Combination therapy for pain management in inflammatoryarthritis (rheumatoid arthritis, ankylosing spondylitis,psoriatic arthritis, other spondyloarthritis)
Sofia Ramiro1,2, Helga Radner3, Désirée van der Heijde4, Astrid van Tubergen5 , Rachelle Buchbinder6 , Daniel Aletaha3, Robert BMLandewé2 ,7
1Department of Rheumatology, Hospital Garcia de Orta, Almada, Portugal. 2Department of Clinical Immunology & Rheumatology,Academic Medical Centre, Amsterdam, Netherlands. 3Division of Rheumatology, Department of Internal Medicine 3, Medical Univer-sity Vienna, Vienna, Austria. 4Department of Rheumatology, Leiden University Medical Center, Leiden, Netherlands. 5Department ofMedicine, Division of Rheumatology, Maastricht University Medical Center, Maastricht, Netherlands. 6Monash Department of Clini-cal Epidemiology at Cabrini Hospital, Department of Epidemiology and Preventive Medicine, School of Public Health and PreventiveMedicine, Monash University, Malvern, Australia. 7Department of Rheumatology, Atrium Medical Centre, Heerlen, Netherlands
Editorial group: Cochrane Musculoskeletal Group.Publication status and date: New, published in Issue 10, 2011.Review content assessed as up-to-date: 30 October 2010.
Citation: Ramiro S, Radner H, van der Heijde D, van Tubergen A, Buchbinder R, Aletaha D, Landewé RBM. Combination therapy forpain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis, other spondyloarthritis).Cochrane Database of Systematic Reviews 2011, Issue 10. Art. No.: CD008886. DOI: 10.1002/14651858.CD008886.pub2.
Despite optimal therapy with disease-modifying antirheumatic drugs, many people with inflammatory arthritis (IA) continue to havepersistent pain that may require additional therapy.
Objectives
To assess the benefits and safety of combination pain therapy for people with IA (rheumatoid arthritis (RA), ankylosing spondylitis (AS),psoriatic arthritis (PsA) and other spondyloarthritis (SpA)). We planned to assess differences in effects between patients on backgrounddisease-modifying antirheumatic drug (DMARD) therapy and patients on no background therapy in subgroup analyses.
Search methods
We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library); MEDLINE; and EMBASE. Wedid not impose any date or language restrictions in the search. We also handsearched conference proceedings of the American Collegeof Rheumatology and the European League against Rheumatism (2008-10).
Selection criteria
Randomised and controlled clinical trials (RCTs and CCTs) assessing combination therapy (at least two drugs from the followingclasses: analgesics, non-steroidal anti-inflammatory drugs (NSAIDs), opioids, opioid-like drugs and neuromodulators (antidepressants,anticonvulsants and muscle relaxants)) compared with monotherapy, for adults with IA (RA, AS, PsA and other SpA). We speficicallyexcluded studies that did not report pain or studies without a standardised pain scale as an outcome measure.
1Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Two review authors independently selected trials for inclusion, assessed risk of bias and extracted data.
Main results
Twenty-three trials (total of 912 patients) met the inclusion criteria (22 in RA; one in a mixed population of RA and osteoarthritis);all except one were published before 1990. Most study populations were not taking DMARDs (e.g. methotrexate, sulphasalazine,hydroxychloroquine and leflunomide) and all studies were performed prior to the introduction of biologic therapies (e.g. etanercept,infliximab and adalimumab). All trials were at high risk of bias, heterogeneity precluded meta-analysis, and we were only able to reporta general description of results.
The majority (18 studies, 78%) found no differences between the combination and monotherapy treatments they studied, while five(22%) reported conflicting results, favouring either the combination or monotherapy arms.
From the 12 trials on NSAID + analgesic vs NSAID, nine reported no significant difference between the interventions, while three did: intwo, the combination therapy achieved better pain control; and the third trial compared combination therapy with two different dosagesof monotherapy (NSAID alone) and reported that a high dose phenylbutazone was superior to combination therapy (paracetamol +aspirin), which was superior to low dose phenylbutazone.
From the five studies on the combination of two NSAIDS vs one NSAID, four reported no significant differences between interventions,and one reported significantly better pain control with combination therapy.
The single trial comparing a combination of opioid + neuromodulator vs opioid reported better pain control with monotherapy.
The remaining trials (NSAID + neuromodulator vs NSAID (3 trials); opioid + NSAID vs NSAID (1 trial); and opioid + analgesic vsanalgesic (1 trial)) found no significant difference between combination therapy and monotherapy.
Information regarding withdrawals due to inadequate analgesia and safety was incompletely reported, but in general there were nodifferences between combination therapy and monotherapy.
No data were available that addressed the value of combination pain therapy or monotherapy for people with IA who have optimaldisease suppression. There were no studies that included patients with AS, PsA or SpA.
Authors’ conclusions
Based on 23 trials, all at high risk of bias, there is insufficient evidence to establish the value of combination therapy over monotherapyfor people with IA. Importantly, there are no studies addressing the value of combination therapy for patients with IA who havepersistent pain despite optimal disease suppression. Well designed trials are needed to address this question.
P L A I N L A N G U A G E S U M M A R Y
Combining two or more drugs vs one drug for pain control in inflammatory arthritis
This summary of a Cochrane review presents what we know from research about the effect of a combination of two pain relievingdrugs for pain control in inflammatory arthritis (IA).
We are uncertain if two pain relieving drugs such as paracetamol (also called acetaminophen) (e.g. Panadol® and Tylenol®) plusnon-steroidal anti-inflammatory drugs (NSAIDs), or paracetamol plus aspirin compared with one drug improved pain, because onlysingle studies of low quality evidence were available. For the same reason, we do not have precise information about side effects andcomplications.
What is IA, and what drugs are used to treat pain?
IA is a group of diseases that includes rheumatoid arthritis (RA), ankylosing spondylitis (AS), psoriatic arthritis (PsA) and otherspondyloarthritis (SpA). When you have IA, your immune system, which normally fights infection, attacks your joints. This makesyour joints swollen, stiff and painful. In RA, the small joints of your hands and feet are usually affected first. In contrast, in AS, the jointsof the spine are the most affected. PsA is characterised by inflammation of the skin, psoriasis, and joints and, depending on the diseasetype, can affect the small joints of the hands and feet or more the spine. There is no cure for IA at present, so the treatments aim torelieve pain and stiffness and improve your ability to move. Patients are started on disease-modifying antirheumatic drugs (DMARDs)
2Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
(e.g. methotrexate, sulphasalazine, hydroxychloroquine and leflunomide) as soon as possible in an attempt to control the inflammationand to prevent the progression of the disease. Many people continue to have pain despite optimal disease treatment and have the needfor specific medication to control pain.
Several drugs can be used to treat pain in IA. Paracetamol/acetaminophen, is used to relieve pain but does not affect swelling; NSAIDssuch as ibuprofen, diclofenac and COX-2s (e.g. celecoxib), are used to reduce pain and swelling; and opioids, such as codeine-containing Tylenol®, hydromorphone (Dilaudid), oxycodone (Percocet and Percodan), morphine and tramadol are powerful pain-relieving substances. Other drugs have some pain relieving properties and can therefore be used to mainly control pain. This is the case ofthe so-called neuromodulators, such as antidepressants (e.g. fluoxetine, paroxetine and amitriptyline), anticonvulsants (e.g. gabapentineand pregabaline) or muscle relaxants (e.g. diazepam). It is not clear if combining two of these drugs offers the best treatment and whichdrugs cause more side effects. It is known that, for instance, high doses of paracetamol/acetaminophen may cause stomach problems,such as ulcers, and NSAIDs may cause stomach, kidney or heart problems.
Best estimate of what happens to people with IA who take combination therapy for pain
There is insufficient evidence to establish the value of combination therapy over monotherapy for people with IA. We included 23studies in this review, all at high risk of bias (i.e. high chance of giving invalid results). Twenty-two of the trials were in patients withRA and one in a mixed population (RA and osteoarthritis). There were no studies in patients with AS, PsA or SpA. Included studieswere old (all but one were published before 1990) and patients were, in general, not on optimal disease-modifying antirheumatic drugs,as is standard current practice. Therefore, it is not possible to draw conclusions about the value of combination pain therapy overmonotherapy for people with IA. Importantly, there are no studies addressing the value of combination therapy for patients with IAwho have persistent pain despite optimal disease suppression. Well designed studies are needed to address this question.
B A C K G R O U N D
Description of the condition
Inflammatory arthritis (IA) is the term given to a group of chronicinflammatory rheumatic diseases that primarily include rheuma-toid arthritis (RA), ankylosing spondylitis (AS), psoriatic arthri-tis (PsA) and undifferentiated spondyloarthritis (SpA). Together,they have an estimated global prevalence of approximately 3%(Bergman 2006), comprising a prevalence of 1% for RA (Gabriel2001) and 1.9% for SpA (including AS and PsA) (Braun 1998).They are progressive diseases, characterised by pain, joint destruc-tion and decreased patient function (Bergman 2006). Further-more, these diseases have a profound impact on the patient’s qual-ity of life and on society in terms of medical costs and work dis-ability (Bergman 2006).The management of IA has dramatically changed over the lastdecade. The current approach focuses on early detection and man-agement at an early stage of the disease with disease-modifying an-tirheumatic drugs (DMARDs) (e.g. methotrexate, sulphasalazine,hydroxychloroquine and leflunomide) (Smolen 2010a), and withthe introduction of the efficacious biological disease-modifyingantirheumatic drugs (bDMARDs) (e.g. etanercept, infliximab andadalimumab), remission of disease is considered the appropriatetreatment target (Smolen 2010b). Nevertheless, despite these sig-
nificant advances, many patients with IA continue to experiencemusculoskeletal pain (Kvien 2004). Evidence from randomisedcontrolled trials (RCTs) indicates that bDMARDs improve mostoutcomes, except pain, an important patient-relevant outcome(Keystone 2004; Mease 2000; Van der Heijde 2005).Addressing pain is an integral part of patient care, which has beenrecognised by the Joint Commission on Accreditation of Health-care Organizations which has designated pain as the fifth vital sign,recommending it be monitored with the same vigilance as bloodpressure, pulse, temperature and respiratory rate (Lanser 2001;Philips 2000). Pain is one of the most commonly reported com-plaints, especially among patients with rheumatic diseases (Kasis1983) and studies have reported that patients with IA perceivepain to be their predominant impairment (Minnock 2003). Im-provement in pain has been identified by patients with IA to beamong their highest priorities (Da Silva 2010; Heiberg 2002), evenamongst those who have achieved adequate disease control andare being treated with biologicals (Ten Klooster 2007). Chronicpain has a significantly negative impact on quality of life and thiseffect increases with the duration of pain (Skevington 1998).
Description of the intervention
The primary goal of treating patients with IA (such as RA) is to
3Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
maximise long-term, health-related quality of life through controlof symptoms, prevention of structural damage, and normalisationof function and social participation (Smolen 2010b). Reducingpain is therefore, also one of the most important aims of ther-apy. Drug therapy to relieve pain may include analgesics, nons-teroidal anti-inflammatory drugs (NSAIDs), opioids, opioid-likedrugs (e.g. tramadol) and neuromodulators (including antidepres-sants, anticonvulsants and muscle relaxants). In treating patientswho have persistent pain, drugs from each of these classes can beprescribed as monotherapy, but may also be combined.
How the intervention might work
Combining drugs from different classes to treat persistent painmay achieve better pain control than monotherapy as a result ofadditive effects or synergistic mechanisms such that the final effectis more than the sum of the individual effects. On the other hand,combination therapy may also be associated with increased risks.Superiority of combination pain therapy could also be expectedfrom what we know about pain management in neuropathic pain.A recent review of pharmacological treatment in neuropathic painshowed evidence in favour of better pain control with combinationtherapy over monotherapy, for example through the combinationof an anticonvulsant with an opioid or with an antidepressant(Finnerup 2010).
Why it is important to do this review
Pain is a significant issue for people with IA, even in those whohave optimally controlled disease. It is not known whether specificcombinations of classes of drugs to treat pain are more effectivethan monotherapy to treat persistent pain in patients with IA orif adverse effects offset any benefits. The aim of this review wasto summarise the existing data on the efficacy and safety of com-bination therapy for pain management in patients with IA. Thiswill provide clinicians with information to guide their decisionson optimal pain management for patients with IA. The reviewwas undertaken according to the previously published protocol(Ramiro 2010). Differences between the protocol and review arenoted in the Differences between protocol and review section.
O B J E C T I V E S
To assess the benefits and safety of combination therapy for painmanagement in patients with IA.
M E T H O D S
Criteria for considering studies for this review
Types of studies
We included all published RCTs and quasi-randomised controlledclinical trials (CCTs) (i.e. where allocation was not truly ran-domised). We imposed no restrictions in length of follow-up orlanguage of the paper. We only included trials that were publishedas full articles or were available as a full trial report.We excluded studies that did not contain pain as an outcomemeasure or that did not include an understandable pain scale.
Types of participants
We selected studies that included adults of at least 18 years ofage with a clinical diagnosis of IA (RA, AS, PsA or SpA). Studiescontaining patients with other diagnoses were only eligible if theresults from patients with IA were presented separately.
Types of interventions
We included studies that compared combination therapy for painmanagement with monotherapy. We considered the followingdrug classes.
Combination therapy was defined as the combination of at leasttwo drugs of any of the drug classes presented. We included all thepossible variations (dosage, intensity, mode of delivery, frequencyof delivery, duration of delivery and timing of delivery).
Types of outcome measures
We chose the primary and secondary pain outcomes based uponthose currently recommended by the Cochrane Pain, Palliative andSupportive Care Systematic Review Group editors and others, forsystematic reviews on chronic pain (Moore 2010).
Main outcomes
1. Benefits: patient reported pain relief of 50% or greater.2. Safety: number of withdrawals due to adverse events.
4Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
1. Pain:◦ patient-reported pain relief of 30% or greater;◦ patient-reported global impression of clinical change
(PGIC) as “much” or “very much improved”;◦ proportion of patients achieving a pain score below
30/100mm on visual analogue scale; and◦ mean change in pain score on visual analogue scale or
numerical rating scale.2. Number of patients with adverse events.3. Number of deaths.4. Function - for example for RA, as measured by the Health
Assessment Questionnaire (HAQ) or modified HAQ (Fries1980; Pincus 1983), or by the Bath AS Functional Index(BASFI) (Calin 1994) for AS.
5. Quality of life - as measured by either generic instruments(such as the Short-Form-36 [SF-36]) (Ware 2001) or disease-specific tools, such as the RA Quality of Life instrument[RAQoL] (De Jong 1997) and AS Quality of Life Instrument[ASQoL]) (Doward 2003).
6. Participant withdrawals due to inadequate analgesia.If available in the included studies, we evaluated the outcomes forthe following end-points. We also chose these time intervals basedupon the recommendations of (Moore 2010).
We chose the short- and long-term outcomes for proportion re-porting pain relief of 50% or greater, total number of withdrawalsdue to adverse effects, number of adverse events, function andquality of life as the most important ones to be presented in thesummary of findings table.
Search methods for identification of studies
Electronic searches
We searched the following computerised bibliographicaldatabases: the Cochrane Central Register of Controlled Trials(CENTRAL) (The Cochrane Library Issue 2, 2010); MEDLINE(1950 to 4 May 2010); and EMBASE (EMBASE classic 1947 to1979 and EMBASE 1980 to 4 May 2010). We imposed no lan-guage restrictions and used the highly sensitive Cochrane Collab-oration search strategy, which aims to identify all RCTs (Lefebvre2011). We used specific MeSH headings and additional keywordsto identify all RCTs on combination therapy for pain managementin IA. The complete search strategies for the database searches areprovided in the Appendices (MEDLINE and CENTRAL searchstrategy Appendix 1, EMBASE search strategy Appendix 2). Inorder to retrieve additional references, we searched for systematic
reviews in the Cochrane Database of Systematic Reviews (CDSR)and the Database of Abstracts of Reviews of Effects (DARE) (TheCochrane Library Issue 2, 2010).
Searching other resources
We screened references from included RCTs and other systematicreviews on combination therapy for pain management in IA inorder to identify all possible studies for this systematic review. Wealso handsearched the conference proceedings for the AmericanCollege of Rheumatology (ACR) and the European League againstRheumatism (EULAR) for 2008 to 2010, to identify unpublishedstudies.
Data collection and analysis
Selection of studies
Two reviewers (SR, HR) independently assessed each title andabstract for suitability for inclusion in the review according tothe predetermined selection criteria (see Criteria for consideringstudies for this review). In cases of doubt, we retrieved and readthe full text article. We discussed disagreements between reviewauthors about the eligibility of the articles in a consensus meetingand if there was non-consensus, a third review author made adecision (RL).
Data extraction and management
Two reviewers (SR and HR) independently extracted the data re-garding study design, study duration, characteristics of study pop-ulation, interventions, outcome measures and timing of outcomeassessment, cointerventions, adverse effects, and loss to follow-up,by using a standardised data extraction form. We resolved differ-ences in data extraction by referring back to the original articlesand establishing consensus. We consulted a third review author(RL) to help resolve differences if needed.For studies published in languages other than English, German,Portuguese, French, Spanish or Dutch, we obtained the help ofa native speaker or translator with content and methodologicalexpertise.
Assessment of risk of bias in included studies
Two reviewers (SR and HR) independently assessed the risk ofbias of each included study with regard to the following items:random sequence generation; allocation concealment; blinding ofparticipants, care provider, and outcome assessor for each outcomemeasure (see Types of outcome measures); incomplete outcomedata; selective outcome reporting; and other potential sources ofbias, conforming to the methods recommended by the Cochrane
5Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Collaboration (Higgins 2011a). To determine the risk of bias ofa study, we evaluated the presence of sufficient information andthe likelihood of potential bias for each criterion. We rated eachcriterion as low risk of bias, high risk of bias, or unclear risk ofbias (either lack of information or uncertainty over the potentialfor bias). In a consensus meeting, we discussed and resolved dis-agreements among the review authors. If consensus could not bereached, a third review author (RL) made the final decision.
Measures of treatment effect
In order to assess efficacy, we extracted, if available, from the pub-lished reports, raw data for outcomes of interest (means and stan-dard deviations (SDs) for continuous outcomes, and number ofevents for dichotomous outcomes) as well as number of partic-ipants. If reported data needed to be converted or imputed, werecorded this in the notes section of the table Characteristics ofincluded studies. The results of each trial were to be plotted aspoint estimates with 95% confidence intervals (CIs). We plannedto present point estimates as risk ratios (RRs) for dichotomousoutcomes, and mean differences (MDs) for continuous outcomes.A RR > 1.0 would indicate a beneficial effect of combination ther-apy (Deeks 2011). RRs are considered clinically relevant if RR< 0.66 or > 1.5, in favour of the intervention or control group,respectively. This resembles an absolute difference of 25%. Forcontinuous data, results were to be analysed as MDs between theintervention and comparator group, with corresponding 95% CIs.The MD between treated group and control group is weighted bythe inverse of the variance in the pooled treatment estimate. How-ever, if different scales were used to measure the same conceptualoutcome (e.g. functional status or pain), we planned to calculatestandardised mean differences (SMDs) with corresponding 95%CIs instead. SMDs are calculated by dividing the MD by the SD,resulting in a unitless measure of treatment effect (Deeks 2011).SMDs > zero would indicate a beneficial effect in favour of com-bination therapy for pain management. SMDs can be interpretedas described by Cohen 1988; i.e. a SMD of 0.2 is considered toindicate a small beneficial effect, 0.5 a medium effect, and 0.8 alarge effect of combination therapy for pain management. SMDis considered to indicate a clinically relevant effect if SMD was> 0.5. Upon completion of the analysis, we planned to translateback any SMDs into a MD, on a scale of 0 to 10, which can bebetter appraised by clinicians.
Unit of analysis issues
In the event that we identified cross-over trials in which the report-ing of continuous outcome data precluded paired analysis, we didnot plan to include these data in a meta-analysis, in order to avoidunit-of-analysis error. Where carry-over effects were thought toexist, and where sufficient data existed, we only included data re-garding the first period in the analysis (Higgins 2011b).
For studies containing more than two intervention groups, mak-ing multiple pair-wise comparisons between all possible pairs ofintervention groups possible, if data were available we plannedto include the same group of participants only once in the meta-analysis.
Dealing with missing data
In cases where individuals were missing from the reported results,we assumed the missing values to have a poor outcome. For di-chotomous outcomes (e.g. number of withdrawals due to adverseevents), the withdrawal rate was calculated using the number ofpatients randomised in the group as the denominator (worst casescenario). For continuous outcomes (e.g. mean change in painscore), we calculated the MD or SMD based on the number of pa-tients analysed at that time point. If the number of patients anal-ysed was not presented for each time point, we used the numberof randomised patients in each group at baseline.Where possible, missing SDs were computed from other statisticssuch as standard errors (SEs), CIs or P values, according to themethods recommended in the Cochrane Handbook for SystematicReviews of Interventions (Higgins 2011c). If we could not calculateSDs, we imputed them, for example from other studies in themeta-analysis (Higgins 2011b).
Assessment of heterogeneity
We first assessed the studies for clinical homogeneity with respectto the condition under study, intervention and control groups(drug classes being combined and compared), and the outcomesand timing of outcome assessment.For clinically homogeneous studies, we planned to test statisticalheterogeneity using the I2 statistic (Deeks 2011), using the follow-ing as a rough guide for interpretation: 0% to 40% might not beimportant; 30% to 60% may represent moderate heterogeneity;50% to 90% may represent substantial heterogeneity; and 75% to100% considerable heterogeneity. In cases of considerable hetero-geneity (defined as I2
≥75%), we had planned to explore the datafurther, including subgroup analyses, in an attempt to explain theheterogeneity.
Assessment of reporting biases
An assessment of reporting biases was planned through the screen-ing of the Clinical Trial Register at the International Clinical Tri-als Registry Platform of the World Health Organization (http://apps.who.int/trialssearch; De Angelis 2004) to determine whetherthe protocol of the RCT was published before recruitment of pa-tients of the study was started. Furthermore, we planned a com-parison between the fixed-effect estimate and the random-effectsestimate, as well as a funnel plot, if data were available, in orderto assess the possible presence of small sample bias and reportingbias, respectively.
6Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
We planned to pool clinically homogeneous studies using the ran-dom-effects model.
Subgroup analysis and investigation of heterogeneity
We had planned the following subgroup analyses if data were avail-able.
1. Effects of age, gender, and disease duration.2. Difference in effects between the diagnoses (RA, AS, PsA,
undifferentiated SpA).3. Comparison of interventions belonging to the same drug
classes (e.g. combination of NSAID + analgesic vs analgesic).4. Difference in effects between patients on background
DMARD therapy and patients on no background therapy.
Sensitivity analysis
We had planned sensitivity analyses for studies with regard toallocation concealment, blinding of outcome assessor, and loss tofollow-up compared to studies with study limitations (low risk ofbias versus high risk of bias or unclear risk of bias). However therewere insufficient data for meta-analysis or sensitivity analysis.
Summary of findings table
The main results of the review were planned to be presented ina ’Summary of findings’ table, which includes an overall gradingof the evidence using the GRADE approach (GRADEpro) anda summary of the available data on the main outcomes (short-and long-term outcomes for proportion reporting pain relief of50% or greater, total number of withdrawals due to adverse effects,number of adverse events, function and quality of life), as described
in the Cochrane Handbook for Systematic Reviews of Interventions(Schünemann 2011).Grading of the evidence involves consideration of within-studyrisk of bias, directness of evidence, heterogeneity, precision of effectestimates and risk of publication bias. However, other factors canaffect the quality of evidence, for example it can be increasedby a large magnitude of effect, plausible confounding, and dose-response gradients. Using this system, we planned to grade thequality of the body of evidence as ’high’, ’moderate’, ’low’ or ’verylow (Atkins 2004).
R E S U L T S
Description of studies
See: Characteristics of included studies; Characteristics of excludedstudies.See: Characteristics of included studies; Characteristics of excludedstudies
Results of the search
We found a total of 14,854 articles in the three databases andobtained an additional 160 meeting abstracts from conferenceproceedings (Figure 1). After de-duplication and title and abstractscreening, we excluded 14,788 studies, leaving 66 articles for fullpaper review. The main reason for exclusion was wrong studytype or wrong intervention. We read the 66 articles in detail andexcluded 43 of them (see Excluded studies). We did not includeany meeting abstracts, not even for detailed review. In the end, weincluded 23 articles (see Included studies).
7Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Figure 1. Flowchart results from the systematic literature search
Included studies
We included a total of 23 trials involving a total of 912 pa-tients (range 12 to 134, mean 40 patients) with a study durationranging from 2 days to 5 months. A full description of the in-cluded studies is provided in the table, Characteristics of includedstudies. Other than the trial by Seideman 1993, all trials werepublished before 1990. Twenty-two trials were reported in En-glish and one trial was reported in German (Brooks 1975). Interms of study design there were five parallel RCTs (Brooks 1975;Coigley 1975; Puttini 1988; Sperryn 1973; Staunton 1980), oneCCT (Beales 1972) and the remaining 17 studies were cross-overtrials (Bedi 1969; Ekstrand 1981; Furst 1987; Grennan 1979;Haslock 1971; Hingorani 1973; Hobkirk 1977; Huskisson 1974;Kean 1981; Malcolm 1974; Mavrikakis 1977a; Mavrikakis 1977b;Mowat 1979; Saarialho-Kere 1988; Seideman 1988; Seideman1993; Sharma 1978). Twenty-two trials included RA study pop-ulations, while one trial included a mixed population of patientswith RA and osteoarthritis (Coigley 1975). For the latter trial, wehave included only the results for the RA subset in this review. We
did not find any trials for other types of IA.All included studies recruited adult participants. Six studies (Furst1987; Hobkirk 1977; Kean 1981; Seideman 1988; Seideman1993; Sharma 1978) reported the mean age of the wholestudy population, which ranged from 44 to 56 years. Eightstudies (Brooks 1975; Hobkirk 1977; Kean 1981; Mavrikakis1977a; Mavrikakis 1977b; Saarialho-Kere 1988; Seideman 1988;Seideman 1993) reported the mean duration of disease for thewhole population, which ranged from 4 to 10 years. The de-gree of disease activity of participants was not reported. Theproportion of females in the study population was reported innine studies (Furst 1987; Hobkirk 1977; Kean 1981; Mavrikakis1977a; Mavrikakis 1977b; Puttini 1988; Saarialho-Kere 1988;Seideman 1993; Sharma 1978) and varied between 55% and 89%.Only six of the 23 studies provided information about concur-rent DMARD therapy. One trial reported that no antirheumaticdrugs were allowed (Brooks 1975), and the other five studies(Coigley 1975; Furst 1987; Saarialho-Kere 1988; Seideman 1988;Seideman 1993) reported allowance of stable doses of DMARDs,
8Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
such as gold, penicillamine, choloroquine or steroids. No informa-tion was given about the proportion of patients taking DMARDtherapy. None of the participants in the included studies was re-ceiving biological DMARDs.The interventions used in the trials were very heterogeneous: dif-ferent drug classes, different drugs and different dosages. Sev-eral of the interventions studied are not in current practice, asfor instance, benorylate or safapryn, which are a combinationof aspirin and paracetamol (Beales 1972; Brooks 1975; Coigley1975; Haslock 1971; Hingorani 1973; Malcolm 1974; Mavrikakis1977a; Mavrikakis 1977b; Mowat 1979; Sperryn 1973). Thedosages were frequently suboptimal according to current recom-mendations (Dougados 2011), in terms of their analgesic effect,for example, ibuprofen 1200 mg/day (Hingorani 1973) or in-domethacin 100 mg/day (Hobkirk 1977). In several of the stud-ies, the monotherapy drug was not part of the combination ther-apy, for example, one trial compared phenylbutazone to aspirinand paracetamol (Brooks 1975) while another trial comparednaproxen to aspirin and paracetamol (Mowat 1979). In other tri-als, despite the monotherapy drug being part of the combinationtherapy, their dosages differed: for example, one trial comparedparacetamol 650 mg and dextropropoxyphene 65 mg to parac-etamol 1000 mg (Huskisson 1974) while another trial compareddextropropoxyphene 65 mg and amitriptyline 25 mg to dextro-propoxyphene 130 mg (Saarialho-Kere 1988).Categorising the interventions of the trials according to the drugclasses being combined, the most prevalent combination wasNSAID + analgesic compared to a NSAID in 12 trials (Beales1972; Brooks 1975; Coigley 1975; Haslock 1971; Hingorani1973; Malcolm 1974; Mavrikakis 1977a; Mavrikakis 1977b;Mowat 1979; Seideman 1988; Seideman 1993; Sperryn 1973),followed by the combination of two NSAIDs vs one NSAID infive trials (Ekstrand 1981; Furst 1987; Grennan 1979; Kean 1981;Staunton 1980) and then a combination of NSAID + neuromod-ulator vs NSAID alone in 3 trials (Hobkirk 1977; Puttini 1988;Sharma 1978). Other combinations were only reported in onetrial each: opioid + NSAID vs NSAID alone (Bedi 1969), opioid+ analgesic vs analgesic alone (Huskisson 1974) and opioid + neu-romodulator vs opioid alone (Saarialho-Kere 1988).Of the 23 included trials, 12 primarily reported pain on a vi-sual analogue scale (VAS) (Ekstrand 1981; Furst 1987; Grennan1979; Hobkirk 1977; Mavrikakis 1977a; Mavrikakis 1977b;Mowat 1979; Puttini 1988; Saarialho-Kere 1988; Seideman 1993;Seideman 1988; Sharma 1978); 10 on a categorical scale (Beales1972; Brooks 1975; Coigley 1975; Haslock 1971; Hingorani1973; Huskisson 1974; Kean 1981; Malcolm 1974; Sperryn 1973;Staunton 1980); and one described the outcome as an ordinal scaleof improvement in pain (Bedi 1969).Studies were heterogeneous with respect to which time points werepresented and most studies failed to adequately report their re-sults. For example, three studies reported both baseline and end ofstudy pain outcomes (Haslock 1971; Puttini 1988; Saarialho-Kere
1988); seven only provided end of study pain outcomes (with-out baseline values) (Grennan 1979; Hobkirk 1977; Kean 1981;Mavrikakis 1977a; Mavrikakis 1977b; Seideman 1993; Sharma1978); six trials reported baseline pain values and change scoresonly (Beales 1972; Coigley 1975; Furst 1987; Hingorani 1973;Mowat 1979; Sperryn 1973); mean values for pain throughoutthe study were reported in two trials (Ekstrand 1981; Seideman1988); a change score only was presented in three studies (Bedi1969; Huskisson 1974; Malcolm 1974); and one trial reportedbaseline pain and mean pain over the course of the trial (Brooks1975). One trial did not provide any numerical outcome data(Staunton 1980).Most trials did not report the time period that participants wereasked to consider when assessing their pain. No study presenteddata for our primary efficacy endpoint (patient reported pain reliefof 50% or greater) and no study reported any other dichotomouspain outcomes.Physical function was only assessed in three trials (Furst 1987;Hingorani 1973; Seideman 1993) measured using functional ca-pacity (Seideman 1993) or activities of daily living score (Furst1987; Hingorani 1973) and none of the trials assessed quality oflife.
Excluded studies
We excluded a total of 43 studies after detailed review. Reasons forexclusion are described in the Characteristics of excluded studiestable. We excluded four studies due to including a mixed pop-ulation without reporting results for patients with IA separately(Jaffé 1973; Lewis-Faning 1972; Mitchell 1984; Murphy 1978).We excluded nine trials as they reported a comparison of twomonotherapies (Barnardo 1966; Lynch 2001; Maldykowa 1983;Pavelka 1972; Perrot 2006; Raptopoulou 2008; Sasisekhar 1973;Sidiropoulos 2008; Vergne-Salle 2009); one trial as it reported acomparison of two triple therapy interventions (Glowinski 1999);and one trial because it compared two different combination ther-apies (Maneksha 1973).We excluded 15 trials due to absence of an outcome of interest.Eight trials did not include a pain assessment (Andrade-Padilla1995; Dalmases 1966; Franke 1972; Jeremy 1970; Rudge 1982;Torgyan 1979; Van Hoek 1973; Willkens 1976); six trials didnot include a global pain assessment (Bain 1970; Cardoe 1970;Chalmers 1978; Ridolfo 1982; Robinson 1975; Roth 1975); andin one trial the reported pain scale was not defined or under-standable (Brooks 1977). Finally, we excluded 13 studies be-cause of wrong study type (De Mattos 1968; Eberl 1968; Famaey1971; Famaey 1972; France 1968; Hernandez Pena 1973; Hersh2007; Huss 1974; Lopez Prats 1968; Moll 1966; Triandaf 1970a;Triandaf 1970b; Wettreich 1966).
Risk of bias in included studies
9Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
We considered all included studies to be at high risk of bias. Adescription of the risk of bias of the included studies is presentedin the Risk of Bias tables (Characteristics of included studies). Arisk of bias graph and a risk of bias summary can be seen in Figure2 and Figure 3, respectively.
Figure 2. Risk of bias graph: review authors’ judgements about each risk of bias item presented as
percentages across all included studies.
10Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
We did not consider any trial to have an adequate sequence gener-ation method. We classified one trial as having an inadequate se-quence generation as patients were all allocated to the same treat-ment order in a cross-over trial (Furst 1987), and in all the othertrials the adequacy of the sequence generation was unclear, due tolack of information reported in the paper.We only considered one trial to have an adequate allocation con-cealment (Mowat 1979), one had an inadequate allocation con-cealment (Furst 1987) and all the others were unclear with respectto this item. We evaluated blinding as low risk of bias in two studies(Seideman 1988; Seideman 1993), as high risk of bias in three stud-ies (due to lack of blinding of participants) (Beales 1972; Coigley1975; Mowat 1979), and all the other trials were unclear. Onetrial presented complete outcome data (Seideman 1993), 13 stud-ies had a high risk of bias on the incomplete outcome data item,mainly because of excluding withdrawals from the analyses (Beales1972; Bedi 1969; Grennan 1979; Haslock 1971; Hingorani 1973;Hobkirk 1977; Malcolm 1974; Mavrikakis 1977a; Mavrikakis1977b; Puttini 1988; Sharma 1978; Sperryn 1973; Staunton1980), and the other trials were unclear on this item. We did notconsider any study to have a low risk of bias in both selective re-porting and other potential sources of bias; the majority of thetrials were considered to have a high risk of bias, and some wereunclear.
Effects of interventions
Due to multiple sources of heterogeneity, a meta-analysis couldnot be performed and we present a narrative summary of pertinentfindings from the individual studies. The main reported findingsfor the individual trials are also summarised in the Notes sectionsin the Table, Characteristics of included studies.
EFFICACY
The majority of studies (18/23, 78%) reported no differences inoutcome between the combination and monotherapy treatmentsthey studied, while five (22%) reported conflicting results, favour-ing either the combination or monotherapy arms.
NSAID + ANALGESIC VS NSAID (12 studies)
Nine studies (75%) reported no differences between the combi-nation and monotherapy treatments with respect to pain (Beales1972; Haslock 1971; Hingorani 1973; Malcolm 1974; Mavrikakis1977a; Mavrikakis 1977b; Mowat 1979; Seideman 1988; Sperryn1973). The other three trials reported a significant differencebetween the treatment groups (Brooks 1975; Coigley 1975;Seideman 1993). Two of these trials demonstrated better pain con-trol with combination therapy: Seideman 1993 reported a lowerpain score at two weeks in participants who received naproxen1000 mg + paracetamol 4 g per day compared to those who
received naproxen 1000 mg alone (mean pain score (SD)(0 to100 VAS) 31.7 (9.6) vs 46.5 (14.6) respectively); P value < 0.05;Coigley 1975 described an overall mean improvement in pain (%)at two weeks of 73% in those who received paracetamol + aspirin(benorylate = 8 g) compared with 32% in those who received in-domethacin 75 mg alone; P value < 0.05.The third trial (Brooks 1975) comparing combination therapywith two different dosages of monotherapy found conflicting re-sults depending on the dose of the monotherapy. This study com-pared paracetamol 3 g and aspirin 3.6 g per day to either 50 mg or300 mg phenylbutazone and reported that high dose phenylbuta-zone was superior to combination therapy which was superior tolow dose phenylbutazone (pain score over two weeks (adjusted forbaseline pain score): 2.8 (0.2) on 300 mg phenylbutazone versus3.1 (0.2) on paracetamol 3 g and aspirin 3.6 g and 3.3 (0.2) onphenylbutazone 50 mg; P value < 0.05).
NSAID + NSAID VS NSAID (5 studies)
Four studies (Furst 1987; Grennan 1979; Kean 1981; Staunton1980) found no significant differences between the treatmentarms. One study (Ekstrand 1981) reported a median pain score(0 to 100 VAS) over three weeks of 43 for those who receivedacetylsalicylic acid 2 g + indomethacin 50 mg compared with ascore of 53 for those who received acetylsalicylic acid 2 g alone (nobaseline values were reported); P value < 0.05.
NSAID + NEUROMODULATOR VS. NSAID (3 studies)
All trials (Hobkirk 1977; Puttini 1988; Sharma 1978) reportedno significant differences between combination therapy andmonotherapy.
OPIOID + NSAID VS NSAID (1 study)
Bedi 1969 found no significant difference between the treatmentarms.
OPIOID + ANALGESIC VS ANALGESIC (1 study)
Huskisson 1974 found no significant difference between the treat-ment arms.
OPIOID + NEUROMODULATOR VS OPIOID ( 1 study)
Saarialho-Kere 1988 reported worse pain control with a combina-tion of dextropropoxyphene 65 mg + amitriptyline 25 mg versusdextropropoxyphene 130 mg alone (mean (SD) pain score (0 to100 VAS) at baseline and 4 hours 50 (6.6) and 44 (6.6) versus 46
12Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
(6.0) and 34 (4.8), respectively); P value < 0.05. However, partici-pants in the monotherapy arm received double the dose of opioidcompared to those in the combination therapy arm.
FUNCTION AND WITHDRAWALS DUE TO
INADEQUATE ANALGESIA
Of the three trials that reported function, none found a significantdifference between combination and monotherapy.Withdrawals due to inadequate analgesia were incompletely as-sessed and/or reported in the included studies. Eight trials did notreport this outcome (Huskisson 1974; Kean 1981; Malcolm 1974;Mavrikakis 1977a; Mavrikakis 1977b; Puttini 1988; Saarialho-Kere 1988; Seideman 1993), seven studies reported no with-drawals due to inadequate analgesia in both study arms, and the re-maining eight trials reported some withdrawals due to inadequateanalgesia in at least one of the study arms (Beales 1972; Brooks1975; Coigley 1975; Furst 1987; Grennan 1979; Haslock 1971;Mowat 1979; Sperryn 1973). A comparison of the withdrawalsdue to inadequate analgesia between the study groups (performedby the authors of the review) revealed no significant differencesbetween the interventions, except for one comparison within oneof the trials (Brooks 1975). One of this study’s monotherapy arms(phenylbutazone 50 mg) had a significantly higher number ofwithdrawals due to inadequate analgesia (2% under combinationtherapy of aspirin + paracetamol vs 33% under phenylbutazone50 mg in monotherapy).
SAFETY
Withdrawals due to adverse events, our primary safety outcome,was incompletely reported in the included trials. Eight trials didnot report this outcome at all, five trials reported that there wereno withdrawals due to adverse events and the remaining ten trialsreported a few withdrawals due to adverse events in at least oneof the study arms (Beales 1972; Bedi 1969; Brooks 1975; Coigley1975; Furst 1987; Grennan 1979; Hobkirk 1977; Malcolm 1974;Mowat 1979; Sperryn 1973). For these ten trials, there was nosignificant difference in the proportion of withdrawals due to ad-verse events between monotherapy and combination therapy arms(analysis performed by the authors of the review).No deaths were reported in the eighteen trials that either directlyor indirectly reported on this outcome.Of the ten trials that reported on the number of participantswith adverse events (Bedi 1969; Coigley 1975; Haslock 1971;Hingorani 1973; Hobkirk 1977; Malcolm 1974; Mowat 1979;Seideman 1988; Sharma 1978; Sperryn 1973) all but one reportedno differences between the combination and monotherapy arms.Seideman 1988 reported that 55% (11/20) of those who receivedcombination therapy (indomethacin 50 mg + paracetamol 4 g)had at least one adverse event compared to 20% (4/20) of thosewho received indomethacin 150 mg alone, P value = 0.05.
Due to lack of data, subgroup or sensitivity analyses could notbe performed and a ’Summary of Findings’ table could not beconstructed.
D I S C U S S I O N
Summary of main results
Overall, based on 23 trials, it is not possible to establish the effi-cacy and safety of combination therapy for pain management inpeople with IA. Pooling of data was not possible and the resultshave been presented descriptively. Five of the trials reported a sig-nificant difference in pain control between combination therapyand monotherapy: in three trials combination therapy was better,in one trial monotherapy was better and the fifth trial reportedmixed results depending upon the dosage used in the monother-apy arm. Only three studies reported functional outcomes, andin neither case was there a difference between monotherapy andcombination therapy. None of the included studies reported qual-ity of life data. Statistically significant differences in safety betweencombination and monotherapy were not reported.
Overall completeness and applicability ofevidence
We did not find any studies of combination pain therapy for AS,PsA and undifferentiated SpA.For RA, there were 23 studies, all published between 1969 and1993, preceding the significant advances that have occurred intherapeutics subsequently. Most study populations were not tak-ing DMARDS, none were exposed to biologic therapy suggestingthat the majority had active disease and the primary pain sourcewas likely to be of inflammatory origin. It is likely that the resultsof these studies cannot be transposed to current clinical practicewithout consideration. In addition, several of the trials includeddrugs, such as benorylate (a combination of aspirin and paraceta-mol) and anti-inflammatory doses of aspirin, that are no longer incommon usage.
Quality of the evidence
The risk of bias of all included trials was high. Generation of anadequate randomisation sequence and concealment of treatmentallocation were poorly performed and/or reported, and partici-pants with missing data were often excluded from the analysis.Furthermore, the included trials were very small and significantresults from small trials can easily be wrong (Peto 1976); while 17of the 23 included studies were cross-over trials, which imply thepossibility of carry-over and period-effects.
13Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
The interventions studied were heterogenous in terms of drugcombinations, treatment duration, drug class, drugs and dosages.The comparator was also a source of heterogeneity and in severalstudies the monotherapy drug was not part of the combinationtherapy, precluding meaningful comparison for the purpose of ourreview.The outcomes measured, how and when they were measured andhow the results were reported also varied widely between trials.For example some trials measured pain while others measuredimprovement in pain; some used visual analogue scales for painwhile others used a categorical scale; and some trials did not reportbaseline and/or, end values and/or change scores.None of the included trials reported any of the four dichotomouspain outcomes we planned to include and that were recommendedby the Cochrane Pain, Palliative and Supportive Care System-atic Review Group editors and others, for systematic reviews onchronic pain (Moore 2010). Group means for pain continuousmeasures are difficult to interpret in terms of their clinical relevanceas the underlying distribution is often skewed (Moore 2010).
Potential biases in the review process
We believe that we identified all relevant studies. We devised athorough search strategy and searched all major databases for rele-vant studies, and applied no language restrictions. Two review au-thors assessed the trials for inclusion in the review and risk of bias,with a third reviewer adjudicating if there was any discrepancy.The biggest limitation of the review process was the heterogeneitybetween the trials and the lack of data in a form that could beextracted for meta-analysis.
Agreements and disagreements with otherstudies or reviews
We did not identify any other review on combination therapy forpain management in IA. However a systematic review of pharma-cological treatment for neuropathic pain also considered the role ofcombination therapy in controlling pain (Finnerup 2010). Theyreported that the combination of an anticonvulsant or an antide-pressant with an opioid achieved better pain control compared tomonotherapy. These findings may be generalisable to patients withIA as similarly to persistent pain in patients with IA, neuropathicpain can be very disabling, severe and intractable (Saarto 2007),although this needs to be confirmed in high quality RCTs.
A U T H O R S ’ C O N C L U S I O N S
Implications for practice
Notwithstanding the dramatic advances that have occurred in themanagement of IA in the last decade, pain remains an importantissue for patients. However, there is currently insufficient data todraw conclusions about the efficacy and safety of combination paintherapy in the management of patients with IA. Importantly, thereare no studies addressing the value of combination pain therapy forpatients with IA who have persisting pain despite optimal diseasesuppression.
Implications for research
More evidence regarding the efficacy and safety of combinationtherapy for pain management in IA is required. Well-designedRCTs in RA as well as other IAs including AS and PsA are neededto address this question. To be of relevance to current practice,included patients should have persistent pain despite optimal dis-ease suppression with DMARD and/or biologic DMARDs. Trialsshould seek to compare the risk:benefit profile of different combi-nation analgesic strategies, different drug classes being combined,different routes of administration and different intervals.
Future trials need to be rigorous in design and delivery, undertakenwith high quality methodology that is described in enough detailto enable appraisal and interpretation of results. They need tohave adequate power and be of sufficient duration (at least 12weeks), report their methods of sequence generation and allocationconcealment, and blind participants (especially important whenthe primary outcome of interest is pain), physicians and outcomeassessors, and have an appropriate method for handling incompletedata.
Clinically-relevant pain outcomes should be collected, with spe-cial emphasis on including dichotomous pain outcomes as recom-mended by the Cochrane Pain, Palliative and Supportive Care Re-view Group editors and others (Moore 2010). Function and qual-ity of life are important outcomes that should also be addressedin trials focusing on pain treatment. Lastly, a rigorous analysis ofadverse events is indispensable to enable a comparison of the risk:benefit profile of combination analgesic therapy.
A C K N O W L E D G E M E N T S
The authors thank Louise Falzon, Trials Search Coordinator of theCochrane Musculoskeletal Group, for assisting with the design ofthe search strategy.
14Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Beales DL, Burry HC, Grahame R. Comparison of aspirinand benorylate in the treatment of rheumatoid arthritis.BMJ 1972;2(5812):483–5.
Bedi 1969 {published data only}
Bedi SS. Comparison of aspirin and dextropropoxyphene-with-aspirin as analgesics in rheumatoid arthritis. BritishJournal of Clinical Practice 1969;23(10):413–7.
Brooks 1975 {published data only}
Brooks PM, Walker JJ, Lee P, Bell AM, Buchanan WW,Fowler PD, et al.Clinical study on a new acetylsalicylicacid/paracetamol preparation with gastric acid resistantcoating (Safapryn), and on two various phenylbutazonedosages in patients with primary chronic polyarthritisas based on a new evaluation method [Erprobung einesneuen acetylsalicylsäure/paracetamol–präparates mitmagensaftresistentem Überzug (Safapryn) und zweiverschiedenen Dosierungen von Phenylbutazon beiPatienten mit primär chronischer Polyarthritis anhand einesneuen Bewertungsverfahrens]. Zeitschrift fur Rheumatologie
1975;34(9-10):350–65.
Coigley 1975 {published data only}
Coigley MH. Comparison of benorylate and indomethacinin the symptomatic control of arthritic disorders.Practitioner 1975;215(1287):348–52.
Ekstrand 1981 {published data only}
Ekstrand R, Alvan G, Magnusson A, Oliw E, PalmerL, Rane A. Additive clinical effect of indomethacinsuppositories during salicylate therapy in rheumatoidpatients. Scandinavian Journal of Rheumatology 1981;10(2):69–75.
Furst 1987 {published data only}
Furst DE, Blocka K, Cassell S, Harris ER, Hirschberg JM,Josephson N, et al.A controlled study of concurrent therapywith a nonacetylated salicylate and naproxen in rheumatoidarthritis. Arthritis and Rheumatism 1987;30(2):146–54.
Grennan 1979 {published data only}
Grennan DM, Ferry DG, Ashworth ME, Kenny RE,Mackinnon M. The aspirin-ibuprofen interactionin rheumatoid arthritis. British Journal of ClinicalPharmacology 1979;8(5):497–503.
Haslock 1971 {published data only}
Haslock DI, Nicholson PA, Wright V. A comparison ofphenylbutazone and benorylate. Clinical Trials Journal
1971;8(1):43–50.
Hingorani 1973 {published data only}
Hingorani K. Double blind study of benorylate andibuprofen in rheumatoid arthritis. Rheumatology and
Rehabilitation 1973;12(suppl):39–47.
Hobkirk 1977 {published data only}
Hobkirk D, Rhodes M, Haslock I. Night medicationin rheumatoid arthritis: II. Combined therapy with
indomethacin and diazepam. Rheumatology & Rehabilitation1977;16(2):125–7.
Huskisson 1974 {published data only}
Huskisson E. Simple analgesic for arthritis. British MedicalJournal 1974;4(5938):196–200.
Kean 1981 {published data only}
Kean WF, Kraag GR, Rooney PJ, Capell HA. Clinicaltherapeutic trial of aspirin and azapropazone in rheumatoidarthritis when prescribed singly and in combination.Current Medical Research and Opinion 1981;7(3):164–7.
Malcolm 1974 {published data only}
Malcolm A. A comparative clinical trial of benorylateand indomethacin in the treatment of active rheumatoidarthritis in general practice. Journal of International Medical
Research 1974;2(1):66–70.
Mavrikakis 1977a {published data only}
Mavrikakis ME, McLeod M, Hernandez LA, El-GhobareyAF, Cappel HA. Comparison of benorylate tablets withibuprofen in rheumatoid arthritis. Pharmatherapeutica
1977;1(9):621–4.
Mavrikakis 1977b {published data only}
Mavrikakis ME, Madkour MM, Spencer DM, Balint GP.Double-blind study comparing the therapeutic effect ofa new non-steroidal anti-inflammatory drug, sulindac,with benorylate in tablet form in rheumatoid arthritis.Pharmatherapeutica 1977;1(10):681–5.
Mowat 1979 {published data only}
Mowat AM, Mowat AG. A comparative trial ofnaproxen and benorylate suspensions in the treatment ofrheumatoid arthritis. European Journal of Rheumatology and
Inflammation 1979;2(1):74–8.
Puttini 1988 {published data only}
Puttini P, Cazzola M, Boccassini L, Ciniselli G, SantandreaS, Caruso I, et al.A comparison of dothiepin versus placeboin the treatment of pain in rheumatoid arthritis andthe association of pain with depression. The Jounal of
International Medical Research 1988;16(5):331–7.
Saarialho-Kere 1988 {published data only}
Saarialho-Kere U, Julkunen H, Mattila M, Seppälä T.Psycomotor performance of patients with rheumatoidarthritis: cross-over comparison of dextroproproxyphene,dextropropoxyphene plus amitriptyline, indomethacin, andplacebo. Pharmacology and Toxicology 1988;63(4):286–92.
Seideman 1988 {published data only}
Seideman P, Melander A. Equianalgesic effects ofparacetamol and indomethacin in rheumatoid arthritis.British Journal of Rheumatology 1988;27(2):117–22.
Seideman 1993 {published data only}
Seideman P. Additive effect of combined naproxen andparacetamol in rheumatoid arthritis. British Journal of
Rheumatology 1993;32(12):1077–82.
15Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Sharma BK, Haslock I. Night medication in rheumatoidarthritis. III. the use of sulindac. Current Medical Research
and Opinion 1978;5(6):472–5.
Sperryn 1973 {published data only}
Sperryn PN, Hamilton EB, Parsons V. Double-blindcomparison of aspirin and 4-(acetamido) phenyl-2-acetoxy-benzoate (benorylate) in rheumatoid arthritis. Annals of the
Rheumatic Diseases 1973;32(2):157–61.
Staunton 1980 {published data only}
Staunton Smith G, Moran CJ. Diflunisal plus indomethacinas concomitant therapy in rheumatoid arthritis. EuropeanJournal of Rheumatology and Inflammation 1980;3(3):187–93.
References to studies excluded from this review
Andrade-Padilla 1995 {published data only}
Andrade-Padilla JM, Aguilar JA, Reynoso P, RomeroGuillermo, Oritiz D. Double-blind trial with ketoprophene+ chlorzoxazone and ketoprophene in most frequentosteomuscular problems in the outpatient practice [Estudiodoble ciego con ketoprofeno + clorzoxazona y ketoprofenoen problemas osteomusculares más frecuentes en la consultaexterna]. Investigacion Medica Internacional 1995;22:52–7.
Bain 1970 {published data only}
Bain LJS, Burt RAP. The treatment of rheumatoid disease.A double blind trial comparing buffered aspirin withbenorylate. Clinical Trials Journal 1970;7(2):307–12.
Barnardo 1966 {published data only}
Barnardo DE, Currey HL, Mason RM, Fox WR, WeatherallM. Mefenamic acid and flufenamic acid compared withaspirin and phenylbutazone in rheumatoid arthritis. BritishMedical Journal 1966;2(5509):342–3.
Brooks 1977 {published data only}
Brooks PM, Khong TK. Flurbiprofen-aspirin interaction: adouble-blind crossover study. Current Medical Research andOpinion 1977;5(1):53–7.
Cardoe 1970 {published data only}
Cardoe N. The treatment of rheumatoid disease.Preliminary assessment of a new drug: benorylate. Clinical
Trials Journal 1970;7:173–8.
Chalmers 1978 {published data only}
Chalmers A, Offer R, Robinson HS. A double-blindcontrolled study comparing the use of ASA and tolmetinwith ASA and placebo in the treatment of rheumatoidarthritis. Current Therapeutic Research - Clinical andExperimental 1978;24:517–23.
Dalmases 1966 {published data only}
Dalmases R, Heredia J, Cots R. Association of butazolidinwith other non-steroid drugs in the treatment of chronicevolutive polyarthritis [Associación de butazolidina conotros medicamentos no esteroideos en el tratamiento dela P.C.P]. Revista Espanola de Reumatismo y Enfermedades
Osteo-Articulares 1966;11(8):40–3.
De Mattos 1968 {published data only}
De Mattos H. Use of paracetamol-phenylbutazonecombination in rheumatic diseases of professional motorists[Uso da associação paracetamol–fenilbutazona em afecçõesreumáticas de motoristas profissionais]. Hospital 1968;74
(2):543–8.
Eberl 1968 {published data only}
Eberl R. Additional treatment of rheumatologic diseaseswith Arcobutina ointment. Hippokrates 1968;39(18):710–1.
Famaey 1971 {published data only}
Famaey JP, Colinet E. Cooperative study of the combinationbufexamac-alclofenac in rheumatic diseases [Etude cliniquede l’association bufexamac – alclofenac dan less affectionsd’origine rhumatismale]. Journal Belge de Rhumatologie et de
Medecine Physique - Belgisch Tijdschrift voor Reumatologie enFysische Geneeskunde 1971;26(6):297–311.
Famaey 1972 {published data only}
Famaey JP, Bauduin MP, Wanet G. Combinedalclofenac-bufexamac in rheumatology [L’associationalclofenac–bufexamac en rhumatologie]. Journal Belge deRhumatologie et de Medecine Physique - Belgisch Tijdschrift
voor Reumatologie en Fysische Geneeskunde 1972;27(3):110–19.
France 1968 {published data only}
France O. Combination of ketophenylbutazone (Ketason),oxyphenbutazone and proteolytic enzymes in rheumaticdiseases [Associacion de Cetofenilbutazone (Ketazon)oxifenbutazona y enzimas proteoliticas en afeccionesreumaticas]. Revista Medica de Chile 1968;96(8):531–4.
Franke 1972 {published data only}
Franke M, Manz G. Benorylate and indomethacin inthe treatment of rheumatoid disease: a double-blindclinical trial. Current Therapeutic Research, Clinical and
Experimental 1972;14(3):113–22.
Glowinski 1999 {published data only}
Glowinski JBE. Placebo-controlled study of the analgesicefficacy of a paracetamol 500 mg/Codeine 30 mgcombination together with low-dose vs high-dose diclofenacin rheumatoid arthritis. Clinical Drug Investigation 1999;18
(3):189–97.
Hernandez Pena 1973 {published data only}
Hernandez Pena M. Clinical experience with anindomethacin-phenylbutazone combination in variousrheumatic diseases [Experiencia clinica con la asociacionindometacina–fenilbutazona en algunas afeccionesreumaticas]. Revista Espanola de Reumatismo y EnfermedadesOsteo-Articulares 1973;16(2):111–31.
Hersh 2007 {published data only}
Hersh EV, Pinto A, Moore PA. Adverse drug interactionsinvolving common prescription and over-the-counteranalgesic agents. Clinical Therapeutics 2007;29(11 Suppl):2477–97.
Huss 1974 {published data only}
Huss V. Combined drug therapy in highly acute attacksof pain in the course of rheumatic diseases [Kombinierte
16Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Jaffe G. Treatment of arthritis in general practice. Acomparison between benorylate and indomethacin. Journalof International Medical Research 1973;1(3):166–71.
Jeremy 1970 {published data only}
Jeremy R, Towson J. Interaction between aspirin andindomethacin in the treatment of rheumatoid arthritis.Medical Journal of Australia 1970;2(3):127–9.
Lewis-Faning 1972 {published data only}
Lewis Faning E, Wilkinson PM. The treatment ofrheumatic and degenerative arthritis. A double blind trial oftandalgesic and aspirin. Clinical Trials Journal 1972;9(2):23–9.
Lopez Prats 1968 {published data only}
Lopez Prats JL. Clinical tests of paracetamol-phenylbutazonecombination [Ensayo clinico de la associacion paracetamol–fenilbutazona]. Espanola de Reumatismo y EnfermedadesOsteo-Articulares 1968;12(8):233–43.
Lynch 2001 {published data only}
Lynch ME. Antidepressants as analgesics: A review ofrandomized controlled trials. Journal of Psychiatry andNeuroscience 2001;26(21):30–6.
Maldykowa 1983 {published data only}
Maldykowa H, Abgarowicz-Milkowska T, Dratwianka B,Grabowska A, Jedryka A. Results of treatment of patientswith rheumatoid arthritis with Voltaren SR and Metindol[Wyniki leczenia voltarenem SR i metindolem chorych nareumatoidalne zapalenie stawow]. Reumatologia 1983;21(3-4):265–70.
Maneksha 1973 {published data only}
Maneksha S. ’Safapryn’ and benorylate-a comparative trialof two new preparations of aspirin and paracetamol inthe treatment of rheumatoid arthritis and osteoarthritis.Current Medical Research & Opinion 1973;1(9):563–8.
Mitchell 1984 {published data only}
Mitchell H, Cunningham TJ, Mathews JD, MuirdenKD. Further look at dextropropoxyphene with or withoutparacetamol in the treatment of arthritis. Medical Journal ofAustralia 1984;140(4):224–5.
Moll 1966 {published data only}
Moll W. Therapeutic efficacy and tolerance of a combinedphenylbutazone-isopyrin preparation [TherapeutischeWirksamkeit und Verträglichkeit eines Phenylbutazon–isopyrin–kombinationspräparates]. Praxis 1966;55(49):1430–5.
Murphy 1978 {published data only}
Murphy JE, Donald JF, Layes Molla A. Analgesic efficacyand acceptability of fenoprofen combined with paracetamoland compared with dihydrocodeine tartrate in generalpractice. Journal of International Medical Research 1978;6(5):375–80.
Pavelka 1972 {published data only}
Pavelka K, Vojtisek O, Bremova A, Handlova D, KankovaD, Kralova M. Bufexamac in the treatment of progressivepolyarthritis. Double-blind test using phenylbutazone[Bufexamac v lécbe progresivni polyartritidy]. Casopis
Lekaru Ceskych 1972;111(22):514–7.
Perrot 2006 {published data only}
Perrot S, Maheu E, Javier R-M, Eschalier A, Coutaux A,LeBars M, et al.Guidelines for the use of antidepressantsin painful rheumatic conditions. European Journal of Pain
2006;10(3):185–92.
Raptopoulou 2008 {published data only}
Raptopoulou A, Sidiropoulos P, Siakka P, Boki K, DrososAA, Aslanidis S, et al.Evidence-based recommendationsfor the management of ankylosing spondylitis: resultsof the Hellenic working group of the 3E Initiative inRheumatology. Clinical and Experimental Rheumatology2008;26(5):784–92.
Ridolfo 1982 {published data only}
Ridolfo AS, Ashbrook EM, Schmid GE, Vogel JA,Rockhold FW, Offen WW. A double-blind study comparingbenoxaprofen, aspirin, and benoxaprofen plus aspirin inpatients with rheumatoid arthritis. European Journal of
Rheumatology and Inflammation 1982;5(2):239–45.
Robinson 1975 {published data only}
Robinson H, Abruzzo JL, Miyara A, Ward JR. Concomitanttolmetin and aspirin therapy for rheumatoid arthritis.Excerpta Medica 1975;10:102.
Roth 1975 {published data only}
Roth SH, Englund DW, Harris BK, Ross HA. Tolmetinwith acetaminophen in the treatment of rheumatoidarthritis. Excerpta Medica 1975;1:112.
Rudge 1982 {published data only}
Rudge SR, Lloyd-Jones JK, Hind ID. Interaction betweenflurbiprofen and indomethacin in rheumatoid arthritis.British Journal of Clinical Pharmacology 1982;13(31):448–51.
Sasisekhar 1973 {published data only}
Sasisekhar PR, Penn RG, Haslock I, Wright V. A comparisonof benorylate and aspirin in the treatment of rheumatoidarthritis. Rheumatology and Rehabilitation 1973;12(suppl):31–8.
Sidiropoulos 2008 {published data only}
Sidiropoulos PI, Hatemi G, Song IH, Avouac J, CollantesE, Hamuryudan V, et al.Evidence-based recommendationsfor the management of ankylosing spondylitis: systematicliterature search of the 3E Initiative in Rheumatologyinvolving a broad panel of experts and practisingrheumatologists. Rheumatology 2008;47(3):355–61.
Torgyan 1979 {published data only}
Torgyan S, Wagner L, Neumann T, Pakuts B, Csanyi M.A comparative study with indomethacin and combinedindomethacin sodium-salicylate in rheumatoid arthritis.International Journal of Clinical Pharmacology and
Biopharmacy 1979;17(11):439–41.
17Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Triandaf I, Bobulesco V. Association of chlorzoxazone,acetaminophen and prednisolone (Predniflex) in thetreatment of rheumatismal diseases [L’association dechlorzoxosone, acetaminophen et prednisolon (predniflex)dans le traitement des affections rhumatismales]. RevistaMedico-Chirurgicala a Societatii de Medici Si Naturalisti Din
Iasi 1970;74(3):723–7.
Triandaf 1970b {published data only}
Triandaf I, Bobulesco V. Clinical trial of an association of5-chlorobenzoxazolinone and N-acetyl-para-aminophenol(Parafon) [Considerations sur léxpérimentation cliniqued’une association entre le 5–chlorobenzoxazolinone et len–acetyl–para–aminophenol (parafon)]. Revista Medico-Chirurgicala a Societatii de Medici Si Naturalisti Din Iasi
1970;74(2):459–63.
Van Hoek 1973 {published data only}
Van Hoek J. Trial of benorilate in rheumatic diseases[Benorylaast getoest in de rheumatologische practijk]. Ars
Medici Internationaal Tijdschrift voor Praktische Therapie1973;2(10):1739–45.
Vergne-Salle 2009 {published data only}
Vergne-Salle P, Mejjad O, Javier RM, Maheu E, FallutM, Glowinski J, et al.Antiepileptic drugs to treat painin rheumatic conditions. Recommendations based onevidence-based review of the literature and expert opinion.Joint, Bone, Spine: Revue du Rheumatisme 2009;76(1):75–85.
Wettreich 1966 {published data only}
Wettreich W. The association of paracetamol andphenylbutazone in rheumatology. Preliminary report [Aassociação paracetamol–fenilbutazona em reumatologia].Hospital 1966;70(2):427–32.
Willkens 1976 {published data only}
Willkens RF, Segre EJ. Combination therapy withnaproxen and aspirin in rheumatoid arthritis. Arthritis andRheumatism 1976;19(4):677–82.
Additional references
Atkins 2004
Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y,Flottorp S, et al.GRADE Working Group. Grading qualityof evidence and strength of recommendations. BMJ 2004;328(7454):1490–4.
Bergman 2006
Bergman MJ. Social and economic impact of inflammatoryarthritis. Postgraduate Medicine 2006;Spec No:5–11.
Braun 1998
Braun J, Bollow M, Remlinger G, Eggens U, RudwaleitM, Distler A, et al.Prevalence of spondylarthropathies inHLA-B27 positive and negative blood donors. Arthritis andRheumatism 1998;41(1):58–67.
Calin 1994
Calin A, Garrett S, Whitelock H, Kennedy LG, O’HeaJ, Mallorie P, et al.A new approach to defining functional
ability in ankylosing spondylitis: the development of theBath Ankylosing Spondylitis Functional Index. Journal ofRheumatology 1994;21(12):2281–5.
Cohen 1988
Cohen J. Statistical Power Analysis for the Behavioral Sciences.Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.
Da Silva 2010
Da Silva JA, Ramiro S, Pedro S, Rodrigues A, VasconcelosJC, Benito-Garcia E. Patients and physicians prioritiesfor improvement. The case of rheumatic diseases. Acta
Reumatologica Portuguesa 2010;35(2):192–9.
De Angelis 2004
De Angelis CD, Drazen JM, Frizelle FA, Haug C, Hoey J,Horton R, et al.Clinical trial registration: a statement fromthe International Committee of Medical Journal Editors.JAMA 2004;292(11):1363–4.
De Jong 1997
De Jong Z, Van Der Heijde D, McKenna SP, WhalleyD. The reliability and construct validity of the RAQoL:a rheumatoid arthritis-specific quality of life instrument.British Journal of Rheumatology 1997;36(8):878–83.
Deeks 2011
Deeks JJ, Higgins JPT, Altman DG (editors). Chapter 9:Analysing data and undertaking meta-analyses. In: HigginsJPT, Green S (editors). Cochrane Handbook for SystematicReviews of Interventions. Version 5.1 [updated March2011]. The Cochrane Collaboration, 2011. Available fromwww.cochrane-handbook.org.
Dougados 2011
Dougados M, Simon P, Braun J, Burgos-Vargas R,Maksymowych WP, Sieper J, et al.ASAS recommendationsfor collecting, analysing and reporting NSAID intakein clinical trials/epidemiological studies in axialspondyloarthritis. Annals of Rheumatic Diseases 2011;70(2):249–51.
Doward 2003
Doward LC, Spoorenberg A, Cook SA, Whalley D,Helliwell PS, Kay LJ, et al.Development of the ASQoL: aquality of life instrument specific to ankylosing spondylitis.Annals of Rheumatic Diseases 2003;62:20–6.
Finnerup 2010
Finnerup NB, Sindrup SH, Jensen TS. The evidence forpharmacological treatment of neuropathic pain. Pain 2010;150(3):573–81.
Fries 1980
Fries JF, Spitz P, Kraines RG, Holman HR. Measurementof patient outcome in arthritis. Arthritis and Rheumatism1980;23(2):137–45.
Gabriel 2001
Gabriel SE. The epidemiology of rheumatoid arthritis.Rheumatic Disease Clinics 2001;27(2):269–81.
Heiberg 2002
Heiberg T, Kvien TK. Preferences for improved healthexamined in 1,024 patients with rheumatoid arthritis: pain
18Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
has highest priority. Arthritis and Rheumatism 2002;47(4):391–7.
Higgins 2011a
Higgins JPT, Altman DG, Sterne JAC (editors). Chapter8: Assessing risk of bias in included studies. In: HigginsJPT, Green S (editors). Cochrane Handbook for SystematicReviews of Interventions Version 5.1 [updated March2011]. The Cochrane Collaboration, 2011. Available fromwww.cochrane-handbook.org.
Higgins 2011b
Higgins JPT, Deeks JJ, Altman DG (editors). Chapter16: Special topics in statistics. In: Higgins JPT, Green S(editors). Cochrane Handbook for Systematic Reviewsof Interventions Version 5.1 [updated March 2011].The Cochrane Collaboration, 2011. Available fromwww.cochrane-handbook.org.
Higgins 2011c
Higgins JPT, Deeks JJ (editors). Chapter 7: Selectingstudies and collecting data. In: Higgins JPT, Green S(editors). Cochrane Handbook for Systematic Reviewsof Interventions Version 5.1 [updated March 2011].The Cochrane Collaboration, 2011. Available fromwww.cochrane-handbook.org.
Kasis 1983
Kazis LE, Meenan RF, Anderson JJ. Pain in the rheumaticdiseases. Investigation of a key health status component.Arthritis and Rheumatism 1983;26(8):1017–22.
Keystone 2004
Keystone EC, Kavanaugh AF, Sharp JT, Tannenbaum H,Hua Y, Teoh LS, et al.Radiographic, clinical, and functionaloutcomes of treatment with adalimumab (a human anti-tumor necrosis factor monoclonal antibody) in patientswith active rheumatoid arthritis receiving concomitantmethotrexate therapy: a randomized, placebo-controlled,52-week trial. Arthritis and Rheumatism 2004;50(5):1400–11.
Kvien 2004
Kvien TK. Epidemiology and burden of illness ofrheumatoid arthritis. Pharmacoeconomics 2004;22(2 Suppl1):1–12.
Lanser 2001
Lanser P, Gesell S. Pain management: the fifth vital sign.Health Benchmarks 2001;8(6):68–70.
Lefebvre 2011
Lefebvre C, Manheimer E, Glanville J. Chapter 6: Searchingfor studies. In: Higgins JPT, Green S (editors). CochraneHandbook for Systematic Reviews of Interventions Version5.1 [updated March 2011]. The Cochrane Collaboration,2011. Available from www.cochrane-handbook.org.
Mease 2000
Mease PJ, Goffe BS, Metz J, VanderStoep A, Finck B,Burge DJ. Etanercept in the treatment of psoriatic arthritisand psoriasis: a randomised trial. Lancet 2000;356(9227):385–90.
Minnock 2003
Minnock P, Fitzgerald O, Bresnihan B. Quality of life,social support, and knowledge of disease in women withrheumatoid arthritis. Arthritis and Rheumatism 2003;49(2):221–7.
Moore 2010
Moore AR, Eccleston C, Derry S, Wiffen P, Bell RF, StraubeS, et al.“Evidence” in chronic pain- establishing best practicein the reporting of systematic reviews. Pain 2010;150(3):386–9.
Peto 1976
Peto R, Pike MC, Armitage P, Breslow NE, Cox DR,Howard SV, et al.Design and analysis of randomized clinicaltrials requiring prolonged observation of each patient. I.Introduction and design. British Journal of Cancer 1976;34
(6):585–612.
Philips 2000
Phillips DM. JCAHO pain management standards areunveiled. Joint Commission on Accreditation of HealthcareOrganizations. JAMA 2000;284(4):428–9.
Pincus 1983
Pincus T, Summey JA, Soraci SA Jr, Wallston KA, HummonNP. Assessment of patient satisfaction in activities ofdaily living using a modified Stanford Health AssessmentQuestionnaire. Arthritis and Rheumatism 1983;26(11):1346–53.
Ramiro 2010
Ramiro S, Radner H, Van Der Heijde D, Van TubergenA, Buchbinder R, Aletaha D, et al.Combination therapyfor pain management in inflammatory arthritis. Cochrane
Database of Systematic Reviews 2010, Issue 12. [DOI:10.1002/14651858.CD008886]
Saarto 2007
Saarto T, Wiffen PJ. Antidepressants for neuropathic pain.Cochrane Database of Systematic Reviews 2007, Issue 4.[DOI: 10.1002/14651858.CD005454.pub2]
Schünemann 2011
Schünemann HJ, Oxman AD, Higgins JPT, Vist GE,Glasziou P, Guyatt GH. Chapter 11: Presenting resultsand ’Summary of findings’ tables. In: Higgins JPT,Green S (editors). Cochrane Handbook for SystematicReviews of Interventions. Version 5.1 [updated March2011]. The Cochrane Collaboration, 2011. Available fromwww.cochrane-handbook.org.
Skevington 1998
Skevington SM. Investigating the relationship between painand discomfort and quality of life, using the WHOQOL.Pain 1998;76(3):395–406.
Smolen 2010a
Smolen JS, Landewé R, Breedveld FC, Dougados M, EmeryP, Gaujoux-Viala C, et al.EULAR recommendations for themanagement of rheumatoid arthritis with synthetic andbiological disease-modifying antirheumatic drugs. Annals ofRheumatic Diseases 2010;69(6):964–75.
19Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Smolen JS, Aletaha D, Bijlsma JW, Breedveld FC, BoumpasD, Burmester G, et al.Treating rheumatoid arthritis totarget: recommendations of an international task force.Annals of Rheumatic Disease 2010;69(4):631–7.
Ten Klooster 2007
Ten Klooster PM, Veehof MM, Taal E, Van Riel PL, VanDe Laar MA. Changes in priorities for improvement inpatients with rheumatoid arthritis during 1 year of anti-tumour necrosis factor treatment. Annals of Rheumatic
Diseases 2007;66(11):1485–90.
Van der Heijde 2005
Van Der Heijde D, Dijkmans B, Geusens P, Sieper J,DeWoody K, Williamson P, et al.Ankylosing SpondylitisStudy for the Evaluation of Recombinant InfliximabTherapy Study Group. Efficacy and safety of infliximabin patients with ankylosing spondylitis: results of arandomized, placebo-controlled trial (ASSERT). Arthritis
and Rheumatism 2005;52(2):582–91.
Ware 2001
Ware JE, Kosinski M. SF-36 Physical and Mental Health
Summary Scales: manual for users version 1. 2nd Edition.Lincoln, Rhode Island, USA: Quality Metric, 2001.
∗ Indicates the major publication for the study
20Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Characteristics of included studies [ordered by study ID]
Beales 1972
Methods Controlled clinical trialStudy duration: 2 weeksRandomisation method: not specified, probably not randomisedBlinding: single blinded study (only physicians blinded)Sample size calculation: not described
Participants 72 patientsInclusion criteria: definite or classical RA (as defined by the ARA criteria)Exclusion criteria: gold therapy in the year preceding the study or current therapy withcorticotrophin or antimalarial drugs; history of renal, hepatic, or cardiac dysfunction orpregnancy; history of dyspepsia or intolerance to aspirinMean age (years): combination therapy arm 51 ± 1; monotherapy arm 54 ± 2Gender (% female): combination therapy arm 66.7; monotherapy arm 44.4Disease duration (years): combination therapy arm 7.5 ± 1; monotherapy arm 9 ± 2Comedication not specifiedWithdrawals: 6 patients
Interventions Group 1 (combination therapy): benorylate (aspirin + paracetamol) 8 g (n = 36)Group 2 (monotherapy): aspirin 4.8 g (n = 36)
Outcomes 1) Diurnal pain score at 2 weeks (0 = nil; 1 = mild; 2 = moderate; 3 = severe);2) Severity of disease;3) Functional grading;4) Grip strength;5) Finger stiffness;6) Total ring size;7) Articular index
Notes Pain at 2 weeks - percentage improvement from baseline (negative scores mean the painimproved):Group 1 (combination therapy): baseline pain level 0: 0%; 1: 39% (n = 14), 2: 44% (n= 16), 3: 14% (n = 5); change from baseline: -41%Group 2 (monotherapy): baseline pain level 0: 3% (n =1); 1: 33% (n = 12), 2: 39% (n= 14), 3: 19% (n = 7); change from baseline: -48%No significant difference between the groups, as reported by the authors, P value > 0.05Withdrawals due to adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 3% (n = 1)Group 2 (monotherapy): 3% (n = 1)Withdrawals due to inadequate analgesia: (interpretation of the authors of the reviewfrom what is reported)Group 1 (combination therapy): 0% (n = 0)Group 2 (monotherapy): 5.5% (n = 2)No deaths were reported (interpretation of the authors of the review from what is re-
21Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
ported)No information on the number of patients with adverse events in the whole trial isreported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Not specified. The word randomised doesnot appear in the text
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
High risk Single-blinded study: “They were asked notto discuss their treatment among them-selves nor to let the examining physi-cian know whether they were receiving thetablets or the suspension”. This could biaspatient-reported outcome assessment
Incomplete outcome data (attrition bias)All outcomes
High risk Withdrawals are excluded from the analysis
Selective reporting (reporting bias) High risk Pain measurement is only presented in %of change and not in absolute value or notindividualised per category of the categor-ical variable, as it has been presented forbaseline
Other bias Unclear risk No specification whether attempts weremade to limit cointerventions
Bedi 1969
Methods Cross-over trialStudy duration: 2 weeks total (1 week in each arm, no washout)Randomisation method: not specifiedBlinding: participants and outcome assessor blindedSample size calculation: not described
Participants 51 patientsInclusion criteria: patients with definite or classical RA with moderate to severe pain(not further defined)Exclusion criteria: not specifiedMean age, gender, disease duration and comedication not specifiedWithdrawals: 5 patients
22Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Interventions Group 1 (combination therapy): aspirin 500 mg + dextropropoxyhene napsilate 50 mg(completing the treatment: n = 23)Group 2 (monotherapy): aspirin 500 mg (completing the treatment: n = 22)Dosage of both arms: 1 tablet 3xday for the first 2 days and 6 tablets in divided dosagedaily thereafterNumber of patients randomised to each group not presented, only the number of patientscompleting the treatment
Outcomes 1) Improvement in pain (better, no change, worse) at the end of each treatment2) Number and type of adverse events (although not specified in the methods)
Notes For the purpose of this review results were extracted from the first period for efficacy andfrom both periods for safetyImprovement in pain (end of treatment, 1 week):Group 1 (combination therapy): worse 2 (9%); no change 7 (32%); better 13 (59%)Group 2 (monotherapy): worse 4 (17%); no change 9 (39%); better 10 (44%)No significant difference between the groups, all P values > 0.05 (calculated by authorsof the review)Withdrawals due to adverse events: (interpretation and calculation by the authors of thereview from what is reported)Group 1 (combination therapy): 3 (6%)Group 2 (monotherapy): 0Number of patients with adverse events:Group 1 (combination therapy): 17 (33%) - lightheadedness (n = 7); constipation (n =6); tinnitus (n = 4); dizziness (n = 4); heartburn (n = 3); nausea (n = 3); vomiting (n =2); allergic rash (n = 1)Group 2 (monotherapy): 19 (37%) - lightheadedness (n = 4); constipation (n = 3);tinnitus (n = 4); dizziness (n = 2); heartburn (n = 4); nausea (n = 6); vomiting (n = 2)No withdrawals due to inadequate analgesia and no deaths were reported (interpretationof the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Not specified. The word randomised doesnot appear in the text
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “Double-blind study. The hospital pharmacist aloneknew what drug each patient had received”.No information on how blinding was as-sured or on an analysis of the patients per-ception of which arm they were in
23Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Incomplete outcome data (attrition bias)All outcomes
High risk 6 patients were excluded from the analysesof the trial, although they were reported tohave been allocated to one of the treatmentgroups. Number of patients randomisedinto each group not clear (only number ofpatients finishing each arm is presented)
Selective reporting (reporting bias) High risk Outcome is not clearly defined. Emphasisis put on the significant difference foundbetween the 2 drugs in the second periodof the cross-over trial
Other bias High risk Outcome not clearly defined, outcomescale not reported. Baseline characteristicsnot presented, so similarity cannot be as-sured. Crossover design with possible carry-over effect. Statistical analysis not detailed.No specification whether attempts weremade to limit cointerventions
Brooks 1975
Methods Parallel RCTStudy duration: 2 weeksRandomisation method: not specifiedBlinding: not specifiedSample size calculation: not described
Participants 134 patientsInclusion criteria: seropositive classical or definite RA (according to the ARA criteria)Exclusion criteria: hypersensitivity to aspirin or phenylbutazone; renal or hepatic diseases;symptoms of gastric or duodenal ulcer; changes in medication in the preceding months;current therapy with gold, corticosteroids, immunosuppressing agents, D-penicillamineor anticoagulantsMean age (years): combination therapy arm: 53.7 ± 1.35; monotherapy arm 1: 54.2 ±1.48; monotherapy arm 2: 50.8 ± 1.57Gender (% female): not specifiedMean disease duration (years): 10.4Comedication: No antirheumatic drugs allowed36 withdrawals
Notes Pain (throughout 2 weeks):Group 1 combination therapy: baseline 3.2 ± 0.1; throughout 2 weeks 3.0 ± 0.1 (adjustedfor baseline pain score 3.1 ± 0.2)Group 2 monotherapy 1: baseline 3.0 ± 0.1; throughout 2 weeks 3.3 ± 0.1 (adjusted forbaseline pain score 3.3 ± 0.2)Group 3 monotherapy 2: baseline 2.8 ± 0.1; throughout 2 weeks 2.7 ± 0.1 (adjusted forbaseline pain score 2.8 ± 0.2)Significant difference between the groups, as reported by the authors, but no PvalueshownWithdrawals due to adverse events:Group 1 (combination therapy): 14% (n = 6)Group 2 (monotherapy 1): 0% (n = 0)Group 3 (monotherapy 2): 2% (n = 1)Withdrawals due to inadequate analgesia:Group 1 (combination therapy): 2% (n = 1)Group 2 (monotherapy 1): 33% (n = 15)Group 3 (monotherapy 2): 4% (n = 2)No deaths were reported (interpretation of the authors of the review from what is re-ported). No information on the number of patients with adverse events
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: allocation byrandomisation
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Not mentioned
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Not clear how the withdrawals were ad-dressed and it appears they were excludedfrom the analyses
Selective reporting (reporting bias) Unclear risk Number of patients with adverse events andtype of adverse events not clear. Statisticalcomparison between the groups not veryclear, not clear which groups are signifi-cantly different in terms of pain control
Other bias High risk Very high drop-out rate (less than 50% outof each group finished the trial)
25Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Parallel RCTStudy duration: 2 weeksRandomisation method: not specifiedBlinding: only outcome assessors blindedSample size calculation: not described
Participants 95 patients, 41 of which with RAInclusion criteria: RA or osteoarthrosis (not further defined)Exclusion criteria: known pregnancy; age < 20 years; presence of renal or hepatic disease,peptic ulceration, or painful neurological conditions; those who needed to continuetaking analgesic/anti-inflammatory agentsMean age (years): combination therapy arm: 62.3; monotherapy arm: 60.6Gender (% female): combination therapy arm: 78; monotherapy arm: 78Disease duration (years): combination therapy arm: 7; monotherapy arm: 5Comedication: patients were permitted to continue maintenance doses of gold or corti-costeroids6 withdrawals
Interventions Group 1 (combination therapy): benorylate (paracetamol + aspirin) 8 g (n = 46; patientswith RA: n = 18)Group 2 (monotherapy): indomethacin 75 mg (n = 49; patients with RA: n = 26)
Outcomes 1) Improvement in pain (clinical assessment of pain lower by one or more categories,within the following: mild, moderate, severe) at the end of treatment2) Morning stiffness
Notes For the purpose of this review results were extracted from the subgroup of patients withRA for efficacy and from the whole population for safety (available data)Improvement in pain (end of treatment, 2 weeks):Group 1 (combination therapy): 73%Group 2 (monotherapy): 32%Significant difference between the groups (P < 0.05)Number of patients with adverse events:Group 1 (combination therapy): 22% (n = 10):tinnitus and other salicylate-based symp-toms, indigestion, diarrhoea (n not specified)Group 2 (monotherapy): 25% (n = 12): headache, diplopia and other CNS effects,indigestion, intermittent infection (n not specified)Withdrawals due to adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 9% (n = 4)Group 2 (monotherapy): 8% (n = 4)Withdrawals due to inadequate analgesia: (interpretation of the authors of the reviewfrom what is reported)Group 1 (combination therapy): 2% (n = 1)Group 2 (monotherapy): 8% (n = 4)No deaths were reported (interpretation of the authors of the review from what is re-ported)
Risk of bias
26Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Unclear risk Only information in the text: “They wereassigned on a random basis”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
High risk Single-blinded study and “The observerdid not know which preparation was be-ing administered”. Besides observer blind-ing, nothing is mentioned regarding partic-ipant blinding, which could bias patient-reported outcome assessment
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Not clear how the withdrawals were ad-dressed and whether or not they were in-cluded in the analyses although it appearsunlikely that they were included as a totalN per outcome is not provided
Selective reporting (reporting bias) Unclear risk 2 outcomes are reported, but it is not clearif these are all the measured ones; the out-come assessment method is not clear either
Other bias High risk Not all the results are separately reportedfor patients with RA (e.g. Baseline value ofpain), so the data need to be interpretedwith caution
Ekstrand 1981
Methods Cross-over trialStudy duration: 2 weeks wash-out period + 6 weeks (3 weeks on each dose of acetylsalicylicacid, no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 12 patientsInclusion criteria: classical or definite RA according to the criteria of the ARA, diseaseactivity stable during the previous 6 monthsExclusion criteria: treatment with corticosteroids, antimalarials or gold during the pre-vious 6 monthsMean age, gender, disease duration and comedication not specifiedNo withdrawals
27Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Interventions Group 1 (combination therapy 1): acetylsalicylic acid 2 g + indomethacin 50 mgGroup 2 (monotherapy 1) acetylsalicylic acid 2 g (+ placebo)Group 3 (combination therapy 2): acetylsalicylic acid 4.5 g + indomethacin 50 mgGroup 4 (monotherapy 2): acetylsalicylic acid 4.5 g (+ placebo)Indomethacin 50 mg was just given on the 13th and 20th day of each 3-week period
Outcomes 1) Overall global assessment2) Preference for one of the suppositories given3) Duration of morning stiffness4) Pain throughout the period and on the day after indomethacin/placebo (0-100mmVAS)5) Grip strength6) Articular index7) Size of PIP joints
Notes Results reflect the average of all cross-over periods (only available ones)Pain (throughout the 3-week treatment): (median values presented, available ones)Group 1 (combination therapy 1): 43Group 2 (monotherapy 1): 53Group 3 (combination therapy 2): 40Group 4 (monotherapy 2): 42Comparison of pain level on the day after indomethacin/placebo between monotherapyand combination therapy only significantly different in the low-dose group (P < 0.05) -data not shown in the paperNo withdrawals due to inadequate analgesia, no withdrawals due to adverse events and nodeaths were reported (interpretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “On the 13thand 20th day of each 3-week period, the pa-tients were instructed to take a suppositorycontaining either 50 mg of indomethacinor placebo in randomised order”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “Each pa-tient was given two dosages of acetylsali-cylic acid double-blind.” No informationon who exactly was blinded, how blindingwas assured or on an analysis of the patients’perception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Incomplete data not clear, N per outcomenot specified
28Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Selective reporting (reporting bias) High risk No measures of spread (SD) are presentedwith the means. Only the outcomes withsignificant differences are reported. Severaloutcomes reported in the methods are thennot presented in the results. Outcomes arenot clearly defined, regarding scales, unitsof measurement, etc. Separate results ofeach cross-over period are not presented
Other bias High risk Baseline characteristics of the groups notpresented. Time points of the outcomes notclear (pain reported to be assessed through-out the period and on the day after in-domethacin/placebo, and then not clear inthe results) Safety data not presented indi-vidually for each treatment arm
Furst 1987
Methods Cross-over trialStudy duration: 4 weeks run-in period (therapeutic salicylate concentrations were ob-tained) + washout until flare + 4 months (1 month in each arm, no washout) + washoutuntil flareRandomisation method: no randomisationBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 84 patientsInclusion criteria: active classic or definite RA according to the ARA criteria. Patientswere considered to have active disease if at least 3 of the following were present: 1) >6 painful or tender joints on motion; 2) >3 swollen joints; 3) > 45 min of morningstiffness; 4) ESR ≥ 28 mm/hExclusion criteria: joint injection with corticosteroids within the last 4 weeks; onset orstop of antimalarial drugs, immunosuppressive drugs, or penicillamine in the past 3months; active peptic ulcer disease; significant renal or cardiovascular disease; functionalclass IV by Steinbrocker’s criteria; pregnancy; other NSAIDs use during the trial; knownhypersensitivity to salicylates or naproxen; inability or unwillingness to comply withrequirements of the study; discontinuation of gold therapy, defined as a stable dose ofgold therapy within the past 3 monthsMean age (years): 56 ± 13Gender (% female): 74Disease duration: not specifiedComedication: stable gold therapy (≥ 20 weeks) was allowed37 withdrawals (19 during the run-in period and 18 during the cross-over periods)
Outcomes 1) Joint tenderness2) Joint swelling3) Duration of time to beginning amelioration of morning stiffness4) Grip strength5) Physician’s global assessment6) Patient’s global assessment7) Patient’s pain evaluation (0 to100 mm VAS) - absolute (percentage) change from flare8) Patient’s subjective evaluation of ability to complete activities of daily living (0 to 180,maximum score 180):absolute (percentage) change from flare
Notes Results reflect the average of all cross-over periods (all the 4 treatment periods on eachdrug) (only available ones)Pain - absolute (percentage) change from flare (negative scores mean the pain improved):Group 1 (combination therapy 1): -26 (-37%)Group 2 (monotherapy 1): -24 (-35%)Group 3 (combination therapy 2): -15 (-22%)Group 4 (monotherapy 2): -15 (-22%)Mean at first flare only reported for the whole study population: 66No significant difference between the groups (P = 0.94)Activities of daily living - absolute (percentage) change from flare (negative scores meanthe pain improved):Group 1 (combination therapy 1): -27 (237%)Group 2 (monotherapy 1): -23 (-21%)Group 3 (combination therapy 2): -18 (-17%)Group 4 (monotherapy 2): -16 (-15%)No significant difference between the groups (P = 0.80)Withdrawals due to adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 11.2% (both groups of combination therapy)Group 2 (monotherapy): 5.2% (both groups of monotherapy)Withdrawals due to inadequate analgesia: (interpretation of the authors of the reviewfrom what is reported)Group 1 (combination therapy): 5% (n = 3)Group 2 (monotherapy): 6% (n = 4)Number of patients with adverse events not specified. No deaths were reported (inter-pretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
30Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
High risk No randomisation. “In this study, a cyclicpermutation Latin-square design was used,and each treatment was always followed bythe same treatment, a cyclic permutationof the numbers 4, 2, 1, 3), except when itfollowed the flare”
Allocation concealment (selection bias) High risk No randomisation
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “Double-blind..”, “CMT, N and matching place-bos were supplied and packaged in appro-priately blinded individual dosage pack-ets”. No information on who exactly wasblinded, how blinding was assured or on ananalysis of the patients’ perception of whicharm they were in
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Description in the methods:“Because somepatients did not complete all phases of thestudy, a general linear model was used forefficacy analysis. It allowed all patients tobe included in the analysis, even those whodid not complete all phases”. But then thenumbers in terms of withdrawals and pa-tients in each treatment arm do not match,not understandable what happened to allthe patients and how they were (or not) in-cluded in the analyses
Selective reporting (reporting bias) High risk Outcomes are presented as a difference be-tween final value and mean at first flare,which emphasises the improvement ob-tained with the drug. Absolute end valuesshould be presented in order to enable thecalculation of the difference to the meanbaseline score. No measures of spread arepresented with the means (SD). Number ofadverse events not clear; not clear to whichpatients they refer to. Separate results ofeach cross-over period are not presented
Other bias High risk Cross-over design with possible carry-over effect, although the authors advocate:“When testing for carry-over effects, 5 of42 instances were significant, but given thenumber of comparisons, this was not sur-prising. We interpreted the results as show-ing that if any carry-over effect occurred, it
31Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
was of limited significance” Comparison ofbaseline characteristics very limited
Grennan 1979
Methods Cross-over trialStudy duration: 16 weeks (8 weeks for each trial part, and each of them with 2 weeks ineach of the 4 arms, being placebo one of them, no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 16 patients (Part I); 14 patients (Part II)Inclusion criteria: definite or classical RA (defined by the ARA criteria)Exclusion criteria: previous history of peptic ulceration, renal or hepatic disease; currenttherapy with gold, penicillamine or corticosteroidsMean age, gender, disease duration not specifiedComedication: not specifiedWithdrawals not specified
Interventions Group 1 (combination therapy): aspirin + ibuprofenGroup 2 (monotherapy 1): aspirinGroup 3 (monotherapy 2): ibuprofenPart I of the trial: 2.4 g aspirin + 800 mg ibuprofenPart II of the trial: 3.6 g aspirin + 1600 mg ibuprofen
Outcomes 1) Articular index2) Duration of morning stiffness3) Pain score at 2 weeks (0 to 10 cm VAS)4) Patient global subjective score5) Global observer score6) Time to walk 50 feet7) Grip strength
Notes Results reflect the average of all cross-over periods (all the 4 treatment periods on eachdrug) (only available ones)Pain (at 2 weeks of treatment):Part I of the trialGroup 1 (combination therapy): 5.6 ± 2.3Group 2 (monotherapy 1): 6.3 ± 2.7Group 3 (monotherapy 2): 6.6 ± 2.8Part II of the trialGroup 1 (combination therapy): 4.7 ± 2.1Group 2 (monotherapy 1): 5.4 ± 2.6.Group 3 (monotherapy 2): 5.3 ± 2.3No significant difference between the groups, as reported by the authors, but no P valueshownWithdrawals due to adverse events and due to inadequate analgesia not clearly reported.
32Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Numbers presented for patients not finishing one study arm and continuing to the next,but not exactly withdrawing from the studyNumber of patients with adverse events not specified. No deaths were reported (inter-pretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “Pre-packed,randomised supplies of 300 mg tabletsof soluble aspirin and dummy soluble as-pirin tablets, 200 mg ibuprofen tablets anddummy ibuprofen tablets were supplied byBoots of Australia”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “…double-blind”, “Pre-packed, randomised suppliesof 300 mg tablets of soluble aspirin anddummy soluble aspirin tablets, 200 mgibuprofen tablets and dummy ibuprofentablets were supplied by Boots of Aus-tralia”. No information on who exactly wasblinded, how blinding was assured or on ananalysis of the patients’ perception of whicharm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk Not clear how the withdrawals were ad-dressed and whether or not they were in-cluded in the analyses although it appearsunlikely that they were included, as a totalN per outcome is not provided
Selective reporting (reporting bias) High risk No baseline values are presented. Separateresults of each cross-over period are not pre-sented
Other bias High risk Baseline characteristics not presented.Number of randomised patients and pa-tients finishing each trial arm not clear.Cross-over design with possible carry-overeffect
33Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Cross-over trialStudy duration: 8 weeks (4 weeks in each arm, no washout)Randomisation method: predetermined random codeBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 33 patientsInclusion criteria: definite or classical RA of at least one year’s duration and an ESR ≥
20 mm/hExclusion criteria: corticosteroids, gold or antimalarial therapy in the preceding 6 monthsMean age and disease duration not specifiedGender (% female): combination therapy arm: 76; monotherapy arm: 73Comedication: “No physiotherapy was undertaken during the trial, and all other anal-gesic and anti-inflammatory drugs were withdrawn three weeks before the trial period,paracetamol being substituted until the therapy commenced”1 withdrawal (severe dizziness not related to the trial drug)
Outcomes 1) Grip strength2) Patient own assessment of overall pain at 4 weeks (0 = no pain, 1 = mild pain, 2 =moderate pain, 3 = severe pain, 4 = unbearably severe pain)3) Duration of morning stiffness4) Adverse events
Notes Results reflect the average of both cross-over periods (only available ones)Pain:Group 1 (combination therapy): baseline 2.6 ± 0.3; 4 weeks 2.4 ± 0.4Group 2 (monotherapy): baseline 2.9 ± 0.4; 4 weeks 2.0 ± 1.0Values read by the review authors from a graphWithdrawals due to inadequate analgesia: (interpretation of the authors of the reviewfrom what is reported)Group 1 (combination therapy): 9% (n = 3)Group 2 (monotherapy): 0% (n = 0)Number of patients with adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 18% (n = 6): brown-coated tongue (n = 1); hungerand losing weight (n = 1); irritation of the skin (n = 1); vomiting and diarrhoea (n = 1);diarrhoea and loss of energy (n = 1); nausea, anorexia and depression (n = 1)Group 2 (monotherapy): 21% (n = 7): abdominal discomfort (n = 1); polyuria (n = 2);dizziness and faintness (n = 1); headache, lack of energy, itchy eyes (n = 1); indigestion(n = 1); rash and swollen neck glands (n = 1)No withdrawals due to adverse events or deaths were reported (interpretation of theauthors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
34Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Unclear risk Only information in the text: “The treat-ment order for each patient was decided bya predetermined random code”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The trial wasa double-blind…” and “All tablets were ofidentical shape, size and colour”. No infor-mation on who exactly was blinded, howblinding was assured or on an analysis ofthe patients’ perception of which arm theywere in
Incomplete outcome data (attrition bias)All outcomes
High risk One patient has withdrawn and was ex-cluded from the analysis, after being in-cluded in the study
Selective reporting (reporting bias) High risk Separate results of each cross-over periodare not presented
Other bias High risk Cross-over design with possible carry-overeffect
Hingorani 1973
Methods Cross-over trialStudy duration: 8 weeks (1 week paracetamol + 3 weeks in one arm + 1 week paracetamol+ 3 weeks in another arm)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 27 patientsInclusion criteria: classical RA with hand involvementExclusion criteria: known hepatic or renal disease; existing pregnancy; current therapywith gold, steroids (> 10 mg/day) or antimalarialsMean age not specifiedGender (% female): combination therapy arm: 100; monotherapy arm: 82Median disease duration (years): combination therapy arm: 6; monotherapy arm: 7Comedication: not specified3 withdrawals
Notes For the purpose of this review results were extracted from the first period for efficacy andfrom both periods for safetyPain:Group 1 (combination therapy): baseline 2.23 ± 0.23; improvement at 3 weeks of treat-ment -0.5 ± 0.2Group 2 (monotherapy): baseline 2.00 ± 0.33; improvement at 3 weeks of treatment -0.8 ± 0.3No significant difference between the groups, as reported by the authors, but no P valueshownNumber of patients with adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 54% (n = 13): vomiting (n = 3); headache (n = 3);constipation (n = 2); nausea (n = 1); tinnitus (n = 4); other (n = 8)Group 2 (monotherapy): 29% (n = 7): headache (n = 1); constipation (n = 2); nausea (n= 3); tinnitus (n = 1); other (n = 4)Functional capacity (presented, but not clearly defined):Group 1 (combination therapy): baseline mild 77% (n = 10); moderate 23% (n = 3);severe 0% (n = 0); improvement at 3 weeks of treatment +0.2 ± 0.1Group 2 (monotherapy): baseline mild 82% (n = 9); moderate 18% (n = 2); severe 0%(n = 0); improvement at 3 weeks of treatment +0.2 ± 0.18No significant difference between the groups, as reported by the authors, but no P valueshownNo withdrawals due to adverse events, withdrawals due to inadequate analgesia or deathswere reported (interpretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “…in ran-dom allocation”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “Double-blind study” and “Each treatment appearedidentical in that while on benorylate theyalso received 6 placebo tablets and whileon ibuprofen, placebo emulsion.” No in-formation on who exactly was blinded, howblinding was assured or on an analysis ofthe patients’ perception of which arm they
36Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Incomplete outcome data (attrition bias)All outcomes
High risk Withdrawals were excluded from the anal-yses: “Three patients failed to complete thestudy for various personal and domesticreasons. These patients were excluded fromthe study”
Selective reporting (reporting bias) High risk Although the study reported baseline andend values for each group, no statisticalcomparison between groups was reported
Other bias High risk Outcomes are not clearly defined, regard-ing scales, units of measurement (e.g. func-tional capacity)
Hobkirk 1977
Methods Cross-over trialStudy duration: 2 days (1 night in each arm, no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 18 patientsInclusion criteria: hospitalisation for treatment of active RA (not further specified)Exclusion criteria: current therapy with corticosteroids or >75 mg indomethacin daily,known hypersensitivity to either trial medicationMean age (years): 54.5 ± 17.4Gender (% female): 76Mean disease duration (years): 6.8 ± 8.9Comedication: not specified1 withdrawal
Outcomes 1) Morning stiffness2) Pain (0 to 10 cm VAS)3) Sleep4) Adverse events
Notes Results reflect the average of both cross-over periods (only available ones)Pain (at 1 day of treatment):Group 1 (combination therapy): 0.83Group 2 (monotherapy): 1.28No significant difference between the groups, as reported by the authors, but no P valueshownWithdrawals due to adverse events:
37Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Group 1 (combination therapy): 6% (n = 1)Group 2 (monotherapy): 0% (n = 0)Number of patients with adverse events:Group 1 (combination therapy): 17% (n = 3): headache (n = 2); dizziness (n = 1)Group 2 (monotherapy): 17% (n = 3): headache (n = 2); nausea (n = 1)No withdrawals due to inadequate analgesia or deaths were reported (interpretation ofthe authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “…, the treat-ment order was randomised..”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “…the blindnature of the trial was ensured by usingappropriate dummies”. No information onwho exactly was blinded, how blinding wasassured or on an analysis of the patients’perception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk There was one withdrawal that was ex-cluded from the analysis
Selective reporting (reporting bias) High risk No measures of spread are presented withthe means (SD). Separate results of eachcross-over period are not reported. No base-line values or change values for the out-comes are presented, only the final values
Other bias High risk Cross-over design with possible carry-overeffect
Huskisson 1974
Methods Cross-over trialStudy duration: 18 days: 1 day on one drug, the 2nd day on another drug, 3rd daywashout; all 6 possible treatment pairs (see below in interventions + placebo) comparedRandomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants Trial 1: 30 patientsTrial 2: 24 patientsInclusion criteria: definite or classical RA (ARA criteria) with pain of sufficient severity
38Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
to require “on demand” analgesics at least once a dayExclusion criteria: not specifiedMean age, gender, disease duration and comedication not specifiedTrial 1: 7 withdrawals; Trial 2: 2 withdrawals
Interventions Trial 1:Group 1 (combination therapy): paracetamol 650 mg + dextropropoxyphene 65 mgGroup 2 (monotherapy): paracetamol 1000 mg OR Pentazocine 50 mgNumber of patients randomised to each group not presentedTrial 2:Group 1 (combination therapy): aspirin 1000 mg + codeine 16 mg OR paracetamol 650mg + dextropropoxyphene 65 mgGroup 2 (monotherapy): aspirin 600 mg
Notes Results reflect the average of all cross-over periods (all the 3 treatment periods on eachdrug) (only available ones)Pain relief score (at 6 hours):Trial 1:Group 1 (combination therapy): 1.18Group 2 (monotherapy): 1.09 (paracetamol); 1.25 (pentazocine)No significant difference between the groups (P > 0.05)Trial 2:Group 1 (combination therapy): 0.95 (paracetamol + codeine); 0.91 (paracetamol +dextropropoxyphene)Group 2 (monotherapy): 0.72No significant difference between the groups (P > 0.05)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: Trial 1 -“Treatment order was randomized and bal-anced so far as possible”; Trial 2 - “The or-der of treatment and the colours used wererandomized…”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: Trial 1 - “Thedouble-placebo method was used to ensurethat the trial was double blind and each pa-tient was therefore given 4 tablets in eachdose”; Trial 2 - “The tablets were identicalin shape and size but each was made in 4colours, red, blue, green and yellow. Four
39Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
colour codes were used to ensure that eachtreatment in a pair comparison appeareddifferent. For a particular patient the samecolour code was used throughout.” No in-formation on who exactly was blinded (es-pecially in Trial 2), how blinding was as-sured (especially in Trial 1) or on an analy-sis of the patients’ perception of which armthey were in
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Not clear how the withdrawals were ad-dressed and whether or not they were in-cluded in the analyses although it appearsunlikely that they were included, as a totalN per outcome is not provided
Selective reporting (reporting bias) High risk Pain outcome measurement is not clear, notexplained. No baseline values are presented.No measures of spread (SD) are presented.Separate results of each cross-over periodare not presented
Other bias High risk Cross-over design with possible carry-overeffects. Safety data not presented
Kean 1981
Methods Cross-over trialStudy duration: 2 weeks (1 week in each arm, no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 24 patientsInclusion criteria: classical or definite RA (not further specified)Exclusion criteria: not specifiedMean age (years): 44.6Gender (% female): 58Mean disease duration (years): 6.4Comedication: not specifiedWithdrawals not stated
3) Severity of morning stiffness4) Grip strength5) Joint circumference6) Articular index
Notes Results reflect the average of both cross-over periods (only available ones)Pain (at 1 week of treatment):Group 1 (combination therapy): 2.08 ± 0.18Group 2 (monotherapy 1): 2.08 ± 0.18Group 3 (monotherapy 2): 1.92 ± 0.16No significant difference between the groups, as reported by the authors, but no P valueshownNo information on withdrawals due to adverse events, withdrawals due to inadequateanalgesia or adverse events reported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “The studywas a double-blind crossover with ran-domisation of treatment sequence order bythe Latin squares method”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The studywas a double-blind crossover …” No infor-mation on who exactly was blinded, howblinding was assured or on an analysis ofthe patients’ perception of which arm theywere in
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Number of patients randomised into eachgroup and patients finishing the treatmentnot clear
Selective reporting (reporting bias) High risk Separate results of each cross-over periodare not reported. No baseline values are re-ported
Other bias High risk Cross-over design with possible carry-overeffect. No specification whether attemptswere made to limit cointerventions. No in-formation on safety
41Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Cross-over trialStudy duration: 8 weeks (1 week paracetamol + 3 weeks trial drug + 1 week intervalunder paracetamol + 3 weeks other trial drug)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 17 patientsInclusion criteria: patients between the ages of 18 and 65 years with active RA of thehands (not further specified)Exclusion criteria: peptic ulcer, impaired renal function, pregnancy, therapy with steroidsor anticoagulants during the previous 6 monthsMean age, gender, disease duration and comedication not specified5 withdrawals
Notes Results reflect the average of both cross-over periods (only available ones)Improvement in pain, at 3 weeks of treatment (negative scores mean the pain improved):Group 1 (combination therapy): -0.83Group 2 (monotherapy): -0.33No significant difference between the groups, as reported by the authors, but no P valueshownWithdrawals due to adverse events (interpretation of the authors of the review from whatis reported):Group 1 (combination therapy): 6% (n = 1)Group 2 (monotherapy): 12% (n = 2)Number of patients with adverse events: (interpretation of the authors of the review fromwhat is reported)Group 1 (combination therapy): 29% (n = 5) - not specifiedGroup 2 (monotherapy): 29% (n = 5) - not specifiedNo withdrawals due to inadequate analgesia or deaths were reported (interpretation ofthe authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “Patients en-tering the trial were consecutively num-bered and the order in which they received
42Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
the drugs was according to a pre-arrangedrandomized code”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “Double-blind…” and“ In order to preserve the dou-ble-blind nature of the trial double place-bos were used”. No information on whoexactly was blinded, how blinding was as-sured or on an analysis of the patients’ per-ception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk 5 withdrawals were excluded from the anal-yses
Selective reporting (reporting bias) High risk Pain measurements are not reported. Nobaseline values are presented. No measuresof spread (SD) are presented. Separate re-sults of each cross-over period are not pre-sented
Other bias High risk No specification whether attempts weremade to limit cointerventions
Mavrikakis 1977a
Methods Cross-over trialStudy duration: 3 weeks (1 week in each of the study arms, being placebo one of them,no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 18 patientsInclusion criteria: classical RA (not further specified), an articular index of joint ten-derness exceeding 15 score units and a degree of clinical reversibility, as judged by onerheumatologistExclusion criteria: current or previous therapy with corticosteroids, chrysotherapy orcytotoxic drugs; pregnancyMean age (years): 46.3 (female) and 47.2 (males)Gender (% female): 67Disease duration (years): 6.1Comedication: not statedWithdrawals not stated
Outcomes 1) Pain at 1 week (scale not described but values ranged from 0 to 10)2) Articular index3) Duration of morning stiffness4) Grip strength5) PIP joint circumference6) Patient’s assessment of change in disease activity7) Physician’s assessment of change in disease activity
Notes Results reflect the average of both cross-over periods - only available onesPain (at 1 week of treatment):Group 1 (combination therapy): 1.77 ± 0.22Group 2 (monotherapy): 1.77 ± 0.19No significant difference between the groups, p-value < 0.95No information on withdrawals due to adverse events, withdrawals due to inadequateanalgesia or adverse events reported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Not specified. The word randomised doesnot appear in the text
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The designof the study was that of a double-blindcrossover nature”. No information on whoexactly was blinded, how blinding was as-sured or on an analysis of the patients’ per-ception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk Number of patients randomised into eachgroup and patients finishing the treatmentnot clear
Selective reporting (reporting bias) High risk Separate results of each cross-over periodare not reported. No baseline values are re-ported. Scale measurement of the outcomesnot mentioned, e.g. pain scale
Other bias High risk Cross-over design with possible carry-overeffect. No information on safety
44Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Cross-over trialStudy duration: 3 weeks (1 week in each of the study arms, being placebo one of them,no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 18 patientsInclusion criteria: classical or definite RA (not further specified) and continued jointpain despite antirheumatic drug therapyExclusion criteria: current or previous therapy with corticosteroids, chrysotherapy orcytotoxic drugsMean age (years): 47.3 (female) and 51.5 (males)Gender (% female): 72Disease duration (years): 6.1Comedication: not statedWithdrawals not stated
Outcomes 1) Pain at 1 week (scale not described but values ranged from 0 to 10)2) Articular indices3) Duration and severity of morning stiffness4) Grip strength5) PIP joint circumference6) Patient’s assessment of change in disease activity
Notes Results reflect the average of both cross-over periods (only available ones)Pain (at 1 week of treatment):Group 1 (combination therapy): 2.05 ± 0.22Group 2 (monotherapy): 1.88 ± 0.29No significant difference between the groups, as reported by the authors, but no P valueshownNo information on withdrawals due to adverse events, withdrawals due to inadequateanalgesia or adverse events reported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Not specified. The word randomised doesnot appear in the text
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The designof the study was that of a double-blindcrossover type”. No information on who
45Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
exactly was blinded, how blinding was as-sured or on an analysis of the patients’ per-ception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk Number of patients randomised into eachgroup and patients finishing the treatmentnot clear
Selective reporting (reporting bias) High risk Separate results of each cross-over periodare not reported. No baseline values are re-ported. Scale measurement of the outcomesnot mentioned, e.g. pain scale
Other bias High risk Cross-over design with possible carry-overeffect. No information on safety
Mowat 1979
Methods Cross-over trialStudy duration: 8 weeks (4 weeks in each arm, no washout)Randomisation method: not specifiedBlinding: single-blinded (only physicians blinded)Sample size calculation: not described
Participants 31 patientsInclusion criteria: definite or classical RA (not further specified)Exclusion criteria: aspirin-hypersensitivity, breast feeding or pregnancy and compromisedhepatic or renal functionMean age (years): combination therapy arm: 59.6; monotherapy arm: 54.4Gender (% female): combination therapy arm: 58.8; monotherapy arm: 50Disease duration not specifiedComedication: all other NSAIDs were withdrawn4 withdrawals
Interventions Group 1 (combination therapy): benorylate (aspirin + paracetamol), starting dose 4 g,maximum 8 gGroup 2 (monotherapy): naproxen, starting dose 500 mg, maximum 1 gPatients were allowed to increase their dose of drug every 4th day up to the maximumdaily dosage
Outcomes 1) Pain at 4 weeks (0 to 10cm VAS)2) Morning stiffness3) Joint pain on full range of active movement4) Adverse events
Notes For the purpose of this review results were extracted from the first period for efficacy andfrom both periods for safetyPain (negative scores mean the pain improved):Group 1 (combination therapy): baseline 4.50 ± 1.48; improvement at 4 weeks of treat-
46Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
ment -0.29 ± 1.69Group 2 (monotherapy): baseline 4.57 ± 2.44; improvement at 4 weeks of treatment -0.29 ± 2.38No significant difference between the groups, as reported by the authors, but no P valueshownWithdrawals due to adverse events:Group 1 (combination therapy): 7% (n = 2)Group 2 (monotherapy): 3% (n = 1)Withdrawals due to inadequate analgesia:Group 1 (combination therapy): 3% (n = 1)Group 2 (monotherapy): 0% (n = 0)Number of patients with adverse events:Group 1 (combination therapy): 55% (n = 17): nausea, constipation, indigestion, vom-iting, hearing upset, headache, dry skin, rash (n per adverse event not specified)Group 2 (monotherapy): 45% (n = 14): nausea, constipation, indigestion, vomiting,hearing upset, dry skin, rash (n per adverse event not specified)No deaths were reported (interpretation of the authors of the review from what is re-ported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “Patientswere randomly divided into 2 groups”
Allocation concealment (selection bias) Low risk “The drugs were supplied in identicalsealed boxes with a medicine measure”
Blinding (performance bias and detectionbias)All outcomes
High risk Single-blinded study and “Patients wereasked not to reveal details of the physi-cal characteristics of the suspensions to thephysician to keep the trial single-blind”.This could bias patient-reported outcomeassessment
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Unclear how incomplete outcome datawere handled
Selective reporting (reporting bias) High risk Only the significant comparisons between2 treatment groups are reported, puttingemphasis on the time points in whichthere is a significant difference between thegroups
Other bias High risk Cross-over design with possible carry-overeffect. Baseline characteristics poorly com-pared (only age and sex)
47Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Parallel RCTStudy duration: 7 weeks (1 week washout period + 4 weeks study treatment + 2 weeksibuprofen)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 60 patientsInclusion criteria: classical or definite active RA (according to the ARA criteria), satisfyingat least 3 of the following criteria: Ritchie’s index >15; ESR > 25; duration of morningstiffness ≥ 30 min; and subjective pain index > 50 mmExclusion criteria: Hamilton rating scale for depression score of 7 to 19Mean age (years): depressed patients 48.1 ± 2.0; not depressed patients 50.2 ± 2.1Gender (% female): 87Disease duration (years): depressed patients 10.1 ± 1.9; not depressed patients 8.1 ± 1.3Comedication: not specified10 withdrawals
Outcomes 1) Hamilton rating scale for depression2) Cassano-Castrogiovanni self-evaluation rating scale for depression3) Ritchie’s index4) Lee index5) Duration of morning stiffness6) Grip strength7) Pain (0-100mm VAS)8) Daytime pain9) Night-time pain10) Physician’s evaluation of the efficacy of treatment11) Patient’s evaluation of the efficacy of treatment
Notes Pain at 4 weeks of treatment:Group 1 (combination therapy):- not depressed patients: baseline 57.0 ± 0.8; 4 weeks of treatment 43.0 ± 1.0- depressed patients: baseline 58.0 ± 0.5; 4 weeks of treatment 43.0 ± 0.5Group 2 (monotherapy):- not depressed patients: baseline 47.0 ± 0.7; 4 weeks of treatment 43.0 ± 1.0- depressed patients: baseline 59.0 ± 0.8; 4 weeks of treatment 51.7 ± 0.8No significant difference between the groups, as reported by the authors, but no P valueshownNo information on withdrawals due to adverse events, withdrawals due to inadequateanalgesia or number of patients with adverse events reported in the paper. No deathswere reported (interpretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
48Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Unclear risk Only information in the text: “…were ran-domly added to the ibuprofen therapy in adouble-blind design”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “…were ran-domly added to the ibuprofen therapy in adouble-blind design”. No information onwho exactly was blinded, how blinding wasassured or on an analysis of the patients’perception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk 10 withdrawals excluded from the analy-ses. Not clear how other missing data wereapproached, since a total n per outcome isnot given, and the number per randomisedgroup not mentioned either
Selective reporting (reporting bias) High risk Not all outcomes are reported (e.g. gripstrength, morning stiffness, although re-ported in the methods)
Other bias High risk Baseline characteristics not presented. Nospecification whether attempts were madeto limit cointerventions. No safety dataprovided
Saarialho-Kere 1988
Methods Cross-over trialStudy duration: 12 days (3 days on paracetamol + 3 days in each of the study arms, nowashout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 16 patientsInclusion criteria: classical RA (according to the ARA criteria) and current daily therapywith anti-inflammatory analgesicsExclusion criteria: gastrointestinal, hepatic, renal or psychic disease; previous intoleranceto the drugs testedMean age (years): not specifiedGender (% female): 88Disease duration (years): 6.2 ± 2.0Co-medication: stable doses of DMARDs (gold, penicillamine, chloroquine) were main-tained and unchanged throughout the study Patients were allowed to take extra parac-etamol on days 1 to 5 to alleviate pain and the amount of paracetamol was recorded
49Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Outcomes 1) Pain at 4 hours after the treatment (0 to 100 mm VAS)2) Duration of morning stiffness3) Grip strength4) Articular index5) Several psychomotor skills
Notes Results reflect the average of both cross-over periods (only available ones)Pain (at 4h after the treatment):Group 1 (combination therapy): baseline 50 ± 6.6; 4 hours 44 ± 6.6Group 2 (monotherapy 1): baseline 46 ± 6.0; 4 hours 34 ± 4.8Group 3 (monotherapy 2): baseline 49 ± 6.9; 4 hours 36 ± 5.6Significant difference between combination therapy and monotherapy 1, as reported bythe authors, P value < 0.05No information on withdrawals due to adverse events, withdrawals due to inadequateanalgesia or adverse events reported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “The trialcomprised four randomized double-blindcrossover treatment periods started at two-week intervals”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The trialcomprised four randomized double-blindcrossover treatment periods started at two-week intervals”. No information on whoexactly was blinded, how blinding was as-sured or on an analysis of the patients’ per-ception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
Unclear risk Not clear how missing data were ap-proached, since a total n per outcome is notgiven
Selective reporting (reporting bias) High risk Not all outcomes are reported (e.g. gripstrength, morning stiffness, although re-ported in the methods). No baseline val-ues are presented. Separate results of each
50Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Other bias High risk No data regarding safety. Cross-over trialwith possible carry-over effects
Seideman 1988
Methods Cross-over trialStudy duration: 3 days wash-out period without NSAIDs + 7 days run-in period (assesstolerability of indomethacin) + 4 weeks (2 weeks in each arm, no washout)Randomisation method: not specifiedBlinding: participants and outcome assessor blindedSample size calculation: not described
Participants 20 patientsInclusion criteria: classic or definite RA (as diagnosed by ARA criteria) with pain ofsufficient degree to require NSAID medicationExclusion criteria: gastrointestinal, hepatic or renal disease, known intolerance to in-domethacinMean age (years): 47 ± 10Gender: not specifiedDisease duration (years): 10 ± 8Comedication: DMARDs that had been given for at least 6 months were kept and notchanged throughout the study3 withdrawals
Interventions Group 1 (combination therapy): indomethacin 50 mg + paracetamol 4 g (randomised:n = 10; completing the treatment: n = 8)Group 2 (monotherapy): indomethacin 150 mg (randomised: n = 10; completing thetreatment: n = 9)
Outcomes 1) Pain (0 to 100 mm VAS); mean pain estimate during the 2nd and 4th weeks2) Duration of morning stiffness3) Grip strength4) Number of joints painful to digital pressure5) Joint circumference
Notes Results reflect the average of both cross-over periods (only available ones)Mean values for pain:Group 1 (combination therapy): 29.9 ± 26.4Group 2 (monotherapy): 29.2 ± 26.4Mean of the estimates made during the second and fourth treatment weeksNo significant difference between the groups, P value = 0.94 (calculated by authors ofthe review)Number of patients with adverse events:Group 1 (combination therapy): 55% (n = 11): headache, tiredness and vertigo (n = 3);anorexia, dyspepsia and vomiting (n = 1)Group 2 (monotherapy): 20% (n = 4): headache, tiredness and vertigo (n = 6); anorexia,
51Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
dyspepsia and vomiting (n = 5)No withdrawals due to inadequate analgesia, no withdrawals due to adverse events and nodeaths were reported (interpretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “Half ofthe patients were subsequently randomizedto…”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Low risk “The study was performed in a double-blind crossover manner”. “Due to differ-ences in appearance of the formulations, adouble-dummy technique was used”
Incomplete outcome data (attrition bias)All outcomes
Unclear risk 3 patients withdrawn from the study, ap-parently for reasons different to the exclu-sion criteria. From the included patients, itis unclear which are the missing data sinceresults are not presented with numbers ofpatients having been assessed per outcomeand per time point
Selective reporting (reporting bias) High risk No baseline values are presented. Separateresults of each cross-over period are not pre-sented
Other bias Unclear risk Cross-over trial with possible carry-over ef-fects
Seideman 1993
Methods Cross-over trialStudy duration: 3 to 7 day “flare” period + 4 weeks (2 weeks in each arm, no washout)Randomisation method: not specifiedBlinding: participants and outcome assessor blindedSample size calculation: not described
Participants 20 patientsInclusion criteria: classic or definite RA (as diagnosed by ARA criteria) with pain ofsufficient degree to require NSAID medicationExclusion criteria: not specifiedMean age (years): 52.4Gender (% female): 55
52Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Disease duration (years): 4Co-medication: DMARDs that had been given for at least 6 months were kept andnot changed throughout the study. 12 patients on 2nd line drugs (penicillamine: 4;aurothiomalate: 4; chloroquine: 4)No withdrawals
Outcomes 1) Number of joints painful to digital pressure or passive movements2) Duration of morning stiffness3) Pain at 2 weeks (0 to 100 mm VAS)4) Global assessment of disease activity5) Activity of daily living assessment at 2 weeks
Notes Results correspond to the average from all cross-over periods and baseline is the averagepre-treatment value, so not individualised for each group (only available ones)Pain at 2 weeks:Pretreatment (all the groups): 75 ± 16Group 1 (combination therapy 1): 45.7 ± 14.6Group 2 (monotherapy 1): 61.5 ± 15.9Significant difference between the groups (P < 0.001)Group 3 (combination therapy 2): 31.7 ± 9.6Group 4 (monotherapy 2): 46.5 ± 14.6Significant difference between the groups (P < 0.05)Function at 2 weeks:Pretreatment (all the groups): 3.8 ± 0.7Group 1 (combination therapy 1): 4.6 ± 0.3Group 2 (monotherapy 1): 4.5 ± 0.3No significant difference between the groups, as reported by the authors, no P valueshownGroup 3 (combination therapy 2): 4.6 ± 0.3Group 4 (monotherapy 2): 4.6 ± 0.3No significant difference between the groups, as reported by the authors, no P valueshownNo withdrawals due to inadequate analgesia, no withdrawals due to adverse events and nodeaths were reported (interpretation of the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “Each dosewas given for 2-week periods in a random-ized order”
Allocation concealment (selection bias) Unclear risk Not mentioned
53Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Blinding (performance bias and detectionbias)All outcomes
Low risk “The study was double-blind with identicalnaproxen tablets and placebo and identicalparacetamol and placebo”. “Due to differ-ences in appearance of the formulations, adouble-dummy technique was used”
Incomplete outcome data (attrition bias)All outcomes
Low risk No missing outcome data
Selective reporting (reporting bias) High risk No baseline values are presented. Separateresults of each cross-over period are not pre-sented
Other bias High risk Cross-over design with possible carry-overeffects. Safety data not presented
Sharma 1978
Methods Cross-over trialStudy duration: 3 days (1 night in each arm, no washout)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 18 patientsInclusion criteria: hospitalisation for the treatment of active RA, classical and definite(as defined by the ARA)Exclusion criteria: current therapy with corticosteroids, > 75 mg indomethacin or anyadditional dose of sulindac; known hypersensitivity to any of the trial medicationsMean age (years): 47.6Gender (% female): 89Disease duration not specifiedComedication: no other analgesic or hypnotic1 withdrawal
Outcomes 1) Duration of morning stiffness2) Night pain at 1 day of treatment (0 to 10 cm VAS)3) Quality of sleep4) Adverse events
Notes Results reflect the average of all cross-over periods (only available ones)Pain (at 1 day of treatment):Group 1 (combination therapy 1): 4.76Group 2 (monotherapy): 4.07
54Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Group 3 (combination therapy 2): 3.39No significant difference between groups 1 and 2, as reported by the authors, but no Pvalue shownNo significant difference between groups 2 and 3, as reported by the authors,P value =0.061Number of patients with adverse events:Group 1 (combination therapy 1): 22% (n = 4); headache (n = 1); drowsiness (n = 3)Group 2 (monotherapy): 22% (n = 4): headache (n = 1); drowsiness (n = 2); depression(n = 1)Group 3 (combination therapy 2): 33% (n = 6): drowsiness (n = 5); depression (n = 1)1 patient withdrew due to adverse events, but it is not clear in which arm he/she wasNo withdrawals due to inadequate analgesia and no deaths were reported (interpretationof the authors of the review from what is reported)
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “…and thetreatment order was randomized”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “The dou-ble-blind nature of the trial was ensured byusing appropriate dummy tablets or cap-sules”. No information on who exactly wasblinded, how blinding was assured or on ananalysis of the patients’ perception of whicharm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk One withdrawal excluded from the analy-sis, after being included in the study. Num-ber of patients randomised into each groupnot clear
Selective reporting (reporting bias) High risk Separate results of each cross-over periodare not reported. No baseline values arereported. No measures of spread are pre-sented with the means (SD)
Other bias High risk Cross-over design with possible carry-overeffect. Number of patients randomised foreach treatment order is not clear
55Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Methods Parallel RCTStudy duration: 13 weeks (1 week of placebo + 12 weeks in one of the study arms)Randomisation method: predetermined random codeBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 41 patientsInclusion criteria: classical or definite RA (not further specified)Exclusion criteria: therapy with steroids, gold or antimalarial drugs within the previousyear; definite history of aspirin intoleranceMean age (years): combination therapy arm: 59; monotherapy arm: 56Gender (% female): combination therapy arm: 81; monotherapy arm: 76Disease duration not specifiedComedication: no other analgesics or anti-inflammatory drugs were permitted and phys-iotherapy was withheld for the duration of the trial8 withdrawals
Interventions Group 1 (combination therapy): benorylate (aspirin + paracetamol) 6 g (n = 20)Group 2 (monotherapy): aspirin 4 g (n = 21)
Outcomes 1) Pain severity (1 = mild, 2 = moderate, 3 = severe) at 12 weeks of treatment2) Number of inflamed joints3) Morning stiffness4) Articular index5) Grip strength6) Functional status
Notes Improvement in pain, at 12 weeks of treatment:Group 1 (combination therapy): baseline pain level 0: 0%; 1: 0%; 2: 30% (n = 6); 3:45% (n = 9); at 12 weeks: better 50% (n = 10); unchanged 20% (n = 4); worse 10% (n= 2)Group 2 (monotherapy): baseline pain level 0: 0%; 1: 0%; 2: 43% (n = 9); 3: 38% (n =8); at 12 weeks: better 48% (n = 10); unchanged 24% (n = 5); worse 10% (n = 2)No significant difference between the groups, P value > 0.05Withdrawals due to adverse events:Group 1 (combination therapy): 10% (n = 2)Group 2 (monotherapy): 10% (n = 2)Withdrawals due to inadequate analgesia:Group 1 (combination therapy): 0% (n = 0)Group 2 (monotherapy): 10% (n = 2)Number of patients with adverse events:Group 1 (combination therapy): 45% (n = 9): indigestion; nausea; flautulence; consti-pation; drowsiness; tinnitus; itchy legs; nocturia (n per adverse event not specified)Group 2 (monotherapy): 71% (n = 15): indigestion; nausea; flautulence; constipation;drowsiness; macular rash; sweating; nocturia (n per adverse event not specified)No deaths were reported (interpretation of the authors of the review from what is re-ported)
Risk of bias
56Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Unclear risk Only information in the text: “...patientswere treated according to a predeterminedrandom code”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the title: “Double-blind comparison...” No information onwho exactly was blinded, how blinding wasassured or on an analysis of the patients’perception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk Withdrawals were excluded from the anal-yses, even after having been assigned to atreatment group
Selective reporting (reporting bias) High risk Not all outcomes are presented, e.g. func-tional status. No measures of spread are pre-sented with some of the means (SD). Someoutcomes (e.g. pain, morning stiffness, gripstrength) presented in a different way thanwhat was described in methods, and alsodifferently compared to the baseline valuepresentation
Other bias Unclear risk Scale measurement of the outcomes not al-ways clear and varying between the presen-tation of baseline results and of the changeresults
Staunton 1980
Methods Parallel RCTStudy duration: 6 weeks (2 weeks on indomethacin 75 mg + 4 weeks in one of the studyarms)Randomisation method: not specifiedBlinding: reference to double-blinded, not specifiedSample size calculation: not described
Participants 18 patientsInclusion criteria: classical or definite RA (according to the ARA criteria). Only thosepatients aged 30 to70 years were included and the diagnosis must have been establishedat least 6 months prior to entry. Clear evidence of active disease had to be apparent whenthe patients were off all anti-inflammatory and analgesic drugsExclusion criteria: pregnancy and lactation; ARA functional class I; therapy with systemicor intra-articular steroid therapy or ACTH within 6 weeks prior to entry, immunosup-
57Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
pressive therapy during the previous year, penicillamine within 6 months prior to entry,anticoagulants, oral hypoglycaemic agents, barbiturates and phenothiazines (crysother-apy was permitted, as long as the dosage regime had not altered within 6 months prior toentry); presence of any other rheumatic or collagen disorder or evidence of concomitantdisease that might affect jointsMean age (years): combination therapy arm: 56.5 ± 10.5; monotherapy arm: 50.3 ± 12.4Gender (% female): combination therapy arm: 56; monotherapy arm: 100Mean disease duration (years): combination therapy arm: 12; monotherapy arm: 6.9Comedication: patients were requested to stop taking any current anti-inflammatory oranalgesic drugs for a withdrawal period of 2 to 7 days and during the study these drugswere not allowed (except for the trial drugs)2 withdrawals (while on indomethacin only, so not during the 4 weeks under studydrugs)
Outcomes 1) Ritchie articular index2) Grip strength3) Morning stiffness4) Pain during the day and during the night (1 = none; to 4 = very severe pain)5) Physician’s global evaluation6) Patient’s global evaluation
Notes Pain:No significant difference between the groups, as reported by the authors; neither theresults, nor the P value for the comparison is shownNo withdrawals due to adverse events, withdrawals due to inadequate analgesia or deathswere reported (interpretation of the authors of the review from what is reported). Noinformation on the number of adverse events is reported in the paper
Risk of bias
Bias Authors’ judgement Support for judgement
Random sequence generation (selectionbias)
Unclear risk Only information in the text: “A double-blind randomized trial…”, “Patients weregiven an allocation number and then re-ceived indomethacin 25 mg capsules threetimes daily for two weeks. They were thenallocated to one of two treatment groupsaccording to their allocation number...”
Allocation concealment (selection bias) Unclear risk Not mentioned
Blinding (performance bias and detectionbias)All outcomes
Unclear risk Only information in the text: “A double-blind randomized trial…”. No informationon who exactly was blinded, how blinding
58Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
was assured or on an analysis of the patients’perception of which arm they were in
Incomplete outcome data (attrition bias)All outcomes
High risk Not clear which are incomplete outcomedata. Data not presented. Number of pa-tients finishing each arm not presented. 2withdrawals, that withdrew during the run-in period, but not even clear if they wereincluded in the analyses
Selective reporting (reporting bias) High risk Not all outcomes are presented, e.g. gripstrength. No measures of spread are pre-sented with some of the means (SD). Nofigures for pain assessment are presented.No baseline value for all the outcomes re-ported
Other bias High risk Differences in baseline characteristics
ARA - American Rheumatism Association; ESR - erythrocyte sedimentation rate; VAS - visual analogue scale; PIP - proximal inter-phalangeal
Characteristics of excluded studies [ordered by study ID]
Study Reason for exclusion
Andrade-Padilla 1995 Outcome - no pain assessment
Bain 1970 Outcome - wrong assessment (no global pain assessment)
Barnardo 1966 Intervention - monotherapy
Brooks 1977 Outcome - no understandable pain scale
Cardoe 1970 Outcome - wrong assessment (no global pain assessment)
Chalmers 1978 Outcome - wrong assessment (no global pain assessment)
Dalmases 1966 Outcome - no pain assessment
De Mattos 1968 Study type - observational study
Eberl 1968 Study type - observational study
59Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Appendix 1. MEDLINE and CENTRAL search strategy (OVID format)
1. exp arthritis, rheumatoid/2. ((rheumatoid or reumatoid or revmatoid or rheumatic or reumatic or revmatic or rheumat$ or reumat$ or revmarthrit$) adj3 (arthrit$or artrit$ or diseas$ or condition$ or nodule$)).tw.3. (felty$ adj2 syndrome).tw.4. (caplan$ adj2 syndrome).tw.5. (sjogren$ adj2 syndrome).tw.6. (sicca adj2 syndrome).tw.7. still$ disease.tw.8. bechterew$ disease.tw.9. exp Spondylarthropathies/10. (ankylos$ or spondyl$).tw.11. (bekhterev$ or bechterew$).tw.12. (Marie adj struempell$).tw.13. Arthritis, Psoriatic/14. (psoria$ adj (arthriti$ or arthropath$)).tw.15. ((arthriti$ or arthropath$) adj psoria$).tw.16. exp Arthritis, Infectious/17. reactive arthritis.tw.18. (reiter$ adj (disease or syndrome)).tw.19. ((sexual$ or chlamydia or yersinia or postyersinia or postdysenteric or salmonella or shigella or b27 or postinfectious or postinfectious) adj5 arthrit$).tw.20. reactive enthesitis.tw.21. undifferentiated oligoarthritis.tw.22. or/1-2123. exp Inflammatory Bowel Diseases/24. exp Arthritis/25. 23 and 2426. ((inflamm$ or ibd or chrohn$ or enteropath$) adj5 (arthrit$ or arthrop$)).tw.27. (granulomatous colitis adj5 (arthrit$ or arthrop$)).tw.28. (ulcerative colitis adj5 (arthrit$ or arthrop$)).tw.29. (granulomatous enteritis adj5 (arthrit$ or arthrop$)).tw.30. (regional enteritis adj5 (arthrit$ or arthrop$)).tw.31. (Ileocolitis adj5 (arthrit$ or arthrop$)).tw.32. (terminal ileitis adj5 (arthrit$ or arthrop$)).tw.33. (regional ileitis adj5 (arthrit$ or arthrop$)).tw.34. or/22,25-3335. exp Analgesics, Opioid/36. exp Narcotics/37. narcotic$.tw.38. opiate$.tw.39. opioid$.tw.
62Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
93. oxymorph$.tw.94. pantopon.tw.95. papaveretum.tw.96. paracymethadol.tw.97. paregoric.tw.98. pentazocine.tw.99. pethidine.tw.100. phenazocine.tw.101. phenbenzorphan.tw.102. phenethylazocine.tw.103. phenoperidine.tw.104. pirinitramide.tw.105. propoxyphene.tw.106. protopine.tw.107. pyrrolamidol.tw.108. remifentanil.tw.109. sufentanil.tw.110. sufentanyl.tw.111. talwin.tw.112. tapentadol.tw.113. thebaine.tw.114. theocodin.tw.115. tilidine.tw.116. tramadol.tw.117. trimeperidine.tw.118. dihydromorphine.tw.119. hydroxycodeinone.tw.120. levomethadyl.tw.121. dihydrocodein$.tw.122. dihydrohydroxycodeinone.tw.123. dipipanone.tw.124. exp Muscle Relaxants, Central/125. exp Neuromuscular Nondepolarizing Agents/126. exp Neuromuscular Blocking Agents/127. exp Benzodiazepines/128. muscle relaxant$.tw.129. benzodiazepine$.tw.130. (Alprazolam or Xanax or Xanor or Tafil or Alprox or Frontal).tw.131. (Bromazepam or Lexotanil or Lexotan or Lexomil or Somalium or Bromam).tw.132. (Chlordiazepoxide or Librium or Tropium or Risolid or Klopoxid).tw.133. (Cinolazepam or Gerodorm).tw.134. (Clonazepam or Klonopin or Rivotril or Iktorivil).tw.135. (Cloxazolam or Olcadil).tw.136. (Clorazepate or Tranxene).tw.137. (Diazepam or Valium or Pax or Apzepam or Stesolid).tw.138. (Estazolam or ProSom).tw.139. (Flunitrazepam or Rohypnol or Fluscand or Flunipam or Rona or Rohydorm).tw.140. (Flurazepam or Dalmadorm or Dalmane).tw.141. (Flutoprazepam or Restas).tw.142. (Halazepam or Paxipam or Ketazolam or Anxon or Loprazolam or Dormonoct or lorazepam or Ativan or Temesta or Tavor orLorabenz or Lormetazepam or Loramet or Noctamid or Pronoctan or Medazepam or Nobrium or Midazolam or Dormicum or Versedor Hypnovel or Dormonid or Nimetazepam or Erimin or Nitrazepam or Mogadon or Alodorm or Pacisyn or Dumolid or Nordazepamor Madar or Stilny or Oxazepam or Seresta or Serax or Serenid or Serepax or Sobril or Pinazepam or Domar or Prazepam or Lysanxia or
64Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
Centrax or Quazepam or Doral or Temazepam or Restoril or Normison or Euhypnos or Tenox or Tetrazepam or Mylostan or Triazolamor Halcion or Rilamir).tw.143. (Orphenadrine or Norflex or Mephenamin or Disipal or Banflex or Flexon or Tizanidine or Zanaflex or Sirdalud or Flupirtine orDantrolene or Dantrium or Dantrolen or Baclofen or Kemstro or Lioresal).tw.144. exp Neurotransmitter Agents/145. neuromodulator$.tw.146. neurohumor$.tw.147. exp Anticonvulsants/148. Anticonvulsant$.tw.149. gabapentin.tw.150. neurontin.tw.151. (Carbamazepine or Tegretol).tw.152. (Clonazepam or Klonopin or Rivotril or Rivatril).tw.153. (Lamotrigine or Lamictal).tw.154. (Oxcarbazepine or Trileptal or Oxaleptal).tw.155. (Phenytoin or Phenytek or Dilantin or Eptoin or Epanutin or Diphenin or Dipheninum or Phydum).tw.156. (Pregabalin or Lyrica).tw.157. (Topiramate or Topamax).tw.158. (Valproic Acid or Valproate or Epilim or Depakote).tw.159. (Ketamine or Ketalar or Ketaset, Ketajet, Vetalar, Rogarsetic).tw.160. (Capsaicin or Zostrix or Capsin or Capzasin-P).tw.161. exp Capsaicin/162. exp Ketamine/163. exp Bupropion/164. (Bupropion or Wellbutrin or Zyban).tw.165. exp Methylphenidate/166. (Dexmethylphenidate or Focalin).tw.167. (Methylphenidate or Ritalin or Concerta).tw.168. Antidepressive Agents/169. (anti-depressant$ or antidepressant$).tw.170. exp Serotonin Uptake Inhibitors/171. exp Serotonin Antagonists/172. SSRI$.tw.173. (Citalopram or Celexa).tw.174. (Fluoxetine or Prozac).tw.175. (Paroxetine or Paxil or Seroxat).tw.176. (Sertraline or Zoloft or Lustral).tw.177. (Escitalopram or Lexapro or Cipralex).tw.178. (Fluvoxamine or Luvox).tw.179. (Desvenlafaxine or Pristiq).tw.180. (Venlafaxine or Effexor).tw.181. (Duloxetine or Cymbalta).tw.182. (Milnacipran or Ixel or Savella).tw.183. (Reboxetine or Edronax).tw.184. (Viloxazine or Vivalan).tw.185. Amitriptyline/186. (Amitriptyline or Elavil or Endep).tw.187. (Clomipramine or Anafranil).tw.188. (Desipramine or Norpramin or Pertofrane).tw.189. (Dosulepin or Dothiepin or Prothiaden).tw.190. (Doxepin or Adapin or Sinequan).tw.191. (Imipramine or Tofranil).tw.192. (Lofepramine or Feprapax or Gamanil or Lomont).tw.
65Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
193. (Nortriptyline or Pamelor).tw.194. (Protriptyline or Vivactil).tw.195. (Trimipramine or Surmontil).tw.196. (Amoxapine or Asendin).tw.197. (Loxapine or Loxapac or Loxitane).tw.198. (Maprotiline or Deprilept or Ludiomil or Psymion).tw.199. (Mazindol or Mazanor or Sanorex).tw.200. (Mianserin or Bolvidon or Norval or Tolvon).tw.201. (Mirtazapine or Remeron or Avanza or Zispin).tw.202. (Setiptiline or Tecipul).tw.203. exp Monoamine Oxidase Inhibitors/204. (Isocarboxazid or Marplan or Moclobemide or Aurorix or Manerix or Phenelzine or Nardil or Selegiline or L-Deprenyl or Eldeprylor Zelapar or Emsam or Tranylcypromine or Parnate or Moclobemide or Aurorix or Manerix).tw.205. Phenelzine/206. Moclobemide/207. Selegiline/208. Tranylcypromine/209. Isocarboxazid/210. Monoamine Oxidase Inhibitor$.tw.211. serotonin-norepinephrine reuptake inhibitor$.tw.212. norepinephrine reuptake inhibitor$.tw.213. serotonin-noradrenaline reuptake inhibitor$.tw.214. noradrenaline reuptake inhibitor$.tw.215. (Duloxetine or Cymbalta or Milnacipran or Ixel or Savella).tw.216. (Reboxetine or Edronax or Viloxazine or Vivalan).tw.217. exp Anti-Inflammatory Agents, Non-Steroidal/218. Anti-Inflammator$.tw. or (Anti adj Inflammator$).tw. or AntiInflammator$.tw.219. nsaid$.tw.220. Ampyrone.tw.221. Antipyrine.tw.222. Apazone.tw.223. Aspirin.tw.224. Bufexamac.tw.225. Clofazimine.tw.226. Clonixin.tw.227. Curcumin.tw.228. Diclofenac.tw.229. Diflunisal.tw.230. Dipyrone.tw.231. Epirizole.tw.232. Etodolac.tw.233. Fenoprofen.tw.234. Flurbiprofen.tw.235. Ibuprofen.tw.236. Indomethacin.tw.237. Ketoprofen.tw.238. Ketorolac.tw.239. Meclofenamic Acid.tw.240. Mefenamic Acid.tw.241. Mesalamine.tw.242. Naproxen.tw.243. Niflumic Acid.tw.244. Oxyphenbutazone.tw.
66Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
245. Phenylbutazone.tw.246. Piroxicam.tw.247. Prenazone.tw.248. Salicylate$.tw.249. Sulfasalazine.tw.250. Sulindac.tw.251. Suprofen.tw.252. Tolmetin.tw.253. exp Acetaminophen/254. acet?minophen.tw.255. paracetamol.tw.256. acetamidophenol.tw.257. acephen.tw.258. acetaco.tw.259. tylenol.tw.260. anacin$.tw.261. datril.tw.262. panadol.tw.263. acamol.tw.264. algotropyl.tw.265. or/35-264266. 34 and 265267. randomized controlled trial.pt.268. controlled clinical trial.pt.269. randomized.ab.270. placebo.ab.271. drug therapy.fs.272. randomly.ab.273. trial.ab.274. groups.ab.275. or/267-274276. (animals not (humans and animals)).sh.277. 275 not 276278. 266 and 277
Appendix 2. EMBASE search strategy (OVID format)
1. exp arthritis, rheumatoid/2. ((rheumatoid or reumatoid or revmatoid or rheumatic or reumatic or revmatic or rheumat$ or reumat$ or revmarthrit$) adj3 (arthrit$or artrit$ or diseas$ or condition$ or nodule$)).tw.3. (felty$ adj2 syndrome).tw.4. (caplan$ adj2 syndrome).tw.5. (sjogren$ adj2 syndrome).tw.6. (sicca adj2 syndrome).tw.7. still$ disease.tw.8. bechterew$ disease.tw.9. exp spondyloarthropathy/10. (ankylos$ or spondyl$).tw.11. (bekhterev$ or bechterew$).tw.12. (Marie adj struempell$).tw.13. psoriatic arthritis/14. (psoria$ adj (arthriti$ or arthropath$)).tw.
67Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
120. dihydrocodein$.tw.121. dihydrohydroxycodeinone.tw.122. dipipanone.tw.123. exp muscle relaxant agent/124. exp benzodiazepine derivative/125. muscle relaxant$.tw.126. benzodiazepine$.tw.127. (Alprazolam or Xanax or Xanor or Tafil or Alprox or Frontal).tw.128. (Bromazepam or Lexotanil or Lexotan or Lexomil or Somalium or Bromam).tw.129. (Chlordiazepoxide or Librium or Tropium or Risolid or Klopoxid).tw.130. (Cinolazepam or Gerodorm).tw.131. (Clonazepam or Klonopin or Rivotril or Iktorivil).tw.132. (Cloxazolam or Olcadil).tw.133. (Clorazepate or Tranxene).tw.134. (Diazepam or Valium or Pax or Apzepam or Stesolid).tw.135. (Estazolam or ProSom).tw.136. (Flunitrazepam or Rohypnol or Fluscand or Flunipam or Rona or Rohydorm).tw.137. (Flurazepam or Dalmadorm or Dalmane).tw.138. (Flutoprazepam or Restas).tw.139. (Halazepam or Paxipam or Ketazolam or Anxon or Loprazolam or Dormonoct or lorazepam or Ativan or Temesta or Tavor orLorabenz or Lormetazepam or Loramet or Noctamid or Pronoctan or Medazepam or Nobrium or Midazolam or Dormicum or Versedor Hypnovel or Dormonid or Nimetazepam or Erimin or Nitrazepam or Mogadon or Alodorm or Pacisyn or Dumolid or Nordazepamor Madar or Stilny or Oxazepam or Seresta or Serax or Serenid or Serepax or Sobril or Pinazepam or Domar or Prazepam or Lysanxia orCentrax or Quazepam or Doral or Temazepam or Restoril or Normison or Euhypnos or Tenox or Tetrazepam or Mylostan or Triazolamor Halcion or Rilamir).tw.140. (Orphenadrine or Norflex or Mephenamin or Disipal or Banflex or Flexon or Tizanidine or Zanaflex or Sirdalud or Flupirtine orDantrolene or Dantrium or Dantrolen or Baclofen or Kemstro or Lioresal).tw.141. neuromodulation/142. neuromodulator$.tw.143. neurohumor$.tw.144. exp anticonvulsive agent/145. Anticonvulsant$.tw.146. gabapentin.tw.147. neurontin.tw.148. (Carbamazepine or Tegretol).tw.149. (Clonazepam or Klonopin or Rivotril or Rivatril).tw.150. (Lamotrigine or Lamictal).tw.151. (Oxcarbazepine or Trileptal or Oxaleptal).tw.152. (Phenytoin or Phenytek or Dilantin or Eptoin or Epanutin or Diphenin or Dipheninum or Phydum).tw.153. (Pregabalin or Lyrica).tw.154. (Topiramate or Topamax).tw.155. (Valproic Acid or Valproate or Epilim or Depakote).tw.156. (Ketamine or Ketalar or Ketaset, Ketajet, Vetalar, Rogarsetic).tw.157. (Capsaicin or Zostrix or Capsin or Capzasin-P).tw.158. capsaicin/159. ketamine/160. (Bupropion or Wellbutrin or Zyban).tw.161. methylphenidate/162. (Dexmethylphenidate or Focalin).tw.163. (Methylphenidate or Ritalin or Concerta).tw.164. exp antidepressant agent/165. (anti-depressant$ or antidepressant$).tw.166. exp serotonin antagonist/
70Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
167. SSRI$.tw.168. (Citalopram or Celexa).tw.169. (Fluoxetine or Prozac).tw.170. (Paroxetine or Paxil or Seroxat).tw.171. (Sertraline or Zoloft or Lustral).tw.172. (Escitalopram or Lexapro or Cipralex).tw.173. (Fluvoxamine or Luvox).tw.174. (Desvenlafaxine or Pristiq).tw.175. (Venlafaxine or Effexor).tw.176. (Duloxetine or Cymbalta).tw.177. (Milnacipran or Ixel or Savella).tw.178. (Reboxetine or Edronax).tw.179. (Viloxazine or Vivalan).tw.180. (Amitriptyline or Elavil or Endep).tw.181. (Clomipramine or Anafranil).tw.182. (Desipramine or Norpramin or Pertofrane).tw.183. (Dosulepin or Dothiepin or Prothiaden).tw.184. (Doxepin or Adapin or Sinequan).tw.185. (Imipramine or Tofranil).tw.186. (Lofepramine or Feprapax or Gamanil or Lomont).tw.187. (Nortriptyline or Pamelor).tw.188. (Protriptyline or Vivactil).tw.189. (Trimipramine or Surmontil).tw.190. (Amoxapine or Asendin).tw.191. (Loxapine or Loxapac or Loxitane).tw.192. (Maprotiline or Deprilept or Ludiomil or Psymion).tw.193. (Mazindol or Mazanor or Sanorex).tw.194. (Mianserin or Bolvidon or Norval or Tolvon).tw.195. (Mirtazapine or Remeron or Avanza or Zispin).tw.196. (Setiptiline or Tecipul).tw.197. (Isocarboxazid or Marplan or Moclobemide or Aurorix or Manerix or Phenelzine or Nardil or Selegiline or L-Deprenyl or Eldeprylor Zelapar or Emsam or Tranylcypromine or Parnate or Moclobemide or Aurorix or Manerix).tw.198. Monoamine Oxidase Inhibitor$.tw.199. serotonin-norepinephrine reuptake inhibitor$.tw.200. norepinephrine reuptake inhibitor$.tw.201. serotonin-noradrenaline reuptake inhibitor$.tw.202. noradrenaline reuptake inhibitor$.tw.203. (Duloxetine or Cymbalta or Milnacipran or Ixel or Savella).tw.204. (Reboxetine or Edronax or Viloxazine or Vivalan).tw.205. exp nonsteroid antiinflammatory agent/206. Anti-Inflammator$.tw.207. (Anti adj Inflammator$).tw.208. AntiInflammator$.tw.209. nsaid$.tw.210. Ampyrone.tw.211. Antipyrine.tw.212. Apazone.tw.213. Aspirin.tw.214. Bufexamac.tw.215. Clofazimine.tw.216. Clonixin.tw.217. Curcumin.tw.218. Diclofenac.tw.
71Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
HR, AvT, RL, DvdH, RB and DA contributed to the final version of the review by providing comments and suggestions on draftversions of the review.
All authors approved the final version of the manuscript.
D E C L A R A T I O N S O F I N T E R E S T
None known.
S O U R C E S O F S U P P O R T
Internal sources
• Department of Medicine, Division of Rheumatology, Maastricht University Medical Centre, Netherlands.In kind support
• Division of Rheumatology, Department of Internal Medicine 3, Medical University Vienna, Austria.In kind support
• Department of Clinical Epidemiology, Cabrini Hospital; Department of Epidemiology and Preventive Medicine, MonashUniversity, Australia.In kind support
73Combination therapy for pain management in inflammatory arthritis (rheumatoid arthritis, ankylosing spondylitis, psoriatic arthritis,
D I F F E R E N C E S B E T W E E N P R O T O C O L A N D R E V I E W
The primary outcome was changed to “Patient reported pain relief of 50% or greater” instead of “Patient reported pain relief of 30% orgreater” on the recommendation of the CMSG. The latter was considered to be a secondary outcome. The change was made because50% improvement in pain is more commonly assessed in IA (e.g. ACR50 improvement in RA).
Clinical heterogeneity of the included studies precluded pooling of the data, statistical heterogeneity assessment and subsequentsensitivity and subgroup analyses, as well as the summary of findings table presentation. Results of the included studies could only bedescribed.
