Complications after Reverse Shoulder Arthroplasty

Thesis

Reference

Complications after Reverse Shoulder Arthroplasty

LAEDERMANN, Alexandre

Abstract

The aim of RSA is to lower the shoulder's center of rotation which increases tension in the deltoid muscle, thus increasing muscle fiber recruitment in its anterior and posterior heads, in order to compensate the deficient rotator cuff. Today, this procedure is commonly performed even if the rate of intra- or postoperative complication remains high. Inadequate deltoid tension is an accepted major risk factor for complication. Postoperatively, the mean humerus lengthening is approximately 2 mm and the arm approximately 24 mm. As for humerus and arm shortening, it is known to increase the risk of dislocation and to lead to poor anterior active elevation, respectively. The type of surgical approach does not play a role in postoperative function. Subclinical neurologic lesions are frequent, are not related to the glenosphere position in the vertical plane, with a drastically increasing prevalence above 40 mm of arm lengthening. A reasonable goal for arm length should thus be between 0 and 20 mm. Preoperative planning seems mandatory in revision cases. Finally, experience can lead to a dramatic decrease in complications over [...]

LAEDERMANN, Alexandre. Complications after Reverse Shoulder Arthroplasty. Thèse de privat-docent : Univ. Genève, 2013

DOI : 10.13097/archive-ouverte/unige:26301

Available at:

http://archive-ouverte.unige.ch/unige:26301

Disclaimer: layout of this document may differ from the published version.

Division of Orthopaedics and Trauma Surgery Department of Surgery

Geneva University Hospitals

___________________________________________________________________

Complications after Reverse Shoulder Arthroplasty

Thesis submitted to the Medical School of the University of Geneva

For the degree of Privat-Docent by Alexandre LÄDERMANN

___________________________________________________________________

Geneva

Table of Contents

Abbreviations ... 3

Acknowledgment ... 4

Disclaimer ... 6

Abstract ... 7

Objectives ... 8

Hypothesis ... 8

Introduction ... 9

Pathologic conditions of the shoulder ... 9

Reverse shoulder arthroplasty ... 12

A) Development ... 12

B) Indications and contraindications ... 13

C) Complications ... 14

Radiological changes ... 17

Infection ... 17

Instability ... 18

Acromial fracture ... 20

Neurological lesion ... 23

Functional impairment ... 23

Surgeon’s experience ... 24

Other complications ... 24

Appendix ... 34

Appendix 1 ... 34

Appendix 2 ... 35

Appendix 3 ... 36

Appendix 4 ... 37

Appendix 5 ... 38

Appendix 6 ... 52

Appendix 7 ... 53

Appendix 8 ... 54

References ... 70

Abbreviations

AAE: Active anterior elevation EMG: Electromyography

ORIF: Open reduction and internal fixation RSA: Reverse shoulder arthroplasty

Acknowledgment

We are not born surgeons, we become surgeons. Such a career is built progressively, and a solid base ground is mandatory. Every step has its importance.

If one is missed, the necessary balance will not be reached.

I would like to thank all persons that have been crucial to obtain this intelectual and technical equilibrium.

First, I thought about my parents. They have offered me a stable environment at home that allowed me to grow peacefully. They also taught me the merriment of work, perseverance, and the will to always try harder.

Second, I am also very indebted to all surgeons who taught me surgery and the pleasure of its practice. I have particularly in mind Nicolas RIAND, Danilo TOGNINALI and Dimitri CERONI. Little would have been achieved without those colleagues who played an important role in my professional life. All of them remain esteemed friends.

Third, I am sincerely grateful to my three mentors for their guidance, teaching, enthusiasm, commitment, and academic leadership. Prof. Pierre HOFFMEYER, who actively supported me in this work and offered me the possiblity to become an orthopaedic surgeon. He let me evolve while maintaining a necessary control. He encouraged me to travel and has sent me to the right places to learn shoulder surgery. I thus had the privilege to perform fellowships in the world’s best centers. I

underwent my first shoulder electroshock with Gilles WALCH, the initiator of prosthetic shoulder surgery in Lyon. Moreover, I had the oportunity to work with Steve BURKHART during another scientific fellowship in San Antonio. This surgeon is also considered a «godfather», one of shoulder arthroscopic surgery.

Disclaimer

The author of this thesis, his immediate family, and any reasearch foundation with which he is affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this thesis.

Several of the co-authors that have been working with the author of this thesis on the various articles that will be mentionned have disclosed that they receive royalties from Tornier for the reverse shoulder arthroplasty (RSA) and Arthrex for shoulder arthroscopic devices.

Abstract

The aim of RSA is to lower the shoulder’s center of rotation which increases tension in the deltoid muscle, thus increasing muscle fiber recruitment in its anterior and posterior heads, in order to compensate the deficient rotator cuff. Today, this procedure is commonly performed even if the rate of intra- or postoperative complication remains high. Inadequate deltoid tension is an accepted major risk factor for complication. Postoperatively, the mean humerus lengthening is approximately 2 mm and the arm approximately 24 mm. As for humerus and arm shortening, it is known to increase the risk of dislocation and to lead to poor anterior active elevation, respectively. The type of surgical approach does not play a role in postoperative function. Subclinical neurologic lesions are frequent, are not related to the glenosphere position in the vertical plane, with a drastically increasing prevalence above 40 mm of arm lengthening. A reasonable goal for arm length should thus be between 0 and 20 mm. Preoperative planning seems mandatory in revision cases.

Finally, experience can lead to a dramatic decrease in complications over time.

Objectives

The purpose of this thesis was (1) to present the research projects that have been achieved by the author in order to better understand biomechanical changes induced by reverse arthroplasty of the shoulder, (2) to propose a technique to measure humerus and arm lengthening (Appendix 1) and (3) a standardized template for RSA in difficult cases like in the case of bone loss or revision surgery (Appendix 1), (4) to determine the ideal soft-tissue tension that provides the best functional outcome without increasing the risk of complications (Appendix 2-4), (5) to find the causes of complications (Appendix 1-8) and, finally, (6) to propose strategies to avoid them (Appendix 1-5, 7, 9).

Hypothesis

The hypothesis is that a greater knowledge of changes induced by RSA helps to prevent troublesome complications related to the use of this implant.

Introduction

Joints of the shoulder include the glenohumeral, acromioclavicular, sternoclavicular and scapulothoracic joints. Pain can arise in or around the shoulder from its joints and surrounding soft tissues. This problem is common, with an estimated prevalence of 4% to 26%.1,2,5,12,91,126 About 1% of adults aged over 45 years consult their general practitioner with a new presentation of shoulder pain every year.⁹⁷

Pathologic conditions of the shoulder

The etiology of shoulder pain is diverse and includes pathologies originating from the neck, glenohumeral joint, acromioclavicular joint, rotator cuff, and other soft tissues around the shoulder girdle. The most common source of shoulder pain is the rotator cuff, accounting for over two thirds of the cases.

The erect posture and the evolution of morphology allowed humans to acquire new skills. However, the anatomical changes in the shoulder may also explain the mechanical failure of the shoulder in modern humans.⁵⁹ During evolution, the permanently orthograde posture has freed the human shoulder girdle of its quadruped functions. The scapulohumeral complex underwent drastic changes to facilitate prehension, the anterior limbs became the upper limbs with the characteristics of a non-weight-bearing joint.¹¹ Major bony and muscular

First, the upper limb evolved from an internal to an external rotation position. The humerus undergoes torsion that has, in turn, affected the position and size of the humeral tubercles. In primitive forms, the bicipital groove lays approximately midway between the two tubercles, which are of almost equal size. The effect of the torsion has displaced the bicipital groove medially, so that it encroaches upon, and reduces the size of the lesser tubercle. The illogical pathway of the long head of the biceps, going from a primarily horizontal to a secondary vertical position, without a strong medial tether, increase the possibility of its medial subluxation and dislocation.

Second, the quadruped human being became a biped; the humerus that formerly gave support to the glenoid, became supported by the latter. The notion “inverted morphological role” (that led later to the notion of reverse arthroplasty) appeared at this point.⁵⁹

Third, obvious modifications occurred in the shape of the scapula that can be expressed by an index known as the scapular index. A relative atrophy of the supraspinatus muscle occurred, as illustrated by a decreased of this index.^68,107 The decrease in the effectiveness of this muscle was at the same time compensated by the increase in size, mass, and lateral extension of the acromion process. The progressive distal migration of the point of insertion of the deltoid muscle and lateralization of the acromion indicates the more dominant position occupied by the deltoid with strengthening of the middle deltoid abduction component (Figure 1).⁵⁹

Figure 1: After a reverse shoulder arthroplasty, the medialization of the center of rotation recruits more of the deltoid fibers (segments I and IV) for active elevation.

From: Boileau P, Watkinson DJ, Hatzidakis AM, Balg F. Grammont reverse prosthesis: design, rationale, and biomechanics. J Shoulder Elbow Surg. 2005 Jan- Feb; 14(1 Suppl S): 147S-161S.

Other degenerative, traumatic, morphologic, biologic and genetic factors are also associated contributors to the genesis of this failure.102,103,140,147 As previously mentioned, the rotator cuff is the most common structure that becomes insufficient and its failure is associated with numerous changes.^75,95,140 When detached from the bone, the musculotendinous unit will retract medially,⁹⁶ the muscle may develop atrophy^140,144 or fatty infiltration.56-58,72,140 In the absence of concavity compression and humeral head lowering exerted by the rotator cuff, the unopposed contraction of the deltoid creates a force vector that displaces the head superiorly rather than in abduction. According to the type of rotator cuff lesion, the patient may present pseudoparalysis,⁷⁶ phenomenon which is characterized by loss of active anterior

Reverse shoulder arthroplasty

A) Development

To compensate the loss of function of the rotator cuff, several options have been proposed; the most reasonable, whenever possible, is to repair the rotator cuff. Good results are obtained in the vast majority of the cases39-43,77,78 with healing of the rotator cuff on the tuberosities¹⁴⁷ and even reversal of the associated pseudo paralysis.³⁸ In some circumstances, rotator cuff repair is however contraindicated or technically impossible.

For instance, a rotator cuff insufficiency associated with pain and pseudo paralysis remains a challenging condition. It is extremely difficult, if not impossible, to obtain a functionally good result with a conventional prosthetic arthroplasty in this situation, where only a “limited goals surgery” is appropriate, a concept introduced by Neer.¹⁰⁴ Effectively, hemiarthroplasty provides satisfactory pain relief but poor motion,

110,115,139,146 whereas total anatomic shoulder arthroplasty is complicated with early loosening of the glenoid component. ^7,49,66

In rotator cuff deficiency, the only remaining muscle able to elevate the arm is the deltoid. In order to allow AAE above 90°, the abduction role of the deltoid has to be increased. This can be obtained by several mechanisms, such as an osteotomy of the scapular spine⁸ or by medializing the rotation center of the joint.⁶¹ The concept of functional surgery is born from the latter option: whereas no effective anatomic solution exists, restoration of function has to be proposed through a novel

morphology.

The first-generation of RSA have been implanted in Germany and France.^53,73 However, high rate of complications forced to abandon their use. Nevertheless, successive improvements imagined by Grammont followed and, in 1991, a RSA called the “Delta III” has been developed.^8,9

The RSA should be inherently stable, the weight bearing part being convex and the supported part concave (reversal of the ball and socket). The arthroplasty position medializes and lowers the glenohumeral center of rotation, thereby increasing the lever of the deltoid muscle. Deltoid tension, increased by the lowered center of rotation, increases muscle fiber recruitment of the anterior and posterior deltoid that compensates for a deficient rotator cuff (Figure 1).

The lever of the deltoid muscle is almost doubled with a RSA, and therefore the abduction efficiency of the deltoid. Under such tension, the metaglene provides the stable fulcrum essential for shoulder AAE and prosthesis stability.²¹ The increase in compressive force component has also a stabilizing effect on the humeral head.⁵⁰  

B) Indications and contraindications

The RSA is a powerful tool that has opened new barriers, especially for reconstructive shoulder surgery. Traditionally, the ideal candidate is a patient above 70 years old with symptomatic cuff tear arthropathy. Many pathologies that could not

currently expanding. It is now used for various conditions such as failed total shoulder arthroplasty or hemiarthroplasty, complex proximal humeral fractures and defective fracture union or nonunion, rheumatoid arthritis with rotator cuff lesions, failed or irreparable massive rotator cuff tears, and tumors.

16,20,21,25,34,37,48,51,63,64,84,86,90,113,134-136

Absolute contraindications include an uncontrolled active infection, neuro- arthropathies, and substantial deltoid insufficiency because of the very high probability of recurrent instability and the minimal potential gain in function (Appendix 8).⁸⁰

C) Complications

The first series of RSA with an at least two years follow-up, confirmed the preliminary results with excellent functional outcome and stable glenoid fixation.9,10,22,113,121,128,136

However, the complexity of this procedure with regards to its singular anatomy and special patient population, is reflected by the large number of reported problems and complications. As defined by Zumstein et al., problems can be defined as intra- or postoperative events that are not likely to affect the patient’s final outcome.¹⁴⁸ This will include hematomas, phlebitis, heterotopic ossification, algodystrophy and will not be part of the treated subjects of this thesis. Complications are defined as any intra- or postoperative events that are likely to have a negative influence on the patient’s final outcome, such as intraoperative cement extravasation, intra- or postoperative fractures, dislocations, infections, neurological lesions, radiographic changes such as glenoid or humeral lucent lines, scapular notching, stress shielding, aseptic

loosening, reinterventions (without replacement of the component) or revisions (with replacement of the component).

RSA, often used in multiple operated patients with distorted anatomy, implies physiological and biomechanical changes that may increase the potential for complications.⁴⁶ The first change is the medialization of the centre of rotation, which is responsible for impingement of the medial border of the humeral component on the scapular neck when the arm is adducted (Appendix 6). Repetitive contact between polyethylene and bone may result in polyethylene wear debris, chronic inflammation and osteolysis,^105,106 radiolucency around the glenoid component,¹³⁶ loosening of the glenoid component,²⁶ presence of an inferior bone spur and ossification in the glenohumeral space (Appendix 6).^20,83

The other change is the semi-constrained design of the RSA; the tortional forces may influence the stability of the humeral component (Appendix 6). Effectively, changes on the humeral side such as osteolysis, osteopenia, the development of medial and lateral cortical bone narrowing associated with osteopenia, condensation lines around the tip of the stem, and a spot weld between the cortical bone and the stem, radiolucent lines and loosening have been reported (Appendix 6).^20,122

Also due to this semi-constrained design, deltoid tension is crucial (Figure 2).

Figure 2: Influence of position of glenosphere in vertical plane. (A) A rather high

position of the glenosphere in the vertical plane allows satisfactory deltoid retensioning.

From: Lädermann A, Walch G, Lubbeke A, Drake GN, Mélis B, Bacle G, Collin P, Edwards TB, Sirveaux F. Influence of arm lengthening in reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2012 Mar; 21(3): 336-41.

Failure to adequately tension the deltoid may result in prosthetic instability, the most common clinically significant complication (Appendix 1). Moreover, other complications following RSA, such as neurological lesions, fractures of the acromion, or permanent abduction of the arm,20,21,50,120,129 have also been described and could also be related to retensioning (Appendix 1). When the author of this thesis began his research, very few had been published about humerus and arm lengthening after RSA, its biomechanical implications and consequences.

Radiological changes

This is the most frequently reported complication after RSA.¹⁴⁸ Only short- to mid- term studies reported their prevalence.83,119,121,122 We thus conducted a retrospective long-term study (Appendix 6). All the current knowledge of this subject is summarized in this work.

Infection

The incidence of infections after primary RSA is around 5%,⁴⁶ which is higher than in anatomic shoulder arthroplasty.^47,117 Reasons are the large dead space caused by the ball-and-socket configuration, the frequent postoperative hematoma, the extensive surgical dissection, and in some patients the compromised general health and the numerous previous surgeries.^32,98 The commonly identified low-virulence organism are Propionibacterium acnes and Staphylococcus epidermidis.20,98,134,136

Proven and suspected infections should be revised operatively. Acute infection of less than three weeks in a stable arthroplasty should be treated with debridement and antibiotics.^124,145 Late infections should be treated with arthroplasty removal, debridement and reimplantation.^98,124 The author of this study advocates a two-stage surgery (Figure 3).

Figure 3: Case illustration. This 54 year old male was known for a below knee amputations due to diabetes mellitus. He sustained a motor vehicle accident in

fracture of the right proximal humerus. B) The patient developed a deep infection that finally required removal of the prosthesis in October 2007. The patient suffered from consequent pain and was unable to walk with crutches. C) A RSA was implanted in February 2008.

From: Lädermann A, Walch G. Reverse shoulder arthroplasty with deltoid insufficiency. Swiss Med Wkly 2012; 142:25S

Instability

Instability at the ball and socket interface is usually detected at its full range, which is anterior or anterolateral dislocation (Figure 4).

Figure 4: A) Dislocation of a right reverse shoulder arthroplasty that requires open revision and humeral lengthening with a spacer. B) At 47 months follow-up, the patient had no recurrence.

From: Lädermann A, Walch G. Reverse shoulder arthroplasty with deltoid insufficiency. Swiss Med Wkly 2012; 142:25S

The latter is one of the most common complications after RSA, with rates as high as 14%, accounting for almost half of the complications in some series (Appendix 1).

20,27,34,35,37,55,84,134,136 Most cases of dislocations occur during the first few months after implantation and result of a technical error. The origin of dislocation is multifactorial. It can occur due to 1) deltoid insufficiency (Appendix 2), 2) lack of anterior restraint in case of subscapularis, conjoint tendon weakness, or pectoralis

impingement, and 5) infection. Instability is more frequent in case of revision arthroplasty.¹⁴⁸ Deltoid insufficiency can result from a postoperative lack of deltoid retentionning and thus compressive forces, polyethylene wear, stem subsidence, or postoperative neurological lesion (Appendix 1-6). This represents the most common cause. Subscapularis integrity has attracted much attention in recent publications.^45,52,134 The utility of its repair, when not discouraged by the implant design, is however still debated. Edwards et al. and Gallo et al. found a compromise of the subscapularis tendon at the time of initial RSA or an irreparable subscapularis in all their cases of dislocation.^45,52 On the other hand, Wall reported that repair of the subscapularis was not significantly related to postoperative complications.¹³⁴ A superolateral approach that spares the anterior structures plays logically a protective role.²¹

Acromial fracture

Pre- or postoperative acromial pathology, which could theoretically compromise deltoid condition and affect the proper function of the prosthesis, is of legitimate concern. Postoperative fractures occur at least in 3% of cases⁹⁸ and their causes are numerous. Preoperatively, the acromion may be subject to a congenital or acquired abnormality such as an os acromiale. It can also already be eroded, fragmented or even fractured from the underlying head in case of cuff tear arthropathy (Figure 5), or osteoporosis-induced insufficiency.⁸⁵

Figure 5: Preoperative insufficiency of a left acromion on an AP view. Note that a

large part of the acromion just seems to have disappeared.

The superior metaglene fixation screw may function as a stress riser that results in acromial fractures.³³ They can be classified as avulsion fractures of the anterior acromion (Type I), fractures of the acromion posterior to the acromioclavicular joint (Type II) and fractures of the scapular spine (Type III).³³

Postoperatively, the arm is lengthened by approximately 24 to 27 mm above normal length (Appendix 1-4). Biomechanically, the tension on the deltoid and the acromion is thus more important due to arm lengthening and increased longer lever arm. The symptomatology usually appears within the first year with sudden pain and decrease of function. The localization of the former is typically posterior. The fracture is best seen on an axillary lateral view to differentiate acromial fracture from scapular spine fracture (Figure 6).

Figure 6: Postoperative acromial fracture (blue arrow).

The author believes that the use of positron emission tomography-computed tomography is helpful in diagnosis of non-displaced fractures. Good results have surprisingly been reported in patients with preoperative acquired or congenital acromial pathology or postoperative acromial fracture.⁹⁹ This can be explained by the persistent attachment of the deltoid to the spine of the scapula and clavicle and the more predominant postoperative scapulothoracic motion compared to the glenohumeral one. The best treatment option for acromial fractures is thus conservative, as it does not lead to major shoulder dysfunction. Outcome of scapular spine fractures are more unpredictable with displacement of the bony support for the entire deltoid, pain and dysfunction. Consequently, some authors recommend open reduction and internal fixation (ORIF) associated with postoperative immobilization on a 60° abduction splint in order to avoid nonunion and acromio-humeral contact secondary to inferior acromial tilt.^33,136

Neurological lesion

Potentially relevant neurological complications involving the brachial plexus or the axillary nerve are considered as rare.13,14,20,89,109 However, the author of the present thesis did a study that compared prospectively the prevalence of neurological lesion after anatomic and reverse shoulder arthroplasty. The conclusion of this study was that the occurrence of peripheral neurologic lesions following RSA is relatively common but usually transient, with arm lengthening as a main risk factor. All the current knowledge about this troublesome complication is summarized in Appendix 4 and 5.

Functional impairment

The evaluation of functional outcomes have shown variable results concerning the range of motion.^20,134 Poor postoperative AAE can be attributed to improper use, poor patient selection, and preoperative AAE as well as postoperative problems. The author postulates that failure to restore sufficient deltoid tension and the superior approach that splits the deltoid (future motor of the prosthesis) could also be responsible for poor postoperative function. The research done on these two hypotheses are resumed in Appendix 2 and 3, respectively. The aim of theses studies was also to correlate humeral and arm lengthening to postoperative AAE and determine the ideal soft-tissue tension to provide the best functional results.

Surgeon’s experience

It has previously been demonstrated that a surgeon’s experience can lead to a dramatic decrease in complications over time when total shoulder arthroplasty is concerned.²⁸ The author made the same observation with RSA (Appendix 7).¹³⁰

Other complications

Glenoid or humeral non- or disassembly, and polyethylene disassociation are minor problems and are mainly due to prosthetic design (Figure 7).

Figure 7: Glenoid non-assembly. Immediate postoperative imaging revealing a non- assembly (black doted lines) of the glenosphere on the metaglene.

From: Courtesy from K. Yamaguchi.

Humeral fractures occurred intra- or postoperatively. Intraoperatively, they can appear in the metaphyseal area (“controlled fracture” according to Walch) and are related to retractor positioning. The author prefers to rim the metaphyseal area after glenoid preparation in order not to weaken the humerus. Humeral diaphyseal fractures occur intraoperatively in case of an incorrect sizing of the component or excessive external rotation during preparation of the glenoid and release. They usually require the use of a longer implant to bypass the fracture line or an ORIF.

Postoperatively, fractures usually result from trauma (Figure 8). They can be treated either conservatively if the component is stable or they require revision in cases of unstable components.

Figure 8: Patient known for a right RSA that sustained a fall on the ipsilateral elbow.

A transverse supracondylar fracture of the distal humerus is noted on anteroposterior and lateral views.

Conclusion

Nowadays, RSA is a commonly performed procedure and its indications continue to widen. The high complication rate of RSA is accepted due to the usual postoperative improvement of shoulder function. Furthermore, prevention of most complications likely to lead to poorer function, still have to be further studied in detail. This has been the goal of the author of the present thesis.

Adequate deltoid tension is accepted as a key to prosthetic function and stability (Appendix 1).^21,81 This tension is determined by arm length. The latter is dependant of 1) the position of the glenosphere in the frontal plane, 2) the size of the glenosphere, 3) the use of an eccentric or inferiorly tilted glenosphere, 4) the use of a spacer, 5) the thickness of the polyethylene, and 6) the height of humeral cut and stem implantation (Figure 9) (Appendix 2).⁸¹

Figure 9: Influence of humeral cut on arm length. (A) Preoperative status with a lack of deltoid tension. (B-C) A low humeral cut induces a low implantation of the stem with a lack of deltoid retensioning. (D-E) A high humeral cut leads to a high implantation of the prosthetic stem with adequate deltoid retensioning.

From: Lädermann A, Walch G, Lubbeke A, Drake GN, Mélis B, Bacle G, Collin P, Edwards TB, Sirveaux F. Influence of arm lengthening in reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2012 Mar; 21(3): 336-41.

The glenosphere has to be implanted on the lower part of the glenoid to avoid notching (Appendix 6).¹⁰⁵ The type of glenosphere (size, eccentricity) allowed the adjustment of arm length by several millimeters (about 1% of arm length).

Consequently, the only key factor for arm length is humerus length (including height of stem implantation, polyethylene thickness and spacer), as these factors permit the correction of arm length by several centimeters (about 10% of arm length).

There is currently no described standard technique to determine deltoid tension or deltoid length with implant position, preoperatively. Intraoperative criteria have been proposed by other authors to assess prosthetic stability. The recommendations are numerous and include 1) a prosthesis implantation in such a way that it is difficult to reduce, 2) the absence of pistoning of the prosthesis when applying axial traction on the arm, 3) stability throughout a full range of motion, 4) passive adduction of the arm with elbow at side, 5) palpation of the tension in the conjoint tendon after reduction with the arm at the side and the elbow extended,²¹ 6) no asymmetric subluxation or tilting of the proximal humeral component on the glenosphere during adduction,⁶² and 7) free glenohumeral motion without scapula-thoracic motion between 0° to 60° of abduction.¹²⁷ Unfortunately, these intraoperative criteria are qualitative, completely subjective, and depend more on patient relaxation (i.e., depth of anesthesia and quality of muscle relaxation) and preoperative scar tissue (i.e., post-traumatic arthritis

in order to assess the appropriate length of the deltoid or the arm. Some authors even recommend the use of a «Jedi skill that involves using the force» rather than the previously mentioned intraoperative and unreliable criteria.¹⁰⁸

A preoperative guide, useful in complex cases such as revision arthroplasty or post- traumatic arthritis where scar tissue and bone loss prevent making an accurate determination of humeral length, has thus been proposed in the Appendix 1.

Preoperative planning is probably not necessary in all primary cases; its use in revision cases seems however mandatory.

In order to guarantee the best possible functional results, restoration of proper humeral and arm length has to be aimed for (Appendix 2-3). Implantation of the humeral stem at the level of the humeral cut with polyethylene thickness compensation in case of low cut, seems to be a reasonable option. The failure to restore sufficient tension in the deltoid can be responsible for poor AAE and prosthetic instability (Appendix 1-3). These results have been recently confirmed by other authors.⁷⁰ However, at the other end of the spectrum, a consequent lengthening may be responsible for neurological lesion, acromial or spine fracture or permanent arm abduction (Appendix 1, 4). One study (Appendix 4) demonstrated a substantially high prevalence of acute postoperative subclinical neurologic lesions after RSA. If one also takes into account subclinical worsening of preoperative lesions, 63% of the patients had postoperative neurologic lesions. The occurrence of peripheral nerve lesions determined by EMG analysis following RSA is thus common, but patients usually recover. The author hypothesized that lengthening of the arm during this procedure, because of its nonanatomic design and/or maneuver of

glenohumeral reduction, might be a major factor responsible for the increased prevalence of neurologic injury, hypothesis that as been confirmed.

Another study (Appendix 5) showed that the axillary nerve is not in close proximity to the glenosphere after implantation of a reverse shoulder arthroplasty. The distance between the nerve and the glenosphere was systematically greater than 15 mm.

Inferior glenosphere overhanging does not appear to decrease the distance to the axillary nerve. Rather, it seems that the lengthening of the arm after a RSA leads to a lowering and lateralization of the nerve and protects it from impingement with the glenosphere component. Therefore, the position of the glenosphere in the vertical plane is probably not related to the development of a neurological lesion due to direct contact.

Another factor contributing to the development of an axillary nerve lesion after lengthening of the arm could be the course of the main anterior circumflex branch of the axillary nerve which is relatively fixed around the humeral metaphysis. Close relationship between the main anterior circumflex branch and the humeral metaphysis was obvious in two out of six specimens (33%) (Figure 10).

Figure 10. Posterior view of right shoulder showing in detail the quadrangular space after implantation of a reverse shoulder arthroplasty. The main anterior circumflex branch of the axillary nerve courses distally around the metaphysis and is in close contact with the humeral trial liners. (AN (black arrows), main anterior circumflex

branch of the axillary nerve; *, blue humeral liners; D, deltoid muscle; TB, long head of the triceps brachii muscle; TM, teres minor).

From: Appendix 5.

Caution should therefore be observed when reaming the metaphysis to avoid posterior humeral cortical violation, particularly when using a large reamer. To prevent such lesions, the use of a combination of polyethylene adaptor systems that allow the use of a large glenosphere with a small metaphysis might be an option.

Lateral offset of glenosphere from the glenoid surface may be another possibility since this approach provides sufficient stability without excessive lengthening. This approach may also theoretically relieve tension in the axillary nerve in the quadrilateral space. Further studies are needed to examine this relationship.

Despite the proximity of the nerve to the prosthesis and the humeral metaphysis, there are few case reports in the literature of clinically evident (as oppose to

electrodiagnostically apparent) postoperative axillary nerve injury. This can be explained by the fact that transient partial lesions (neurapraxia) of the main anterior circumflex branch of the axillary nerve, which leads to weakness of the deltoid muscle, can remain unnoticed in the postoperative period. Additionally, more severe lesions such as axonotmesis can be compensated for either by other motor branches in a process of intramuscular collateral reinnervation, or by muscle hypertrophy.

The risk of neurological lesions increases drastically above 4 cm of lengthening. A reasonable goal for arm lengthening should thus be between 0 and 2 cm. As a result strategies have been developed to limit the lengthening. In difficult cases with a high risk of dislocation, large glenoid size with shallow concave components, superior approach and prosthetic or bony lateralization of the glenosphere should be considered to avoid excessive tension.^17,133

The radiological analysis suggests that, compared with anatomic shoulder arthroplasties, the constraints associated with the RSA are predominantly located on the humeral side rather than on the glenoid side, which is protected by the medialization of the glenoid implant. The RSA does have congruent joint surfaces making it a semi-constrained prosthesis, which explains these mechanical complications. We observed eight cases of humeral loosening and no cases of glenoid loosening. The results confirm that the rate of humeral loosening is much higher than after inserting an anatomical prosthesis, and much greater than glenoid loosening after RSA. However, with cemented as well as with uncemented components, resorption of the tuberosities and stress shielding had no effect on the

The best way to avoid complication in RSA is probably acquired surgical experience (Appendix 7). Experience led surgeons to maintain the indications involving the best functional outcomes and to limit the indications with bad functional outcome that had a high complication rate. Experience leads to better results with a statistically significant difference in the Constant score. This improved shoulder function could be explained by three phenomena: a better preoperative shoulder function involving an obviously better capacity to gain postoperative functional condition; a careful patient selection with a decrease of revision arthroplasty; and, finally, a better technical mastery as a consequence of the learning curve. Moreover, this increase in postoperative function was associated with a statistically significant decrease in overall complications. The three main complications of the first study (dislocations, infection, and glenoid loosening) diminished drastically. The main causes of instability reported in the literature are the release of the anteroinferior soft tissues during deltopectoral approach, shortening of the humerus, and failure to restore correct tension of the deltoid. The repair of the subscapularis tendon, whenever possible, and a better tension of the deltoid obtained by a correct restoration of the length of the humerus could explain this decline. The decrease in infections could mainly be explained by an increase of the surgeons’ experience, which involved less operative time, and by the reduction in the number of revision arthroplasty cases performed in the second series. The decrease in glenoid loosening is most likely to be related to 1) a better surgical technique with improvement in baseplate position and fixation, 2) correct positioning of the central peg into the native bone, 3) the use a central screw fixation, 4) and inferior tilt in order to limit shear forces causing loosening.

This decrease in complication rate allowed to progressively extend the indications.

Thus, preoperative deltoid insufficiency, in certain circumstances, is not no longer an absolute contraindication to RSA. This treatment option can yield reliable improvements in AAE and functional outcome (Appendix 8).

Appendix

Appendix 1

Objective evaluation of lengthening in reverse shoulder arthroplasty

Alexandre La¨dermann, MD^a, Matthew D. Williams, MD^b, Barbara Melis, MD^c, Pierre Hoffmeyer, MD^a, Gilles Walch, MD^c,*

aDepartment of Orthopaedic Surgery, University Hospital of Geneva, Switzerland

bAcadiana Orthopaedic Group, Lafayette, LA

cDepartment of Orthopaedic Surgery, centre orthope´dique Santy, Lyon, France

Background:Reverse shoulder arthroplasty requires a re-tensioning of the deltoid to obtain active eleva- tion and implant stability. Currently, there is no objective and reliable technique described for the preop- erative planning of reverse shoulder prosthesis or the postoperative evaluation of deltoid tension and arm lengthening. The purpose of this investigation was to outline a standardized technique for measuring deltoid length and to preoperatively plan a reverse shoulder arthroplasty, and to determine whether compli- cations are related to inadequate deltoid lengthening.

Methods:Fifty-eight patients were included in this radiographic review. Variations in humeral length, overall arm length, and the height of the subacromial space were evaluated before and after reverse shoulder arthroplasty.

Results:The average postoperative lengthening of the humerus was 2!7 mm (range, -9-16,P¼.243) and the arm was lengthened 23!12 mm (range, 1-47,P<.001). Measured preoperative and postoperative differences of the subacromial space were statistically significant when comparing the operated and contra- lateral arm (P<.0001). Lengthening was not correlated to sex (P¼.242), acromial fractures, or neuro- logical complications (P¼.83). However, in cases of postoperative instability, both humeral and overall arm lengthening were statistically lower (P<.0001).

Conclusion:A technique to preoperatively plan adequate deltoid tensioning using radiographs of the contralateral arm is described. This technique is critical in challenging cases and postoperatively in cases of complication to assess the deltoid length. Subjective intraoperative criteria to evaluate deltoid tension should be replaced by objective measures to prevent insufficient or excessive deltoid tension.

Level of evidence:Level 3.

!2009 Journal of Shoulder and Elbow Surgery Board of Trustees.

Keywords:Shoulder; reverse arthroplasty; lengthening; planning; dislocation; complications

The Grammont designed reverse shoulder arthroplasty reverses the ball-and-socket relationship of the shoulder joint.^{3, 4}The prosthesis position medializes and lowers the glenohumeral center of rotation, thereby increasing the

lever arm of the deltoid muscle. Deltoid tension, produced by the lowered center of rotation, increases muscle fiber recruitment of the anterior and posterior deltoid that compensates for a deficient rotator cuff. The tensioned deltoid provides the stable fulcrum essential for active elevation of the shoulder and prosthesis stability.²Due to the semi-constrained design of the Grammont reverse shoulder, deltoid tensioning is crucial. Failure to adequately

*Reprint requests: Gilles Walch, MD, Department of Orthopaedic Surgery, Centre Orthope´dique Santy, 24 Avenue PAUL SANTY, 69008 Lyon, France.

E-mail address:walch.gilles@wanadoo.fr(G. Walch).

J Shoulder Elbow Surg (2009) 18, 588-595

www.elsevier.com/locate/ymse

1058-2746/2009/$36.00 - see front matter!2009 Journal of Shoulder and Elbow Surgery Board of Trustees.

doi:10.1016/j.jse.2009.03.012

Appendix 2

Influence of arm lengthening in reverse shoulder arthroplasty

Alexandre L€adermann, MD^a,*, Gilles Walch, MD^b, Anne Lubbeke, MD, DSc^a, Gregory N. Drake, DO^c, Barbara Melis, MD^b, Guillaume Bacle, MD^b,

Philippe Collin, MD^d, T. Bradley Edwards, MD^c, Franc¸ois Sirveaux, MD^e

aService de Chirurgie Orthop"edique et Traumatologie de l’Appareil Moteur, H^opitaux Universitaires de Gen#eve, Geneva, Switzerland

bD"epartement de Chirurgie Orthop"edique, Centre Orthop"edique Santy et H^opital Priv"e Jean Mermoz, Lyon, France

cFondren Orthopedic Group, Houston, TX, USA

dCentre Hospitalier Priv"e Saint Gr"egoire, Saint-Gr"egoire, France

eCentre Chirurgical Emile Gall"e, Nancy, France

Background:Reverse shoulder arthroplasty (RSA) can improve anterior active elevation (AAE) by length- ening of the deltoid and hence increasing its lever arm. However, evaluations of functional outcomes of RSA have shown variable improvements in the range of motion. The aim of our study was to correlate humeral and arm lengthening to postoperative AAE.

Methods:We reviewed 183 RSAs with a minimum follow-up of 1 year. Lengthening of the humerus and the arm was evaluated in relation to the contralateral side.

Results:We observed mean humeral lengthening of 0.2!1.4 cm (range,"4.7 toþ5.2 cm) and mean arm lengthening of 1.6!1.9 cm (range,"5.1 toþ5.4 cm). Postoperative AAE was 140^$!27^$ (range, 30^$to 180^$). We found no significant correlation between lengthening or shortening of the humerus and AAE (P¼.169). Shortening of the arm led to a mean AAE value of 122^$; lengthening of 0 and 1 cm, mean AAE of 140^$; lengthening of greater than 1 cm to 2.5 cm, mean AAE of 144^$; and lengthening of greater than 2.5 cm, mean AAE of 147^$. When we compared patients with lengthening of the arm and those with shortening, the postoperative AAE was significantly greater after arm lengthening, 145^$versus 122^$, with a mean difference of 23^$(95% confidence interval, 13^$to 33^$) (P<.001).

Conclusion:This study shows that shortening of the arm reduced AAE. With respect to arm lengthening, a lengthening threshold was not found. An objective assessment of deltoid lengthening is possible preop- eratively, intraoperatively, and postoperatively, and this measure seems to correlate with the functional outcome.

Level of evidence:Level IV, Case Series, Treatment Study.

!2012 Journal of Shoulder and Elbow Surgery Board of Trustees.

Keywords:Reverse shoulder arthroplasty; function; humeral and arm lengthening

Ethical Committee approval was received from the Institutional Review Board of H^opital Priv"e Jean Mermoz and Centre Orthop"edique Santy, Lyon, France (Reference Study 20).

*Reprint requests: Alexandre L€adermann, MD, Department of Surgery, Division of Orthopaedics and Trauma Surgery, Geneva University Hospitals, Rue Gabrielle-Perret-Gentil 4, CH-1211 Geneva 14, Switzerland.

E-mail address:Alexandre.laedermann@hcuge.ch(A. L€adermann).

J Shoulder Elbow Surg (2012) 21, 336-341

www.elsevier.com/locate/ymse

1058-2746/$ - see front matter!2012 Journal of Shoulder and Elbow Surgery Board of Trustees.

doi:10.1016/j.jse.2011.04.020

Appendix 3

Orthopaedics & Traumatology: Surgery & Research (2011)97, 579—582

ORIGINAL ARTICLE

Influence of surgical approach on functional outcome in reverse shoulder arthroplasty

A. Lädermann^a,∗, A. Lubbeke^a, P. Collin^b, T.B. Edwards^c, F. Sirveaux^d, G. Walch^e

aDepartment of Orthopaedic Surgery and Traumatology, Geneva University Hospitals, 4, rue Gabrielle-Perret-Gentil, 1211 Genève, Switzerland

bSaint-Grégoire Private Hospital Center, 6, boulevard Boutière, 35768 Saint-Grégoire cedex, France

cFondren Orthopedic Group, Houston, Texas, USA

dCentre chirurgical Émile-Gallé Surgical Center, 49, rue Hermite, 54000 Nancy, France

eOrthopaedic Center and Jean-Mermoz Private Hospital, 24, avenue Paul-Santy, 69008 Lyon, France

Accepted: 11 April 2011

KEYWORDS Reverse shoulder arthroplasty;

Function;

Surgical approaches

Summary

Introduction:Reverse shoulder arthroplasties (RSA) can be performed using a Deltopectoral (DP) or alternatively a Transdeltoid (TD) approach.

Hypothesis:Although the humeral cut is lower by TD approach, this should not affect postop- erative functional results.

Material and methods:This retrospective multicentric study evaluated the complete medical records of RSA implanted between October 2003 and December 2008. Inclusion criteria were:

follow-up of at least 1 year, a complete file including a comparative radiological work-up mak- ing it possible to analyze eventual arm and humeral lengthening. Evaluation of postoperative function was based on Active Anterior Elevation (AAE).

Results:We studied 144 RSA in 142 patients. One hundred and nine RSA were implanted by the DP approach and 35 by the TD approach. Mean lengthening of the humerus compared to the controlateral side by DP approach was 0.5±1.3 cm while there was a mean shortening of

−0.5±1.0 cm by TD approach (P< 0.001). The difference in cut was partially compensated by using thicker polyethylene inserts with the TD approach. Mean arm lengthening compared to the controlateral side was 1.7±1.7 cm by DP approach and 1.2±1.4 cm by TD approach (mean difference 0.5 cm; (95% CI−0.1; 1.2). AAE for RSA by DP approach was 145±22^◦and 135±29^◦ by TD approach (mean difference 10^◦, 95% CI−1; 21).

∗Corresponding author. Tel.: +41 22 372 79 08; fax: +41 22 372 79 03.

E-mail address:alexandre.laedermann@hcuge.ch(A. Lädermann).

1877-0568/$ – see front matter © 2011 Elsevier Masson SAS. All rights reserved.

doi:10.1016/j.otsr.2011.04.008

Appendix 4

Prevalence of Neurologic Lesions After Total Shoulder Arthroplasty

A. L¨adermann, MD, A. L¨ubbeke, MD, DSc, B. M´elis, MD, R. Stern, MD, P. Christofilopoulos, MD, G. Bacle, MD, and G. Walch, MD

Investigation performed at the Division of Orthopaedics and Trauma Surgery, Department of Surgery, Geneva University Hospitals, Geneva, Switzerland, and the Department of Orthopaedic Surgery, Centre Orthop´edique Santy, Lyon, France

Background: Clinically evident neurologic injury of the involved limb after total shoulder arthroplasty is not uncommon, but the subclinical prevalence is unknown. The purposes of this prospective study were to determine the subclinical prevalence of neurologic lesions after reverse shoulder arthroplasty and anatomic shoulder arthroplasty, and to evaluate the correlation of neurologic injury to postoperative lengthening of the arm.

Methods: All patients undergoing either a reverse or an anatomic shoulder arthroplasty were included during the period studied. This study focused on the clinical, radiographic, and preoperative and postoperative electromyographic evalu- ation, with measurement of arm lengthening in patients who had reverse shoulder arthroplasty according to a previously validated protocol.

Results: Between November 2007 and February 2009, forty-one patients (forty-two shoulders) underwent reverse shoulder arthroplasty (nineteen shoulders) or anatomic primary shoulder arthroplasty (twenty-three shoulders). The two groups were similar with respect to sex distribution, preoperative neurologic lesions, and Constant score. Electromyog- raphy performed at a mean of 3.6 weeks postoperatively in the reverse shoulder arthroplasty group showed subclinical electromyographic changes in nine shoulders, involving mainly the axillary nerve; eight resolved in less than six months. In the anatomic shoulder arthroplasty group, a brachial plexus lesion was evident in one shoulder. The prevalence of acute postoperative nerve injury was significantly more frequent in the reverse shoulder arthroplasty group (p=0.002), with a 10.9 times higher risk (95% confidence interval, 1.5 to 78.5). Mean lengthening (and standard deviation) of the arm after reverse shoulder arthroplasty was 2.7±1.8 cm (range, 0 to 5.9 cm) compared with the normal, contralateral side.

Conclusions: The occurrence of peripheral neurologic lesions following reverse shoulder arthroplasty is relatively common, but usually transient. Arm lengthening with a reverse shoulder arthroplasty may be responsible for these nerve injuries.

T

dissection, compression secondary to retractors or postoper- ative hematoma, excessive mobilization of the limb, vascular injury, humeral shaft fractures, cement extrusion, and possibly interscalene block^2-4. All of these factors are similar whether a reverse or an anatomic shoulder replacement is implanted.

The prevalence of clinically evident neural impairment of the involved limb has been reported to be 2% in patients after reverse shoulder arthroplasty⁵and 1% to 4.3%^1,6-8following an- atomic shoulder arthroplasty. Nagda et al. documented episodes of nerve disturbance in up to 57% of patients undergoing Disclosure:None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The completeDisclosures of Potential Conflicts of Interestsubmitted by the authors of this work are available with the online version of this article at jbjs.org.

1288

C^OPYRIGHT!2011^BYT^HEJ^{OURNAL OF}B^{ONE AND}J^OINTS^URGERY, INCORPORATED

J Bone Joint Surg Am.2011;93:1288-93 ^d doi:10.2106/JBJS.J.00369

Appendix 5

Elsevier Editorial System(tm) for Journal of Shoulder and Elbow Surgery Manuscript Draft

Manuscript Number:

Title: Injury to the Axillary Nerve During Reverse Shoulder Arthroplasty: An Anatomic Study Article Type: Original Article

Keywords: Axillary nerve lesion; reverse shoulder arthroplasty; complications; prosthesis.

Corresponding Author: Dr Alexandre Lädermann, M.D.

Corresponding Author's Institution: Latour Hospital,Faculty of Medicine, University of Geneva First Author: Alexandre Lädermann, M.D.

Order of Authors: Alexandre Lädermann, M.D.; Bojan V Stimec; Patrick J Denard; Gregory Cunningham; Philippe Collin

Abstract: Background

Subclinical neurological lesions after reverse shoulder arthroplasty are frequent, mainly those involving the axillary nerve. One of the major reported risk factors is postoperative lengthening of the arm. The purpose of this study was to evaluate the anatomical relationship between the axillary nerve and prosthetic component after reverse shoulder arthroplasty. The study hypothesis was that inferior overhang of the glenosphere relative to glenoid could put this nerve at risk.

Material and methods

Six shoulder specimens were dissected after having undergone reverse shoulder arthroplasty.

Results

The mean distance from the inferior border of the glenoid to the inferior edge of the glenosphere was 5.7±4.0 mm (range, 2.0 to 12.6 mm). The axillary nerve was never within 15 mm of the glenosphere.

The main branch of the axillary nerve was in close contact with the posterior metaphysis or humeral prosthetic implant. The mean distance between the nerve and the humeral implants was 5.8±2.4 mm (range, 1.2 to 8.1 mm).

Conclusions

The proximity of the axillary nerve to the posterior metaphysis or humeral implants may be a risk factor for axillary nerve injury after reverse shoulder arthroplasty.

Introduction

The Grammont designed shoulder arthroplasty reverses the ball-and-socket relationship of the shoulder. The medialization of the center of rotation optimizes the deltoid lever arm, and by lowering the humerus relative to the acromion, reestablishes the tension of the deltoid thus allowing this muscle to produce shoulder range of motion even in the absence of the rotator cuff. Lowering the humerus lengthens the arm, which can be increased by using a larger or eccentric glenosphere, or positioning the glenosphere on the lower part of the glenoid surface.

The latter can lead to an inferior overhang of the glenosphere that may decrease the rate of scapular notching.^21,105 It has been shown that lengthening of the arm is necessary to obtain good postoperative function.⁸¹ However, lengthening is also one of the major risk factors for postoperative neurological lesions.⁷⁹

Neurological lesions are frequent following reverse shoulder arthroplasty; half of these cases involve the axillary nerve that provides innervation for the essential deltoid muscle.⁷⁹ Although most of these axillary nerve lesions do not occur as flaccid paralysis, they may be responsible for postoperative pain, weakness, dislocation, or impair rehabilitation. These lesions could theoretically occur from direct nerve damage during surgical dissection, compression secondary to retractors or postoperative hematoma, excess mobilization of the limb, vascular injury, humeral shaft fractures, cement extrusion, and possibly interscalene block.^29,101,118 However, a previous study demonstrated that the previously mentioned factors did not seem to play a significant role.⁷⁹ The vulnerability of the axillary nerve compared to the rest of

course around the humerus. Previous studies that examined its location during other types of glenohumeral procedures did not take into account the changes induced by reverse shoulder arthroplasty.44,69,92,142 Therefore, the position of the nerve relative to the glenosphere, humerus and implants, has not been quantified. The anatomical position of the axillary nerve could make it more specifically vulnerable to injury due to lengthening of the arm and eventually to compression in cases of secondary impingement. An appreciation of this proximity may help shoulder surgeons avoid iatrogenic injuries that can be devastating.

The aim of this cadaveric study was to evaluate the anatomical relationship between the axillary nerve and the prosthetic component after reverse shoulder arthroplasty.

Our hypothesis was that inferior overhang of the glenosphere would decrease the distance between reverse shoulder arthroplasty implants and the axillary nerve. This relationship may explain the high rate of axillary nerve lesions following reverse shoulder arthroplasty.

Materials and Methods

Specimens

Six fresh-frozen human cadaveric shoulders were dissected after thawing. The mean donor age was 90.0 years (range, 85 to 91 years). Five of the donors were male and one was female. The specimens were mounted in a simulated beach-chair position, secured with a clamp on the medial scapula, and mounted onto an aluminum frame.

Surgical technique

The surgical technique was a standard deltopectoral approach and a subscapularis tenotomy was used to provide access to the glenohumeral joint.¹³² A Delta 3 reverse shoulder arthroplasty implant was used in all cases (Delta III; DePuy, Johnson &

Johnson, Leeds, UK). The size of the metaglene was 27 mm in all cases. The metaglene was implanted low on the glenoid to simulate the ideal position to avoid scapular notching. A 38 mm glenosphere was implanted in 3 cases, and a 42 mm glenosphere was implanted in 3 cases. Concentric glenospheres were implanted in 3 cases and eccentric glenospheres of 4 mm were used in 3 specimens. The lateral landmark for the humeral cut was the top of the greater tuberosity. All the humeral stems were non-cemented. The retroversion of the stem was determined according to the anatomy of the patient, but a maximum of 20° was tolerated. Non-constrained humeral liners of 6 mm were then placed on the humeral components.

Dissection

The axillary nerve anatomy has been well described.³ The nerve originates from the posterior cord of the brachial plexus, runs anteriorly towards the subscapularis muscle and posterior to the axillary artery, passing under the inferior capsule between the glenoid rim and the humeral metaphysis, supplying a branch to the shoulder joint, and crosses the quadrilateral space. At this point, the nerve splits into a main anterior circumflex division, which innervates the deltoid muscle and provides a few cutaneous filaments. A posterior division gives the superior-lateral brachial