22.1 Introduction

Reverse shoulder arthroplasty (RSA) can provide excellent function to patients with a spectrum of glenohumeral pathology; however, implant positioning remains a potential challenge of this nonanatomic prosthesis. Glenosphere malposition has been associated with a variety of complications including impingement, scapular notching, dislocation, component loosening, and early mechanical failure. Scapular notching is perhaps the most well-studied complication of RSA. Notching has a reported prevalence of 0 to 96%, depending on implant positioning and design, and signifies impingement of the humeral cup against the scapular neck during arm adduction and/or rotation of the humerus. Although notching is often clinically insignificant, it may represent suboptimal implant placement. Multiple studies have shown that glenosphere position relative to the scapula directly influences the incidence of notching.1,2,3,4,5,6

The glenosphere defines the center of rotation and biomechanical properties of the new joint. Clinical outcome may be influenced by both the design attributes of the selected implant system and the position of the glenosphere after surgical implantation. Design features such as increased glenosphere lateralization or a more varus humeral component neck-shaft angle have been associated with increased range of motion.1 The position of the glenosphere is determined by surgical technique. Positioning is straightforward in patients with minimal glenoid bone loss and a scapular neck of at least 1cm, but becomes more challenging in patients with significant glenoid bone loss or reduced scapular neck length (► Fig. 22.1). Preoperative planning with three-dimensional (3D) computed tomography (CT) imaging and virtual implants is especially helpful in these patients. Templating helps the surgeon choose the correct implant size and determine the optimal placement of the glenosphere relative to the scapula. Templating can also help the surgeon decide whether bone graft or patient-specific implant design features will help facilitate glenosphere placement. Patient-specific instrumentation (PSI) offers further customization to help the surgeon transfer information from the preoperative planning software to the patient.

This chapter is divided into three sections. The first section discusses the principles behind ideal implant placement. Emphasis is placed on how to position the glenosphere relative to the glenoid and scapula to optimize impingement-free range of motion and implant stability. Different implant design features that assist with glenosphere placement in specific circumstances are also discussed. The second section discusses how variations in scapular anatomy and glenoid bone loss affect surgical decision-making. This section highlights the utility of 3D preoperative planning software and the glenoid vault model to assist with implant placement. The third section discusses the role of PSI as a way to improve execution of the preoperative plan. This section reviews the current literature on PSI for anatomic and RSA. Several commercially available systems for 3D templating and PSI are also discussed.

22.2 Principles in Glenosphere Placement

The glenosphere and baseplate may be positioned in six degrees of freedom relative to the glenoid and scapula. This includes superior-inferior translation, superior–inferior tilt (inclination), anteroposterior translation, anteroposterior tilt (version), mediolateral translation (offset), and rotation (► Fig. 22.2). The location and orientation of the glenosphere directly influence glenohumeral range of motion, stability, implant impingement, and scapular notching. Scapular notching signifies impingement of the humeral cup against the scapular neck during arm adduction and/or rotation of the humerus. Ideally, the glenosphere should be placed in the location that maximizes impingement-free range of motion and stability while avoiding notching. This location is ultimately determined by each patient’s individual anatomy; however, there are several general principles in placing the glenosphere. Ideal placement usually includes inferior translation with neutral or inferior tilt relative to the scapula centerline, which is defined as a line between the center of the glenoid and the scapula trigonum. In contrast, superior tilt, superior translation, and excessive medialization of the glenosphere should be avoided. Gutiérrez et al developed a computational model to assess the hierarchy of surgical factors that affect motion after RSA. That study identified increased glenosphere lateral offset and inferior translation as the two most important factors to maximize abduction. Those factors along with a more varus humeral–neck shaft angle were also associated with superior adduction.1

The superior–inferior location of the glenosphere is determined by placement of the guide pin, which later becomes the location of the center of the of the glenoid baseplate. Multiple studies have shown that inferior placement of the glenosphere results in increased range of motion, and reduced scapular notching.3,5,6 Ideally, the guide pin should be placed in the location that will allow the glenosphere to slightly overhang the glenoid rim. Poon et al found that inferior glenosphere overhang > 3.5 mm prevented notching in a randomized controlled trial of 50 patients receiving concentric or eccentric glenospheres.4 The distance between the central peg/screw of the baseplate and the inferior glenoid rim has been defined as the peg-glenoid rim distance (PGRD).5 It is important to anticipate how this distance will change after reaming. PGRD lengthens proportionally to the amount of bone removed during reaming and is affected by the angle of the scapular neck (► Fig. 22.3). Patients with a short scapular neck that requires asymmetric inferior reaming are particularly prone to this issue. Asymmetric inferior reaming is a common strategy to create a flat surface for the baseplate in patients with superior glenoid bone loss due to cuff-tear arthropathy. Guide pin placement must also account for the dimensions of the glenoid vault so that the central peg of the baseplate does not perforate the medial wall. Vault perforation may destabilize the baseplate and is a risk factor for early failure. Different manufacturers offer different design features to avoid perforation while facilitating inferior offset of the glenosphere. Eccentricity may be built into the baseplate (Exactech) or glenosphere (Arthrex, Depuy, Exactech, Lima, Tornier) in order to allow adjustments in glenosphere position relative to the central peg of the baseplate.

Guide pin placement also determines glenosphere tilt. Superior–inferior tilt (inclination) is defined relative to the coronal axis of the scapula. This axis extends along a line from the scapular trigonum to the center of the glenoid, the scapula centerline. Anterior–posterior tilt (version) is defined relative to the transverse axis of the scapula. This axis also extends from the scapular trigonum to the center of the glenoid. The glenosphere should be positioned in neutral or slightly inferior tilt. Li et al performed a computer simulation study of the Biomet Comprehensive RTSA (reverse total shoulder arthroplasty) System that examined the effect of inferior tilt on internal and external rotation across all degrees of scaption. The authors found that both 15 and 30 degrees of inferior tilt produced increased range of motion when compared with neutral tilt, but that 30 degrees of inferior tilt limited external rotation at 60 degrees of scaption.6 Others have suggested that the “ideal” amount of inferior tilt is 10 degrees relative to the coronal axis of the scapula.7,8,9,10 The optimal amount of anteroposterior tilt (version) is less well understood. Glenosphere version and humeral component version likely have a complementary effect in determining the amount of glenohumeral internal and external rotation. Several authors have defined the “ideal” glenosphere version as 0 degrees relative to the transverse axis of the scapula.8,9,10 In retroverted glenoids, this may require significant anterior reaming. We prefer to use posterior glenoid bone grafting to avoid anterior reaming. When there is minimal glenoid bone loss, the surgeon may use the plane of the glenoid fossa as a reference to place the guide pin in the correct amount of tilt. However, this becomes increasingly difficult with more significant glenoid bone loss. In the study by Verborgt et al, standard instrumentation without advanced imaging resulted in an error range of 16 degrees for glenosphere inclination and 12 degrees for glenosphere version.10

The ideal mediolateral glenosphere offset is an area of active debate. Paul Grammont’s design places the center of rotation medial to the glenoid baseplate and includes a humeral component with a valgus neck-shaft angle around 155 degrees. These features are believed to enhance implant stability and decrease shear stress across the glenoid baseplate; however, they have also been associated with reduced range of motion and increased scapular notching. Alternative designs lateralize the center of rotation and include a humeral component with a more anatomic neck-shaft angle.1,6 Grammont-style implants work best in patients with minimal glenoid bone loss and a long scapular neck. Conversely, these implants can be problematic in patients with severe glenoid bone loss or a congenitally short scapular neck. The Grammont design is predisposed to excessive medialization in this scenario. This may result in reduced range of motion, impingement, and scapular notching.6 Excessive medialization also shortens the remaining rotator cuff and potentially weakens external rotation strength. Therefore, it is the authors’ practice, when using a Grammont-style implant, to lateralize the glenoid baseplate with a bone graft taken from the resected humeral head. The baseplate is ideally positioned at the premorbid joint line and the scapular neck length is restored to at least 1 cm. 3D preoperative imaging, application of the vault model to define premorbid anatomy, and implant templating are critical for determining when to use a graft and the size of graft required (► Fig. 22.4, ► Fig. 22.5, ► Fig. 22.6).

A variety of manufacturers offer options to adjust the amount of lateral glenosphere offset. The Zimmer Trabecular Metal Reverse Shoulder System is a Grammont-style system that offers 2.5- and 4.5-mm-width baseplates, which can be used to compensate for central glenoid bone loss or mild joint line medialization. The Exactech Equinoxe is another Grammont system that offers baseplates with superior or posterior augments for these common areas of bone loss. The Tornier Aequalis offers a 10-degree tilted glenosphere option that compensates for superior glenoid wear and lateralizes the center of rotation within the baseplate. The DJO AltiVate reverse system offers a wide range of glenosphere designs that lateralize the center of rotation from 2 to 10 mm. There is currently no consensus on the ideal amount of lateralization. Increasing amounts of lateralization result in increased range of motion, reduced scapular notching, and improved shoulder strength, but expose the glenosphere fixation to more torque.1 Technological advances in baseplate fixation such as locking screws, a central compression screw, and porous ingrowth coating appear to have been successful in reducing the high failure rate associated with early lateralized designs. 3D imaging and templating can help maximize impingement-free range of motion and function with any implant design by defining the optimal location for the implant relative to the patient’s specific anatomy.

22.3 Preoperative Planning

Scapular anatomy has an important role in glenosphere positioning, and implant choice. As previously discussed, scapular neck length can change how inferior glenoid reaming affects PGRD. The surgeon can adjust for this by placing the central pin lower on the glenoid or using an eccentric baseplate or glenosphere. Even when the glenosphere is adequately inferior, short scapular neck length has been shown to correlate with increased notching. For this reason, Paisley et al suggested lateralizing the glenosphere in patients with a scapular neck length less than 9 mm measured on a true anteroposterior radiograph.2 Glenoid bone loss has the same functional consequences. Asymmetric bone loss commonly occurs posteriorly in primary glenohumeral arthritis or superiorly in rotator cuff arthropathy. This may be addressed with asymmetric reaming or glenoid bone grafting in order to create a flat surface for the glenoid baseplate. The same functional results can be accomplished with different surgical strategies provided that the glenosphere ends up in the same location relative to the scapula. For instance, the surgeon may choose to address superior glenoid bone loss with bone grafting and a Grammont-style implant or with asymmetric inferior reaming and a lateralized glenosphere.

3D CT imaging and implant templating provide superior accuracy over 2D CT imaging for quantifying glenoid bone loss and guiding surgical decision-making.11,12,13 3D CT enables the surgeon to define the planes of the glenoid and scapula. This allows more precise measurement of glenoid version and inclination and eliminates measurement error due to the plane of image acquisition (gantry angle). The glenoid plane is the surgeon’s primary reference for placing the guide pin at the time of surgery, but the scapular plane defines glenosphere position. Preoperative planning with 3D imaging enables the surgeon to understand how the glenoid plane and scapular plane are related. 3D imaging also provides useful information about the surface anatomy of the glenoid. Surface irregularities such as indentations and osteophytes seen on 3D images can be used as a map for placing standard or patient-specific pin guides in the correct superior–inferior and anteroposterior locations. These surface features are also helpful for determining the proper rotational orientation of certain patient-specific pin guides (► Fig. 22.7).

The internal shape of the normal glenoid vault is highly consistent across individuals. A virtual model of this shape has been created along with rules to place this model into a pathologic glenoid. This method can be used with 3D planning to estimate native glenoid version, inclination, and joint line position in the setting of glenoid bone loss.14,15,16,17 Sizing and placing the vault model requires adjustments in three planes. In the axial and sagittal planes, the vault is aligned with the inner cortical margin of the anterior wall, and in the coronal plane, the vault is aligned with the inner cortical margin of the superior wall at the suprascapular notch. When the vault model is positioned correctly, its medial portion will closely follow the preserved 3D internal architecture of the patient’s medial glenoid. In pathologic shoulders, the lateral portion of the vault model represents the anatomy of the premorbid glenoid and defines where bone loss has occurred (► Fig. 22.4, ► Fig. 22.5, ► Fig. 22.6).

The vault model is a helpful tool for determining surgical strategy and implant placement in RSA. It is the authors’ practice to place the glenoid baseplate in the mediolateral location represented by the vault model. This is especially important when using a Grammont-style implant, where excessive medialization results in reduced motion and increased impingement. Bone loss may be addressed by bone grafting the glenoid or selecting an implant system with a lateralized glenosphere design. Ultimately, all designs become limited when severe bone loss makes the glenoid vault too shallow to accommodate the baseplate peg or central screw without perforation.

Virtual templating allows the surgeon to simulate different RSA components on a 3D reconstruction of the patient’s scapula. This helps the surgeon determine the optimal implant design and location, with awareness that glenosphere size may need to be adjusted at the time of surgery based on soft-tissue tension. After placing the baseplate, the software calculates the location and orientation of the central guide pin. Some software allows the surgeon to perform simulated reaming to plan reaming depth and location. After reaming, the surgeon confirms that the baseplate is sufficiently inferior on the glenoid. If it is not, the implant is repositioned and these steps are repeated. Occasionally the guide pin may need to be positioned slightly anterior or posterior to the center of the glenoid in order to ensure that the central peg of the baseplate is contained within the glenoid vault. Eccentric glenosphere and baseplate designs offer additional flexibility to choose a contained location for the central peg/screw while positioning the glenosphere in the desired location. When the vault is too shallow to contain the central peg in any location, central glenoid bone grafting is required. The width of the bone graft can be calculated by determining how much bone loss has occurred relative to the vault model.

Virtual templating has been shown to improve the accuracy of guide pin position and orientation using standard instrumentation. Iannotti et al compared standard and PSI with 3D templating to standard instrumentation with 2D imaging in placing the guide pin for anatomic TSA in nine bone models from patients with a variety of glenohumeral arthritis. They found that 3D templating with standard instrumentation improved the accuracy of pin position by 4.5 ± 1.0 degrees in version, 3.3 ± 1.3 degrees in inclination, and 0.4 ± 0.2 mm in location.18 The authors followed this study with a randomized clinical trial of 46 patients undergoing anatomic TSA. They again demonstrated that 3D templating with standard instrumentation or PSI was superior to standard instrumentation with 2D imaging. 3D templating significantly improved the accuracy of implant placement to within 5 degrees of desired inclination or 10 degrees of version.11