Abstract
Rheumatoid arthritis (RA) patients receiving biological and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs) often experience treatment changes due to inefficacy or adverse events. The purpose of this study was to clarify the incidence and reasons for change of b/tsDMARDs in a cohort of Japanese patients with RA and to identify the predictors of treatment change. This was a retrospective observational study of RA patients prescribed b/tsDMARD between April 2011 and December 2020 at the Kindai University Nara Hospital. We focused on the change of first-line b/tsDMARDs and identified the reasons for change using the electronic medical records. Logistic regression analysis was performed to identify predictors of treatment change as the objective variable and baseline characteristics as the explanatory variable. The reasons for treatment change were inefficacy in 69.6% of cases and adverse events in 29.7% of cases. Concomitant administration of higher dose prednisolone at baseline (adjusted odds ratio: [95% confidence interval]: 2.52 [1.19–5.33]) and old age (2.00 [1.03–3.87]) were associated with change in b/tsDMARD treatment due to inefficacy within 2 years of initiation. A better understanding of b/tsDMARDs persistence and elucidating the predictors of treatment change can help improve treatment outcomes for RA.
INTRODUCTION
Rheumatoid arthritis (RA) is an autoimmune disease characterized by inflammatory arthritis. Patients with RA are often prescribed disease-modifying antirheumatic drug (DMARD) therapy to control disease activity, achieve early clinical remission, and improve long-term outcomes and prognosis.1,2) The recent introduction of biological and targeted synthetic DMARDs (b/tsDMARDs) represents a major advance in the treatment of RA. The proportion of RA patients in Japan being treated with b/tsDMARDs has been increasing over time.3,4) As a result, more than 70% of RA patients in Japan achieved the treatment target of clinical remission or low disease activity.4) For patients who achieve remission, withdrawal of b/tsDMARD treatment is considered.5) The lack of achievement of the treatment target often requires changing the b/tsDMARD. Additionally, the development of an adverse event may necessitate treatment interruption or a change in the b/tsDMARD. Thus, there are various reasons for discontinuation of b/tsDMARD treatment in patients with RA.
Persistence is used as a measure to evaluate the benefit-risk balance of a treatment in studies using real-world data.6,7) Medication persistence is defined as the duration of time from initiation to discontinuation of therapy.8) Several studies have evaluated the persistence and drug retention rate for b/tsDMARDs.9,10) A more accurate understanding of the persistence of b/tsDMARDs and identification of predictors of discontinuation can potentially lead to improved outcomes in RA.11,12) The reasons for discontinuation of b/tsDMARDs can be classified into inefficacy, remission, adverse events, and other reasons.9) Among these, inefficacy and adverse events are undesirable events for treatment. Change of the b/tsDMARD is an important outcome to consider for improving persistence because it is more likely to be due to inefficacy and adverse events. Identifying factors associated with the change of b/tsDMARDs and preventing its occurrence may help improve treatment outcomes in patients with RA. Some studies have investigated the incidence and reasons for the discontinuation of b/tsDMARDs in Japanese patients.9,13,14) However, these studies have not reported on change of b/tsDMARDs. Some studies conducted in other countries have reported on reasons and predictors of the change of b/tsDMARD among RA patients receiving b/tsDMARD therapy.15,16) However, the types of b/tsDMARDs approved for RA and the healthcare environment in Japan differ from those in other countries.2) The purpose of this study was to clarify the incidence and reasons for the change of b/tsDMARDs in Japanese RA patients and to identify the predictors of treatment change.
MATERIALS AND METHODS
Study Design and Data SourceThis was a retrospective observational study using the electronic medical records (EMR) at the Kindai University Nara Hospital through December 2022. This is a 518-bedded academic hospital located in Ikoma, Nara, Japan. The following information was extracted from the EMR: disease records using International Classification of Diseases tenth revision (ICD-10) codes, hospital visit dates, admission dates, discharge dates, patient characteristics, laboratory data, and prescription data. After data collection, personal data were anonymized to protect patient privacy. RA patients were identified by ICD-10 codes M05 and M06, and these patients were confirmed as RA patients in the EMR.
EthicsThis study was conducted in accordance with the Ethical Guidelines for Medical and Biological Research Involving Human Subjects established by the Ministry of Health, Labor, and Welfare in Japan. This study was approved by the Ethics Committee of Kindai University Nara Hospital on January 23, 2023 (Approval Number: 689), and the Ethics Committee of Kindai University School of Pharmacy on April 8, 2023 (Approval Number: 22-224). Patients included in this study were guaranteed the opportunity to be excluded from the study using an opt-out consent approach.
Study PopulationThe study population consisted of RA patients aged ≥18 years who were prescribed b/tsDMARD therapy between April 2011 and December 2020. The details of the medications are presented in Supplementary Table S1. The date of the first prescription of b/tsDMARDs at our institution was defined as the baseline. The exclusion criteria were as follows: patients with a history of b/tsDMARD treatment before initiation at our institution; patients who had received b/tsDMARD treatment both at our institution and at other institutions after baseline; and patients for whom C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) levels were not measured before baseline.
Outcome of InterestThe treatment plan for discontinuing b/tsDMARDs has three options: treatment change, end of treatment, and interruption of treatment. The primary outcome was the change of the b/tsDMARD treatment. Treatment change is one of the treatment plans when b/tsDMARDs are discontinued. We focused on discontinuing the first-line b/tsDMARDs to distinguish “treatment change” from “end or interruption of treatment” (Fig. 1). The date of the first discontinuation was defined as the event date. The event date and reasons for discontinuation were identified using EMRs. Patients who were repeatedly prescribed first-line b/tsDMARDs were considered to be continuing treatment of first-line b/tsDMARDs, regardless of the prescription interval, unless there was a record indicating discontinuation. The discontinuation was classified as interruption if b/tsDMARD treatment was reinitiated after the event date, regardless of the duration of interruption period. The following events were considered as a continuation: (1) switching between originator and biosimilar of the same component; (2) adverse events that did not require hospitalization or referral to another department regardless of whether or not the administration was postponed; and (3) drug withdrawal with surgery. Reasons for discontinuation of first-line b/tsDMARDs were classified into four categories based on previous studies: (1) inefficacy, (2) remission, (3) adverse events, and (4) other reasons.9) For patients who discontinued on their self-judgment (noncompliance) or patients who had no reason recorded on the date of their last visit (lost to follow-up), the date of last prescription of b/tsDMARDs was considered as the date of discontinuation, because the exact date of discontinuation could not be determined. If a patient had more than one reason for discontinuation, each reason was counted separately. Patients who continued first-line b/tsDMARD treatment until December 2022 were censored at the end of December 2022.
Baseline CharacteristicsBaseline data on age, sex, initiation date of b/tsDMARDs, age at RA onset, disease duration, concomitant medications, CRP, and ESR level were collected. Conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), nonsteroidal anti-inflammatory drugs (NSAIDs), and prednisolone (PSL) were selected as concomitant medications. The only glucocorticoid used in this study population was PSL. PSL users were classified into lower dose (<5 mg/d) and higher dose (≥5 mg/d) groups as defined by the European League Against Rheumatism task force.17) The date of initiation of first-line b/tsDMARDs was classified into three terms to ensure an even distribution of the number of patients. Old age was defined as age ≥65 years at baseline. Elderly-onset RA was defined as onset at the age of ≥65 years.18) If the date of RA onset was not indicated in the EMR, the date of the first prescription of DMARDs was defined as the date of RA onset. Disease duration was calculated as the duration between the date of RA onset and baseline, and divided into two categories based on the median value. For baseline CRP and ESR, values closest to the baseline were adopted. Since the ESR data analyzed did not distinguish values of 100 or more, any value of ≥100 was treated as 100. Baseline CRP and ESR were classified into two groups using cutoff values based on a previous study.14)
Statistical AnalysisThe cumulative incidence of the discontinuation of b/tsDMARDs was assessed using the cumulative incidence curve. However, patients with multiple reasons for discontinuation had their data replicated based on the number of reasons for discontinuation. To explore the predictors of treatment change, we identified patients who experienced treatment change within 2 years of baseline and those who continued first-line b/tsDMARDs for 2 years, using data for 2 years from baseline. Univariate or multivariate logistic regression analyses were used to assess the association between baseline covariates and change of first-line b/tsDMARDs in 2 years and the results were reported as odds ratios (ORs) and 95% confidence intervals (CIs).
Data management was performed using Visual Mining Studio (version 8.7; NTT DATA Mathematical Systems Inc., Tokyo, Japan) and SAS Studio (version 3.8, Enterprise Edition; SAS Institute, Cary, NC, U.S.A.). All statistical tests were two-sided, and p < 0.05 was considered indicative of statistical significance using EZR (version 1.61; Saitama Medical Center, Jichi Medical University, Japan).
RESULTS
Study ParticipantsA total of 385 RA patients aged ≥18 years who were prescribed b/tsDMARDs between April 2011 and December 2020 were included. After excluding patients with a history of b/tsDMARD treatment at other institutions before initiation at our institution (n = 54), those who had received b/tsDMARD treatment both at our institution and at other institutions after baseline (n = 1), and patients without CRP and ESR levels measured before baseline (n = 5), a total of 325 new b/tsDMARD users were included in this study. Table 1 summarizes the baseline characteristics of the study population. Of the 325 new b/tsDMARD users, 76.0% were female, the median age at baseline was 68 years, 41.5% had elderly-onset RA, and the median disease duration was 3 years. Tumor necrosis factor inhibitors (TNFi) were the most commonly prescribed first-line b/tsDMARD (n = 234, 72.0%), followed by abatacept (n = 56, 17.2%), interleukin-6 inhibitors (IL-6i) (n = 20, 6.2%), and Janus kinase inhibitors (JAKi) (n = 15, 4.6%). csDMARDs were the most commonly prescribed concomitant medications (n = 246, 75.7%) followed by PSL (n = 151, 46.5%) and NSAIDs (n = 147, 45.2%). Methotrexate was the most frequently used csDMARD in combination with b/tsDMARDs (n = 148, 45.5%). The median CRP and ESR levels were 0.98 mg/dL and 42 mm/h, respectively.
Table 1. Baseline Characteristics
| New b/tsDMARD users (n = 325) |
|---|
| Female, n (%) | 247 (76.0) |
| Age, years, median [IQR] | 68 [54–76] |
| Elderly-onset RA, n (%) | 135 (41.5) |
| Disease duration, years, median [IQR] | 3 [1–10] |
| Initiation date of first-line b/tsDMARDs |
| 1st term: 2011–2015, n (%) | 111 (34.2) |
| 2nd term: 2016–2018, n (%) | 95 (29.2) |
| 3rd term: 2019–2020, n (%) | 119 (36.6) |
| Type of first-line b/tsDMARDs |
| TNFi, n (%) | 234 (72.0) |
| Abatacept, n (%) | 56 (17.2) |
| IL-6i, n (%) | 20 (6.2) |
| JAKi, n (%) | 15 (4.6) |
| Concomitant medication |
| csDMARDs, n (%) | 246 (75.7) |
| NSAIDs, n (%) | 147 (45.2) |
| Prednisolone | |
| Non-user, n (%) | 174 (53.5) |
| <5 mg/d, n (%) | 66 (20.3) |
| ≧5 mg/d, n (%) | 85 (26.2) |
| Laboratory data |
| CRP, mg/dL, median [IQR] | 0.98 [0.42–2.68] |
| ESR, mm/h, median [IQR] | 42 [27–62] |
Elderly-onset RA was defined as onset at 65 years or older. Disease duration was calculated as the duration between the date of RA onset and baseline.Abbreviations: b/tsDMARD, biological and targeted synthetic disease-modifying antirheumatic drug; CRP, C-reactive protein; csDMARD, conventional synthetic disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; IL-6i, interleukin-6 inhibitor; IQR, interquartile range; JAKi, janus kinase inhibitor; NSAID, nonsteroidal anti-inflammatory drug; RA, rheumatoid arthritis; TNFi, tumor necrosis factor inhibitor.
Of the 325 new b/tsDMARD users, 272 patients discontinued first-line b/tsDMARDs during the study period, and 53 patients were censored at the end of December 2022. Of the 272 patients who discontinued, 12 patients had more than one reason for discontinuation and their data was replicated based on the number of reasons for discontinuation. After data replication, there were 337 new b/tsDMARDs users and 284 patients who discontinued (Fig. 2). At the event date, 138 patients were managed through treatment change, and 146 patients were managed through end or interruption of treatment (Fig. 3). One patient who discontinued due to noncompliance after the first administration had an observation period of 0 d. Figures 4a and 4b show the cumulative incidence of treatment change and the end or interruption of treatment disaggregated by reason, respectively. The number of patients with treatment change (cumulative incidence rate of treatment change) due to inefficacy was 71 (21.1%) at 2 years, 88 (24.9%) at 4 years, and 96 (32.6%) at 12 years; the corresponding data for treatment change due to adverse events were 35 (8.9%) at 2 years, 39 (10.4%) at 4 years, and 41 (13.2%) at 12 years (Fig. 4a). The number of patients with end or interruption of treatment (cumulative incidence rate of end or interruption of treatment) due to adverse events was 59 (17.5%) at 2 years, 71 (22.1%) at 4 years, and 75 (25.1%) at 12 years; the other predominant reasons for the end or interruption of treatment were “other reasons” (26 [7.7%] at 2 years, 34 [10.5%] at 4 years, and 38 [14.1%] at 12 years] and remission [17 [5.0%] at 2 years, 27 [8.5%] at 4 years, and 30 [10.5%] at 12 years) (Fig. 4b). Table 2 shows the distribution of reasons for treatment change and end or interruption of treatment. The predominant reasons for treatment change were inefficacy (69.6%) and adverse events (29.7%). The most common reasons for end or interruption of treatment were adverse events (51.4%), other reasons (26.0%), and remission (20.5%). Inefficacy was the least common reason for end or interruption of treatment (2.1%). The median period from baseline to the event date in the 99 patients who discontinued the first-line b/tsDMARDs due to inefficacy was 266 d (interquartile range 123–773). Supplementary Table S2 shows the detailed distribution of adverse events leading to discontinuation of first-line b/tsDMARDs. The main adverse events leading to treatment change were rash (68.4%) and hepatic toxicity (12.2%). The main adverse events leading to end or interruption of treatment were infections (30.7%), interstitial pneumonia (14.7%), and malignancy and lymphoma (12.0%). Since treatment changes mostly occurred within 2 years and were mostly due to inefficacy or rash, treatment changes due to rash and inefficacy within 2 years of initiating b/tsDMARDs were selected as the outcome of interest for further analysis.
Table 2. Reasons for Discontinuation of First-Line b/tsDMARDs
| Reasons, n (%) | Total
(n = 284) | Treatment plan |
|---|
Treatment change
(n = 138) | End or interruption of treatment
(n = 146) |
|---|
| Inefficacy | 99 (34.9) | 96 (69.6) | 3 (2.1) |
| Remission | 30 (10.6) | 0 (0.0) | 30 (20.5) |
| Adverse events | 116 (40.8) | 41 (29.7) | 75 (51.4) |
| Other reasons | 39 (13.7) | 1 (0.7) | 38 (26.0) |
| Transferred | 12 (4.2) | 0 (0.0) | 12 (8.2) |
| Lost to follow-up | 8 (2.8) | 0 (0.0) | 8 (5.5) |
| Noncompliance | 7 (2.5) | 0 (0.0) | 7 (4.8) |
| Patient’s request | 5 (1.8) | 1 (0.7) | 4 (2.7) |
| Economic reasons | 4 (1.4) | 0 (0.0) | 4 (2.7) |
| Pregnancy | 2 (0.7) | 0 (0.0) | 2 (1.4) |
| Safety reason | 1 (0.4) | 0 (0.0) | 1 (0.7) |
Data for patients with multiple reasons for discontinuation were replicated based on the number of reasons for discontinuation. Abbreviation: b/tsDMARD, biological and targeted synthetic disease-modifying antirheumatic drug.
Table 3 shows the comparison of baseline characteristics between patients who experienced treatment change due to rash within 2 years of baseline (n = 26) and those who continued treatment for 2 years (n = 126). Compared with patients who continued treatment for 2 years, all patients who experienced treatment change due to rash within 2 years were using TNFi.
Table 3. Predictors for Treatment Change of First-Line b/tsDMARDs Due to Rash
| Treatment change due to rash in 2 years
(n = 26) | Continuation for 2 years
(n = 126) | Crude odds ratio
(95% CI) |
|---|
| Female, n (%) | 22 (84.6) | 104 (82.5) | 1.16 (0.37–3.71) |
| Age≧65 years, n (%) | 13 (50.0) | 61 (48.4) | 1.07 (0.46–2.48) |
| Early-onset RA, n (%) | 13 (50.0) | 49 (38.9) | 1.57 (0.67–3.67) |
| Disease duration <3 years, n (%) | 13 (50.0) | 65 (51.6) | 0.94 (0.40–2.18) |
| Initiation date of first-line b/tsDMARDs |
| 1st term: 2011–2015, n (%) | 7 (26.9) | 42 (33.3) | Reference |
| 2nd term: 2016–2018, n (%) | 11 (42.3) | 31 (24.6) | 2.13 (0.74–6.12) |
| 3rd term: 2019–2020, n (%) | 8 (30.8) | 53 (42.1) | 0.91 (0.30–2.70) |
| Type of first-line b/tsDMARDs |
| TNFi, n (%) | 26 (100.0) | 85 (67.5) | – |
| Abatacept, n (%) | 0 (0.0) | 18 (14.3) | |
| IL-6i, n (%) | 0 (0.0) | 14 (11.1) | |
| JAKi, n (%) | 0 (0.0) | 9 (7.1) | |
| csDMARDs, n (%) | 20 (76.9) | 96 (76.2) | 1.04 (0.38–2.83) |
| NSAIDs, n (%) | 12 (46.2) | 60 (47.6) | 0.94 (0.40–2.20) |
| Prednisolone |
| Non-user, n (%) | 13 (50.0) | 72 (57.1) | Reference |
| <5 mg/d, n (%) | 5 (19.2) | 28 (22.2) | 0.99 (0.32–3.03) |
| ≧5 mg/d, n (%) | 8 (30.8) | 26 (20.6) | 1.70 (0.63–4.58) |
| CRP |
| <1.50 mg/dL, n (%) | 18 (69.2) | 78 (61.9) | Reference |
| ≧1.50 mg/dL, n (%) | 8 (30.8) | 48 (38.1) | 0.72 (0.29–1.79) |
| ESR |
| <50 mm/h, n (%) | 19 (73.1) | 75 (59.5) | Reference |
| ≧50 mm/h, n (%) | 7 (26.9) | 51 (40.5) | 0.54 (0.21–1.38) |
Abbreviations: b/tsDMARD, biological and targeted synthetic disease-modifying antirheumatic drug; CI, confidence interval; CRP, C-reactive protein; csDMARD, conventional synthetic disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; IL-6i, interleukin-6 inhibitor; JAKi, janus kinase inhibitor; NSAID, nonsteroidal anti-inflammatory drug.
Table 4 shows the comparison of baseline characteristics between patients who experienced treatment change due to inefficacy within 2 years of baseline (n = 71) and those who continued for 2 years (n = 126). Higher dose PSL (adjusted OR, 2.52; 95% CI, 1.19–5.33) and old age (adjusted OR, 2.00; 95% CI, 1.03–3.87) were associated with a significantly higher incidence of treatment change due to inefficacy within 2 years. Compared with TNFi, IL-6i was associated with a significantly lower incidence of treatment change due to inefficacy within 2 years (adjusted OR, 0.09; 95% CI, 0.01–0.76).
Table 4. Predictors for Treatment Change of First-Line b/tsDMARDs Due to Inefficacy
| Treatment change due to inefficacy in 2 years
(n = 71) | Continuation for 2 years
(n = 126) | Crude odds ratio
(95% CI) | Adjusted odds ratio
(95% CI) |
|---|
| Female, n (%) | 51 (71.8) | 104 (82.5) | 0.54 (0.27–1.08) | 0.71 (0.33–1.51) |
| Age≧65 years, n (%) | 49 (69.0) | 61 (48.4) | 2.37 (1.29–4.38) | 2.00 (1.03–3.87) |
| Early-onset RA, n (%) | 30 (42.3) | 49 (38.9) | 1.15 (0.64–2.08) | |
| Disease duration <3 years, n (%) | 37 (52.1) | 65 (51.6) | 1.02 (0.57–1.83) | |
| Initiation date of first-line b/tsDMARDs |
| 1st term: 2011–2015, n (%) | 23 (32.4) | 42 (33.3) | Reference | |
| 2nd term: 2016–2018, n (%) | 20 (28.2) | 31 (24.6) | 1.18 (0.55–2.51) | |
| 3rd term: 2019–2020, n (%) | 28 (39.4) | 53 (42.1) | 0.97 (0.49–1.91) | |
| Type of first-line b/tsDMARDs |
| TNFi, n (%) | 53 (74.6) | 85 (67.5) | Reference | Reference |
| Abatacept, n (%) | 16 (22.5) | 18 (14.3) | 1.43 (0.67–3.04) | 1.08 (0.48–2.42) |
| IL-6i, n (%) | 1 (1.4) | 14 (11.1) | 0.12 (0.01–0.90) | 0.09 (0.01–0.76) |
| JAKi, n (%) | 1 (1.4) | 9 (7.1) | 0.18 (0.02–1.45) | 0.15 (0.02–1.28) |
| csDMARDs, n (%) | 50 (70.4) | 96 (76.2) | 0.74 (0.39–1.43) | |
| NSAIDs, n (%) | 37 (52.1) | 60 (47.6) | 1.20 (0.67–2.14) | |
| Prednisolone |
| Non-user, n (%) | 28 (39.4) | 72 (57.1) | Reference | Reference |
| <5 mg/d, n (%) | 15 (21.1) | 28 (22.2) | 1.38 (0.64–2.96) | 1.09 (0.50–2.42) |
| ≧5 mg/d, n (%) | 28 (39.4) | 26 (20.6) | 2.77 (1.39–5.52) | 2.52 (1.19–5.33) |
| CRP |
| <1.50 mg/dL, n (%) | 41 (57.7) | 78 (61.9) | Reference | |
| ≧1.50 mg/dL, n (%) | 30 (42.3) | 48 (38.1) | 1.19 (0.66–2.15) | |
| ESR |
| <50 mm/h, n (%) | 38 (53.5) | 75 (59.5) | Reference | |
| ≧50 mm/h, n (%) | 33 (46.5) | 51 (40.5) | 1.28 (0.71–2.30) | |
Abbreviations: b/tsDMARD, biological and targeted synthetic disease-modifying antirheumatic drug; CI, confidence interval; CRP, C-reactive protein; csDMARD, conventional synthetic disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; IL-6i, interleukin-6 inhibitor; JAKi, janus kinase inhibitor; NSAID, nonsteroidal anti-inflammatory drug.
DISCUSSION
This single-center study investigated the incidence and reasons for treatment change in RA patients who initiated first-line b/tsDMARDs. While the incidence rates of treatment change and end or interruption of treatment were similar, the reasons for treatment change and end or interruption of treatment were different. The most common reason for treatment change was inefficacy, while the most common reason for end or interruption of treatment was adverse events. Therefore, to investigate the predictors for discontinuation of b/tsDMARDs, it is important to distinguish between treatment change and end or interruption of treatment.
In a previous study using Japanese national database, the median age of new b/tsDMARD users among RA patients was 60–64 years and females accounted for 78.6% of these patients.19) In the present study, females accounted for 76% of new b/tsDMARD users and the median age of patients was 68 years, which was higher than that in the previous study. In the present study, TNFi was the most common type of first-line b/tsDMARD, and abatacept was more commonly prescribed than IL-6i. In previous studies investigating the trends in b/tsDMARDs prescription in Japan, TNFi were the most common type of b/tsDMARDs, and tocilizumab, one of the IL-6i, was more frequently prescribed than abatacept.3,19) Old age has been reported to be a determinant of first-line prescription of abatacept.20) Thus, the difference in the type of b/tsDMARDs between our study and previous studies is likely attributable to the difference in age. In our study, 75.7% of patients were concomitantly receiving csDMARDs and 46.5% were concomitantly receiving PSL at baseline. These findings were similar to those of a multicenter study of new b/tsDMARD users in Japan.13) Therefore, although the patients in our single-center study were relatively older, they were likely treated in line with the general trend in Japanese clinical settings.
Infections, interstitial pneumonia, and malignancy and lymphoma were the frequent reasons for the end or interruption of treatment in our study population. These results are similar to a previous study in Japan.13) Furthermore, our results indicate that adverse events have consistently been the most common reason for the end or interruption of treatment since the early phase of b/tsDMARDs treatment. This suggests the need to be cautious of the occurrence of adverse events such as infections throughout the duration of treatment with b/tsDMARDs. Guidelines of the Japan College of Rheumatology recommend vaccination against influenza vaccine and pneumococcal vaccine in RA patients using DMARDs.2) This practice may help improve persistence by reducing the incidence of end or interruption of treatment due to infections. However, live herpes zoster vaccine should be avoided in bDMARDs users. Preventing end or interruption of treatment due to interstitial pneumonia, malignancy, and lymphoma can improve persistence, but the corresponding preventive interventions are inherently challenging. On the contrary, inefficacy is the most common reason for treatment change, which is consistent with previous studies conducted in other countries.15,21) Furthermore, our study indicates that inefficacy has consistently been the most common reason for treatment change since the early phase of b/tsDMARDs treatment. Therefore, although both treatment change and end or interruption of treatment in b/tsDMARDs treatment represent discontinuation, the underlying reasons are different. Interestingly, rash was the most common adverse event leading to treatment change. Rash may occur due to hypersensitivity reaction (HSR), infusion-related reaction (IRR), or injection site reaction (ISR), and tends to occur shortly after administration.22) In this study, the incidence of treatment change due to adverse events was found to be markedly higher in the early phase of b/tsDMARDs initiation, suggesting the need for due caution in this phase. Therefore, preventing treatment changes that occur in the early phase of treatment may help improve the persistence of first-line b/tsDMARDs.
Exploring baseline characteristics that may help predict treatment change in the early phase of therapy is of much clinical significance. We sought to identify the factors that predicted treatment changes within the first 2 years of treatment initiation. All 26 patients who experienced treatment change due to rash within 2 years of baseline were TNFi users. ISRs are the most commonly reported adverse effects of etanercept.23) In a previous study, etanercept was shown to be associated with the highest risk of developing ISRs among the bDMARDs.24) In the present study, out of the 26 TNFi users who experienced treatment change due to rash within 2 years, 23 were using etanercept, 2 were using certolizumab pegol, and 1 patient was using adalimumab (data not shown). These findings are consistent with those of previous studies suggesting that etanercept use is associated with a higher risk of developing ISRs and that TNFi are more likely to induce ISRs than other b/tsDMARDs. ISRs are generally transient and well tolerated.22) However, IRRs and HSRs may also manifest as rash. Rash caused by HSR is likely to worsen with each successive infusion.25) Therefore, the prescribing physician at our institution may have decided on treatment change to avoid the severity of the rash. ISRs can be prevented by measures such as topical corticosteroids, cool packs, and rotation of injection sites. The change to biosimilar may also reduce the occurrence of ISRs.26) The risk of IRRs is reduced by slowing the infusion rate or by premedication with antihistamines and acetaminophen.27) These measures can potentially increase the persistence rate of first-line TNFis. However, it is unclear why rash is more common in patients receiving TNFi than other b/tsDMARDs. Further studies are required to utilize TNFi as a predictor of treatment change due to rash.
Old age, higher dose PSL, and IL-6i were identified as predictors of treatment change due to inefficacy. The significantly lower odds ratio for IL-6i suggests that IL-6i is associated with a lower risk of treatment change due to inefficacy compared to TNFi. All IL-6i users who continued treatment for 2 years were receiving tocilizumab (data not shown). This finding is consistent with a previous study reporting a higher rate of persistence with tocilizumab compared to other bDMARDs.28) However, further research is required to investigate whether the class of IL-6i is a predictor of higher rate of persistence. In contrast, a significantly higher odds ratio for old age suggests that elderly patients are at a higher risk of treatment change due to inefficacy. This finding is consistent with a previous study that showed a lower efficacy of bDMARDs in elderly patients compared to young patients.29) However, the type of b/tsDMARDs and age are variables that are not amenable to intervention after the initiation of treatment. In contrast, PSL use is a predictor that is amenable to intervention. The use of PSL at baseline has been identified as a risk factor for poor outcomes of b/tsDMARDs.30,31) Herein, the OR for higher dose PSL was substantially higher than non-PSL users, which is consistent with the results of previous studies. However, PSL is a treatment for RA, and it is more likely to be used in patients with more severe disease. High disease activity and elevated ESR at baseline have also been reported to be associated with poor response to bDMARDs.32) Therefore, PSL may not directly influence the efficacy of b/tsDMARDs. In addition to this possibility, Ochi et al. speculated that glucocorticoids may also influence the responsiveness to TNFi.30) Clarifying the influence of glucocorticoids on b/tsDMARDs treatment may help improve the outcomes of b/tsDMARD therapy.
This is the first study to clarify the reasons for and the incidence of change of b/tsDMARDs during the treatment of Japanese patients with RA. In addition, we identified predictors that may help reduce the incidence of treatment change due to inefficacy and rash. However, some limitations of our study should be acknowledged. First, this was a single-center retrospective observational study with a small sample size, and caution should be exercised while interpreting the results. Second, the median age of the patients (68 years) and the prescribed b/tsDMARDs type in the study were characteristic, which potentially limit the generalizability of the results. Third, data on disease activity scores were not available. In our clinical practice, the treatment of RA was based on symptoms, signs (joint swelling), and laboratory indices. Disease activity scores at baseline were not recorded in the EMR for most of the patients. The data on comorbidities was also incomplete. These incomplete data were not included as baseline characteristics as it could have introduced an element of bias. Fourth, discontinuation due to adverse events that did not require hospitalization or referral to other department were not included in the definition of discontinuation in this study. This may have potentially underestimated the incidence of interruption. However, some patients were determined by their physicians as suitable for continuation of first-line b/tsDMARDs at the time of discontinuation due to such adverse events, and the date of next administration was determined. Inclusion of such patients as interruptions could have overestimated interruptions. Therefore, we prioritized avoiding overestimation of interruption by limiting them to adverse events requiring hospitalization or referral to other departments. Finally, there are no standardized criteria to determine discontinuation of b/tsDMARDs, so discontinuation was based on the patient’s opinion and the physician’s judgment. A study based on a large clinical database can help overcome the limitation of sample size and physician preference bias. However, the reason for the discontinuation of b/tsDMARDs is not recorded in the claims database commonly used in Japan. Despite some limitations, we clarify the incidence and reasons for change of b/tsDMARDs using the EMR. This study contributes to a better understanding of the persistence of b/tsDMARDs, which may lead to improved treatment outcomes in RA patients.
In conclusion, treatment change with first-line b/tsDMARDs were discontinuation due to undesirable reasons such as inefficacy and rash. TNFi use and PSL use were identified as predictors of treatment change due to rash and inefficacy, respectively. Further research considering the characteristics of patients is required to utilize TNFi and PSL as predictors for treatment change. A more accurate understanding of b/tsDMARDs persistence and elucidating the predictors for treatment change can help improve treatment outcomes for RA.
Acknowledgments
This study was supported by JSPS KAKENHI (Grant No. JP23K14397).
Author Contributions
All authors contributed to the study’s conception and design. Material preparation and data collection and analysis were performed by C.N. Analysis and interpretation of the data were performed by C.N., R.O., A.H., S.Y., T.U., and K.H. The first draft of the manuscript was written by C.N., and all authors commented on previous versions of the manuscript. All authors have read and approved the final version of the manuscript.
Conflict of Interest
K.H. received a research funding from GEXVal Inc. The remaining authors declare no conflict of interest.
Data Availability
The data used in the present study are available from the corresponding author on reasonable request.
Supplementary Materials
This article contains supplementary materials.
REFERENCES
- 1) Di Matteo A, Bathon JM, Emery P. Rheumatoid arthritis. Lancet, 402, 2019–2033 (2023).
- 2) Kawahito Y, Morinobu A, Kaneko Y, Kohno M, Hirata S, Kishimoto M, Seto Y, Sugihara T, Tanaka E, Ito H, Kojima T, Matsushita I, Nishida K, Mori M, Murashima A, Yamanaka H, Nakayama T, Kojima M, Harigai M. Drug treatment algorithm and recommendations from the 2020 update of the Japan College of Rheumatology clinical practice guidelines for the management of rheumatoid arthritis-secondary publication. Mod. Rheumatol., 33, 21–35 (2023).
- 3) Hirata A, Ota R, Hata T, Hamada T, Nishihara M, Uchiyama K, Katsumata T. Prescribing trends of biologic disease-modifying anti-rheumatic drugs using a claims database from 6 million people in Japan. Clin. Drug Investig., 41, 967–974 (2021).
- 4) Yamanaka H, Tanaka E, Nakajima A, Furuya T, Ikari K, Taniguchi A, Inoue E, Harigai M. A large observational cohort study of rheumatoid arthritis, IORRA: Providing context for today’s treatment options. Mod. Rheumatol., 30, 1–6 (2020).
- 5) Schlager L, Loiskandl M, Aletaha D, Radner H. Predictors of successful discontinuation of biologic and targeted synthetic DMARDs in patients with rheumatoid arthritis in remission or low disease activity: a systematic literature review. Rheumatology (Oxford), 59, 324–334 (2020).
- 6) Luttropp K, Dozier M, Justo N, Cornillie F, Kachroo S, Govoni M, Salomonsson S, Black CM, Khalifa A. Real-world treatment persistence of golimumab in the management of immune-mediated rheumatic diseases in Europe: a systematic literature review. BMJ Open, 9, e027456 (2019).
- 7) Koncz T, Pentek M, Brodszky V, Ersek K, Orlewska E, Gulacsi L. Adherence to biologic DMARD therapies in rheumatoid arthritis. Expert Opin. Biol. Ther., 10, 1367–1378 (2010).
- 8) Cramer JA, Roy A, Burrell A, Fairchild CJ, Fuldeore MJ, Ollendorf DA, Wong PK. Medication compliance and persistence: terminology and definitions. Value Health, 11, 44–47 (2008).
- 9) Ebina K, Hirano T, Maeda Y, Yamamoto W, Hashimoto M, Murata K, Takeuchi T, Shiba H, Son Y, Amuro H, Onishi A, Akashi K, Hara R, Katayama M, Yamamoto K, Kumanogoh A, Hirao M. Drug retention of 7 biologics and tofacitinib in biologics-naive and biologics-switched patients with rheumatoid arthritis: the ANSWER cohort study. Arthritis Res. Ther., 22, 142 (2020).
- 10) Acar M, Juneja P, Handel M. Treatment persistence of subcutaneous TNF inhibitors among Australian patients with immune-mediated rheumatic disease (IMRD). Open Access Rheumatol., 10, 151–160 (2018).
- 11) Letarouilly JG, Salmon JH, Flipo RM. Factors affecting persistence with biologic treatments in patients with rheumatoid arthritis: a systematic literature review. Expert Opin. Drug Saf., 20, 1087–1094 (2021).
- 12) Blum MA, Koo D, Doshi JA. Measurement and rates of persistence with and adherence to biologics for rheumatoid arthritis: a systematic review. Clin. Ther., 33, 901–913 (2011).
- 13) Terabe K, Takahashi N, Asai S, Hirano Y, Kanayama Y, Yabe Y, Oguchi T, Fujibayashi T, Ishikawa H, Hanabayashi M, Hattori Y, Suzuki M, Kishimoto K, Ohashi Y, Imaizumi T, Imagama S, Kojima T. Reasons and risk factors for discontinuation of treatment with any biological disease-modifying antirheumatic drugs in patients with rheumatoid arthritis: a long-term observational study. Mod. Rheumatol., 33, 891–898 (2023).
- 14) Aramaki T, Ichinose K, Umetsu A, Terada K, Hara K, Eguchi K, Takatani A, Ueki Y, Iwamoto N, Kawakami A. Drug retention rate of TNF inhibitors in bio-naïve patients with rheumatoid arthritis and prognostic factors of the various reasons for discontinuation—a single-center observational study. Clin Rheumatol Rel Res, 33, 253–263 (2021).
- 15) Ouardi NE, Maghraoui AE, Djossou HJ, et al. Discontinuation of biologic therapy in patients with rheumatoid arthritis and ankylosing spondylitis: analysis from multicenter cohort study. Rheumatol. Int., 43, 79–87 (2023).
- 16) Jin Y, Landon JE, Krueger W, Liede A, Desai RJ, Kim SC. Predictors of treatment change among patients with rheumatoid arthritis treated with TNF inhibitors as first-line biologic agent in the USA: a cohort study from longitudinal electronic health records. BioDrugs, 36, 521–535 (2022).
- 17) Strehl C, Bijlsma JW, de Wit M, et al. Defining conditions where long-term glucocorticoid treatment has an acceptably low level of harm to facilitate implementation of existing recommendations: viewpoints from an EULAR task force. Ann. Rheum. Dis., 75, 952–957 (2016).
- 18) Sugihara T. Treatment strategies for elderly-onset rheumatoid arthritis in the new era. Mod. Rheumatol., 32, 493–499 (2022).
- 19) Takabayashi K, Ando F, Ikeda K, Fujita S, Nakajima H, Hanaoka H, Suzuki T. Trend in prescription and treatment retention of molecular-targeted drugs in 121,131 Japanese patients with rheumatoid arthritis: a population-based real-world study. Mod. Rheumatol., 32, 857–865 (2022).
- 20) Miyashiro M, Asano T, Ishii Y, Miyazaki C, Shimizu H, Masuda J. Treatment patterns of biologic disease-modifying antirheumatic drugs and janus kinase inhibitors in patients with rheumatoid arthritis in Japan: a claims-based cohort study. Drugs Real World Outcomes, 11, 285–297 (2024).
- 21) Rashid N, Lin AT, Aranda G Jr, Lin KJ, Guerrero VN, Nadkarni A, Patel C. Rates, factors, reasons, and economic impact associated with switching in rheumatoid arthritis patients newly initiated on biologic disease modifying anti-rheumatic drugs in an integrated healthcare system. J. Med. Econ., 19, 568–575 (2016).
- 22) Mocci G, Marzo M, Papa A, Armuzzi A, Guidi L. Dermatological adverse reactions during anti-TNF treatments: focus on inflammatory bowel disease. J. Crohns Colitis, 7, 769–779 (2013).
- 23) Batycka-Baran A, Flaig M, Molin S, Ruzicka T, Prinz JC. Etanercept-induced injection site reactions: potential pathomechanisms and clinical assessment. Expert Opin. Drug Saf., 11, 911–921 (2012).
- 24) Tank ND, Karelia BN, Vegada BN. Biological Response Modifiers in Rheumatoid Arthritis: Systematic Review and meta-analysis of Safety. J. Pharmacol. Pharmacother., 8, 92–105 (2017).
- 25) Pintea I, Petricau C, Dumitrascu D, Muntean A, Branisteanu DC, Branisteanu DE, Deleanu D. Hypersensitivity reactions to monoclonal antibodies: Classification and treatment approach. Exp. Ther. Med., 22, 949 (2021).
- 26) Yamanaka H, Tanaka Y, Hibino T, Unmesh G, Shah C, Bakhle D, Stefanidis D. Lower injection-site reactions and long-term safety, immunogenicity, and efficacy of etanercept biosimilar YLB113: Results from a post-hoc analysis of a double-blind, randomized, phase III comparative study and its open-label extension in patients with rheumatoid arthritis. Int. J. Rheum. Dis., 26, 108–115 (2023).
- 27) Waldron JL, Schworer SA, Kwan M. Hypersensitivity and immune-related adverse events in biologic therapy. Clin. Rev. Allergy Immunol., 62, 413–431 (2022).
- 28) Takabayashi K, Ando F, Suzuki T. Comparing the effectiveness of biological disease-modifying antirheumatic drugs using real-world data. Mod. Rheumatol., 29, 87–97 (2019).
- 29) Dalal DS, Duran J, Brar T, Alqadi R, Halladay C, Lakhani A, Rudolph JL. Efficacy and safety of biological agents in the older rheumatoid arthritis patients compared to Young: a systematic review and meta-analysis. Semin. Arthritis Rheum., 48, 799–807 (2019).
- 30) Ochi S, Saito K, Mizoguchi F, Kato S, Tanaka Y. Insensitivity versus poor response to tumour necrosis factor inhibitors in rheumatoid arthritis: a retrospective cohort study. Arthritis Res. Ther., 22, 41 (2020).
- 31) Atzeni F, Antivalle M, Pallavicini FB, Caporali R, Bazzani C, Gorla R, Favalli EG, Marchesoni A, Sarzi-Puttini P. Predicting response to anti-TNF treatment in rheumatoid arthritis patients. Autoimmun. Rev., 8, 431–437 (2009).
- 32) Khader Y, Beran A, Ghazaleh S, Lee-Smith W, Altorok N. Predictors of remission in rheumatoid arthritis patients treated with biologics: a systematic review and meta-analysis. Clin. Rheumatol., 41, 3615–3627 (2022).
