Effect of arthroscopic subacromial decompression during arthroscopic repair of small to medium-sized rotator cuff tears

Abstract

Background

Although arthroscopic subacromial decompression (ASD) is a commonly used procedure during arthroscopic rotator cuff repair (ARCR), the effect of ASD on the clinical outcomes for ARCR is controversial. The purpose of this study was to compare the clinical outcomes of ARCR with or without ASD.

Methods

Patients (n = 315 with 320 shoulders) who underwent ARCR for small to medium-sized rotator cuff tears were followed for at least 24 months. ARCR was performed with ASD (180 shoulders, group A) or without ASD (140 shoulders, group N). There were no significant differences in patient demographics, including mean age and mean follow-up time. Rotator cuff repair was performed using the suture-bridge technique in all shoulders, and all patients were treated using the same rehabilitation protocol after surgery. University of California at Los Angeles score, Constant score, re-tear rates, revision surgery rates, and operating time were compared between groups. Re-tear was defined as Sugaya classification Types 4 and 5 using postoperative magnetic resonance imaging at more than 12 months.

Results

There was no statistically significant difference in clinical outcomes before and after ARCR between groups. However, the University of California at Los Angeles scores and Constant scores significantly improved in both groups after surgery (P < .001). Furthermore, there was no major difference in the re-tear rates between groups A (7/180 shoulders, 3.9%) and N (11/140 shoulders, 7.9%) (P = .146). Revision surgeries were performed on 3/180 shoulders (1.7%) in group A (due to postoperative deep infection in one shoulder and revision ARCR for re-tear in two shoulders). No revisions surgeries were needed in group N patients (P = .259). The mean surgical time for group A was 62.0 ± 27.0 minutes (29-138 min.) and 52.4 ± 26.1 minutes (17-124 min.) for group N (P = .007).

Conclusion

These results suggest that ASD has a limited effect on clinical outcomes of ARCR for small to medium-sized rotator cuff tears.

Arthroscopic subacromial decompression (ASD) is a commonly used procedure during arthroscopic rotator cuff repair (ARCR) since it was first described by Neer in 1972.¹⁷ The value of ASD is pain relief with removal of subacromial impingement and improved visualization during ARCR.⁸ Although ASD is thought to be useful in the treatment of partial or small-sized rotator cuff tears,⁸ there have been some reports on a lack of the beneficial effect.³ Beard et al reported that improvement in clinical score with persistent subacromial shoulder pain due to subacromial impingement did not differ in patients with or without ASD (placebo surgery).³

ASD generally includes resection of the coracoacromial ligament (CAL) and flattening of the inferior and posterior aspects of the acromion. The CAL has a role in maintaining the stability of the glenohumeral joint^11,16 and may prevent superior and anterosuperior translation of the humeral head in cases of failure after rotator cuff repair.⁵ Therefore, we have not performed ASD during ARCR for large to massive rotator cuff tears because re-tear rates in these patients are higher than that of small to medium-sized tears.²¹ The effect of ASD on clinical outcomes of ARCR recently has been controversial.^1,9,12, 13, 14^,23,24 However, most studies included large to massive rotator cuff tears, so the effect of ASD on the clinical outcomes of ARCR for small to medium-sized tears is still not fully known.

The purpose of this study was to compare the clinical outcomes of ARCR for small to medium-sized rotator cuff tears with or without ASD. We hypothesized that the clinical outcomes of ARCR for small to medium-sized rotator cuff tears are not different with or without ASD.

Materials and methods

Patients (n = 503 with 508 shoulders) who underwent ARCR for small to medium-sized rotator cuff tears from January 2013 to December 2019 were evaluated retrospectively. Exclusion criteria included (1) concomitant complete tears of the subscapularis tendon, (2) isolated subscapularis tendon tears, (3) follow-up less than 24 months, (4) unavailable magnetic resonance imaging (MRI) on the affected shoulder at least 12 months after surgery, and (5) previous surgery on the affected shoulder. We excluded concomitant complete tears of the subscapularis tendon because we consider that concomitant a complete tear of the supraspinatus and subscapularis tendons is classified as a large or massive tear. Concomitant complete tears of the subscapularis tendon or isolated subscapularis tendon tears were determined based on intraoperative observation and defined as a Lafosse classification Type 2 or greater tear. Among the 508 shoulders, 102 shoulders were excluded because of concomitant complete subscapularis tendon tears or isolated tears, 82 shoulders because of too short duration follow-up, and four shoulders without MRI at least 12 months after surgery. No patients had previous surgery on the affected shoulder. The final sample included 315 patients with 320 shoulders (Fig 1). We conducted multivariate analyses of patient demographics to adjust for selection bias because there were roughly 40% of patients excluded by the significant exclusion criteria. ARCR was performed with ASD (group A, n = 180 shoulders) and without ASD (group N, n = 140 shoulders). From 2013 to 2017, all cases underwent ARCR with ASD and from 2018 to 2019, all cases underwent ARCR without ASD. In group A, the subacromial bursal débridement was performed without resecting the CAL, and the bony débridement was performed until the inferior aspect of the acromion was flat. In group N, only the subacromial bursal débridement was performed without resecting the CAL. There were six surgeons, but a single supervisor guided all surgeries. This study was approved by the Institutional Review Board of Chiba University Graduate School of Medicine (approval ID number: M10131).

Figure 1.

The flow diagram of patient selection. ARCR, arthroscopic rotator cuff repair; MRI, magnetic resonance imaging.

Clinical outcomes using the University of California at Los Angeles Shoulder score and Constant score, re-tear rates, revision surgery rates, surgical time, and acromial morphology were compared between groups at the final follow-up. MRI on the affected shoulder was evaluated at least 12 months postoperatively. Re-tears were evaluated using postoperative MRI at least 12 months after surgery, and Sugaya classification Types 4 and 5 were defined as re-tears.²² The Sugaya classification classified postoperative MRI according to five types. Type 1: tendon appears to have sufficient thickness, and a homogeneously low-intensity signal. Type 2: sufficient thickness with the partial high-intensity area. Type 3: insufficient thickness, less than half that of the normal tendon, without discontinuity, suggesting a partial-thickness delaminated tear. Type 4: minor discontinuity on only one or two slices, suggesting a small full-thickness tear. Type 5: major discontinuity in more than two slices suggesting a medium or large full-thickness tear. X-rays and computed tomography (Revolution Maxima; GE Healthcare, Tokyo, Hino, Asahigaoka, Japan) were examined before surgery in all patients. Acromial morphology was evaluated according to the acromial type (AT),⁴ acromion index,¹⁹ and critical shoulder angle (CSA).¹⁵ AT was evaluated using preoperative computed tomography, and acromion index and CSA were evaluated using preoperative anterior-posterior view X-rays in a neutral position. These parameters were evaluated by two orthopedic surgeons. Measurements were made according to the agreement of both examiners. If there is a difference of opinion, it can be resolved through discussion or by referring to a third person's opinion.

Furthermore, a multivariate analysis was performed to investigate factors related to re-tear. Logistic regression analysis was performed using explanatory variables with or without ASD, older age (≥65 years), tear size, and acromial morphology (AT III or CSA >35 degrees), which were considered risk factors for re-tear in other studies.^7,9

ARCR was performed in all cases in the beach chair position under general anesthesia using a suture-bridge technique. The numbers of suture anchors during ARCR in this study were between 2 and 6 (mean 2.73 ± 1.09). All patients used an abduction sling after surgery for three weeks, started passive range of motion exercises one week after surgery, assisted active range of motion four weeks after surgery, and active range of motion five weeks after surgery with formal physical therapy.

Statistical analysis

Independent-samples Student t-tests, paired t-tests, and Fisher's exact tests were used to compare patient demographics, clinical outcomes, re-tear rates, and acromial morphology. Surgical time did not follow a normal distribution, so that was analyzed using the Mann-Whitney U test. P < .05 was considered significant in tests of statistical inference. Logistic regression analysis and multiple regression analysis were used to evaluate patient demographics between the two groups. Furthermore, a logistic regression analysis was used to evaluate a risk factor for re-tear. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan).

Results

The mean age of group A was 64.6 ± 9.5 (39-83) years and that of group N was 63.0 ± 11.0 (27-83) years (P = .276). The mean follow-up of group A was 30.6 ± 12.6 (24-88) months and that of group N was 29.0 ± 11.1 (24-86) months (P = .263). There were no statistically significant differences in sex, tear size, concomitant subscapularis tendon joint-side tear, and anchor numbers between groups (Table I). The results of multivariable analyses of patient demographics were shown in Tables II and III. The results showed no significant differences in any of the demographics between the two groups.

Table I.

Patient demographics of groups A and N.

	Group A	Group N	P value
N (shoulders)	180	140	-
Age (mean ± SD) (y)	64.6 ± 9.5 (39-83)	63.0 ± 11.0 (27-83)	.276
Follow-up (mean ± SD) (mo)	30.6 ± 12.6 (24-88)	29.0 ± 11.1 (24-86)	.263
Sex (male: female)	98:82	78:62	.910
Tear size (small: medium)	140:40	105:35	.296
SSc∗ tear (nothing: joint side†)	137:43	101:39	.441
Anchors (mean ± SD) (n)	2.70 ± 0.9	2.79 ± 1.2	.420

SD, standard deviation.

^∗Subscapularis tendon.

^†Joint side tear.

Table II.

Multivariable analyses of sex, tear size, and SSc^∗ tear between groups A and N.

	Odds ratio (95% CI)	P value
Intercept	1.24 (0.61-2.55)	.550
Sex
Female	Reference
Male	0.95 (0.61-1.49)	.824
Tear size
Small	Reference
Medium	1.06 (0.78-1.45)	.750
SSc∗ tear
Nothing	Reference
Joint side†	0.81 (0.49-1.34)	.415

CI, confidence interval.

^∗Subscapularis tendon.

^†Joint side tear.

Table III.

Multivariable analyses of age, follow-up, and the number of anchors between groups A and N.

	Lower of 95% CI	Upper of 95% CI	t value	P value
Intercept	−0.098	0.675	1.470	.142
Age	−0.001	0.010	1.578	.115
Follow-up	−0.002	0.007	.987	.324
Anchors	−0.082	0.027	1.001	.317

CI, confidence interval.

There was no statistical difference in clinical outcomes before or after ARCR between groups. Preoperative University of California at Los Angeles scores were (A/N group) 15.1/15.0 (P = .849) and postoperatively were 34.5/34.3 (P = .358). Preoperative Constant scores were (A/N group) 54.9/54.9 (P = .968) and postoperatively were 97.6/96.9 (P = .126). However, scores significantly improved in both groups after ARCR compared to those preoperatively (P < .001). Furthermore, although the re-tear rate was slightly higher in group N (11/140 shoulders (7.9%) than in group A (7/180 shoulders, 3.9%), there was no statistical difference (P = .146). According to the Sugaya classification, there were 191 shoulders in Type 1, 90 in Type 2, 21 in Type 3, 12 in Type 4, and 6 in Type 5. Revision surgery was performed on 3/180 shoulders (1.7%) in group A, due to postoperative deep infection in one shoulder and revision ARCR for re-tear in two shoulders. Although the remaining 15 shoulders were re-tears, there were no specific symptoms so those patients were only followed up. No revision surgeries were required in group N. There was no statistical difference between the two groups (P = .259) (Table IV). The mean surgical time for group A was 62.1 ± 27.0 (29-138) minutes and 52.4 ± 26.1 (17-124) minutes in group N (P = .007). Thus, the mean surgical time of group N was significantly shorter than that of group A (Fig 2). There was no statistical difference in acromial morphology between groups (Table V). The results of a multivariable analysis on the re-tears were shown in Table VI. Although medium-sized tear was a risk factor for re-tear (odds ratio 3.84 and 95% confidence interval 1.22-12.1), the presence of ASD was poorly associated with re-tear.

Table IV.

Clinical outcomes in groups A and N.

	Group A (n = 180)	Group N (n = 140)	P value
UCLA score (mean ± SD)
Pre	15.1 ± 4.1	15.0 ± 3.8	.849
Post	34.5 ± 1.2	34.3 ± 1.6	.358
Constant score (mean ± SD)
Pre	54.9 ± 14.6	54.9 ± 14.9	.968
Post	97.6 ± 3.6	96.9 ± 5.0	.126
Re-tear rates (n)	3.9% (7)	7.9% (11)	.146
Revision rates (n)	1.7% (3)∗	0% (0)	.259

UCLA, University of California at Los Angeles; SD, standard deviation.

^∗Revision ARCR in two shoulders and débridement for deep infection in one shoulder. ARCR, arthroscopic rotator cuff repair.

Figure 2.

The mean operating time of groups A and N. The mean operating time of group A was about 10 minutes longer than for group N (P = .007).

Table V.

Acromial morphology in groups A and N.

	Group A (n = 180)	Group N (n = 140)	P value
AT (Type Ⅰ:Ⅱ:Ⅲ) (n)	59:71:50	44:60:36	.757
AI (mean ± SD)	0.759 ± 0.08	0.770 ± 0.07	.314
CSA (mean ± SD) (deg)	36.8 ± 4.5	37.2 ± 4.3	.515

AT, acromial type; AI, acromion index; CSA, critical shoulder angle; SD, standard deviation.

Table VI.

Multivariable analysis on re-tears.

	Odds ratio (95% CI)	P value
Intercept	0.01 (0.001-0.102)	<.001
ASD
Without	Reference
With	0.71 (0.23-2.20)	.557
Age
<65 years	Reference
≥65 years	2.04 (0.64-6.50)	.226
Tear size
Small	Reference
Medium	3.84 (1.22-12.1)	.022
AT
Type I, II	Reference
Type III	0.79 (0.20-3.08)	.736
CSA
≤35 degrees	Reference
>35 degrees	1.25 (0.31-5.02)	.226

ASD, arthroscopic subacromial decompression; AT, acromial type; CSA, critical shoulder angle.

Discussion

This study evaluated only small to medium-sized rotator cuff tears and compared ARCR clinical outcomes with and without ASD. Furthermore, postoperative cuff integrity was evaluated in all cases using MRI at least 12 months after surgery. This study suggests that the effectiveness of ASD on clinical outcomes (clinical score) of ARCR for small to medium-sized rotator cuff tears is small. Although previous studies showed that the effect of ASD on clinical outcomes of ARCR was small, most of the studies included large to massive-sized rotator cuff tears.^1,9,13,14 Shin et al²³ compared the clinical outcomes of ARCR for small to medium-sized rotator cuff tears with or without ASD. Although their study included small numbers of patients, they randomly allocated 120 shoulders into an ARCR with ASD group and a group without ASD. They concluded there was no significant difference in clinical outcomes between these groups. Our study included 320 shoulders assigned to groups with ASD or without ASD. This larger case series had very similar results to Shin et al's study showing that the effectiveness of ASD during ARCR is small in small to medium-sized rotator cuff tears.^9,13

This study compared the mean surgical time between groups A and N and showed that the surgical time of group N was significantly shorter (about 10 minutes) than that in group A. ASD has the potential benefit of improving visualization during arthroscopy,⁸ so abbreviating the ASD procedure may result in poor visualization. As a good visualization is critical for arthroscopic surgery, abbreviating the ASD procedure has the risk of making arthroscopic surgery more difficult to perform, resulting in a longer surgical time. However, the surgical time of group N was significantly shorter (about 10 minutes) than that of group A. Therefore, the effect of ASD on visualization during ARCR is small.

In general, the CAL might be resected or damaged by the ASD procedure. The CAL maintains the stability of the glenohumeral joint^11,16 and may prevent superior and anterosuperior translation of the humeral head in cases of failure after rotator cuff repair.⁵ Lee et al performed a biomechanical study to evaluate the role of CAL in glenohumeral stability and resulted that the CAL had a role in the static stabilization of the glenohumeral joint at lower levels of abduction and release of the CAL resulted in increased anterior and inferior translation of the internally and externally rotated glenohumeral joint.¹¹ Furthermore, Rothenberg et al reported that the CAL is an integral component of the coracoacromial arch. CAL release during acromioplasty remains controversial because of its role as a glenohumeral joint stabilizer.²⁰ However, the extent of CAL damage due to ASD is unclear. Therefore, further studies are needed to investigate the effect of CAL damage due to ASD and the resulting joint instability on clinical symptoms.

Re-tear in some patients after ARCR is inevitable. The re-tear rate was 5.6% in this study, which was the same or slightly lower than that of previous studies of ARCR for small to medium-sized tears.^6,19 Although the reason for this difference is not clear, differences in repair methods and postoperative treatment might be the main causes of the difference. There was no major difference between the two groups in the re-tear rate and a multivariable analysis resulted in that larger tear size was a risk factor for re-tear but not in presence of ASD, so the influence of ASD on re-tear might be small. However, we conducted a post hoc power analysis using a 3.9% re-tear rate in group A and a 7.9% in group N to determine the sample size (a beta of 0.2 and an alpha of 0.05), and it indicated that 594 shoulders per group were necessary. Therefore, we cannot state definitively that there was no significance due to the insufficient sample size. In revision surgery, there was no case that required revision surgery in group N. Although there was no significant difference in follow-up between the two groups, we consider that further follow-up is needed in group N, which had surgery in 2018-2019, as the number of cases of reoperation may increase in the future. Furthermore, our study evaluated acromial morphology as a patient factor associated with rotator cuff tears and re-tears.^2,7,10,18,25 Acromial morphology did not significantly differ between groups, and the mean CSA of both groups was larger than that of healthy subjects.¹¹ Garcia et al reported that a large CSA was associated with re-tears.⁷ Zhang et al reviewed nine studies of ARCR combined with acromioplasty and concluded that the acromioplasty could significantly reduce CSA, notably in cases of high preoperative CSA.²⁶ Although Zhang's study suggested that ASD has the potential to reduce the re-tear rate, in the present study there was no major difference in the re-tear rate between the two groups. Therefore, the effect of ASD on preventing re-tears is considered to be small, regardless of acromial morphology.

This study showed that the effect of ASD on clinical outcomes of ARCR for small to medium-sized tears might be minor. From our result, we concluded that ARCR without ASD might be reasonable.

There were several limitations to this study. First, postoperative clinical outcomes were evaluated at various time points. However, the follow-up period in all cases was at least two years, which was long enough to evaluate outcomes. Second, our study did not evaluate other factors such as smoking, diabetes mellitus, or worker's compensation, among others, so further studies are necessary to evaluate these covariates. Furthermore, roughly 40% of patients were excluded by the significant exclusion criteria, so the effect of selection bias cannot be denied. Third, we could not use patient-reported outcomes to evaluate clinical outcomes. Fourth, we could not rule out the possibility that the temporal effect, such as group A performing ARCR from 2013 to 2017 and group N from 2018 to 2019, might have influenced the results of this study. In particular, surgical time might be affected by the surgeon's skill level. Fifth, the number of cases was insufficient to statistically conclude that there was no significant difference in the re-tear rate with or without ASD. Therefore, further study is necessary.

Conclusion

There was no statistical difference between ARCR with and without ASD in clinical outcomes and there was no major difference in re-tear rates between the two groups. Therefore, ASD is considered to have a limited effect on the clinical outcomes of ARCR for small to medium-sized rotator cuff tears.

Disclaimers

Funding: No funding was disclosed by the authors.

Conflicts of interest: The authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.