Surgery for shoulder osteoarthritis
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To determine the effectiveness and safety of surgery for shoulder osteoarthritis (OA).
Background
Description of the condition
Shoulder osteoarthritis (OA) is characterized by narrowing of the glenohumeral (shoulder joint). It presents primarily with shoulder pain, limitation of shoulder function and disability. Shoulder pain is common, affecting 5% to 21% adults in U.S. and Western countries Bjelle 1989; Miller 1993; Bergenudd 1988; Chakravarty 1990; Chard 1991; Breivik 2006). Shoulder OA is the underlying cause of shoulder pain in 2% to 5% of this group (Meislin 2005), although few truly population‐based studies have been done. Other common causes of shoulder pain include rotator cuff tendinitis and adhesive capsulitis. Shoulder pain is associated with shoulder‐related disability in more than half of patients reporting pain (Chard 1991; Croft 1996; Pope 1997), as well as increased health care utilization (Wofford 1997) and disability claims (Nygren 1995). Thus, shoulder osteoarthritis leads to significant morbidity, especially in the aging population.
Description of the intervention
Current non‐surgical treatment options for chronic shoulder pain associated with shoulder OA include oral analgesic and non‐steroidal anti‐inflammatory drugs (NSAIDs), intra‐articular injections (corticosteroids and hyaluronic acid), and physical therapy. NSAIDs may help to alleviate the pain, but may cause systemic side effects, including renal insufficiency and gastrointestinal problems, especially in the elderly (ACR 2000; Shamoon 2000). Intra‐articular corticosteroid injections and physiotherapy may provide benefits as they do for other shoulder conditions (Green 2003; van der Windt 2003), but their benefits in shoulder OA have not been proven. Intra‐articular hyaluronic acid injections have been recently investigated for treatment of persistent shoulder pain including, but not limited to, patients with shoulder OA (Blaine 2008).
The surgical treatment options for shoulder OA, which are the focus of this review, include total shoulder arthroplasty (Fenlin 1998); hemiarthroplasty ) (Smith 1998); arthroscopic debridement (Bishop 2003); interpositional arthroplasty, i.e., putting in material between two ends of the bone; and cartilage repair/implant. These procedures are usually performed for patients who have failed conservative management and have disabling pain and limitation of shoulder function. Total shoulder arthroplasty is the surgical replacement of both glenoid and the humeral head with implants. Shoulder hemiarthroplasty is the surgical replacement of the humeral head with an implant. Arthroscopic debridement of the shoulder joint involves introduction of an arthroscope into the shoulder joint for joint debridement and synovectomy, i.e., removal of synovium (joint lining).
How the intervention might work
The removal of damaged bone, cartilage, synovium (the tissue lining of the joint) and replacement with prosthetic components (artificial man‐made joint parts) is associated with improved biomechanics, decreased inflammation and removal of the underlying disease.
Why it is important to do this review
Shoulder pain due to OA is a painful, disabling condition. Surgical treatment of shoulder OA has been reported to be associated with a significant improvement in pain, function, and quality of life (Fehringer 2002). Therefore, a synthesis of data is needed to assess the effectiveness and safety of surgical approaches as compared to each other, to placebo, or to other conservative options.
Objectives
To determine the effectiveness and safety of surgery for shoulder osteoarthritis (OA).
Methods
Criteria for considering studies for this review
Types of studies
We will consider all randomized or quasi‐randomized (methods of allocating participants to a treatment which are not strictly random, e.g., date of birth, hospital record number or alternation) clinical trials (RCTs) for inclusion. There will be no language restrictions on included studies and we will translate non‐English articles. We will not include any abstracts in this review.
Types of participants
We will include studies of adults with osteoarthritis of the shoulder joint, confirmed by radiographic or clinical examination. We will exclude studies of adults undergoing surgery for benign or malignant tumors, adhesive capsulitis, shoulder instability or fractures. We include patients with OA secondary to rheumatoid arthritis (RA).
Types of interventions
We will include surgical techniques (total shoulder arthroplasty, hemiarthroplasty, arthroscopy with debridement and others such as Interpositional arthroplasty, cartilage repair/grafting) versus placebo or sham surgery, non‐surgical modalities (e.g. intra‐articular corticosteroid injections, physical therapy, acupuncture, etc.), no treatment, or comparison of one type of surgical technique to another.
Types of outcome measures
Primary outcomes
We will examine all outcomes at the latest follow up in each study.
The co‐primary efficacy outcomes include the following.
-
Pain, on visual analog scale (VAS), numeric rating scale (NRS) or semi‐quantitative descriptive scales or other instruments
-
Disability/function measured using shoulder‐specific instruments (e.g. Constant score, University of California at Los Angeles Shoulder Scale (UCLA), American Shoulder and Elbow Surgeons Shoulder Score (ASES), Neer rating, disability of the arm, shoulder and hand (DASH) questionnaire etc.)
Primary safety outcomes include the following.
-
Adverse effects, total, serious and specific (local such as shoulder stiffness, instability, infection, nerve damage and systemic such as thromboembolism, other pulmonary, cardiac, gastrointestinal and others); number of withdrawals and deaths
Secondary outcomes
Secondary outcomes will include the following.
-
Quality of life, assessed by generic instruments such as Short‐Form 36 (SF‐36) and other similar instruments
-
Patient‐ or physician‐evaluated success of treatment, including patient satisfaction
-
Function as assessed by range of motion (active and passive), strength, recurrence of symptoms, return to work and sport
-
Revision/re‐operation
Search methods for identification of studies
Electronic searches
We will search the following databases for randomized or quasi‐randomized trials.
1. The Cochrane Central Register of Controlled Trials (CENTRAL), via The Cochrane Library, Wiley InterScience (www.thecochranelibrary.com), current issue;
2. OVID MEDLINE, 1966‐present;
3. CINAHL (via EBSCOHost), 1982‐present;
4. OVID SPORTdiscus, 1949‐present;
5. EMBASE 1980‐present;
6. Science Citation Index (Web of Science) 1945‐present.
We have provided the complete search strategies for the database searches in Appendix 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6. We will not apply any date or language restrictions.
Data collection and analysis
Selection of studies
Following identification of potential trials for inclusion, two review authors (JS, DO) will independently review the methods sections of all identified trials according to predetermined criteria (see 'Selection criteria'). We will resolve any disagreements by consensus.
Data extraction and management
Two review authors (JS, DO) will independently extract data from the included trials including source of funding, study population, number of centers, number and experience of surgeons in each trial, duration of operation, intervention, analyses and outcomes using standardized data extraction forms. We will also assess whether trials meet the recently described modified CONSORT criteria for non‐pharmacological treatment (Boutron 2008). These include, among others, description of eligibility criteria for centers and surgeons; details of both the experimental treatment and comparator; description of different components of the interventions and when applicable; and descriptions of the procedure for tailoring the interventions to individual participants. We will contact the authors of original studies to obtain more information if needed.
In order to assess efficacy, we will extract raw data for outcomes of interest (means and standard deviations for continuous outcomes and number of events for dichotomous outcomes) where available in the published reports. When possible, we will extract numbers based on intention‐to‐treat analysis.
Assessment of risk of bias in included studies
Two review authors (JS, DO) will independently assess the risk of bias of each included trial. We will resolve disagreements by consensus and, if needed, referral to the third review author (RB).
For each included study, we will assess risk of bias against key criteria: random sequence generation; allocation concealment; blinding of participants, personnel and outcomes; incomplete outcome data; selective outcome reporting; and other sources of bias, in accordance with methods recommended by the Cochrane Collaboration (Higgins 2008). Each of these criteria will be explicitly judged using: Yes (low risk of bias); No (high risk of bias); or unclear (either lack of information or uncertainty over the potential for bias).
Measures of treatment effect
We will plot the results of each RCT as point estimates: i.e., mean and standard deviation for continuous outcomes and risk ratios (RRs) with corresponding 95% confidence interval for dichotomous outcomes. We will calculate mean differences if possible for continuous measures, as these results are easier for clinicians/readers to interpret. If individual outcome measures vary but the construct being measured is the same (i.e. use of different scales across trials and/or inability to convert data into the same scale), then we will use standardized mean differences. To expedite rapid and easier updating of the review, we will extract all results that can be extracted from the included trials. When the results cannot be shown in this way, for example if reported as median scores only, we will describe them in the 'Characteristics of included studies' table.
Unit of analysis issues
The patient will be the unit of analysis. If analyses are presented only by joints (we anticipate that some RCTs may only provide this), we will perform additional analysis considering the joint as the unit of analysis. In case eligible data are presented only by joints (unit of analysis error), we will record the data as from a cluster‐randomized trial.
Dealing with missing data
We will present this in the 'Characteristics of included studies' table and summarization of the risk of bias tables. We will not undertake any imputation.
Assessment of heterogeneity
We will first assess studies for clinical homogeneity with respect to the duration of the disorder, control group, and outcomes. For studies judged to be clinically heterogeneous, we plan to describe them separately and not combine them in a meta‐analysis. For studies judged as clinically homogeneous, we plan to test statistical heterogeneity using the I2 statistic. For interpreting I2, we will use the following method, which is in accordance withthe Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2008): 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90%, substantial heterogeneity; 75% to 100%, considerable heterogeneity. We will use a random‐effects model as the default option, independent of the I2. However, for sensitivity we plan to discuss the results based on a fixed‐effect analysis as well.
Assessment of reporting biases
We will assess publication bias using a funnel plot, if 10 or more studies are available for meta‐analysis. A funnel plot is a scatter plot with sample size along the y‐axis and the treatment effect along the x‐axis. An asymmetric appearance of this plot in either axis indicates a potential level of publication bias in the literature.
Data synthesis
For clinically homogeneous studies, we will pool outcomes in a meta‐analysis using the random‐effects model as a default. We will also examine the results from the fixed‐effect model as sensitivity analyses, to evaluate any bias from smaller studies. We will calculate the control (placebo) event rate (expressed as a percentage), which is the sum of all events in the placebo group divided by the total patient numbers in the placebo group. We will calculate the number needed to treat to benefit (NNTB) and number needed to treat for harm (NNTH) as one divided by the absolute risk difference.
For continuous outcomes (for example, pain measured on a visual analogue scale), we will calculate absolute change (benefit) from the mean difference. We will calculate relative difference in the change from baseline as the absolute benefit divided by the baseline mean of the control (placebo) group.
Subgroup analysis and investigation of heterogeneity
We plan the following sub‐group analyses, where data are available, to determine if outcomes differ for the following factors.
1. Age of the patient.
2. Presence or absence of significant rotator cuff disease.
Sensitivity analysis
If there are sufficient data, we will conduct a sensitivity analysis to assess for any bias attributed to allocation concealment, as described in section "Assessment of risk of bias" above. We will also perform an analysis of all studies, including the quasi‐randomized studies versus only RCTs, if sufficient data exist.
Summary of findings tables
We will present the main results of the review in summary of findings tables, which provide key information concerning the quality of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the main outcomes, as recommended by The Cochrane Collaboration (Schunemann 2008a). We plan to include the following main outcomes in the summary of findings tables: pain; disability/function; adverse events ‐ total; number of withdrawals due to adverse events; adverse events ‐ serious (as defined in 'Types of outcome measures', above); revision rate.
In addition to the absolute and relative magnitude of effect provided in the summary of findings table, we will calculate the number needed to treat (NNT) from the control group event rate (unless the population event rate is known) and the risk ratio using the Visual Rx NNT calculator (Cates 2004). For continuous outcomes, we will calculate the NNT using the Wells calculator software available at the CMSG editorial office. We will determine the minimal clinically important difference (MCID) for each outcome for input into the calculator. The summary of findings table includes an overall grading of the evidence related to each of the main outcomes, using the GRADE approach (Schunemann 2008b).
