Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial
BMJ 2018; 362 doi: https://doi.org/10.1136/bmj.k2860 (Published 19 July 2018) Cite this as: BMJ 2018;362:k2860
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
We thank our colleagues for their numerous comments on our FIMPACT trial (1). We also sincerely apologise for our belated reply: while still working on the final edits of our trial, we were approached by the panels of both the Cochrane Reviews and the BMJ Rapid Recommendation. They both requested our data to be added to their systematic reviews (2, 3) and the BMJ panel also for an ensuing clinical practice guideline (4) on the use of subacromial decompression surgery. As we felt that the findings of FIMPACT trial should be considered within the context of the totality of evidence, we decided to delay our response to the rapid responses submitted on our FIMPACT trial until the two systematic reviews (2, 3) and the BMJ Rapid Rec guideline (4) was published.
The BMJ Rapid Rec guideline panel that included patients, clinicians, and methodologists produced their recommendation in adherence with standards for trustworthy guidelines and the GRADE system. The recommendation, basing on two linked systematic reviews on the benefits and harms of subacromial decompression surgery (2, 3) and the minimally important differences for patient reported outcome measures (5), summarised the evidence as follows: Surgery did not provide important improvements in pain, function, or quality of life compared with placebo surgery or other options. The panel then went on to make a strong recommendation against surgery, stating that almost all informed patients would choose to avoid surgery because there is no benefit but there are harms and it is burdensome (4).
Although we feel that the above noted captures the essence about the issue at stake, we would like to take this opportunity to highlight a few additional aspects:
A favourable long-term prognosis of patients with SAPS is a consistent finding across the number of different interventions carried out to treat this condition (e.g., Fig. 2 of the systematic review by Lahdeoja et al. (ref. 2)). We were delighted to note that one of the rapid responses also acknowledged this, stating: “At 24 months the mean VAS at rest was 9.9/100 in the arthroscopy group and 12.8/100 in the exercise therapy group. Even if subacromial decompression had produced perfect results and removed all pain it could not have produced a pain score that was 15 points (i.e., the minimal important difference) lower than that achieved either of the control groups.” To us, this is the ultimate proof that there is little point in subjecting patients with such a benign condition with a favourable overall prognosis to the risks and burden of a surgical procedure, particularly to a procedure that has never been proven effective in a randomised setting.
Berntzen et al. recommend reporting the NNT as a more intuitive and meaningful for clinicians and patients compared to mean VAS score differences. While we agree with their comment, we respectfully remind that there is no point in trying to provide confidence intervals for a comparison that shows no statistically significant difference (ASD vs. DA), as the upper boundary of the confidence interval is bound to be infinite.
Dr. Barfod and his colleagues call attention to the fact that it took us almost ten years to complete the trial despite having three high-volume centres recruiting, asserting that this could have compromised the external validity (generalisability) of our findings. We beg to disagree. The FIMPACT trial was designed as an efficacy (proof-of-concept) trial. Although the design has been elaborated in detail previously 6, we briefly remind that an efficacy trial assesses whether an intervention — here, arthroscopic subacromial decompression — can even theoretically work. This means that it will be tested under the most ideal circumstances. Accordingly, it was never our intention to recruit typical patients undergoing arthroscopic subacromial decompression, rather to find the patients most likely to have an optimal response to surgery. We refined the recruitment process with this goal in mind: To obtain a sample, based on existing scientific literature, most likely to show a positive treatment response to surgery. If the results of this kind of efficacy trial are negative – as proved to be the case here, then there is no point entering into a discussion as to whether the intervention could work in less optimal, routine settings. To truly question the generalisability of our findings, one should provide explicit evidence that questions the validity of our efficacy design.
Dr. Symonds (and other correspondents) raised concerns about how the diagnosis of impingement syndrome was confirmed. Dr. Symonds also provides an insightful elaboration on the irrationalities or even paradoxes related to the nomenclature, diagnosis, and treatment of a condition coined shoulder impingement or subacromial pain syndrome. We really have little to add to his comments, except… as also noted in our Introduction, arthroscopic subacromial decompression is the most common shoulder surgery in the world and “diagnostic criteria” similar to those used in our trial are used to justify placing hundreds of thousands of patients each year to this surgery.
Dr. Singh and colleagues provide interesting and quite provocative assertions, such as “previous published evidence and clinical experience indicates that this procedure improves quality of life of certain patients with shoulder impingement” and “the presence or absence of impingement lesion… has been shown to be associated with improvement at surgery”. Unfortunately, it is very difficult to corroborate or refute these claims, as the authors have chosen not to provide references for their assertions.
The authors also argue that our inability to demonstrate any clinically significant benefit of ASD could be attributable to us operating on patients too early or the marginal resection of subacromial bursal tissue carried out on a few of the patients randomised to diagnostic arthroscopy group. We remind our colleagues that the mean duration of symptoms in the patients of the FIMPACT trial was 18 months in the ASD and DA groups, and 22 months in the ET group. Also, in our pre-specified subgroup analysis comparing those with symptoms less than 12 months to those with symptoms longer than 12 months, no between-group difference was found between DA and ASD groups. Further, we carried out a pre-specified post-hoc analysis to assess the potential effect of bursal tissue resection on our findings. Although underpowered, the analysis did not show any statistically significant differences in the primary outcomes between those who had resection carried out versus those who did not. If anything, the observed differences suggested that bursectomy had a marginal negative effect. Finally, the authors called for long-term evidence of ASD. FIMPACT is an ongoing study and we are committed to publishing the long-term (5- and 10-year follow-up) results of this trial.
Drs. Singh and Reilly and their colleagues suggest that our trial is underpowered. As noted in the article, our point estimates exclude clinically significant treatment effects. In essence, our findings are not based on absence of evidence, as in an underpowered study, but rather on evidence of absence of a clinically significant treatment benefit.
Given the overwhelming totality of evidence that suggest at best a marginal benefit of arthroscopic subacromial decompression surgery, the burden of proof now rests on those who wish to argue that the surgery is more effective than non-surgical alternatives. To suggest that there is a subgroup of patients who possibly benefit from the surgical management, this hypothesis should be tested in a randomised (placebo-surgery) controlled study.
References
1. Paavola M, Malmivaara A, Taimela S, et al. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ 2018;362:k2860. doi: 10.1136/bmj.k2860 [published Online First: 2018/07/22]
2. Lahdeoja T, Karjalainen T, Jokihaara J, et al. Subacromial decompression surgery for adults with shoulder pain: a systematic review with meta-analysis. Br J Sports Med 2019 doi: 10.1136/bjsports-2018-100486 [published Online First: 2019/01/17]
3. Karjalainen TV, Jain NB, Page CM, et al. Subacromial decompression surgery for rotator cuff disease. Cochrane Database Syst Rev 2019;1:CD005619. doi: 10.1002/14651858.CD005619.pub3 [published Online First: 2019/02/02]
4. Vandvik PO, Lahdeoja T, Ardern C, et al. Subacromial decompression surgery for adults with shoulder pain: a clinical practice guideline. Bmj 2019;364:l294. doi: 10.1136/bmj.l294 [published Online First: 2019/02/08]
5. Hao Q, Devji T, Zeraatkar D, et al. Minimal important differences for improvement in shoulder condition patient-reported outcomes: a systematic review to inform a BMJ Rapid Recommendation. BMJ Open 2019 doi: doi:10.1136/ bmjopen-2018-028777
6. Paavola M, Malmivaara A, Taimela S, et al. Finnish Subacromial Impingement Arthroscopy Controlled Trial (FIMPACT): a protocol for a randomised trial comparing arthroscopic subacromial decompression and diagnostic arthroscopy (placebo control), with an exercise therapy control, in the treatment of shoulder impingement syndrome. BMJ Open 2017;7(5):e014087. doi: 10.1136/bmjopen-2016-014087 [published Online First: 2017/06/08]
Competing interests: No competing interests
A systematic review with meta-analysis concluded that subacromial decompression surgery provides no benefit when compared to placebo surgery or exercise, and probably carries a small risk of serious harms!
Reference
https://bjsm.bmj.com/content/early/2019/01/15/bjsports-2018-100486
Competing interests: No competing interests
Dear editors,
Recently we evaluated the study by Paavola M et al. [1] as an in-class example for critical appraisal. This paper provided comprehensive information (including appendices and protocol paper) on a trial of treatments for shoulder impingement.
The primary outcome measures were shoulder pain at rest and on arm activity at 24 months, measured with the visual analogue scale (VAS). The considered “minimal clinically important improvement” (MCII) for patient benefit was a 15-point decrease in VAS. Even more relevant to the patient is the “patient acceptable symptomatic state” (PASS); defined as VAS < 30.
Unfortunately, MCII and PASS were not presented in the main results of the paper. We found a comparison of these between surgery (ASD) and exercise therapy (ET) in Supplementary appendix 2, Table S11. However, MCII and PASS outcomes were not available for diagnostic arthroscopy.
We calculated the absolute risk reduction (ARR) and number needed to treat (NNT) from data in Table S11 for MCII in VAS at rest (ASD vs. ET) as follows:
Proportion of unwanted outcomes after ET: (68-42)/68 = 0.38 and ASD: (59-45)/59 = 0.24
ARR = 0.38 – 0.24 = 0.14
NNT = 1/0.14 ≈ 7
Using the same approach, NNT for PASS ≈ 14.
Thus, 7 patients need surgery to reach one more MCII compared to ET, and 14 to reach an additional PASS for shoulder pain at rest. This would perhaps warrant a slightly more positive interpretation of surgery effects. We recommend reporting the NNT, because this statistic is more intuitive and meaningful for clinicians and patients compared to mean VAS score differences.
We would like to encourage reporting of patient-relevant outcome measures against a do-nothing option, with NNT whenever possible, because not everyone has the skills nor the time required to evaluate all data provided. This will facilitate clinical decision-making.
Reference
1 Paavola M, Malmivaara A, Taimela S, et al. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ 2018;362:k2860. doi:10.1136/bmj.k2860
Competing interests: No competing interests
We read the results of the FIMPACT trial with great interest and must commend the authors for their excellent work (1). The study compared patients with shoulder impingement syndrome who were randomised to one of three groups: (1) active treatment of arthroscopic shoulder decompression (ASD), (2) a placebo operation of diagnostic arthroscopy (DA) or (3) exercise therapy. The trial demonstrated no statistically significant difference in clinical outcomes between patients undergoing ASD and DA. The results were very similar to the recently published Can Shoulder Arthroscopy Work (CSAW) study which also failed to demonstrate a benefit of ASD over a placebo operation (2).
One strength of the study was the use of a clinically relevant primary outcome. Many studies simply look for a statistically significant difference between placebo and active treatment irrespective of the effect size which can often be achieved if the sample size is large enough. For the FIMPACT trial to achieve its primary endpoint a minimally clinically important difference (MCID) of 15 units on a visual analogue scale (VAS) out of 100 was required ie the active treatment (ASD) had to produce an improvement that was at least 15 points greater than that achieved by the control groups (DA or exercise therapy). This is important as many patients would not wish to undertake an operation if they knew it would only improve their pain score by a few points on a scale of 0-100. However, in this particular trial the performance of the two control groups was so good that it meant an MCID of 15 at rest was impossible to achieve. At 24 months the mean VAS at rest was 9.9/100 in the arthroscopy group and 12.8/100 in the exercise therapy group. Even if subacromial decompression had produced perfect results and removed all pain it could not have produced a pain score that was 15 points lower than that achieved either of the control groups.
The FIMPACT trial had two primary endpoints. The second looked was pain VAS on arm activity. The mean score was unsurprisingly higher on activity than at rest and while this outcome was at least mathematically achievable it would have required an extremely well performing operation.
One of the surprising things about the FIMPACT trial is how well the patients did in the exercise therapy group. Patients in this group had a mean duration of symptoms of 22 months and had already undergone a trial of non-operative treatment which included physiotherapy and steroid injection prior to enrolment (3). One might assume that this group of patients had already “failed” exercise therapy and a different treatment strategy is needed yet 90% (95% CI: 81% to 98%) of patients allocated to this group were considered to be responders and 84.9% (95% CI: 79.9% to 89.8%) reported being satisfied with the treatment. Perhaps one reason this group did so well is the added motivation and supervision that comes with being part of a clinical trial and this is something we need to try and replicate in routine clinical practice.
References:
1. Paavola M, Malmivaara A, Taimela S, Kanto K, Inkinen J, Kalske J, et al. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ (Clinical research ed). 2018;362:k2860.
2. Beard DJ, Rees JL, Cook JA, Rombach I, Cooper C, Merritt N, et al. Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial. Lancet (London, England). 2018;391(10118):329-38.
3. Paavola M, Malmivaara A, Taimela S, Kanto K, Jarvinen TL, Investigators F. Finnish Subacromial Impingement Arthroscopy Controlled Trial (FIMPACT): a protocol for a randomised trial comparing arthroscopic subacromial decompression and diagnostic arthroscopy (placebo control), with an exercise therapy control, in the treatment of shoulder impingement syndrome. BMJ open. 2017;7(5):e014087.
Competing interests: No competing interests
This is a Letter to the editor concerning the article ‘Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial’, by Paavola et al [1], which we have read with great interest.
The purpose of this letter is to get clarity regarding the patients included in the FIMPACT trial, as we are concerned that they do not represent the majority of patients who are offered arthroscopic subacromial decompression (ASD) in high volume centers. By giving Paavola et al. the opportunity to provide full transparency on the sampling strategy and patient flow, we hope to allow clinicians and policy makers, on an informed basis, to decide for themselves whether the FIMPACT trial has sufficient external validity to make conclusions and claims in relation to current practice.
Firstly, we would like to congratulate the authors for completing a large and long-lasting randomized controlled trial (RCT) on a very disabling problem. The trial assesses the efficacy of ASD in a three-armed RCT: ASD, diagnostic arthroscopy (placebo control), and exercise therapy. The authors conclude that: “The findings do not support the current practice of performing subacromial decompression in patients with shoulder impingement syndrome”.
As a high-volume shoulder arthroscopy center, we commend the large effort related to the FIMPACT trial, and believe that such initiatives are the only way to bring the orthopedic profession forward. We have focused intensively on the shoulder impingement syndrome (SIS) the past 5 years, and we know the challenges and difficulties of running RCT studies on this patient group [2–5]. Based on our similar experience with running large SIS cohort studies and RCT’s, we have a query regarding the selection of patients in the FIMPACT trial.
In order to better understand the generalisability of the findings, we would like the authors to describe the total number of patients with subacromial shoulder pain referred to the three recruiting clinics in the study period. In the protocol paper [6] it is described that all patients complaining of subacromial shoulder pain, at any of the three participating clinics, would be assessed for eligibility. Over an eight year period only 281 patients were assessed [1]. This implies that each clinic only had 12 patients with subacromial shoulder pain referred per year. As the FIMPACT trial only included high volume surgeons with +500 shoulder arthroscopies as investigators, it seems quite contradictory that such an experience would be present if only 12 patients on average were referred on a yearly basis. For comparison, our clinic in Copenhagen, Denmark, which we believe is similar in size, standards and uptake area to the clinics in Helsinki, Finland, had 400-600 referrals of subacromial shoulder pain per year and performed 100-150 ASD’s per year in the trial period. Based on the size of the uptake area, and the fact that our clinics seem comparable, it implies that less than 5% of patients referred with subacromial shoulder pain were accounted for in the FIMPACT trial.
Therefore, we urge the authors of the FIMPACT trial to provide full transparency on the number of patients referred with subacromial shoulder pain in the study period (meaning all patients that, according to the protocol paper, should have been assessed for eligibility), and ideally provide an explanation why they were not included in the patient flow or mentioned as a limitation in the paper. A typical challenge we have experienced in our center when running consecutive sampling trials, such as the FIMPACT trial, is that the clinicians forget, or do not have the time to screen or recruit patients to ongoing trials. This has lead us to provide a better daily support for the clinicians in an ongoing shoulder impingement trial at our center [5] to make sure all eligible patients are accounted for. In a present study [5] we have managed to include 200 patients with SIS over a 2 year period from our center alone, again highlighting our main point, that the majority of patients with subacromial pain must have been missed in the three recruiting clinics in the FIMPACT trial. Further indication of this is seen in the first FIMPACT Clinical trials registration from January 2007. This documents the FIMPACT trial starting date to be October 2005, and the expected completion date to be December 2009, suggesting that the expected recruitment period, which we can only assume must be based upon the patient flow in the three clinics, was large enough to finish in 4 years.
In the peer review process, available at BMJ’s homepage, the authors argue that the protracted inclusion period was due to very stringent eligibility criteria made to include the patients that were most likely to benefit from surgery. The authors argue having performed an efficacy trial and thereby increasing the generalizability of the negative finding. This, however, requires that the selective recruiting is based on relevant criteria such as the mentioned exclusion criteria. If the selective recruiting occurs before patients are assessed for eligibility, it will negatively affect the generalizability of the findings.
We hope this letter will open for insight on the sampling strategy and patient flow of the FIMPACT trial, as such vital information should be transparently reported by Paavola et al. before conclusions are drawn and policy implications related to current practice can be made.
References
1 Paavola M, Malmivaara A, Taimela S, et al. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. Bmj 2018;:k2860. doi:10.1136/bmj.k2860
2 Clausen MB, Witten A, Holm K, et al. Glenohumeral and scapulothoracic strength impairments exists in patients with subacromial impingement, but these are not reflected in the shoulder pain and disability index. BMC Musculoskeletal Disorders 2017;18. doi:10.1186/s12891-017-1667-1
3 Clausen MB, Merrild MB, Witten A, et al. Conservative treatment for patients with subacromial impingement: Changes in clinical core outcomes and their relation to specific rehabilitation parameters. PeerJ 2018;6:e4400. doi:10.7717/peerj.4400
4 Witten A, Clausen MB, Thorborg K, et al. Patients who are candidates for subacromial decompression have more pronounced range of motion deficits, but do not differ in self-reported shoulder function, strength or pain compared to non-candidates. Knee Surg Sports Traumatol Arthrosc Published Online First: 17 March 2018. doi:10.1007/s00167-018-4894-6
5 Clausen MB, Bandholm T, Rathleff MS, et al. The Strengthening Exercises in Shoulder Impingement trial (The SExSI-trial) investigating the effectiveness of a simple add-on shoulder strengthening exercise programme in patients with long-lasting subacromial impingement syndrome: Study protocol for a p. Trials 2018;19:1–17. doi:10.1186/s13063-018-2509-7
6 Paavola M, Malmivaara A, Taimela S, et al. Finnish Subacromial Impingement Arthroscopy Controlled Trial (FIMPACT): A protocol for a randomised trial comparing arthroscopic subacromial decompression and diagnostic arthroscopy (placebo control), with an exercise therapy control, in the treatment of . BMJ Open 2017;7:1–14. doi:10.1136/bmjopen-2016-014087
Competing interests: All authors work at the Department of Orthopedic Surgery, at the public Amager-Hvidovre Hospital in Copenhagen. KWB and PH are orthopedic surgeons specialized in arthroscopy, AW is a resident in orthopedic surgery, and MBC and KT are physiotherapists specialised in the primary treatment of patients with SIS. None of the authors have competing financial interest related to the treatment of patients at the hospital.
This paper is another in the long list of trials in disorders of the musculoskeletal system where no significant difference was found between treatments. The reason for this is hardly surprising: no diagnosis was made.
Confusion is evident in the very phrase ‘shoulder impingement syndrome’. A syndrome is a group of symptoms or signs which consistently occur together. But in this paper all we are told of the clinical features is that ‘the pathognomonic clinical sign…is subacromial shoulder pain while lifting the arm’.
Here we immediately have a difficulty. Since the acromion is only found in the shoulder region, the word ‘shoulder’ is redundant. So we have ‘subacromial pain on lifting the arm’. But where is subacromial pain felt? There is no mention of where the patients indicated the location of their pain. As for ‘pain on lifting the arm’, virtually all disorders of the shoulder region will produce this symptom.
Or did all the ten authors of this paper intend to refer to pain arising from one or more of the structures in the subacromial region? In this case, which one(s)? The subacromial bursa, or those components of the rotator cuff situated below the acromion (the supraspinatus and infraspinatus tendons), or the glenohumeral joint?
Further, we are told that in this syndrome the entirely non-discriminatory symptom of pain on lifting the arm is only ‘attributed’ to impingement of the rotator cuff. There is further speculation that the under surface of the acromion somehow has become roughened and is in need of being surgically ‘smoothened’.
Even if this theory were true, since the patients were almost certainly suffering from a heterogeneous group of disorders, the trial would not be expected to produce meaningful results.
Competing interests: No competing interests
In the BMJ (2018; 362:k2860), Paavola et al. reported a randomised, placebo surgery controlled clinical trial involving patients with shoulder impingement syndrome (SIS), and concluded that arthroscopic subacromial decompression (ASD) provided no benefit over diagnostic arthroscopy (DA)1.
Multi-centre randomised controlled trials (RCTs) are commonly considered as the highest level of medical evidence. Influenced by their findings, the orthopaedic community will propose to limit the surgical indications, and thus significantly reduce the medical burden, economic pressure and medical injuries. Because these findings will have a great influence on clinical decision-making, great care needs to be taken in using their conclusions.
In fact, their conclusions are based on an important hidden premise. Both the ASD and DA groups underwent very good postoperative rehabilitation and performed home exercises. Previously, many researchers have found that a specific exercise strategy can significantly reduce clinical symptoms and the need for surgery 2 3. In this study, the authors gave a secondary comparison (ASD versus exercise therapy) and found statistically significant differences. However, exercise therapy also improved the visual analogue scale (VAS) scores, while the P values (P = 0.023 and P = 0.008) are in fact not particularly striking. If the authors had provided the missing third comparison (DA versus exercise therapy), they may have found that there was no difference between DA and exercise therapy, and then realized that good postoperative rehabilitation and home exercises may have had a very big impact. The data presented cannot rule out whether the postoperative rehabilitation and exercise program may have caused the lack of a significant difference between the ASD and DA groups.
As a consequence, their conclusions may only apply in some developed countries, in which there is very good provision for rehabilitation and patient compliance is also good, but not for the whole world. Doctors in different countries, especially in countries where it is difficult to provide suitable rehabilitation conditions or to ensure patient compliance in the short term, should make decisions on the basis of specific national circumstances, instead of blindly following the results of studies performed under different conditions.
References:
1. Paavola M, Malmivaara A, Taimela S, et al. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ (Clinical research ed) 2018;362:k2860. doi: 10.1136/bmj.k2860 [published Online First: 2018/07/22]
2. Hallgren HC, Holmgren T, Oberg B, et al. A specific exercise strategy reduced the need for surgery in subacromial pain patients. British journal of sports medicine 2014;48(19):1431-6. doi: 10.1136/bjsports-2013-093233 [published Online First: 2014/06/28]
3. Holmgren T, Bjornsson Hallgren H, Oberg B, et al. Effect of specific exercise strategy on need for surgery in patients with subacromial impingement syndrome: randomised controlled study. BMJ (Clinical research ed) 2012;344:e787. doi: 10.1136/bmj.e787 [published Online First: 2012/02/22]
Competing interests: No competing interests
There has been a significant increase in the number of subacromial decompressions and the use of this procedure needs to be rationalised FIMPACT trial was designed to address this issue. The authors concluded that shoulder decompression surgery might not provide clinically significant benefit over non-operative treatment, and that there is no benefit of decompression over arthroscopy only. Based on the conclusions, many may recommend that this procedure should no longer be offered routinely to patients, but previous published evidence and clinical experience indicates that this procedure improves quality of life of certain patients with shoulder impingement. The study design of this trial has selection bias and the results are valid only in the trial population that are not generalisable to all conditions included under the diagnostic umbrella term of shoulder subacromial pain with impingement syndrome. We believe that the limitations of any research need to be considered before knee jerk reactions and unreasonable extrapolations.
Randomised controlled trials (RCTs) are generally considered the highest level of evidence to inform clinical practice. There are increasing numbers being performed and the orthopaedic surgeons have been engaging actively in these multicentre clinical trials. The findings of these trials get widely disseminated and are considered by the orthopaedic community to inform change in their clinical practice. Therefore it becomes the responsibility of the researchers to acknowledge that the study design could contain potential biases and the results are valid only in the study population and may not be generalisable to all conditions included under a diagnostic umbrella term such as shoulder subacromial pain and impingement syndrome.
The FIMPACT (Finnish Subacromial Impingement Arthroscopy Controlled Trial) was a multi-centre, randomised, pragmatic, parallel group, placebo-controlled trials comparing surgery with non-operative or no treatment for shoulder impingement syndrome. The authors concluded that arthroscopic shoulder decompression (ASAD) surgery might not provide clinically significant benefit over non-operative treatment, and that there is no benefit of decompression over arthroscopy only. The authors need to be commended on a well-presented study and on completing difficult RCT including placebo arm. There has been significant increase in number of arthroscopic subacromial decompressions (ASAD) in the last decade. The use of this procedure needs to be rationalised and for this it is necessary to clarify the evidence to identify the patients who benefit from surgery rather than withdrawing surgery for all patients with subacromial shoulder pain. The limitations of any research need to be considered before knee jerk reactions and unreasonable extrapolations where it may be recommended that this procedure should no longer be offered to any patient routinely.
The major strength of this study is the use of a randomised placebo-controlled design with three groups (including both placebo and nonoperative treatment arm), multiple follow-up assessments, and the use of valid patient-reported outcome measures. The researchers went to great length to ensure masking of the assessors and patients to the specific surgical intervention and ensuring that the placebo surgery was conducted in the same manner as decompression. The procedure involved inspection and irrigation of the glenohumeral joint (arthroscopy) and the subacromial bursa (bursoscopy). At first reading it appears very convincing that surgery has no benefit till the articles are read closely when one realises that it is not the case. There are certain limitations of the trial that need to be considered.
The inclusion criterion was subacromial pain for 3 months with intact rotator cuff tendons, patients who had previously completed a non-operative management programme that included exercise therapy and at least one steroid injection. The authors excluded patients with a full-thickness tear of rotator cuff or any other shoulder pathology. It needs to be highlighted that according to previous published research and clinical experience a large subgroup of patients with shoulder subacromial pain settle over six months to one year of onset of symptoms without intervention. Therefore, a period of three months is a short time for assessment, diagnosis and adequate physiotherapy treatment for a patient to be offered surgical treatment.
The diagnosis of subacromial pain with impingement syndrome encompasses a myriad of pathologies that may include rotator cuff disease, inflammatory shoulder pathology, frozen shoulder and acromioclavicular joint (ACJ) pathologies. Both the studies did not stipulate clear specific diagnostic criterion for impingement syndrome and it was left to the surgeons to decide which patient had impingement syndrome. This adds variability to the conditions included in the trial and calls into question how good clinicians are at differentiating the causes of shoulder subacromial pain. It needs to be critically assessed how the diagnosis of impingement syndrome is confirmed and what is included under this diagnostic umbrella. Studies investigating shoulder impingement syndrome have shown inconsistencies in participant selection criteria between health disciplines and there is a need for harmonization of the selection criteria in the form of an international editorial consensus.
This trial only considered patients who were having arthroscopic decompression surgery for impingement syndrome and did not include patients who were having decompression as part of lateral clavicle excision, or as part of a debridement of the cuff, or as part of a cuff repair or other intra-articular pathology. It would not be reasonable to extend the results of these trials to all patients with impingement. The patients with symptoms related to the lateral end of clavicle benefit from surgery if conservative measures have failed.
The actual number of patients that underwent treatment as per the protocol and randomisation is below the number of patients required as per sample size calculations and there was significant crossover of patients in treatment arms. This has been a major issue in most of the previous trials published in orthopaedics with a high-crossover rate or a high rate of non-adherence to treatments like the SPORT trial for lumbar spine and those for knee arthroscopy. The authors present sensitivity analysis to address this issue where the findings of the trial were consistent when analysed both as randomised and per protocol. However, the per protocol analyses suffer from indication bias and attrition bias as would be seen in a cohort study and thus the authors lose the benefits of randomisation. As the size of the study groups was lower than planned, it may involve a substantial risk for a type II error (beta error).
There was a benefit of surgery -either shoulder arthroscopy alone or shoulder scope with the ASAD- over nonoperative management although it may not have reached clinical significance. Patients who had surgery were better on primary and secondary outcomes in surgical groups compared to non-operative treatment. The mechanism by which the surgery works, however, needs further research. It can be argued whether the arthroscopy only group is truly a placebo group as surgeons performed a bursoscopy to visualise the subacromial space and would have performed a partial bursal debridement to visualise and confirm that the rotator cuff was intact especially in cases of partial rotator cuff tears. Previous studies support the theory that the subacromial pain due to rotator cuff tendinosis might be mainly caused by inflammation of the bursa. In the patients who underwent only arthroscopy, a significant amount of fluid flowed through the glenohumeral joint and subacromial space to allow adequate visualisation of the rotator cuff and control bleeding that could have improved the joint inflammation. In addition, surgeons were not asked to document the presence or absence of impingement lesion: significant scuffing of the under surface of the acromion and the CA ligament and the presence of kissing lesion of the tendon; the presence of which has been shown to be associated with improvement at surgery.
The long-term effects of this procedure could not be compared with non-operative management. Recently published randomised controlled trial has shown that after a minimum of 10-year follow-up, the surgical treatment of subacromial impingement syndrome appears to render better clinical results than physical therapy alone.
The natural course of shoulder impingement syndrome in patients without rotator cuff tears is unclear. Both patient groups (surgical and nonsurgical groups) do improve over one year of the onset of symptoms. Whilst treatment can have a cumulative effect with one physiotherapy course followed by another course, there remains a certain group of patients whose symptoms would be no better or worse and it would be reasonable to offer these patients surgery if they have failed to show improvement for more than six months to one year. There are certain variables that are shown to be associated with good outcome after subacromial decompression surgery: shoulder pain with overhead activities, persistent symptoms for more than six months, symptoms persistent despite a 3-month course of supervised physiotherapy, a consistently positive Hawkins test, radiological changes of impingement on both acromion and humerus in the subacromial region and improvement for more than one week following a steroid injection. If four or more of these factors are present, patients have better outcome with surgery. Further research is necessary to identify subgroup of these patients who continue to have long term symptoms despite conservative management and could benefit from surgical management. It also needs to be noted that if patients are not treated, some patients develop chronic disease over the long term and progress to rotator cuff tears. We suggest that patients with impingement syndrome (an umbrella diagnostic term) without rotator cuff tears should be offered full conservative management with physiotherapy and/or steroid injection and observed for improvement over six months to one year. If these patients continue to have symptoms and they fulfil certain criterion they could benefit from surgery and should be offered the opportunity.
Competing interests: No competing interests
This published study is very much of its time: a pragmatic randomised trial that shows no benefit from surgery. Unfortunately, the nuances of decision making in surgical practice are removed from the design of such trials and their findings represent questionable evidence that policy makers can use to ration treatment. There are significant limitations to the design of this study that may not be obvious to the non-specialist reader.
The inclusion of patients with at least 3 months of impingement symptoms is problematic. The natural history of the condition without any intervention is symptom resolution seen in the majority of patients after six months. Furthermore, routine management of symptoms before consideration of surgical intervention would usually include image guided injections and a period of rehabilitation. In practice surgery is rarely performed before six months.
It is not clear from the methodology whether routine imaging in this study included a 30 degree caudal tilt radiograph to assess enthesophyte formation as a driver for extrinsic compression, which has been suggested as a predictor of outcome from mechanical decompression.
The placebo group is a misnomer: intervention included entry into the subacromial bursa; pumping saline into the bursa which may well in isolation have some therapeutic benefit; and 15 placebo patients had ‘minimal bursa resection’, the extent of which is not described. Three more had ‘extensive’ resections. None of these interventions represent a placebo surgical procedure.
The trend demonstrated in this study appears to show that arthroscopic subacromial decompression (ASD) is superior to arthroscopy, which is in turn superior to exercise in modifying Visual Analogue score for pain. These differences were not statistically significant, yet by the authors own admission the study is underpowered. Concluding that there is no difference in the groups, when the study is undersized by 14 patients introduces the risk of type II error, and this weakens the authors’ assertion that ‘arthroscopic subacromial decompression provides no clinically relevant benefit over diagnostic arthroscopy in patients with shoulder impingement syndrome.’
ASD may well be an over performed operation, however pragmatic RCTs such as this should not automatically be given the platform of a BMJ publication simply because they represent level I evidence. Even when ignoring the fact that this paper was underpowered, the inclusion criteria and management approach do not reflect patient selection and the myriad of factors that inform the decision to treat patients in whom non-operative treatment has failed.
The BMJ and its reviewers have a responsibility to ensure that conclusions drawn do not mislead readers: their patients deserve better!
Competing interests: No competing interests
Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: a 5-year follow-up of a randomised, placebo surgery controlled clinical trial
Dear Editors
I would like to flag an update of this study at 5 year followup “ Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: a 5-year follow-up of a randomised, placebo surgery controlled clinical trial ” (Ref 1) by the FIMPACT group.
Thank you
Reference
1. https://bjsm.bmj.com/content/early/2020/10/04/bjsports-2020-102216.long
Competing interests: No competing interests