Making a prescribing choice for rheumatoid arthritis: a focus on small molecule drugs vs. biologics for the most favourable patient outcome

ABSTRACT Introduction The molecular targeted therapies available for rheumatoid arthritis include 10 types of biological disease-modifying antirheumatic drugs (bDMARDs) and five types of Janus kinase (JAK) inhibitors. This article reviews the differential and proper use of bDMARDs and JAK inhibitors, focusing on their efficacy and safety, based mainly on phase III clinical trials. Area covered The JAK inhibitors approved for treating rheumatoid arthritis are compared with bDMARDs based on the evidence derived from global phase III trials. Expert opinion In patients with inadequate responses to bDMARDs and JAK inhibitors, switching between these drugs is comparable in efficacy in both directions. Head-to-head comparison demonstrated that baricitinib and upadacitinib are more efficacious than tumor necrosis factor (TNF) inhibitors. However, the ORAL Surveillance study demonstrated that JAK inhibitors are associated with higher incidences of death, major adverse cardiovascular events, malignancies and thrombosis than TNF inhibitors in at-risk patients. The need for risk assessment by pre-treatment screening, regular monitoring during treatment, and appropriate systemic management in adverse events should be recognized. Meanwhile, some JAK-inhibitors were efficacious even for difficult-to-treat disease. These results suggest the need for establishing therapeutic strategies considering the balance between safety and efficacy in individual patients.


Introduction
In 1975, Köhler and Milstein introduced monoclonal antibody technology, for which they were awarded the Nobel Prize in Physiology and Medicine in 1984 [1].Since then, monoclonal antibodies have been applied to treating cancers, transplantation-associated conditions, and intractable diseases.In 1998, monoclonal antibodies targeting tumor necrosis factor (TNF) were marketed for treating rheumatoid arthritis.In treating rheumatoid arthritis, oral glucocorticoids and non-steroidal anti-inflammatory drugs were used in the 20th century to alleviate pain but failed to prevent joint destruction.In the late 20th century, immune abnormalities were found to be involved in the pathogenesis of rheumatoid arthritis, and immunosuppressants were used to suppress immune abnormalities and control the disease.These immunosuppressants are referred to as disease-modifying antirheumatic drugs (DMARDs), which are classified into conventional synthetic DMARDs (csDMARDs) as represented by methotrexate and biological DMARDs (bDMARDs) as represented by monoclonal antibodies targeting TNF and others.While bDMARDs are administered by drip infusion and injection, Janus kinase (JAK) inhibitors were developed using orally available low-molecularweight compounds.JAK inhibitors were evaluated to be highly efficacious and approved as targeted synthetic DMARDs (tsDMARDs) in 2012.
Rheumatoid arthritis is a typical autoimmune disease with polyarthritis manifested as a primary pathological feature [2][3][4][5].It inevitably causes irreversible structural joint damage without appropriate treatment and is often accompanied by various organ dysfunctions.Joint damage progresses soon after the onset, and joint deformity causes irreversible physical dysfunction.Thus, prompt and appropriate diagnosis and treatment are necessary.For rheumatoid arthritis, treatment is initiated with methotrexate soon after diagnosis if it is not contraindicated (Figure 1) [6].When remission is not achieved within 6 months of treatment with methotrexate and when poor prognostic factors are present, bDMARDs or JAK inhibitors in patients without risk factors are added.If the target is not achieved, the added drugs are switched to other bDMARDs or JAK inhibitors in three to six months to achieve remission.These drugs have allowed molecular targeted therapy based on pathological mechanisms, and remission induction before joint destruction has become the therapeutic goal.If remission is maintained, the progression of irreversible structural joint damage and physical dysfunction can be controlled for an extended period.As for molecular targeted drugs, 10 types of bDMARDs and five types of JAK inhibitors are available worldwide, and the differential use and proper use of these drugs have been discussed.This article reviews the differences between bDMARDs and JAK inhibitors in terms of efficacy and safety.

Biological DMARDs and JAK inhibitors
The pathogenesis of rheumatoid arthritis is understood as follows.Genetic and environmental factors interact, and extracellular matrix molecules, such as filaggrin and fibrinogen, are citrullinated, which leads to epigenomic modification, especially ACPA-positive rheumatoid arthritis.Several other factors or mechanisms also may intervene in disease pathological processes.Consequently, immune tolerance to antigens fails, and autoimmunity is induced.Tissues affected by synovitis are characterized by angiogenesis/vascular dilation, accumulation of lymphocytes, and proliferation of synovial cells that are caused by the onset of autoimmune disorders, and various cytokines such as TNF and interleukin 6 (IL-6) are produced in inflamed synovium [2][3][4][5].These cytokines cause extraarticular organ dysfunction such as dry keratoconjunctivitis, sialadenitis, interstitial lung disease, osteochondral destruction and atherosclerosis through intercellular and intracellular signal transduction.In treatment targeting significant molecules involved in this pathogenic process, molecular targeted therapy, bDMARDs and tsDMARDs (JAK inhibitors) are used.
Drugs using molecules existing in the body, such as antibodies, are called biologics.Monoclonal antibodies include human-mouse chimeric antibodies in which mouse antibody constant regions are replaced with human antibody constant regions, humanized antibodies in which the entire region is replaced with that of a human antibody except for the complementarity-determining regions important for antigen binding affinity, and human antibodies that are produced by the phage display or transgenic animals.There are also pegylated antibodies with in vivo stability improved by conjugating polyethylene glycol to the variable region, next-generation heavy-and light-chain antibodies in which two human TNF-binding domains are bound to one human albumin-binding domain, and fusion protein in which soluble receptors are bound to the human antibody constant regions [7].bDMARDs include those targeting TNF (e.g.infliximab, etanercept, adalimumab, golimumab, certolizumab, and ozoralizumab [Japan]), those targeting IL-6 (e.g.tocilizumab and sarilumab), a selective T-cell costimulation modulator (i.e.abatacept), and an anti-CD20 antibody for B cells (i.e.rituximab), all of which are used for patients with rheumatoid arthritis who inadequately respond to methotrexate.Some TNF-targeted drugs inhibit the progression of structural joint damage and physical dysfunction over 10 years.

Article highlights
• This article reviews the differences between bDMARDs and JAK inhibitors in terms of efficacy and safety in patients with rheumatoid arthritis and with inadequately responses to methotrexate.• Switching between bDMARDs and JAK inhibitors is suggested to be comparable in efficacy in both directions in patients with inadequate responses to each of them.• In phase III clinical trials of JAK inhibitors, baricitinib and upadacitinib were more efficacious than TNF inhibitors, while tofacitinib and filgotinib were comparable to them by head-to-head comparison.
• The ORAL Surveillance study that the incidence of MACEs and malignant tumors was higher for the tofacitinib group than the TNF inhibitors group in patients aged 50 years or older who had at least one risk factor for MACEs.Conversely, JAKs are typical tyrosine kinases involved in the signal transduction of approximately 40 types of cytokines and growth factors [8][9][10][11].Cytokines and growth factors bind to cell surface receptors and activate associating JAKs, which phosphorylate the intracellular domain of the receptors and bind to transcription factors, signal transducers and activators of transcription (STATs).When STATs are phosphorylated, they translocate into the nucleus and induce transcription of the next gene.Four types of JAK isoforms, JAK1, JAK2, JAK3, and tyrosine kinase 2, form heterodimers or homodimers and transmit intracellular signals through combination with seven types of STATs to induce transcription of various cytokines and cell functional molecules.Tofacitinib, which is the first drug approved for rheumatoid arthritis treatment, is an orally available lowmolecular-weight targeted drug with a molecular weight of approximately 312 that competitively binds to the adenosine triphosphate binding site of JAK3 and specifically inhibits signal transduction of these cytokines.Because its half-life is approximately three hours, this drug is pharmacologically characterized by rapid in vivo metabolism [12].In addition to tofacitinib, other JAK inhibitors, including baricitinib, peficitinib (Japan), upadacitinib, and filgotinib, are approved for rheumatoid arthritis treatment in patients with inadequate responses to methotrexate.In these patients some JAK inhibitors appeared superior to TNF-inhibitors in efficacy but comparable to them in safety by head-to-head comparison studies.However, post-marketing ORAL surveillance study showed TNF inhibitors were superior to tofacitinib in safety in patients with moderate to high disease activity despite methotrexate and with risk factors such as older than 50 years and cardiovascular risks.The following sections will discuss these points in detail.

JAK inhibitors for patients with inadequate responses to bDmards
TNF inhibitors were first approved in developing molecular targeted drugs for rheumatoid arthritis treatment.Consequently, the subsequently developed non-TNF bDMARDs and JAK inhibitors were tested in phase III clinical trials, including patients with inadequate responses to TNF inhibitors or bDMARDs (Table 1).
The ORAL Step trial compared tofacitinib at 5 or 10 mg twice daily with placebo in addition to methotrexate in 399 patients with rheumatoid arthritis who inadequately responded to TNF inhibitors [13].After three months of treatment, the American College of Rheumatology 20% improvement (ACR20) response rate in the tofacitinib 5-mg group was 41.7%, significantly higher than the 24.4% in the placebo group.The changes in the Health Assessment Questionnaire Disability Index scores were −0.43 in the active drug groups, showing a significant improvement compared to −0.18 in the placebo group.The rate of achieving a 28-joint Disease Activity Score (DAS28) using an erythrocyte sedimentation rate of < 2.6 in the active drug groups was 6.7%, significantly higher than the 1.7% in the placebo group.Both findings met the primary endpoint.After this trial, tofacitinib was approved in the United States in 2012 as the first JAK inhibitor efficacious for patients with rheumatoid arthritis who inadequately respond to TNF inhibitors.
The RA-BEACON trial compared baricitinib at 2 or 4 mg once daily plus csDMARD with placebo in 527 patients with rheumatoid arthritis who inadequately responded to TNF inhibitors [14].After 12 weeks of treatment, the ACR20 response rates were 49% in the baricitinib 2-mg group and 55% in the baricitinib 4-mg group.These rates were significantly higher than the placebo group (27%) and met the primary endpoint.Accordingly, baricitinib was approved.
The SELECT-BEYOND trial compared upadacitinib at 15 or 30 mg once daily plus csDMARD with placebo in 499 patients with rheumatoid arthritis who inadequately responded to bDMARDs [15].After 12 weeks of treatment, the ACR20 response rate in the upadacitinib 15-mg group was 64.6%, significantly higher than the 28.4% in the placebo group.The rate of achieving a DAS28 using C-reactive protein (CRP) of 3.2 or higher in the active drug groups was 43.3%, which was also significantly higher than the 14.2% in the placebo group.Both findings met the primary endpoint.This drug showed similar ACR20 response rates, even in patients with inadequate responses to three or more bDMARDs with the same action mechanisms or two or more bDMARDs with different action mechanisms and patients with inadequate responses to anti-IL-6 receptor antibodies.
The FINCH2 trial compared filgotinib at 100 or 200 mg once daily plus csDMARD with placebo in 448 patients with rheumatoid arthritis who inadequately responded to bDMARDs [16].After 12 weeks of treatment, the ACR20 response rate in the filgotinib 200-mg group was 66.0%, significantly higher than the 31.1% in the placebo group and met the primary endpoint.This drug showed a similar efficacy even in patients with inadequate responses to three or more bDMARDs and anti-IL-6 receptor antibodies.In contrast, no studies have compared the efficacy of bDMARDs with placebo in patients with rheumatoid arthritis who inadequately respond to JAK inhibitors.The SELECT-COMPARE trial compared upadacitinib at 15 mg once daily, adalimumab at 40 mg once every two weeks, and placebo, in combination with methotrexate, in 1629 patients with rheumatoid arthritis who inadequately responded to methotrexate [17].After the double-blind comparison phase of this trial, the effects of reciprocal switching between upadacitinib and adalimumab were investigated in patients with < 20% improvement in swollen or tender joint counts at weeks 14, 18, and 22 and patients with moderate disease activity, as defined as a clinical disease activity index (CDAI) of < 10.0 at week 26.The efficacy of switching between these drugs was compared in patients with inadequate responses to each drug.The results revealed that 39% of the patients inadequately responded to upadacitinib and switched to adalimumab, and 49% inadequately responded to adalimumab and switched to upadacitinib.At week 48, 62% of the patients assigned to upadacitinib at the baseline achieved CDAI low disease activity, and 28% achieved remission.Of the patients assigned to adalimumab, 59% achieved low disease activity, and 25% achieved remission.Although these percentages tended to be higher in the upadacitinib group, no significant differences were observed.These results showed that the effects of reciprocal switching between upadacitinib and adalimumab are almost comparable [18].
Thus, in patients with inadequate responses to bDMARDs and JAK inhibitors, switching between these drugs is suggested to be almost comparable in efficacy in both directions.

Comparison of efficacy between bDmards and JAK inhibitors
In phase III clinical trials of JAK inhibitors, their efficacy has been directly compared with that of bDMARDs such as TNF inhibitors.
The ORAL Strategy trial compared tofacitinib at 5 mg twice daily with or without methotrexate and subcutaneous injection of adalimumab at 40 mg once every two weeks with methotrexate in 1252 patients with rheumatoid arthritis who inadequately responded to methotrexate [19].ACR50 response was attained in 38% of patients with tofacitinib monotherapy, 46% of patients with tofacitinib and methotrexate, and 44% of patients with adalimumab and methotrexate at 6 months.Non-inferiority was declared for tofacitinib and methotrexate versus adalimumab and methotrexate but not for tofacitinib monotherapy versus either adalimumab and methotrexate or tofacitinib and methotrexate.
The RA-BEAM trial compared baricitinib at 4 mg once daily, subcutaneous injection of adalimumab at 40 mg once every two weeks, and placebo in 1307 patients with rheumatoid arthritis who inadequately responded to methotrexate [20].At week 12, more patients had an ACR20 response with baricitinib (70%) than with placebo (40%) and the study, therefore, met primary end point.In addition, the difference of ACR20 response rate between the baricitinib and adalimumab groups was 8.4 by the Newcombe-Wilson method, and the 95% confidence interval was 1.7 to 15.1, which did not cross 0, indicating that baricitinib was the first low-molecularweight JAK inhibitor that was demonstrated to be superior in efficacy to bDMARDs.
The SELECT-COMPARE trial, as described above, compared upadacitinib at 15 mg once daily, adalimumab at 40 mg once every two weeks, and placebo, in combination with methotrexate, in 1629 patients with rheumatoid arthritis who inadequately responded to methotrexate [17].After 12 weeks of treatment, the upadacitinib 15-mg group showed an ACR20 response rate of 71%, significantly higher than the 36% in the placebo group and the 63% in the adalimumab group.The DAS28(CRP)≤3.2 achievement rate in the upadacitinib 15-mg group was 29%, significantly higher than the 6% in the placebo group and the 18% in the adalimumab group.Both rates in the upadacitinib group met the primary endpoint, and upadacitinib was demonstrated to be more efficacious than the placebo and adalimumab.
The SELECT-CHOICE trial compared upadacitinib and intravenous injection of abatacept in patients with rheumatoid arthritis who inadequately responded to csDMARDs [21].The mean DAS28(CRP) change from baseline to week 12 in the upadacitinib group was −2.52, significantly greater than −2.00 in the abatacept group.The non-inferiority analysis also showed the superiority of upadacitinib with an intergroup difference of −0.52 and a 95% confidence interval of −0.69 to −0.35.
The FINCH1 trial compared filgotinib at 100 or 200 mg once daily, subcutaneous injection of adalimumab at 40 mg once every two weeks, and placebo in 1755 patients with rheumatoid arthritis who inadequately responded to methotrexate [22].After 12 weeks of treatment, the ACR20 response rates were 49.4%, 69.8%, 76.6%, and 70.5% in the placebo, filgotinib 100-mg, filgotinib 200-mg, and adalimumab groups, respectively.These rates in all filgotinib and adalimumab groups were significantly higher than that in the placebo group and met the primary endpoint.Based on the DAS28(CRP)≤3.2achievement rates at week 12, filgotinib at 200 mg was also demonstrated to be non-inferior to adalimumab.
The RAJ3 trial compared peficitinib at 100 or 150 mg once daily, subcutaneous injection of etanercept at 50 mg once weekly, and placebo in 507 patients with rheumatoid arthritis who inadequately responded to DMARDs, including bDMARDs [23].Etanercept was used for reference and not directly compared to peficitinib.After 12 weeks of treatment, the ACR20 response rates were 30.7%, 57.7%, 74.5%, and 83.5% in the placebo, peficitinib 100-mg, peficitinib 150-mg, and etanercept groups, respectively.These rates in all peficitinib and etanercept groups were significantly higher than that in the placebo group and met the primary endpoint.Non-inferiority analysis was not performed.
Based on these trials, baricitinib and upadacitinib were more efficacious than TNF inhibitors, while tofacitinib and filgotinib were comparable to TNF inhibitors.These results showed the superiority of orally available low-molecularweight drugs to bDMARDs.

Comparison of safety between bDmards and JAK inhibitors
Based on the post-marketing all-case surveillance of bDMARDs in Japan, the most common serious adverse drug reaction was bacterial pneumonia, and the risk factors identified for this reaction included advanced age, history of respiratory diseases, and concomitant use of glucocorticoid [24,25].These results also revealed the need for medical management, prevention, and treatment of serious adverse drug reactions such as opportunistic infections.
Alternatively, the short-and long-term safety of JAK inhibitors was assessed based on results and findings from the extension phase of each clinical trial.A survey with the most prolonged follow-up period of 9.5 years on the longterm safety of tofacitinib showed no temporal changes in the incidences of infections, opportunistic infections, serious infections, malignant tumors, thrombosis, and cardiovascular disorders [26].In five phase III trials using upadacitinib, pooled safety analyses showed that upadacitinib was comparable to methotrexate and adalimumab in short-and long-term safety, with the exception that the risks of herpes zoster and elevated creatine phosphokinase levels were higher in patients treated with upadacitinib than in those treated with adalimumab [27].In addition, analyses of randomized controlled trials on tofacitinib and baricitinib suggested a possible dose-dependent pattern of the risk of infections and identified the development of herpes zoster, mainly in Asians, as one of the common safety issues with either JAK inhibitor [28,29].None of the previously performed clinical trials has indicated the risks of JAK inhibitors concerning MACEs and malignant tumors.However, because aging and other complications (e.g.hypertension and smoking) are apparent risk factors for these conditions, their incidences differ entirely between patients with and without these risk factors.It has been challenging to investigate the true safety of JAK inhibitors.
In these circumstances, the ORAL Surveillance study was conducted to compare the safety of tofacitinib and TNF inhibitors in patients aged 50 years or older with rheumatoid arthritis who had at least one risk factor for MACEs (e.g.smoking, hypertension, diabetes mellitus, and a history of coronary artery disease) [30].It was a large-scale trial with the primary endpoint of the incidences of MACEs and malignant tumors, in which 4362 patients were followed up for approximately six years.The incidence of MACEs was 2.5% in the TNF inhibitor group receiving adalimumab or etanercept and 3.4% in the tofacitinib group receiving either 5 or 10 mg twice daily.These results failed to demonstrate the noninferiority of tofacitinib to TNF inferiors and did not meet the primary endpoint specified by the 95% confidence interval.The incidence of MACEs was shown to be higher for tofacitinib.
Similarly, the incidence of malignant tumors was 2.9% in the TNF inhibitor group and 4.2% in the tofacitinib group, significantly higher for tofacitinib.Based on these results, the Food and Drug Administration of the United States and the European Medicines Agency issued the latest boxed warning regarding increased risks of death, MACEs, malignant tumors, and thrombosis with JAK inhibitors compared to TNF inhibitors and recommended that JAK inhibitors should be used in patients aged 65 years or older, those who previously smoked or currently smoke, and those with cardiovascular risk factors only when no appropriate alternative therapies are available.Furthermore, because of concerns about the class effects of JAK inhibitors, these recommendations are applied to all JAK inhibitors approved for treating inflammatory diseases and all approved uses of the inhibitors in the United States and Europe.
However, because the immunological mechanisms by which JAK inhibitors increase the incidences of MACEs and malignant tumors are unknown, there is an increasing need to evaluate in vivo mechanisms.In any case, it seems essential to thoroughly perform pre-treatment screening and subsequent monitoring in the real-world clinical practice.

Conclusions
The advent of bDMARDs and JAK inhibitors has allowed molecular targeted therapy based on pathological mechanisms and remission induction has become the therapeutic goal in most cases of rheumatoid arthritis.Ten types of bDMARDs and five types of JAK inhibitors are available globally.We have reviewed the differential use and proper use of these drugs, focusing on both efficacy and safety, based mainly on the results of phase III clinical trials.In patients with inadequate responses to bDMARDs and JAK inhibitors, switching between these drugs has been demonstrated to be comparable in efficacy in both directions.Studies directly comparing efficacy have demonstrated that baricitinib and upadacitinib are more efficacious than TNF inhibitors and that tofacitinib and filgotinib are comparable to TNF inhibitors.Meanwhile, in clinical trials on safety, the short-and long-term safety of JAK inhibitors was assessed based on results and findings from the extension phase of each clinical trial.However, the ORAL Surveillance trial results demonstrated that JAK inhibitors are associated with a higher risk of death, MACEs, malignant tumors, and thrombosis than TNF inhibitors in at-risk patients.In any case, it seems important to thoroughly perform a risk assessment based on pre-treatment screening and subsequent regular monitoring in the real-world setting.

Safety
Many adverse events associated with JAK inhibitors are predicted by the mechanisms associated with blockade of cytokines that use the JAK-STAT signal transduction, and these mechanisms can explain the risk of serious and opportunistic infections, such as herpes zoster.Many contraindications for using JAK inhibitors are associated with their pharmacokinetic and pharmacodynamic profiles and adverse events.Those contraindications include severe active infections (e.g.latent tuberculosis and opportunistic infections), active malignant tumors, severe organ dysfunction, pregnancy, and lactation.However, the development of thromboembolism is a relatively rare, unexpected, and inexplicable event.It remains unknown whether it is involved in activating the coagulation and fibrinolytic system or activating platelets and endothelial cells.
The regulatory authorities in the United States and Europe pointed out risks such as cardiovascular disorders, malignant tumors, and thrombosis, based on results of ORAL surveillance using tofacitinib, and extended the risks to the class of JAK inhibitors.The risks of infections, cardiovascular disorders, thrombosis, and malignant tumors should also be assessed sufficiently, and the need for systemic management in the event of adverse events should be recognized.Rheumatoid arthritis is associated with various organ dysfunctions, including dry eyes/oral dryness, interstitial lung disease, and mononeuritis multiplex.Furthermore, there are concerns about adverse drug reactions to the long-term use of glucocorticoids, such as lipid, glucose, and bone metabolism disorders, immunocompromised conditions due to immunosuppression, and risks of cardiovascular and cerebrovascular disorders.The long-term use of methotrexate is suggested to be associated with hepatic and renal disorders, bone marrow suppression and infections.Thus, while drug-induced immunosuppression, metabolic disorders and organ dysfunction associated with drugs or rheumatoid arthritis are comprehensively evaluated, therapeutic strategies should be determined and systematically implemented.
Approximately 5000 patients with rheumatoid arthritis who initiate or switch to treatment with bDAMRDs and/or JAK inhibitors have been all registered in the real-world clinical practice at our department (FIRST registry) since 2003 [31][32][33].Enrolled patients have been placed on the clinical-path during hospitalization for safe treatment and undergo systemic screening before treatment to determine whether they have contraindications or risk factors.Nurses provide inpatient care and instructions for self-injection, designated administrative staff members manage data, and research assistants handle and measure samples.After indicated drugs are identified, the most effective and safe drug is selected through shared decision making between rheumatologists and patients.When malignant tumors or infection, such as tuberculosis, is accidentally detected, treatment of such conditions is prioritized to minimize adverse events.After discharge, the patients are thoroughly and regularly monitored at the outpatient clinics of our hospital and affiliated hospitals.In case of adverse drug reactions, the patients are promptly placed under systemic management so that they can be treated.
Conversely, it is necessary to re-recognize that the disease activity of rheumatoid arthritis is associated with increased incidence and mortality of organ dysfunction due to rheumatoid arthritis, such as interstitial lung disease and vasculitis, increased incidence of malignant tumors such as malignant lymphoma, increased incidences of cardiovascular and cerebrovascular disorders due to prolonged arteriosclerotic inflammation, increased incidence of thrombosis due to inflammation-induced coagulation disorder, and the risk of exacerbating infection such as increased incidence, exacerbation rates, and mortality of Covid-19.Effective disease control by appropriate treatments should be emphasized even in patients presenting risk factors such as older age, active smoking, CV risk factors, etc.Thus, thorough consideration seems necessary while balancing the reduction in risk factors by controlling disease activity with JAK inhibitors and bDMARDs and the increase in risk factors due to these drugs.Therapeutic strategies should be established considering the balance between safety and efficacy in individual patient.

Difficult-to-treat rheumatoid arthritis
Despite the advent of many types of molecular targeted drugs, multidrug resistance remains an important unmet need in treating rheumatoid arthritis.The European League Against Rheumatism defined difficult-to-treat rheumatoid arthritis (D2T-RA) in patients unresponsive to csDMARD by resistance to two or more bDMARDs with different targets or tsDMARDs [34].The existing bDMARDs were less effective in patients unresponsive to one bDMARD than in patients unresponsive to csDMARDs, and the rate of treatment refractoriness further increased in patients with D2T-RA.However, as described above, upadacitinib and filgotinib, JAK inhibitors, remained effective even for patients despite two or more bDMARDs [16,17].In the FIRST registry of about 5,000 realworld clinical setting, patients who met the criteria for D2T-RA accounted for 16.6%.When the background factors of patients using bDMARDs and patients using JAK inhibitors were adjusted by inverse probability of treatment weighting based on the propensity score, patients treated with JAK inhibitors showed a significant improvement in CDAI scores at 6 months, compared to patients treated with bDMARDs (Figure 2) [35].When the trajectories of therapeutic effects were analyzed using Latent Class analysis, the proportion of patients with poor responses was significantly lower in patients treated with JAK inhibitors than in patients treated with bDMARDs.In particular, the significant difference was more pronounced in patients not treated with methotrexate or glucocorticoids.The hazard ratio for serious adverse events tended to be lower for JAK inhibitors.Based on the above, JAK inhibitors are useful in patients with multidrug-resistant rheumatoid arthritis.However, the use of JAK inhibitors may not be relevant in D2T-RA patients without objective inflammation and careful differential diagnosis would be required.

Reviewer disclosures
Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Treatment duration (years) Screening of risk factors Before joint damage intensive treatment with methotrexate Methotrexate + bDMARD Methotrexate + JAK-inhibitor Regularly monitoring safety Adherence Safety management Efficacy maintenance Patient-reported outcomes De-escalation of treatment Remission induction Remission maintenance Drug holiday? Methotrexate + JAK-inhibitor Difficult-to-treat RA Figure 1. Treatment
strategies of rheumatoid arthritis.Note: Intensive treatment is required for inducing remission in RA, but subsequently maintaining remission with high adherence and safety is prerequisite for the good long-term outcome.

Table 1 .
Biological DMARDs vs. JAK-inhibitors in patients with rheumatoid arthritis.