Cochrane in CORR®: Shoulder Replacement Surgery For Osteoarthritis And Rotator Cuff Tear Arthropathy

Importance of the Topic Shoulder pain is the third most common musculoskeletal condition that general practitioners treat, with a population prevalence estimated at 16% to 26% [11]. Osteoarthritis (OA) causes the pain in 2% to 5% of those patients [10]. While the exact incidence of glenohumeral OA is unknown, studies in cadavers suggest it is present in more than 30% of patients older than 60 years of age [7, 13], and it is expected to continue to become more common as life expectancy continues to rise; the rapid growth in the number of shoulder replacements performed supports this projection [5, 8, 9]. Demand for shoulder arthroplasty is projected to rise by more than 750% by 2030 in the United States [12]. Furthermore, projections indicate that the growth in demand for shoulder arthroplasty may exceed that of hip and knee replacements [3]. In several joint replacement registries, OA was reported as the primary diagnosis in 34% to 72% of patients and rotator cuff tear arthropathy in 4% to 21% [1], indicating lack of precision which is noticed from the wide ranges reported. Surgeons are using reverse shoulder arthroplasty more commonly, as indications for this device have expanded [4]. Despite the rise in diagnosis of shoulder OA, rotator cuff tear arthropathy, and shoulder arthroplasties performed, there is no clear consensus regarding the best type of replacement when it comes to managing these shoulder conditions, nor enough data on the harms in order to fully appreciate the risks and benefits of each type of surgery. This Cochrane review and meta-analysis evaluated all randomized controlled trials (20 studies; 1105 shoulders) of shoulder replacement surgery in adults with OA of the shoulder, including rotator cuff tear arthropathy. The Cochrane review found that total shoulder arthroplasty did not provide a clinically important advantage over hemiarthroplasty in terms of patient-reported pain, function, nor adverse effects; however, the evidence on this topic was of low quality. These findings fail to address the dilemma of what the ideal surgical options in these cases are, keeping decision making by surgeons an ongoing challenge. Upon Closer Inspection All of the included trials enrolled participants underwent some type of shoulder replacement surgery including total shoulder arthroplasty, reverse shoulder arthroplasty, hemiarthroplasty, humeral head resurfacing arthroplasty shoulder replacement surgery due to either primary glenohumeral OA, or rotator cuff tear arthropathy. We noticed considerable differences in the type of arthroplasty and surgical techniques performed across the trials, mostly in terms of methods of implant fixation and approaches to the subscapularis tendon that made pooling of study results for meta-analysis impossible. Out of the 20 studies included, none compared reverse shoulder arthroplasty to another type of shoulder arthroplasty despite the rise in reverse shoulder arthroplasty performed. Only three studies that compared total shoulder arthroplasty to stemmed humeral hemiarthroplasty for OA were eligible for pooling and meta-analysis. The pooled result of the three studies showed improvement in pain measured on visual analogue scale (VAS) 0-10, and function measured by different scales represented on Western Ontario Osteoarthritis of the Shoulder Index (WOOS Index), 0 to 100 scale in favor of conventional stemmed total shoulder arthroplasty for OA. The mean difference was -1.49 (95% confidence interval [CI] -2.88 to -0.10); minimum clinically important difference (MCID) 1.5 for pain, and mean difference 10.57, (95% CI 2.11 to 19.02); MCID 10 for function. Despite these results suggesting statistical differences in improvements in pain and function, they fell below the accepted MCID. This is important, because patients perceive effect sizes—scores for pain and function reflect this—not p values. Importantly, p values do not inform surgeons and patients about the importance of treatment effects. Effect sizes below the MCID may not be large enough for patients to care about, and regardless of p values, such small effects should not justify surgical choices; this is especially true when smaller MCIDs are chosen; some studies suggest that pain MCIDs are two points or larger [2], rather than 1.5 as was chosen here. The width of a CI for a meta-analysis depends on the precision of the individual study estimates and on the number of studies combined [6]. The wide CI in the three studies suggests the estimates here should be considered imprecise. This is a function of the few studies on the topic. More studies will help us to render those estimates more precisely, and will be important, because it’s hard for surgeons to help patients anticipate what to expect when confidence intervals bracketing the effect sizes varied by more than an order of magnitude, as was the case here. Some good news: There were 12 ongoing trials at the time of this Cochrane review was published, and so we hope that future such efforts will provide the greater precision that surgeons and their patients need. publication that could be included in future systematic reviews. One of these is comparing standard hemiarthroplasty to total shoulder arthroplasty in degenerative shoulder joints, and the rest mainly compare different surgical techniques of the same type of arthroplasty. We hope many of these will be completed so that future meta-analysts can include them. Take-home Messages This Cochrane review found that total shoulder arthroplasty did not provide a clinically important benefit over hemiarthroplasty in terms of patient-reported pain or function; however, the evidence on this topic was of low quality. As mentioned, the confidence intervals were wide, and so it’s possible that future studies will find that the statistical difference observed in favor of total shoulder arthroplasty will be clinically important. However, given the substantial difference in complexity and risk associated with implanting a total shoulder replacement, until or unless those future studies deliver in that way, the best-available (albeit relatively low quality) evidence we have suggests that hemiarthroplasty is as good in terms of improving pain and function, which is a finding that cuts against the current dominant paradigm. The durations of follow-up were insufficient to show whether total shoulder arthroplasty or hemiarthroplasty was associated with increased or decreased risk of harm. This deficiency could change conclusions in either direction. On one hand, the glenoid component is more likely to loosen than is the humeral component (a fact that favors hemiarthroplasty); by contrast, cartilage wear and increases in pain over time sometimes occur in patients with hemiarthroplasties (which might favor total shoulder replacement over time). Given the limitations of the available evidence, surgeon experience as well as individual patient expectations should play an important role in guiding the management of shoulder OA and rotator cuff tear arthropathy. Future research and high-quality randomized controlled trials are still needed to compare long-term results of reverse shoulder arthroplasty to total shoulder arthroplasty and hemiarthroplasty with consideration and focus on indications, harm, and benefit of each procedure to provide surgeons managing these conditions with critical information to form treatment decisions.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.