Rotator cuff and capsule healing after shoulder arthroscopy: A second look arthroscopic study

Abstract

BACKGROUND

Shoulder arthroscopy is commonly used for the repair of glenohumeral ligament avulsions or tendon tears. The success of the operation depends on the ability of the ligaments or rotator cuff tendon to heal to their original attachment site. Soft tissue healing can be evaluated with imaging methods or alternatively with second-look arthroscopy.

AIM

To investigate shoulder tendon and capsule healing after arthroscopic rotator cuff and instability repair using second-look arthroscopy.

METHODS

In this study, 24 adult patients with rotator cuff tears (13 patients) or anterior shoulder instability (11 patients) were included. All patients were initially subjected to arthroscopic repair using suture anchors and were re-evaluated with second-look arthroscopy for reasons not related to the original pathology. The second operation was performed in 8 patients due to mild but persistent pain or stiffness, in 3 patients for recalcitrant stiffness, in 5 patients for secondary biceps tenotomy, in 6 patients for persistent acromioclavicular joint pain and in 2 patients for suture anchor prominence causing shoulder grinding. Soft tissue healing was evaluated visually and by probing, whereas clinical outcomes were evaluated using the University of California–Los Angeles (UCLA) and Rowe rating scales.

RESULTS

In almost all patients, complete soft tissue healing occurred at the site of tissue reattachment, either on the glenoid articular surface or the greater humeral tuberosity. The strongest repair, as confirmed by probe palpation, was encountered at the site of suture passage through the soft tissue. All suture material was covered with bursal synovial tissue, with no cases of knot impingement or cartilage fraying. The mean preoperative and postoperative UCLA scores for rotator cuff repair patients were 1354 ± 3205 and 2931 ± 2898, respectively (P < 0.001), whereas for shoulder instability patients, the mean Rowe scores preoperatively and postoperatively were 2591 ± 1338 and 9272 ± 754, respectively (P < 0.001). The use of bioabsorbable implants did not cause synovitis or other tissue reactions.

CONCLUSION

Soft tissue healing in the shoulder is successful and strongest at the site of suture anchor placement.

Key Words: Arthroscopy; Instability; Rotator cuff; Shoulder; Soft tissue healing; Tendon

Core Tip: This is the first paper describing the macroscopic appearance of soft tissue healing in the shoulder using shoulder arthroscopy. A series of patients who had had shoulder arthroscopy for instability or rotator cuff tear were subjected to second look arthroscopy. The macroscopic findings of the study are described.

INTRODUCTION

Shoulder arthroscopy has significantly evolved over the last 20 years following the refinement of surgical indications and the advancement of instrumentation, surgical techniques, and implanted materials. The available information regarding the progression and efficiency of in vivo soft tissue and tendon healing in the human shoulder remains sparse and is primarily based on imaging techniques[1,2]. Tendon-to-bone healing is challenging because of the limited regenerative capacity of tendon-bone entheses[3]. In fact, the normal tendon-to-bone enthesis following surgical repair is not regenerated, but a biomechanically and histologically inferior enthesis is reformed, composed primarily of type I and III collagen fibers[2-4].

The primary surgical goal of arthroscopic shoulder surgery is to reposition the detached glenohumeral ligaments in unstable shoulders or to reattach the torn rotator cuff tendons to the bone to facilitate ligament or tendon-to-bone healing. Consequently, the success of arthroscopic shoulder surgery depends on the successful healing of the repaired soft tissues[1-5]. Although watertight healing is not an absolute prerequisite to achieve favorable clinical outcomes, it can be presumed that successful soft tissue healing may provide a better functional outcome[6,7]. The rate of rotator cuff retears after arthroscopic repair varies significantly, reaching 94% 12 months postoperatively[8], and it occurs in 27% of full-thickness tears and 5% of partial-thickness tears[9]. The average retear rates for patients with single-row and double-row repairs are 14.5% and 12.7%, respectively[10]. Glenohumeral ligament[5-7] healing on the glenoid is essential to ensure the restoration of glenohumeral stability[8-11].

The primary purpose of this study was to describe the arthroscopic morphology of the shoulder capsule and rotator cuff healing viasecond-look arthroscopy in a series of patients previously subjected to shoulder arthroscopy. The secondary goals are the description of possible cartilage or soft tissue injury caused by prominent arthroscopic knots and the possible reaction of synovial shoulder tissue to surgery and surgical implants.

MATERIALS AND METHODS

In this study, 24 adult patients with rotator cuff tears (13 patients, 8 males and 5 females, mean age 60.85 years ± 11.81 years) or anterior shoulder instability (11 patients, 9 males and 2 females, mean age 24 years ± 4.53 years) were included. All patients were initially subjected to arthroscopic repair and were re-evaluated with second-look arthroscopy for various reasons not directly related to the initial pathology. The study was approved by the Institutional Review Board Committee.

The second operation was performed in 8 patients on a volunteer basis, with mild, nonspecific discomfort in overhead sports activities resistant to treatment, in 3 patients for shoulder stiffness, in 5 patients due to persistent biceps tendinopathy to perform secondary biceps tenotomy, in 6 patients for persistent acromioclavicular joint pain and in 2 patients for anchor prominence and mechanical symptoms. Detailed data on the patients included in this study are presented in Tables 1 and 2. Magnetic resonance imaging (MRI) scans were reviewed prior to the first and second operations (Figures 1 and 2), which revealed complete healing of the repaired capsule or rotator cuff tendons in almost all the patients.

Figure 1 Magnetic resonance imaging evaluation of a repaired rotator cuff 11 months after the first operation in a 63-year-old patient. A: Coronal magnetic resonance imaging (MRI) of the repaired rotator cuff. The rotator cuff tendons were reattached using a double row technique and absorbable suture anchors (arrowheads). Complete healing with restoration of tendon continuity is observed; B: Sagittal MRI showing restoration of the signal intensity of the rotator cuff tendons (arrowheads).

Figure 2 Transverse magnetic resonance imaging of the shoulder in a patient who underwent revision Bankart repair 5 months earlier. The anterior capsulolabral structures were reattached on the anterior glenoid rim using suture anchors (arrowhead). Remplissage was performed to cover a sizeable engaging Hill-Sachs lesion using another suture anchor (arrowhead).

Table 1 Demographic data, outcome scores and intraoperative details of the patients with shoulder instability who underwent second look arthroscopy.

Case number	Gender	Age (years)	Side	Chronicity	Arthroscopic findings on the initial operation	Number of anchors	Follow-up (months)	Reason for reoperation	Findings at the second operation	Rowe score preoperatively	Rowe score prior to second look arthroscopy
1	Male	19	Right	Acute	Bankart, Hill-Sachs, SLAP II	4	19	Biceps tendon pain	Partial healing SLAP	45	100
2	Male	23	Right	Acute	Bankart, Hill-Sachs	3	14	Stiffness	Firm healing	45	100
3	Male	21	Right	Chronic	Bankart, Hill-Sachs	4	19	Activity related pain	Partial healing superior glenohumeral ligament	40	100
4	Male	22	Left	Acute	Bankart, Hill-Sachs	3	22	New injury with persistent synovitis	Firm healing	40	100
5	Male	20	Right	Chronic	Bankart, Hill-Sachs	3	24	ACJ pain	Firm healing	20	95
6	Female	28	Left	Chronic	Bankart, Hill-Sachs	3	12	Occasional locking of the shoulder	Partial healing superior glenohumeral ligament, anchor pull out	20	95
7	Male	32	Left	Chronic	Bankart, Hill-Sachs	3	18	ACJ pain	Firm healing	15	90
8	Male	31	Left	Chronic	Bankart, Hill-Sachs, SLAP II	4	13	Biceps tenodesis	Mild synovitis, complete healing	15	90
9	Male	24	Right	Chronic	Bankart, Hill-Sachs	2	26	Anchor prominence	Firm healing, no cartilage injury	15	90
10	Male	25	Right	Acute	Bankart, Hill-Sachs	3	21	Stiffness	Firm healing, anchor pull out	15	80
11	Male	20	Left	Chronic	Bankart, Hill-Sachs	3	19	Persistent activity pain	Partial healing superior glenohumeral ligament	15	80

ACJ: Acromioclavicular joint; SLAP: Superior labral tears.

Table 2 Demographic data, outcome scores and intraoperative details of the patients with rotator cuff pathology who underwent second look arthroscopy.

Case number	Gender	Age (years)	Side	Tear size (cm)	Number of anchors	Follow-up (months)	Reason for reoperation	Findings at the second operation	UCLA score preoperatively	UCLA score prior to second look arthroscopy
1	Male	44	Right	2	3	22	Biceps tenodesis	Complete healing	30	44
2	Female	52	Right	1.5	2	17	Biceps tenodesis	Complete healing	28	52
3	Female	59	Left	2	3	12	Anchor prominence	Local synovitis, partial healing locally	27	59
4	Female	64	Left	2.5	3	12	ACJ pain	Complete healing	25	64
5	Female	57	Right	3.2		17	Biceps tenodesis	Complete healing	30	57
6	Female	69	Right	2.2	4	29	ACJ pain	Complete healing	30	69
7	Female	74	Left	3.3	3	17	Stiffness	Synovitis, complete healing	33	74
8	Female	63	Right	2.9	4	11	Biceps tenodesis	Complete healing	33	63
9	Male	35	Left	1	2	15	Mild, persistent stiffness and discomfort	Complete healing	31	35
10	Female	77	Left	4	4	24	ACJ pain	Complete healing	31	77
11	Male	70	Right	3.5	4	20	ACJ pain	Partial healing	30	70
12	Male	66	Right	2.4	3	16	ACJ pain	Complete healing, mild synovitis	30	66
13	Female	61	Right	3	3	13	Stiffness	Mild synovitis	23	61

ACJ: Acromioclavicular joint; UCLA: University of California–Los Angeles.

First shoulder arthroscopy

All operations were performed in the beach chair position. In rotator cuff repair patients, three arthroscopic portals were used (posterior viewing and anterior and lateral working portals), and subacromial bursectomy and acromioplasty were performed. Tear size was evaluated after debridement of the degenerated tendon edges ranging between 1.5 cm and 4 cm. Double-row fixation was performed in all patients with one or two medial row anchors and two lateral row absorbable anchors (Healix, DePuy Mitek, Raynham, MA, United States). The rotator cuff tendons were released until approximation to the greater tuberosity was possible, taking care to avoid undue tension at the repair site. The greater tuberosity was lightly abraded with a rasp, avoiding removal of the humeral cortex to prevent a reduction in the anchor pull-out strength. Postoperatively, the upper limb was placed in a sling, and physiotherapy started after 7–10 days and continued for 12–16 weeks as necessary.

In the instability group, 11 patients with anterior instability were included: (1) 8 patients with chronic instability (> 4 dislocations); and (2) 3 patients who underwent surgery after the first dislocation. The operation was carried out in the lateral decubitus position with one posterior view and one anterior and one anteroinferior working portals. During the operation, the avulsed anterior capsule (Bankart lesion) or the superior labrum [superior labral tears (SLAP) II lesion in 2 cases] were reattached using double-loaded suture anchors (Lupine, DePuy Mitek, Raynham, MA, United States). The upper limb was immobilized postoperatively in a sling, and after one week, passive range of movement exercises commenced. External rotation beyond 45 degrees was avoided for 6 weeks, and contact sports and overhead weight-bearing activities were avoided for 6 months.

Second-look arthroscopy

The reasons for second-look arthroscopy are listed in Tables 1 and 2. All patients underwent extensive conservative treatment and were reoperated only when it failed to relieve the symptoms. The procedure was performed in the beach chair position using a posterior viewing and an anterior or lateral working portal.

Soft tissue healing was evaluated visually and by probing the tendons or the anterior capsule (Figure 3). Three types of healing were established: (1) Complete, watertight healing; (2) Incomplete or partial healing; and (3) No healing.

Figure 3 Arthroscopic view of a left shoulder from the posterior viewing portal. A switching stick is inserted through the anterosuperior portal and used to palpate the area of capsule reattachment.

The rotator cuff repair patients (mean follow-up duration of 1907 months ± 552 months) were evaluated preoperatively and prior to second-look arthroscopy using the University of California–Los Angeles (UCLA) rating scale[12,13]. The shoulder instability patients (mean follow-up of 18.23 months ± 3.39 months) were evaluated preoperatively and prior to second-look arthroscopy using the Rowe rating scale[14]. The mean time between the first and second operations was 17.49 months ± 4.22 months in the rotator cuff group and 15.29 months ± 3.09months in the instability group.

Statistical analysis

Statistical evaluation of the rating scale scores and external rotation in patients with instability was performed using the paired t tests. The level of significance was set at P < 0.05. Statistical analysis was performed with IBM Statistical Package for the Social Sciences statistical software (version 29.0; SPSS, Chicago, IL, United States).

RESULTS

There were no cases of recurrence of shoulder instability. Compared with their preoperative status, all the unstable shoulders significantly improved, with mean Rowe scores of 25.91 ± 13.38 and 92.73 ± 7.54, respectively, before and before the second operation (P < 0.001). Mild synovial proliferation, i.e., hypervascularity of the synovium without hypertrophic villi, was noted and removed in 3 patients at the rotator interval, whereas complete healing of the anterior capsulolabral structures was noted in 7/11 patients (Figure 4). Upon direct palpation using the arthroscopic probe, the strongest repair was encountered at the site of suture passage, i.e., passing the probe between the soft tissue and the bone was not possible. There were no cases of knot impingement or humeral head cartilage injury, and all the suture material was covered with bursal synovial tissue.

Figure 4 Arthroscopic evaluation of the glenohumeral joint 14 months after the first operation. A: The anterior capsule is viewed from the anterosuperior portal revealing complete, watertight healing to the glenoid rim; B: The knots are covered with synovium without causing articular surface fraying.

In 4 patients, there was incomplete healing in the superior glenohumeral ligament area (1 and 3 o’clock positions). In those cases, external rotation was better than in cases with complete healing but without a positive apprehension test or instability, probably because of sufficient tension afforded by inferior glenohumeral ligament healing. The degree of external rotation with the arm at the side was 78° ± 5.1° in the complete healing group and 85°± 5.3° (P < 0.01) in the incomplete healing group, whereas the degree of external rotation at 90° of abduction was 98°± 6.2° and 86°± 7.1° (P < 0.01). Finally, the SLAP lesion was incompletely healed despite complete healing of the anterior capsule.

In rotator cuff repair patients, the mean preoperative and final UCLA scores were 13.54 ± 3.20 and 29.31 ± 2.89, respectively (P < 0.001). In all rotator cuff repair cases except one, tendon healing was complete with continuity of the tendon and bursal coverage without visual or palpable signs of incomplete healing (Figure 5). In a 3.5 cm long chronic rotator cuff, tear healing was incomplete in the anterior part of the greater tuberosity (Figure 6). The undersurface of the acromion and the bursal surface of the rotator cuff were covered with synovium without signs of inflammation or synovitis in all but 3 patients. The arthroscopic knots were covered with synovium and laid flat on the surface of the tendon. In three patients, prominent knots were removed without compromising the stability of the tendon repair. In one case, there was partial breakage of an absorbable lateral row anchor that was removed. Suture failure, primarily partial knot untying, was noted in 4 instances, and they were removed without compromising the integrity of the rotator cuff repair.

Figure 5 Arthroscopic view of the repaired rotator cuff tendons in the right shoulder from the posterior portal. A: The repaired rotator cuff is covered with synovium without evidence of synovitis or fraying. This patient was re-operated due to persistent acromioclavicular joint pain; B: The repaired rotator cuff is covered with synovium and an arthroscopic knot is evident (arrowhead). This patient underwent re-operation for stiffness.

Figure 6 Partial rotator cuff healing failure in the left shoulder viewed from the posterior portal. The suture cut through the tendon and the anchor is seen uncovered. The orthocord suture appears white because its purple-colored phytoene desaturase coating has been absorbed.

In five patients with evidence of partial discontinuity of the tendon on preoperative MRI, a visible or palpable tear or rent was observed in three patients, whereas no visible gap was noted in the other 2 patients.

DISCUSSION

The most important finding of the present study was that soft tissue healing in the shoulder following arthroscopic repair is accomplished successfully and can also be evaluated using second-look arthroscopy. Partial failure of the suture knots or the suture anchors may not cause symptoms.

Healing of rotator cuff tendons after surgical reattachment is associated with a high rate of incomplete healing[15,16]. Structural failure of the rotator cuff after arthroscopic repair is not uncommon, but it is not always associated with clinical failure[7,8,15]. Nevertheless, patients[7,8,15] demonstrate excellent pain relief and improved function[16-18]. Patients with an intact repaired rotator cuff display better shoulder function and active range of motion than patients with a recurrent defect[8,19,20]. In patients with recurrent rotator cuff defects, the size of the defect increases with time along with progressive loss of muscle strength, despite improvements in pain level, function and satisfaction[21-23]. Tears smaller than 2 cm² are twice as likely to heal than tears greater than 6 cm²[23].

Shoulder instability repair may fail because of unrecognized bony defects, failure to address capsular laxity, technical errors, inappropriate implants and neglected associated pathology to secondary structures of the glenohumeral joint, such as the articular cartilage or the rotator cuff[24]. The lifetime recurrence rate following arthroscopic Bankart repair ranges between 3.4% and 33.3%, although the average recurrence rate is 13.1%[25]. Following arthroscopic revision Bankart repair, the failure rate ranges between 6.1% and 46.8%, reflecting diminished soft tissue quality and reduced healing potential[26]. In revision arthroscopic Bankart repair, only 47.5% of patients return to their preinjury level of play[27].

MRI is the imaging method of choice for studying the efficiency of soft tissue and tendon healing following arthroscopic shoulder surgery, but it does not provide information on soft tissue tension or the robustness of healing. Second-look arthroscopy is useful in selected cases with unexplained postoperative pain because it can address issues such as anchor protrusion, synovitis or osteolysis. In our cases, knot protrusion or bioabsorbable anchor disengagement did not cause significant clinical symptoms, and neither synovitis nor osteolysis around the bioabsorbable anchors was noticed on MRI or second-look arthroscopy. The SLAP lesion was incompletely healed despite complete healing of the anterior capsule, probably because the supraglenoid area was not decorticated; thus, most of the superior labrum was reattached to the glenoid articular cartilage.

Suture failure and knot untying were noted in 6 patients. Park et al[28] reported the treatment of knot-induced pain in 11 patients after SLAP II lesion repair. The patients reported sharp pain (100%) and clicking (64%) with failed SLAP repair. Knots placed on the glenoid side of the superior labrum caused glenoid cartilage erosion, and arthroscopic removal improved the symptoms.

We noticed that soft tissue healing is strongest at the site of the suture anchor location, although the area available for healing is reduced because of the presence of the anchor.

Although second-look arthroscopy[29-31] has been used to evaluate the success of various arthroscopic repair techniques for the knee and shoulder in asymptomatic patients[32-35], it cannot be recommended as a routine arthroscopic procedure because of its invasive nature. It is justified when there is persistent clinical symptomatology and when the imaging techniques (MRI, computed tomography scan or ultrasound) either locate or fail to reveal the source of the symptoms. In our series, 8 patients experienced persistent mild pain and stiffness; these patients were unresponsive to conservative treatment and chose to undergo a second-look arthroscopy to resolve their symptoms. Second-look shoulder arthroscopy[5,36-39] has been used in patients with recurring problems[29-31,40,41] or to evaluate the results of a previously performed arthroscopic operation[32-35,42].

Warner et al[5] performed second-look arthroscopy in 15 patients whose arthroscopic Bankart repair failed. In 6/7 patients with recurrent instability, the capsule was lax, and complete healing was observed in 3 patients. In two patients, biopsy revealed histiocytic infiltration with foreign body giant cells around the polyglyconate polymer debris. Notably, in 8 patients with pain but without recurrence of instability, the anterior capsule healed in only 5 of the patients. Sugimoto et al[34] evaluated 30 patients following successful open anterior shoulder instability repair and reported that the magnetic resonance findings were in agreement with second-look arthroscopic findings in 95% of the patients.

Wilson et al[42] performed arthroscopic staple fixation in 35 patients with rotator cuff repair and evaluated the healing of the rotator cuff with second-look arthroscopy on staple removal 3 months after surgery. Complete healing in 22 (67%) patients was reported, and a small area of necrosis under the staple head was observed in 2 patients (6%). Failure to heal was reported in 3 patients (9%). Wolf and Bayliss[35] reported that 70% of rotator cuff tears appeared normal during second-look arthroscopy, which was performed at an average of five months postoperatively.

Huberty et al[37] treated 24 patients with postoperative stiffness after rotator cuff repair with arthroscopic lysis of adhesions and reported complete healing of the rotator cuff in 23 of 24 patients and 60% footprint coverage in the remaining patient. Mormino et al[38] performed second-look arthroscopy in 13 patients with shoulder stiffness after rotator cuff repair at an average of 37 weeks after the first operation. The rotator cuff healed in all patients, and subacromial adhesions were released without capsular release.

Nakagawa et al[39] performed second-look arthroscopy in 20 patients 3–13 months after arthroscopic repair of partial articular surface tendon avulsion lesions due to pain, stiffness, or weakness. Healing was complete in most cases with conversion of the partial tear to a full-thickness tear, as opposed to the cases where trans-tendon repair was performed. Weber[40] reported that there was no healing response after arthroscopic debridement and acromioplasty of partial-thickness rotator cuff tears.

Lee and Harryman[31] performed arthroscopic bone grafting of a large posterior subchondral cyst adjacent to a malunited glenoid fracture via a cancellous bone graft harvested from the posterior bare area of the humeral head to fill the cavity. One year later, second-look arthroscopy revealed complete healing of the posterior capsular release and integration of the bone graft, which was covered with fibrocartilage. Kanbe et al[29] evaluated the results of arthroscopic cuff repair using a polytetrafluoroethylene patch to cover the lost tissue 1 year after the first operation, and Mori et al[41] used second-look arthroscopy to evaluate the healing of a fascia lata autograft. Kholinne et al[30] evaluated 6 patients who underwent superior capsular reconstruction using an Achilles tendon-bone allograft for the treatment of irreparable rotator cuff tears at a mean of 7.5 months postoperatively. They reported an 83.3% graft failure rate despite the significant improvement in the American Shoulder and Elbow Surgeons and visual analog scale scores and the range of motion. Scheibel et al[32] assessed the healing of osteochondral autologous transplantations in the shoulder using second-look arthroscopy in two patients 6 months after the operation. They noted integration of the graft and preservation of the overlying articular cartilage.

The evaluation of soft tissue healing after arthroscopy typically involves a combination of clinical examination, imaging, and functional assessment. Each method provides valuable insight into different aspects of healing. Both MRI and ultrasound are valuable tools for evaluating the rotator cuff after surgery, but each has strengths and weaknesses[43,44]. MRI is generally considered the gold standard for evaluating the rotator cuff postoperatively because of its superior ability to detect retears and structural changes to the muscle and the tendon. It provides detailed anatomical information, has high sensitivity and can provide quantitative evaluation[43,44]. On the other hand, it is time-consuming and has increased cost, limited accessibility and several contraindications, such as the presence of metallic implants. Finally, the intra- and interobserver agreement of MRI assessment of rotator cuff healing using the 5-type Sugaya classification is only fair or substantial[45]. Ultrasound is a valuable alternative and is preferred for routine follow-up and serial monitoring of tendon healing due to its cost-effectiveness, portability and radiation-free, dynamic, real-time imaging[43,46]. Ultrasound is operator dependent, provides limited depth evaluation, and the imaging is less detailed[46]. A multifaceted approach that incorporates both imaging methods offers the best evaluation of healing, guiding decision-making for further rehabilitation or interventions if necessary. Finally, needle arthroscopy, which uses a 1–1.9 mm diameter arthroscope under local anesthesia, can be used to evaluate patients who are ineligible for an MRI and shows comparable accuracy with MRI for diagnosing intra-articular shoulder pathologies or rotator cuff pathologies[47].

The present study has several limitations. Second-look shoulder arthroscopy was performed in a limited number of patients without clinical failure or significant functional symptoms related to the operation. Revision shoulder arthroscopy cases for rotator cuff structural failure, infection or instability recurrence were not included in the study. Another limitation is the lack of standardized MRI or ultrasound evaluations of the repaired tendons or capsules in all patients, although MRI revealed healing of the capsule or rotator cuff tendons in most of the patients.

CONCLUSION

Our study adds to the knowledge regarding soft tissue healing in humans using direct palpation and arthroscopic visualization. Soft tissue healing following arthroscopic shoulder surgery is successful and is strongest at the site of suture anchor placement. Bioabsorbable implants do not cause significant synovitis or other tissue reactions. Second-look shoulder arthroscopy might be useful in patients with persistent symptoms following shoulder arthroscopy because the source of persistent symptoms can be evaluated and treated simultaneously.