Evaluation of Real-World Evidence to Assess Effectiveness Outcomes of Janus Kinase Inhibitors for Rheumatoid Arthritis: A Systematic Review of US Studies

Introduction

Rheumatoid arthritis (RA) is a chronic, common autoimmune inflammatory arthritis disease affecting multiple joints, potentially leading to joint damage and permanent disability.¹ An estimated 1.3 million people in the US, accounting for 0.6–1.0% are living with RA.² RA is associated with annual direct medical costs up to $12,509.³ The core pharmacotherapy for RA consists of disease-modifying antirheumatic drugs (DMARDs) and involve conventional synthetic DMARDs (csDMARDs), biological DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs).^4,5 For patients with RA, the current guidelines recommend to use csDMARDs as the first-line therapy and then adding or switching to bDMARDs or tsDMARDs following the csDMARDs.^4,5 The tsDMARDs, also known as Janus kinase inhibitors (JAKi), are the newest class of approved oral treatment options for individuals with RA. The therapeutic effect of JAKi is through effects on the JAK-STAT pathway that is responsible for the transduction of cytokines and growth factor signals that play a crucial role in inflammation and autoimmune diseases.⁶ Tofacitinib, baricitinib, and upadacitinib are the three JAKi to receive Food and Drug Administration (FDA) approval for RA treatment in 2012, 2018, and 2019, respectively.⁷ There are key differences in the therapeutic targets of each of available JAKi. Tofacitinib and baricitinib inhibit Janus kinase 1, 2, and 3 non-selectively, while upadacitinib selectively targets the Janus kinase 1, reducing dose-related toxicity and side effects without a significant loss of efficacy.⁸

Clinical improvement with JAKi have been demonstrated in the results from the randomized controlled trials (RCTs) suggesting superior clinical efficacy of novel JAKi treatments as compared to other biologics for RA.^9–11 However, patients enrolled in RCTs are different from those seen in the routine clinical setting.^12–14 In a specific analysis focused on RA patients, a large systematic review of RCTs and real-world studies, patients enrolled in the trials had substantial differences in the characteristics from those seen in the real-world routine clinical setting.¹⁵ Therefore, it potentially suggested that the efficacy outcomes associated with RA treatment demonstrated in the RCTs may not be reflective of real-world effectiveness outcomes. Hence, as more JAKi becomes available, there is a critical need to evaluate the real-world effectiveness of JAKi, and such real-world evidence (RWE) can complement the efficacy data from RCTs.

Furthermore, the FDA is issuing guidance to leverage RWE to support the decision making of the regulatory process, to complement evidence from RCTs.^16,17 The US 21^st Century Cures Act mandates the US FDA to include RWE into the regulatory approval process.¹⁸ In addition, the global regulatory bodies are launching programs to leverage RWE.^19,20 Findings from systematic reviews of RCTs related to JAKi have examined the comparable efficacy of JAKi vs other DMARDs for RA.^21–25 While the majority of systematic reviews compared the clinical response to JAKi against csDMARDs or bDMARDs,^21–24 Hernández-Cruz et al compared patient-reported outcomes and persistence rates between baricitinib and other DMARDs.²⁵ To the best of our knowledge, there are no systematic review based on the real-world effectiveness of JAKi among the RA population. However, given the complementary roles of RWE, there is a need to specifically focus on real-world effectiveness outcomes of JAKi for RA.

To inform evidence-based decision-making in clinical practice, this review aims to synthesize the real-world effectiveness outcomes of JAKi from the real-world observational studies among individuals with RA in the US healthcare settings. This review seeks to provide a better knowledge of evidence on the clinical effectiveness of JAKi in a broader group of individuals with RA from daily routine practice compared to those only seen in the RCTs.

Methods

In accordance with the Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA) guidelines, this systematic review was conducted to summarize the exploratory evidence on the clinical effectiveness of JAKi in patients with RA, in a real-world setting in the US.²⁶ The PRISMA checklist is provided in the Electronic Supplementary Doc (ESD) II. The eTable 1 of the ESD shows the results of the PRISMA reporting checklist.

Data Sources and Literature Searches

This review searched PUBMED, CINAL, and Embase, the three key databases composed of biomedical literature, to identify all relevant publications indexed as of June 2023. The search strategy was designed by the senior author (Y.H.), the clinician (S.B.) and the librarians. The search strategy used both keywords and Medical Subject Heading (MeSH) terms to identify potentially relevant citations identified in all databases. Search terms involved “rheumatoid arthritis”, “Janus kinase inhibitors”, and “real-world databases”. Search queries were crafted for each of these concepts, and results were combined to identify publications of interest (eTable 2 of the Electronic Supplementary Doc [ESD]).

Study Selection and Eligibility Criteria

The focus of this systematic review was to identify US publications that reported the effectiveness data of JAKi for RA patients seen in clinical practice. After the searches, the eligible citations were imported into the COVIDENCE platform, which has been used to manage citations in previous systematic reviews.²⁷ Using the COVIDENCE platform, two authors (C.G. and S.B.) independently conducted the title/abstract screening from literature searches and reviewed the full-text publications for final inclusion based on the eligibility criteria shown in eTable 3 of the ESD. Any conflicts were addressed by the senior author.

Study eligibility criteria were set based on PICO, which stands for Population, Intervention, Comparator(s), and Outcomes. We also considered the setting, as we aimed to only focus on real-world setting. We included studies with adult patients (age ≥18) with a diagnosis of RA as defined by the individual study. Secondly, we included studies that were investigating the use of JAKi including baricitinib, tofacitinib, and upadacitinib regardless of the specific lines of treatment as intervention. JAKi that are not FDA-approved were excluded from consideration, as our focus is on US-based studies. We did not consider any comparator in our search, as we aimed to conduct an exploratory review.

When defining the study eligibility criteria for outcomes, we included all studies that included effectiveness data with RA. Specific outcomes that we collected were: (1) effectiveness patterns; (2) clinical outcomes such as clinical disease activity index (CDAI) and disease activity score using 28 joint counts (DAS28) and patient-reported outcomes including pain, fatigue, morning stiffness, multidimensional health assessment questionnaire | physician global assessment (MDHAQ PGA), multidimensional health assessment questionnaire | patient global assessment (MDHAQ PtGA), multidimensional health assessment questionnaire (MDHAQ functional index), health assessment questionnaire – disability index (HAQ-DI), and routine assessment of patient index data 3 (Rapid3). From effectiveness patterns, we included adherence, persistence, and effectiveness algorithm. Adherence refers to how closely a patient follows the medication regimen recommended by their healthcare provider. It is often represented as the medication possession ratio (MPR) or the percentage of days covered (PDC). Both MPR and PDC are capped to 1, which is equal to complete adherence.²⁸ Persistence typically denotes the time span from the start to the discontinuation of a treatment. It is commonly expressed as the number of consecutive days the therapy is used. Alternatively, persistence can be described as the rate at which patients maintain their treatment over a set period without significant treatment gaps.²⁹ Effectiveness algorithm in RA patients includes following points: (1) high adherence; (2) no bDMARDs or JAKi switch or addition; (3) no csDMARD switch or addition; (4) no increase in dose or frequency of index drug; (5) no more than one glucocorticoid joint injection; (6) no new/increased oral glucocorticoid dose.^30,31

Overall, the inclusion criteria were: (1) analyzed RA patient-level population; (2) evaluated patients with RA exposed to JAKi; (3) used observational healthcare data of routine clinical practice (prospective, retrospective, cross-sectional, longitudinal or survey) which is shown in eTable 3 of the ESD. To capture broad use of JAKi in real-world setting, this review did not limit studies to those also considering another comparator, such as csDMARDs or bDMARDs. In this regard, this review included all empirical observational studies if they reported the effectiveness data with JAKi for RA, regardless of the specific lines of treatment. We excluded animal studies, narrative/systematic reviews, and RCTs. Studies were limited to US settings and those published after 2012. This review also excluded non-English publications.

Data Extraction

Two investigators (S.B. and C.G.) independently conducted data extraction using a Microsoft Excel worksheet. First, we extracted the characteristics of each study: (1) the author/year, (2) publication type, (3) data sources, (4) study design, (5) study sponsor, and (6) the characteristics of study participants, including study inclusion/exclusion, number of patients, female (%), RA duration and major comorbidities.

In addition, we extracted the detailed information of the real-world effectiveness outcomes associated with JAKi. These included (1) treatment patterns such as adherence, persistence, (2) clinical disease activities, and (3) patient reported outcomes. For each outcome, to ease the synthesis of results, the value was recorded and summarized as a range. Because this review summarized a variety of real-world outcomes of jaki in RA including too much data, we decide not to perform a meta-analysis analyzing the comparison between jaki and biologics class. Hence, the systematic review summarizing different real-world outcomes of JAKi, with the meta-analysis is performed.

Quality Assessment and Certainty of Evidence

Two reviewers (C.G and S.B) conducted the quality assessment, using the Microsoft Excel spreadsheet. Only full-text studies were used for the quality assessment. The quality assessment was conducted using the Newcastle-Ottawa Scale (NOS).³² NOS was developed specifically as a quality assessment tool for prospective non-randomized studies. It has been used by prior systematic reviews on efficacy before.

NOS has specific categories for cohort studies. It contains 3 categories, selection of cohort (4 subcategory), comparability of cohort (2 subcategory), and assessment of outcome (3 subcategory). A maximum of one “star” for each item within the “Selection” and “Exposure/Outcome” categories can be marked, while maximum of two ‘stars’ for “Comparability” category can be considered. The complete list of considered criteria with their definition is available in eTable 4 of the ESD.

The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) criteria was used to assess the confidence level of the data.³³ GRADE is a system used to assess the quality of evidence and the strength of recommendations in healthcare and other sectors. It provides a transparent and systematic approach for evaluating the certainty of evidence and translating that into recommendations for clinical practice or policy.

The GRADE system considers factors such as study design, risk of bias, inconsistency, indirectness, imprecision, and publication bias when assessing the quality of evidence. The evidence is then graded as high, moderate, low, or very low quality.

Results

Included Publications

After the searches, 1236 citations were imported into COVIDENCE. A total number of 1115 studies were moved to title/abstract screening, after removing 121 duplicates. 89 full-text studies were then assessed for eligibility, after removing 1026 studies that were not relevant. Upon the review of the full-text publications, 54 studies were excluded, with the following reasons: not based in the US/Not US patients (n=23), RCT/Long Term Extension (LTE) (n=15), JAKi is not the intervention (n=14) or no effectiveness dependent variables (n=2). Overall, a total number of 35 studies were included. The PRISM diagram is shown in Figure 1.

Figure 1 PRISMA Flow Diagram. It shows study selection results. Notes: PRISMA figure adapted from Liberati A, Altman D, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Journal of clinical epidemiology. 2009;62(10). Creative Commons.34

Across these 35 studies, sample sizes ranged from 455–30,556 patients for administrative claims databases, 252–4044 patients for electronic medical record (EMR) databases and 103–8572 patients for patient registry databases (Table 1). Across these studies, a majority used administrative claims datasets (n=23, 65.71%), including claims only, claims plus EMR, or claims plus registry.^35–57 A total of 9 studies used patient registry datasets (25.71%),^58–66 and only 3 studies used electronic medical records (8.57).^67–69 Across these studies, 10 of them evaluated JAKi as a whole (n=28.57%),^{35,37,41,43,57,58,62–65} many only evaluated the JAKi of tofacitinib (n=20, 57.14%),^{38–40,42,44–56,59,66,67} while others only evaluated upadacitinib (n=5, 14.28%).^{36,60,61,68,69}

Table 1 Study Characteristics

Effectiveness Outcomes

Treatment Pattern Related Measures

Table 2 shows the treatment pattern-related measures of RA patients in the US. These included adherence, persistence, and the claims-based effectiveness algorithms. Adherence of JAKi among patients with RA was reported in 11 studies.^{35,37,39–41,48,50–53} These studies all used administrative claims datasets, including Medicare, MarketScan, Symphony Health, Optum, and Truven claims databases. Adherence had a wide range across these studies, with an adherence rate between 0.53–0.83 in terms of PDC.

Table 2 Treatment Pattern Related Measures of RA Patients in the US

Persistence of JAKi in RA patients was reported in 14 studies, of which these studies also all used administrative claims datasets.^{35,37–40,42–45,48,50,53,57} The median time in treatment without discontinuation was 121–516 days with the discontinuation rate for JAKi being between 11% and 72.4%.

In addition, there were six studies that used the effectiveness algorithms^{40,46,47,54–56} which involved a set of specific criteria to measure adherence. The criteria used in those studies were: 1. High adherence; 2. No biologic or tofacitinib switch or addition; 3. No DMARD switch or addition; 4. No increase in dose or frequency of index drug; 5. No more than one glucocorticoid joint injection; 6. No new/increased oral glucocorticoid dose. The percentage of patients being effectively treated ranged from 14.8% to 26%.

Clinical Response Measures

Table 3 shows the clinical response measures related to JAKi use among RA patients in the US. The most common measure of disease activity throughout the studies was Clinical Disease Activity Index (CDAI) with 10 articles reporting it.^{36,39,58,59,61,64–66,68} Of these studies, 9 used registry datasets^{36,39,58,59,61,63–66} and only 1 used electronic medical records (EMR).⁶⁸ The mean change in CDAI after 6 months for JAKi ranged between −4.7 and 5.1.

Table 3 Clinical Response Measures in Patients with RA Treated with JAKi in the US

Other reported outcomes include disease activity score using 28 joint counts (DAS28 (ESR), DAS28 (CRP)), American College of Rheumatology 20/50/70% improvement criteria (ACR 20/50/70), Tender joint count (TJC), and swollen joint count (SJC).^{39,58,59,66–69}

Patient Reported Outcomes

Table 4 shows the patient reported outcome measures related to JAKi use among RA patients in the US. There were 12 studies that measured the patient reported outcomes among patients with RA using JAKi.^{36,39,58,59,61–66,68,69} A total of 10 studies utilized registry datasets^{36,39,58,59,61–66} and 2 used EMR.^68,69 The most widely reported patient reported outcome was pain, with all 12 of the studies including it in their findings.^{36,39,58,59,61–66,68,69} The mean change in pain ranged from −9.3 to 8.9 across these studies.

Table 4 Patient Reported Outcomes in Patients with RA Treated with JAKi in the US

Other PRO outcomes included fatigue, morning stiffness, multidimensional health assessment questionnaire | physician global assessment (MDHAQ PGA), multidimensional health assessment questionnaire | patient global assessment (MDHAQ PtGA), multidimensional health assessment questionnaire (MDHAQ functional index), health assessment questionnaire – disability index (HAQ-DI), and routine assessment of patient index data 3 (Rapid3).

Risk of Bias of Included Studies

Six articles got a total of 8 stars out of 9 stars, while only two articles got 6 stars. Overall, the articles can be considered high quality. The eTable 5 of the ESD shows the results of the quality assessment across all full-text studies.

Discussions

Several recent RCTs have proved the efficacy and safety of JAKi agents including tofacitinib, baricitinib, and upadacitinib for treatment of RA by showing their long-term treatment benefit in terms of decrease in disease activity and achieving clinical remission in RA over a follow-up period of 3–6 months.^{9–11,70–72} The present evidence from real-world observational studies adds supplementary information of treatment effectiveness in a broader range of patient populations with diverse demographics and clinical characteristics profiles, which are not representative in the RCTs. In this present systematic review, we included adherence/persistence/effectiveness, clinical response outcomes, and patient reported outcomes data from 35 observational studies, further demonstrating the real-world effectiveness of JAKi agents for the treatment of individuals with RA seen in daily routine clinical practice. These studies utilized three main data sources: administrative claims, EMRs, registries. Administrative claims are data collected by insurance companies for billing purposes, including information on diagnoses, treatments, and services. While useful for assessing adherence, persistence and effectiveness, they often lack detailed clinical information for clinical outcomes and PROs. EMRs are digital records of patients’ clinical data gathered during visits. They provide comprehensive information about patient history, diagnoses, medications, and lab results, making them valuable for assessing clinical outcomes and PROs. Registries systematically collect data on individuals with specific conditions over time, tracking outcomes and treatment effects. They offer insights into the long-term effectiveness and safety of treatments, making them suitable for evaluating clinical outcomes and PROs. Treatment with JAKi agents is associated with good treatment effectiveness, with 12-month adherence rates up to 56%, and treatment persistence rates up to 42%, respectively. Overall, the clinical outcome measured by the CDAI were also good, with mean change in terms of CDAI ranging from −4.7–5.1.

This systematic review adds to the existing body of reviews indicating the superior efficacy of JAKi in RA and other rheumatology disease areas.^7,22,73–76 In Wang et al, a systematic review of 20 RCTs, an improvement was shown for ACR20 for all JAKi including tofacitinib, baricitinib, and upadacitinib (RR, 2.03 [1.87 to 2.20]) and Health Assessment Questionnaire-Disability Index scores (mean differences, −0.31 [−0.34 to −0.28]) over treatment with placebo.⁷³ In another systematic review, He et al, they showed that there was a significant proportion of patients with RA reaching the ACR 20 responses for the tofacitinib treatment group.⁷⁴ Another two reviews and meta-analyses of RCTs by Liu et al and Wang et al also evaluated the achievement of Health Assessment Questionnaire disability index (HAQ-DI) and remission outcomes for JAKi in treatment of RA.^22,75 Our systematic review further confirms the real-world clinical effectiveness of JAKi in patients with RA across treatment adherence/persistence, improvement in disease activity scores, and PROs.

Of note, the definition of real-world treatment related outcomes varied among the administrative claims studies included in this systematic review. For example, the most common measurement of adherence was PDC greater than 80%, and others used MPR greater than 80%. Regardless of the differences in definitions of adherence measures, the treatment with JAKi yielded high adherence rates throughout the timepoints from 6 months to 12 months among individuals with RA. In addition, studies in this analysis reported long persistence with JAKi in patients with RA, with median time in treatment without discontinuation ranging from 121 to 516 days. There are differences of the definition of persistence. The primary outcome definition for persistence was time to discontinuation of the drug, with some studies wording this differently (survival time, time on therapy, duration of treatment). Although some studies defined persistence differently, they were all measuring the same outcome. Another outcome measured for persistence was gaps greater than a specified number of days.

Strengths of this systematic review included a comprehensive evaluation of real-world effectiveness outcomes of JAKi in RA. These novel outcomes associated with JAKi for RA based on the real-world studies are particularly meaningful and provide insights on the real-world clinical performance of JAKi for RA population, which complements RCT based outcomes. RCTs usually require patients to have full compliance and may overestimate its efficacy. Findings on real world-based effectiveness outcomes of JAKi help clinicians to optimize the use of JAKi for RA care.

Limitations of this systematic review should be noted as follows. Due to the observed differences in design and patient population, there are many heterogeneities across studies; therefore, the results in the present review are only based on description of real-world effectiveness outcomes for JAKi in RA. Second, given that we aimed to conduct an exploratory review of JAKi effectiveness in RA, we did not consider any comparator requirement in our search. Future systematic reviews could also examine the comparative effectiveness of JAKi versus other DMARDs using observational data, as more relevant studies become available. Finally, because the medication approval process differs in the US compared to the rest of the world, we focused on US-based studies evaluating JAKi approved by the FDA for RA, specifically tofacitinib, upadacitinib, and baricitinib. Other JAKi, such as filgotinib, which are not FDA-approved for RA treatment, were excluded from this review. Future, larger studies encompassing patients from other regions and including additional JAKi will further provide insights about the real-world effectiveness of JAKi among the RA population.

Conclusions

This systematic review confirms the effectiveness of JAKi but is limited in that most of the studies included in the analysis are claims-based data and limited individual patient information is available. Nevertheless, this review describes the ever-growing number of studies evaluating the clinical utilization of these therapies. While the efficacy of JAKi is demonstrated in the RCTs, the observational studies show the real-world effectiveness of JAKi for individuals with RA in clinical practice through evaluation of a variety of real-world outcomes, including treatment adherence/persistence/effectiveness, clinical disease activity changes, or PROs. This systematic literature review showed notable adherence and persistence rates in individuals with RA, proving the clinical effectiveness of JAKi in real-world clinical practice. Real-world treatment benefit with JAKi was observed across all included studies despite the variation in design, patient populations, and real-world outcome definitions. Findings in this review complement the efficacy data observed in RCT populations and confirm the treatment benefit of JAKi for treatment in RA in a broader group of patients other than the RCT setting. Future reviews focusing on the comparative effectiveness studies between JAKi and other biologics would help to further inform clinical practice.