Lateralized versus nonlateralized reverse total shoulder arthroplasty

Abstract

Throughout the history of reverse total shoulder arthroplasty, the extent of lateral offset has changed considerably from “too lateral” to “too medial” and has been lately swinging back towards a point somewhere in between. Nonlateralized designs minimize shear forces on the glenoid and decrease force required by the deltoid. Glenoid lateralization decreases impingement and scapular notching and improves range of motion. Humeral lateralization achieves a more anatomic position of the tuberosities while maintaining a nonlateralized center of rotation. Several factors play a role in choosing the extent of lateral offset and method of lateralization.

Introduction

The principles associated with reverse total shoulder arthroplasty (RTSA) are being continuously modified in order to optimize performance and minimize complications. The purpose of this review is to summarize the biomechanical concepts of lateralized and nonlateralized RTSA and outline the clinical outcomes of each.

Definitions

In the native shoulder, the lateral humeral offset is the distance between the base of the coracoid and the most lateral point of the greater tuberosity.^1–3 Harman et al.⁴ defined the lateral offset in RTSA as the distance between the baseplate and the center of articular contact between the glenosphere and the polyethylene cup. The latter method measures the distance between different parts of the prosthesis, so it helps compare different RTSA designs. However, it does not consider humeral component influence on the lateral offset. Moreover, it does not measure the distance between bony landmarks which is important to assess changes in muscle tension. Valenti et al.⁵ defined the lateral offset in RTSA as the distance between the center of rotation (COR) and the greater tuberosity.

Werthel et al.⁶ provided a detailed description of different components of the total lateral offset, which is believed to be the most accurate in the literature. However, their definitions still do not account for differences in the amount of glenoid erosion and/or reaming. The measurements they described included humeral stem offset, humeral insert offset, perceived radius of the glenosphere, and COR offset. Humeral stem offset is the horizontal distance between the humeral stem diaphysis (point A) and the midpoint of the humeral implant bearing surface at the level of the humeral cut (point B). Humeral insert offset is the horizontal distance between point B and the deepest point of the articular surface of the humeral insert (point C). Perceived radius of the glenosphere is the distance between point C and the COR (point D). The COR offset, as defined by Werthel et al.,⁶ is the distance between the COR (point D) and the glenoid-baseplate interface (point E) (Figure 1). We suggest redefining the COR offset to be the distance between the COR and the native glenoid after glenoid reaming (point E) in order to incorporate the increase in lateral offset achieved by bony increased offset implants. Sum of humeral and glenoid lateralization is the global implant lateral offset. The greater tuberosity lateral offset is the horizontal distance between point E and the lateral-most point of the greater tuberosity with the shoulder at 0 degrees of abduction. There are large variations in current RTSA designs in terms of means and extent of lateralization. Total lateral offset can range from 13.1 mm (Delta III, Depuy International Ltd, Leeds, UK) to 50.7 mm (maximal lateral offset in Biomet Comprehensive RTSA (Biomet, Warsaw, IN, USA)).⁶

Figure 1.

Components of the lateral offset in RTSA. Republished from Werthel et al.⁶ with permission from Springer Nature.

Boutsiadis et al.⁷ described the lateralization shoulder angle (LSA) for postoperative radiographic evaluation of the lateral offset. On anteroposterior views, the LSA vertex is at the most lateral border of the acromion and the two rays of the angle pass through the superior glenoid tubercle medially and the most lateral border of the greater tuberosity laterally. In their study, higher LSA correlated with increased implant lateralization. Routman et al.⁸ considered a 5 mm COR offset the cutoff to classify glenoid components into medialized and lateralized glenoids and 15 mm humeral offset to categorize humeral components to medialized and lateralized designs.

In this review, as with most reports in the current literature, the term “lateralization” is used to describe any implant with a lateral offset greater than that of the Delta III prosthesis. Despite failure of the early lateralized RTSA implants, lateralization has regained much attention in the past few years owing to its potential advantages.

Methods of lateralization

Implants that follow Grammont’s principles are nonlateralized on both glenoid and humeral sides. Most implants in current use allow lateralization on the glenoid and/or the humerus.⁶ Factors affecting the lateral offset are summarized in Table 1.

Table 1.

Factors affecting the lateral offset in RTSA.

Glenoid and glenoid component factors	Humerus and humeral component factors
Pre-existing glenoid erosion Extent of glenoid reaming Autograft bone spacers Thickness of baseplate Shape of glenosphere Glenosphere size (radius of curvature)	Thickness of polyethylene insert Level of humeral cut Inclination angle of humeral cut (neck shaft angle) Humeral bearing (inlay vs eccentric onlay) Humeral stem design (straight vs curved)

Size of glenosphere

Lateral offset increases with increased glenosphere radius. Werthel et al.⁶ demonstrated that increasing the size of the glenosphere to the next available size increases glenoid lateralization by a mean of only 1.14 mm. This is a very limited increase in lateral offset relative to COR lateralization or humeral lateralization. A larger glenosphere has been shown to increase arc ranges of abduction, internal, and external rotation.^9,10 A large glenosphere can introduce difficulties with surgical exposure and carries the potential risk of overstuffing the joint. Given the minimal lateralization that can be achieved and the other variables that are more sensitive to upsizing the glenosphere than the lateral offset, glenosphere size is rarely used as a method for lateralization. Glenosphere size is the only method of glenoid lateralization that does not lateralize the COR. Therefore, glenoid lateralization is often referred to as COR lateralization.

Center of rotation

Unlike the native shoulder, the COR in RTSA is fixed due to increased constraint and matched radii of curvature. This creates a combination of compression and shear forces passing through the fixed COR to the glenoid-implant interface. Maximizing the compressive forces and minimizing the shear forces is encouraged for the durability of glenoid component fixation. Although a lateralized COR improves rotational movements,¹¹ the lever arm through which destabilizing and shear forces act on the glenoid-implant interface increases, resulting in increased torque¹² (Figure 2). The Grammont-style design medializes the COR and therefore converts shear forces into compressive forces.¹³

Figure 2.

(a) Compressive (Fc) and shear (Fs) components of the resultant force vector (Fv), which act through a fixed center of rotation (COR). (b) COR lateralization increases the lever arm length which decreases compressive forces, increases destabilizing shear forces, and creates a new moment (M) at the glenoid-implant interface. Republished from Berliner et al.¹⁴ with permission from Elsevier.

Gutiérrez et al.¹⁵ analyzed the in vitro effect of multiple implant factors on range of motion (ROM). COR offset had the greatest effect. COR can be lateralized either by the implant design or with insertion of bone graft.¹⁶ Lateralization of COR by implant design can be achieved by either glenosphere shape modification¹⁵ or baseplate lateralization.^5,17 Bony lateralization of COR was developed with the hypothesis that increasing the length of the scapular neck would lateralize both, COR and glenoid-implant interface. This type of “biologic lateralization” would achieve the advantages of COR lateralization with minimizing the undesired forces generated at the glenoid-implant interface with prosthetic COR lateralization. Bony increased offset (BIO) RTSA was introduced by Boileau et al.¹⁶ in 2011 using Aequalis Reverse Shoulder Prosthesis (Tornier, Saint-Ismier Cedex, France) in which a cylinder of autologous bone graft was harvested from the resected humeral head and placed between the reamed glenoid surface and baseplate. This maintained the COR at the glenoid-implant interface minimizing the torque on the glenoid component.

Bony COR lateralization is not without drawbacks. Denard et al.¹⁸ compared stress and displacement in bony and prosthetic lateralization. Stress was lower with prosthetic lateralization through the glenosphere or baseplate. They concluded that the maximum mechanically acceptable limit is at least 10 mm for prosthetic glenoid lateralization and only 5 mm for bony lateralization. Other theoretical advantages of prosthetic over bony lateralization include less surgical time and more accurate control of lateralization in contrast to bone graft which adds surgical time for graft harvest and impaction and also can compress with impaction decreasing the desired extent of lateralization.

Despite the biomechanical advantages of glenoid lateralization, the extent of lateralization that can be achieved through the glenoid component is limited. In a review that included 28 currently used RTSA configurations, Werthel et al.⁶ demonstrated that none of the implants lateralizes more than 8.3 mm on the glenoid.

Humeral neck shaft angle

Valgus neck shaft angle (NSA) improves abduction but increases scapular notching and limits extension.^19–21 Decreasing the NSA has been shown to increase lateral offset²² and improve impingement-free ROM,¹⁹ particularly inferior impingement, which helps avoid an adduction deficit²⁰ and lowers the risk of scapular notching.²³ Using computed tomography scan-based three-dimensional computer templating, Werner et al.¹⁹ analyzed the influence of NSA and bony COR lateralization on ROM. Impingement-free adduction, extension, and rotation were mostly influenced by NSA, whereas COR lateralization had the largest effect on impingement-free abduction and forward flexion. They concluded that a humeral NSA of 135° along with 5 mm glenoid lateralization provides the best combination for impingement-free abduction, adduction, and global function. Gutiérrez et al.¹⁵ studied the effect of glenosphere diameter, COR offset, glenosphere position on the glenoid, and humeral NSA on inferior scapular impingement. Reduction of NSA to 130° had the largest effect on inferior scapular impingement. Similarly, de Wilde et al.²⁴ demonstrated that decreasing the NSA from 155° to 145° improved impingement-free adduction by 10°. In a systematic review of 2222 shoulders, scapular notching was significantly higher with 155° prostheses than with 135° prostheses²³ (16.8% vs 2.8%). Lädermann et al.²⁵ compared the effect of humeral stems with different NSAs on ROM. Stems with 155° NSA achieved better abduction but worse adduction, extension, and external rotation compared to 135° and 145° designs.

Polyethylene insert and level of humeral cut

Since the humeral surface upon which humeral bearing is inserted is oblique when the arm is in the anatomic position, changing the level of humeral cut or polyethylene cup thickness not only affects the lateral offset but also the distal offset. The distal offset has an influence on the deltoid length and tension. Virani et al.⁹ demonstrated that polyethylene thickness did not have a significant effect on ROM. Polyethylene thickness and level of humeral cut can be adjusted intraoperatively to optimize passive joint tension and stability. Care should be taken not to overstuff the joint with thicker polyethylene inserts as this may have negative effects on joint load and deltoid force required for abduction.²⁶

Humeral bearing

Grammont-style RTSA design features an inlay design of the humeral component. Onlay designs, in which the humeral tray rests on top of the humeral stem, were developed to facilitate conversion of an anatomic arthroplasty to RTSA without replacing the humeral stem. Another advantage of onlay designs is the ability to adjust the lateral offset by placing the tray eccentric over the stem (Figure 3). All designs with humeral lateralization are onlay designs.⁶ Potential disadvantages of onlay designs include the need to resect more bone from the humerus as well as the modularity which creates another potential junction for failure of the implant.²² Additionally, humeral lateralization using an onlay design decreases the acromiohumeral distance (AHD) which may lead to acromial impingement.²²

Figure 3.

Illustration showing how eccentricity of an onlay humeral tray can lateralize the humerus. Republished from Lädermann et al.²² with permission from Springer Nature.

The position of the humeral tray does not change relative to the glenosphere.²⁷ Therefore, a medially placed tray lateralizes the humerus whereas a laterally placed tray shifts the humerus medially. Similarly, the anterior and posterior offsets can be changed by eccentric placement of the tray in the sagittal plane. An anteriorly placed tray shifts the humerus posteriorly while posterior placement of the tray shifts the humerus anteriorly. Berhouet et al.,²⁸ in a virtual model, studied the effect of onlay humeral tray position on impingement-free ROM and muscle moment arms. Placing the tray laterally (medializing the humerus) decreased superior impingement during abduction and elevation in the scapular plane. Placing the tray posteriorly shifted the humerus anteriorly increasing the subscapularis moment arm while anterior placement of the tray increased infraspinatus and teres minor moment arms. Therefore, eccentric placement of the tray in the sagittal plane can be adjusted according to preoperative deficits. A posteriorly placed tray maybe preferable in preoperative internal rotation deficit and an anteriorly placed tray may better correct preoperative external rotation deficit. Glenday et al.,²⁷ in a similar study, demonstrated that a laterally placed tray increases abduction and forward flexion ranges of motion before acromial impingement. A 5 mm laterally placed tray increased abduction by 10° and increased forward flexion ROM by 6°. A posterolaterally placed tray maximized the overall impingement-free ROM, whereas a medially placed tray maximized muscle moment arms. They identified an antagonistic relationship between impingement-free ROM and muscle moment arms due to tray placement. Both studies showed that eccentric placement of the tray does not affect the position of the cup relative to the COR or to the inferior scapular neck. Therefore, there was no effect on inferior impingement. Lädermann et al.²⁵ showed that a medially placed tray (lateralized humerus) decreased abduction due to decreased AHD.

Humeral stem design

Lädermann et al.²² demonstrated that using a curved stem can increase the lateral offset to a greater extent than that could be achieved by decreasing NSA or by using an onlay design (Figure 4). However, unlike other factors that influence the lateral offset, lateralization by stem design (curved vs straight) is predetermined by the implant design and cannot be modified by the surgeon intraoperatively. Awareness of shoulder surgeons about these factors help choose the appropriate implant preoperatively and should be considered in future implant designs.

Figure 4.

Illustration showing how a curved stem can lateralize the humerus relative to the center of polyethylene. Republished from Lädermann et al.²² with permission from Springer Nature.

Effect of lateral offset on biomechanics of RTSA

Scapular notching

Inferior impingement leading to scapular notching is one of the main complications of RTSA. High rates of scapular notching have been reported with the use of nonlateralized prostheses, reaching up to 96% in some series.²⁹ Notching occurs due to friction between the polyethylene humeral bearing and the scapular pillar.³⁰ This can lead to foreign body reaction from polyethylene debris and higher risk of glenoid loosening.^31,32 Clinically, scapular notching can result in increased pain, in addition to decreased strength, ROM and functional outcome scores.^33–36

As mentioned earlier, Lateralization can minimize inferior impingement and scapular notching through decreasing NSA and/or COR lateralization.²⁴ Eccentric placement of humeral tray does not affect inferior scapular impingement.²⁷ de Wilde et al.²⁴ demonstrated that prosthetic overhang is the most effective method to prevent scapular notching. Additionally, with only 1 mm overhang, the positive effect of COR lateralization in preventing scapular notching was eliminated. However, in another study, NSA had the largest effect on inferior scapular impingement.¹⁵

Range of motion

Gutierrez et al.³⁷ demonstrated a positive linear correlation between COR offset and abduction ROM before impingement on the acromion. Werner et al.¹⁹ showed similar results in terms of increased impingement-free abduction with COR lateralization. Additionally, they demonstrated that superior impingement with abduction occurred at the superior edge of the glenoid with medialized COR and at the acromion with lateralized COR.

Several studies have shown the COR lateralization improves adduction ROM.^15,20 However, Henninger et al.³⁸ disputed this effect. Authors of the latter study explained the different outcomes were related to different types of implant (hemispherical glenosphere with a cylindrical spacer (Tornier Aequalis) versus a more spherical glenosphere (Reverse Shoulder Prosthesis; DJO Surgical, Austin, TX, USA)), experimental methods (accounting for native deltoid tension) and surgical technique (glenoid tilt).

Rotational movements improve with both glenoid^16,39,40 and humeral lateralization.²² Both methods restore the tension of the anterior and posterior cuff. Additionally, glenoid (COR) lateralization recruits more deltoid fibers medial to the COR and increases rotational moment arm of the deltoid. It also decreases impingement on the coracoid with internal rotation and the scapular spine with external rotation. On the other hand, humeral lateralization, in addition to the improved cuff tension, increases the distance between the COR and cuff insertion which increases the moment arm of the anterior and posterior cuff and thereby improves rotation.

Giles et al.⁴¹ demonstrated that both internal and external rotation improved with either humeral lateralization or glenosphere lateralization. However, for internal rotation, combining both methods were not synergistic and resulted in internal rotation less than that achieved by each method alone. Overstuffing of the joint and overtensioning of the posterior soft tissues tether the posterior structures limiting internal rotation. They highlighted the effect of intraoperative assessment of internal rotation and the possibility of releasing the contracted posterior soft tissue. On the other hand, external rotation was not limited by overtensioning anterior soft tissues when both humeral and glenoid lateralization were applied.⁴¹ Similarly, Ferle et al.⁴² showed that increased soft tissue tension had no negative impact on passive external rotation ROM.

Using a virtual shoulder model, Virani et al.⁹ evaluated the effects of different methods of glenoid and humeral lateralization on ROM. They concluded that COR lateralization increases ROM in all planes. A higher NSA maximized abduction while a lower NSA maximized flexion/extension. In their study, inferior glenosphere placement had the most impact on rotational ROM.

Joint loads

Lateralization of COR increases the overall joint contact forces including compression, shear, and bending moments.^26,43 Cuff repair in the setting of COR lateralization additionally increases joint load by 29%.⁴¹ The increased overall joint loads improve joint stability possibly by increased soft tissue tension.⁴³ However, this can have a negative impact on prosthetic wear and implant survival.³³ The increase in shear and bending moments creates a destabilizing torque at the glenoid-implant interface.⁴³ Several studies demonstrated a linear correlation between increasing COR offset and baseplate micromotion.^4,44,45 This potentially leads to a higher risk of baseplate loosening, glenosphere unscrewing, and migration.⁴⁶ The importance of adequate screw fixation of the baseplate with COR lateralization has been emphasized in multiple studies.^4,47,48

Studies have shown that joint load magnitude is either decreased or unchanged with humeral lateralization.^41,49 Giles et al.⁴¹ demonstrated that humeral lateralization (eccentric tray) and rotator cuff loading significantly affect the joint load angle to become more compressive and therefore contribute to concavity compression and enhance joint stability. They found no effect of glenoid lateralization on joint load angle.⁴¹

Deltoid

Grammont named his second-generation prosthesis “Delta” in order to emphasize the importance of the deltoid in RTSA.⁵⁰ All current RTSA designs have a medialized COR relative to the native shoulder. This increases the moment arm of the deltoid which helps decreasing the load required by the deltoid for abduction. Disadvantages of nonlateralized designs include impairment of normal shoulder contour^16,40 and loss of physiological wrapping of the deltoid around the greater tuberosity.⁸ Lateralized COR designs usually minimize the positive influence of a medialized COR on the deltoid by decreasing the deltoid moment arm in elevation and abduction.⁵¹ This increases the abduction force required by the deltoid often leading to deltoid fatigue and acromial fracture.^{26,38,43,52–56} In contrast, humeral lateralization achieves the advantages of lateralizing the greater tuberosity and the deltoid insertion while keeping the COR in its medialized position. This increases the abductor lever arm of the deltoid decreasing the force required for abduction.⁵⁷ Additionally, with humeral lateralization, more fibers of the anterior and posterior deltoid are lateral to COR, which can theoretically assist in abduction. Therefore, the influences of glenoid and humeral lateralization on the abduction force required by deltoid are opposite. This has been confirmed in several studies where the deltoid force required for abduction increases with COR lateralization^26,38,58 and decreases with humeral lateralization.^26,49 In two different in vitro studies by the same authors, Giles et al.^26,41 reported that humeral lateralization using an eccentric humeral tray can partly reverse the negative effect of glenosphere lateralization on deltoid force. Additionally, rotator cuff load exacerbated the negative effect of glenosphere lateralization on the deltoid.⁴¹ They suggested that glenosphere lateralization causes the cuff to act as an adductor, antagonizing the effect of the deltoid. The negative effect of cuff repair in the setting of glenosphere lateralization on deltoid force and joint loads makes this combination undesirable.

Stability

Instability is the most common cause for RTSA revision with a prevalence up to 48%.⁵⁹ Instability in nonlateralized designs is possibly due to insufficient soft tissue tension as well as inferior impingement on the scapular neck⁶⁰ and loss of physiological wrapping angle of the deltoid.⁸ Inferior impingement can be minimized by glenoid lateralization or decreasing NSA. Any method of lateralization can provide more tension on the deltoid and rotator cuff muscles during flexion and abduction. Glenoid lateralization increases overall joint loads, including the compressive component of joint reaction force (JRF) which improves stability.^43,49 Humeral lateralization via eccentric humeral tray placement has been shown to change the joint load angle to become more compressive.⁴¹ Ferle et al.⁴² compared anterior dislocation force with different magnitudes of glenosphere lateralization and with varying NSA. Higher dislocation forces were required with 6 mm and 9 mm lateralized compared to nonlateralized glenosphere. Configurations with NSA of 135° required higher dislocation forces than those with NSA of 145°.

The biomechanical aspects of different RTSA lateral offset designs are summarized in Table 2.

Table 2.

Summary of biomechanical aspects according to the magnitude of lateral offset and method of lateralization.

	Nonlateralized RTSA	Glenoid (COR) lateralization	Humeral lateralization (Nonlateralized COR)
Biomechanical concept	Nonlateralized COR to minimize shear forces on glenoid and decrease force required by deltoid	Lateralized COR to decrease impingement and improve ROM	More anatomic position of tuberosities while maintaining a nonlateralized COR
Deltoid	Less force required by deltoid due to: • Higher abduction lever arm • More fibers of the anterior and posterior deltoid lateral to COR, thereby can assist in abduction	COR closer to deltoid line of pull leads to: • Decreased moment arm of deltoid in elevation and abduction • Increased abduction force required for the deltoid • Increased acromial stress fracture	Less force required by deltoid due to: • Higher abduction lever arm • More fibers of the anterior and posterior deltoid lateral to COR, thereby can assist in abduction
ROM	Less impingement-free ROM-Poor restoration of IR and ER due to: • Slack cough • Peripheral impingement • Less deltoid fibers medial to COR	Improves ROM: • Increased tension of the remaining cuff • Less peripheral impingement • More deltoid fibers medial to COR which can help with rotational movements	Improves ROM • Increased tension of the remaining cuff • Increased abductor lever arm • Increased rotational moment arm of anterior and posterior cuff-More acromial impingement due to less AHD (less abduction)
Scapular notching	Higher risk of scapular notching leading to • glenoid osteolysis, loosening • polyethylene wear	Lower risk of scapular notching	• Lower risk with decreasing NSA • No effect with eccentric humeral tray placement
Stability	Instability due to • slack cough • inferior impingement loss of physiological wrapping angle of the deltoid	Improved stability • Increased soft tissue tension which increases compressive force • Increased overall joint load magnitude • Partial restoration of physiologic wrapping of the deltoid • Less peripheral impingement	Improved stability • Increased soft tissue tension which increases compressive force • Affects joint load angle which increases compressive component of JRF • Partial restoration of physiologic wrapping of the deltoid
Joint loads	Increased compressive component of joint load	Increased shear forces which increases risk of glenoid loosening	Increased compressive component of joint load
		Increased total JRF due to: • Increased soft tissue tension, including rotator cuff • Increased force required by deltoid • Deltoid wrapping around greater tuberosity Leads to: • Improved stability • Negative impact on prosthetic wear and implant survival	Increased total JRF due to: • Increased soft tissue tension, including rotator cuff • Deltoid wrapping around greater tuberosity Leads to: • Improved stability • Negative impact on prosthetic wear and implant survival
Shoulder contour	Loss of normal shoulder contour	More anatomic shoulder contour
Soft tissue tension	Slack rotator cuff	• Retensions the remaining cuff. • A more anatomic vector of pull of the deltoid as it wraps around the humeral component

COR: center of rotation; RTSA: reverse total shoulder arthroplasty; ROM: range of motion; AHD: acromiohumeral distance; JRF: joint reaction force; IR: internal rotation; ER: external rotation.

The subscapularis

The influence of subscapularis repair on biomechanics of RTSA is controversial. Some advocate repair in order to decrease the risk of dislocation and improved internal rotation.^21,61–64 Repair also improves joint protection with muscle closure around the joint. Other studies showed that subscapularis repair antagonizes the posterior cuff in external rotation and does not improve stability.^65–67 Additionally, the distalization of the humerus creates an adductor moment arm of the subscapularis,⁶³ which increases the abduction force required by the deltoid.⁶⁷

The decision to repair the subscapularis may be influenced by the lateral offset. Werner et al. reported that subscapularis repair in the setting of glenosphere lateralization can negatively affect functional outcomes.⁶⁸ In contrast, Roberson et al.,⁶⁹ using a lateralized glenoid design, reported no differences in functional outcomes with or without repair. Humeral lateralization increases the distance between COR and humerus which increases the rotational moment arms. Friedman et al.,⁷⁰ with a humeral lateralized design, reported better internal rotation and functional outcome scores but less active abduction and passive external rotation with subscapularis repair. Eno et al.⁷¹ demonstrated that repairing the subscapularis in a more superior position can help reduce its adductor moment arm and minimize the antagonistic effect on the deltoid.

Published outcomes in clinical studies

Nonlateralized RTSA

Sirveaux et al.,³⁶ in a multicenter study, reported the outcomes of 80 RTSAs for treatment of glenohumeral arthritis with massive rotator cuff tears using the Delta III prosthesis at a mean follow-up of 44 months. The mean gain in active external rotation at the side was 7.7°. The mean gain in passive external rotation at the side was 9°. Both were not statistically significant. However, statistically significant improvements were achieved in active and passive external rotation at 90° of abduction. Rate of scapular notching was 63.6%. Several other studies reported high rates of scapular notching and limited improvements in external rotation with nonlateralized designs.^29,72,73

Lateralized RTSA

Frankle et al.⁴⁰ reported outcomes of RTSA using the first-generation Reverse Shoulder Prosthesis (Encore Medical, Austin, TX, USA) which has a lateralized COR. They achieved a 29.1° postoperative improvement in external rotation and no notching at a mean of 33 months follow-up. However, out of 60 shoulders, there were 7 cases of glenoid loosening and 2 cases of stress fracture of the metal baseplate. To minimize the incidence of baseplate failure that was observed in the latter study, a modification of the Encore Reverse Shoulder Prosthesis was made by designing a 5 mm peripheral locking screws for use with the baseplate instead of 3.5 mm nonlocking screws. Additionally, the surgical technique was modified to inferiorly tilt the glenosphere 10–15° which had been proved to produce the most uniform compressive forces and limit micromotion at the glenoid-implant interface.⁷⁴

Following these important modifications, another study was performed in the same institution on 96 shoulders using the second-generation Reverse Shoulder Prosthesis. At a mean follow-up of 27.5 months, they reported no mechanical baseplate failure and no scapular notching. They achieved a 54.5° gain in forward flexion, 48.5° in abduction and 14.8° in external rotation.⁷⁵ In their five-year follow-up report, there were no baseplate loosening or failure and 9% scapular notching. Additionally, forward flexion, abduction, and external rotation ranges of motion continued to improve significantly.⁷⁶ However, they suggested that improved ROM may be due to different methods of assessment at the two- and five-year follow-up reports. In their 10-year follow-up report, 42 shoulders were available for follow-up. There was no mechanical baseplate failure or glenoid component loosening. There was a slight decrease in all planes of motion compared to the five-year report. Implant survivorship was 90.7% using revision as an endpoint.⁷⁷ Following these reports, several studies reported improved external rotation and lower rates of scapular notching with COR lateralization compared to nonlateralized designs.^5,16,17

Comparative studies

Streit et al.,⁷⁸ in a prospective study compared the outcomes of 9 grammont-style prostheses (Aequalis) (nonlateralized COR and a 155° NSA) and 9 Encore Reversed Shoulder Prostheses (DJO) (6- or 10-mm lateralized COR and a 135° NSA). Mean follow-up was 8.1 months. Postoperative ER was 35° in the lateralized group and 28.3° in the nonlateralized group (p = 0.07). However, higher forward flexion values were achieved in the nonlateralized group (144° vs 116°) (p = 0.05). They explained the latter observation by the less distalization achieved with the Encore prosthesis. Despite the small sample size and the short follow-up period, the observed differences in ROM are worth considering.

Athwal et al.⁷⁹ retrospectively compared outcomes of a standard Grammont-style RTSA with a BIO-RTSA (40 patients; 20 in each cohort). They utilized the Aequalis system in both groups. Scapular notching was 75% in the standard RTSA group and 40% in the BIO-RTSA group (p = 0.022). Rate of complete graft incorporation in the BIO-RTSA group was 80%. No differences were noted in active flexion, external rotation, internal rotation, or any of the functional outcomes tested.

Collin et al.⁸⁰ compared outcomes of standard (69 shoulders) and BIO-RTSA (61 shoulders) by a single surgeon. They used the Aequalis system for both groups. There were no differences in internal/external rotation and in radiographic outcomes including scapular notching. Forward flexion was better in the standard group. However, the difference was not clinically important (145° vs 138°).

Franceschetti et al.⁸¹ recently performed a comparative study of 29 nonlateralized and 30 BIO-RTSAs. All surgeries were performed by a single surgeon using Aequalis Ascend Flex with a NSA 145°. Graft incorporation 90% in the BIO group. External rotation gain was 23° in BIO and 17° in standard (statistical analysis was not performed on external rotation gain)

Merolla et al.⁸² compared outcomes of 36 Aequalis Reverse II (155°, inlay, straight stem, follow up (FU) 35.1 months) versus 38 Aequalis Ascend flex (short curved stem, eccentric tray, FU 29.1 months). In the second group, 20 patients had BIO lateralization. External rotation gain was 15° in the first group and 33° in the second. (p = 0.003). Three scapular fractures occurred with the curved short stem and BIO implant. Scapular notching was 39% in the first group and 5% in the second.

Level I randomized controlled trials

Greiner et al.³⁹ compared the outcomes of 10 mm BIO versus standard Aequalis Reverse Shoulder Prosthesis (34 patients; 17 in each group). Mean follow-up was 22 months. Two patients died and 1 was lost to follow-up leaving 15 patients in the standard group and 16 in the BIO group. External rotation gain in standard versus BIO at 0° was 18° versus 28° and at 90° was 33° versus 42°. Differences did not reach statistical significance, possibly due to the small sample size. However, it did reach statistical significance after excluding patients with degenerative changes of teres minor; 16° versus 42° at 0 and 27° versus 60° at 90. Graft healed in all BIO patients.

Gobezie et al.⁸³ enrolled 100 primary reverse shoulder arthroplasty patients to compare outcomes according to NSA using otherwise identical implants (Univers Revers prosthesis (Arthrex, Naples, FL, USA) with a nonlateralized glenosphere). This prosthesis allows placing the humeral cup at either 135° or 155°. Thirty-seven patients in the 135° group and 31 patients in the 155° group completed a minimum follow-up of two years. No significant differences in ROM and functional outcomes. Postoperative external rotation was similarly not improved in both groups. Rate of scapular notching was significantly higher in the 155° group (58%) versus 21% in the 135° group. They attributed the lack of improvement in external rotation and the relatively high rate of scapular notching in the 135° group to the nonlateralized glenosphere which they now recommend against using.

Long-term outcomes

It is still not clear whether lateralization influences long-term outcomes and implant survivorship. Functional outcomes of nonlateralized RTSA were reported to deteriorate after six to eight years.^36,53 Similarly, there was slight deterioration in all planes of motion at 10-year follow-up compared to the 5-year report in the study by Cuff et al.⁷⁷ In two studies with minimum 10-year-follow-up, implant survivorship at 10 years was 90.7% with the lateralized Encore/DJO Reversed Shoulder Prosthesis⁷⁷ and 93% with the nonlateralized Delta III prosthesis.⁸⁴ Kennon et al.⁸⁵ recently compared outcomes of nonlateralized and glenoid-lateralized RTSA at 10 years. There were no significant differences in complications or reoperations. Nonlateralized group had better active forward flexion and higher scapular notching rate.

Conclusion

The ideal extent and method of lateralization are yet to be determined and are probably variable from one patient to another. Surgeons can choose from several RTSA designs with various features. Patient factors usually influence this choice including age, bone quality, strength of glenoid fixation, cuff condition, patient activities, and patient expectations. Most current implants allow different degrees and methods of lateralization within the same implant to help individualize the lateral offset for different patients. Excessive lateralization can be problematic in smaller patients and those with soft tissue contractures. This can result in difficult joint reduction and subscapularis repair. Moreover, overstuffing can limit ROM and accelerate polyethylene wear.

In case glenoid-lateralized implants are used, glenoid component fixation should be optimal. Anterior or posterior screws are commonly omitted in small-sized glenoids or in patients with remarkable glenoid erosion. In this case, risk of glenoid complications might be higher and the surgeon’s threshold to not lateralize through the glenoid should be lowered. In case optimal baseplate fixation is achievable, maximizing the function and ROM may be reasonable to aim at. Humeral lateralization can be applied to improve soft tissue tension, improve stability, and decrease force required by deltoid. However, the risk of acromial impingement and overstuffing the joint should be considered.

Ethical approval and Patient Consent

Not needed.