Treatment Patterns of Biologic Disease-Modifying Antirheumatic Drugs and Janus Kinase Inhibitors in Patients with Rheumatoid Arthritis in Japan: A Claims-Based Cohort Study

Rheumatoid arthritis (RA) is a chronic, inflammatory autoimmune joint disease that can cause cartilage and bone damage as well as disability, which may negatively impact patients’ ability to perform daily activities and reduce health-related quality of life [1, 2]. The global prevalence of RA between 1980 and 2019 is estimated to be 0.46% based on a recent systematic review [3].

RA therapies have slowly evolved over the last century. Previously, most were based on conventional synthetic disease-modifying antirheumatic drugs (csDMARDs). However, over the past 20 years, RA treatment has undergone an innovative improvement, including the approval of effective biologic and targeted synthetic DMARDs (b/tsDMARDs) [4, 5]. With these improvements, the recommended therapeutic strategy for RA has been a treat-to-target (T2T) approach, aiming to achieve and sustain disease remission [6]. With T2T, csDMARDs are recommended as first-line treatment, most commonly methotrexate; bDMARDs or tsDMARDs such as Janus kinase inhibitors (JAKi) are recommended for use in patients with inadequate response to csDMARDs [6]. Following this dramatic improvement in RA therapies, RA experts anticipate further therapeutic advances that include the differential use of therapeutic drugs and de-escalation of treatment after remission induction [7].

In Japan, a tumor necrosis factor inhibitor (TNFi) was approved for the indication of RA in 2003 as the first bDMARD introduced in real-world clinical practice, followed by an interleukin-6 inhibitor (IL-6i) and cytotoxic T lymphocyte-associated antigen-4 immunoglobulin fusion proteins (CTLA4-Ig), approved in 2008 and 2010, respectively [8‐10]. These three classes of bDMARDs are currently clinically available, covered by insurance, and widely used in Japanese clinical practice [11]. An estimate based on the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB) in 2017 shows that 22.9% of 825,772 patients with RA aged ≥ 16 years received treatment with bDMARDs, increasing to 50.9% in those aged 16–19 years, and decreasing to 13.7% in those aged ≥ 85 years [11]. In Japan, the first JAKi was approved for RA in 2013 [12]. Subsequently, JAKi are increasingly used as second-line treatment following methotrexate, and the total number of JAKi prescriptions reported increased approximately 8-fold during 2016–2019, based on an analysis of the Japan Medical Data Center (JMDC) claims database containing 10,399 patients with RA [13].

Recent availability of several different types of bDMARD/JAKi now provides opportunities for multiple therapeutic strategies in Japanese clinical practice, with different treatment patterns. In harmony with EULAR endorsement of T2T [6], Japanese clinical practice guidelines for RA management recommend the use of bDMARDs/JAKi, regardless of csDMARD use [14]. Reports on Japanese treatment patterns of bDMARD/JAKi use are still sparse, especially in combination with csDMARDs. We thus conducted a retrospective study of a claims-based database between April 2008 and August 2020, to investigate bDMARD/JAKi treatment patterns, with or without csDMARDs, in Japanese patients with RA. We also analyzed factors that contribute to initiation, switching, temporary discontinuation, and discontinuation of therapies with bDMARDs/JAKi.

In the present study, using the MDV claims database, we reported the treatment patterns of bDMARDs and JAKi in clinical practice in Japan from 2008 to 2020, with or without csDMARDs. TNFi were the main first-line therapy, especially in combination with csDMARDs, and were likely to be switched to a different class regardless of combination with csDMARDs. Temporary discontinuation and discontinuation of bDMARDs/JAKi were also observed during the follow-up periods, irrespective of combination with csDMARDs.

Several articles have recently reported treatment patterns of bDMARDs/JAKi in real-world clinical practice in Japan [11, 13, 15]. Takabayashi et al. reported prescription trends and treatment retention of molecular-targeted drugs using the NDB from 2010 to 2016 [15]. A report by Kaneko et al. from the JMDC database demonstrated JAKi/bDMARD treatment patterns from 2016 to 2019 [13]. However, these reports did not fully examine treatment patterns related to the combination of advanced therapies with csDMARDs. In addition, our study using the MDV claims database covered the older RA patient population, which differed from the study by Kaneko et al. of a primarily working-age RA patient population (no enrollees aged ≥ 75 years) [13].

Our study results showed that 3.5% or 4.1% of patients were prescribed JAKi among bDMARD/JAKi with or without csDMARDs, respectively. It was a slightly higher rate than a previous report from NDB where JAKi were prescribed in 0.9% of the patients [11]. This may be caused by the difference in data source between MDV and NDB databases and in the analysis period between our study (2008‒2020) and the previous report from 2017. This might also suggest a recent increase in the use of JAKi. Among the study cohort of RA patients who were prescribed any DMARDs, the proportion of patients who were prescribed each bDMARD/JAKi generally continued to increase over time, after the introduction of each drug in Japan, suggesting that bDMARDs/JAKi play a key role in real-world clinical practice in Japan. Among different types of bDMARDs/JAKi available in clinical practice in Japan, TNFi were the most commonly prescribed drugs between 2008 and 2020. Accumulated evidence in Japanese RA patients, including post-marketing surveillance (PMS), supports increasing TNFi prescription numbers since 2008 [16, 17]. Among the bDMARDs/JAKi, TNFi were most commonly prescribed as first-line bDMARDs. In line with the prescribing information, TNFi were more commonly used in combination with csDMARDs, with more prescription of non-TNFi biologics (IL-6i, CTLA4-Ig, and JAKi) as monotherapy. This finding could be influenced by the guideline recommending an IL-6i and a JAKi to be used in patients who should not receive methotrexate/csDMARD [6, 14]. Regardless of combination with csDMARDs, JAKi were more frequently prescribed as second- or third-line bDMARDs/JAKi compared with first-line prescription. As in a previous report [13], the JAKi prescription numbers have continued to increase. However, as long-term safety of JAKi has not been fully established, the guidelines recommend consideration of JAKi use only after risk assessment, such as cardiovascular events and malignancies, regardless of concomitant use of csDMARDs [6, 14]. This may reflect our findings.

Approximately 30% of patients starting bDMARDs/JAKi therapy regardless of combination with csDMARDs are thought to switch to another bDMARDs or JAKi in real-world clinical practice in Japan. Sankey diagrams indicate varied switching pathways from first-line bDMARDs/JAKi to second- or third-line therapy, either switching to another drug class or within a class. Compared with intra-class switching, different class switching from TNFi was observed with second or third-line prescription, regardless of csDMARD combination. The guidelines suggest switching to non-TNFi biologics rather than another TNFi in RA patients failing TNFi therapy [6, 14]. This may explain our finding. We explored factors contributing to initiation and switching of therapies with bDMARDs/JAKi. Ages 61–75 years and > 75 years were significant negative determinants of first-line prescription of TNFi and IL-6i, and were otherwise a significant positive determinant of CTLA4-Ig use. A previous study using NDB Japan was consistent with our findings, indicating decreased TNFi or IL-6i use and increased CTLA4-Ig use with age [11]. Reasons for the difference in the treatment trend between TNFi/IL-6i and CTLA4-Ig are unknown, but some of the following evidence might indicate that this is a trend in treatment. Japanese studies have reported the risk of serious infection, especially in elderly patients with RA treated with TNFi or IL-6i [18, 19], while the 2015 ACR guidelines suggested the use of CTLA4-Ig in patients with previous serious infections, due to the lower risk of subsequent hospitalized infections among patients prescribed abatacept [2]. Interstitial lung disease (ILD) may occur more frequently among older RA patients, and CTLA4-Ig might have a lower risk of worsening pre-existing ILD [20]. In this study, a higher rate of chronic lung disease was seen in patients prescribed CTLA4-Ig, as compared with other bDMARDs/JAKi. In contrast, data from patients with older-onset RA (mean age 74.9 years) from a prospective registry study conducted in Japan indicate significant benefits from treatment initiated with non-biologic DMARDs, followed by TNFi or IL-6i, in order to achieve structural remission, functional remission, and low disease activity in clinical practice [21]. As suggested in the guideline, bDMARDs/JAKi should be used in elderly RA patients with particular consideration of safety, despite csDMARD therapy [14].

In the present study, temporary discontinuation and discontinuation of bDMARDs/JAKi were observed in patients both in combination with csDMARDs and without csDMARDs. Although reasons for temporary discontinuation and discontinuation were not available in this study, factors such as achievement of remission, adverse events, comorbidities, and treatment cost could be associated with this observation. The latest Japanese treatment guidelines suggest tapering of bDMARD/tsDMARD therapies in RA patients who maintain remission or low disease activity after glucocorticoid tapering [14]. Although, according to this recommendation, csDMARDs should not be stopped, we found some patients who experienced temporary discontinuation and discontinuation of bDMARDs/JAKi during the study period. In this study, multivariate analyses showed age 61–75 years and > 75 years were likely to be associated with discontinuation but not temporary discontinuation. In elderly patients, rheumatologists are required to consider patients’ safety carefully, including the presence of comorbidities such as ILD, infections, cardiovascular events, methotrexate toxicity, and effectiveness of re-treatment following discontinuation. A lower rate of temporary discontinuation was observed with the TNFi infliximab, possibly supported by the likelihood of infusion reactions with infliximab reintroduction [16, 22], a potential barrier to reintroduction. Although discontinuation of bDMARDs/JAKi ultimately leads to a flare, recent reports document discontinuation [23‐30]. There may be various reasons for discontinuation; however, our study still provides real-life data on temporary discontinuation or discontinuation of bDMARDs/JAKi by drug. As the guidelines recommend, there needs to be a shared decision between rheumatologists and patients about treatment, taking into consideration disease activity, safety issues, and other patient factors, such as comorbidities and progression of structural damage.

Several limitations should be discussed. First, generalizability of the findings in our study should be interpreted with caution. Our data were generated in large Japanese DPC hospitals, we cannot rule out the possibility of a bias toward patients whose RA was more severe and thus were more likely to be treated with advanced therapies. The results are therefore not necessarily representative of RA treatment in Japan in daily practice. Second, regarding the definition of ‘first-line’ in this study, it is possible that patients in this cohort already started the first-line treatment in another hospital/clinic. Additionally, this MDV database does not allow RA patients that changed hospitals during the treatment to be followed. Third, we could not determine reasons for temporary discontinuation and discontinuation of treatment, potentially due to adverse events, lack of efficacy, or even clinical remission, because claims data analysis in general can utilize only a very limited set of medical information and parameters. Fourth, we explored the determinants of the first-line prescription, switching, temporary discontinuation, and discontinuation, using candidate predictors of age, sex, CCI score, and concomitant csDMARDs use. There may be other candidate predictors such as disease severity and RA duration, however, we could not evaluate these candidates because of the limitation of the study database. In addition, methotrexate was included in csDMARDs, because we focused on treatment pattern of bDMARDs/JAKi in this study. In the guideline, positioning of methotrexate differs from that of other csDMARDs and it is possibility that the difference between concomitant use of methotrexate and other csDMARDs could impact on the prescription of bDMARDs/JAKi. Fifth, determinants for prescription and the mean days of temporary discontinuation or discontinuation can be affected by available treatment options, prescription number, and follow-up period. Finally, sample sizes were heterogeneously distributed in each drug subclass, and were especially small in the JAKi group because of its recent launch in Japan. We anticipate future analysis of more RA patient data in real-world clinical practice in Japan, especially those including more patients treated with JAKi.

The real-world treatment trajectories of advanced therapies, with and without csDMARDs, were analyzed in Japanese RA patients using a claims-based database between 2008 and 2020. TNFi were the main first-line therapy, and likely to be switched to different classes. Understanding the current treatment patterns, including discontinuation, is important to find an optimal treatment strategy for RA patients.