The Influence of Different Rotator Cuff Deficiencies on Shoulder Stability Following Reverse Shoulder Arthroplasty

Abstract

Background:

The primary indication for reverse shoulder arthroplasty (RSA) is rotator cuff arthropathy caused by a deficient rotator cuff. Cuff deficiency in patients is highly variable in its distribution and extent, with mechanical implications that may significantly affect post-operative recovery. This study investigated the effects of variable cuff deficiency on the propensity for impingement between the scapula and humeral component and resulting subluxation, the source of two common complications (scapular notching and instability).

Methods:

Five different finite element models of an RSA were analyzed with varying degrees of rotator cuff deficiency: (1) baseline, with intact subscapularis, infraspinatus and teres minor, (2) no subscapularis, (3) no subscapularis or infraspinatus, (4) no infraspinatus, and (5) no infraspinatus or teres minor. The supraspinatus was not included in any models, as it is absent in rotator cuff arthropathy. Each model was moved through a prescribed arc of 45° internal/ external rotation originating from neutral.

Results:

Greater rotator cuff deficiency was associated with more impingement and larger magnitudes of subluxation. The largest subluxation (7.5 mm) and highest impingement-related contact stress (479 MPa) was in the model lacking all rotator cuff muscle groups. Posterior subluxation was present in most models lacking the infraspinatus, while anterior subluxation was present in all models lacking the subscapularis.

Conclusions:

This study helps clarify how different rotator cuff deficiencies influence shoulder stability following RSA and can ultimately help predict which patients may be at greater risk for impingement-related scapular notching and subluxation.

Clinical Relevance:

Surgeons should carefully consider the nature of the rotator cuff deficiency and its influence on impingement and instability when planning for RSA.

Level of Evidence: V

Introduction

Since its FDA approval in 2003, reverse shoulder arthroplasty (RSA) has continually increased in popularity, accounting for 33% of all shoulder arthroplasties in 2007.^1,2 RSA reverses the ball-in-socket-design of the native shoulder by implanting a hemispherical component onto the glenoid surface and a cup/ stem component into the humerus. By moving the center of rotation distally and medially, a larger moment arm is created, enabling the deltoid muscle to move the humerus with increased efficiency and stability.^3,4 RSA is performed primarily to restore pain-free function and range of motion for patients with rotator cuff arthropathy, shoulder arthritis resulting from rotator cuff deficiency and characterized by degenerative changes of the glenohumeral joint, superior migration of the humeral head, and often severe pain and disability.^5-8 Based on its success in relieving pain and restoring function, the indications for RSA have expanded to include younger populations with rotator cuff deficiency, arthritis, or previously failed shoulder implants.^9-17

Despite the initial success associated with RSA, complications such as instability are common.^18,19 Instability, accounting for 38% of all complications, is due to an unbalanced force coupling associated with rotator cuff deficiency and often accompanies impingement.^20,21 Reduced control of motion associated with rotator cuff deficiencies leads to less predictable motion, increasing the risk of dislocation which can ultimately necessitate revision surgery.

The nature of the cuff deficiency varies considerably across rotator cuff arthropathy patients, with implications for shoulder stability and function after RSA. It is not simply a matter of whether a patient has muscle deficiency, since each patient presents with a variable amount of intact cuff tissue. Numerous factors such as which muscles are deficient/intact, quality of muscle/ tendon tissue, and strength of a given muscle may impact the overall stability and function of the shoulder before and after implantation. Orthopedic surgeons currently rely on their best judgement to accommodate these variable cuff deficiencies when deciding on implant positioning and tension.

This study built upon previous work using finite element analysis of RSA to investigate the influence of specific rotator cuff deficiencies on impingement-related contact stress and shoulder stability during a full external/internal rotation motion. With RSA still being relatively new, the association between rotator cuff deficiency and its influence on motion post-operatively can serve to identify the role that each muscle group plays and their contribution to instability and contact stress at the impingement site. Results can also provide guidance on which muscles could benefit most from muscle transfer, and how implants should be positioned based on which muscles are intact. Ultimately, this can provide a better understanding of stability and functionality post-op. We hypothesized that a complete lack of the rotator cuff musculature would produce the most subluxation and highest contact stress values at the impingement site, and that subluxation would occur on the same side of the joint as the cuff deficiency.

Methods

A finite element (FE) model of an RSA-implanted shoulder previously implemented in Abaqus/Explicit 6.14-2 (Dassault Systêmes, Paris, France)^22,23 was modified to simulate varying degrees of rotator cuff deficiency. The scapula was segmented from CT scans of the Visible Human Female (National Library of Medicine; https://www.nlm.nih.gov/research/visible/visible_human.html) using OsiriX software (Pixmeo, Geneva, Switzerland). The FE model incorporated a contemporary RSA implant design (Tornier Aequalis Ascend Flex Reversed; Wright Medical, Memphis, TN) placed according to manufacturer-supplied guidelines. The FE model consisted of a deformable 15 mm section of the lateral-most portion of the scapula, a rigid glenosphere and humeral stem, and a deformable polyethylene insert. All FE meshes were generated using hexahedral meshing software (TrueGrid; XYZ Scientific Applications, Livermore, CA).

Points representing where the lines of action of rotator cuff muscle groups traverse over the bony surfaces were mapped onto the RSA-implanted shoulder model using Geomagic Studio software ( 3D Systems, Rock Hill, USA). The surface geometry of the implant and scapular bone were combined in Geomagic by a shoulder surgeon specializing in RSA to simulate the correct positioning of the device.^24,25 Slipring elements were used in the FE model to produce cable and pulley systems that approximated the lines of action of individual rotator cuff muscle groups, with in-line stiffness values taken from prior studies to represent the associated active resistance to elongation for each of the muscle groups (Figure 1).^24,25 The deltoid was also modeled using slipring elements, with in-line stiffness values taken from prior work and the proximal end held fixed against displacement to support loading while undergoing motion.^26,27 Point of reference locations on the ends of the muscles and the implant were chosen using a combination of geometric calculations to determine spacing and location, based upon the anatomy of the human shoulder, to ensure contact points accurately portrayed the location of each muscle group. Slipring elements were distributed evenly within each muscle, dividing the outermost edges of the area into lines of action with equal spacing to one another for each muscle group. The numbers of elements used for each muscle group were based on their cross-sectional area: 5 elements for the subscapularis, 4 for the infraspinatus, and 2 for the teres minor. The insertions of all muscles were tied to the humerus, while their origins were fixed on the scapula.

Figure 1.

The rotator cuff muscles were modeled using slipring elements. Each set of colored slipring elements represented a different rotator cuff muscle group.

Five different models with varying degrees of rotator cuff deficiency were created. Because the vast majority of rotator cuff arthropathy patients lack the supraspinatus, it was not included in any model. A baseline model included all muscles except the supraspinatus, while four additional models were created using varying combinations of intact muscle tendon units: no subscapularis (No Sub), no subscapularis or infraspinatus (No Sub or Inf), no infraspinatus (No Inf), and no infraspinatus or teres minor (No Inf or TM).

An internal/external rotation motion was simulated, as these motions have been shown to produce the most impingement/subluxation and would likely be the most affected A by muscle deficiencies.²⁸ Prior to simulating any arm motion, the sliprings were pre-tensioned by pulling the medial-most ends to represent in-vivo active muscle tensioning. Sliprings representing deltoid and rotator cuff musculature were kept at this position throughout the movement. A 40 N load was applied to the distal end of the humerus to represent the weight of an arm.²² The motions prescribed for the FE model simulated an external/internal rotation with the shoulder abducted 25° from the neutral position. Humeral rotation was defined about its long axis. The model was rotated from neutral to 45° of external rotation, and then back through the neutral position to 45° of internal rotation.

Subluxation of the humeral component was defined by a vector connecting the normally concentric centers of rotation of the humeral polyethylene liner and the glenosphere. The magnitude and direction of subluxation were tracked, with decomposition into antero-posterior, supero-inferior, and medio-lateral displacements. The location of any impingement and the maximum impingement-related contact stress were recorded.

Results

The nature of the impingements and subluxations varied with the type of rotator cuff deficiency modeled (Table 1). Whereas a posterior subluxation of 4 mm was observed in most models lacking the infraspinatus, an anterior subluxation of 4 mm was observed in all models lacking the subscapularis. All models had inferior subluxation, except for the baseline model, which had 0.7 mm superior subluxation. The No Inf or TM model and No Inf model contained the highest magnitudes of supero-inferior subluxation at 1 mm and 0.9 mm inferior subluxation, respectively. No Sub or Inf model exhibited 0.6 mm of inferior subluxation, while No Sub had 5 mm of inferior subluxation. No Sub and No Sub or Inf each contained the highest amount of antero-lateral subluxation with 6.3 mm lateral subluxation. This was followed by Baseline with 4.5 mm lateral subluxation, No Inf or TM with 3.3 mm lateral subluxation, and No Inf with 3.3 mm lateral subluxation. Larger magnitudes of subluxation were seen in all models lacking a subscapularis, while minimal changes in subluxation were seen whether the teres minor was present or not. Elevated amounts of subluxation were found with increased rotator cuff deficiency, with the No Sub or Inf model and No Sub model both having a total subluxation of 7.5 mm. The baseline model had a total subluxation of 6 mm, followed by the No Inf or TM model with 5.3 mm and then the No Inf model at 5.2 mm. The model lacking all rotator cuff musculature experienced the most subluxation and greatest impingement-related contact stress.

Table 1.

Subluxations Present Through 45° of External/Internal Rotation from the Neutral Position for Five Models Varying in Cuff Deficiency

Cuff Deficiency	Antero-Posterior Sublux. (mm)	Supero-Inferior Sublux. (mm)	Medio-Lateral Sublux. (mm)	Total Sublux. at 45 ° (mm)
Baseline (No Sup)	-3.8	0.7	-4.5	6.0
No Sub	4.0	-0.5	-6.3	7.5
No Inf	-4.0	-0.9	-3.3	5.3
No Sub or Inf	4.0	-0.6	-6.3	7.5
No Inf or TM	-4.0	-1.0	-3.3	5.3

Sublux. = subluxations, No Sub = No Subscapularis, No Inf = No Infraspinatus, TM = Teres Minor

Impingement of the poly insert on the scapula in models lacking only posterior muscles occurred medially relative to the baseline model lacking only the supraspinatus, while for all models lacking anterior musculature impingement sides were located laterally to the baseline model (Figure 2). Absence of the supraspinatus, subscapularis, and infraspinatus produced a maximum contact stress of 479 MPa. Absence of the supraspinatus and subscapularis gave a maximum stress of 474 MPa, absence of supraspinatus and infraspinatus gave a maximum stress of 340 MPa, absence of the supraspinatus, infraspinatus and teres minor gave a maximum stress of 265 MPa, and the absence of only the supraspinatus gave a maximum stress of 260 MPa.

Figure 2.

Inferior glenoid view of impingement-related contact stress during external rotation plotted on a single scapula for all models. The No Sub model impinged in the same region as the No Sub or Inf model, while the No Inf or TM model impinged in the same region as the No Inf model. No Sub = No Subscapularis, No Inf = No Infraspinatus, TM = Teres Minor

Discussion

Rotator cuff arthropathy is the primary indication for RSA, yet the mechanical effects of rotator cuff deficiency in the context of RSA are still not fully understood. This study aimed to build upon previous FE studies^22,23,29 to examine the effects of variable rotator cuff deficiencies on subluxation and impingement after RSA. As expected, increasing cuff deficiency led to greater amounts of subluxation. Integrity of the subscapularis had the greatest influence on subluxation, whereas the teres minor had the least effect. These findings aligned with data from prior studies indicating that the subscapularis is the primary contributor to internal rotation, while the teres minor played a minor role in that movement.^30-34 Contrary to our expectation, anterior muscle deficiency was associated with anterior subluxation and posterior deficiency was associated with posterior subluxation. This is most likely due to the presence of the muscle, lacking compressive force in this direction, blocking the subluxation path as it moves closer to the muscle group. Differences between subluxation found with anterior versus posterior muscle deficiency can be attributed to the relative control of specific rotation directions by certain muscles as well as this blocking characteristic.

The FE analyses also showed that absence of the subscapularis produced higher, more focal contact stress at the impingement site as opposed to the baseline model (lacking only a supraspinatus). Elevated amounts of stress found in this study ranging from 260 MPa to 479 MPa, which would likely result in high wear rates at the impingement site. As a point of reference, the compressive yield stress of polyethylene is on the order of 15 MPa,³⁵ suggesting that the polyethylene component could be damaged by this impingement.

Internal rotation is primarily controlled by the subscapularis; hence, absence of the primary muscle involved with internal rotation movement might be expected to create larger amounts of stress. The sites of impingement for models without a subscapularis were located superiorly to those with an intact subscapularis. The smallest contact stress values were observed in the baseline model. This supported the consensus that the subscapularis, infraspinatus, and teres minor play a role in internal/ external motion.^30-34

Several factors limited the scope of this study. Mapping the rotator cuff musculature in the models required some degree of subjectivity due to the lines of action being placed along the geometry of the bone. This may have produced minor discrepancies in the direction of the pulling force and location of the coordinates. Use of only one implant geometry and scapular surface precluded any study of possible influences on scapular contact stress and shoulder subluxation owing to individual variation in geometry. Mechanical properties of each muscle group were taken from prior studies using intact shoulders from subjects aged 58 to 98 years, which may not be the case for all rotator cuff arthropathy patients. The modeling approach also does not account for fatty and epithelial tissue surrounding the shoulder, or pain experienced during major subluxations, both of which might limit the amount of subluxation experienced by patients. Forcing the shoulder to rotate through a prescribed range of motion with no regard for the amount of muscle force that would be required to do so could result in impingement-related contact stresses and accompanying subluxations greater than would actually be experienced. Variation in soft tissue composition, musculature, and the specific nature of the cuff tears may also affect the results between individuals.

This study determined the influence of variable rotator cuff deficiency on shoulder stability following RSA. Future studies may include physical testing of material properties in each rotator cuff muscle for validation. Testing the models through different types of motion other than internal/ external rotation may provide further insight as to how the muscle deficiencies affect subluxation and contact stress. The results of this study can aid in understanding how different rotator cuff deficiencies influence shoulder stability following reverse shoulder arthroplasty, and the results can ultimately help predict which patients may be at greater risk for dislocation and impingement-related subluxation.