Aerobic exercise probably has a small clinically important benefit on function (mean difference (MD) in HAQ DI −0.25 points; 95% CI −0.38 to −0.11) with no significant difference between long-term and short-term interventions.
There are no major adverse effects reported during exercise interventions, nor does exercise increase the disease activity or pain. On the contrary, evidence indicates that strength exercise may reduce disease activity (DAS28: SMD -0.69 [-1.26, -0.11]; ESR: SMD -0.86 [-1.65, -0.07;] and aerobic exercise may slightly reduce pain compared with other exercise types (SMD −0.46; 95% CI −0.90 to −0.01), but the results are inconclusive. Exercise may result to little or no difference in the quality of life in people with rheumatoid arthritis.
The quality of evidence is downgraded due to overall risk of bias of included studies, imprecision and inconsistency.
The evidence is applicable to adults and elderly people with rheumatoid arthritis.
| Reference | Study type | Population n Country |
Intervention (I) and comparison (C) | Outcomes (outcome measures) | Risk of bias «Additional comments for included studies...»2 |
|---|---|---|---|---|---|
| Ye et al. 2022 «Ye H, Weng H, Xu Y, et al. Effectiveness and safet...»1 |
SR + MA | Adults with rheumatoid arthritis 13 RCT studies 967 participants USA, Netherlands, Turkey, France, Saudi-Arabia, UK, Sweden, Ireland, Mexico |
I: Aerobic exercise C: Usual care or other types of exercise |
Function (HAQ DI) Disease activity (DAS28 CRP, ESR) Pain (VAS, SFMPQ) |
Moderate |
| Wen et al. 2021 «Wen Z, Chai Y. Effectiveness of resistance exercis...»2 |
SR + MA | Adults with rheumatoid arthritis without other serious diseases 17 RCT studies 610 participants USA, UK, Finland, Sweden, Netherlands, Bosnia, New Zealand, Ireland, Korea |
I: Strength + rehabilitation exercise C: usual care (regular nursing), non-aerobic exercise or range of movement exercise. |
Function (HAQ DI) Disease activity (DAS28, ESR) Pain (VAS) |
High |
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
SR + MA | Adults with rheumatoid arthritis 9 RCT studies 1343 participants |
I: Exercise therapy C: Usual care |
Function (HAQ) Disease activity (VAS, DAS28) Pain (VAS) Quality of life (SF-36, EQ-5D) |
High |
| Andonian et al. 2023 «Andonian BJ, Ross LM, Sudnick AM, et al. Effect of...»4 |
RCT | Physically inactive older adults with rheumatoid arthritis 20 participants |
I: Combination exercise (aerobic + strength) + diet C: Diet and exercise counselling |
Disease activity (DAS28, CRP, ESR) |
Low |
| Cerasola et al. 2023 «Cerasola D, Argano C, Chiovaro V, et al. Physical ...»5 |
RCT | Adults with rheumatoid arthritis 80 participants Italy |
I: Occupational therapy and home exercise C: brochure about home exercises and general information |
Function (HAQ) Disease activity (DAS28, CRP, ESR) |
Low |
| Tauwen et al. 2024 «Teuwen MMH, van Weely SFE, Vliet Vlieland TPM, et ...»6 |
RCT | Adults with rheumatoid arthritis and severe limitations in daily activities 217 participants Netherlands |
I: Combination exercise (aerobic, strength, mobility, functional) C: Usual care (might include rehabilitation) |
Function (PSC1 NRS, PROMIS, HAQ-DI) Quality of life (RAQoL) |
Low |
| Bilberg et al. 2024 «Bilberg A, Mannerkorpi K, Borjesson M, et al. High...»7 |
RCT | Adults with stable rheumatoid arthritis and low physical activity 87 participants Sweden |
I: Combination exercise (HIIT + strength) C: Education on exercise and home exercise program |
Disease activity (DAS28, CRP, ESR) Pain (VAS) |
Low |
| RCT = randomized controlled trial; SR = systematic review; MA = meta-analysis; HAQ = Health Assessment Questionnaire; HAQ DI = Health Assessment Questionnaire disability index; DAS28 = Disease Activity Score in 28 joints; VAS = visual analogue scale; SFMPQ = Short Form McGill Pain Questionnaire; SF-36 = Short-Form Health Survey 36; SF-12 = Short-Form Health Survey 12; PROMIS = Patient-Reported Outcomes Measurement Information System; PSC1 NRS = Patient-Specific Complaints Numerical Rating Scale; RAQoL = Rheumatoid Arthritis Quality of Life Questionnaire; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; HIIT = high-intensity interval training | |||||
| Reference | Comments |
|---|---|
| Ye et al. 2022 «Ye H, Weng H, Xu Y, et al. Effectiveness and safet...»1 |
Interventions: aerobic exercise interventions performed at 50–90% of the maximal heart
rate, including walking, cycling, jogging, etc. Duration of interventions: 2-96 weeks. Meta-analysis subgroups: long term > 16weeks and short term ≤ 16 weeks. Review lacks an assessment of publication bias. 8/13 of the original studies have high risk of bias due to lack of allocation concealment. No other concerns identified. |
| Wen et al. 2021 «Wen Z, Chai Y. Effectiveness of resistance exercis...»2 |
Exercise interventions were mostly land-based leg and arm strength training, performed
1-3 times/week, 5-75 minutes/ training session. Duration of the interventions varied
from 4 to 104 weeks. Review lacks a pre-registered protocol and sensitivity analysis. The risk of bias assessment is not reported, only total Jadad-scores. Risk of bias in the meta-analysis due to variations in the length of interventions in the same analysis. |
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
Intervention: Diverse exercise programs, including dynamic exercise, progressive resistance,
water-based aerobic exercise and sensorimotor training. Duration of the interventions:
4-16 weeks. Review lacks a pre-registered protocol, sensitivity analysis and a clear estimate of publication bias. 2/6 of the original studies have an unclear risk of bias due to allocation concealment and all studies have high risk of bias due to lack of blinding of the outcome assessment. |
| Andonian et al. 2023 «Andonian BJ, Ross LM, Sudnick AM, et al. Effect of...»4 |
Inclusion criteria: adults aged 60 to 80 years with overweight/obesity (body mass
index: 28–40), with seropositive or erosive RA meeting 2010 ACR/EULAR criteria for
RA,28 not currently meeting the 2018 Physical Activity Guidelines for Americans,29
and without absolute contraindications to exercise were enrolled. Interventions: aerobic training (150 minutes/week moderate-to-vigorous intensity), resistance training (two days/week), and a hypocaloric diet (7% weight loss goal. Control group completed two lifestyle counseling sessions followed by monthly check-ins. Duration of the interventions: 16 weeks. Unable to blind participants due to intervention, might have some effect on PROMIS. No other biases detected. |
| Cerasola et al. 2023 «Cerasola D, Argano C, Chiovaro V, et al. Physical ...»5 |
The selection criteria included a diagnosis of rheumatoid arthritis, aged 18 years
or older. Patients aged under 18 and subjects with formed deformations of joints that
prevent physical exercise or severe anemia were excluded. Intervention included occupational therapy: (1)physical exercise to manage balance to maintain muscle tone and joint restart after the night, and (2) joint protection and decreased joint load during daily actions. The physical exercise included the flexion and extension of the upper and lower limbs for the main joints (ankle, knee, cox-femoris, elbow, wrist, and shoulder) with or without weight tools for 20/30 min day. Duration of the intervention: 30 days. No study protocol available. Unable to blind participants due to intervention, might have some effect on HAQ. No other biases detected. |
| Tauwen et al. 2024 «Teuwen MMH, van Weely SFE, Vliet Vlieland TPM, et ...»6 |
Eligible individuals were adults (aged ≥18 years) with a clinical diagnosis of RA
made by a rheumatologist. Individuals had self-perceived severe limitations in daily
activities involving self-care (eg, dressing, washing), and/or transfers (eg, getting
in and out of bed, rising from a chair or using the toilet), and/or mobility indoors
or outdoors. Individuals who had received physical therapy in the past 3 months, either
or not in the context of a multidisciplinary team intervention or were shortly in
need of admission to a hospital or rehabilitation centre, were excluded. Intervention: longstanding, personalised exercise therapy for 52 weeks. Active treatment comprised exercises (aerobic, muscle strengthening, flexibility/joint range of motion and functional/ neuromotor exercises), patient education and the promotion of physical activity, including the provision of a simple waist pedometer. Control group received usual care, with the content and delivery determined by the treating clinician(s) and participants themselves. Unable to blind participants due to intervention, might have some effect on patient-reported outcomes. Allocation concealment was not held during the study for all participants despite respectful efforts for concealment, outcome assessors became aware of the allocation to some amount. This might be considered inevitable in long studies, and probably has only minor effect on patient-reported outcomes. Control group might have participated in physiotherapy, as this was not denied. |
| Bilberg et al. 2024 «Bilberg A, Mannerkorpi K, Borjesson M, et al. High...»7 |
Patients fulfilling the American College of Rheumatology/European Alliance of Associations
for Rheumatology (ACR/EULAR) 1987/2010 criteria for RA,23 24 with disease duration
>1 year, aged 20–60 years, stable treatment on antirheumatic drugs for the past 3
months, and low-to-moderate disease activity (DAS28 <5.1) were eligible for inclusion.
DAS28 <5.1 was chosen as upper limit for inclusion for safety considerations. Exclusion
criteria were symptoms of CVD, other comorbidities, physical disabilities and pregnancy
precluding participation in high-intensity exercise, engagement in regular cardiorespiratory
exercise on high intensity level (>1 hour/week for the past 6 months), inability to
understand and speak Swedish. Interventions: Exercise program consisted of high intensity interval training two times/week and muscle strength exercise two times/week (supervised) and non-supervised aerobic exercise one/week. Control group received individual information of the general health recommendations for physical activity, with encouragement to be physically active on moderate intensity level ≥150 min/week. A home exercise programme was provided, and verbal and written instructions were given. Duration of the interventions: 12 weeks Unable to blind participants due to intervention, but this has no relevant effect on the objective outcomes used. No biases detected. |
| Clinically meaningful means HAQ DI change >0.22, which is a frequently used Minimal
clinically important difference (MCID) for HAQ-DI in people with rheumatoid arthritis
(Behrens et al. 2019). However, MCID may be as low as 0.07 in people with arthralgia (Claasen
et al. 2025), whereas discrimination from random long-term variations in patients
on stable therapy may require a MCID as high as 0.68 (Claasen et al. 2025). The Minimal Important Difference for the PSC1 NRS (on a scale from 0 to 10) ranged from 1.3 (small change) to 2.3 (medium change) to 2.7 (large change) in people with musculoskeletal disorders (Abbott 2014). Minimal important difference in VAS is defined as -1,5 = small, -3 = medium and -3,5 = large (Abbott et al. 2014) |
|
| Summary of the risk of bias | Evidence is based on three systematic reviews and meta-analysis, and four RCT studies, one large (n=217), two medium-sized (n≈80) and one small (n=20). The risk of bias in the reviews is moderate to high, whereas the risk of bias in the RCT studies is low. The grading of evidence is done individually for every four outcomes. |
Results
| Reference | Number of studies and number of participants (intervention (I) / control (C)) | Follow-up (F) | Relative effect (95% CI) (results after intervention) |
|---|---|---|---|
| Level of evidence: moderate The quality of evidence is downgraded due to (study limitations, inconsistency, imprecision). I= intervention; C=comparison; CI=confidence interval; HAQ DI=Health Assessment Questionnaire disability index; HAQ = Health Assessment Questionnaire; MD = Maen Difference; WMD = Weighted Maen Difference (WMD); OT = occupational therapy; PSC1 NRS = Patient-Specific Complaints Numerical Rating Scale |
|||
| Ye et al. 2022 «Ye H, Weng H, Xu Y, et al. Effectiveness and safet...»1 |
11 RCT studies n=735 I (aerobic exercise) 378 participants C (usual care) 357 participants |
None (assessments were performed immediately at the end of the training interventions, with no follow-up measurements). | Exercise had a clinically important effect on function HAQ DI MD -0.25 [-0.38. -0.11]
No significant difference between long-term and short-term interventions |
| Wen et al. 2021 «Wen Z, Chai Y. Effectiveness of resistance exercis...»2 |
13 RCT studies n=745 I (strength and rehab exercises) 380 participants Control (usual care) 365 participants |
None (assessments were performed immediately at the end of the training interventions, with no follow-up measurements). | No statistical difference between groups (HAQ WMD -0.10 [-0.26, 0.06] |
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
5 RCT studies n=347 I (exercise) 170 participants C (usual care) 177 participants |
F: n/a | Exercise had a clinically important effect on function HAQ DI MD -0.35 [-0,60, -0.10] |
| Cerasola et al. 2023 «Cerasola D, Argano C, Chiovaro V, et al. Physical ...»5 |
1 RCT (single study) n=160 I (OT + comb. exer.) 80 participants C (education) 80 participants |
None | Exercise had a clinically important effect on function HAQ MD 2.05 ± 0.83 |
| Tauwen et al. 2024 «Teuwen MMH, van Weely SFE, Vliet Vlieland TPM, et ...»6 |
1 RCT (single study) n=202 I (comb. exercise) 104 participants C (education) 98 participants |
None | Exercise had a large effect on function compared to education alone (PSC1 NRS MD -1.7 [-2.4, -1.0]; ES 0.7). Similar results in secondary outcome measures. |
| Reference | Number of studies and number of participants (intervention (I) / control (C)) | Follow-up (F) | Relative effect (95% CI) (results after intervention) |
|---|---|---|---|
| Level of evidence: low The quality of evidence is downgraded due to (study limitations, inconsistency, imprecision). I= intervention; C=comparison; CI=confidence interval; MD = Mean Difference; SMD = Standardized Mean Difference; DAS28 = Disease Activity Score in 28 joints;CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; TJC = tender joint count; SJC = swollen joint count; RAI = Ritchie Articular Index |
|||
Ye et al. 2022 «Ye H, Weng H, Xu Y, et al. Effectiveness and safet...»1 |
DAS28 8 RCT studies I (aerobic exercise) 289 participants C (usual care) 284 participants CRP 5 RCT studies I (aerobic exercise) 149 participants C (usual care) 149 participants ESR 5 RCT studies I (aerobic exercise) 157 participants C (usual care) 148 participants |
n/a | No statistically significant difference in disease activity between groups, when measured
with DAS28: MD -0.55 [-1.12, 0.01] or CRP: MD -1.08 [-2.20,0.05] or ESR: SMD -0.76 [-1.67, 0.14] Subgroup analysis showed that long-term aerobic exercise significantly decreased CRP levels: MD = -2.28 [-3.24, -1.33] |
| Wen et al. 2021 «Wen Z, Chai Y. Effectiveness of resistance exercis...»2 |
8 RCT studies (DAS28) 6 RCT studies (ESR) I (strength and rehab exercises) DAS28: 252 participants ESR: 178 Control (usual care) DAS28241 participants ESR 158 participants |
n/a | Exercise had a moderate effect on decreased disease activity when measured with DAS28:
SMD -0.69 [-1.26, -0.11]. and a large effect when measured with ESR: SMD -0.86 [-1.65,
-0.07] |
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
3 RCT studies I (exercise) 79 participants C (usual care) 88 participants |
n/a | No statistically significant difference between groups on disease activity: MD DAS28
-0.52 [-1.25, 0.21] |
| Andonian et al. 2023 «Andonian BJ, Ross LM, Sudnick AM, et al. Effect of...»4 |
1 RCT (single study) I (exercise) 12 participants C (education) 12 participants |
none | Statistically significant disease activity reduction in the exercise group compared
to control group in DAS28-CPR but not in DAS28-ESR. No difference between groups in other outcomes (TJC, SJC, RAI, hs-CRP, ESR) |
| Cerasola et al. 2023 «Cerasola D, Argano C, Chiovaro V, et al. Physical ...»5 |
1 RCT (single study) I (OT + comb. exer.) 80 participants C (education) 80 participants |
none | Statistically significant disease activity reduction in the exercise group compared
to control group: DAS28-ESR MD -1.95±0.84 vs. -0.37±1.69 and DAS28-CRP MD -2.0±0.88 vs. -0.7±1.69 |
| Bilberg et al. 2024 «Bilberg A, Mannerkorpi K, Borjesson M, et al. High...»7 |
1 RCT (single study) I (comb. exercise) 44 participants C (education) 43 participants |
none | No statistically significant difference between groups (DAS28-ESR, ESR, CRP) |
| Reference | Number of studies and number of participants (intervention (I) / control (C)) | Follow-up (F) | Relative effect (95% CI) (results after intervention) |
|---|---|---|---|
| Level of evidence: very low The quality of evidence is downgraded due to (study limitations, inconsistency and imprecision. I= intervention; C=comparison; CI=confidence interval; MD = Mean Difference; SMD = Standardized Mean Difference; VAS = Visual Analogue Scale; |
|||
| Ye et al. 2022 «Ye H, Weng H, Xu Y, et al. Effectiveness and safet...»1 |
4 RCT studies I (aerobic exercise) 128 participants C (usual care) 119 participants |
D: 2-96 wk (subgroups long term > 16wk and short term ≤ 16 wk) F: n/a |
Exercise had a small effect on pain (SMD -0.46 [-0.90, -0.01] |
| Wen et al. 2021 «Wen Z, Chai Y. Effectiveness of resistance exercis...»2 |
6 RCT studies I (strength and rehab exercises) 159 participants Control (usual care) 159 participants |
D: 4-104 weeks F: n/a |
No statistical difference between groups in pain scores when measured with VAS SMD -0.61[-1.49, 0.27] |
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
3 RCT studies I (exercise) 123 participants C (usual care) 121 participants |
D: 4-16 weeks F: n/a |
Exercise had a small effect on pain when measured with VAS MD -2.04 [-3.77, -0.32] |
| Bilberg et al. 2024 «Bilberg A, Mannerkorpi K, Borjesson M, et al. High...»7 |
1 RCT (single study) I (comb. exercise) 44 participants C (education) 43 participants |
D: 12 weeks F: none |
No statistical difference between groups in pain scores when measured with VAS MD -4.0 [-13.07, 5.06] |
Notes: Minimal important difference in VAS is defined as -1,5 = small, -3 = medium and -3,5 = large (Abbott et al. 2014 «Abbott JH, Schmitt J. Minimum important difference...»8)
| Reference | Number of studies and number of participants (intervention (I) / control (C)) | Follow-up (F) | Relative effect (95% CI) (results after intervention) |
|---|---|---|---|
| Level of evidence: low The quality of evidence is downgraded due to (study limitations, inconsistency, imprecision). I= intervention; C=comparison; CI=confidence interval; MD = Mean Difference; SMD = Standardized Mean Difference; VAS = Visual Analogue Scale; RAQoL = Rheumatoid Arthritis Quality of Life; SF-36 = 36-Item Short-Form Health Survey |
|||
| Sobue et al. 2021 «Sobue Y, Kojima T, Ito H, et al. Does exercise the...»3 |
2 RCT studies I (exercise) 78 participants C (usual care) 84 participants |
n/a | Exercise had a beneficial effect on 6/8 categories of the SF36: Functional skills MD 21.39 [13.71, 29.07] Pain MD 13.21 [7.36, 19.05] General health MD 13.75 [5.77, 21.73] Vitality MD 7.51 [1.41, 13.62] Social aspects MD 11.04 [0.64, 21.43] Mental health MD 10.34 [2.11, 18.57] and no difference between groups in 2/8 categories: Physical aspects MD 22.26 [-2.40, 46.92] Emotional aspects MD 15.25 [-3.18, 33.68] |
| Tauwen et al. 2024 «Teuwen MMH, van Weely SFE, Vliet Vlieland TPM, et ...»6 |
1 RCT (single study) I (comb. exercise) 104 participants C (education) 98 participants |
F: none | Exercise improved quality of life when measured with RAQoL: MD -2.03 [-3.38, -0.69] and SF-36 PCS: MD 3.83 [1.49, 6.17] There were no difference between groups when measured with SF-36 MCS: MD 2.54 [-0.47,5.54] |