Clinical Outcomes of Arthroscopic Superior Capsular Reconstruction Using Fascia Lata Autograft Versus Reverse Shoulder Arthroplasty in Patients 65 Years and Older With Irreparable Rotator Cuff Tears: A Retrospective Cohort Study

Abstract

Background:

Arthroscopic superior capsular reconstruction (ASCR) and reverse shoulder arthroplasty (RSA) have both shown favorable outcomes in patients with irreparable rotator cuff tears (IRCTs).

Purpose:

To (1) compare the clinical outcomes of ASCR versus RSA in patients aged ≥65 years with IRCTs and (2) compare serial changes in clinical outcomes between treatment groups.

Study Design:

Cohort study; Level of evidence, 3.

Methods:

This study included patients with IRCTs without glenohumeral osteoarthritis who underwent either ASCR or RSA between March 2013 and December 2020 and had at least 2 years of follow-up data. We assessed active range of motion, a visual analog scale (VAS) pain score, the American Shoulder and Elbow Surgeons (ASES) score, and the Single Assessment Numeric Evaluation (SANE) score at the preoperative, short-term (postoperative 6-12 months), and final follow-up times.

Results:

In total, 64 patients (ASCR, 31 patients; RSA, 33 patients) were included. The mean age of patients was 71.3 ± 4.4 and 72.9 ± 4.1 years, and the mean final follow-up duration was 42 ± 21.8 and 37.7 ± 21.7 months in the ASCR and RSA groups, respectively. At the short-term follow-up, RSA achieved significant improvements in all clinical outcomes, except for internal rotation (IR), while ASCR only showed significant improvements in VAS pain, ASES, and SANE scores. Compared with the preoperative period, both ASCR and RSA achieved significant improvements in all clinical outcomes, except for IR in the RSA group at the final follow-up. The ASCR group achieved better IR and ASES scores at the final follow-up, while the time taken to recover from pseudoparalysis was shorter after RSA. The ASCR group showed a 67.8% graft healing rate at the 1-year follow-up, while the RSA group showed 12.1% of scapular notching at the final follow-up. No other postoperative complications were observed in either group.

Conclusion:

ASCR and RSA achieved favorable clinical outcomes in the study cohort. At the short-term follow-up, RSA showed significant improvements in all clinical outcomes, except for IR, while ASCR only showed significant improvements in VAS pain, ASES, and SANE scores. At the final follow-up, however, ASCR had better IR and ASES scores compared with RSA.

Irreparable rotator cuff tears (IRCTs) cause pain and limitation of motion, resulting in loss of function in the affected shoulder; moreover, they could lead to glenohumeral osteoarthritis (OA) if left untreated.^6,11,16 Possible surgical options for managing IRCTs include debridement, partial repair, graft interposition, tendon transfer, arthroscopic superior capsular reconstruction (ASCR), and reverse shoulder arthroplasty (RSA). However, there is a lack of consensus regarding the best surgical option for treating IRCTs.^{4,5,9,24,26,43,48}

ASCR, first proposed by Mihata et al,^34-36 is a joint-preserving surgery performed to restore the glenohumeral joint stability by blocking the proximal humeral head migration for IRCTs. It has shown promising healing rates and clinical outcomes in patients with IRCTs, compared with primary arthroscopic rotator cuff repair with a 78% retear rate. ⁴⁰ However, favorable long-term outcomes after ASCR have not been proven yet, and relatively higher retear rates have been reported in older patients.^{13,23,25,31,41} RSA, which tensions the deltoid by medializing the center of rotation and lengthening the humerus, is a reliable treatment option to improve active elevation, pain, and function in IRCTs and cuff tear arthropathy. However, because of the nature of prosthetic replacement surgery, the longevity and potential risk for complications remain a concern.^8,28,43,45

Based on the advantages and disadvantages of each procedure, we developed an algorithm to be used as a guideline for determining the best surgical option for IRCTs, considering that younger age is related to poor clinical outcomes and higher concerns about implant longevity in RSA (Figure 1).^14,22,25,37 No previous study has compared clinical outcomes after ASCR using the fascia lata autograft and RSA for IRCTs. Therefore, in this study, we aimed to (1) compare the clinical outcomes of ASCR using the fascia lata autograft versus RSA in IRCTs without glenohumeral OA for patients aged ≥65 years and (2) assess serial changes of clinical outcomes in each group. We hypothesized that ASCR and RSA would show comparable outcomes at the short-term follow-up, while ASCR would show overall better outcomes than RSA at the final follow-up.

Figure 1.

Algorithm for determining the best surgical management option for irreparable rotator cuff tears. ASCR, arthroscopic superior capsular reconstruction; RSA, reverse shoulder arthroplasty.

Methods

Our institutional review board approved the protocol of this study. Informed consent was waived due to the retrospective study design. This single-center retrospective cohort study included patients aged ≥65 years with IRCTs who underwent ASCR or RSA between March 2013 and December 2020. IRCTs were diagnosed by defining the greatest dimension of tear ≥5 cm or full-thickness tear of ≥2 tendons on a preoperative magnetic resonance imaging (MRI) scan. ³⁰ The diagnosis was confirmed by the inability to achieve a direct repair of the rotator cuff tendon to the humerus intraoperatively. As there is no consensus regarding the best surgical option for managing IRCTs in patients without glenohumeral OA, which surgery to perform was determined by an agreement between the patients and the surgeon after an explanation of the benefits and risks of each surgery and thorough discussion with the patients. Patients were classified into 2 groups according to their surgery: ASCR versus RSA. All of the clinical processes—including the physical examination, diagnosis, and operation—were performed by a single senior shoulder surgeon (I.-H.J.).

Inclusion and Exclusion Criteria

Patients were included if they met the following criteria: (1) diagnosed with IRCTs on preoperative MRI and confirmed intraoperatively; (2) underwent either ASCR or RSA; and (3) aged ≥65 years. Patients were excluded if they had any of the following: (1) signs of glenohumeral OA with bony deformity (Hamada grades 4 and 5) on preoperative plain radiography; (2) use of grafts other than the fascia lata autograft for patients who underwent ASCR; (3) history of any surgery on the same shoulder; and (4) being observed for <2 years.

Surgical Technique

Arthroscopic Superior Capsular Reconstruction

We performed ASCR using fascia lata autografts between March 2013 and September 2016 and using fascia lata autografts augmented with 0.020–inch (0.058-milimeter) thick polypropylene mesh (Ethicon Inc) between October 2016 and December 2020. The modification of surgical technique was based on better postoperative outcomes shown by mesh augmentation. ²⁶

With the patient under general anesthesia with an interscalene brachial plexus block, a standard diagnostic arthroscopic examination and anterolateral acromioplasty were performed in a beach-chair position. After confirmation of IRCT, the fascia lata was harvested at the ipsilateral side and prepared by the assisting surgeon (S.-P.S., E.K., J.-B.L.) at the back table. The fascia lata was double-folded, and the margin was sealed using a 2-0 Ethibond suture (Ethicon Inc). In the case of mesh augmentation, a single layer of propylene mesh was placed between the double-folded fascia lata. ²⁵ The bony bed at the glenoid and greater tuberosity was then prepared using an arthroscopic shaver, burr, and curette. If present, the long head of the biceps was tenotomized. The graft was fixed at the glenoid side using 3 suture anchors (2.5-mm JuggerKnot, Zimmer Biomet; or 1.7-mm Suturefix Ultra Suture Anchor, Smith & Nephew), which were put at the 10-, 12-, and 2-o’clock positions of the glenoid. Two PEEK threaded anchors (4.5-mm Healicoil PK; Smith & Nephew) were inserted for medial row fixation, and sutures from the medial anchors were passed through the lateral end of the graft and then passed through the remaining bursa and rotator cuff tendon known as the “over-the-top technique.” Two knotless anchors (4.5-mm Footprint Ultra; Smith & Nephew) were used to place suture limbs for lateral row fixation. Further repair to adjacent subscapularis or infraspinatus/teres minor was not performed. After surgery, an abduction brace was applied for 6 weeks. With brace removal, pendulum and passive shoulder range of motion (ROM) exercises were started 6 weeks postoperatively, and rotator cuff and periscapular muscle strengthening exercises were started 3 months postoperatively under the guidance of a dedicated physical therapist.

Reverse Shoulder Arthroplasty

We performed RSA using the Comprehensive Reverse Shoulder Arthroplasty System (Zimmer Biomet) with cemented humerus stem between March 2013 and December 2017 and SMR Reverse Shoulder System (Lima Corporate) with cementless humerus stem between January 2018 and December 2020.

After general anesthesia with an interscalene brachial plexus block, a standardized deltopectoral approach was used to expose the glenohumeral joint in a beach-chair position. The subscapularis was cut 1 cm medial to the lesser tuberosity (LT), and if present, the biceps was tenotomized. After exposure and dislocation of the glenohumeral joint, osteophytes at the humeral head were removed, and the humeral head was cut. Afterward, the glenoid was fully exposed and prepared by adequate labrum removal and capsular release. A guide pin was placed at the center of the glenoid with 10° of inferior tilt, and the glenoid was reamed to assume the neutral version. If glenoid retroversion was >15°, autograft bone grafting using a cut humeral head was placed to reduce retroversion, and a baseplate was inserted, followed by screws and glenosphere insertion. The humerus was then re-exposed and reamed sequentially, maintaining 30° of retroversion using a retroversion guide. A humeral stem trial was inserted, and implant stability and muscle tension were assessed after the reduction of the glenohumeral joint. An intraoperative plain radiograph was taken to check whether the implant was appropriately inserted and reduced. After confirmation of adequate stability and satisfactory findings on an intraoperative plain radiograph, the final humeral component was inserted, and the glenohumeral joint was reduced. For cases with a reparable subscapularis muscle, the muscle was repaired at the LT using a No. 2 polyester suture (Ethibond). After surgery, an abduction brace was applied for 3 weeks. Pendulum and passive shoulder ROM exercises started 3 weeks postoperatively with brace removal, while muscle strengthening exercises started 6 weeks postoperatively under the guidance of a dedicated physical therapist.

Assessment of Clinical Outcomes

Descriptive data—including age, sex, history of smoking, diabetes mellitus, osteoporosis, and pseudoparalysis—were obtained through chart review. We assessed active ROM and patient-reported outcome measures—including a visual analog scale (VAS) for pain, the American Shoulder and Elbow Surgeons (ASES) score, and the Single Assessment Numeric Evaluation (SANE) score. Forward elevation (FE) and external rotation (ER) at 0° abduction were measured using a goniometer, and IR was measured at the vertebral level that could be reached by the patient in the sitting position, consecutively numbered for data analysis. According to the vertebral level, the degree of internal rotation was numbered from 1 (1st thoracic vertebra, T1) to 18 (sacrum), with 7 indicating the 7th thoracic vertebra (T7), 12 indicating the 12th thoracic vertebra (T12), 13 indicating the first lumbar vertebra (L2), and 17 representing the 5th lumbar vertebra (L5).

Complications, including postoperative infections, neurologic deficit, and donor site morbidity (for the ASCR group), and requirements for revision surgery were determined by reviewing the patients’ charts. All clinical outcomes were assessed at 3 time points—preoperatively and short-term (postoperative 6-12 months) and final follow-up (at least 2 years after surgery—by a physician assistant who was not involved in the surgery). The time taken to recover from pseudoparalysis was checked.

Assessment of Radiologic Outcomes

The acromiohumeral distance and the Hamada classification were evaluated using the preoperative plain radiograph taken a day before surgery. Tear sizes in both coronal (anteroposterior) and sagittal (mediolateral) planes and global fatty degeneration index (GFDI) were measured using MRI performed within 6 months before surgery.^15,19,20,49 In the ASCR group, MRI was taken at postoperative 1 year to check the graft status, while graft failure was defined as any sign of graft discontinuity. In RSA, plain radiographs at the final follow-up were reviewed to check implant-related complications, including scapular notching, implant loosening, dislocation, and periprosthetic fracture.

Statistical Analysis

Continuous variables were reported as means and standard deviations and categorical variables were reported as numbers and percentages. An independent t test and the Mann-Whitney U test were utilized to compare continuous variables and categorical variables in patient descriptive data and clinical outcomes between the 2 groups, respectively. A paired t test was used to assess clinical outcomes serially, from the preoperative period to the short-term and final follow-up. A power of 83.9% was calculated for the postoperative ASES score using the post hoc analysis—with a 2-sided significance level of .05, an effect size of 0.67, and a total number of 64 patients. All statistical analyses were performed using IBM SPSS Statistics for Windows Version 21.0 (IBM Corp). Statistical significance was set at P < .05.

Results

Study Population

A total of 101 patients (ASCR, 48 patients; RSA, 53 patients) who met the inclusion criteria were reviewed. Of all patients, 37 were excluded: 8 patients with grafts other than a fascia lata autograft, 3 patients with previous surgery, and 6 patients observed for <2 years in the ASCR group and 11 patients with Hamada grades 4 and 5, 3 patients with previous surgery, and 6 patients observed for <2 years in the RSA group. Therefore, 64 patients (ASCR, 31 patients; RSA, 33 patients) were finally included (Figure 2). In the ASCR group, fascia lata autografts without mesh augmentation were used for 12 patients, while fascia lata autografts with mesh augmentation were used for 19 patients. In the RSA group, the Comprehensive Reverse Shoulder Arthroplasty System was used for 8 patients, while the SMR Reverse Shoulder System was used for 25 patients. The mean age was 71.3 ± 4.4 years and 72.9 ± 4.1 years (P = .138), and the mean final follow-up duration was 42 ± 21.8 months and 37.7 ± 21.7 months (P = .432) in the ASCR and RSA groups, respectively. The ASCR and RSA groups did not show significant differences in preoperative patient descriptive data and radiologic severity of rotator cuff tears. Detailed patient descriptive data and preoperative radiologic severity are summarized in Table 1.

Figure 2.

Flowchart of the patient inclusion process. ASCR, arthroscopic superior capsular reconstruction; FL, fascia lata; IRCTs, irreparable rotator cuff tears; OA, osteoarthritis; RSA, reverse shoulder arthroplasty.

Table 1.

Preoperative Patient Characteristics and Radiological Severity of Rotator Cuff Tears in the ASCR and RSA Groups ^a

Variable	ASCR (n = 31)	RSA (n = 33)	P
Age, y	71.3 ± 4.4	72.9 ± 4.1	.138
Follow-up, mo
Short-term	7.7 ± 2	7.9 ± 2.5	.826
Final	42 ± 21.8	37.7 ± 21.7	.432
BMI, kg/m2	25.9 ± 3.8	24.1 ± 3.4	.052
Male sex	5 (16.1)	11 (33.3)	.309
Affected side, dominant	26 (83.9)	26 (78)	.605
Smoking history	2 (6.5)	6 (18.2)	.159
Diabetes mellitus	4 (12.9)	5 (15.2)	.798
Osteoporosis	4 (12.9)	6 (18.2)	.564
Heavy worker	4 (12.9)	6 (18.2)	.564
AHD, mm	4.9 ± 2	4.8 ± 3.2	.947
Hamada classification	2 ± 0.8	2.2 ± 0.7	.321
1	9 (29)	6 (18.2)
2	14 (45.2)	16 (48.5)
3	8 (25.8)	11 (33.3)
Tear size, mm
Anteroposterior	40.7 ± 4.5	39.1 ± 4.1	.146
Mediolateral	35.4 ± 5.1	36.9 ± 3.3	.168
GFDI	2.6 ± 0.6	2.6 ± 0.7	.862

^aCategorical data are reported as n (%) and continuous data as mean ± SD. AHD, acromiohumeral distance; ASCR, arthroscopic superior capsular reconstruction; BMI, body mass index; GFDI, global fatty degeneration index; RSA, reverse shoulder arthroplasty.

Clinical Outcomes

ROM Measurements and Pseudoparalysis

Preoperative ROM was comparable between the 2 groups (Table 2, Figure 3). At the short-term follow-up, the ASCR group did not show any significant improvements in all ROM measurements, compared with the preoperative period, while all ROM measurements improved significantly from the short-term period to the final follow-up (Table 2, Figure 3). The RSA group showed significant serial improvements in FE and ER from the preoperative period to the short-term and final follow-up, while IR showed a small but significant decrease from the preoperative period to the short-term period (Table 2, Figure 3). All ROM measurements in the ASCR group and FE and ER in the RSA group showed significant improvements at the final follow-up compared with their respective preoperative values. The ASCR group showed significantly better IR than the RSA group at the short-term and final follow-up.

Table 2.

Comparison of Pre- and Postoperative Clinical Outcomes Between the ASCR and RSA Groups at the Preoperative, Short-Term, and Final Follow-up Times ^a

Clinical Outcome	ASCR (n = 31)	RSA (n = 33)	P
Forward elevation, deg
Pre	128.6 ± 37.3	124.1 ± 36	.628
Short-term	138 ± 35.4	146.7 ± 18.1	.752
Final	153.2 ± 19.6	149.7 ± 20.7	.486
P b	.001	.001
External rotation, deg
Pre	29.7 ± 19.1	28.6 ± 17.6	.821
Short-term	33.1 ± 13.9	33.5 ± 13.4	.805
Final	45.3 ± 12.7	40.6 ± 13.7	.159
P b	<.001	<.001
IR c
Pre	13.4 ± 2.7	14.3 ± 2.8	.176
Short-term	12.8 ± 2.5	15.8 ± 2.4	<.001
Final	11.2 ± 2.3	15 ± 2.5	<.001
P b	.001	.252
VAS pain score
Pre	6.1 ± 1.7	5.7 ± 1.9	.386
Short-term	2.2 ± 1.2	2.1 ± 1	.490
Final	1.1 ± 1.1	1.6 ± 1.6	.160
P b	<.001	<.001
ASES score
Pre	41.8 ± 16.4	37.5 ± 9.2	.199
Short-term	69.3 ± 12.3	67 ± 12.1	.382
Final	80.1 ± 12.9	71.5 ± 15.4	.020
P b	<.001	<.001
SANE score
Pre	39.5 ± 17.5	36.7 ± 13.8	.481
Short-term	63.3 ± 12.1	68.9 ± 17.2	.110
Final	77.6 ± 15.3	74.5 ± 13.2	.399
P b	<.001	<.001
Pseudoparalysis
Pre	7 (22.6)	12 (36.4)	.231
Final	0 (0)	1 (3)	.595
P b	.008	.001
Time to recover, mo	7 ± 3	3.3 ± 1.2	.004

^aCategorical data are reported as n (%) and continuous data as mean ± SD. Bold P values indicate statistically significant differences between groups (P < .05). ASES, American Shoulder and Elbow Surgeons; ASCR, arthroscopic superior capsular reconstruction; IR, internal rotation; Pre, preoperative; RSA, reverse shoulder arthroplasty; SANE, Single Assessment Numeric Evaluation; VAS, visual analog scale.

^bP value of the difference between the preoperative and final follow-up values.

^cThe degree of IR was numbered according to the vertebral level reached by the patient, from 1 (first thoracic vertebra, T1) to 18 (sacrum).

Figure 3.

(A-C) Comparison of pre- and postoperative ROM measurements between the ASCR and RSA groups at each time point and (D-F) a serial assessment of ROM measurements in each group; (A, D) forward elevation, (B, E) external rotation, and (C, F) internal rotation. The degree of internal rotation was numbered according to the vertebral level reached by the patient, from 1 (first thoracic vertebra, T1) to 18 (sacrum). ASCR, arthroscopic superior capsular reconstruction; Pre, preoperative; ROM, range of motion; RSA, reverse shoulder arthroplasty. Statistically significant difference *between the time points as shown and †between the preoperative and final follow-up time points (P < .05).

The number of patients with pseudoparalysis decreased significantly after surgery in both groups and those at the final follow-up were comparable between the 2 groups. However, the time taken to recover from pseudoparalysis was significantly shorter in the RSA group than the ASCR group (Table 2).

Patient-Reported Outcome Scores

Preoperative and short-term VAS pain, ASES, and SANE scores were comparable between the 2 groups (Table 2, Figure 4). For the serial assessment, both ASCR and RSA groups showed significant serial improvements in VAS pain, ASES, and SANE scores from the preoperative period to the short-term and final follow-up (Table 2, Figure 4). VAS pain, ASES, and SANE scores significantly improved at the final follow-up in both ASCR and RSA groups, compared with their respective preoperative values. The ASCR group showed a significantly better ASES score than the RSA group at the final follow-up (Table 2, Figure 4).

Figure 4.

(A-C) Comparison of pre- and postoperative VAS pain, ASES, and SANE scores between the ASCR and RSA groups at each time point and (D-F) a serial assessment of VAS pain, ASES, and SANE scores in each group; (A, D) VAS pain scores, (B, E) ASES scores, and (C, F) SANE scores. ASES, American Shoulder and Elbow Surgeons; ASCR, arthroscopic superior capsular reconstruction; Pre, preoperative; RSA, reverse shoulder arthroplasty; SANE, Single Assessment Numeric Evaluation; VAS, visual analog scale. Statistically significant difference *between the time points as shown and †between the preoperative and final follow-up time points (P < .05).

Radiologic Outcomes and Complications

Graft healing was observed in 21 patients (67.7%) in the ASCR group on MRI 1 year postoperatively. At the final follow-up, scapular notching was observed in 4 patients (12.1%) without other radiologic signs of mechanical complications in the RSA group. No other postoperative complications, including donor site morbidity, were observed in both ASCR and RSA groups.

Discussion

This study showed that both ASCR and RSA provide favorable clinical outcomes in IRCTs without glenohumeral OA for patients aged ≥65 years. ASCR achieved better postoperative IR and ASES scores than RSA at the final follow-up. At the short-term follow-up, RSA showed significant improvements in ROM measurements, including FE and ER, and VAS pain, ASES, and SANE scores, compared with the preoperative period, while ASCR only showed significant improvements in VAS pain, ASES, and SANE scores.

ASCR was first introduced in 2013, which was more recent than RSA. ³⁵ Despite that, both ASCR and RSA showed favorable outcomes in patients with IRCTs without glenohumeral OA, and both have been considered reliable treatment options.^{1,2,12,25,33,38,39} However, the surgical alternative that is better for IRCTs is still unknown, and the decision of the surgical option to be used is often made based on the surgeon’s preference and various patient-related factors, including sex, age, activity level, severity of RCTs, presence of pseudoparalysis, and glenohumeral OA. ⁷ To our knowledge, only 1 previous study compared outcomes between ASCR and RSA in IRCTs. ²⁹ Lacheta et al ²⁹ compared clinical outcomes between ASCR using dermal allograft and RSA in patients aged <70 years and reported comparable ASES, SANE, and Quick Disabilities of the Arm, Shoulder and Hand scores at a mean of 2.1 and 2.9 years of follow-up after ASCR and RSA, respectively. However, patient characteristics—including age and follow-up interval—were significantly different between the 2 groups, and preoperative factors that could affect surgical outcomes—including tear size, GFDI, and Hamada classification—were not assessed, while this study showed comparable preoperative patient characteristics and radiologic severity of rotator cuff tears between the 2 groups. In this study, ASCR using fascia lata autografts showed significantly better IR and ASES than RSA at the final follow-up, and all clinical outcomes, except for IR, showed significant improvements in the RSA group. Previous studies reported unpredictable postoperative IR in RSA using Grammont prosthesis.^3,47 IR is known to be essential for activities of daily living and closely related to the patient functional status and satisfaction. Therefore, nonachievement of significant IR after RSA might have led to worse ASES scores than ASCR in this study. ²⁷

In addition to better clinical outcomes in ASCR at the final follow-up, advantages of ASCR as a joint preserving procedure include the fact that complications related to implants or concerns regarding revision arthroplasty could be avoided. However, in terms of graft healing, older age (≥65 years) was related to worse graft healing and clinical outcomes than younger age (<65 years) after ASCR.^17,25 Kholinne et al ²⁵ compared graft healing rate and clinical outcomes after ASCR between younger and older patients and between grafts using only fascia lata autografts and fascia lata autografts augmented with mesh. The overall graft healing rate was 64.9% in patients aged ≥65 years, similar to this study (67.7%). However, subgroup analysis in patients aged ≥65 years showed that mesh augmentation resulted in better graft healing rate (85% vs 41.2%) and clinical outcomes—including ASES scores, VAS pain scores, and ROM—than using only the fascia lata. Moreover, when the graft was augmented by mesh, the graft healing rate (85% vs 83.3%) and clinical outcomes were comparable between older and younger patients. Although we did not perform a subgroup analysis about whether graft healing affects clinical outcomes, we do believe that graft healing is important to achieve better clinical outcomes based on previous literature. Therefore, every effort—including mesh augmentation and firm immobilization after surgery—should be made to ensure that the graft heals well, especially for older patients. In addition, because of the time needed for graft healing, ASCR needs a relatively longer immobilization period than RSA, leading to relatively gradual recovery of function. In this study, all ROM measurements did not improve significantly after ASCR at the short-term follow-up, while all ROM measurements showed significant improvements from the short-term period to the final follow-up.

In contrast, RSA has its advantages of shorter immobilization and earlier rehabilitation than ASCR in IRCTs without glenohumeral OA. In this study, RSA showed already significant improvements in all clinical outcomes, except for IR, at the short-term follow-up compared with the preoperative period. Compared with ASCR, active FE was better in RSA; however, it did not reach statistical significance. Moreover, the time taken to recover from pseudoparalysis was nearly 4 months shorter in RSA than ASCR (3.3 vs 7 months), which was significant, even considering the relatively longer immobilization period for ASCR (3 vs 6 weeks). RSA showed its strength in overall early recovery. However, the most important issue after RSA is implant-related failure and consequent functional decline because revision RSA has been known to be related to higher complication rates and lower patient satisfaction compared with primary RSA.^{10,21,42,44,46} After RSA, 4 patients (12.1%) had scapular notching at the final follow-up; however, no other complications or need for revision surgery were observed in this study. Longer follow-up would be needed to assess implant-related complications after RSA, graft retear after ASCR, and need for revision surgery.

Previously, pseudoparalysis in IRCTs was an important indicator to choose RSA as a surgical option because RSA was considered as the only surgical procedure that could reverse pseudoparalysis; however, ASCR also showed high possibility for reversing pseudoparalysis, which means pseudoparalysis itself is not the indication for RSA. ³⁴ In this study, the time taken to recover from pseudoparalysis was significantly shorter in RSA. However, the number of patients with pseudoparalysis at the final follow-up was comparable between the 2 groups, and both groups showed significant decreases in the number of patients with pseudoparalysis.

This study aimed to compare outcomes of ASCR using fascia lata autografts versus RSA in patients aged ≥65 years with IRCTs. Considering its better ASES scores and IR at the final follow-up after ASCR and less possibility for implant-related complications, we recommend performing ASCR using fascia lata autografts in patients ≥65 years old with IRCTs. However, as both surgeries have different advantages and disadvantages, making patient-specific decisions regarding which surgery to perform and conducting further randomized controlled trials are required to avoid any possible selection bias.

Limitations

This study had several limitations. First, this was a retrospective, nonrandomized study. The authors did not have a preference in deciding which surgery to perform; however, the possibility of selection bias could not be excluded. Second, we did not perform matching between the 2 groups because the study population was not large enough in each group. However, preoperative patient characteristics, radiologic severity of RCTs, and clinical outcomes were comparable between the 2 groups. Third, this study analyzed clinical and radiologic outcomes based on medium-term follow-up. The follow-up period might not have been long enough to evaluate graft failure or deterioration to glenohumeral OA after ASCR and implant-related complications or failure after RSA. Fourth, different techniques of graft preparation and different implant systems were used for ASCR and RSA, respectively. These might affect clinical outcomes after each surgery; however, all surgeries were routinely performed by 1 surgeon. Fifth, we could not prove whether statistically significant differences in ASES at the final follow-up and IR at both short-term and final follow-up are clinically meaningful. However, based on previous studies, we believe that these differences were clinically meaningful.^18,32 Despite the above limitations, this was the first study to compare clinical outcomes between ASCR using the fascia lata autograft and RSA in patients with IRCTs without glenohumeral OA. Moreover, we assessed clinical outcomes serially, which helped us understand the natural course after each surgery.

Conclusion

Both ASCR and RSA achieved favorable clinical outcomes for managing IRCTs without glenohumeral OA for patients aged ≥65 years. At the short-term follow-up, RSA showed significant improvements in all clinical outcomes, except for IR, while ASCR only showed significant improvements in VAS pain, ASES, and SANE scores. However, ASCR showed better IR and ASES scores than RSA at the final follow-up.