INTRODUCTION
Since reverse total shoulder arthroplasty (RTSA) was introduced in 1987, its indications have been consistently expanded by implant development, advancements in surgical technique, and accumulation of experience [1]. Originally, medializing the center of rotation (COR) and lengthening the humerus increased deltoid muscle tension, making the deltoid lever arm more efficient [2]; however, issues such as scapular notching [3-5] and reduction of the external rotation moment arm occurred [6-8], limiting clinical outcomes. Furthermore, with lengthening of the humerus, changes in the muscle vector negatively affected glenohumeral joint stability frequently [9]. In concomitant subscapularis (SSC) tendon repair in RTSA with a medialized design, a 9% dislocation rate was reported in procedures that did not achieve SSC tendon repair compared with a 0% rate following procedures in which repair was achieved (relative risk, 1.9), and hence, attempted repair was advocated in every case [10].
However, the need for SSC tendon repair with lateralized implant designs remains unclear. In several reports on the biomechanical effects of lateralization, the lateralized prosthesis was found to reduce scapular notching and to preserve the remaining rotator cuff muscles, thereby allowing for an increased moment arm while maintaining the original anatomical vector of the muscles around which the deltoid wraps the shoulder [11-17]. This more normalized vector maximizes the overall arc of motion, especially external rotation, and improves stability when compared to a medialized design [16,18-20]. Lateralized designs reportedly have lower dislocation rates (0%–4.2%) [21-24] than medialized designs (0%–8.6%) [25-29]. A previous study examining the role of concomitant SSC tendon repair in dislocation rates associated with lateralized designs found no difference between the repair and no-repair groups [21,30-34].
The SSC is a strong muscle that strengthens the internal rotators after RTSA; however, it may antagonize the posterior deltoid and external rotators, especially the teres minor (TM) [35]. The original insertion of the SSC migrates inferiorly after RTSA; hence, the vector moves perpendicularly and acts as an adductor, which counteracts the abductor moment arm of the deltoid muscle [36,37]. Some researchers have suggested performing SSC repair at a superiorly migrated position of the humerus rather than reattaching it to the original footprint, as this approach restores the preoperative vector of the SSC tendon and improves glenohumeral abduction [38,39]. However, patient satisfaction after RTSA may be reduced due to external rotation weakness; hence, the advantages of concomitant repair of SSC tendon in lateralized designs remain controversial.
Therefore, the aims of this research were to determine: (1) The effects of SSC repair on abduction and rotation following RTSA; (2) the effects of superior repair of the SSC tendon on glenohumeral abduction and rotation compared with original repair; and (3) the effects of complete loss of the posterior cuff on SSC repair.
METHODS
This cadaveric biomechanical study does not involve living human subjects or identifiable private information and therefore does not constitute human subjects research according to 45 CFR 46. Institutional review board approval was not required.
Specimen Preparation
Eight fresh-frozen cadaveric shoulders (four left and four right; mean age, 60.3±7.6 years; range, 51–70) were tested in a custom, validated shoulder testing system (Fig. 1). The specimens were stored at -20°C and thawed overnight at room temperature before the experiment. The humerus was transected 2 cm distal to the deltoid tuberosity and the scapula was removed from the chest wall. The clavicle, serratus anterior, pectoralis minor, coracobrachialis, short and long heads of the biceps brachii, trapezius, triceps, and brachialis muscles were removed completely along with all the neurovascular structures. Dissection was performed carefully to preserve the tendinous insertions of the pectoralis major, latissimus dorsi, deltoid, TM, and SSC muscles, and the coracoacromial ligament. All specimens were macroscopically intact with no history of trauma or prior surgery.
Surgical Procedures for RTSA
RTSA was performed using an Exactech Equinoxe Reverse System (Exactech), which features a lateralized humeral stem with a 145° neck-shaft angle and a lateralized COR. The humeral procedure was performed first. Humeral implant retroversion was determined using the relationship between the humeral axis and biceps groove described by Doyle and Burks [40] by placing the humeral implant 12-mm posterior to the bicipital groove, reproducing normal anatomic humeral retroversion. The humerus was then sectioned along the anatomical neck with care to prevent damage to the tendinous insertions of the rotator cuff. The intramedullary canal was reamed and the humeral stem was inserted. Humeral stems measuring 7, 9, and 11 mm were used for each specimen. A stem which has mild resistance during insertion was selected and press-fitted into the humeral shaft to prevent rotation during the experiment. A standard humeral adaptor tray and a standard polyethylene liner for 36-mm glenosphere were used.
The glenoid procedure was performed next. Normal glenoid version and anatomy were confirmed preoperatively. Bone loss or abnormal version of the glenoid muscles was not observed. The capsule-labral tissue located 2 to 10 hours from the glenoid was removed using a blade. The central hole for the peg reamer was established carefully using a standard baseplate guide to place the glenosphere at an inferior tilt of 10° with a 3-mm inferior overhang to minimize prosthesis-scapular impingement. Subsequently, the glenoid surface was reamed until the subchondral bone was exposed and the central peg was reamed. A standard baseplate was inserted onto the glenoid surface and superior, inferior, and 4 or 8 O’ clock screw holes were drilled and fixed with locking screws. The screws were advanced to penetrate the far cortex and then secured with locking caps. A standard 36-mm glenosphere was implanted onto the baseplate and fixed with a central locking screw. Proper fixation was confirmed by pulling the glenosphere with a bone hook. Finally, the inferior overhang and tilt of the glenosphere were examined.
Testing Setup
Tendinous insertions of the muscles were tagged using locking sutures made with No. 2-0 Fiber-Wire (Arthrex). The TM, latissimus dorsi, and pectoralis major muscles were tagged with a single locking suture at the center of the tendinous insertion. The anterior, middle, and posterior deltoid insertions were tagged with three locking sutures. The SSC was tagged with three locking sutures at the upper and lower ends of the tendinous portion and between (in the middle) the two ends. A bone tunnel was created 3 mm posterior to the bicipital groove to facilitate superiorly migrated SSC insertion for optimal tendon repair (Fig. 1). The migrated position was chosen because, following RTSA, inferior displacement of the subscapularis insertion of approximately 5–10 mm is typically observed, as reported in biomechanical studies. This location was selected to allow superior repositioning and suturing of the SSC within a range that prevents tearing of the remaining tendon tissue.
The scapula was mounted on the shoulder testing system with 0° of scapular abduction (medial border perpendicular to the floor) and 20° anterior scapular tilt to mimic the anatomic position of scapula-thoracic articulation. An intramedullary rod (400 g) was inserted and distally weighted to simulate the weight of the arm. The previously sutured tags were tied with braided low-stretch fishing line (Izorline) which was fed through an adjustable metal guide in the testing system, simulating anatomic muscle vectors. The muscle loading direction was validated based on the original muscle vector direction [41]. The ends of these lines were connected to a free hanging weight that represented muscle loading. The muscle load was determined according to previous studies, using the cross-sectional surface of each muscle [42,43]. The muscle force for balanced load was as follows: three SSCs (2.5 N each); TM (2.5 N); latissimus dorsi (2.5 N); pectoralis major (2.5 N), and three deltoids (anterior, posterior; 10 N each, middle; from 10 to 20 N with 2.5-N increments).
A three-dimensional (3D) referencing system was established using the MicroScribe 3DLX device (Revware) (accuracy, <0.3 mm) with the testing system to measure humeral abduction and rotation. The scapular reference was established using three screws inserted into the tip of the coracoid and the anterior and posterior corners of the acromion. The humeral reference was established using three screws: 2 cm distal to the greater tuberosity and 0.5 cm posterior to the biceps groove, 3 cm distal to the greater tuberosity and 5 mm posterior to the biceps groove, 3 cm distal to the greater tuberosity, and 1 cm posterior to the biceps groove. The MicroScribe 3DLX device was referenced to the scapular screws at the beginning of each test condition and the initial humeral position was measured. The initial position was set by the humerus at 0° abduction (humerus perpendicular to the floor) and 0° rotation (the central screw of the glenosphere and lateral marking of the humeral stem apex in the same line).
First, the three deltoid lines (tagged on insertion of the anterior, middle, and posterior deltoid; 10 N each) were loaded and the other lines were unloaded to confirm the deltoid muscle vector. After placing the humerus in the initial position, it was moved freely by releasing the hand. If there was humeral rotation, flexion, or extension during shoulder abduction, we adjusted the deltoid fishing line’s metal guide anteriorly or posteriorly to position the humerus in neutral rotation and abduction in the scapular plane. The experiment was continued after confirming the absence of humeral rotation or flexion/extension imbalance.
Testing Conditions
We tested six conditions sequentially as follows: (1) unrepaired SSC/intact TM (SSC unloaded [SU]/TM loaded [TL]); (2) original SSC repair/intact TM (SSC loaded [SL]/TL); (3) superiorly repaired SSC/intact TM (superiorly repaired SSC loaded [SSL]/TL); (4) unrepaired SSC/deficient TM (SU/TU); (5) original SSC repair/deficient TM (SL/TM unloaded [TU]); and (6) superiorly repaired SSC/deficient TM (superiorly repaired SSC unloaded/TU).
Conditions SU/TL, SL/TL, and SSL/TL represented large to massive rotator cuff tears with intact TM. Conditions SU/TU, SL/TU, and SSL/TU represented large to massive rotator cuff tears with TM deficiency. When SSC repair was not performed with an intact TM (SU/TL), three SSC fishing lines were unloaded, and the TM fishing line was loaded. In complete repair of the SSC and intact TM (SL/TL), three SSC fishing lines and the TM fishing line were loaded. Superior repair of the SSC with an intact TM (SSL/TL) was simulated by the supraspinatus fishing line and upper two SSC fishing lines being loaded. Conditions SU/TU, SL/TU, and SSL/TU were the same as SU/TL, SL/TL, and SSL/TL except that the TM was unloaded.
Testing Protocol and Statistical Analysis
After equalization of the testing system, the weights were loaded to the fishing lines for each condition. The humerus was placed in the initial position and was freely movable by releasing the hand. The movement from initial position to released position was detected using the MicroScribe 3DLX device and the reference screw of the humerus to calculate its abduction and rotation. Three trials were repeated for each condition. Mid-deltoid weight was raised at 2.5-N increments from 10 N to 20 N, and humeral abduction and rotation were measured repeatedly (Fig. 2).
Data were exported to a Microsoft Excel spreadsheet (Microsoft Corp.), and the mean and standard deviation were calculated. Repeated measures analysis of variance (ANOVA) was performed along with post-hoc tests with Bonferroni correction to compare the abduction and rotation for the six different conditions using SPSS statistical software version 25 (IBM Corp.). All P-values less than 0.05 were considered to indicate statistical significance.
RESULTS
Humeral Abduction and Rotation with Intact TM
In the simulated intact TM (TL) scenario, humeral abduction and rotation was tested under three conditions: (unrepaired SSC (SU/TL), original SSC repair (SL/TL), and superiorly repaired SSC (SSL/TL). At a 10-N mid-deltoid load (initial load), the SU/TL had less abduction than the SL/TL and SSL/TL (P=0.013 and P=0.019, respectively). However, abduction in SU/TL increased steeply as the mid-deltoid load increased. SU/TL had more abduction than SL/TL and SSL/TL at 15 to 20 N of mid-deltoid load. SU/TL showed significantly higher abduction than the SL/TL at mid-deltoid loads of 17.5 and 20 N (P=0.014 and P=0.010, respectively). At 10 and 12.5 N, SU/TL showed less abduction than the SSL/TL (P=0.001 and P=0.009, respectively). However, there were no significant differences from 15 to 20 N, as the mid-deltoid load increased (P=0.18). In a comparison of the SSC repair site (original vs. superior repair), SSL/TL showed more abduction than the SL/TL at all mid-deltoid loads (statistically significant at the 10 to 17.5 N mid-deltoid loads increased at 2.5-N increments; P=0.009, P=0.004, P=0.011, and P=0.023, respectively).
In SU/TL, the internal rotation (IR) of the humerus was minimal (0.6°–8.1°) at all mid-deltoid loads. However, both SL/TL and SSL/TL showed greater IR (47.8°–61.2° and 52.9°–73.5° respectively) than SU/TL (P<0.05). SSL/TL showed more IR than the SL/TL at 10- to 20-N mid-deltoid loads (statistically significant at the 10- to 15-N mid-deltoid loads increased at 2.5-N increments; P=0.024, P=0.03 and P=0.04, respectively) (Fig. 3).
Humeral Abduction and Rotation with Deficient TM
In the simulated deficient TM (TU) scenario, humeral abduction and rotation was tested in three conditions: unrepaired SSC (SU/TU), original SSC repair (SL/TU), and superiorly repaired SSC (SSL/TU). The abduction ranges in the deficient TM and intact TM were similar. SU/TU showed less abduction with the 10- and 12.5-N mid-deltoid loads than the SSL/TU (P=0.001 and P=0.048, respectively). Humeral abduction of SU/TU increased steeply as the mid-deltoid load increased. A significant difference was observed at 17.5- and 20-N mid-deltoid loads of SL/TU and 17.5 N of SSL/TU. In the SSC repair scenario (original vs. superior repair), SSL/TU showed more humeral abduction than the SL/TU for all mid-deltoid loads (P<0.05).
The IR was lower in SU/TU (range, 25.0°–31.8°) than in SL/TU (range, 51.0°–72.6°) and SSL/TU (range, 60.8°–85.5°) for 10- to 17.5-N mid-deltoid loads (P<0.05). SSL/TU showed higher IR than SL/TU in 10-, 12.5-, and 15-N mid-deltoid loads (P=0.001, P=0.001, and P=0.001, respectively). SSL/TU showed IR exceeding 80° at 10- to 15-N mid-deltoid loads and SL/TU exhibited IR exceeding 70° at 10- to 12.5-N mid-deltoid loads (Fig. 4). Table 1 presents data on humeral abduction and rotation under different conditions.
DISCUSSION
In our study, we found that SSC repairs (using both the original footprint and superiorly migrated insertion) limited glenohumeral abduction and external rotation while aiding humeral IR in lateralized RTSA. Greater glenohumeral abduction and IR was seen with superior SSC repair than with the original SSC footprint repair, irrespective of posterior rotator cuff status (intact or deficient TM). SSC repairs (both original footprint and superiorly migrated insertion) led to excessive humeral IR, especially with a deficient posterior rotator cuff. The most favorable movement arc (increased abduction and reduced resistance to external rotation) was demonstrated by unrepaired SSC with an intact TM.
Changing RTSA designs from medialized to lateralized implants to avoid scapular notching and improve overall arc of motion, especially external rotation, has also improved stability [3,4,11,12,16,17,20,21,38]. Concomitant SSC repair in RTSA may have increased dislocation rates, if the SSC is not repaired, especially in medialized designs [10]. In our study, the dislocation rates were similar to those of previous studies, suggesting that SSC repair is not critical to stability when using lateralized designs [31,32,34]. Nevertheless, the benefits of performing SSC repair during RTSA are still being debated, especially in lateralized RTSA.
The SSC is a biphasic muscle and contracting the upper portion of the SSC causes abduction, while contracting the lower portion causes adduction on initiating abduction [44-48]. Downward migration of the humerus after RTSA causes the original footprint of rotator cuff insertion (i.e., the lesser tuberosity) to move downward, below the COR [38,39]. Accordingly, the force vector of the SSC becomes steeper. Much of the SSC is involved in adduction as a negative moment arm can generate antagonistic action on the deltoid during abduction.
Routman [49] reported that at low levels of abduction, the SSC acts as an adductor antagonistically against the deltoid, with a 132% increase in deltoid force being required for abduction when the SSC was intact versus released. Hansen et al. [37] performed a cadaveric biomechanical study and demonstrated that concomitant SSC repair was biomechanically unfavorable in medialized and lateralized RTSA designs. Therefore, SSC repair is sometimes performed superiorly to the original footprint to recover the preoperative vector of the SSC, similar to the original SSC.
A few studies have focused on the SSC repair site. King et al. [39] described superior SSC repair at the greater tuberosity, referred to it as over-the-top SSC repair. They used a 3D computer model for muscle moment arm analysis after RTSA and compared the abduction moment arm length of the SSC tendon for original footprint repair and over-the-top repair with three different RTSA designs. They demonstrated that the abduction moment arm of the SSC was better in over-the-top SSC repair than that with original footprint repair, irrespective of the medialized or lateralized RTSA design [39]. Moreover, lower SSC tenotomy is beneficial for the abduction moment arm because the lower portion of the SSC always functions as an adductor during glenohumeral abduction [39]. Eno et al. [38] also investigated the moment arms and muscle lengths of the SSC after superior repair during RTSA using a 3D-biomechanical computer shoulder model. They investigated only the lateralized glenoid design and found that superior SSC repair reduced the adductive moment arm of the SSC and resulted in lower tension at the repair site compared with repair at the original footprint [38].
Our study showed increased abduction and IR in superior repair of SSC, consistent with the findings of prior studies. In conditions with TM deficiency, SSC repairs caused excessive humeral IR of 51.0°–72.6° in original SSC footprint repairs and 60.8°–85.5° in superior SSC repairs, whereas unrepaired SSC showed 25.0°–31.8° IR for all mid-deltoid loads.
The main concerns regarding postoperative range of motion (ROM) are weakness in external rotation and limitation in internal rotation. SSC repair could improve IR force. However, it remains unclear whether SSC repair can resolve the limitation of IR at the terminal range caused by implant-bone impingement, tensioning during reduction, and restricted excursion of the residual posterior cuff muscle postoperative. Increased IR force might resist external rotation. External rotation is frequently used and essential in daily living, and associated deficits affect clinical outcomes and patient satisfaction following RTSA. However, in the comparison between the original repair and the superior repair, the differences in abduction and internal rotation (IR) were small, it might be difficult to prove the significant difference between two groups in clinical setting.
This study has several limitations. First, since this was a cadaveric biomechanical study, the load and vector of the rotator cuff and the scapulo-thoracic muscle cannot accurately represent the real human shoulder. The vector of muscles can generate different degrees of force during dynamic active motion, and thus the cross-surface method cannot adequately simulate real action of the muscles under dynamic circumstances. The central tagging suture at the point of tendon insertion could have potentially changed the muscle vector, since muscles typically have a continuous and broad insertion. Additionally, the retracted SSC tendon and lateralized RTSA design may increase tension on the repaired SSC tendon during surgery, which could not be reproduced in this experiment. Thus, the clinical applications of this study are limited. Second, humeral impingement was observed between the greater tubercle and acromion during terminal abduction due to the absence of scapular motion. Humeral implant-glenoid neck impingement was observed at the terminal range of IR (exceeding 80°), although we used a lateralized implant and created an inferior glenosphere overhang. During the increase of mid-deltoid load, impingement occurred, and no further angular increase was observed thereafter, so additional loading did not contribute to obtaining meaningful experimental results. Third, we did not measure humeral flexion/extension even though it was observed during rotation. Fourth, the small sample size and the use of a single implant design may further impact the applicability of the results to other implant configurations.
Our study’s strength is that it is the first cadaveric biomechanical study to reveal the effects of the SSC repair options (repair on the native footprint, superiorly migrated insertion, or none) in lateralized RTSA. Our findings also describe the indications for SSC repairs for massive rotator cuff tears with TM deficiency.
In conclusion, concomitant SSC repair in lateralized RTSA decreases glenohumeral abduction and increases IR. Superior SSC repair to restore the native vector of the SSC slightly increases humeral abduction and IR compared with original SSC footprint repair. However, increased IR might cause external rotation difficulty, a major complaint of patients undergoing RTSA. SSC repair might also cause excessive humeral IR in massive rotator cuff tears involving TM deficiency. Concomitant SSC repair on the original and superiorly migrated footprints in lateralized RTSA should be carefully considered, especially when the posterior rotator cuff is deficient.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
