RA = rheumatoid arthritis; NR = not reported; PsA = psoriatric arthritis; DMARD = disease-modifying antirheumatic drug; MTX = methotrexate.
Evidence for the benefit of aerobic and strengthening exercise in rheumatoid arthritis
Article first published online: 3 JUN 2003
Copyright © 2003 by the American College of Rheumatology
Arthritis Care & Research
Volume 49, Issue 3, pages 428–434, 15 June 2003
How to Cite
Stenström, C. H. and Minor, M. A. (2003), Evidence for the benefit of aerobic and strengthening exercise in rheumatoid arthritis. Arthritis & Rheumatism, 49: 428–434. doi: 10.1002/art.11051
- Issue published online: 3 JUN 2003
- Article first published online: 3 JUN 2003
- Manuscript Accepted: 2 NOV 2002
- Manuscript Received: 20 JUN 2002
Rheumatoid arthritis (RA) is a chronic systemic inflammatory disease of unknown etiology. It usually presents as symmetric polyarthritis, but also confers general inflammatory symptoms. RA is present in 0.5–1% of the general population, twice as often in women, and the age at disease onset is mainly between 45 and 65 years. The American College of Rheumatology (formerly American Rheumatism Association [ARA]) has suggested diagnostic criteria (1) for RA. The disease course varies and prediction of the prognosis is difficult in any particular case. In the long run, a reduced function, difficulties in activities of daily living (ADL), and a negative psychosocial impact are often seen. RA also is connected with increased risk for osteoporosis (2), cardiovascular disease, and premature death (3). Criteria for the classification of functional loss have been suggested (4). Functional class (FC) I includes individuals without difficulties in daily life, FC II includes those with symptoms but minor limitations only, FC III includes those who are partly dependent, and FC IV includes those who are totally dependent on other persons in daily life. The vast majority of individuals with RA belong to FC II.
Pain, stiffness, and fatigue generally occur early in the course of RA. Reduction of body function, such as range of motion (ROM), muscle strength, and aerobic capacity, may follow. As previously summarized by van den Ende et al (5), 50% of patients with RA displayed decreased hand ROM at their first rheumatology visit. Two years later, reduced ROM was found in large joints; the decreased ROM varied between 25% and 35% in different joints (6). Patients with RA and some functional loss (ARA FC II) have been found to have a 25–50% reduction in muscular strength compared with age-matched healthy controls (7–10), and 55% deficit in muscular endurance (9). In patients with more serious RA, reductions of muscular strength up to 70% have been reported (11). Patients taking long-term oral steroids have a more pronounced loss of muscle strength (12, 13). Reduced muscle function in patients with RA may also present itself as loss of functional balance and coordination (14). Among patients with RA who were able to perform bicycle ergometer tests, the aerobic capacity was reduced by 20–30% (7, 9, 15, 16). The aerobic capacity is probably even lower among those who are not able to perform such tests. A portion of the reduced physical capacities found among individuals with RA may be attributed to inadequate levels of physical activity.
The treatment of RA focuses on decreasing inflammatory activity and symptoms, limiting joint destruction and disability, and improving health related quality of life. It includes a rich variety of medication, surgery, and rehabilitation. Despite earlier fear of aggravation of symptoms, increased disease activity, and joint destruction, there is now scientific support that various forms of exercise are both safe and beneficial (17). Exercise has become an important part of rehabilitation during the last decades. A number of studies of various quality have been carried out to investigate the effects of aerobic and strengthening exercise in RA, but no recent reviews have been published.
The aim of the present study was to investigate the evidence for the benefit of aerobic and strengthening exercise in RA.
Materials and Methods
Randomized controlled trials (RCTs) investigating the effects of exercise to improve aerobic capacity and muscle strength in people with RA were searched in Medline via Ovid. The search was performed in the beginning of June 2001 and limited to studies involving humans and written in the English language.
A combination of the keywords “arthritis, rheumatoid” and either “exercise/exercise therapy” or “physical education and training” resulted in 208 articles. Those reporting nonoriginal research, effects of interventions other than exercise, exercise in connection with surgery, and studies with designs other than RCT were excluded in a subsequent manual screening process.
After this process, 30 articles reporting on 26 RCTs remained (18–47). Four of these studies, reported in 5 articles, were excluded because they evaluated ROM exercise (21, 22, 25, 27, 36) and another 3 were eliminated because the exercise put too little emphasis on aerobic capacity or strength (30, 35, 43). Another study was excluded because its focus was not on the exercise applied (34). Further, 3 original studies reported in 4 articles (18–20, 31) and 1 followup (38) were excluded because between-group analyses of exercise effects were not reported. An additional paper (48) on a study already identified in the database search (39) was found in its reference list and included here.
Seventeen papers reporting results from 15 RCTs remained for further assessment (Table 1).They had originally been published between 1989 and 2000, but in 2 cases, followups had been reported in additional articles, the latest in March 2001.
|Author, year (reference)||Diagnoses, characteristics||Functional class||Female, %||Age, years mean||Numbers at start (end/followup)|
|Minor et al 1989 (23)||RA, lower limb involvement||NR||85||54||40 (28)|
|Ekdahl et al 1990 (24)||RA, lower limb involvement||II||64||53||67 (62/56)|
|Baslund et al 1993 (26)||RA or PsA, ≥2 swollen joints, ≥15 min morning stiffness||NR||89||48||18 (18)|
|Hansen et al 1993 (28)||RA||I–II||65||52†||75 (65)|
|Hoenig et al 1993 (29)||RA||II–III||NR||57||57 (41/NR)|
|Häkkinen et al 1994 (32)||RA or PsA, recent-onset||NR||51||∼43||43 (39)|
|Lyngberg et al 1994 (33)||RA, fragile on low-dose steroids||NR||92||∼67||24 (24)|
|Stenström et al 1996 (48)||RA, PsA or other peripheral arthritis||I–II||72||54||54 (48)|
|Stenström et al 1997 (39)||RA, PsA or other peripheral arthritis||I–II||72||54||54 (34)|
|Van den Ende et al 1996 (37)||RA, no joint replacement in lower limbs||NR||63||52||100 (99/93)|
|Komatireddy et al 1997 (40)||RA||II–III||76||61||49 (41)|
|Boström et al 1998 (41)||RA, shoulder pain and/or functional limitation||NR||100||66||45 (37)|
|Häkkinen et al 1999 (42)||RA, recent-onset, no DMARD or glucocorticoids||NR||62||∼49||70 (65)|
|Häkkinen et al 2001 (47)||RA, recent-onset, no DMARD or glucocorticoids||NR||62||∼49||70 (62)|
|McMeeken et al 1999 (44)||RA, no joint replacement or fragile skin||NR||83||∼51||36 (35)|
|van den Ende et al 2000 (45)||RA, active disease, functional disability, no knee replacement||NR||63||60||64 (61/47/52/48)|
|Westby et al 2000 (46)||RA, continous low-dose prednisone, no previous high-dose prednisone or recent MTX use||I–II||100||∼56||30 (21)|
Participants were mainly those with RA, but some studies also included a minor number with peripheral arthritis of other origin. Most of the studies included individuals in ARA FC II with low to moderate disease activity and on stable medication. However, a couple of studies investigated patients in an acute disease stage or those who were considered fragile due to long-term steroid treatment. The percentage of females varied from 51 to 100, and mean ages ranged from 43 to 67 years. The number of participants included in each study varied between 18 and 100, and retention rates varied between 63% and 100% (Table 1). Some studies reported only the participants that completed the study protocol, but others described the initial selection procedure, analyzed the dropouts, and applied an intention to treat analysis.
The exercise and control conditions are presented in Table 2. Two to five different interventions were compared in the various studies. Some studies included untreated control groups; others included control groups that were given attention only, ROM exercise, or relaxation. Two studies compared different types of aerobic exercise (23, 28), and 2 studies compared static with dynamic strengthening exercises (24, 41). The length of the interventions varied between 1 and 24 months. Exercise was carried out at minimum twice weekly with a maximum of twice daily in 1 study. The aerobic exercise was often progressive, the intensity varied between 50% and 85% of maximum, which was defined as the percentage of either measured maximum heart rate (23, 26, 28), age-predicted maximum heart rate (37, 45, 46), or predicted maximum pulse increment (33). The exercise loads of most of the strengthening exercise programs were progressively adjusted, starting at 30–50% and increasing to 80% of maximum, which was defined as the percentage of either 1 repetition maximum (32, 42, 47), one maximum voluntary contraction (41, 45), maximum speed (44), or as maximal subjective exertion (40). A couple of studies reported their functional exercise program in appendices or references (24, 48) and 1 study gave rather limited information on exercise dose and content (29).
|Author, year (reference)||Exercise type||Length, months||Frequency, times/week||Duration, minutes||Intensity, % of maximum||Load, % of maximum|
|Minor et al 1989 (23)||a) Aerobic aquatics||3||3||60||60–80||NA|
|b) Progressive aerobic walking||3||3||60||60–80||NA|
|Ekdahl et al 1990 (24)||a) Dynamic lower limb exercise||1.5||7||60||NR||NR|
|b) Static lower limb exercise||1.5||7||60||NR||NR|
|Baslund et al 1993 (26)||a) Progressive interval bicycling||2||4–5||25||∼85||NA|
|Hansen et al 1993 (28)||a) Self training||24||≥3||≤90||NR||NR|
|b) a + weekly PT training||24||≥3||≤90||NR/70||NR|
|c) a + weekly group training||24||≥3||≤90||NR/70||NR|
|d) c + weekly pool exercise||24||≥3||≤90||NR||NR|
|Hoenig et al 1993 (29)||a) Hand ROM||3||14||10–20||NA||NA|
|b) Hand resistive training||3||14||10–20||NA||NR|
|c) Hand ROM + resistive training||3||14||10–20||NA||NR|
|Häkkinen et al 1994 (32)||a) Progressive dynamic strength training of major muscle groups||6||2–3||NR||NA||40–80|
|Lyngberg et al 1994 (33)||a) Progressive interval training: bicycling, heel lifting, step climbing, stretching||3||2||45||50/70||NA|
|Stenström et al 1996 (48)||a) Upper and lower limb strengthening, aerobic walking||3||5||30||NR||NR|
|Stenström et al 1997 (39)||a) Upper and lower limb strengthening, aerobic walking||3||5||30||NR||NR|
|Van den Ende et al 1996 (37)||a) Progressive interval and bicycle training||3||3||60||70–85||NR|
|b) ROM and non–weight-bearing exercise||3||2||60||NA||no|
|c) As b but individually supervised||3||2||60||NA||no|
|d) As b but written instructions only||3||2||15||NA||no|
|Komatireddy et al 1997 (40)||a) Progressive resistive muscle training (circuit with home video)||3||≥3||20–27||NA||30–40|
|Boström et al 1998 (41)||a) Progressive dynamic shoulder exercise||2.5||3||40–60||NA||30|
|b) Progressive static shoulder exercise||2.5||3||40–60||NA||30|
|Häkkinen et al 1999 (42)||a) Dynamic whole-body strengthening home exercises||12||2||45||NA||50–60|
|Häkkinen et al 2001 (47)||a) Dynamic whole-body strengthening||24||2||45||NA||50–70|
|McMeeken et al 1999 (44)||a) Knee extensor and flexor muscle training||1.5||2–3||45||NA||70|
|Van den Ende et al 2000 (45)||a) Progressive dynamic and isometric strength training||1 (mean)||5||NR||NA||70|
|+ bicycle training||3||15||60||NA|
|+ ROM and isometric strength training||5||NR||NA||NR|
|b) ROM + isometric strength training||1 (mean)||5||NR||NA||NR|
|Westby et al 2000 (46)||a) Aerobic exercise||12||3||45–60||60–75||NA|
The results of the exercise interventions are displayed in Table 3. Five studies supported evidence for improvements of aerobic capacity (23, 24, 26, 37, 46), whereas 3 did not (28, 33, 40). Eight studies reported, in 9 papers, at least some kind of improved muscle function after strengthening exercises (24, 29, 32, 33, 37, 42, 44, 45, 47), whereas 6 did not (23, 28, 39–41, 48). Pain remained unchanged in all but 3 studies (40, 44, 45), ADL in all but 3 (42, 44, 47), and health-related quality of life in all but 1 (23). Some indication of reduced disease activity was reported in 7 studies (24, 29, 37, 40, 42, 45, 47), while it remained unchanged in 9 (23, 26, 28, 32, 33, 39, 41, 46, 48). Joint destruction remained unchanged in all 3 studies including such evaluations (28, 38, 47).
|Author, year (reference)||Aerobic capacity||Muscle strength||ROM||Performance tests||Pain||ADL||HRQoL||Disease activity|
|Minor et al, 1989 (23)||+||0||0||+||0||NR||+||0|
|Ekdahl et al, 1990 (24)||+||+||0||+||0||0||NR||+|
|Baslund et al, 1993 (26)||+||NR||NR||NR||NR||NR||NR||0|
|Hansen et al, 1993 (28)||0||0||NR||0||0||0||NR||0|
|Hoenig et al, 1993 (29)||NR||+||0||+||NR||NR||NR||+|
|Häkkinen et al, 1994 (32)||NR||+||NR||NR||NR||0||NR||0|
|Lyngberg et al, 1994 (33)||0||+||NR||0||0||0||NR||0|
|Stenström et al, 1997, 1996 (39, 48)||NR||−||0||0||NR||NR||0||0|
|Van den Ende et al, 1996 (37)||+||+||+||0||0||0||0||+|
|Komatireddy et al, 1997 (40)||0||0||NR||+||+||0||0||+|
|Boström et al, 1998 (41)||NR||0||0||NR||0||0||0||0|
|Häkkinen et al, 1999, 2001 (42, 47)||NR||+||NR||0||0||+||NR||+|
|McMeeken et al, 1999 (44)||NR||+||NR||+||+||+||NR||NR|
|Van den Ende et al, 2000 (45)||NR||+||0||0||−/0†||0||NR||+|
|Westby et al, 2000 (46)||+||NR||NR||NR||NR||0||NR||0|
Results of 2 followup studies were reported separately (39, 47) and are included in Table 3. Another 5 studies, which had indicated positive immediate exercise results, also included followup results. In 1 of them, the participants with RA were not evaluated separately at followup (23), another simply stated that “follow-up evaluation at 3 months after the exercise period revealed no statistically significant changes in any of the intervention groups” (29). In 1 study of exercise in active RA, some improvements in muscle strength remained 12 weeks after discharge from the hospital (45). While Ekdahl et al reported that 3-month followup results were basically similar to those directly after the exercise (24); van den Ende et al found that the gain in physical capacity had disappeared 12 weeks after discontinuation of exercise (37).
The results of the present literature review suggest that there is a growing body of well-designed research to support the benefit of aerobic and strengthening exercise in RA. Early studies (18–20, 49, 50), although not meeting the inclusion criteria of our review, have been of great importance in introducing the idea that individuals with RA could perform exercise safely and effectively and in forming a basis for future questions and RCTs.
Patients investigated in the studies included in the present review were generally functionally independent, as indicated by ARA functional class, and in a stable phase of their disease. Some support for the safety and benefit of exercise for patients with recent diagnosis, active disease, or fragile bones also was found. Moreover, a recent case study suggests that some individuals with very severe rheumatic diseases may be motivated to exercise, and with professional guidance may benefit emotionally as well as physically (51).
Exercise in healthy individuals produces specific and predictable effects. However, in individuals with rheumatic diseases, a nonspecific overflow may occur because more general adaptations to training take place in a deconditioned state. This more generalized effect, even from specifically targeted exercise regimens, was reported in a couple of the RA exercise studies (24, 37). Studies presented here were therefore not categorized as aerobic or strengthening. More detailed information describing the actual exercise prescription (intensity, frequency, duration, and mode) would have been helpful in understanding the content and dose of the exercise programs. It seems, though, that the programs resulting in benefits for individuals with RA were mainly in accordance with the recommended quantity and quality of exercise for the development and maintenance of cardiorespiratory and muscular fitness for healthy adults (52). Most exercise programs were performed in a clinical environment, but the results of 2 studies suggest that individuals with RA may also benefit from professionally supported community-based or home-based exercise (24, 47).
A number of studies reported using 1 repetition maximum or 1 maximal voluntary contraction as a method to determine strength (32, 41, 42, 45, 47). There is some evidence, however, that this may not be a valid or appropriate method to describe strength in RA (24). Valid, well-described, and standardized methods to assess strength to determine initial exercise loads and to measure progress are required.
The statistical power of most studies included in this review was probably limited due to few participants or many subgroups. Still, results suggest that aerobic capacity and strength improved without negative effects on pain and disease activity. The studies failing to show exercise benefits might either have been using too low an exercise load (39–41, 48) or had poor statistical power (28, 33). The specific effects of exercise on ADL, health related quality of life, and disease progression still remain unclear. Regarding ADL, it might be that the Stanford Health Assessment Questionnaire (53), which frequently was used in the studies reviewed, is not sensitive enough to capture exercise effects additional to medication, for which this measure is known to be sensitive.
This review has covered clinical trials of exercise in RA, but has not included the applied research related to the effects of exercise on metabolic, physiologic, or mechanical factors. This latter field of research, which is crucial to our understanding of the mechanisms through which exercise supplies benefit, provides tantalizing yet scanty results to date. It would be tempting to speculate on how aerobic exercise seems to reduce joint swelling in RA or why strengthening exercise reduces pain reports. Rheumatology-specific research suggests the possibility of exercise-related changes in synovial circulation (54), immune response and inflammatory factors (55–57), and neuropeptide levels (58). General exercise research also may pertain to this population in terms of the benefits of neuromuscular learning and improved elasticity and strength of periarticular structures. A thorough review of the results to date is certainly needed, but is beyond the scope of this article.
Evidence-based recommendations for aerobic exercise in RA.
In light of the studies reviewed in this article, the goal for the intensity level of the aerobic exercise should be moderate to hard (i.e., 60–85% of maximum heart rate), and exercise be performed 3 times weekly for a duration of 30–60 minutes. Exercise can be performed in water or be land-based and carried out in a clinical environment with continuous supervision or in a community-based environment with professional support. Typical activities might include aquatics, walking, cycling, or aerobic class participation. Progressive adjustment of the intensity is recommended
Evidence-based recommendations for strengthening exercises in RA.
Information derived from the reviewed studies indicates that the goal for the load level of strengthening exercises should be moderate to hard (i.e., 50–80% of a maximal voluntary contraction), and exercise be performed 2 to 3 times a week. Exercises may be static or dynamic and performed against body weight or with various types of equipment, including resistance training equipment, pulley apparatus, dumbbells, or elastic bands. Progressive adjustment of the load is recommended and exercises may be performed in a supervised clinical environment or at home with professional support
Suggestions for future research.
For the most part, research investigating exercise safety and benefit for people with RA is of good quality and has provided useful information to guide future research. Based on the present review, some suggestions for future study are outlined below.
Pain, swelling, and willingness to exert a maximal voluntary contraction influence measurements of strength and endurance. These factors may vary considerably in RA, be unpredictable over time, and confound the results of muscle testing. Establishment of valid and stable measurement procedures of strength and endurance are needed to obtain true baseline values for exercise prescription, appropriate exercise progression, and response assessments.
Major difficulties with pain, fatigue, loss of function, and diminished quality of life are well-known problems for many people with RA. To date, exercise research findings are scanty in these important areas. Further identification and evaluation of disease-specific, client-centered measures that may be modifiable with exercise are needed.
Response criteria have been published for drug trials in which laboratory and radiographic information are available. However, these methods are rarely applicable to the majority of exercise studies, are not particularly germane to the goals of an exercise or physical activity intervention, and are not client centered. Now that general safety and benefit of exercise have been established from group comparisons in short-term clinical trials, it is important to begin to document exercise response in terms of characteristics of individuals. For example, individuals with more serious joint involvement may respond more favorably to a moderate rather than an intense training program. Appropriate response criteria will need to be determined to allow this area of inquiry to proceed.
Evidence to date supports the clinical usefulness of exercise in RA to improve both muscle function and fitness. It is now time to add to our clinical knowledge a better understanding of the underlying mechanisms by which exercise supplies benefit. We need more information regarding the physiologic, metabolic, mechanical, and behavioral avenues through which exercise improves function, as well as a greater knowledge of the mechanisms by which exercise reduces pain and joint effusion. The evidence will come from applied research that includes various types of exercise in the test environment and from clinical trials that ask questions and measure results at the cellular level as well as at levels of body function and activity.
Self management education as well as exercise instruction will need to be tested in settings other than the controlled clinical and research settings that have been used to date. Long-term followup of those who continue and do not continue to exercise is needed, as well as study of the effectiveness of methods to promote appropriate health behaviors. Development and evaluation of programs and strategies to promote long-term, self directed physical activity in community-based settings are thus other important areas for future research.
- 43Serum levels of hyaluronan, antigenic keratin sulfalate, matrix metalloproteinase 3, and tissue inhibitor of metalloproteinase 1 change predictably in rheumatoid arthritis patients who have begun activity after a night of bed rest. Arthritis Rheum 1999; 42: 1861–9., , , , , , et al.
- 44The effects of knee extensor and flexor muscle training on the timed-up-and-go test in individuals with rheumatoid arthritis. Physiother Res Int 1999; 4: 55–67., , , .
- 47A randomized two-year study of the effects of dynamic strength training on muscle strength, disease activity, functional capacity, and bone mineral density in early rheumatoid arthritis. Arthritis Rheum 2001; 44: 515–22., , , .
- 49Effects of isometric exercise on the quadriceps muscle in patients with rheumatoid arthritis. Arch Phys Med Rehabil 1966; 47: 737–41., .