ARTHROSCOPIC SUBACROMIAL DECOMPRESSION: ACROMIOPLASTY VERSUS BURSECTOMY ALONE-DOES IT REALLY MATTER? A SYSTEMATIC REVIEW

Abstract

Background

Subacromial impingement is a common disorder mat in some cases results in surgical management. Arthroscopic subacromial bursectomy alone or in combination with acromioplasty are treatment options when non-operative measures fail.

Methods

A systematic review of all level-I and level-II studies regarding subacromial bursectomy and acromioplasty for impingement was performed. Medline publications were reviewed for appropriate studies.

Results

A total of six studies that met inclusion criteria were identified. However, only one randomized study was identified that directly compared the treatments in question. Additionally, only one prospective study of subacromial bursectomy was identified. A comparison of similar outcome measures revealed bursectomy alone provided similar results to bursectomy with acromioplasty.

Discussion

Limited high-level studies are available regarding arthroscopic treatment of subacromial impingement. Data available currently suggests that bursectomy alone provides similar outcomes to bursectomy with acromioplasty.

INTRODUCTION

Subacromial impingement is the most common disorder of the shoulder, accounting for 44-65% of all shoulder complaints.^1,2 The etiology of subacromial impingement is controversial, with two main theories described: A degenerative (intrinsic) theory, where symptoms are thought to result from overload on degenerating rotator cuff tendons; and a mechanical (extrinsic) theory, where symptoms are caused by compression of the rotator cuff. Conservative, nonoperative management is often successful, but when it fails, current surgical treatment is arthroscopic subacromial decompression including bursectomy and acromioplasty. The rationale for acromioplasty is based on Neer's extrinsic impingement theory, which describes irritation of the subacromial tissue by impingement of the rotator cuff under the cora-coacromial (CA) arch.^3,4 Alternatively, bursectomy alone without acromioplasty may be considered adequate by those who subscribe to the intrinsic theory, since symptoms are felt to be caused by degenerative tendinopathy and subsequent inflammation of the bursa with observed changes in the acromion felt to be secondary.^5,6 Avoiding acromioplasty is felt to be preferable by some due to concerns regarding alteration of the CA arch which could predispose to the loss of active glenohumeral elevation due to uncontrolled anterosuperior migration of the humeral head. In addition, there could be an economic benefit from bursectomy without acromioplasty as a the result of shorter procedures and decreased equipment costs. Although basic-science and clinical data exist in support of both perspectives, no evidence-based consensus is available to guide decision making.

The purpose of this systematic review is to collect and present data comparing arthroscopic acromioplasty and bursectomy alone for treatment of subacromial impingement syndrome using an evidence-based medicine approach. Our hypothesis is that there is not sufficient data that demonstratesing a clinical benefit of acromioplasty compared with bursectomy alone when performing an arthroscopic subacromial decompression.

MATERIALS AND METHODS

A literature search was performed to identify all English-language studies evaluating the management of subacromial impingement syndrome without a full-thickness rotator cuff tear with: (1) Arthoscopic debridement and bursectomy without acromioplasty; and (2) arthroscopic decompression with acromioplasty. The literature search utilized PubMed MEDLINE, Embase, and Cochrane Reviews and was performed in December 2009 (PubMed, Cochrane) and April 2010 (Embase). PubMed and Embase search terms were “subacromial decompression”, “acromioplasty”, “bursectomy”, “subacromial decompression without acromioplasty”, “arthroscopic subacromial decompression”, and “impingement”. Cochrane search terms were “shoulder” and “impingement”. Only prospective, randomized, controlled trials (Levels of Evidence I and II) with at least two years mean follow-up were included. Only one study met criteria and compared the two treatments.⁷ Level-I and level-II studies that evaluated either technique, but did not directly compare the two techniques, were included (Table 1). There were five studies that met criteria and evaluated decompression with acromioplasty. Other than the study by Henkus et al.⁷, no studies evaluating bursectomy without acromioplasty were of sufficient quality to be included. Thus, there were six studies included in the review. Each study was individually analyzed and the data collected (Table 1). Four major sources of bias were evaluated for each study when sufficient information was included in the publication: Selection, performance, attrition and detection. Minimal accepted follow-up was defined as 70% of the study population, and all six studies acheieved this.

TABLE 1.

Authors	Journal, Year	Country	Surgeons	Study Method	Population	Randomization Method	Number at Follow Up (%)	Acr (a) or Burs (b)	CA Iigement Resection	Outcome	Reoperation	Followup	Level of	Biases
Henkus et al7	JBJS Br, 2009	Netherlands		Prospective, 1 randomized	n=57, mean age 47 yo	Radmonization code	56/57 (98%)	Both, randomized	Not stated	Constant b:69.6 a: 75.8* SST: b: 9.0 a: 9.9* VAS: b:4.3 a: 3.6*	a:3/30, b:2/26	Mean 2.5 yrs 1-5)	I	Selection: no Performance: no Detection bias: no Attrition bias: no
Brox et al8	JSES, 1999	Norway		Prospective, randomized, controlled (vs. PT and 2	n=125, 45 in acromioplast y group, avg age 48 yo	Random permuted blocks	39/45 (87%)	Acromioplast y	Yes	Neer 93.0	Not indicated	2.5 yrs	I	Selection: no Performance: no Detection bias: unknown
Lindh et al9	CORR, 1993	Sweden		Prospective, randomized (vs open 1 acromioplasty)	n=20,10 in artrhroscopic group, avg age 42 yo	Not specified	10/10 (100%)	Acromioplast y	Yes	UCLA 29.0	Not indicated	2 years	I/II	Not enough information to assess
Haahr et al10	Scand J Rheumatol, 2006	Denmark		Prospective, randomized (vs 2 PT)	n=90, 45 in acromioplast y group, avg age 44 yo	Computer generated random sequence, sealed opaque envelope selected from closed bag	40/45 (89%)	Acromioplast y	Yes	PRIM Score 9.1	Not indicated	4-8 years	I	Selection: no Performance: no Detection: no Attrition biasis: no
Husby et al11	Acta Orthop Scand, 2003	Norway	Not specified	Prospective, randomized (vs open acromioplasty	n=39, 20 in acromioplast y group, avg age 42 yo	Not specified	15/20 (75%)	Acromioplast y	Yes	UCLA 32, VAS 0	Not indicated	8 years	I	Not enough information to assess
Spangehl et al12	J Shoulder Elbow Surg, 2002	Canada		Prospective randomized (vs open 3 acromioplasty)	n=62, 32 in acromioplast	Unclear	62/87 (71%)	Accromioplast y	CA ligament “divided”	UCLA 28.8, VAS 4.3	7 of 32	Mean 25 months, min 12 mo	I/II	Selection: unknown Performance: no Detection: Interview based, blinded Attrition: no

$=statistically significant difference

*=no statistically significant difference

DEMOGRAPHICS

Consistent with their randomization, the included studies reported no difference in groups in terms of gender and age, with the exception of the study by Henkus et al.⁷ In that study, which was randomized by randomization code, there were higher percentages of females and younger patients (average age of 43 verses 50) in the bursectomy group compared to the acromioplasty group. A multivariate analysis found that male gender had a positive effect on the baseline constant and SST scores, but otherwise gender and age had no statistically significant effect on outcome measures.

SURGICAL TECHNIQUE

Most studies indicated that the coracoacromial ligament was resected as part of the acromioplasty (see Table 1). The amount of acromion resected was not reported in most of the included studies. All studies used a motorized shaver for bursectomy, with or without electrocautery, and a motorized burr for acromioplasty (when performed).

OUTCOMES

A wide variety of outcome measures was used, thus preventing meaningful pooling of data. There was a lack of consistent evaluation tools, with studies utilizing the visual analog scale for pain (VAS), the Constant score, the UCLA shoulder score, the Simple Shoulder test (SST), the Neer score, and the Project on Research and Intervention in Monotonous Work (PRIM) score, among others. In the one study that directly compared bursectomy alone to acromioplasty, there was not a statistically significant difference between the VAS, SST, and Constant scores of the two groups at a mean of 2.5 years.⁷ The only other study that included one of these outcome measures which would allow further comparison between bursectomy alone and acromioplasty was published by Spangehl et al., and the VAS score for acromioplasty in that study was equal to that found by Henkus et al. (4.3) in their bursectomy-only cohort.⁷,¹²

REOPERATION

Unfortunately, only two studies reported their reoperation rates.^7,12 The reoperation rate in the bursectomy-only group was 7.7% compared to 10% in the acromioplasty group in one study, and 21% in the arthroscopic acromioplasty group in the study by Spangehl et al.

COMPLICATIONS

Of the included studies, only Spangehl et al. commented specifically on complications.¹² They reported no neurovascular complications, but 4/32 treated with arthroscopic acromioplasty had significant stiffness.

EVIDENCE-BASED EVALUATION OF BIAS

Selection bias was minimized by study inclusion criteria of randomized controlled trials, although studies by Lindh et al.⁹, Husby et al.¹¹ and Spangehl¹² et al. failed to provide sufficient information about their randomization method to allow assessment.

Performance bias was minimal, as the included studies excluded patients who underwent additional procedures, and only studies of patients undergoing SAD without rotator cuff repair were included. The procedure was fairly standardized, with some variation in regard to resection of the CA ligament, as discussed above.

Attrition bias was also minimized since follow-up was generally excellent in the included studies, ranging from 71-100% with a mean of 84% (222/264).

Detection bias was minimized by utilization of an unbiased evaluation method of results by an independent examiner blinded to treatment, or by using a validated patient-relevant outcome questionnaire. The studies by Henkus et al., Haahr et al., and Spangehl et al. used blinded, independent examiners. However, all studies used interviewer-based assessments as part of their outcomes data, which may be subject to detection bias.

DISCUSSION

In this systematic review using principals of evidence-based medicine to compare arthroscopic subacromial decompression with and without acromioplasty for management of subacromial impingement, we found a profound deficiency of data. Only one study met criteria and directly compared the two treatments. This study, by Henkus et al., found no statistically significant differences.⁷ Although this was a randomized, controlled study, one relatively small study is not adequate grounds upon which to make evidence-based management decisions. There was one additional study identified that compared the two treatments but did not meet criteria for inclusion due to lack of randomization.³³ In³³ In that study, which was performed in Brazil and published in Portuguese with the abstract in English, at minimum two- year follow- up of 18 patients they found no difference in ASES scores between those treated with or without acromioplasty.³³

Our review also included randomized controlled trials that included a group of patients who underwent arthroscopic subacromial decompression with acromioplasty. No other randomized, controlled trials were identified that included bursectomy without acromioplasty. Comparing the one cohort of bursectomy-alone to the acromioplasty groups in these other studies failed to identify any significant differences in comparable outcomes, although comparison was only possible with one other study due to different outcome measures utilized. Therefore, based on available data, no statistically sig nificant difference in VAS scores, Constant scores, or Simple Shoulder Test scores was identified for patients with subacromial impingement treated with bursectomy alone compared with patients treated with arthroscopic decompression with acromioplasty.

It is surprising that such little data exists for a problem so common (subacromial impingent) and a treatment so frequently performed (arthroscopic subacromial decompression). The benefit of acromioplasty is believed to result from removal of the extrinsic cause of impingement—prominence and spurring of the anterior one-third of the acromion, the CA ligament, and the bursa.³ However, the data supporting this “extrinsic theory” is in contrast to evidence that suggests that primary intrinsic degeneration is responsible for impingement syndrome and rotator cuff disease. ^6,13-16 Budoff et al. published their cohort of patients treated with rotator cuff debridement without acromioplasty at four years and then at nine-and-a-half years.^13,17 They found good long-term results at four years (87% good or excellent results) and at nine-and-a-half years (79% good or excellent results). They concluded that the preponderance of evidence suggests that in the majority of cases, extrinsic acromial compression is not the primary cause of impingement syndrome or rotator cuff disease. In support of this conclusion is evidence that rotator cuff pathology predates acromion pathology, rotator cuff dysfunction can lead to spurring of the acromion, and that functional outcome of subacromial decompression does not correlate with the amount of acromion resection.⁶,¹⁴,¹⁸,¹⁹,³² Perhaps the strongest sup port for an intrinsic rather than an extrinsic etiology for impingement is the success of nonoperative treatment in managing impingement syndrome. Clearly, rehabilitation and therapy cannot remove osseous structures.²⁸ The effectiveness of bursectomy alone can be explained by evidence suggesting an inflamed and thickened bursa generates pain.²³27 The nociceptive feedback relayed by free nerve ends may be responsible for the pain associated with impingement syndrome, thus explaining the success in pain relief observed when the bursa is removed.

The theoretical benefit of bursectomy without acromioplasty includes preservation of the CA arch. Destruction of the CA arch may predispose to loss of active glenohumeral elevation and anterosuperior migration of the humeral head.²⁰,²¹ The CA ligament and acromion are secondary stabilizers of the humeral head against anterosuperior migration.²²,²⁹,³⁰ An increase in translation of the humeral head has been hypothesized to predispose to rotator cuff pathology.³¹ Current arthroscopic acromioplasty technique generally recommends some preservation of the CA ligament, however, resection of spurring and bony resection inevitably lead to at least some CA ligament thinning and disruption. Thus, if surgical alteration of the acromion through acromioplasty does not result in improved outcome compared to bursectomy and debridement alone, avoiding damage to the CA arch with bursectomy alone would be preferred. In addition, there would be expected economic benefits would be expected through decreased operating time and equipment costs.

Our systematic review has several limitations. The most significant limitation is the lack of more than one comparative study meeting evidence-based medicine criteria. The cohort of patients treated with one of the procedures being compared (bursectomy alone) consisted of 26 patients. A few of the other studies in this review suffered from inadequate reporting of randomization, surgical technique and blinding, making evaluation of bias incomplete. Reported outcomes were extremely variable and mostly interview based. In addition, our search was limited to the English language, and there may be well-designed level-I studies in the non-English literature.

SUMMARY

In summary, there is a marked sparsity of level-I and -II evidence available that compares arthroscopic bursectomy alone to arthroscopic bursectomy with acromio-plasty for the treatment of subacromial impingement, despite some evidence that avoiding acromioplasty may be preferable due to preservation of the coracoacromial arch. Only one level I study comparing these two treatments was identified in the English literature, and it found no statistically significant difference. Clearly this is an area that would benefit from prospective, randomized controlled studies using validated outcomes. The ubiquity of this disease and the frequency of this surgical treatment should make these studies feasible.