Rheumatoid arthritis is a chronic, systemic, inflammatory, autoimmune disease characterized by progressive joint destruction and systemic manifestations (1–3). The pathogenesis of rheumatoid arthritis is largely attributed to the dysregulation of both innate and adaptive immune responses (1–3). In recent years, an advanced understanding of pro-inflammatory and anti-inflammatory effects of specific cytokines involved in the pathogenesis of rheumatoid arthritis has driven the development of targeted disease-modifying antirheumatic drugs (DMARDs), including biologic DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs), i.e., Janus kinase (JAK) inhibitors.

JAK inhibitors, including tofacitinib, baricitinib, upadacitinib, and filgotinib, inhibit JAKs by competing with the biological substrate, the adenosine triphosphate (ATP) (4–6). The ATP concentration in patients with rheumatoid arthritis has been linked to the level of inflammation (7–9), which is related to rheumatoid arthritis disease activity. JAK inhibitors are indicated, after risk assessment, for the treatment of moderate to high active rheumatoid arthritis in adult patients who have shown an inadequate response to, or who are intolerant to one or more DMARDs (10–13), as well as being a treatment option for “difficult-to-treat” rheumatoid arthritis patients in real-world clinical practice (14–16).

Patients with rheumatoid arthritis receiving treatment with JAK inhibitors in real-world settings often differ from those typically included in randomized controlled trials (RCTs), where study populations are meticulously selected and rigorously monitored under predefined environments (17, 18). In clinical practice, real-world patients often exhibit interindividual differences, including diverse sociodemographic characteristics, rheumatoid arthritis disease activities, comorbidities, concomitant treatments, and treatment histories. This interindividual variability can contribute to the fact that not all eligible patients achieve the primary goal in rheumatoid arthritis treatment. According to the treat-to-target (T2T) recommendations (19, 20), which are supported by several rheumatoid arthritis guidelines (14, 21, 22), the defined primary goal is to achieve and maintain clinical remission or, at least, low disease activity.

The present study aims to evaluate patient- and drug-related factors to help understand interindividual variations in the treatment effectiveness of JAK inhibitors. Real-world evidence, derived from both prospective or retrospective studies, significantly complements the information obtained from RCTs, providing valuable insights that enhance healthcare decision-making (17, 18).

One hundred fifty rheumatoid arthritis patients treated with a JAK inhibitor were included in this study. Table 1 provides a summary of their demographic and clinical characteristics.

Similar distribution of JAK inhibitor type was observed between patients achieving and not achieving the T2T goal at 6 months of treatment. At JAK inhibitor treatment initiation, T2T goal achievers and non-achievers presented comparable years of age, sex distribution, BMI, years of rheumatoid arthritis disease duration, seropositivity considering RF and anti-CCP antibodies, prior csDMARD use, and prior JAK inhibitor use.

In terms of prior bDMARD use, patients who did not achieve the T2T goal at 6 months of the JAK inhibitor treatment, exhibited a higher discontinuation of prior treatment with adalimumab (p = 0.021), certolizumab (p = 0.022), etanercept (p = 0.001), abatacept (p = 0.049), and anakinra (p = 0.043). No significant differences were noted between the T2T goal achievers and non-achievers regarding the discontinuation of prior treatment with golimumab, infliximab, tocilizumab, sarilumab, and rituximab.

Concerning rheumatoid arthritis disease activity, patients who did not achieve the T2T goal presented higher disease activity both at baseline (p < 0.001) and at 6 months of the JAK inhibitor treatment (p < 0.001).

Regarding concomitant rheumatoid arthritis treatment, patients who did not achieve the T2T goal presented a higher use of concomitant glucocorticoids (p < 0.001) and received higher equivalent doses of prednisone (p < 0.001) compared to those who did achieve the T2T goal at the first 6 months of the JAK inhibitor treatment. About concomitant csDMARD use, sulfasalazine use was found to be higher among T2T goal non-achievers compared to achievers, whereas the concomitant use of methotrexate, leflunomide, and hydroxychloroquine was comparable between both groups of patients.

The results from the logistic regression analyses are presented in Table 2.

High disease activity at baseline, with respect to moderate activity, was identified as a potential unfavorable factor for achieving the T2T goal at the first 6 months of the JAK inhibitor treatment [unadjusted odds ratio (OR): 0.94; 95% confidence interval (CI): 0.91–0.98; p = 0.001]. Similarly, the prior number of JAK inhibitor treatments (OR: 0.61; 95% CI: 0.38–0.99; p = 0.046), the prior number of bDMARD treatments (OR: 0.79; 95% CI: 0.68–0.91; p = 0.001), the prior use of a TNF-α inhibitor (OR: 0.67; 95% CI: 0.52–0.87; p = 0.002), the prior use of a IL6 inhibitor (OR: 0.59; 95% CI: 0.36–0.96; p = 0.032), and the prior use of abatacept (OR: 0.51; 95% CI: 0.27–0.99; p = 0.047) were considered as potential unfavorable factors, while age was determined as a potential favorable factor (OR: 1.03; 95% CI: 1.00–1.05; p = 0.043).

Through multivariate model, high disease activity at baseline was established as an unfavorable factor for achieving the T2T goal at the first 6 months of the JAK inhibitor treatment, with an adjusted odds ratio (ORadj) of 0.96 (95% CI: 0.92–0.99; p = 0.028). None of the following factors were identified as predictive, either favorable or unfavorable: age; female sex; BMI; rheumatoid arthritis disease duration; seropositivity for rheumatoid factor or anti-CCP; JAK inhibitor selectivity; concomitant use of csDMARDs; the number of csDMARDs, bDMARDs, or JAK inhibitors previously used; or prior use of a TNF-α inhibitor, a IL6 inhibitor, abatacept, rituximab, or anakinra.

The retention of JAK inhibitor treatment, due to discontinuation for lack of treatment effectiveness, is depicted in Figure 1.

There were no observed differences in treatment retention rates between patients with high versus moderate disease activity at baseline (p = 0.103). Rheumatoid arthritis disease activity at baseline was not found to impact the survival of the JAK inhibitor treatment effectiveness [hazard ratio (HR) for patients with high disease activity at baseline: 1.58; 95% CI: 0.91–2.74; p = 0.106].

By considering the interindividual variability in the demographic and clinical characteristics of real-world patients, this research assessed both patient- and drug-related factors that may influence the effectiveness of JAK inhibitors in rheumatoid arthritis treatment.

Among factors associated with the effectiveness of JAK inhibitors, baseline rheumatoid arthritis disease activity has been identified. Logistic regression analyses have determined it to be a predictive factor for resistance to achieving remission or low disease activity at the first 6 months of treatment (ORadj for high disease activity: 0.96; 95% CI: 0.92–0.99; p = 0.028). These findings align with those of Hayashi et al. (26), who linked baseline rheumatoid arthritis disease activity to resistance in achieving low disease activity in patients treated with tofacitinib or baricitinib. According to the Cheng-Prusoff equation, high intracellular ATP concentrations—which correspond to a higher degree of inflammation or rheumatoid arthritis disease activity (7–9)—displace competitive inhibitors, such as tofacitinib, baricitinib, upadacitinib, or filgotinib, from binding to their target, thereby influencing the clinical response to treatment (27–29). However, it is important to note that our long-term analysis, assessed through treatment retention, points out that high disease activity does not preclude an effective response to these small molecules in rheumatoid arthritis treatment (HR for patients with high baseline disease activity: 1.58; 95% CI: 0.91–2.74; p = 0.106). Our findings lead us to suggest that tailoring the dosing regimen of JAK inhibitors according to the level of rheumatoid arthritis disease activity might enhance treatment effectiveness. In this context, Takeuchi et al. (30) demonstrated that reducing the dose of baricitinib in patients with sustained disease activity effectively maintains disease control, a finding now included as a recommendation in the drug’s product information (11).

With regard to age, Fleischmann et al. (31) found that the effectiveness of JAK inhibitors was comparable between younger and older rheumatoid arthritis patients. Furthermore, indirect studies have not identified age as a factor associated with the discontinuation of these small molecules due to lack of effectiveness (16, 32). Consistent with these findings, our results suggest that age does not influence the effectiveness of tofacitinib, baricitinib, upadacitinib, or filgotinib treatments in rheumatoid arthritis. It is worth mentioning that the age of our patients ranges between 64 years in subjects who did not achieve the T2T goal at 6 months of JAK inhibitor treatment and 66 years in T2T achievers.

About sex, prior research has shown no significant differences in the effectiveness of JAK inhibitors between female and male patients (16, 32–34). Martinez-Molina et al. (34) highlighted the importance of employing a comprehensive set of validated composite indices to assess rheumatoid arthritis disease activity. This recommendation is particularly relevant given the potential misclassification of disease activity in female patients when using the 28-joint Disease Activity Score (DAS28) with erythrocyte sedimentation rate (ESR), which may be influenced by sex-related differences in ESR levels (34–37). In line with these insights, the present study did not identify female sex as a factor affecting the effectiveness of JAK inhibitors in the treatment of rheumatoid arthritis.

In reference to BMI, the current evidence suggests that BMI does not impact the treatment effectiveness of these tsDMARDs in adult patients with rheumatoid arthritis (38, 39). This observation is consistent with the findings of our study.

Another consideration is the rheumatoid arthritis disease duration. Previous evidence has suggested that it is not linked to JAK inhibitor treatment effectiveness (16, 32, 33). Similarly, our study found no connection between disease duration and the effectiveness of tofacitinib, baricitinib, upadacitinib, or filgotinib treatments.

Concerning rheumatoid arthritis seropositivity, published studies indicate that neither RF nor anti-CCP positivity influences the treatment effectiveness of these small molecules (16, 33, 40). RF and/or anti-CCP positivity has been identified in several studies to be associated with enhanced treatment effectiveness of rituximab, abatacept, and TNF-α inhibitors in patients with rheumatoid arthritis (41–46). The largest study, a registry involving 2,583 patients, identified a positive association between seropositivity and treatment effectiveness to rituximab and abatacept, but found no association with TNF-α inhibitors (45). Consistent with previous evidence, regarding JAK inhibitors, the present study did not find rheumatoid arthritis seropositivity to be associated with treatment effectiveness.

As for JAK inhibitor selectivity, there is currently no conclusive answer due to the lack of head-to-head randomized clinical trials comparing JAK inhibitors (47). However, Traves et al. (48) conducted a study that integrated in vitro and clinical pharmacokinetics to differentiate the inhibition and selectivity of tofacitinib, baricitinib, upadacitinib, and filgotinib within the context of rheumatoid arthritis. These authors concluded that the four tsDMARDs exhibit similar effectiveness, which can be attributed to their comparable inhibition of JAK-1 signaling (48). The evidence supports the notion that in our study, no differences were observed in the effectiveness of these small molecules with respect to JAK-1 selectivity in rheumatoid arthritis treatment.

Other important factors to consider include the number of previous csDMARD treatments and the concomitant use of csDMARDs. Both JAK inhibitors and csDMARDs are used in the treatment of rheumatoid arthritis, each targeting distinct immune pathways through different mechanisms of action. To date, there is a lack of strong evidence directly comparing JAK inhibitor monotherapy (49–54) with the concomitant use of csDMARDs (51, 54–66). The EULAR Task Force recommends maintaining csDMARDs when initiating treatment with a JAK inhibitor (14). According to their guidelines, the csDMARD dose can be reduced to mitigate the risk of adverse events (14). Consistent with our findings, neither the number of previous csDMARD treatments nor the concomitant use of csDMARDs seems to determine the effectiveness of JAK inhibitors in the treatment of rheumatoid arthritis.

In terms of the number of previous bDMARDs treatments, Hayashi et al. (26) identified this factor as a predictor of resistance to achieving low disease activity with baricitinib, but not with tofacitinib. Other studies have not recognized it as a factor associated with the effectiveness of JAK inhibitor treatment (16, 33). The immune dysregulation in rheumatoid arthritis may involve a complex interplay among various cytokines, suggesting that blocking a single cytokine does not necessarily lead to the achievement of the T2T goal. In this context, it may be necessarily to target another specific molecule or even an entire pathway. The findings of the present study indicate that neither the number of previous bDMARD treatments nor the specific mechanism of action of a prior bDMARD significantly impacts the treatment effectiveness of tofacitinib, baricitinib, upadacitinib, or filgotinib.

Regarding the number of previous JAK inhibitor treatments, both the JAK inhibitor cycling strategy and several studies suggest that this factor does not influence the effectiveness of subsequent small molecule treatments (16, 32, 67). The different chemical structure of each JAK inhibitor provides a specific selective inhibition profile toward its target, which likely accounts for the variability in clinical responses observed in real-world practice. As recommended by the EULAR Task Force, if treatment with a JAK inhibitor fails, a change in treatment to another tsDMARD or bDMARD should be considered, given the lack of reliable predictors of treatment response in individual patients (14).

This study has several limitations that may affect the interpretation of the results. Firstly, its observational retrospective nature means that the findings should be viewed as hypothesis-generating and confirmed in prospective studies. Secondly, the population size and the fact that the study was conducted at a single centre should be considered when extrapolating the results to the broader population. Thirdly, our study did not include small molecules other than tofacitinib, baricitinib, upadacitinib, or filgotinib —the JAK inhibitors presently approved in Europe for rheumatoid arthritis treatment— such as peficitinib. Fourthly, adjustments in administration frequency or in dosage of JAK inhibitor treatment were not recorded; it was presupposed that all patients received their required administration frequency and dosage as per the approved guidelines. Fifthly, the imbalance in sex distribution, which mirrors real-world clinical data of a disease that primarily affects females more than males. Sixthly, patient comorbidities were not considered; it was assumed that the JAK inhibitor treatment was prescribed in accordance with the recommendations of the approved guidelines. Nonetheless, the findings are consistent with existing evidence.

The main strength of our study is its identification of factors influencing the effectiveness of JAK inhibitors in rheumatoid arthritis treatment, achieved by examining the interindividual variability in the demographic and clinical characteristics of real-world patients. This is particularly significant as it provides insights that can complement those derived from RCTs.

In summary, the results of our retrospective real-world study suggest that in rheumatoid arthritis, high disease activity at the initiation of treatment with tofacitinib, baricitinib, upadacitinib, or filgotinib does not preclude an effective treatment response but is associated with an increased risk of therapeutic failure. This leads us to suggest that tailoring the dosing regimen of JAK inhibitors according to the level of rheumatoid arthritis disease activity might enhance treatment effectiveness. By contrast, the following factors were not related to JAK inhibitor effectiveness in rheumatoid arthritis treatment: age, female sex, BMI, disease duration, seropositivity, JAK inhibitor selectivity, type and number of prior bDMARDs, concomitant use and number of prior csDMARDs, and number of prior JAK inhibitors. These findings can substantially enhance the insights obtained from randomized controlled trials, offering critical information that can improve healthcare decision-making.