effectiveness of nonpharmacological and nonsurgical interventions for patients with rheumatoid arthritis an overview of systematic reviews.pdf

doi: 10.2522/ptj.20070039Originally published online September 25, 2007

2007; 87:1697-1715.PHYS THER. H Moe, Espen A Haavardsholm and Kre Birger HagenAnne Christie, Gro Jamtvedt, Kristin Thuve Dahm, RikkeReviewsRheumatoid Arthritis: An Overview of SystematicNonsurgical Interventions for Patients With Effectiveness of Nonpharmacological and

http://ptjournal.apta.org/content/87/12/1697found online at: The online version of this article, along with updated information and services, can be

Collections

Systematic Reviews/Meta-analyses Rheumatoid Arthritis

Evidence-Based Practice in the following collection(s): This article, along with others on similar topics, appears

e-Letters

"Responses" in the online version of this article. "Submit a response" in the right-hand menu under

or click onhere To submit an e-Letter on this article, click

E-mail alerts to receive free e-mail alerts hereSign up

by guest on May 1, 2012http://ptjournal.apta.org/Downloaded from

Effectiveness of Nonpharmacologicaland Nonsurgical Interventions forPatients With Rheumatoid Arthritis:An Overview of Systematic ReviewsAnne Christie, Gro Jamtvedt, Kristin Thuve Dahm, Rikke H Moe,Espen A Haavardsholm, Kre Birger Hagen

Conclusions based on systematic reviews of randomized controlled trials are consid-ered to provide the highest level of evidence about the effectiveness of an interven-tion. This overview summarizes the available evidence from systematic reviews onthe effects of nonpharmacological and nonsurgical interventions for rheumatoidarthritis (RA). Systematic reviews of studies of patients with RA (aged 18 years)published between 2000 and 2007 were identified by comprehensive literaturesearches. Methodological quality was independently assessed by 2 authors, and thequality of evidence was summarized by explicit methods. Pain, function, and patientglobal assessment were considered primary outcomes of interest. Twenty-eight sys-tematic reviews were included in this overview. High-quality evidence was found forbeneficial effects of joint protection and patient education, moderate-quality evi-dence was found for beneficial effects of herbal therapy (gamma-linolenic acid) andlow-level laser therapy, and low-quality evidence was found for the effectiveness ofthe other interventions. The quality of evidence for the effectiveness of most non-pharmacological and nonsurgical interventions in RA is moderate to low.

A Christie, PT, MSc, is ResearchFellow, National Resource Centrefor Rehabilitation in Rheumatol-ogy, Diakonhjemmet Hospital, POBox 23 Vindern, 0319 Oslo, Nor-way. Address all correspondenceto Ms Christie at: [email protected].

G Jamtvedt, PT, MPH, is Re-searcher, Norwegian KnowledgeCentre for the Health Services,Oslo, Norway.

KT Dahm, PT, MSc, is ResearchAssistant, Norwegian KnowledgeCentre for the Health Services.

RH Moe, PT, is Research Fellow,National Resource Centre forRehabilitation in Rheumatology,Diakonhjemmet Hospital.

EA Haavardsholm, MD, is ResearchFellow, Department of Rheuma-tology, Diakonhjemmet Hospital.

KB Hagen, PT, PhD, is Researcher,National Resource Centre forRehabilitation in Rheumatology,Diakonhjemmet Hospital.

[Christie A, Jamtvedt G, Dahm KT,et al. Effectiveness of nonpharma-cological and nonsurgical inter-ventions for rheumatoid arthritis:an overview of systematic reviews.Phys Ther. 2007;87:16971715.]

2007 American Physical TherapyAssociation

CARE IVConference Series

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

December 2007 Volume 87 Number 12 Physical Therapy f 1697 by guest on May 1, 2012http://ptjournal.apta.org/Downloaded from

Rheumatoid arthritis (RA) is achronic, systemic inflammatorydisease of unknown etiology. Itis present in 0.5% to 1% of the gen-eral population, twice as often inwomen, and the age at disease onsetis mainly between 45 and 65 years.1

The clinical picture of RA is charac-terized by pain, fatigue, disability,and reduced quality of life. Thecourse of the disease is often unpre-dictable, and the symptoms may varyfrom day to day.

The main goals of treatment for RAare to prevent or control joint dam-age, prevent loss of function, anddecease pain.2 Despite substantialprogress in the pharmacological andsurgical interventions over the lastdecade, many patients with RA willstill experience disability, pain,psychological distress, fatigue, andpoor quality of life.3

Besides pharmacological and surgi-cal interventions, conventional ther-apies such as physical therapy,occupational therapy, and compre-hensive rehabilitation and self-management programs are commonlyand frequently used interventions. Notsurprisingly, given the chronic andpersistent nature of the disease, pa-tients with RA tend to be particularlyhigh users of complementary and al-ternative medicine (CAM) thera-pies.3 Complementary medicine isused together with conventionalmedicine, whereas alternative medi-cine is used in place of conventionalmedicine. Research indicates thatpeople with RA use a broad range ofCAM therapies, such as dietary sup-plements (herbs and vitamins),movement therapies (yoga and taichi) and manual therapy, homeopa-thy, and acupuncture.

Decisions on the provision and re-imbursement of health care are in-creasingly based on the availableevidence. Thus, purchasing organiza-tions and policymakers in health

care are in need of information onthe effectiveness of interventions.Similarly, patients, health care pro-fessionals, and researchers are inneed of this information to improveself-management strategies, to im-prove clinical practice, and to setpriorities for research, respectively.Conclusions based on a systematicreview of randomized controlled tri-als (RCTs) are considered to providethe highest level of evidence aboutthe effectiveness of an intervention.

While systematic reviews summarizethe effectiveness of a specific treat-ment for a specific condition, anoverview of overviews (sometimescalled an umbrella review) typi-cally summarizes the evidence fromseveral systematic reviews on differ-ent treatment options for the samecondition. Because the number ofsystematic reviews is rapidly increas-ing, there might be a need to providepatients and health care providerswith synthesized and easily accessi-ble information on different treat-ment options for a particularcondition.

The aim of this overview is to sum-marize the available evidence fromsystematic reviews on the effect ofnonpharmacological and nonsurgi-cal interventions for patients withRA.

MethodsSystematic reviews were consideredif they were published from January2000 to January 2007 and had theprimary aim of investigating the ef-fects of nonpharmacological andnonsurgical interventions for peoplewith RA (aged18 years). More spe-cifically, the following inclusion cri-teria were used:

Participants: People with RA ac-cording to the American College ofRheumatology criteria4 or other ac-ceptable diagnostic criteria. Re-views including participants with

various rheumatic diagnoses wereaccepted only if results for RAcould be extracted separately.

Interventions: All types of nonphar-macological and nonsurgical inter-ventions. Excluded were interven-tions such as gene therapy, all typesof invasive interventions (eg, injec-tions, arthroscopy), therapeuticapheresis, or interventions relatedto pharmacological or surgical in-terventions (ie, therapeutic exer-cises after total joint replacement).

Outcomes: For the purpose of thisoverview, the primary outcomemeasures were function, pain, andpatient global assessment. The con-cept of function is based on theInternational Classification ofFunctioning, Disability andHealth (ICF),5 where function isan umbrella term for body function,body structure, activities, and par-ticipation. As secondary outcomevariables, we considered the rest ofthe outcomes in the preliminarycore set recommended by the In-ternational League of Associationsfor Rheumatology (ILAR)/OutcomeMeasures for Arthritis Clinical Trials(OMERACT).6

Search StrategyThe following databases weresearched from January 2000 to Janu-ary 2007: MEDLINE, CINAHL, AMED,EMBASE, PsycINFO, The CochraneLibrary, and PEDro.

The search strategy was formulatedin Ovid (MEDLINE, CINAHL,EMBASE, and AMED) in cooperationwith a medical librarian to make itapplicable to all the databases. Abroad computerized search strategywas developed (Appendix 1).

Retrieved hits were assessed by oneof the authors (AC), who screenedthe titles and abstracts to identify rel-evant studies. If there was doubtabout a studys relevance, one of theother authors (KBH) was consulted.

Effectiveness of Nonpharmacological and Nonsurgical Interventions for RA

1698 f Physical Therapy Volume 87 Number 12 December 2007 by guest on May 1, 2012http://ptjournal.apta.org/Downloaded from

Relevant full-text articles were readby 2 authors (AC, KBH).

Assessment ofMethodological QualityThe methodological quality of in-cluded reviews was independentlyassessed by 2 reviewers (AC, KBH)using a modified version of a previ-ously validated checklist consistingof 9 criteria7 (Appendix 2). Disagree-ment was resolved by discussion.Based on a summary of the 9 criteria,an overall scientific quality was ap-plied to each review, as follows:minor limitations (at least 7 of thecriteria were met), moderate limita-tions (at least 4 of the criteria weremet), and major limitations (fewerthan 4 of the criteria were met). Re-views with major limitations wereexcluded.

Data Extraction and SynthesisData were extracted by one of theauthors (AC). If doubt occurred, oneof the other authors (KBH) was con-sulted. The following criteria wereapplied when data on effects wereextracted:

Adequate quantitative pooling ofdata in reviews was regarded asmore valid than a qualitative datasynthesis approach.

If no direct comparisons betweentreatments were undertaken or noquantitative pooling of data wasdone, the results are reported asno quantitative pooling, and theauthors statements were reported.

When we found that the resultswere reported inconsistently in dif-ferent sections of a review, the ef-fects were extracted from the mainresults section.

If possible, the 3 primary outcomeswere collected and presented inthe result tables. The secondaryoutcomes were collected if the pri-mary outcomes were not reported.A maximum of 3 outcomes arereported.

Finally, principles from Grading ofRecommendations Assessment, De-velopment, and Evaluation (GRADE)were used to assess an overall qualityof evidence for each interventionand outcome.8 The quality of evi-dence indicates the extent to whicha person can be confident that theestimate of effect is correct. Basedon judgments considering quality ofprimary studies, design of primarystudies, consistency (similarity of es-timates of effect across studies), anddirectness (the extent to which peo-ple, interventions, and outcomemeasures were similar to those ofinterest), the evidence for each inter-vention was classified as high,moderate, low, or no evidencefrom systematic reviews. The defi-nitions are listed in Table 1. In thesummary of findings (Tab. 2), thefollowing statements were used toindicate direction of effect: improves(function) or reduces (pain) (ie,beneficial effects), no difference,and unclear (inconsistent or unclearevidence of effect).

ResultsThe literature search identified 1,189references, which were first exam-ined on the basis of titles and ab-stracts. Of these, 1,078 referenceswere clearly not relevant, and 111references were retrieved in full

text. Eighty-three reviews were ex-cluded: 41 because of major limita-tions (Appendix 3), 6 because of du-plicate publications, 16 because ofmixed populations, 9 because of useof nonrelevant interventions, and 11were not reviews or mixed reviewand single studies. Twenty-eight re-views were included in this over-view (Fig. 1).

AcupunctureTwo reviews9,10 reported the effectof acupuncture (Tab. 3). The re-views described the intervention asa kind of herbal acupuncture (beevenom acupuncture)9(p79) and as atechnique based on Chinese medicalpractice whereby needles are in-serted into specific exterior body lo-cations to relieve pain and for othertherapeutic purposes.10 Study pop-ulations were adult patients withclassic or definite rheumatoidarthritis.

The methodological quality of theprimary studies either was not as-sessed or was of low to moderatequality. One review9 included 1 RCTand 2 uncontrolled studies and re-ported a significant decrease in pain,but did not report on function orpatient global assessment. Casimiroet al10 found conflicting results re-garding pain, but no significant im-

Table 1.Quality of Evidence

Level Based on

High-quality evidence One or more updated, high-quality systematicreviews that are based on at least 2 high-qualityprimary studies with consistent results

Moderate-quality evidence One or more updated systematic reviews of high ormoderate quality Based on at least 1 high-quality primary study Based on at least 2 primary studies of

moderate quality with consistent results

Low-quality evidence One or more systematic reviews of variable quality Based on primary studies of moderate quality Based on inconsistent results in the reviews Based on inconsistent results in primary

studies



provements in function or patientglobal assessment. Both reviews con-cluded that there is little evidencethat acupuncture relieves RA symp-toms, but underscored that this con-clusion is limited by methodologicalflaws. We find this conclusion to bereasonable and conclude that the ef-fect of acupuncture is unclear(Tab. 2). One of the reviews re-ported no harmful side effects, whilethe other concluded that a greaterunderstanding of the risks and bene-fits of bee venom acupuncture isneeded.

BalneotherapyTwo reviews11,12 included bal-neotherapy, reporting on interven-tions as bathing in water containingminerals (added or natural)11 orany type of balneotherapy (mudpacks, sulphur baths, Dead Seabaths).12 Patients were included ifthey had clinically confirmed RA11

or only RA.12 Again, the method-ological quality of the primary stud-ies was low to moderate. One re-view12 provided quantitative poolingof results (Tab. 3). The pooled re-sults (mudpacks versus control)

were based on 2 studies with smallsample sizes and showed no statisti-cally significant improvements inpain or patient global assessment.Results from single primary studiesshowed conflicting evidence relatedto pain and function (grip strength[force-generating capacity]). Both re-views concluded that firm conclu-sions on the effectiveness of bal-neotherapy for people with RAcannot be drawn, mainly because ofthe heterogeneity of the interven-tions and the poor methodologicalquality of the included studies. We

Table 2.Summary of Findings

Intervention Comparison Results Quality ofEvidence

Acupuncture Control/placebo Unclear Low

Balneotherapy Different types of balneotherapy/other interventions/placebo/no intervention

Unclear Low

Diets Control/usual diet Reduces pain Low

Electrical stimulation Control Unclear Low

Herbal therapy Placebo/alternative herbal intervention Reduces pain and patient globalassessment

Moderate

Occupational therapy

Advice/instruction aboutassistive devices

Usual care/no control Unclear Low

Comprehensiveoccupational therapy

No intervention Improves function, nodifference in pain

Low

Joint protection Alternative intervention/no intervention Improves function, nodifference in pain

High

Provision of splints No intervention/control/different types of splints Unclear Low

Training of motorfunction

No intervention/alternative intervention Unclear Low

Hand/foot orthosis Placebo/no intervention/other intervention Unclear Low

Patient educationalintervention

No intervention/usual care/other educationalintervention/waiting list controls

Improves function and patientglobal assessment

High

Exercise No intervention/alternative intervention/differenttypes of exercises

Reduces pain and improvesfunction, no difference inpatient global assessment

Low

Low-level laser therapy Alternative intervention/placebo Reduces pain and improvesfunction

Moderate

Therapeutic ultrasound Placebo/alternative intervention Reduces pain and improvesfunction

Low

Thermotherapy Placebo/no intervention/alternative interventions Unclear Low

Transcutaneous electricalnerve stimulation

Placebo/alternative interventions Unclear Low



conclude, therefore, that the effectof balneotherapy is unclear (Tab. 2).Possible harmful aspects of bal-neotherapy, such as cardiac prob-lems or risk of bacterial contamina-tion, were not reported in any of thereviews.

DietsOne review13 reported on the effectof fasting followed by a vegetariandiet for at least 3 months (Tab. 4).Patients included were diagnosedwith RA. The primary studies wereof mixed quality, but the pooling ofdata from 2 RCTs showed significantimprovement in pain. However,based on one review of moderatequality, we conclude that there islow-quality evidence that fasting fol-lowed by a vegetarian diet (3months) reduces pain (Tab. 2). Dietsmight have a negative effect on nu-tritional status, but the review didnot report on this issue.

Electrical Stimulation (ES)One review14 based on 1 RCT of lowquality reported the outcome of ES(Tab. 4). The review defined the in-tervention as electrical stimulationapplied to the motor point of a mus-cle and is used to recruit motor unitsthat are not activated at a given mo-ment by voluntary recruitment. Thepatient population (age3075years) had RA affecting the meta-carpophalangeal joint. The authorsconcluded that the evidence for theuse of ES to improve muscle strengthand resistance to fatigue is limited.The study included 6 participants inthe intervention group and 3 partic-ipants in the control group. The Ot-tawa Panel excluded the RCT be-cause: Evidence with acceptableresearch design, interventions,group comparisons, or outcomescould not be identified.15(p1026)

Thus, we conclude that the effect ofES is unclear (Tab. 2). The review didnot comment on safety and possibleharmful effects.

Herbal TherapyEffect of herbal therapy was re-ported in 3 reviews1618 (Tab. 4),defining the interventions asAyurvedic medicines as usuallycomplex mixtures of multiple plantsadministered orally,8(p705) any wholeplant extract except homeotherapy,aroma therapy or any preparation ofsynthetic origin or consisting only ofplant derivates,16(p2) or herbal prep-arations administered orally or topi-cally for RA.17(p652) Patient popula-tions included were RA patients17,18

or all persons diagnosed with RA.16

All reviews found conflicting evi-dence for the effect of herbal ther-apy on pain and function. When re-porting on the effect of gamma-linolenic acid (GLA) specifically, 2

reviews16,17 found statistically sig-nificant improvement in pain andpatient global assessment based onpooled results from 3 RCTs ofmoderate quality, but they empha-sized that further studies areneeded to examine the efficacy,safety, and potential drug interac-tions. We conclude, therefore, thatthere is moderate-quality evidencethat herbal therapy (GLA) reducespain and improves patient global as-sessment, while the evidence forother herbals is unclear (Tab. 2). Fre-quently reported adverse eventsfrom ayurvedic medicines includedanorexia, nausea, diarrhea, constipa-tion, and abdominal pain, but in gen-eral the reviews concluded that theadverse effects reported were fewand minor. Thus, the reviews con-

Reviews clearly

not relevant

(n=1,078)

Potentially relevant

reviews identified and

screened for retrieval

( n=1,189 )

Reviews

retrieved in

full text

(n=111) Reviews excluded:

low quality (n=41),* duplicate publications/no review/ mixed patient populations/notrelevant (n=42)

Reviews with usuable

information by outcome

(n=28)

Figure.Selection process of eligible reviews from all identified citations. *Excluded reviews listedin Appendix 3.



cluded that further research isneeded, not only research on the ef-ficacy of herbal interventions butalso research on safety and potentialdrug interactions.

Occupational TherapyInterventionsOne review,19 including 38 studies,explored several occupational ther-apy intervention categories for RA(Tab. 5). Occupational therapy inter-ventions either were classified into6 specific intervention categories

(training of motor function, trainingof skills, instruction on joint protec-tion and energy conservation, coun-seling, advice and instruction in theuse of assistive devices, or provisionof splints and hand orthoses) or wereregarded as comprehensive occupa-tional therapy (when all 6 interven-tions were part of the evaluated oc-cupational therapy treatment). Thereview found no studies concerningthe interventions training of skillsand counseling. Studies with pa-tients who fulfilled a clinical diagno-

sis of RA were included. Each inter-vention category is presentedseparately below.

Advice and instruction in the useof assistive devices. Two studies(non-RCTs) evaluated advice and in-struction in the use of assistive de-vices versus alternative treatment.Only one study reported function asan outcome variable and found nostatistically significant improvement.The 2 studies reported conflictingresults regarding pain. Based on the

Table 3.Description of Included Reviews: Acupuncture/Balneotherapya

Reference No. of IncludedStudies andParticipants

Methodological Qualityof Review and PrimaryStudies

Results

An overview of beevenom acupuncture inthe treatment ofarthritis (Lee et al)9

1 RCT2 CTs(152 participants)

ReviewModerate limitationsPrimary studiesMean quality score (1 RCT)4

(Jadad scale: 05)2 CTs: not assessed

No quantitative poolingDue to paucity and methodological flaws in

the existing clinical studies, it is prematureto draw any firm conclusions at this time.However, the effectiveness of bee venomacupuncture for arthritis is likely to be apromising area of future research.

Acupuncture andelectroacupuncture forthe treatment ofrheumatoid arthritis(Casimiro et al)10

2 RCTs (84 participants) ReviewMinor limitationsPrimary studiesMean quality score3 (Jadad

scale: 05)

No quantitative poolingWith a silver level of evidence, we conclude

that there is little evidence thatacupuncture relieves RA symptoms.

Balneotherapy forrheumatoid arthritisand osteoarthritis(Verhagen et al)11

6 RCTs (355 participants) ReviewMinor limitationsPrimary studiesMean quality score4.8

(Delphi list: 09)

No quantitative poolingOne cannot ignore the positive findings

reported in most trials. However, thescientific evidence is insufficient becauseof the poor methodological quality, theabsence of an adequate statistical analysis,and the absence, for the patient, of mostessential outcome measures. Therefore, thenoted positive findings should be viewedwith caution. Because of themethodological flaws, an answer about theapparent effectiveness of balneotherapycannot be provided at this moment. Alarge, methodological sound trial isneeded.

Efficacy of balneotherapyfor rheumatoid arthritis:a meta-analysis(Brosseau et al)12

7 RCTs (374 participants) ReviewMinor limitationsPrimary studiesMean quality score2.4 (Jadad

scale: 05)

PainSwollen/tender joints (Ritchie Index Scale):

WMD7.07, 95% CI11.98 to 2.19 (2RCTs)Patient global assessmentVAS: WMD0.2, 95% CI1.08 to 1.8 (2RCTs)

a RCTrandomized controlled trial, CTclinical trial, RArheumatoid arthritis, WMDweighted mean difference, CIconfidence interval, VASvisual analogscale.



conflicting results and the low qual-ity of the primary studies, we con-clude that the effect of advice andinstruction in the use of assistive de-vices is unclear (Tab. 2). No safety orside effects were assessed in the in-cluded studies.

Comprehensive occupational ther-apy. Four studies evaluated com-prehensive occupational therapyversus no treatment or alternativetreatment. Comprehensive occupa-

tional therapywas defined as whenall 6 interventions (training of motorfunction, training of skills, instruc-tion on joint protection and energyconservation, counseling, advice andinstruction in the use of assistive de-vices, and provision of splints andhand orthoses) were part of the eval-uated occupational therapy treat-ment. One RCT of high quality re-ported a statistically significantpositive effect on functional ability,whereas 3 low-quality studies re-

ported no effect. No statistically sig-nificant results were found for pain.Based on conflicting results, we con-clude that there is low-quality evi-dence that comprehensive occupa-tional therapy improves function andmakes no difference in pain (Tab. 2).

Instruction on joint protectionand energy conservation. Eightstudies evaluated instruction in jointprotection versus no treatment or al-ternative treatment. Two RCTs of

Table 4.Description of Included Reviews: Diets/Electrical Stimulation/Herbal Therapya

Reference No. of Included Studiesand Participants

MethodologicalQuality of Reviewand PrimaryStudies

Results

Fasting followed byvegetarian diet inpatients withrheumatoid arthritis: asystematic review(Muller et al)13

2 RCTs2 CCTs(143 participants)

ReviewModerate limitationsPrimary studiesNot reported

PainEffect size0.58, r.28, P.01 (2 RCTs)

Electrical stimulation forthe treatment ofrheumatoid arthritis(Brosseau et al)14

1 RCT (15 participants) ReviewMinor limitationsPrimary studiesMean quality score2

(Jadad scale: 05)

No quantitative poolingES was shown to have a clinically beneficial effect

on grip strength and fatigue resistance for RApatients with muscle atrophy of the hand.However, these conclusions are limited by thelow methodological quality of the trial included.More well-designed studies are therefore neededto provide further evidence of the benefits of ESin the management of RA.

Herbal therapy fortreating rheumatoidarthritis (Little andParsons)16

11 RCTs (398 participants) ReviewMinor limitationsPrimary studiesMean quality

score3.4 (Jadadscale: 05)

PainVAS: WMD32.83, 95% CI56.25 to 9.42 (3

RCTs)Pain scale (04): WMD25.88, 95% CI46.73

to 5.02 (3 RCTs)Patient global assessmentScale (04): WMD20.87, 95% CI39.43 to2.31 (3 RCTs)

Herbal medicines for thetreatment ofrheumatoid arthritis: asystematic review(Soeken et al)17

14 RCTs (493 participants) ReviewModerate limitationsPrimary studiesMean quality


PainVAS: WMD0.76, 95% CI0.37 to 1.15 (3 RCTs)

Ayurvedic medicine forrheumatoid arthritis: asystematic review (Parkand Ernst)18

7 RCTs (508 participants) ReviewMinor limitationsPrimary studiesMean quality


No quantitative poolingThere is a paucity of RCTs of Ayurvedic

medicines for RA. The existing RCTs fail toshow convincingly that such treatments areeffective therapeutic options for RA.

a RCTrandomized controlled trial, CCTcontrolled clinical trial, ESelectrical stimulation, RArheumatoid arthritis, VASvisual analog scale,WMDweighted mean difference, CIconfidence interval.



high quality found statistically signif-icant improvements in functionalability, and the findings were sup-ported by 2 studies of lower quality.Two RCTs of high quality found nostatistically significant improvementin pain. There was insufficient infor-mation about the results from thelow-quality studies. The authors con-cluded that there is strong evidencethat instruction on joint protectionleads to an improvement of func-tional ability. We support this con-clusion and conclude that there ishigh-quality evidence for a positiveeffect on function, but no differencein pain (Tab. 2). Only one of theincluded studies reported decreasesin grip strength and range of motion(ROM) as possible effects, but the

authors questioned whether this wasdue to improved joint protection be-havior or a determinant of increasedjoint protection behavior.

Provision of splints and handorthoses. Sixteen studies relatedto provision of splints (hand, finger,or wrist) versus other types of splintsor no treatment. Three non-RCTsfound a statistically significant de-crease in pain while participantswere wearing working splints,whereas the effects on pain aftersplinting were conflicting. Only 2 ofthe studies were RCTs of high qualityand reported no difference in pain.Statistically significant improve-ments in grip strength while wearinga splint were reported (2 non-RCTs),

whereas 2 RCTs of high quality re-ported no statistically significant in-crease in grip strength after a periodof time. The authors concluded thatthere are indicative findings thatsplints are effective in reducing pain.Furthermore, they stated that thereare indicative findings for a gain ingrip strength immediately after pro-vision of the splints.

Egan et al20 included 12 studies as-sessing the same interventions ver-sus placebo or alternative interven-tions. Splints and orthoses weredefined as any medical deviceadded to a persons body to support,align, position, immobilize, preventor correct deformity, assist weakmuscles, or improve function. The

Table 5.Description of Included Reviews: Occupational Therapy/Orthosisa


Methodological Quality ofReview and PrimaryStudies

Results

Occupational therapy forrheumatoid arthritis(Steultjens et al)19

16 RCTs6 CCTs16 ODs(1,789 participants)

ReviewMinor limitationsPrimary studiesMean quality score (RCT/CCT)

9.3 (van Tulder scale, 1997)b

Mean quality score (OD)8.1(adapted van Tulder scale)

No quantitative poolingWe found strong evidence for the efficacy of

instruction of joint protection on functionalability. Studies that evaluatedcomprehensive OT showed limitedevidence for the effectiveness on functionalability. Studies that evaluated splintinterventions reported indicative findingsfor the effectiveness on pain.

A critical review of footorthoses in therheumatoid arthriticfoot (Clark et al)21


ReviewModerate limitationsPrimary studiesCriteria used to assess

methodological qualityreported

No quantitative poolingFrom the review there is limited and

conflicting evidence upon which to baseclinical practice. The suggestion is that footorthoses may reduce pain and improveability, but these outcomes are notachieved by all studies.

Splints and orthosis fortreating rheumatoidarthritis (Egan et al)20

12 studies (RCT/CCT/OD) ReviewMinor limitationsPrimary studiesMean quality score2.3 (Jadad

scale: 05)

No quantitative poolingThere is insufficient evidence to make firm

conclusions about the effectiveness ofworking wrist splints in decreasing pain orincreasing function for people withRA. . . . There is evidence that extra-depthshoes and molded insoles decrease painduring weight-bearing activities such asstanding, walking, and stair climbing. Extra-depth shoes with semi-rigid insoles providebetter pain relief than extra-depth shoesalone.

a RCTrandomized controlled trial, CCTcontrolled clinical trial, ODother designs, OToccupational therapy, RArheumatoid arthritis.b Van Tulder MW, Assendelft WJJ, Koes BW, Bouter LM. Method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group forSpinal Disorders. Spine. 1997;22:23232330.



primary studies were of low to mod-erate quality. The studies showedconflicting results regarding painand function, and the authors con-cluded that there is insufficient evi-dence to make firm conclusionsabout the effectiveness of workingsplints or resting splints on pain andfunction. Based on the conflicting re-sults in both of the reviews,19,20 weconclude that the effect of splints(hand, finger, or wrist) is unclear(Tab. 2). Twelve of the includedstudies reported on safety or side ef-fects,19 and both reviews19,20 con-cluded that there were some indica-tive findings that splinting has anegative effect on dexterity.

Training of motor function. Sixstudies evaluated training of motorfunction versus no treatment or al-ternative treatment. One RCT withhigh methodological quality re-ported no significant differences be-tween groups on pain and functionalability after training of hand func-tion. The other studies were of lowquality, showed conflicting results,and did not report sufficient data tocalculate effect sizes. The authorsconcluded that there is no evidencefor the effectiveness of training ofmotor function on pain or function.Based on the conflicting results, weconclude that there is unclear evi-dence for the effect of motor func-tion training on pain and function(Tab. 2). One study reported prob-lems with the upper extremity afterresistance exercises.

Foot orthoses. Effects of foot or-thoses and special shoes versus pla-cebo, no intervention, or other inter-vention were reported in 2reviews,20,21 with 4 and 11 studiesincluded, respectively (Tab. 5). Footorthoses are prescribed by cliniciansas a form of intervention for thesymptomatic foot in patients withRA to relieve forefoot, midfoot, andrear-foot pain and to normalize thepain pattern.21 All patients included

had a confirmed diagnosis of RA.One review20 reported that 2 low-quality studies found significant im-provements in pain, but not in func-tion, and 2 high-quality studies foundno significant changes in either painor function (up to 3 years of orthosiswear). Clark et al21 reported conflict-ing results on both pain and func-tion, but no exact results were pre-sented. Thus, we conclude, based onthe 2 reviews, that the effect of or-thosis is unclear (Tab. 2). Safety orside effects were not reported in anyof the reviews.

Patient education. Four re-views3,2224 reported on the effectsof patient education for patientswith RA (Tab. 6), describing the in-terventions as any set of plannededucational activities designed to im-prove patients health behavior orhealth status, a multidisciplinaryintervention delivered by a team ofhealth care professionals, providinga systematic approach to care, andincluding a patient education com-ponent, and psychosocial interven-tions. Patients included had eithera clinical confirmation of RA oradult RA.

Riemsma et al22 included 31 RCTs.The interventions were patient edu-cation interventions versus controland were categorized into 3 groups:information only, counseling,and behavioral treatment. No sig-nificant effects of information onlyor counseling were reported. Signif-icant effects on disability, patientglobal assessment, and depressionwere found for behavioral treatmentinterventions after treatment. Theauthors concluded that patient edu-cation had small short-term effectson disability, joint counts, patientglobal assessment, psychological sta-tus and depression, but there was noevidence of long-term benefits.

Niedermann et al23 concluded thatthe results on physical health status

are conflicting, and no long-termchanges in disability and physicalfunction were found in any study.Badamgarav et al24 reported thatpooled effect sizes were small andnonsignificant, but studies with in-terventions of greater than 5 weeksduration showed significant differ-ences in functional status. Astin et al3

concluded that there are small, butstatistically significant, effect sizesfor pain, functional disability, de-pression, coping, and self-efficacy af-ter treatment.

All included reviews concluded thatdifferent types of patient educationinterventions have small short-termeffects on different health outcomes,but there is no evidence for the long-term benefits. Based on the manystudies with consistent results, weconclude that there is high-qualityevidence that patient education im-proves function and patient globalassessment (Tab. 2).

Physical Therapy InterventionsExercises. Five reviews2529 re-ported the effects of therapeutic ex-ercises (Tab. 7), including tai chiinstructions, any form of exercisefor the hand, therapeutic exer-cises, with an emphasis on the inten-sity of exercise program, and aer-obic activities. Target populationswere ambulatory adults with a diag-nosis of RA or clinically confirmedRA. Generally, the methodologicalquality of the primary studies waslow or not reported, and the inter-ventions were heterogeneous. Oneof the included reviews28 providedquantitative pooling of results basedon 2 primary studies of very lowquality. Most of the reviews statedthat further studies were needed, butthe results indicated that exercisesmay have some beneficial effects onpain and function. One review25

rated the evidence on pain and over-all function as good, but the con-clusion is based on studies with con-flicting results and no quantitative



Table 6.Description of Included Reviews: Patient Education Interventiona


Methodological Quality ofReview and Primary Studies

Results

Psychological interventions forrheumatoid arthritis: a meta-analysis of randomizedcontrolled trials (Astinet al)3

25 RCTs (1,676participants)

ReviewMinor limitationsPrimary studiesMean quality score2.24 (Jadad

scale: 05)Mean quality score5.84 (van

Tulder scale: 010)

PainPooled effect size

(postintervention)0.22,95% CI0.07 to 0.37 (13 RCTs)

Pooled effect size (follow-up)0.06,95% CI0.17 to 0.29 (6 RCTs)

FunctionPooled effect size

(postintervention)0.27,95% CI0.12 to 0.42 (12 RCTs)

Pooled effect size (follow-up)0.12,95% CI0.09 to 0.33 (7 RCTs)

Self-efficacyEffect size (postintervention)0.35,

95% CI0.11 to 0.59 (5 RCTs)Effect size (follow-up)0.20,

95% CI0.08 to 0.48

Effects of disease managementprograms on functionalstatus of patients withrheumatoid arthritis(Badamgarav et al)24


ReviewModerate limitationsPrimary studiesType of quality score not reported

FunctionEffect size0.27, 95% CI0.01 to

0.54 (7 RCTs/1 CCT)Interventions lasting 5 wk: effect

size0.49, 95% CI0.12 to 0.86(2 RCTs/1 CCT)

Interventions lasting 5 wk: effectsize0.13, 95% CI0.25 to 0.52(4 RCTs)

Gap between short- and long-term effects of patienteducation in rheumatoidarthritis patients: asystematic review(Niedermann et al)23

11 RCTs (931participants)

ReviewModerate limitationsPrimary studiesMean quality score5.6

(Cochrane/Amsterdam-Maastricht, range011)

No quantitative poolingMethodologically better-designed

studies had more difficultiesdemonstrating positive outcomeresults. Short-term effects inprogram targets are generallyobserved, whereas long-termchanges in health status are notconvincingly demonstrated.

Systematic review ofrheumatoid arthritis patienteducation (Riemsma et al)22

31 RCTsNo. of included

participants: notpossible to assess (8,632participants?)

ReviewMinor limitationsPrimary studiesMean quality score2.6 (modified

Jadad scale: 08)

PainPosttreatment SMD0.08,

95% CI0.16 to 0.00 (37 CTs)b

Follow-up SMD0.07,95% CI0.19 to 0.05 (19 RCTs)

FunctionPosttreatment SMD0.17,

95% CI0.25 to 0.09(37 RCTs)b

Follow-up SMD0.09,95% CI0.20 to 0.02 (23 RCTs)

Patient global assessmentPosttreatment SMD0.28,

95% CI0.49 to 0.07 (6 RCTs)Follow-up SMD0.06,

95% CI0.22 to 0.10 (7 RCTs)

a RCTrandomized controlled trial, CCTcontrolled clinical trial, CTclinical trial, CIconfidence interval, SMDstandardized mean difference.b Some studies contained more than 2 comparative interventions and were included as independent studies in the statistical pooling.



Table 7.Description of Included Reviews: Physical Therapy Interventions (Exercises/Electrotherapy/Thermotherapy)a


MethodologicalQuality of Reviewand Primary Studies

Results

Ottawa PanelEvidence-BasedClinical PracticeGuidelines forTherapeuticExercises in theManagement ofRheumatoidArthritis inAdults (OttawaPanel)25


ReviewMinor limitationsPrimary studiesMean quality score1.2

(Jadad scale: 05)

No quantitative poolingGood evidence (level I, RCT) exists that therapeutic

exercises, including functional strengthening andlow- or high-intensity exercises, relieve pain andimprove overall function.

Tai chi for treatingrheumatoidarthritis (Hanet al)28

3 RCTs1 CCT(206 participants)

ReviewMinor limitationsPrimary studiesMean quality


FunctionADL index: WMD0.001, 95% CI2.94 to 2.97

(1 RCT/1 CCT)50-ft walking test: WMD0.35, 95% CI1.14 to

1.84 (1 RCT/1 CCT)

The effectivenessof handexercises forpersons withrheumatoidarthritis: asystematicreview(Wessel)26

9 CCTs or case series (262participants)

ReviewModerate limitationsPrimary studiesMean quality score28

(type of scale notreported: 048)

No quantitative poolingThere is not strong research evidence for or against

the value of hand exercise in the treatment ofpersons with rheumatoid arthritis, although resultsof this review suggest that appropriate exercisemight lead to long-term strength changes and veryshort-term changes in stiffness.

Training effects onpain inrheumatoidarthritis (Engeret al)27

6 RCTsUnknown number of

participants

ReviewModerate limitationsPrimary studiesQuality score not

reported (Jamtvedtand Hilde, 2000)b

No quantitative poolingThe results indicate that exercise leads to

unchanged or reduced self-reported pain and jointtenderness for patients with rheumatoid arthritis.Further studies are needed in order to demonstrateto what extent and how different kinds of exerciseinfluence pain and joint tenderness in rheumatoidpatients.

A healthprofessionalsguide toexerciseprescription forpeople witharthritis: areview ofaerobic fitnessactivities(Westby)29

11 RCTs7 CCTs or pretest-posttest

studies

ReviewModerate limitationsPrimary studiesQuality score not

reported (Squires,1989)c

No quantitative poolingAlthough more comprehensive evaluation is

necessary, present evidence suggests that aerobicfitness activities have minimal to moderate positiveshort-term effects on arthritis-related impairment,disease activity, or accelerating joint damage. Thelong-term effects of aerobic exercise are less wellunderstood. Available data do not suggest that onemode of aerobic exercises is better than anotherwhen comparing the changes in aerobic capacity,disease activity, pain, physical activity, anddepression.

Low level lasertherapy (classesI, II and III) inthe treament ofrheumatoidarthritis(Brosseauet al)30

6 RCTs (222 participants) ReviewMinor limitationsPrimary studiesMean quality score: 3.0

(Jadad scale: 05)

FunctionROM: WMD18.03, 95% CI31.80 to 4.27

(2 RCTs)Flexibility-tip to palm distance: WMD1.28,

95% CI1.72 to 0.85 (2 RCTs)PainVAS: WMD1.10, 95% CI1.82 to 0.39

(3 RCTs)

(Continued)



Table 7.Continued



Results

Ottawa PanelEvidence-BasedClinical PracticeGuidelines forElectrotherapyandThermotherapyInterventions inthe Managementof RheumatoidArthritis inAdults (OttawaPanel)15

Low-level laser therapy5 RCTs (204 participants)Therapeutic ultrasound1 RCT (50 participants)TENS3 RCTs (94 participants)Electrical stimulationNo reviews found

ReviewMinor limitationsPrimary studiesLow-level laser therapy:

mean qualityscore4.0 (Jadadscale: 05)

Therapeutic ultrasound:mean qualityscore3.0 (Jadadscale: 05)

TENS: mean qualityscore2.3 (Jadadscale: 05)

Low-level laser therapyFunctionROM: WMD1.26, 95% CI1.72 to 0.85

(2 RCTs)PainVAS: WMD1.05, 95% CI1.58 to 0.53

(4 RCTs)Therapeutic ultrasoundNo quantitative poolingThe Ottawa Panel found good evidence (level I,

RCT) of the effects of therapeutic ultrasound forRA of the hand.

TENSNo quantitative pooling The Ottawa Panel found

good evidence (level I, RCT) on the effects ofTENS for management of RA in the hand and wrist.

Electrical stimulationEvidence with acceptable research design,

interventions, group comparisons, or outcomescould not be identified.

Therapeuticultrasound forthe treatment ofrheumatoidarthritis(Casimiroet al)31

2 RCTs (80 participants) ReviewMinor limitationsPrimary studiesMean quality score2

(Jadad scale: 05)

No quantitative poolingUltrasound in combination with exercises, faradic

current, and wax bath treatment modalities is notsupported and cannot be recommended. Ultrasoundalone can, however, be used on the hand to increasegrip strength, and to a lesser extent and based onborderline results, increase wrist flexion, decreasemorning stiffness, reduce the number of swollenjoints, and reduce the number of painful joints. It isimportant to note that these conclusions are limitedby methodological considerations.

Transcutaneouselectrical nervestimulation(TENS) for thetreatment ofrheumatoidarthritis in thehand (Brosseauet al)34

3 RCTs (78 participants) ReviewMinor limitationsPrimary studiesMean quality score: 2.3

(Jadad scale: 05)

No quantitative poolingThere are conflicting effects on pain outcomes in

patients with RA. AL-TENS is beneficial for reducingpain intensity and improving muscle power scoresover placebo, while, conversely, C-TENS resulted inno clinical benefit on pain intensity compared withplacebo. However, C-TENS resulted in a clinicalbenefit on patient assessment of change in diseaseover AL-TENS. More well-designed studies with astandardized protocol and adequate number ofsubjects are needed to fully conclude the effect onC-TENS and AL-TENS in the treatment of RA of thehand.

Efficacy of paraffinwax baths forrheumatoidarthritic hands(Ayling andMarks)32

4 RCTs (303 participants) ReviewModerate limitationsPrimary studiesMean quality score10

(Beckerman et al,1992: 025)d

No quantitative poolingA critical review of the literature concerning the

clinical efficacy of paraffin wax as a treatment forhands afflicted with RA has not revealed any studywithout serious bias. Thus, no definitive evidenceexists either in support of this treatment approach.In addition, studies suggesting that the techniques ofwax baths may be temporarily useful, or not useful atall, did not necessarily employ optimal applicationmethods. Related data do suggest, though, thatphysiotherapists should avoid using this modality whentheir patients joint disease is active.

(Continued)



pooling. Patient global assessmentwas not statistically significant in thestudies that reported this outcome.Based on the conflicting results, weconclude that there is low-quality ev-idence that exercises reduce painand improve function and that thereis no difference in patient globalassessment (Tab. 2).

Tai chi instructions caused somejoint and muscle pain complaintsthat diminished during the course ofthe study, but it did not cause with-drawals. One review29 found thataerobic fitness activities do not exac-erbate disease activity or acceleratejoint damage, whereas another re-view25 concluded that, although noharmful side effects were reported inthe original studies, the effects of

high-intensity exercise on pain raiseconcern.

Low-level laser therapy (LLLT).Two reviews were identified,15,30 de-fining low-level laser therapy as alight source that generates extremelypure light, of a single wavelength.All patients included had clinicallyconfirmed RA. The methodologicalquality of the primary studies variedfrom low to good, and both reviewsprovided quantitative pooling of re-sults (Tab. 7). Three RCTs of moder-ate quality showed statistically signif-icant improvements in pain, and 2RCTs of moderate quality found sig-nificant improvements in function(ie, ROM, flexibility). Patient globalassessment was not reported. Thereviews agreed on the conclusion

that LLLT could be considered forrelief of pain and rated the evidenceto be of silver quality30 and goodquality,15 respectively. We conclude,therefore, that there is moderate-quality evidence that LLLT reducespain and improves function (Tab. 2).None of the reviews reported onpossible side effects and safety.

Therapeutic ultrasound. Two re-views15,31 reported the effect of ther-apeutic ultrasound (pulsed or contin-uous), including 1 and 2 primarystudies, respectively (Tab. 7). Pa-tients included had clinically con-firmed RA. Results from the 2 pri-mary studies could not be combinedin a meta-analysis. One RCT of mod-erate quality showed statistically sig-nificant improvements in pain (ten-

Table 7.Continued



Results

Ottawa PanelEvidence-BasedClinical PracticeGuidelines forElectrotherapyandThermotherapyInterventions inthe Managementof RheumatoidArthritis inAdults (OttawaPanel15)

2 RCTs (76 participants) ReviewMinor limitationsPrimary studies

Mean qualityscore1 (Jadad scale:05)

No quantitative poolingThe Ottawa Panel found good evidence (level I,

RCT) that thermotherapy, especially waxcombined with exercise, benefits ROM, pain, andstiffness in the management of RA.

Thermotherapy fortreatingrheumatoidarthritis(Robinsonet al)33

7 RCTs (328 participants) ReviewMinor limitationsPrimary studiesMean quality score1.6

(Jadad scale: 05)

No quantitative poolingThe results of this systematic review on

thermotherapy in the treatment of RA found thatthere was no significant effect of hot or ice packapplications or faradic baths on objective measuresof disease activity, including joint swelling, pain,medication intake, ROM, grip strength, or handfunction, when compared to a control or alternatetreatment.

a RCTrandomized controlled trial, SMDstandardized mean difference, CIconfidence interval, CTclinical trial, ADLactivities of daily living,WMDweighted mean difference, ROMrange of motion, VASvisual analog scale, TENStranscutaneous electrical nerve stimulation, RArheumatoidarthritis, AL-TENSacupuncture-like transcutaneous electrical nerve stimulation, C-TENSconventional transcutaneous electrical nerve stimulation.b Jamtvedt G, Hilde G. Evidence-based physiotherapy: critical appraisal of a randomized controlled trial [in Norwegian]. Norwegian Physiotherapy Journal.2000;67:712.c Squires BP. Biomedical review articles: what editors want from authors and peer reviewers. Can Med Assoc J. 1989;141:195197.d Beckerman H, de Bie RA, Bouter LM, et al. The efficacy of laser therapy for musculoskeletal and skin disorders: a criteria-based meta-analysis of randomizedclinical trials. Phys Ther. 1992;72:483491.



der or painful joints) and function(grip strength, ROM). Patient globalassessment was not reported. Basedon one primary study of moderatequality, therefore, we conclude thatthere is low-quality evidence thattherapeutic ultrasound reduces painand improves function (Tab. 2). Nei-ther of the reviews reported sideeffects.

Thermotherapy. Three re-views15,32,33 reported the effects ofthermotherapy (Tab. 7), includingthe interventions melted paraffinwax applications and superficialmoist heat and cryotherapy. All pa-tients included had clinically con-firmed RA. The methodological qual-ity of the primary studies was low,and none of the included reviewsprovided quantitative pooling of theresults. One primary study of lowquality reported statistically signifi-cant improvements in pain and func-tion (grip strength). The study wasincluded in all 3 reviews, and theresults were obtained in the groupthat received a combination of par-affin wax and exercise therapy.When paraffin wax alone was com-pared with a control, there were nosignificant differences in any of theoutcomes. All 3 reviews agreed thatthermotherapy is more effective asan adjunct therapy than it is alone.Patient global assessment was not re-ported as an outcome in any of theincluded studies. On these grounds,we find it reasonable to concludethat the effect of thermotherapyalone is unclear (Tab. 2). No sideeffects were reported.

Transcutaneous electrical nervestimulation (TENS). Two re-views,15,34 including the same pri-mary studies, assessed the outcomesof TENS (low frequency, high fre-quency) (Tab. 7). All patients in-cluded had clinically confirmed RA.Based on a large variation in patientsand methodological issues in the in-cluded studies, the results were not

combined in a meta-analysis. Resultsregarding pain were conflicting.Two studies, one of low quality andone of moderate quality, found sta-tistically significant improvement inpain (visual analog scale score, jointtenderness), whereas one study ofmoderate quality found no signifi-cant improvement in pain (visual an-alog scale score). Because of the in-consistent results, we conclude thatthe effect of TENS is unclear (Tab. 7).No side effects were reported.

DiscussionBased on the evidence from 28 sys-tematic reviews, our overview indi-cates that the quality of evidence forthe effectiveness of most of the in-terventions was low, except for pa-tient educational programs, jointprotection, LLLT, and GLA (herbaltherapy).

Although our intention with thepresent overview was to provideuseful information to patients, clini-cians, researchers, and policymak-ers, it surely has several limitations.First, the results from this overviewcannot be interpreted as treatmentrecommendations. Recommenda-tions for practice are normally partsof clinical guidelines, and guidelinerecommendations should be basedon both research evidence and clin-ical experience. Recommendationsshould be developed through a pro-cess involving both methodologistsand clinicians and taking other fac-tors, such as resources and culture,into account. This overview summa-rizes the empirical evidence, and thebottom line is what is known fromsystematic reviews. Thus, this over-view could be the core element of aclinical guideline, but not a guidelinein itself.

Second, based on our results, it is notpossible to make conclusions regard-ing the effects of modalities on pa-tients with various classes of diseasebecause none of the included re-

views were distinct on these topics.An overview summarizes results ofexisting reviews. It does not re-review the literature or add moreoutcomes or studies. The results ofan overview are dependent on thenumber of high-quality, updated sys-tematic reviews and the quality ofthe primary studies included. Sys-tematic reviews usually give us theanswers in general terms, due to theprimary studies heterogeneity in di-agnosis, interventions, and out-comes, and seldom help the clinicianwho wants to know what type ofintervention is most effective intreating a specific patient in a clinicalpractice. Given these limitations, anoverview might be more valuable forpolicymakers than for clinicians.

Third, possible harmful effects orside effects of the interventions wereseldom reported, and no conclusionswere drawn in any of the includedreviews. This is in accordance withEthgen et al,35 who found that harmwas less often described in reports ofnonpharmacological treatment trialsthan in reports of pharmacologicaltreatment trials. This might partly beexplained by a presupposed lowerdegree of harmful effects; however,most therapy might imply the risk ofadverse effects. Unexpected sideeffects can only be detected if dataon all events are systematicallycollected.

Our overview shows that the evi-dence for effects of the most com-mon nonpharmacological and non-surgical interventions for patientswith RA is of low quality. Low-quality evidence is not the same asthe intervention does not have ef-fect, rather that there is insufficientevidence to draw firm conclusionswith regard to the effectiveness ofmany of the interventions usually of-fered to patients with RA. The over-all methodological quality of theprimary studies was frequently rela-tively weak. However, there are



probably better conducted newlypublished RCTs not captured in theincluded reviews, which might en-hance the quality of evidence for theeffects of the interventions. This ispossibly the case especially for exer-cise therapy.

Exercise therapy is considered to bea cornerstone in the treatment of RAin all stages of the disease.36 TheAmerican College of Rheumatology2

underscores the necessity of an in-terdisciplinary approach to the com-prehensive management of RA. Theyrecommend patient education, in-struction in joint protection, conser-vation of energy, and a home pro-gram of ROM and strengtheningexercises as important in achievingthe treatment goals. Furthermore,the American College of Rheumatol-ogy reports that regular participationin dynamic and aerobic conditioningexercise programs improves jointmobility, muscle strength, aerobicfitness, and function as well as psy-chological well-being without in-creasing fatigue or joint symptoms.Apart from the recommendations ofpatient education and joint protec-tion, their strong recommendationsof exercise therapy are not sup-ported by the present overview,which found low-quality evidencefor exercise therapy.

Our results are supported by Smidtet al,37 who conducted a best-evidence summary of systematic re-views on the effectiveness of exer-cise therapy. They concluded thatthere was insufficient evidence tosupport or refute the effectiveness ofexercise therapy for patients withRA. The Ottawa Panel,25 however,recommend the use of exercise ther-apy for RA in their guidelines, butthey emphasize that their guidelinesare limited by generally poorly re-ported descriptions of therapeuticexercise programs and the outcomesin the included studies.

The latest Cochrane review on exer-cise therapy for RA38 was conductedin 1998. It concluded that dynamicexercise therapy had positive ef-fects, but that research on long-termeffects was needed. Lately, manyhigh-quality studies have been car-ried out and concluded that exerciseis both effective and safe in patientswith RA.3943 Thus, it is likely that anupdated, high-quality review on ex-ercise therapy for patients with RAwill make it possible to draw firmerconclusions on the effect of exercisetherapy for this group of patients.

As pointed out earlier, better con-ducted primary studies are necessaryto draw firm conclusions on the ef-fectiveness of nonpharmacologicaland nonsurgical interventions for pa-tients with RA. Foley et al44 assessedthe quality between pharmacologi-cal and nonpharmacological studies.They found that the greatest differ-ence was for subject masking, inwhich virtually all drug trials suc-ceeded, whereas only a small per-centage (35%) of nondrug trialssucceeded. It is obvious that mask-ing the patients is difficult in most ofthe interventions in our study, butmasking of the assessors is possiblein most cases and should be carriedout. Blinding is particularly impor-tant when the outcome measures in-volve patient-reported symptomssuch as pain.45 Outcome measuresalso should be standardized, usingvalid and reliable tools, andfollow-up should be of sufficientlength to assess long-term effects.

The summarized evidence may beused by patients, health care practi-tioners, policymakers, and research-ers. The policymakers need research-based evidence to guide their budgetdecisions on which interventions toreimburse and what kind of researchto fund. For the researchers, ouroverview might identify areas whereresearch is missing and improve-ments are needed. Clinicians and pa-

tients are usually interested in whattype of intervention is the most ef-fective in a specific context. An over-view of systematic reviews will mostoften not add new evidence for asingle intervention, but rather syn-thesize available evidence on differ-ent interventions for the same con-dition and thus provide cliniciansand patients with an overview of theevidence for the most commonlyused interventions. Because theoverview gives the answers in gen-eral terms, it might be more valuablefor policymakers and researchersthan for clinicians and patients.

The increasing need for valid, rele-vant health care information empha-sizes the need for rigorous clinicalresearch to guide health care deci-sions. In the present overview, wefound that the quality of evidence formost nonpharmacological and non-surgical interventions for patientswith RA is low. This does not meanthat most of the interventions areineffective, but rather that the qual-ity of research evidence is low. Aslong as the primary studies have se-rious limitations, our confidence inthe estimate of effect is low. Betterconducted primary studies are war-ranted to provide stakeholders withhigh-quality information needed intheir health care decisions.

Ms Jamtvedt and Dr Hagen providedconcept/idea/project design. All authorsprovided writing and data analysis. MsChristie, Ms Dahm, Ms Moe, Dr Haavardsh-olm, and Dr Hagen provided data collection.Ms Christie and Dr Hagen provided projectmanagement. Ms Jamtvedt and Dr Hagenprovided consultation (including review ofmanuscript before submission).

This work was inspired by the CARE III andCARE IV International Conferences.

This article was submitted January 31, 2007,and was accepted July 17, 2007.

DOI: 10.2522/ptj.20070039



References1 Stenstrom CH, Minor MA. Evidence for the

benefit of aerobic and strengthening exer-cise in rheumatoid arthritis. ArthritisRheum. 2003;49:428434.

2 American College of Rheumatology Sub-committee on Rheumatoid Arthritis.Guidelines for the management of rheu-matoid arthritis: 2002 update. ArthritisRheum. 2002;46:328346.

3 Astin JA, Beckner W, Soeken K, et al. Psy-chological interventions for rheumatoidarthritis: a meta-analysis of randomizedcontrolled trials. Arthritis Rheum.2002;47:291302.

4 Arnett FC, Edworthy SM, Bloch DA, et al.The American Rheumatism Association1987 revised criteria for the classificationof rheumatoid arthritis. Arthritis Rheum.1988;31:315324.

5 International Classification of Function-ing, Disability and Health: ICF. Geneva,Switzerland: World Health Organization;2001.

6 Brooks P, Hochberg M. Outcome mea-sures and classification criteria for therheumatic diseases: a compilation of datafrom OMERACT (Outcome Measures forArthritis Clinical Trials), ILAR (Interna-tional League of Associations for Rheuma-tology), regional leagues and othergroups. Rheumatology (Oxford). 2001;40:896906.

7 Oxman AD, Guyatt GH. Validation of anindex of the quality of review articles.J Clin Epidemiol. 1991;44:12711278.

8 Atkins D, Best D, Briss PA, et al. Gradingquality of evidence and strength of recom-mendations. BMJ. 2004;328(7454):1490.

9 Lee J-D, Park H-J, Chae Y, Lim S. An over-view of bee venom acupuncture in thetreatment of arthritis. Evid Based Comple-ment Alternat Med. 2005;2:7984.

10 Casimiro LY, Brosseau L, Milne S, et al.Acupuncture and electroacupuncture forthe treatment of rheumatoid arthritis.Cochrane Database Syst Rev. 2005;19(4):CD003788.

11 Verhagen AP, de Vet HC, de Bie RA, et al.Balneotherapy for rheumatoid arthritisand osteoarthritis. Cochrane DatabaseSyst Rev. 2003;(4):CD000518.

12 Brosseau L, Robinson VI, Leonard G, et al.Efficacy of balneotherapy for rheumatoidarthritis: a meta-analysis. Physical TherapyReviews. 2002;7:6787.

13 Muller H, De Toledo FW, Resch KL. Fast-ing followed by vegetarian diet in patientswith rheumatoid arthritis: a systematic re-view. Scand J Rheumatol. 2001;30:110.

14 Brosseau L, Pelland LU, Casimiro LY, et al.Electrical stimulation for the treatment ofrheumatoid arthritis. Cochrane DatabaseSyst Rev. 2002;(2):CD003687.

15 Ottawa Panel Evidence-Based ClinicalPractice Guidelines for Electrotherapy andThermotherapy Interventions in the Man-agement of Rheumatoid Arthritis inAdults. Phys Ther. 2004;84:10161043.

16 Little C, Parsons T. Herbal therapy fortreating rheumatoid arthritis. CochraneDatabase Syst Rev. 2001;(1):CD002948.

17 Soeken KL, Miller SA, Ernst E. Herbal med-icines for the treatment of rheumatoid ar-thritis: a systematic review. Rheumatol-ogy (Oxford). 2003;42:652659.

18 Park J, Ernst E. Ayurvedic medicine forrheumatoid arthritis: a systematic review.Semin Arthritis Rheum. 2005;34:705713.

19 Steultjens EM, Dekker J, Bouter LM, et al.Occupational therapy for rheumatoidarthritis. Cochrane Database Syst Rev.2004;(1):CD003114.

20 Egan M, Brosseau L, Farmer M, et al.Splints and orthosis for treating rheuma-toid arthritis. Cochrane Database SystRev. 2006;(1):CD004018.

21 Clark H, Rome K, Plant M, et al. A criticalreview of foot orthoses in the rheumatoidarthritic foot. Rheumatology (Oxford)2006;45(2):139145.

22 Riemsma RP, Taal E, Kirwan JR, Rasker JJ.Systematic review of rheumatoid arthritispatient education. Arthritis Rheum.2004;51:10451059.

23 Niedermann K, Fransen J, Knols R, Uebel-hart D. Gap between short- and long-termeffects of patient education in rheumatoidarthritis patients: a systematic review.Arthritis Rheum. 2004;51:388398.

24 Badamgarav E, Croft JD Jr, Hohlbauch A,et al. Effects of disease management pro-grams on functional status of patients withrheumatoid arthritis. Arthritis Care Res.2003;41:377387.

25 Ottawa Panel Evidence-Based ClinicalPractice Guidelines for Therapeutic Exer-cises in the Management of RheumatoidArthritis in Adults. Phys Ther. 2004;84:934972.

26 Wessel J. The effectiveness of hand exer-cises for persons with rheumatoid arthri-tis: a systematic review. J Hand Ther.2004;17:174180.

27 Enger KJ, Bjornstad K, Rodevand E, Skoms-voll JF. Training effects on pain in rheuma-toid arthritis [in Norwegian]. Tidsskrift forDen Norske Laegeforening. 2003;123:15081510.

28 Han A, Robinson VI, Judd MG, et al. Tai chifor treating rheumatoid arthritis.Cochrane Database Syst Rev. 2004;(3):CD004849.

29 Westby MD. A health professionals guideto exercise prescription for people witharthritis: a review of aerobic fitness activ-ities. Arthritis Rheum. 2001;45:501511.

30 Brosseau L, Welch V, Wells G, et al. Lowlevel laser therapy (classes I, II and III) inthe treatment of rheumatoid arthritis.Cochrane Database Syst Rev. 2005;(4):CD002049.

31 Casimiro LY, Brosseau L, Judd MG, et al.Therapeutic ultrasound for the treatmentof rheumatoid arthritis. Cochrane Data-base Syst Rev. 2002;(3):CD003787.

32 Ayling J, Marks R. Efficacy of paraffin waxbaths for rheumatoid arthritic hands. Phys-iotherapy. 2000;86:190201.

33 Robinson VI, Brosseau L, Casimiro LY,et al. Thermotherapy for treating rheuma-toid arthritis. Cochrane Database SystRev. 2002;(2):CD002826.

34 Brosseau L, Judd MG, Marchand S, et al.Transcutaneous electrical nerve stimula-tion (TENS) for the treatment of rheuma-toid arthritis in the hand. Cochrane Data-base Syst Rev. 2006;(1):CD004377.

35 Ethgen M, Boutron I, Baron G, et al. Re-porting on harm in randomized, con-trolled trials of nonpharmacologic treat-ment for rheumatic disease. Ann InternMed. 2005;143:2025.

36 van den Ende CH, Vliet Vlieland TP, Mun-neke M, Hazes JM. Dynamic exercise ther-apy in rheumatoid arthritis: a systematicreview. Br J Rheumatol. 1998;37:677687.

37 Smidt N, de Vet HC, Bouter LM, et al. Ef-fectiveness of exercise therapy: a best-evidence summary of systematic reviews.Aust J Physiother. 2005;51:7185.

38 van den Ende CH, Vliet Vlieland TP, Mun-neke M, Hazes JM. Dynamic exercise ther-apy for rheumatoid arthritis. CochraneDatabase Syst Rev. 2000;(2):CD000322.

39 Hakkinen A. Effectiveness and safety ofstrength training in rheumatoid arthritis.Curr Opin Rheumatol. 2004;16:132137.

40 Hakkinen A, Sokka T, Kautiainen H, et al.Sustained maintenance of exercise-induced muscle strength gains and normalbone mineral density in patients with earlyrheumatoid arthritis: a 5-year follow-up.Ann Rheum Dis. 2004;63:910916

41 de Jong Z, Munneke M, Lems WF, et al.Slowing of bone loss in patients with rheu-matoid arthritis by long-term high-intensity exercise. Arthritis Rheum.2004;50:10661076.

42 de Jong Z, Munneke M, Zwinderman AH,et al. Is a long-term high-intensity exerciseprogram effective and safe in patients withrheumatoid arthritis? Arthritis Rheum.2003;48:24152424.

43 Hakkinen A, Hannonen P, Nyman K, et al.Effects of concurrent strength and endur-ance training in women with early or long-standing rheumatoid arthritis: comparisonwith healthy subjects. Arthritis Rheum.2003;49:789797.

44 Foley NC, Bhogal SK, Teasell RW, et al.Estimates of quality and reliability with thePhysiotherapy Evidence-Based DatabaseScale to assess the methodology of ran-domized controlled trials of pharmacolog-ical and nonpharmacological interven-tions. Phys Ther. 2006;86:817824.

45 Boutron I, Tubach F, Giraudeau B, RavaudP. Methodological differences in clinicaltrials evaluating nonpharmacological andpharmacological treatments of hip andknee osteoarthritis. JAMA. 2003;290:10621070.



Appendix 1.Search Strategy

(A) Study type: systematic reviews

1. controlled.ab. [ab.all searchable words from the abstract]

2. design.ab.

3. evidence.ab.

4. extraction.ab.

5. randomized controlled trials/[MESH]

6. meta-analysis.pt. [pt.publication type]

7. review.pt.

8. sources.ab.

9. studies.ab.

10. OR/19

11. letter.pt.

12. comment.pt.

13. editorial.pt.

14. OR/1113

15. 10 NOT 14

(B) Participants: rheumatoid arthritis

Arthritis, rheumatoid/OR (arthritis adj2 rheumat$).tw. [Textword]

(C) Interventions: nonpharmacological and nonsurgical

exp Therapeutics/(exploded MESH term] OR therap$.tw. OR dh.fs. [Diet Therapy as floating subheading to a MESH term]

exp behavior and behavior mechanisms/OR exp psychological phenomena and processes/OR exp mental disorders/OR expbehavioral disciplines and activities/

The following MESH terms and floating subheading were excluded from the search result with NOT:

exp Specialties, Surgical/OR su.fs [Surgery as floating subheading to a MESH term]

exp inorganic chemicals/OR exp organic chemicals/OR exp heterocyclic compounds/OR exp polycyclic compounds/OR expmacromolecular substances/OR exp hormones, hormone substitutes, and hormone antagonists/OR exp enzymes and coenzymes/OR exp carbohydrates/OR exp lipids/OR exp amino acids, peptides, and proteins/OR exp nucleic acids, nucleotides, andnucleosides/OR exp complex mixtures/OR exp biological factors/OR exp biomedical and dental materials/OR exppharmaceutical preparations/OR exp chemical actions and uses/

(D) Language restrictions: English or Scandinavian language

(E) Publication year from 2000 to January 2007



Appendix 2.Quality Assessment Checklist for Systematic Reviews

The following 9 criteria was rated as met, unclear/partly met, or not met according to a criteria list modified from a previouslyvalidated checklist7:

1. Is the search strategy described in enough detail for the search to be reproducible?

2. Was the search for evidence reasonably comprehensive?

3. Were the criteria used for deciding which studies to include in the review reported?

4. Was bias in the selection of articles avoided?

5. Were the criteria used for assessing the validity of the studies that were reviewed reported?

6. Was the validity of all of the studies referred to in the text assessed using appropriate criteria in analyzing the studies that are cited?

7. Were the methods used to combine the findings of the relevant studies (to reach a conclusion) reported?

8. Were the findings of the relevant studies combined (or not combined) and analyzed appropriately relative to the primary question thereview addresses and the available data?

9. Were the conclusions made by the author(s) supported by the data and/or the analysis reported in the review?

Appendix 3.Excluded Reviews (n41)

Bartels EM, Lund H, Nneskiold-Samsoe B. Pool exercise therapy of rheumatoid arthritis [in Danish]. Ugeskrift for Laeger. 2001;163:55075513. [Review] [49 references]

Berman BM, Swyers JP, Ezzo J. The evidence for acupuncture as a treatment for rheumatologic conditions. Rheum Dis Clin North Am.2000;26:103115. [Review] [54 references]

Cleland LG, James MJ, Proudman SM. The role of fish oils in the treatment of rheumatoid arthritis. Drugs. 2003;63:845853. [Review] [47references]

Clyman B. Sports, exercise, and arthritis. Bull Rheum Dis. 2001;50:13. [Review] [22 references]

de Jong Z, Vlieland TP. Safety of exercise in patients with rheumatoid arthritis. Curr Opin Rheumatol. 2005;17:177182. [Review] [43references]

Ernst E. Chinese herbal medicines for consumers in the West. Focus on Alternative & Complementary Therapies. 2005;10:9497.

Ernst E. Complementary medicine. Curr Opin Rheumatol. 2003;15:151155. [Review] [19 references]

Ernst E. Musculoskeletal conditions and complementary/alternative medicine. Best Pract Res Clin Rheumatol. 2004;18:539556.[Review] [100 references]

Fransen M. When is physiotherapy appropriate? Best Pract Res Clin Rheumatol. 2004;18:477489. [Review] [54 references]

Grant KL, Schneider CD. Alternative therapies: turmeric. Am J Health Syst Pharm. 2000;57:11211122.

Hakkinen A. Effectiveness and safety of strength training in rheumatoid arthritis. Curr Opin Rheumatol. 2004;16:132137. [Review] [46references]

Hammond A. What is the role of the occupational therapist? Best Prac Res Clin Rheumatol. 2004;18:491505. [Review] [68 references]

Hardware B, Lacey A. Acupuncture and other alternative therapies in rheumatoid arthritis. Professional Nurse. 2002;17:437439.

Jonas WB, Linde K, Ramirez G. Homeopathy and rheumatic disease. Rheum Dis Clin North Am. 2000;26:117123.

Kettunen JA, Kujala UM. Exercise therapy for people with rheumatoid arthritis and osteoarthritis. Scand J Med Sci Sports. 2004;14:138142. [Review] [31 references]

Li LC, Iversen MD. Outcomes of patients with rheumatoid arthritis receiving rehabilitation. Curr Opin Rheumatol. 2005;17:172176.[Review] [28 references]

Lodha R, Bagga A. Traditional Indian systems of medicine. Ann Acad Med Singapore. 2000;29:3741. [Review] [52 references]

(Continued)



Appendix 3.Continued

Madsen OR. Significance of physical activity for bone mass and fracture risk in patients with rheumatoid arthritis [Danish]. Ugeskrift forLaeger. 2002;164:45284531. [Review] [40 references]

Mulligan K, Newman S. Psychoeducational interventions in rheumatic diseases: a review of papers published from September 2001 toAugust 2002. Curr Opin Rheumatol. 2003;15:156159.

Munneke M, De Jong Z. The role of exercise programs in the rehabilitation of patients with rheumatoid arthritis. International SportMed Journal. 2000;1:112.

Nasermoaddeli A, Kagamimori S. Balneotherapy in medicine: a review. Environmental Health & Preventive Medicine. 2005;10:171179.

OGrady M, Fletcher J, Ortiz S. Therapeutic and physical fitness exercise prescription for older adults with joint disease: an evidence-based approach. Rheum Dis Clin North Am. 2000;26:617646. [Review] [137 references]

Paternostro-Sluga T, Stieger M. Hand splints in rehabilitation. Crit Rev Phys Rehabil Med. 2004;16:233256.

Pedersen BK, Saltin B. Evidence for prescribing exercise as therapy in chronic disease. Scand J Med Sci Sports. 2006;16(suppl 1):363.[Review] [735 references]

Ruxton C. Health benefits of omega-3 fatty acids. Nurs Stand. 2004;18:3842. [Review] [44 references]

Schrieber L, Colley M. Patient education. Best Pract Res Clin Rheumatol. 2004;18:465476. [Review] [20 references]

Soeken KL. Selected CAM therapies for arthritis-related pain: the evidence from systematic reviews. Clin J Pain. 2004;20:1318.

Stamp LK, James MJ, Cleland LG. Diet and rheumatoid arthritis: a review of the literature. Semin Arthritis Rheum. 2005;35:7794.

Stenstrom CH, Minor MA. Evidence for the benefit of aerobic and strengthening exercise in rheumatoid arthritis. Arthritis Rheum. 2003;49:428434. [Review] [58 references]

Stenstrom CH, Sturk N. Moderate intensive exercise has a positive effect in rheumatoid arthritis [Swedish]. Lakartidningen. 2004;101:35163517, 3519. [Review] [20 references]

Taibi DM, Bourguignon C. The role of complementary and alternative therapies in managing rheumatoid arthritis. Fam CommunityHealth. 2003;26:4152.

Taibi DM, Bourguignon C, Taylor AG. Valerian use for sleep disturbances related to rheumatoid arthritis. Holist Nurs Pract.2004;18:120126.

Tidow-Kebritchi S. Effects of diets containing fish oil and vitamin E on rheumatoid arthritis. Nutr Rev. 2001;59:335338.

Trieb K. Management of the foot in rheumatoid arthritis. J Bone Joint Surg Br. 2005;87:11711177. [Review] [79 references]

Uhlig T, Finset A, Kvien TK. Effectiveness and cost-effectiveness of comprehensive rehabilitation programs. Curr Opin Rheumatol. 2003;15:134140. [Review] [55 references]

Vliet Vlieland TP. Multidisciplinary team care and outcomes in rheumatoid arthritis. Curr Opin Rheumatol. 2004;16:153156. [Review][27 references]

Vliet Vlieland TP. Rehabilitation of people with rheumatoid arthritis. Best Pract Res Clin Rheumatol. 2003;17:847861. [Review] [96references]

Weiner DK, Ernst E. Complementary and alternative approaches to the treatment of persistent musculoskeletal pain. Clin J Pain. 2004;20:244255. [Review] [80 references]

Yocum DE, Castro WL, Cornett M. Exercise, education, and behavioral modification as alternative therapy for pain and stress inrheumatic disease. Rheum Dis Clin North Am. 2000;26:145159. [Review] [54 references]

Zan-Bar T, Aron A, Shoenfeld Y. Acupuncture therapy for rheumatoid arthritis. APLAR Journal of Rheumatology. 2004;7:207214.

Zochling J. Complementary and alternative medicines and arthritis. Medicine Today. 2004;5(9):6366.



doi: 10.2522/ptj.20070039Originally published online September 25, 2007

2007; 87:1697-1715.PHYS THER. H Moe, Espen A Haavardsholm and Kre Birger HagenAnne Christie, Gro Jamtvedt, Kristin Thuve Dahm, RikkeReviewsRheumatoid Arthritis: An Overview of SystematicNonsurgical Interventions for Patients With Effectiveness of Nonpharmacological and

References

http://ptjournal.apta.org/content/87/12/1697#BIBLfor free at: This article cites 44 articles, 7 of which you can access

Cited by

http://ptjournal.apta.org/content/87/12/1697#otherarticlesThis article has been cited by 7 HighWire-hosted articles:

Information Subscription http://ptjournal.apta.org/subscriptions/

Permissions and Reprints http://ptjournal.apta.org/site/misc/terms.xhtmlInformation for Authors http://ptjournal.apta.org/site/misc/ifora.xhtml

by guest on May 1, 2012http://ptjournal.apta.org/Downloaded from

effectiveness of nonpharmacological and nonsurgical interventions for patients with rheumatoid arthritis an overview of systematic reviews.pdf

Documents