI at that time point exceeded the assumed threshold response rate of 60%, demonstrating that discontinuing concomitant MTX is clinically feasible in RA patients treated with TCZ. The clinical feasibility of MTX discontinuation was also supported by secondary efficacy analyses, which showed no significant changes in serum MMP-3 level, HAQ-DI, and EQ-5D throughout the course of study.

Our results are consistent with findings from previous studies. The COMP-ACT study [21] and ACT-TAPER study [22] have shown that discontinuing or tapering MTX was noninferior to continuing MTX in maintaining clinical response in patients who had achieved low disease activity with TCZ plus MTX. In those studies, clinical responses were primarily assessed based on DAS28 scores. On the other hand, our assessment approach relied on the use of CDAI scores as well as opinions of the investigators and patients themselves, and thus is likely more closely related to clinical practice. Indeed, six of the 12 patients who had a flare by week 36 received rescue treatment at the investigator’s discretion or upon patient request, even though they had maintained a CDAI score of ≤10. Moreover, since DAS28 contains an acute-phase reactant component, it may not provide sufficiently reliable results when patients are taking TCZ, as TCZ directly reduces acute-phase reactants [1,23–25]. Our study is the first to demonstrate the clinical feasibility of discontinuing MTX in combined therapy with TCZ, based on CDAI scores without the influence of acute-phase reactants.

The proportion of patients maintaining low disease activity in the present study (75.5% at 24 weeks after MTX discontinuation) was higher compared to that in the COMP-ACT study (63.3% at 16 weeks after MTX discontinuation). Some differences were observed in patient background, study protocols, and the definition of disease flare between the two studies, which might explain why the results of the present study were superior. The present study included patients who ‘sustained’ low disease activity for ≥12 weeks with TCZ plus MTX in clinical practice, whereas the COMP-ACT study included patients who ‘achieved’ low disease activity with TCZ plus MTX at a single time point (and subsequently discontinued MTX). Some patients have flares after achieving treatment targets, even while continuing on DMARDs in clinical practice [26]. Thus, the duration of sustained stable disease control is a decisive factor for starting DMARD tapering. Our findings support recent guidelines that state that tapering the csDMARD could be considered for a patient in persistent remission [1,2].

Notably, the dose of MTX used in the present study (mean ± SD, 8.2 ± 2.3 mg/week) was lower than those used in the COMP-ACT study (mean, 17–18 mg/week) [21], even after considering that Japanese patients on average weigh 20–30% less than Western patients. The MTX dose approved by the Japanese Ministry of Health, Labor, and Welfare has an upper limit of 16 mg/week. However, a prospective study on certolizumab pegol in Japan, in which the protocol recommended MTX to be initiated at a dose of 8 mg/week and increased to 16 mg/week at week 8, showed that the MTX dose used in actual clinical settings remains ≤12 mg/week, owing to AEs [27]. Concentrations of erythrocyte MTX polyglutamate, a potential biomarker for MTX therapy, were reported to be markedly higher in the Japanese study than what was reported by a US study, suggesting that a lower dose of MTX may be sufficient for Japanese patients [28].

Interestingly, tapering MTX resulted in a decreased prevalence of GERD. This was consistent with our previous study, which demonstrated that high-dose MTX was associated with upper GI symptoms in Japanese RA patients [6]. GI symptoms reportedly reduce quality of life (QOL) in RA patients [29]; therefore, improvement of GI symptoms is important for improving QOL during RA treatment. The SURPRISE study [18] showed a higher incidence of AEs in patients randomized to receive TCZ plus MTX in spite of the low dose MTX (mean ± SD, 8.6 ± 2.5 mg/week), as compared to TCZ monotherapy. Accordingly, discontinuation of concomitant MTX may be beneficial from the perspective of reducing AEs in TCZ-treated patients.

The SURPRISE study compared the efficacy between TCZ added to MTX and TCZ switched from MTX in patients with inadequate response to MTX [18]. Clinical remission rate in the switch strategy was lower than that in the add-on strategy up to week 24, and caught up with that in the add-on strategy by week 52. The delay in achieving remission led to more progressed structural damage in the switch strategy. We previously reported that, in TCZ treatment, concomitant MTX use was associated with an increased remission rate in patients with high disease activity, but not in patients with low to moderate disease activity [30]. Thus, concomitant MTX use is preferable at least until patients achieve treatment targets during treatment with TCZ; although, TCZ monotherapy has been demonstrated to be effective [12,13]. Patients with RA require long-term treatment to maintain disease control, because RA is a chronic inflammatory disease with unknown etiology. MTX is an anchor drug for RA patients, but due to its toxicity, 30–50% of patients reportedly discontinue MTX therapy after 5 years [3,4]. Our results suggest that discontinuation of MTX following response to TCZ plus MTX may be highly desirable from efficacy and safety perspectives in patients with MTX intolerance.

The present study has some limitations worth noting. First, this was an open-label study without a control group. Both patients and evaluators were aware of MTX reduction and/or discontinuation, which might have affected the assessments. The definition of disease flare included rescue treatment interventions upon patient request, so that investigator bias could be minimized. Second, the observation period was limited despite the fact that patients required long-term treatment. The COMP-ACT study reported that the proportions of patients maintaining low disease activity were comparable between week 16 and week 28 following MTX discontinuation [21]. Accordingly, we considered the duration of 36 weeks (i.e. 24 weeks after MTX discontinuation) to be sufficient for assessing the effect of MTX discontinuation as the primary outcome. Third, given the prospective nature of this study, there may have been potential bias in patient selection. However, to address this possibility, we applied stringent inclusion criteria such as a serum CRP level of ≥0.6 mg/dL at the time of initiating TCZ therapy. Finally, as the sample size was small for secondary analyses, the significance of some of the findings may change with a larger dataset.

In conclusion, the results of the present study demonstrate that discontinuation of MTX is clinically feasible for maintaining low disease activity in Japanese RA patients who achieve sustained low disease activity with TCZ plus MTX. Moreover, discontinuation of MTX may be beneficial from the perspective of reducing AEs including GI symptoms.

Conflict of interest

SA has received speakers’ fees from AbbVie, Astellas, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eisai, Janssen, Takeda, and UCB Japan. MH has received speakers’ fees from AbbVie, Astellas, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eli Lilly, Mitsubishi Tanabe, and Pfizer. YY has received speakers’ fees from Bristol-Myers Squibb, Chugai, and Pfizer. YY has received speakers’ fees from Chugai, Daiichi-Sankyo, Mitsubishi Tanabe, and Ono. YH has received speakers’ fees from AbbVie, Asahi Kasei, Astellas, Ayumi, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eisai, Eli Lilly, Janssen, Mitsubishi Tanabe, Nippon Kayaku, Pfizer, Takeda, and UCB Japan. TF has received speakers’ fees from AbbVie, Astellas, Bristol-Myers Squibb, Chugai, and Mitsubishi Tanabe. TK has received speakers’ fees from Chugai, Daiichi-Sankyo, Eisai, Janssen, Mitsubishi Tanabe, Ono, Pfizer, and Takeda. NT has received speakers’ fees from AbbVie, Asahi Kasei, Astellas, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eisai, Eli Lilly, Janssen, Mitsubishi Tanabe, Ono, Pfizer, Takeda, and UCB Japan. YS has received speakers’ fees from Astellas, Bristol-Myers Squibb, and Ono. MS has received speakers’ fees from Bristol-Myers Squibb. NI has received grant/research support, consulting fees, and/or speakers’ fees from AbbVie, Asahi Kasei, Astellas, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eisai, Eli Lilly, Kaken, Mitsubishi Tanabe, Ono, Otsuka, Pfizer, Taisho Toyama, Takeda, and Zimmer Biomet. TK has received grant/research support and/or speakers’ fees from AbbVie, Astellas, Bristol-Myers Squibb, Chugai, Daiichi-Sankyo, Eli Lilly, Janssen, Mitsubishi Tanabe, Novartis, Pfizer, and Takeda. The other authors declare no conflicts of interest.

Acknowledgements

We thank Drs. Katsuyoshi Kato, Hiroyuki Miyake, Atsushi Kaneko, and Takeshi Oguchi, for their suggestions, and Mr. Tomokazu Morii for data management.

Additional information

Funding

This work was supported by Chugai Pharmaceutical Co., Ltd.