Abstract
Objective
To compare the effectiveness of JAK inhibitors (JAKis) and biologic disease‐modifying antirheumatic drugs (bDMARDs) on pain in patients with rheumatoid arthritis.
Methods
In this retrospective study, we investigated patients with a diagnosis of rheumatoid arthritis, starting treatment with a JAKi (n = 1,827), a tumor necrosis factor inhibitor (TNFi; n = 6,422), an interleukin‐6 inhibitor (n = 887), abatacept (n = 1,102), or rituximab (n = 1,149) in 2017 to 2019, using data from several linked Swedish national registers. Differences in change in pain, assessed with a visual analog scale (0–100 mm), from baseline to 3 months, as well as proportions of patients remaining on initial treatment with low pain (visual analog scale pain <20) at 12 months, were compared between treatments. Comparisons of treatment responses between JAKis and bDMARDs were evaluated using multivariable linear regression, adjusted for patient characteristics, comorbidities, current comedication, and previous treatment.
Results
JAKi treatment was associated with a greater decrease in pain at 3 months compared with TNFi treatment (adjusted mean additional decrease 4.0 mm; 95% confidence interval 1.6–6.3), with similar trends in comparisons with non‐TNFi bDMARDs. More patients achieved low pain at 12 months on JAKis compared with TNFis, in particular among those previously treated with at least two bDMARDs (adjusted change contrast 5.3 percentage points; 95% confidence interval 1.0–9.6).
Conclusion
JAKis had a slightly better effect on pain outcomes at 3 and 12 months compared with TNFis, with significantly greater differences in patients previously treated with at least two bDMARDs. The effect of JAKis on pain reduction was at least similar to that of non‐TNFi bDMARDs.
INTRODUCTION
JAK inhibitors (JAKis) have been used for treating rheumatoid arthritis (RA) in the European Union since 2017. Several randomized clinical trials (RCTs) have compared the efficacy of biologic disease‐modifying antirheumatic drugs (bDMARDs) and targeted synthetic (ts)DMARDs (ie, JAKis) as add‐on treatments after inadequate response to methotrexate (MTX). Better or equivalent effect on disease activity was found for the JAKis baricitinib and upadacitinib compared with the tumor necrosis factor inhibitor (TNFi) adalimumab, 1 , 2 , 3 whereas tofacitinib and filgotinib were reported as noninferior to adalimumab. 4 , 5 In particular, it has been reported that JAKis may have a special impact on pain. 6 Specifically, in the RA‐BEAM and SELECT‐COMPARE trials, treatment with baricitinib and upadacitinib, respectively, resulted in greater pain reduction compared with adalimumab. 2 , 7 In a systematic review and meta‐analysis of clinical trials, the authors concluded that pain reduction was significantly greater for JAKis compared with bDMARDs. 8 Furthermore, baricitinib has been associated with a greater pain improvement compared with adalimumab also in a subanalysis of patients with well‐controlled RA. 9 This suggests that JAKis may have direct analgesic effects beyond its antiinflammatory properties.
In a previous Swedish study, one‐third of patients with RA had unacceptably high pain levels (visual analog scale [VAS] pain >40) up to 2 years after diagnosis. 10 Pain has also been shown to persist in patients with well‐controlled disease. Several studies have supported that noninflammatory pain, presumably because of central sensitization, is a cause for remaining pain in RA. 11 , 12 , 13 Furthermore, pain relief has been identified as the most important goal for a substantial proportion of patients. 14 , 15 This taken together points to a need for improvement in pain management in RA.
So far, only limited results are available from observational studies comparing the effectiveness of JAKis and bDMARDs on pain reduction in clinical practice, especially when it comes to non‐TNFi bDMARDs. In one large observational cohort study from the United States, no difference in mean pain at 6 months were observed between patients initiating treatment with third/fourth line of TNFi combination therapy and third/fourth line of tofacitinib combination therapy. 16 However, the reduction in pain from baseline was not evaluated. Furthermore, although the study 16 used propensity score matching to ensure comparable groups, it did not consider differences in patient characteristics such as comorbidities and previous conventional treatments. The aim of this study was therefore to compare the effect of JAKis and each class of bDMARDs on pain in patients with RA seen in clinical practice, treated according to standard of care.
MATERIALS AND METHODS
Patients
The present cohort study, using Swedish national register data, is similar in design (but not in outcome) to a previous report. 17 Patients with a diagnosis of RA, starting treatment in 2017 to 2019 with a TNFi, rituximab, abatacept, interleukin‐6 inhibitor (IL‐6i), or a JAKi, regardless of previous treatment, with data in the Swedish Rheumatology Quality Register (SRQ), were included in this study. The SRQ is a national clinical register of patients with rheumatic disorders, where data have been prospectively registered by the treating physician at visits in standard clinical care. The register has an estimated national coverage of 95% for patients with RA treated with bDMARD. 18 The study was approved by the Ethical Review Board in Stockholm (DNR: 2016/1986‐32).
Data sources and data assessment
Data on patient characteristics, disease activity measures, and patient‐reported outcomes (PROs) were retrieved from the SRQ at baseline, and at 3 and 12 months following treatment start, with time‐windows defined as 90 days before to 30 days after treatment start for baseline measurements, 60 days to 183 days after treatment start for 3‐month measurements, and 275 to 455 days after treatment start for the 12‐month measurements, consistent with those used previously. 17 If there were more than one visit in each window, the measurement closest to the selected follow‐up time point (0, 3, and 12 months) was used. Data retrieved from the SRQ included the patient's assessment of pain using a VAS (0–100 mm), Health Assessment Questionnaire Disability Index (HAQ‐DI), Disease Activity Score in 28 joints (DAS28) measured with erythrocyte sedimentation rate, Clinical Disease Activity Index (CDAI), and serologic status (rheumatoid factor [RF] and anticitrullinated peptide antibodies).
Dates of initiation and discontinuation of bDMARD or JAKi therapy were also extracted from the SRQ. One patient could participate with more than one treatment episode, including several episodes of the same drug and class. If a patient restarted the same drug within 90 days (or 270 days for rituximab) of a previous stop, the two episodes were counted as one. Data on cotreatment with conventional synthetic (cs)DMARDs, glucocorticosteroids, and nonsteroidal anti‐inflammatory drugs (NSAIDs), as well as previous csDMARD treatment and other prescribed drugs were retrieved from the SRQ and/or from the Prescribed Drug Register— a national register covering dispensations from community pharmacies. Cotreatment with a csDMARD was defined as at least one prescription from 180 days before to 30 days after treatment start of the b/tsDMARD.
History of joint surgeries and days of hospitalization were retrieved from the National Patient Register, covering all inpatient care and a vast majority of specialized outpatient care since 2001. Data on demographics were obtained from registers at Statistics Sweden. Comorbidity data were retrieved from the National Patient Register, the Swedish Cancer Register, and/or from the Prescribed Drug Register, and was defined based on International Classification of Diseases tenth revision, and/or Anatomic Theraputic Chemical (ATC) classification system‐codes.
Outcomes
Pain responses were evaluated at 3 and 12 months from treatment initiation. Drug survival was assessed throughout the follow‐up. Start of a new bDMARD or tsDMARD was counted as discontinuation, if earlier than a recorded stop. A switch between originator and biosimilar was not considered discontinuation. Patients discontinuing therapy because of remission were considered on therapy until the start of another b/tsDMARD. Patients were excluded in the case of death, emigration from Sweden, or discontinuation because of pregnancy.
Treatment response at 3 months
The change in pain from baseline was evaluated at 3 months, assuming that the proportion who discontinued therapy because of lack of effectiveness before the follow‐up visit was low. Treatment episodes that were discontinued before the defined window for the evaluation visit were excluded.
Treatment response at 12 months
The response measure at 12 months was the proportion of patients attaining low pain, defined as VAS pain less than 20 mm, in accordance with the cutoff used in definitions of minimal disease activity in RA. 19 , 20 However, because excluding patients who discontinued therapy before the evaluation visit could lead to a selection bias, the response measure at 12 months was defined as the proportion attaining low pain and remaining on therapy (LUNDEX corrected response). 21
Additional analyses
In addition, treatment responses at 3 months and at 12 months were evaluated in patients stratified by monotherapy or combination therapy with csDMARDs, and by line of treatment with b/tsDMARDs (first line, second line, and third line or later, respectively). In exploratory analyses, we calculated the proportions of patients in each treatment group that, at 12 months, were in remission, low disease activity, and moderate/high disease activity, based on their CDAI (ie, CDAI ≤2.8, CDAI >2.8 and ≤10 and CDAI >10, respectively), among those with 12‐month VAS pain less than 20 mm.
Covariates
Confounders accounted for in the models were selected as in previous studies, taking into account factors that have been described to influence treatment in Swedish patients with RA. 22 The baseline covariates included demographics (ie, age, sex, health care region, year of treatment start, smoking, education level, and country of origin), disease history (days of hospitalization during the last 5 years, number of drug classes prescribed during the last year, and comorbidities), previous and current RA treatment (line of treatment with b/tsDMARD, cotreatment with csDMARDs, glucocorticosteroids, or NSAIDs, previous treatment with csDMARDs and b/tsDMARDs, and average daily dose of glucocorticosteroid use), and RA characteristics (symptom duration, RF, joint surgeries, baseline DAS28, CDAI, HAQ‐DI, and VAS pain). All baseline covariate definitions can be found in Supplementary Table 1. Among the comorbidities that were adjusted for, we included disorders that may have a special impact on pain (ie, pain‐related and psychiatric comorbidities) (Supplementary Table 2).
Statistics
Comparisons between treatment arms for differences in mean change in pain at 3 months and in proportions attaining low pain at 12 months were evaluated with crude and adjusted linear regression models with Huber‐White standard errors. All models were adjusted for potential baseline confounders. Multiple imputation with the fully conditional specification method 23 was used to account for missing data for covariates and pain at baseline, and for the outcome measures at 3 and 12 months, with 25 imputed datasets, to ensure that results were not biased by nonrandom missingness linked to measured variables. Imputation models included all variables included in the analysis models. Nonresponder imputation for the 12‐month outcome was applied after multiple imputation. Treatment outcome at 3 and 12 months and corresponding standard errors were calculated from each imputed data set and pooled using Rubin's rules. Separate imputation models were used for analyses stratified by monotherapy or combination therapy, and by line of treatment. In supplementary analyses, outcomes among patients on drug at 12 months were evaluated (without LUNDEX corrected response), and separate analyses were performed, including only complete cases (without multiple imputation for missing data).
The individual patient data comes entirely from national Swedish registers. According to Swedish law, access to national register data is granted on a restrictive basis and may not be shared without additional specific permissions from the Swedish register‐holding authorities.
RESULTS
Patients
A total of 8,430 patients were included in this study, with a total of 11,387 treatment episodes (1,102 with abatacept, 887 with an IL‐6i, 1,827 with a JAKi [82% baricitinib], 1,149 with rituximab, and 6,422 with a TNFi) (Table 1). The distribution of treatment starts with different IL‐6is, JAKis and TNFis are listed in Supplementary Table 3. TNFis and rituximab were more frequently used as first or second line of b/tsDMARD treatment, whereas abatacept, IL‐6is, and JAKis were primarily used as third/fourth line of treatment. Demographics were similar between treatment groups. Disease duration and disease activity/severity measures at baseline (ie, DAS28, HAQ‐DI, CDAI, and pain) were slightly lower in patients starting treatment with a TNFi compared with patients starting treatment with any other b/tsDMARD, and combination therapy with a csDMARD was somewhat more frequent among those starting TNFi. The proportion of patients initiating treatment as monotherapy but who started csDMARD treatment within 12 months were as follows: 13% for abatacept, 8% for IL‐6is, 9% for JAKis, 21% for rituximab, and 16% for TNFis. Conversely, the proportion of patients starting treatment at baseline as combination therapy but who discontinued csDMARD treatment during the 12‐month follow‐up were 21% for abatacept, 35% for IL‐6is, 21% for JAKis, 16% for rituximab, and 20% for TNFis. Pain values were missing for about 30% of the patients at baseline and for 50% to 60% at 12 months (amount of missing data are shown in Supplementary Table 4). Sixty‐seven percent of the patients initiating a JAKi were still on this treatment at 12 months from baseline; for other treatments, this proportion ranged from 58% for IL‐6is to 80% for rituximab (Supplementary Table 5).
Table 1.
Patient characteristics at baseline between treatment arms*
| Variable | Abatacept | IL‐6 inhibitor | JAK inhibitor | Rituximab | TNF inhibitor |
|---|---|---|---|---|---|
| n | 1,102 | 887 | 1,827 | 1,149 | 6,422 |
| Sex, female, n (%) | 873 (79.2) | 722 (81.4) | 1,496 (81.9) | 869 (75.6) | 5,013 (78.1) |
| Age | 63 (53–72) | 59 (49–70) | 61 (52–71) | 64 (54–73) | 59 (47–69) |
| RA characteristics | |||||
| Symptom duration, years | 12.8 (5.4–22.1) | 10.3 (4.7–18.5) | 13.2 (6.8–22.7) | 12.8 (6.2–22.3) | 8.1 (2.9–16.3) |
| RF seropositivity, n (%) | 835 (77.9) | 649 (74.4) | 1,353 (75.6) | 968 (86.0) | 4,429 (70.3) |
| MTX, n (%) | 546 (49.5) | 401 (45.2) | 792 (43.3) | 585 (50.9) | 4,092 (63.7) |
| Non‐MTX csDMARD n (%) | 220 (20.0) | 152 (17.1) | 356 (19.5) | 277 (24.1) | 1,572 (24.5) |
| Glucocorticoids, n (%) | 775 (70.3) | 631 (71.1) | 1,282 (70.2) | 879 (76.5) | 4,104 (63.9) |
| NSAID, n (%) | 204 (24.6) | 193 (26.9) | 447 (30.9) | 213 (24.3) | 1,037 (20.6) |
| Joint surgeries, n (%) | 187 (17.0) | 135 (15.2) | 344 (18.8) | 191 (16.6) | 677 (10.5) |
| Treatment line | 3 (2–4) | 3 (2–5) | 4 (2–5) | 2 (1–4) | 1 (1–2) |
| DAS28 | 4.8 (3.9–5.6) | 4.9 (4.0–5.7) | 4.7 (3.8–5.6) | 4.8 (3.9–5.6) | 4.4 (3.4–5.2) |
| SJC | 4 (2–7) | 4 (2–7) | 4 (2–7) | 4 (2–7) | 3 (1–6) |
| TJC | 5 (2–9) | 6 (3–10) | 5 (2–10) | 5 (2–9) | 4 (2–8) |
| CDAI | 21 (15–28) | 22 (15–29.5) | 20 (14–28) | 21 (14–28.5) | 18 (12–25) |
| HAQ‐DI | 1.1 (0.8–1.6) | 1.1 (0.8–1.6) | 1.3 (0.8–1.6) | 1.1 (0.8–1.6) | 0.9 (0.5–1.4) |
| VAS pain | 63 (42–77) | 65 (44–80) | 61 (40–76) | 60.5 (38–76) | 56 (34–73) |
| General health | |||||
| Never smoker, n (%) | 325 (37.3) | 313 (42.6) | 601 (39.8) | 334 (35.6) | 2,041 (42.9) |
| Previous smoker, n (%) | 458 (52.6) | 339 (46.1) | 736 (48.8) | 488 (52.0) | 2,141 (45.0) |
| Current smoker, n (%) | 88 (10.1) | 83 (11.3) | 172 (11.4) | 117 (12.5) | 572 (12.0) |
| Days in hospital | 3 (0–11) | 1 (0–7) | 2 (0–8) | 3 (0–10) | 0 (0–5) |
| Education level, years, n (%) | |||||
| ≤9 | 248 (22.6) | 165 (18.8) | 363 (19.9) | 251 (22.1) | 1,173 (18.4) |
| 10–12 | 507 (46.2) | 447 (51.0) | 885 (48.5) | 548 (48.2) | 2,977 (46.7) |
| >12 | 342 (31.2) | 265 (30.2) | 575 (31.5) | 337 (29.7) | 2,219 (34.8) |
| Country of birth, n (%) | |||||
| Swedish | 953 (86.5) | 754 (85.0) | 1,573 (86.1) | 954 (83.0) | 5,474 (85.2) |
| Scandinavian, non‐Swedish | 60 (5.4) | 48 (5.4) | 82 (4.5) | 67 (5.8) | 282 (4.4) |
| Other | 89 (8.1) | 85 (9.6) | 172 (9.4) | 128 (11.1) | 666 (10.4) |
| Comorbidities, n (%) | |||||
| Cancer | 105 (9.5) | 53 (6.0) | 134 (7.3) | 192 (16.7) | 396 (6.2) |
| Acute coronary syndrome | 35 (3.2) | 12 (1.4) | 41 (2.2) | 30 (2.6) | 92 (1.4) |
| Congestive heart failure | 43 (3.9) | 14 (1.6) | 38 (2.1) | 40 (3.5) | 73 (1.1) |
| Stroke | 39 (3.5) | 20 (2.3) | 33 (1.8) | 27 (2.3) | 118 (1.8) |
| Diabetes | 145 (13.2) | 87 (9.8) | 187 (10.2) | 130 (11.3) | 585 (9.1) |
| Liver disease | 17 (1.5) | 14 (1.6) | 22 (1.2) | 23 (2.0) | 67 (1.0) |
| Kidney disease | 24 (2.2) | 16 (1.8) | 28 (1.5) | 23 (2.0) | 82 (1.3) |
| Obstructive pulmonary disease | 83 (7.5) | 33 (3.7) | 91 (5.0) | 67 (5.8) | 186 (2.9) |
| Interstitial lung disease | 52 (4.7) | 21 (2.4) | 40 (2.2) | 77 (6.7) | 44 (0.7) |
| Severe infection | 103 (9.3) | 23 (2.6) | 95 (5.2) | 73 (6.4) | 178 (2.8) |
| Herpes | 49 (4.4) | 37 (4.2) | 64 (3.5) | 46 (4.0) | 203 (3.2) |
| Venous thromboembolism | 35 (3.2) | 21 (2.4) | 58 (3.2) | 38 (3.3) | 102 (1.6) |
| Pain‐related comorbidities | 522 (47.4) | 418 (47.1) | 892 (48.8) | 520 (45.3) | 2,508 (39.1) |
| Psychiatric comorbidities | 334 (30.3) | 250 (28.2) | 550 (30.1) | 365 (31.8) | 1,683 (26.2) |
Values are median (interquartile range) unless otherwise indicated. Only observed values are shown (not imputed values). CDAI, clinical disease activity index; csDMARD, conventional synthetic disease‐modifying antirheumatic drug; DAS28, disease activity score in 28 joints; HAQ‐DI, Health Assessment Questionnaire Disability Index; IL‐6, interleukin‐6; MTX, methotrexate; NSAID, nonsteroidal antiinflammatory drug; RA, rheumatoid arthritis; RF, rheumatoid factor; SJC, swollen joint count; TJC, tender joint count; TNF, tumor necrosis factor; VAS, visual analog scale.
Crude and adjusted differences in Δ pain between baseline and 3 months
Overall, pain scores improved from baseline to 3 months in all treatment arms—with mean changes ranging from −20.1 mm (95% confidence interval [95% CI] –23.1 to –17.2) for IL‐6is to –16.6 mm (95% CI –19.1 to –14.0) for rituximab (Table 2). In patients treated with a JAKi, the mean change in pain at 3 months was –18.3 mm (95% CI –20.2 to –16.5). The distribution of changes in pain per treatment arm is presented in Supplementary Figure 1. In all groups, pain was reduced for most patients, but a small subset experienced increased pain. In analysis of differences in change in pain between JAKis and bDMARDs adjusted for demographics, RA characteristics and for current comedication and previous b/tsDMARD treatment, JAKis decreased pain scores by 4.0 mm (95% CI 1.7–6.3) more than TNFis, and by 3.9 mm (95% CI 0.9–6.9) more than rituximab, at 3 months (Table 2). Further adjustment for comorbidities did not have a major impact on the results (Figure 1). There was no statistically significant difference in change in VAS pain at 3 months when comparing JAKis with abatacept or IL‐6is (Table 2, Figure 1).
Table 2.
Differences in mean change in pain from baseline to 3‐months follow‐up*
| Treatment initiation | n | Baseline mean | Mean change | Crude change contrast | Adj. change contrast |
|---|---|---|---|---|---|
| Abatacept | 1,040 | 58.7 (56.7–60.7) | –18.3 (–20.5 to –16.1) | 0.0 (–2.9 to 3.0) | 2.2 (–0.5 to 4.8) |
| IL‐6 inhibitor | 821 | 60.8 (58.9–62.8) | –20.1 (–23.1 to –17.2) | –1.8 (–5.4 to 1.8) | 1.2 (–2.1 to 4.5) |
| Rituximab | 1,101 | 56.4 (54.6–58.1) | –16.6 (–19.1 to –14.0) | 1.8 (–1.4 to 4.9) | 3.9 (0.9–6.9) |
| TNF inhibitor | 6,118 | 53.2 (52.4–54.0) | –18.1 (–19.2 to –17.1) | 0.20 (–2.1 to 2.5) | 4.0 (1.7–6.3) |
| JAK inhibitor | 1,699 | 57.6 (56.2–59.0) | –18.3 (–20.2 to –16.5) | Ref | Ref |
Values are estimates (95% confidence interval) unless otherwise indicated. Univariate and multivariate linear regression with JAK inhibitor as reference. Adj. for demographics, RA characteristics and previous and current RA treatment. Adj., adjusted; IL‐6, interleukin‐6; RA, rheumatoid arthritis; Ref, reference; TNF, tumor necrosis factor.
Figure 1.
Fully adjusted differences in mean change in pain from baseline to 3‐months follow‐up. Multivariate linear regression with JAK inhibitor as reference. Adj. for demographics, RA characteristics, current and previous RA treatment, and further for comorbidities. Adj., adjusted; IL‐6, interleukin‐6; RA, rheumatoid arthritis; TNF, tumor necrosis factor.
The proportion of patients who discontinued treatment before the 3‐month time frame, and therefore were excluded from this analysis, were 6% for abatacept, 7% for IL‐6is, 4% for rituximab, 5% for TNFis, and 7% for JAKis. The mean change in pain from baseline to 3 months was missing for 57% of patients treated with a JAKi. For bDMARDs, this ranged from 56% for IL‐6is to 67% for rituximab (Supplementary Table 4).
Proportions remaining on drug and in low pain at 12 months
Twenty percent of patients treated with a JAKi remained on drug and reached low pain at 12 months from baseline (the corresponding proportions for the bDMARDs ranged from 17% for IL‐6is to 26% for rituximab) (Table 3). Adjusted proportions remaining on drug and with low pain were numerically higher for patients treated with a JAKi compared with patients treated with a TNFi or an IL‐6i, although the differences did not reach statistical significance (Figure 2). Among those with low pain at 12 months, the proportion in remission was lower and the fraction with active disease higher in patients treated with JAKi compared with the TNFi group (Supplementary Table 6).
Table 3.
Differences in proportions attaining low pain score at 12 months*
| Treatment initiation | n | Crude proportion | Crude contrast | Adj. contrast |
|---|---|---|---|---|
| Abatacept | 1,102 | 20.3 (17.0–23.6) | 0.8 (–3.2 to 4.8) | –0.3 (–4.4 to 3.8) |
| IL‐6 inhibitor | 887 | 17.0 (13.8–20.1) | –2.6 (–6.8 to 1.7) | –3.3 (–7.4 to 0.8) |
| Rituximab | 1,149 | 26.3 (21.9–30.6) | 6.7 (1.8–11.7) | 3.7 (–1.7 to 8.1) |
| TNF inhibitor | 6,422 | 25.7 (24.1–27.3) | 6.2 (2.7–9.7) | –3.4 (–7.2 to 0.3) |
| JAK inhibitor | 1,827 | 19.5 (16.7–22.3) | Ref | Ref |
Values are estimates (95% confidence interval) unless otherwise indicated. Univariate and multivariate linear regression with JAK inhibitor as reference. Adj. for demographics, RA characteristics and previous and current RA treatment. Low pain: visual analog scale pain <20. Adj., adjusted; IL‐6, interleukin‐6; RA, rheumatoid arthritis; TNF, tumor necrosis factor.
Figure 2.
Fully adjusted differences in proportions attaining low pain score at 12 months. Multivariate linear regression with JAK inhibitor as reference. Adj. for demographics, RA characteristics, current and previous RA treatment, and further for comorbidities. Low pain: visual analog scale pain <20. Adj., adjusted; IL‐6, interleukin‐6; RA, rheumatoid arthritis; TNF, tumor necrosis factor.
VAS pain at 12 months was missing for 62% of patients treated with a JAKi. For bDMARDs this proportion ranged from 54% for abatacept to 62% for TNFis (Supplementary Table 4).
Pain responses by line of treatment with b/tsDMARD
Analyses restricted to individuals who started their first ever b/tsDMARD were affected by the small sample size as only a limited number of patients were treated with a JAKi as first b/tsDMARD treatment. In this analysis, the reduction in pain at 3 months was significantly greater for IL‐6is compared with JAKis (Figure 3). By contrast, in the analysis restricted to those starting their second ever b/tsDMARD, JAKis decreased pain significantly more at 3 months after full adjustment compared with both TNFis and rituximab. The same was found when comparing JAKis with TNFis among patients starting their third or more b/tsDMARD (Figure 3).
Figure 3.
Differences in mean change in pain from baseline to 3 months, stratified by line of treatment. Univariate and multivariate linear regression with JAK inhibitor as reference. Adjustment for demographics, RA characteristics, current and previous RA treatment, and further for comorbidities. Adj., adjusted; b/tsDMARD, biologic/targeted synthetic disease‐modifying antirheumatic drug; IL‐6, interleukin‐6; RA, rheumatoid arthritis; Ref, reference; TNF, tumor necrosis factor.
In the analysis of pain response at 12 months (attaining VAS pain <20) restricted to first line of treatment, the results were pointing toward higher proportions with low pain for bDMARDs compared with JAKis, especially for IL‐6is and abatacept in which differences were statistically significant (Supplementary Figure 2). In the analysis restricted to second line of treatment, there were no significant differences in fractions reaching low pain at 12 months between treatments. However, in patients treated with a JAKi as third line of b/tsDMARD or later, proportions with low pain at 12 months were significantly higher compared with TNFis and IL‐6is, with a similar trend in the comparison with abatacept (Supplementary Figure 2).
Pain responses by monotherapy or combination therapy with csDMARDs
When stratifying the b/tsDMARD initiator cohorts by monotherapy versus combination therapy with csDMARDs, differences in mean change in pain at 3 months between JAKis and bDMARDs were similar to the main analysis, but were in general more prominent when comparing monotherapies, with adjusted mean differences for JAKi monotherapy of 5.4 mm (95% CI 2.9–8.8) compared with TNFi monotherapy, and 9.7 mm (95% CI 4.1–15.3) compared with rituximab monotherapy. In those receiving combination therapy, only the difference between JAKis and TNFis remained statistically significant (Supplementary Figure 3).
Pain response at 12 months stratified by monotherapy or combination therapy were also similar to the main analysis, but fully adjusted differences were again more pronounced in the analysis restricted to patients on monotherapy, with a 5.5% (95% CI 0.1–10.8) higher response rate for JAKi monotherapy compared with TNFi monotherapy. Statistically significant differences in proportions reaching low pain were also found for JAKi monotherapy compared with IL‐6i monotherapy, but not to rituximab or abatacept monotherapy (Supplementary Figure 4). There were no statistically significant differences between JAKi combination therapy and bDMARD combination therapy in proportions with low pain at 12 months.
Supplementary analyses
Restriction to patients on drug at 12 months
When restricting the analysis to patients still on drug at 12 months, crude proportions for low pain at 12 months increased for all treatment groups, which could be anticipated as the nonresponder imputation was not applied. In this analysis, there were no differences in treatment responses at 12 months between JAKis and bDMARDs (Supplementary Figure 5).
Complete case analysis
In the complete case analysis, the difference in the reduction of pain at 3 months were similar to the main analysis, although slightly more pronounced (Supplementary Figure 6). Pain response at 12 months, including only complete cases, also supported the main analysis results (Supplementary Figure 7).
DISCUSSION
In this observational retrospective study, comparing the effectiveness of JAKis and bDMARDs while taking patient characteristics and treatment history into account, JAKis had a significantly better effect on the 3‐month pain outcome. Although there were major improvements in pain with all treatments, only a minority of patients achieved a pain score of less than 20 at 12 months. The effect of JAKis was numerically better also regarding the pain outcome at 12 months compared with TNFis. The superior effect of JAKis was more pronounced for monotherapy and when used as third line of treatment with b/tsDMARD or later. Moreover, the effect of JAKis on pain was at least similar compared with non‐TNFi bDMARDs.
Superior effectiveness of JAKis on pain compared with TNFis has also been found in clinical trials. In the RA‐BEAM study, significantly greater improvements in pain were seen for patients treated with baricitinib compared with adalimumab, already at weeks 2 to 4. 1 In the SELECT‐COMPARE trial, upadacitinib was superior to adalimumab regarding pain improvement from baseline to week 12, and also at the 3‐year follow‐up. 24 , 25 However, results from ORAL Strategy showed similar efficacy in pain improvement at 12 weeks for tofacitinib and adalimumab, and similar proportions with low pain at 12 months. 26 In our study, the adjusted mean difference in VAS pain from baseline to 3 months between JAKis and TNFis was 4.0 mm, which is slightly lower than that found in SELECT‐COMPARE (mean difference 6.5 mm for upadacitinib compared with adalimumab) at 12 weeks, and to RA‐BEAM (mean pain reduction at week 12: −31.5 mm and −26.4 mm for baricitinib and adalimumab, respectively). 1 , 24
Because baricitinib and upadacitinib were found to be superior to adalimumab in the RA‐BEAM and SELECT‐COMPARE trials respectively, although no such superiority was found for tofacitinib vs adalimumab in ORAL Strategy, it is possible that the effect on pain reduction differs between the different classes of JAKis. In the present study, more than 80% of the patients were treated with baricitinib, and the remainder with tofacitinib.
Although statistically significant differences in pain reduction at 3 months between JAKis and TNFis were found in our study, and in the above mentioned RCTs, it could be debated whether the estimated difference is of clinical relevance. The minimal clinically important difference for improvement or worsening of pain has been reported as somewhere between 5 and 11 mm, 20 or a 15% relative change. 27 Yet, with a mean difference in pain of 3 to 4 mm for JAKis compared with TNFis in the whole cohort, it is plausible that JAKis could entail a clinically relevant advantage regarding pain in certain subgroups.
The line of treatment with b/tsDMARD had a considerable impact on the results in this study. As first line treatment, all therapies were numerically better regarding pain reduction compared with being given as a later treatment line, with no superiority for JAKis, suggesting that the first advanced therapy is clinically effective in most patients. By contrast, in the stratified analysis, pain improvement at 3 months and fraction of patients with low pain at 12 months were significantly greater for JAKis compared with TNFis when given as third line with b/tsDMARD or later. This is compatible with the known efficacy of JAKis in bDMARD refractory patients. 28 , 29 It could also imply that JAKis are particularly effective in reducing pain in patients that are more resistant to therapy. It has been reported that JAKis may have analgesic effects beyond antiinflammatory properties, 9 which could possibly explain the better effect of JAKis compared with TNFis in the more treatment refractory patients observed in this study, because a higher share of patients with pain causes other than nociceptive could be expected in this group. In an experimental rat model, JAK/signal transducer and activator of transcription signaling has been suggested to increase mechanical pain sensitivity, and subsequent blocking of that pathway, to reduce mechanical allodynia, suggesting a role for JAKis in modulating central pain processing pathways. 30 , 31 The higher proportion with low pain on JAKi despite active disease in this study is in line with this concept.
In this study, we found better pain responses at 3 and 12 months with JAKi monotherapy than with TNFi monotherapy, with numerically greater differences in pain responses for JAKi monotherapy vs TNFi monotherapy compared with JAKi combination therapy vs TNFi combination therapy. This is compatible with previous studies that found JAKi monotherapy to be effective in reducing signs and symptoms of RA, including pain. 32 , 33 Furthermore, JAKi monotherapy appears to be more effective than MTX in reducing PROs. 34 As TNFis in monotherapy were shown to have similar effects as MTX on PROs, 35 , 36 it could be expected that when it comes to monotherapy, JAKis are more efficient for reducing pain than TNFis. Greater pain reduction for baricitinib monotherapy than with adalimumab monotherapy was also found in a matching‐adjusted indirect comparison from a prior systematic literature review. 37
Only a limited number of studies have compared the overall effectiveness of JAKis and non‐TNFi bDMARDs. In a previous analysis of a partly overlapping sample, we found overall equivalent or better treatment responses with baricitinib compared with both TNFis and non‐TNFi bDMARDs when analyzing the outcome measures HAQ‐DI, CDAI, and DAS28 at 3 and 12 months. 17 In the SELECT‐CHOICE trial, in which patients with inadequate response to TNFi treatment were randomized to either upadacitinib or abatacept, greater improvements in pain at 12 weeks were observed in the upadacitinib‐treated patients, 38 while at the same time, there were no differences in the number of swollen joints between the treatment groups. 39 Taken together, this suggests that the observed superior effect of JAKis on composite outcomes 17 is partly driven by a greater reduction in pain.
Regarding comparisons with non‐TNFi bDMARDs, the proportion of patients with low pain at 12 months in the current study were numerically higher for JAKis compared with abatacept when initiated as third treatment line or later, but the difference was not statistically significant. However, when compared with IL‐6i monotherapy, a significantly larger fraction of patients treated with JAKi monotherapy achieved low pain at 12 months. In line with this, greater improvements in pain at 6 months were observed for patients treated with baricitinib compared with tocilizumab in indirect comparisons based on a systematic literature review. 37 In the present study, the share of patients with low pain at 12 months was also greater for JAKis compared with IL‐6is when comparing patients receiving treatment as third line or later with b/tsDMARD. Overall, JAKis were at least as effective as non‐TNFi bDMARDs for pain outcomes at 3 and 12 months.
Several limitations of this study should be mentioned. The majority of patients were treated with baricitinib, potentially limiting generalizability to other JAKis. Although we adjusted for multiple relevant patient characteristics, residual confounding cannot be excluded. Because of a significant amount of missing data, in particular for the follow‐up evaluations, we used multiple imputation to avoid selection bias from potential factors influencing missingness. However, additional bias is possible if factors not accounted for in the model correlated with missingness, and if data were complete, one cannot exclude that the results could have been somewhat different. Nevertheless, the analysis restricted to complete cases supported the main results.
We defined all patients discontinuing treatment as nonresponders in regard to reaching low pain at 12 months, irrespective of the cause of termination, presumably leading to an underestimation of the proportion of responders. However, restricting the analysis to patients on drug would have favored treatments with high discontinuation rates. Notably, the sensitivity analysis, including only patients on drug at 12 months, favored abatacept and IL‐6is, which had somewhat higher discontinuation rates, whereas rituximab was disadvantaged.
Finally, the classifications of monotherapy and combination therapy with csDMARD may not always be correct, as some patients could have stopped csDMARD treatment during the first year. Nevertheless, between 65% and 80% of the patients in the different treatment groups that were classified as comedicating were still on csDMARD treatment at 12 months, and 79% to 92% of patients starting treatment as monotherapy had no registered concurrent csDMARD at 12 months.
The strengths of this study include the real‐world setting for the evaluation of the effectiveness of JAKis on pain compared with various bDMARDs, through the use of a nationwide register in which patients have been observed and treated according to usual care, and hence corresponding to patients seen in daily clinical practice. The main novelty that emerged from this study is the structured estimation of the efficacy of JAKis on pain in patients with RA treated in usual clinical care.
In conclusion, although differences were modest, JAKis had a slightly better effect on pain outcomes at 3 and 12 months in this study compared with TNFis, particularly in patients previously treated with at least two bDMARDs and when given as monotherapy. In addition, JAKis were at least as effective in reducing pain as non‐TNFi bDMARDs.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr Eberhard and Dr Di Giuseppe had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design
Eberhard, Askling, Frisell, Turesson.
Acquisition of data
Di Giuseppe, Chatzidionysiou, Forsblad‐d'Elia, Kastbom, Olofsson, Turesson.
Analysis and interpretation of data
Eberhard, Di Giuseppe, Askling, Bergman, Bower, Chatzidionysiou, Forsblad‐d'Elia, Kastbom, Olofsson, Frisell, Turesson.
Supporting information
Disclosure form
Supplementary Table 1: Baseline covariate definitions
Supplementary Table 2. Pain related disorders defined as one or more of the following before treatment initiation, based on ICD‐10 codes
Supplementary Table 3. Distribution of treatment starts with different b/tsDMARDs
Supplementary Table 4. Missing data
Supplementary Table 5. Proportion of patients on drug at 1 year
Supplementary Table 6. Proportions of patients in CDAI remission, low disease activity or moderate/high disease activity, among patients with VAS pain <20 at 12 months
Supplementary Figure 1. Distribution of changes in VAS pain from baseline to 3 months per treatment arm
Supplementary Figure 2. Differences in proportions attaining low pain score at 12 months, stratified by line of treatment
Supplementary Figure 3. Differences in delta VAS pain from baseline to 3 months, stratified by mono‐ and combination therapy
Supplementary Figure 4. Differences in proportions with low pain at 12 months from baseline, stratified by mono‐ and combination therapy
ACKNOWLEDGMENTS
The authors would like to thank Andrei Barbulescu for his contributions to the statistical analysis plan and his comments on the study.
Supported by The Swedish Research Council (grant 2015‐02228), The Swedish Rheumatism Association (grant R‐664091), and Lund University (grant ALFSKANE‐446501).
Drs Eberhard and Di Giuseppe are co‐first authors and contributed equally to this work. Drs Frisell and Turesson are co‐last authors and contributed equally to this work.
Additional supplementary information cited in this article can be found online in the Supporting Information section (http://onlinelibrary.wiley.com/doi/10.1002/art.43014).
Author disclosures are available at https://onlinelibrary.wiley.com/doi/10.1002/art.43014.
REFERENCES
- 1. Taylor PC, Keystone EC, van der Heijde D, et al. Baricitinib versus placebo or adalimumab in rheumatoid arthritis. N Engl J Med 2017;376:652–662. [DOI] [PubMed] [Google Scholar]
- 2. Fleischmann RM, Genovese MC, Enejosa JV, et al. Safety and effectiveness of upadacitinib or adalimumab plus methotrexate in patients with rheumatoid arthritis over 48 weeks with switch to alternate therapy in patients with insufficient response. Ann Rheum Dis 2019;78:1454–1462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Lee YH, Song GG. Relative efficacy and safety of tofacitinib, baricitinib, upadacitinib, and filgotinib in comparison to adalimumab in patients with active rheumatoid arthritis. Z Rheumatol 2020;79:785–796. [DOI] [PubMed] [Google Scholar]
- 4. Fleischmann R, Mysler E, Hall S, et al; ORAL Strategy investigators . Efficacy and safety of tofacitinib monotherapy, tofacitinib with methotrexate, and adalimumab with methotrexate in patients with rheumatoid arthritis (ORAL Strategy): a phase 3b/4, double‐blind, head‐to‐head, randomised controlled trial. Lancet 2017;390:457–468. [DOI] [PubMed] [Google Scholar]
- 5. Combe B, Kivitz A, Tanaka Y, et al. Filgotinib versus placebo or adalimumab in patients with rheumatoid arthritis and inadequate response to methotrexate: a phase III randomised clinical trial. Ann Rheum Dis 2021;80:848–858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Harrington R, Al Nokhatha SA, Conway R. JAK inhibitors in rheumatoid arthritis: an evidence‐based review on the emerging clinical data. J Inflamm Res 2020;13:519–531. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Taylor PC, Lee YC, Fleischmann R, et al. Achieving pain control in rheumatoid arthritis with baricitinib or adalimumab plus methotrexate: results from the RA‐BEAM Trial. J Clin Med 2019;8:831. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Toth L, Juhasz MF, Szabo L, et al. Janus kinase inhibitors improve disease activity and patient‐reported outcomes in rheumatoid arthritis: a systematic review and meta‐analysis of 24,135 patients. Int J Mol Sci 2022;23:1246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Fautrel B, Kirkham B, Pope JE, et al. Effect of baricitinib and adalimumab in reducing pain and improving function in patients with rheumatoid arthritis in low disease activity: exploratory analyses from RA‐BEAM. J Clin Med 2019;8:1394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Eberhard A, Bergman S, Mandl T, et al. Joint tenderness at 3 months follow‐up better predicts long‐term pain than baseline characteristics in early rheumatoid arthritis patients. Rheumatology (Oxford) 2024;63(3):734–741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Altawil R, Saevarsdottir S, Wedren S, et al. Remaining pain in early rheumatoid arthritis patients treated with methotrexate. Arthritis Care Res (Hoboken) 2016;68:1061–1068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Olofsson T, Wallman JK, Joud A, et al. Pain over two years after start of biologic versus conventional combination treatment in early rheumatoid arthritis: results from a Swedish Randomized Controlled Trial. Arthritis Care Res (Hoboken) 2021;73:1312–1321. [DOI] [PubMed] [Google Scholar]
- 13. Heisler A, Song J, Muhammad L, et al. Association between changes in pain sensitization and changes in disease activity after 12 weeks of diease modifying anti‐rheumatic drug therapy in rheumatoid arthritis [abstract]. Arthritis Rheumatol 2020;72(suppl 10). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Taylor P, Manger B, Alvaro‐Gracia J, et al. Patient perceptions concerning pain management in the treatment of rheumatoid arthritis. J Int Med Res 2010;38:1213–1224. [DOI] [PubMed] [Google Scholar]
- 15. ten Klooster PM, Veehof MM, Taal E, van Riel PL, van de Laar MA. Changes in priorities for improvement in patients with rheumatoid arthritis during 1 year of anti‐tumour necrosis factor treatment. Ann Rheum Dis 2007;66:1485–1490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Reed GW, Gerber RA, Shan Y, et al. Real‐world comparative effectiveness of tofacitinib and tumor necrosis factor inhibitors as monotherapy and combination therapy for treatment of rheumatoid arthritis. Rheumatol Ther 2019;6:573–586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Barbulescu A, Askling J, Chatzidionysiou K, et al. Effectiveness of baricitinib and tofacitinib compared with bDMARDs in RA: results from a cohort study using nationwide Swedish register data. Rheumatology (Oxford) 2022;61:3952–3962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Wadström H, Eriksson JK, Neovius M, et al. ARTIS Study Group . How good is the coverage and how accurate are exposure data in the Swedish Biologics Register (ARTIS)? Scand J Rheumatol 2015;44:22–28. [DOI] [PubMed] [Google Scholar]
- 19. Wells GA, Boers M, Shea B, et al. Minimal disease activity for rheumatoid arthritis: a preliminary definition. J Rheumatol 2005;32:2016–2024. [PubMed] [Google Scholar]
- 20. Wolfe F, Michaud K. Assessment of pain in rheumatoid arthritis: minimal clinically significant difference, predictors, and the effect of anti‐tumor necrosis factor therapy. J Rheumatol 2007;34:1674–1683. [PubMed] [Google Scholar]
- 21. Kristensen LE, Saxne T, Geborek P. The LUNDEX, a new index of drug efficacy in clinical practice: results of a five‐year observational study of treatment with infliximab and etanercept among rheumatoid arthritis patients in southern Sweden. Arthritis Rheum 2006;54:600–606. [DOI] [PubMed] [Google Scholar]
- 22. Frisell T, Baecklund E, Bengtsson K, et al. Patient characteristics influence the choice of biological drug in RA, and will make non‐TNFi biologics appear more harmful than TNFi biologics. Ann Rheum Dis 2018;77:650–657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lydersen S. Multiple imputation of missing data. Tidsskr Nor Laegeforen 2022;142. [DOI] [PubMed] [Google Scholar]
- 24. Fleischmann R, Pangan AL, Song IH, et al. Upadacitinib versus placebo or adalimumab in patients with rheumatoid arthritis and an inadequate response to methotrexate: results of a phase III, double‐blind, randomized controlled trial. Arthritis Rheumatol 2019;71:1788–1800. [DOI] [PubMed] [Google Scholar]
- 25. Fleischmann R, Mysler E, Bessette L, et al. Long‐term safety and efficacy of upadacitinib or adalimumab in patients with rheumatoid arthritis: results through 3 years from the SELECT‐COMPARE study. RMD Open 2022;8:e002012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Strand V, Mysler E, Moots RJ, et al. Patient‐reported outcomes for tofacitinib with and without methotrexate, or adalimumab with methotrexate, in rheumatoid arthritis: a phase IIIB/IV trial. RMD Open 2019;5:e001040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Salaffi F, Stancati A, Silvestri CA, et al. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain 2004;8:283–291. [DOI] [PubMed] [Google Scholar]
- 28. Tanaka Y, Fautrel B, Keystone EC, et al. Clinical outcomes in patients switched from adalimumab to baricitinib due to non‐response and/or study design: phase III data in patients with rheumatoid arthritis. Ann Rheum Dis 2019;78:890–898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Fleischmann R, Blanco R, Van den Bosch F, et al. Long‐term efficacy and safety following switch between upadacitinib and adalimumab in patients with rheumatoid arthritis: 5‐year data from SELECT‐COMPARE. Rheumatol Ther 2024;11:599–615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Busch‐Dienstfertig M, González‐Rodríguez S. IL‐4, JAK‐STAT signaling, and pain. JAKSTAT 2013;2:e27638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Dominguez E, Mauborgne A, Mallet J, et al. SOCS3‐mediated blockade of JAK/STAT3 signaling pathway reveals its major contribution to spinal cord neuroinflammation and mechanical allodynia after peripheral nerve injury. J Neurosci 2010;30:5754–5766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Genovese M, Westhovens R, Meuleners L, et al. Effect of filgotinib, a selective JAK 1 inhibitor, with and without methotrexate in patients with rheumatoid arthritis: patient‐reported outcomes. Arthritis Res Ther 2018;20:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Lee EB, Fleischmann R, Hall S, et al. ORAL Start Investigators . Tofacitinib versus methotrexate in rheumatoid arthritis. N Engl J Med 2014;370:2377–2386. [DOI] [PubMed] [Google Scholar]
- 34. Schiff M, Takeuchi T, Fleischmann R, et al. Patient‐reported outcomes of baricitinib in patients with rheumatoid arthritis and no or limited prior disease‐modifying antirheumatic drug treatment. Arthritis Res Ther 2017;19:208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. van der Heijde D, Klareskog L, Singh A, et al. Patient reported outcomes in a trial of combination therapy with etanercept and methotrexate for rheumatoid arthritis: the TEMPO trial. Ann Rheum Dis 2006;65:328–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Singh JA, Hossain A, Mudano AS, et al. Biologics or tofacitinib for people with rheumatoid arthritis naive to methotrexate: a systematic review and network meta‐analysis. Cochrane Database Syst Rev 2017;5:CD012657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Fautrel B, Zhu B, Taylor PC, et al. Comparative effectiveness of improvement in pain and physical function for baricitinib versus adalimumab, tocilizumab and tofacitinib monotherapies in rheumatoid arthritis patients who are naive to treatment with biologic or conventional synthetic disease‐modifying antirheumatic drugs: a matching‐adjusted indirect comparison. RMD Open 2020;6(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Bergman M, Tundia N, Martin N, et al. Patient‐reported outcomes of upadacitinib versus abatacept in patients with rheumatoid arthritis and an inadequate response to biologic disease‐modifying antirheumatic drugs: 12‐ and 24‐week results of a phase 3 trial. Arthritis Res Ther 2022;24:155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Rubbert‐Roth A, Enejosa J, Pangan AL, et al. Trial of upadacitinib or abatacept in rheumatoid arthritis. N Engl J Med 2020;383:1511–1521. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Disclosure form
Supplementary Table 1: Baseline covariate definitions
Supplementary Table 2. Pain related disorders defined as one or more of the following before treatment initiation, based on ICD‐10 codes
Supplementary Table 3. Distribution of treatment starts with different b/tsDMARDs
Supplementary Table 4. Missing data
Supplementary Table 5. Proportion of patients on drug at 1 year
Supplementary Table 6. Proportions of patients in CDAI remission, low disease activity or moderate/high disease activity, among patients with VAS pain <20 at 12 months
Supplementary Figure 1. Distribution of changes in VAS pain from baseline to 3 months per treatment arm
Supplementary Figure 2. Differences in proportions attaining low pain score at 12 months, stratified by line of treatment
Supplementary Figure 3. Differences in delta VAS pain from baseline to 3 months, stratified by mono‐ and combination therapy
Supplementary Figure 4. Differences in proportions with low pain at 12 months from baseline, stratified by mono‐ and combination therapy