Evidence based management of acute musculoskeletal pain

ACUTE SHOULDER PAIN — Table of Excluded Studies (Interventions) Appendix E • Tables of Included and Excluded Studies STUDY REASON Aaras 2001 Not shoulder-specific — shoulder data not presented separately; not acute; not randomised Arslan 2001 Mean duration of symptoms greater than 3 months Bain 2001 Trial of intra-operative pain relief. Pain > 6 months. Not relevant Bang 2000 Mean duration of symptoms > 5 months Barber 2001 Does not fit inclusion criteria Berry 1980 Duration of symptoms greater than 3 months; used for mixed populations Binder 1984b Duration of symptoms greater than 3 months; used for mixed data Binder 1986 Chronic pain Blockey 1954 Frozen shoulder, periarthritis not defined; not acute (some acute patients but not reported separately) Bottoni 2002 Fracture dislocation rather than acute (non-specific) shoulder pain Brosseau 2002a Exclude as relates to knee tendonitis only Brox 1993/7 Duration of symptoms greater than 3 months 238 Bulgen 1984 Duration of symptoms greater than 3 months Carter 2001 Case series Ceccherelli 2001 Chronic pain Evidence-based Management of Acute Musculoskeletal Pain Chiou 2001 All chronic subjects Dacre 1989 Duration of symptoms at least 4 weeks and periarthritis. Duration of symptoms not reported Dahan 2000 Mean duration of pain is 1 year; inclusion criteria is at least 1 month of surgery Dal Conte 1990 Duration of symptoms greater than 3 months Downing 1986 Duration of symptoms greater than 3 months England 1989 Duration of symptoms greater than 3 months Fankhauser 2002 Not an RCT Ferrante 1998 Results not separate for shoulder; subjects had pain duration > 6 months Gam 1995 Exclude as includes many musculoskeletal conditions; not specific to shoulder Ginn 1997 Duration of symptoms greater than 3 months (used for mixed data) Green 2002 Mixture of acute and chronic pain duration (used to obtain data on mixed populations) Haake 2002 Exclude due to chronic pain > 6 months

ACUTE SHOULDER PAIN — Table of Excluded Studies (Interventions) continued Appendix E • Tables of Included and Excluded Studies STUDY REASON Horneij 2001 Heterogeneous group — prospective prevention trial Invargarsson 1996 Surgery trial. Duration of symptoms unlikely to be less than 3 months Johansson 2002 Not acute shoulder pain (based on references included). Not stated in paper but studies included in SR are on chronic patients Jones 1999 Likely not to be acute but no duration given (2nd or 3rd stage capsulitis) Karatas 2002 Duration of symptoms not defined except for > 4 weeks; likely to be > 12 weeks; duration of study f/u is 60 minutes Karjalainen 2002 Greater than 3 weeks; chronic Ketola 2002 Neck/shoulder pain. Topic not relevant Kivimaki 2001 Pain duration > 12 weeks; no control (placebo); poor inclusion definition; inadequate manipulation Klein 2002 Polio survivors therefore not representative of the general population presenting to primary practice; maximum duration of symptoms unclear Leclaire 1999 Duration of symptoms greater than 3 months Lee 1973 Duration of symptoms greater than 3 months Lesprit 2001 Case series 239 Lindh 1993 Surgery trial. Duration of symptoms unlikely to be less than 3 months Moore 1976 Mean duration of symptoms not reported Nicholson 1985 Duration of symptoms greater than 3 months Evidence-based Management of Acute Musculoskeletal Pain Nykanen 1995 Duration of symptoms reported to be greater than 2 months; mean duration not reported but unlikely to be less than 3 months Oldervoll 2001 Not a randomised trial; not acute duration; not just shoulder pain Perron 1997 Duration of symptoms greater than 3 months Price 2002 Post stroke Reid 1996 Unlikely to be acute Ritchie 1995 Not a RCT Rizk 1991 Chronic pain (mean duration = 13.2 weeks) Rompe 2001 Pain duration greater than 3 months (> 12 months) Saunders 1995 Duration of symptoms greater than 3 months Schmitt 2001 Chronic pain (> 6 months) Shibata 2001 Chronic pain Sileghem 1991 Chronic pain — definition of diagnosis unclear

ACUTE SHOULDER PAIN — Table of Excluded Studies (Interventions) continued Appendix E • Tables of Included and Excluded Studies STUDY REASON Snels 2000a Hemiplegic shoulder pain therefore exclude Snels 2000b Survey Spangehl 2002 Exclude: chronic pain; mean duration > 12 months Speed 2002 Not acute — useful for chronic Sperber 2001 For post-traumatic instability Sun 2001 Duration of pain > 3/12 based on mean duration of symptoms Taverna 1990 Duration of symptoms greater than 3 months van der Heijden 1996 3/16 RCTs included are acute; included studies poor quality; unable to pool results. The three studies were reviewed separately van der Heijden 1997 Physiotherapy for soft tissue disorders van der Windt 1998 Half the participants had pain for greater than three months (used for mixed data) van der Windt 1995 Efficacy of NSAIDs for shoulder pain van der Windt 1999 Not shoulder-specific 240 Waling 2002 Not acute — includes cervical spine Walsh 2001 Case report Wang 2001 No comparison group (case series). All had pain longer than 6 months Evidence-based Management of Acute Musculoskeletal Pain Weldon 2001 Not a systematic review or a RCT Winters 1997b Heterogeneous, variable treatments. Different diagnostic groups Winters 1999a Not acute Withrington 1985 RCT lignocaine and steroid vs saline; outcomes pain and paracetamol count; no significant difference; exclude due to chronic pain Zuinen 1993 Met inclusion criteria but not relevant question (i.e. comparing diclofenac and misoprostol to diclofenac alone. Concluded equal benefit with regard to shoulder symptoms, increased gastro-intestinal adverse events with additional misoprostol ACUTE SHOULDER PAIN — Table of Excluded Studies (Cost Effectiveness) STUDY REASON Bongers 2001 Exclude because not a cost effectiveness study; a narrative review

SUMMARY TABLES Appendix E • Tables of Included and Excluded Studies ANTERIOR KNEE PAIN — Table of Included Studies (Diagnosis) STUDY TESTS PARTICIPANTS RESULTS QUALITY SUMMARY NOTES Chapman- Retrospective study (chart review) to n = 103 patients with non-specific Demonstrated that patients with non-specific knee pain N/A Not acute pain only; only those who went Jones 1998 investigate utility of MRI vs plain knee pain on for MRI included. Those who did not have radiography for non-specific knee pain. having plain xray will have a high probability of a negative MRI (regardless of whether plain film was All had plain film and then MRI and the normal or abnormal) were excluded. corresponding results were compared finding irrespective of any anatomical derangement present The paper highlights the role of MRI but use of MRI is not commonly required in non-specific knee pain ANTERIOR KNEE PAIN — Table of Excluded Studies (Diagnosis) STUDY REASON Abrahams 2001 Personal opinion only Baker 2000 Opinion only Cesarelli 2000 Requires EMG; pain duration unknown 241 Davies 2000 Orthopaedic based sample; little detail regarding subjects; not relevant to primary care setting Ferrari 1998 Not a systematic review. Clinical examination, diagnosis recommendation largely unreferenced. Useful MRI references Fouts 1999 General overview — not a systematic review or primary research; < 7% of all knee injuries involve a clinically significant fracture Evidence-based Management of Acute Musculoskeletal Pain Fulkerson 2000 Not a systematic review. No recommendation of one diagnostic strategy over another Fulkerson 2002 Orthopaedic review — not relevant in primary care. Not specifically acute pain. Not systematic or cross sectional. Opinion only Grelsamer 2000 Individual opinion Hayes 2001 Same as Petersen & Hayes 2000; chronic knee pain Irrgang 2001 Validation of tool; no data on duration Lee 2001 Mechanistic study on cadavers Livingston 1999 Aetiologic; pain duration not reported

ANTERIOR KNEE PAIN — Table of Included Studies (Prognosis) Appendix E • Tables of Included and Excluded Studies STUDY METHOD PARTICIPANTS RESULTS QUALITY SUMMARY NOTES Nimon 1998 Case series Adolescent girls, mean age 15.5 yrs Average follow-up 16 years (14–20): 73% improved; 27% Not applicable (small series; Provides information on natural history presenting to orthopaedic surgeons. same or worse. Suggests that pain will persist for many Duration of acute knee pain unknown years but eventually the majority improve without any no statement about length of symptoms but likely to be longer term significant long-term problems at diagnosis) ANTERIOR KNEE PAIN — Table of Excluded Studies (Prognosis) STUDY REASON Callaghan 2000a No long-term follow-up; no data on prognosis Kaeding 2001 Equated acute knee pain/patellofemoral pain with chondromalacia patella. General opinion based overview Lam 2001 Mechanistic study; no follow-up therefore no prognostic data; chronic pain > 6/12 Petersen 2000 Chronic knee pain — duration, mean 33.9 months (SD > 52.1) median 7.7 months (range 0.4–234) Sanchis-Alfonso 1998 Histological study of knee tissue; not relevant 242 Selfe 2001 Not relevant to these guidelines Wolfe 2002 Not relevant in primary care — specific diagnosis of osteoarthritis. Symptom duration is > 2 weeks at first visit Evidence-based Management of Acute Musculoskeletal Pain

ANTERIOR KNEE PAIN — Table of Included Studies (Interventions) Appendix E • Tables of Included and Excluded Studies STUDY METHODS PARTICIPANTS INTERVENTIONS RESULTS QUALITY SUMMARY NOTES Arroll 1997 A critical review of clinical trials 5 studies included: n = 219. Non-operative therapy was evaluated Five RCTs were located (two comparing Number of studies = 5 An old review so relevant of non-operative therapy Age range 14.8–32.3; (shoe orthoses, knee sleeves, NSAIDs, injection therapies, one comparing wedged Quality assessment = Y recent papers not included. 38% female injection therapy) vs flat shoe insoles, one comparing knee Data pooled = N Problem with lack of stan- sleeves and one comparing two NSAID Heterogeneity = Y dardised outcome measures agents) Search 1966–1995 of broader pain conditions than this guideline addresses Brosseau Cochrane Review on therapeutic N = 29 1. ice massage vs 2. ultrasoundplus ice In group 2 there was 46% (6/13) with Number of RCTs = 1 Assessed side effects and 2002 ultrasound for treating massage improved pain relief compared with Quality assessment = Y effectiveness of ultrasound for patellofemoral pain syndrome 31%(4/13); the difference was not clinically Data pooled = N/A patellofemoral pain *may be or statistically significant. One study only; Heterogeneity = N/A Type II error due to small insufficient evidence (quality and quantity) Well-designed review but only had one sample size and cannot draw for inclusion of ultrasound in evidence- study which scored poorly for conclusions. One study only so based recommendations methodological quality insufficient evidence (quality and quantity) for inclusion of ultrasound in evidence based recommendations Callaghan Pilot study to compare two types N = 14 (M = 2; F = 12) patients Participants randomised to receive either Main outcomes were related to muscle Adequate allocation concealment = U No mention of duration of 2001b of electrical stimulation of the with patellofemoral pain, sequential mixed frequency stimulation or function not pain. Both groups improved; Blinding = U pain so unclear how quadriceps in the treatment of quadriceps atrophy and simultaneous mixed frequency electrical no difference between the two. Concluded ITT analysis = N generalisable. Tertiary patellofemoral pain normal gait parameters stimulation of quadriceps that electrical stimulation of quadriceps LTFU < 15% = no follow-up referral setting so likely to be may help Bias = High more severe than general 243 Groups comparable at baseline practice setting Crossley 2001 Systematic review of physical Studies: N = 456; Taping: No strong evidence for one superior phys- RCTs = 16 Well designed review. All interventions for patellofemoral N = 240; Foot orthoses: ical intervention. There is no data for phys- Quality assessed = Y studies in the review except pain N = 20; patellofemoral ical interventions in treatment of patello- Data pooled = N one involved > 12 week dura- orthoses/braces: N = 218 femoral pain however there was a trend of Heterogeneity = Y (narrative summary tion of pain; other interven- Evidence-based Management of Acute Musculoskeletal Pain (N represents total number some treatment effect from education, of various interventions) tions involved mixed of participants, including stretching and quadriceps strengthening acute/chronic populations, control and treatment (including concentric exercises). but were predominantly groups) 2 studies for taping found insufficient chronic evidence for taping in treatment of patellofemoral pain but cannot discount effectiveness of this measure. 2 studies on patellofemoral orthoses. Reviewed studies on use of laser, acupuncture, sacroiliac joint manipulation and chiropractic patello- femoral joint mobilisation — more research is needed. Progressive resistance brace or corrective foot orthoses also looked at and more studies needed — poor quality of existing studies

ANTERIOR KNEE PAIN — Table of Included Studies (Interventions) continued Appendix E • Tables of Included and Excluded Studies STUDY METHODS PARTICIPANTS INTERVENTIONS RESULTS QUALITY SUMMARY NOTES Crossley 2002 Randomised controlled trial N = 71 with non-specific Group 1 n = 33 received physical therapy Physical therapy group demonstrated signifi- Adequate allocation concealment = Y Long-term follow-up of both groups was not possible as patellofemoral pain of 1 (quadriceps muscle retraining, cantly greater reduction in scores for Blinding = double blind placebo group offered phys- ical therapy intervention at month or longer and no other patellofemoral joint mobilisation, patella average, worst pain and disability than ITT analysis = Y completion of 6 week trial Those receiving physical specific pathologic condition. taping and daily home exercises); Group 2 placebo group after 6 weeks of physical LTFU < 15% = Y (6%) therapy were followed up at 3 months and their pain had Participants were aged 40 or n = 34 received placebo (sham ultrasound, therapy; better outcomes seen in younger Bias = Low continued to abate younger to reduce chance of light application of nontherapeutic gel, participants. Physical therapy appears Groups similar at baseline degenerative joint disease as placebo taping) effective in reducing short-term pain and cause of pain disability in patellofemoral pain D’hondt 2002 Cochrane Review on orthotic 5 trials involving n = 362 Evidence from RCTs too limited to draw Inconclusive; no special recommendation Number of RCTs = 5 Lack of standard outcome Quality assessment = Y measures makes it difficult devices for treating patellofemoral definitive conclusions about the use of knee for practice are made. Only 1 trial was Data pooled = N to compare between studies; Heterogeneity = Y no global rating of change. pain syndrome and foot orthotics for the treatment of considered high quality which compared A number of important references have yet to be patellofemoral pain. Compared McConnell’s McConnell to Couman’s and there was a included. Reviewers graded the strength of the evidence taping and modified Couman’s bandage, the trend to favour McConnell regimen but the as limited therefore unable to make firm recommendations Protonics orthosis and the Palumbo dynamic study only involved 18 subjects. Palumbo vs re foot and knee orthotics patellar brace and the Cho-pat strap Cho-pat showed no difference n = 59. Protonics vs no trt n = 100 showed all outcomes significantly favouring the Protonics orthosis. 1. home exercise vs 2. same with monitored therapy vs 3. same with McConnell and biofeedback showed that biofeedback with McConnell and home 244 exercise produced better outcomes than home exercise alone or home exercise with monitored therapy. Concentric vs eccentric exercises in 76 participants; 60 completed the study with non significant difference between the groups on favour of concentric Evidence-based Management of Acute Musculoskeletal Pain Harrison 2001 Review article on patellofemoral 12 articles reviewed Advice to consider quadriceps strength- Number of RCTs = 12 Useful for epidemiological pain syndrome — looks at ening especially in the short term. For most Quality assessment = N data aetiology, diagnosis and patients other interventions may be consid- Data pooled = N non-surgical management ered when a basic home program has been Heterogeneity = Y ineffective over the short term. Taping English language search between appears to reduce pain in some patients; 1995–2000 because of its low cost and simplicity it should be considered particularly for patients with changes in patellar orientation and mobility. There is a lack of valid clinical measures for patellar mobility and orienta- tion that make it hard to tell whether the effect of taping results from mechanical improvement of from other factors. Biofeedback, manual therapy, NSAIDs and sleeves have not been found to be more effective than simple exercises

ANTERIOR KNEE PAIN — Table of Excluded Studies (Interventions) Appendix E • Tables of Included and Excluded Studies STUDY REASON Brosseau 2002 Not relevant to anterior knee pain. Discusses ilio-tibial band syndrome Cohen 2001 Exclude due to a theoretical study on cadavers Dursun 2001 Pain not < 12 weeks. Gotlin 2000 Not a systematic overview Herrington 2000 Narrative review Mears 2001 For selected patellofemoral conditions not non-specific patellofemoral pain; chronic conditions; narrative review Neptune 2000 Not relevant; mechanistic study not involving patients in pain Powers 1998 Narrative review Rillman 2000 Exclude due to surgical treatment for patellofemoral instability Roush 2000 Study does not specifically address patellofemoral pain — see inclusion criteria. This study is not limited to the specific diagnosis — it includes individuals with a number of diagnostic entities including patellar tendonopathy, plica syndrome etc. Salsich 2002 Exclude due to not an RCT; small sample n = 10 245 Selfe 2001 Not relevant to these guidelines; a validation of outcome measurement rather than clinical assessment Suter 1999 Underlying assumptions are unproven; chronic pain Suter 2000 Basic science (construct validity); chronic pain Evidence-based Management of Acute Musculoskeletal Pain Tang 2001 Basic science — not relevant Tobin 2000 Mechanistic study in asymptomatic study; not relevant for this guideline although a good study Yeung 2001 Review of studies of primary prevention. Not relevant

Appendix E • Tables of Included and Excluded Studies 246 Evidence-based Management of Acute Musculoskeletal Pain