Arthrodesis of the Shoulder Abstract Shoulder arthrodesis is an end-stage salvage option for the failing, painful joint that cannot undergo or has failed reconstruction. It is indicated for irreversible and nonreconstructible massive rotator cuff tears and deltoid muscle denervation as well as for detachment of the deltoid from its origin. Rarely, arthrodesis is done to stabilize the glenohumeral joint after many failed attempts at shoulder reconstruction. Arthrodesis for failed prosthetic arthroplasty or tumor resection presents additional challenges because of the associated bone loss on the humeral and/or glenoid side of the joint. Primary arthrodesis requires rigid internal plate fixation and both an extra- and an intra-articular site of fusion. Depending on bone volume and quality needed, the patient may require bracing for 8 to 10 weeks, autogenous or allograft bone grafting, or a vascularized fibular bone graft to reconstruct the bone deficiency, along with prolonged spica cast immobilization. The optimal position for arthrodesis is 20° of forward flexion, 20° of abduction, and 40° of internal rotation, with modifications based on patient body size or other patient-specific factors. Bone fusion is attained in nearly all patients, with marked pain reduction and improved function. Postoperatively, the patient should be able to lift the arm to near shoulder height and to reach the top of the head, the mouth, the ipsilateral back pocket, and the groin. Complications include nonunion, malposition, pain associated with prominent hardware, and periarticular fractures. T he development of prosthetic shoulder arthroplasty has nearly eliminated the need for arthrodesis in primary arthritic joints. Arthrodesis has become an end-stage salvage pro- cedure for shoulder pain, weakness, and instability not suitable for soft- tissue or prosthetic reconstruction. 1 The principal indications include conditions that result in severe and irreparable deltoid and rotator cuff deficiency, caused by irreversible pa- ralysis of these muscles with preser- vation of the scapular muscles. In such refractory cases, arthrodesis may offer significant pain reduction and some functional use of the upper extremity in what would other wise be an unsolvable situation. Indications Currently, shoulder arthrodesis is in- dicated for brachial plexus injury, failed prosthetic arthroplasty, recon- struction after tumor resection, chronic infection, and refractory in- stability and pseudoparalysis of the shoulder secondary to combined ro- tator cuff and deltoid muscle dys- function. Motor vehicle accidents account Ori Safran, MD Joseph P. Iannotti, MD, PhD Dr. Safran is Senior Orthopaedic Surgeon, Department of Orthopaedic Surgery, Hadassah-Hebrew University Medical School, Jerusalem, Israel. Dr. Iannotti is Professor and Chairman, Department of Orthopaedic Surgery, The Cleveland Clinic Foundation, Cleveland Clinic Lerner School of Medicine of Case Western Reserve University, Cleveland, OH. None of the following authors or the departments with which they are affiliated has received anything of value from or owns stock in a commercial company or institution related directly or indirectly to the subject of this article: Dr. Safran and Dr. Iannotti. Reprint requests: Dr. Iannotti, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. J Am Acad Orthop Surg 2006;14:145- 153 Copyright 2006 by the American Academy of Orthopaedic Surgeons. Volume 14, Number 3, March 200 6 145 for most traumatic brachial plexus injuries. These devastating injuries affect young patients in almost 90% of cases. 2 When spontaneous recov- ery of the deltoid and rotator cuff muscles does not occur, procedures such as neurolysis, nerve grafting, and muscle transfer are warranted in an effort to reestablish shoulder function. Shoulder arthrodesis may be considered in the presence of se- verely restricted shoulder passive range of motion, significant bone loss, or failed prior reconstructive measures. 3 Because the trapezius and levator scapulae muscles are al- most always intact after traumatic brachial plexus injury, active arm abduction can occur through the scapulothoracic articulation. When the serratus anterior is preserved, forward elevation of the arm through scapular rotation is also pos- sible. 4 Severe humeral and glenoid bone loss, poor deltoid and rotator cuff function, and refractory instability (after multiple surgical procedures) may render functional improvement or pain relief via revision arthroplas- ty impossible in patients with failed shoulder arthroplasty. When arthro- desis is considered for these patients, special techniques are needed to compensate for the significant gle- noid and humeral bone loss. Limb-sparing resection for malig- nant and locally destructive tumors of the proximal humerus may result in significant tissue deficiency. The choice between prosthetic recon- struction and glenohumeral arthrode- sis is made based on the quality of the soft tissue and bone after completion of the curative resection. Frequently, tissue loss is severe, and arthrodesis is the only valid option for regaining shoulder stability. 5,6 To achieve fusion in these patients, vascularized au- tograft or bulk allograft must be used to compensate for bone deficiency. Patients with the combination of infection and a painful, damaged joint present a difficult challenge. Prosthetic reconstruction is con- traindicated when the septic process is not eradicated. Surgical débride- ment and glenohumeral fusion may yield a painless stable joint in many patients. 7 Most unstable shoulder pathology is managed through one or a combi- nation of the following procedures: soft-tissue balancing, muscle trans- fer, and bone block. With persistent refractory instability after multiple failed stabilization procedures, pa- tients may be left with a painful, dysfunctional shoulder devoid of competent soft-tissue or bony re- straints. Fusion is one option for re- gaining stability and some degree of function in these patients. 8 Arthropathy resulting from rota- tor cuff tear alone has not been con- sidered an indication for shoulder fu- sion. Currently, patients with a rotator cuff tear may be treated suc- cessfully with shoulder hemiarthro- plasty or reverse total shoulder ar- throplasty with acceptable results. However, in the presence of irrepara- ble rotator cuff tear coinciding with irreparable deficiency of the deltoid muscle or dysfunction, some au- thors recommend fusion. 9,10 Fusion is especially advocated in younger patients with demands for substan- tial strength at low angles of shoul- der flexion. Glenohumeral arthrodesis is gen- erally contraindicated in patients who lack functional scapulothoracic motion, which may be caused by pa- ralysis of the trapezius, levator scap- ulae, and serratus anterior muscles. Fusion is a less favorable option in patients at high risk of pseudarthro- sis, such as those with Charcot arthropathy. 10-12 Patients with ad- vanced bilateral shoulder disease should not undergo bilateral arthro- desis because positioning cannot be achieved to allow the normal perfor- mance of activities of daily living (ADLs). Finally, elderly patients or those with progressive neurologic disease generally do not achieve sat- isfactory results with shoulder ar- throdesis. Surgical Considerations A variety of periarticular fusion techniques, fixation methods, and immobilization strategies have been advocated for shoulder arthrodesis. Shoulder Position The ideal position for shoulder ar- throdesis remains a matter of debate. Most surgeons agree that the proper position of shoulder arthrodesis should enable the patient to reach the face for washing, the midline for dressing and hygiene, and the back pocket (Figure 1). It is important to minimize scapular winging in the resting position in order to minimize fatigue and periscapular muscle pain. 10 The exact position of fusion re- mains elusive. It is difficult to objec- tively and accurately measure the different angles, and the ideal posi- tion depends on patient body size. One of the first guidelines was made in 1942 by a committee of the Amer- ican Orthopaedic Association, 13 which recommended 50° of abduc- tion, 15° to 25° of flexion, and 25° of internal rotation. This recommenda- tion was later disputed by Rowe, 14 who found that the degree of abduc- tion suggested by the committee caused periscapular muscle pain be- cause of excessive loads and winging of the scapula in the resting position. He explained that excessive abduc- tion shifted the extremity away from the fulcrum of the shoulder, thus weakening its lifting and functional strength. Rowe suggested 20° to 25° of abduction (just enough to clear the axilla), 30° of forward flexion, and 45° to 50° of internal rotation to get to the midline of the body . Elbow flexion, in his opinion, would allow the patient to reach the head and face. 14,15 In contrast, in their review of 71 shoulder fusions, Cofield and Briggs 16 found that the amount of abduction and flexion did not correlate with ei- ther periscapular pain or patient sat- isfaction and ADL function. They did Arthrodesis of the Shoulder 146 Journal of the American Academy of Orthopaedic Surgeons find a correlation between the amount of internal rotation and var- ious ADLs. Later studies supported Rowe’s findings, although they fa- vored less internal rotation. 12,17,18 Groh et al 19 advocated a lesser degree of abduction and forward flexion of 10° to 15°, compensated with 45° of internal rotation, enabling the pa- tient to reach the mouth, belt buckle, and contralateral shoulder and axilla. Fusion of the shoulder joint in 20° of abduction, 20° of forward flexion, and 40° of internal rotation in most mesomorphic patients usually al- lows acceptable position. In obese patients, significantly greater abduc- tion is both required and well-toler- ated; the body habitus will not allow <20° of abduction. In all patients, at the conclusion of the procedure the surgeon should be able to bring the hand to the forehead with a combi- nation of shoulder elevation and el- bow flexion. In addition, the pa- tient’s arm should be able to rest at the side without excessive scapular winging. Extra-articular Versus Intra-articular Fusion Shoulder arthrodesis may be ac- complished by fusing the glenohu- meral joint, the acromiohumeral in- terface, or both. Extra-articular fusion involves fusion between the proximal humerus and the undersur- face of the acromion, scapula, and clavicle without entering the gleno- humeral joint. Extra-articular fusion alone was one of the first methods of shoulder arthrodesis. The technique involved decorticating the upper part of the proximal humerus, the undersurface of the acromion, the lateral portion of the spine of the scapula, and the distal clavicle. This procedure was initially advocated for treating tuberculosis patients with glenohumeral joint destruction dur- ing the era in which anti- tuberculosis antibiotics were not available. 20 Extra-articular fusions are rarely done as isolated proce- dures today. Intra-articular fusion involves preserving the rotator cuff tendons without attempting to fuse the hu- merus to the acromion. Limited in- ternal fixation is used to preserve most of the bone architecture and soft tissues. Some surgeons recom- mend intra-articular fusion for younger patients with glenohumeral joint destruction and intact rotator cuff tendons, with the expectation of possible future revision into shoul- der arthroplasty. 21 In 1992, Morgan and Casscells 22 reported on an arthroscopically assisted intra- articular glenohumeral arthrodesis in a 39-year-old woman with refrac- tory multidirectional shoulder insta- bility, intractable pain, and deltoid paralysis. After aggressive arthro- scopic débridement and decortica- tion of the glenohumeral joint, two cannulated compression screws were inserted across the glenohu- meral joint, followed by an acromio- humeral screw for extra-articular stabilization. The patient was placed in a foam abduction pillow for 4 weeks. Radiographs at 10 weeks confirmed solid glenohumeral bony Figure 1 The arthrodesis position should enable the patient to reach the face and hair for washing and combing (A), the midline for dressing (B), and the back pocket (C). Ori Safran, MD, and Joseph P. Iannotti, MD, PhD Volume 14, Number 3, March 2006 147 union. The patient was pain-free and able to perform ADLs with the fused shoulder. Currently, combined extra- and intra-articular fusion techniques are most commonly used. 4,10,12,16-18,23 Decorticating the humeral head, gle- noid, and acromion maximizes the surface area for bone contact, there- by increasing the probability of achieving solid fusion. Fixation Methods Arthrodesis fixation methods have undergone marked evolution. Historically, postoperative spica casting for 3 to 4 months was the only so-called fixation used, and loss of position was common. In 1964, Charnley and Houston 24 introduced external bone compression for achieving solid fusion of the shoul- der. Later, Beltran et al 25 introduced glenohumeral compression screws, which reduced the length of postop- erative immobilization to 4 weeks. Subsequently, compression screws became the accepted standard, 12,15,16 but most patients still required rela- tively prolonged periods of postoper- ative spica cast immobilization. In the past two decades, the additional stability provided by the combina- tion of compression screws and plates has minimized the need for strict postoperative immobilization without a concurrent increase in nonunion rates; this has become the accepted standard for glenohumeral arthrodesis. 4,7,8,10,18,23 Richards et al 4 used a single con- toured 4.5-mm dynamic compres- sion plate (DCP) placed over the spine of the scapula, the acromion, and the lateral portion of the humer- al shaft for shoulder arthrodesis in 14 patients. Postoperatively, patients were placed in a spica cast for an av- erage of 8 weeks. All patients achieved solid fusion; the two who had only acromiohumeral fusion re- mained stable and asymptomatic. Five years later, Richards et al 23 re- ported their use of a single 4.5-mm pelvic reconstruction plate on 11 pa- tients with brachial plexus palsy who underwent shoulder arthrode- sis; they were immobilized postoper- atively in a spica cast for 6 weeks. All achieved union by 5 months. The authors concluded that the more malleable reconstruction plate provided sufficient stability, a high union rate, and minimal skin prob- lems. Stark et al 18 used a long DCP in 15 patients undergoing arthrodesis, with no postoperative immobiliza- tion aside from an abduction pillow. Fusion was achieved in 14 of 15 pa- tients, and extremity position was lost in only one patient, in whom fixation was inadequate. Four pa- tients had local skin irritation and pain, with two ultimately requiring removal of the screws but not the plate. Clare et al 10 recommended the routine use of a 4.5-mm reconstruc- tion plate but favored the 4.5-mm DCP plate for patients weighing more than 100 kg. Use of the ex- tremity was allowed as soon as tol- erated. Routine use of compression screws traversing the plate and crossing the glenohumeral joint were used by all authors. External fixators also have been used in shoulder arthrodesis. Charn- ley and Houston 24 used an external fixator, achieving union in 18 of 19 patients, most with tuberculous joint destruction. Postoperatively, spica casts were used for additional stability in all patients. Johnson et al 26 and, later, Kocialkowski and Wallace, 27 reported using a combina- tion of compression glenohumeral screws and an external fixator. Johnson et al 26 used a Hoffman ex- ternal fixator and compression screws in four patients without addi- tional immobilization, and all fused. The fixator was removed 7 to 14 weeks after surgery; the authors re- ported only one pin tract infection (which responded to antibiotics and local wound care) and one fracture after frame removal. Nagano et al 28 reported on 11 patients who under- went shoulder arthrodesis for brachi- al plexus injury using an external fixator and temporary internal fixa- tion. All patients had solid fusion within 3 months, and the external fixator was removed 3 months after surgery. No complications were re- ported. Technique Preoperative planning is crucial for successful arthrodesis. Often, shoulder arthrodesis is performed af- ter several failed previous surgical attempts or in conjunction with a radical bone and soft-tissue resec- tion. Extensive bone loss may exist because of injury, infection, osteoly- sis after shoulder arthroplasty, or tu- mor resection. Consideration should be given to careful assessment of re- maining bone stock and soft-tissue deficiencies. Small cancellous bone deficiencies may be managed with local or iliac cancellous bone graft. More substantial deficits may re- quire a structural graft, such as a tri- cortical iliac graft, structural al- lograft, or vascularized fibular graft. The patient is anesthetized using a combination of a scalene block and general anesthesia, followed by su- pine beach chair positioning. An in- cision is made over the spine of the scapula, curving it anteriorly over the midacromion toward the antero- lateral corner of the acromion, then continuing over the lateral aspect of the arm toward the deltoid tuberos- ity. The deltopectoral interval is identified and developed. The anteri- or and middle portions of the deltoid are detached from the lateral third of the clavicle, anterior acromion, and lateral acromion. The deltoid is then retracted laterally and distally, hinged on its neurovascular pedicle. The subscapularis tendon is sharply dissected from the lesser tuberosity, and the supraspinatus tendon is ex- cised from its musculotendinous junction to the greater tuberosity. The glenoid is prepared by remov- ing its articular cartilage and cutting its bony surface flat and parallel with its original plane. The humeral Arthrodesis of the Shoulder 148 Journal of the American Academy of Orthopaedic Surgeons head is placed in contact with the prepared glenoid fossa, and the arm is placed in the preferred position and temporarily held in place by two Steinmann pins. Proper positioning is examined by taking the arm through a range of scapulothoracic motion, verifying a functional range. With the arm in the correct position, an oscillating saw is used to cut the medial portion of the humeral head. The undersurface of the acromion is cut to a flat decorticated surface, and the superior portion of the humeral head is cut parallel with it (Figure 2, A and B). A 4.5-mm pelvic reconstruction plate is contoured to accommodate the spine of the scapula, the lateral surface of the acromion, and the lat- eral portion of the proximal humer- us. Under fluoroscopic guidance, two partially threaded cancellous screws are placed through the hu- meral head into the glenoid surface and neck; at least one is placed through the plate. One or two simi- lar screws are inserted through the acromion into the humeral head, with at least one through the plate. An additional screw is placed through the plate and spine of the scapula into the base of the glenoid neck. The remaining screws are placed through the plate into the spine of the scapula and proximal humerus using a standard AO tech- nique (Figure 2, C and D). The fusion site is packed with local morcellized bone using pieces of bone removed during humeral and acromial prepa- ration. The subscapularis is then re- paired to the lesser tuberosity. The deltoid is reattached to the clavicle and to the anterior and lateral por- tions of the acromion through bone holes, as well as to the trapezius fas- cia. In the presence of severe bone stock deficiency (eg, after failed ar- throplasty), augmentation of the bone stock is needed. In these cases, bone contact may be improved by using the tuberosities as a local vas- cularized bone graft; they are fixed to the fusion region between the gle- noid and neck. Further stability may be achieved by adding a structural graft of tricortical iliac bone or al- lograft between the lateral humeral shaft and the decorticated undersur- face of the acromion. A pelvic recon- struction plate is used to fix the al- lograft in place (Figure 3). When >6 cm of the proximal hu- meral shaft is missing, vascularized fibular grafts are used to replace the absent segment, along with addi- tional iliac bone graft or allograft. For optimal docking, the fibular graft should be 4 to 5 cm longer than the humeral length needed. The distal 3 cm of the fibular graft is stripped of its soft tissue and is either placed in the humeral medullary canal or at- tached to the humeral outer cortex and fixed with cortical screws. The proximal part of the fibular graft is keyed into a trough burred within the glenoid fossa. The construct is stabilized with a 4.5-cm pelvic recon- struction plate. Partially threaded cancellous screws are placed through the fibula into the glenoid (Figures 4 and 5). After stable fixation is achieved, the fibular graft is revascu- larized. Thereafter, augmentation is done with autologous graft or frac- tionated bone marrow aspirate mixed with demineralized cortical fibers and cancellous chips. The material is used to fill the space between the proximal fibula, the glenoid, and the Figure 2 Stages of bone preparation and fixation. A, The medial aspect of the humeral head is cut parallel with the glenoid surface. The glenohumeral pins are removed before completing the cut. B, The undersurface of the acromion and superior surface of the humeral head are cut in parallel. C, A 4.5-mm pelvic reconstruction plate is used to firmly stabilize the arthrodesis. Partially threaded screws are used to compress the humerus to the glenoid and acromion. D, Anteroposterior radiograph taken a few weeks after arthrodesis. (Panels A, B, and C copyright Cleveland Clinic Foundation, 2003.) Ori Safran, MD, and Joseph P. Iannotti, MD, PhD Volume 14, Number 3, March 2006 149 undersurface of the acromion. After surgery, the shoulder is im- mobilized using a Southern Califor- nia Orthopaedic Institute (SCOI) brace (DonJoy, Vista, CA), abduction pillow, or spica cast. The extent of immobilization depends on the qual- ity and quantity of bone as well as the stability of the fixation. In a pri- mary fusion with good bone stock and stable fixation, the SCOI brace or abduction pillow is used for 8 to 10 weeks. In complex cases with poor bone quality, poor fixation, or structural grafting, the patient is im- mobilized in a spica cast for 3 to 4 months or until there is radiograph- ic evidence of fusion. When it is dif- ficult to determine bony fusion on radiographs, computed tomography scan reconstructions may be neces- sary. Scapular exercises for range of motion and strength are started after radiographic signs of fusion are found. Functional Results The objective of shoulder arthrode- sis is a painless, functional extremi- ty sufficient for most ADLs. Studies indicate that most patients report marked pain relief after surgery, al- though few are completely pain-free. Hawkins and Neer 12 reported that only 4 of their 16 patients were pain- free, whereas 9 needed analgesics on a daily basis for moderate or severe pain. Cofield and Briggs 16 reported better pain control in their series. Of the 65 patients, 25 (38%),were pain free, 24 (36%) had mild pain, 15 (23%) had moderate pain, and only 2 (3%) had severe pain. Of the 17 pa- tients with moderate or severe pain, 10 (59%) had pain located to the sur- gical area, 5 (29%) had pain in the periscapular region, and 2 (12%) had diffuse pain. Extremity function was limited by pain in most of the pa- tients with moderate or severe pain. Rouholamin et al 29 described ex- cellent pain relief in 10 of 15 pa- tients with brachial plexus injury Figure 3 A 65-year-old man presented with failed humeral arthroplasty of the right shoulder, an irreparable rotator cuff tear, and impaired deltoid muscle function. A, Preoperative anteroposterior radiograph demonstrating noticeable prosthetic loosening. B, Postoperative anteroposterior radiograph after application of a pelvic reconstruction plate. C and D, Functional range of motion was achieved. Figure 4 Arthrodesis with large proximal humeral bone deficiency. The fibular graft is 4 to 5 cm longer than the humeral length needed, for optimal docking. The proximal part of the fibular graft is keyed into a burred trough in the glenoid. The trough is fashioned to optimize bone contact and allow sufficient abduction. (Copyright Cleveland Clinic Foundation, 2003.) Arthrodesis of the Shoulder 150 Journal of the American Academy of Orthopaedic Surgeons who underwent shoulder fusion. Three patients reported aching with prolonged use of the arm; the pain was relieved with rest. Two patients with preoperatively diagnosed neu- rogenic pain continued to have pain. Excellent pain relief also was report- ed by Rybka et al 30 in their series of 41 patients with rheumatoid arthri- tis. Only four patients reported tran- sitory pain, which was thought to be caused by increased strain on the periscapular muscles. In a recent study, Wick et al 7 reported marked pain reduction after arthrodesis in 15 patients with septic arthritis. Four- teen of the 15 patients had active in- fection (positive cultures) at the time of arthrodesis. Ninety percent of pa- tients reported marked pain reduc- tion; however, no patient was pain- free. Functional outcome after shoul- der arthrodesis has lagged behind pain improvement. Of the 17 pa- tients in the Hawkins and Neer study, 12 5 functioned reasonably well at head level (eg, hair combing, face washing), while 4 others had great difficulty. Fourteen patients could use their hands satisfactorily at waist level, but only 3 could reach the hand behind the back for hy- giene. Overall, 7 of the 17 patients were dissatisfied because of func- tional disability. Wick et al 7 found that patients were unable to reach behind the back, and most of them had difficulty with ADLs around the face. In contrast, Cofield and Briggs 16 reported that, of the 65 pa- tients available for follow-up, 70% could lift moderate weights, dress themselves, tend to personal hy- giene, and eat using the extremity with the fused shoulder. Although only 21% could use their arm for light work at shoulder level, 82% found their arthrodesis to be func- tionally beneficial. Richards and colleagues 4,23 and Rouholamin et al 29 reported func- tional results similar to those of Cofield and Briggs. 16 All of their pa- tients (except those with distal ex- tremity paralysis or amputees) could easily reach the mouth. Although none of Richards’ patients could per- form overhead work, all were satis- fied with their functional improve- ment. Hawkins and Neer 12 reported that no patient was able to work overhead or with the arms abducted (eg, hammering, painting, climbing a ladder). Four of 17 patients could not return to jobs requiring manual la- bor. Five of the 17 patients returned to manual labor, but not at their pre- injury level. It is worth noting that Hawkins and Neer 12 as well as Cofield and Briggs 16 found extreme internal rotation to be detrimental for shoulder function. Complications Among the complications of shoul- der arthrodesis are nonunion, mal- positioning of the fused shoulder, perifusion fractures, infection, con- tinued pain, and soft-tissue irritation Figure 5 A 21-year-old man presented several years after resection of the right proximal humerus for osteogenic sarcoma. The proximal humerus was reconstructed with segmental allograft and a custom-made total shoulder prosthesis. A, Preoperative anteroposterior radiograph demonstrating dislocation of the shoulder arthroplasty as well as a recent periprosthetic fracture, necessitating shoulder arthrodesis. B, Postoperative radiograph demonstrating a long vascularized fibular graft compensating for a deficient humerus and arthrodesis of the acromion, glenoid, and fibula. C and D, Functional range of motion was achieved after fusion. Ori Safran, MD, and Joseph P. Iannotti, MD, PhD Volume 14, Number 3, March 2006 151 caused by prominent fixation devic- es. The frequency of these complica- tions is related to the quality and quantity of bone available for fusion as well as to the condition of the soft-tissue envelope. Nonunion In most series, the rate of non- union varies from 0 to 20%. 4,7,8,12,16,23,30 The relatively small number of patients in most studies, along with the complexity of the ini- tial diagnosis, makes it difficult to compare results between studies. With modern fixation methods, however, the nonunion rate is clos- er to 10%, with the exception of fu- sion after septic arthritis, which re- mains at 20%. 7 Most reported nonunions were symptomatic, painful, and required further treatment. Revision surgery with secondary bone grafting and/or refixation achieved union in most patients. Only a small percentage of patients demonstrated partial fusion of either the glenohumeral or the acromiohumeral articulation. In nearly all cases, the fusion mass was stable, the shoulder was functional, and no further treatment was neces- sary. Malposition of the Fused Shoulder It is difficult to define malposi- tion because no consensus exists as to optimal position. However, sever- al guidelines may assist in making this assessment. Painful winging of the scapula with the arm in the rest- ing position is a functional defini- tion of malposition associated with excessive abduction or flexion of the fusion. In addition, in the absence of other causes, inability to reach the face or anterior midline is most like- ly the result of rotational malposi- tion of the arthrodesis. Surgical cor- rection should be sought when the malposition causes scapular pain or leads to dysfunction in ADLs. A closing wedge corrective osteotomy distal to the fusion mass combined with plate fixation and grafting is recommended for managing a signif- icant and symptomatic malposi- tioned fusion. 10 Groh et al 19 reported that osteotomy markedly improved pain and function in all nine patients who underwent corrective osteoto- my for malpositioned shoulder fu- sion. Periarthrodesis Fractures Elimination of motion at the gle- nohumeral joint combined with lo- cal osteopenia increases the proba- bility that relatively minor trauma will cause fracture around the ar- throdesis. Typically, fracture occurs distal to the fusion mass either at the point at which the plate fixation ends or at the entry point of the most distal compression screws. Cofield and Briggs 16 reported on eight postoperative humeral frac- tures in their series of 71 patients, with most occurring in paralytic pa- tients. All fractures healed with ex- ternal immobilization. In their se- ries of 41 patients, Rybka et al 30 reported only one late postoperative humeral neck fracture, which oc- curred after a fall 1.5 years after sur- gery. No fractures were reported by Richards and colleagues 4,23 or by Stark et al. 18 Nondisplaced humeral fractures are managed with immobi- lization; displaced or unstable frac- tures require internal fixation for solid union to occur. 17 Prominence of Fixation Device The superficial location of inter- nal fixation devices, particularly compression screws and plates, may cause local skin problems and skin penetration. Typically, this occurs over the spine of the scapula and acromion because of a poor soft- tissue envelope secondary to muscle atrophy and multiple surgeries. In many cases, partial or complete re- moval of the fixation device after consolidation of the fusion mass is necessary to solve the problem. Cofield and Briggs 16 reported late screw removal in 17 of 71 shoulder operations (compression screws were used). Higher rates of soft-tissue prominence have been reported when DCP plates were used. Rich- ards et al 4 had to remove eight DCP plates from their 14 patients (57%). Stark et al 18 reported four symptom- atic plate removals in 15 patients. In the later study by Richards et al, 23 lower rates of metal-related symp- toms were reported when recon- struction plates were used. However, in a series using only reconstruction plates for fixation, local discomfort or prominence necessitated hard- ware removal in five of eight pa- tients. 8 Summary Shoulder arthrodesis is an end-stage salvage procedure for managing pain and loss of function when other joint-sparing reconstructive options have been exhausted. It is an alterna- tive to resection arthroplasty in the patient who desires the ability to lift to chest or shoulder level. A solid fu- sion achieves good pain relief and ac- ceptable ADL function in most pa- tients. Return to work is less consistently achieved, however. Both the surgeon and the patient should be aware of the limitations of shoulder arthrodesis and set realistic goals. With proper indications and surgical techniques, solid shoulder fusion and a satisfactory result are achieved in the vast majority of cas- es, with improvement in pain and functional outcome. References Evidence-based Medicine: Evidence- based reports consist of level III (case-contracted studies) or level IV (case series) only. Citation numbers printed in bold type indicate references published within the past 5 years. 1. González-Díaz R, Rodriguez- Merchán EC, Gilbert MS: The role of Arthrodesis of the Shoulder 152 Journal of the American Academy of Orthopaedic Surgeons shoulder fusion in the era of arthro- plasty. Int Or thop 1997;21:204-209. 2. Tung TH, Mackinnon SE: Brachial plexus injuries. Clin Plast Surg 2003; 30:269-287. 3. Rühmann O, Gossé F, Wirth CJ, Schmolke S: Reconstructive opera- tions for the paralyzed shoulder in brachial plexus palsy: Concept of treatment. Injury 1999;30:609-618. 4. Richards RR, Waddell JP, Hudson AR: Shoulder arthrodesis for the treat- ment of brachial plexus palsy. Clin Orthop 1985;198:250-258. 5. O’Connor MI, Sim FH, Chao EY: Limb salvage for neoplasms of the shoulder girdle: Intermediate recon- structive and functional results. J Bone Joint Surg Am 1996;78:1872- 1888. 6. Amin SN, Ebeid WA: Shoulder recon- struction after tumor resection by pedicled scapular crest graft. Clin Orthop 2002;397:133-142. 7. Wick M, Müller EJ, Ambacher T, He- bler U, Muhr G, Kutscha-Lissberg F: Arthrodesis of the shoulder after sep- tic arthritis: Long-term results. J Bone Joint Surg Br 2003;85:666-670. 8. Diaz JA, Cohen SB, Warren RF, Craig EV, AllenAA: Arthrodesis as a salvage procedure for recurrent instability of the shoulder. J Shoulder Elbow Surg 2003;12:237-241. 9. Arntz CT, Matsen FA III, Jackins S: Surgical management of complex ir- reparable rotator cuff deficiency. J Arthroplasty 1991;6:363-370. 10. Clare DJ, Wirth MA, Groh GI, Rock- wood CA Jr: Shoulder arthrodesis. J Bone Joint Surg Am 2001;83:593- 600. 11. Wilde AH, Brems JJ, Boumphrey FR: Arthrodesis of the shoulder: Current indications and operative technique. Orthop Clin North Am 1987;18:463- 472. 12. Hawkins RJ, Neer CS II: A functional analysis of shoulder fusions. Clin Orthop 1987;223:65-76. 13. Barr JS,Freiberg JA,Colonna PC,Pem- berton PA: A survey of end results on stabilization ofthe paralyticshoulder: Report of the Research Committee of the American Orthopaedic Associa- tion. J Bone Joint Surg 1942;24:699- 707. 14. Rowe CR: Re-evaluation of the posi- tion of the arm in arthrodesis of the shoulder in the adult. J Bone Joint Surg Am 1974;56:913-922. 15. Rowe CR: Arthrodesis of the shoulder used in treating painful conditions. Clin Orthop 1983;173:92-96. 16. Cofield RH, Briggs BT: Glenohumeral arthrodesis: Operative and long-term functional results. J Bone Joint Surg Am 1979;61:668-677. 17. Richards RR: Glenohumeral arthro- desis, in Iannotti JP, Williams GR (eds): Disorders of the Shoulder: Diag- nosis and Management. Philadelphia, PA: Lippincott Williams & Wilkins, 1999, pp 501-520. 18. Stark DM, Bennett JB, Tullos HS: Rig- id internal fixation for shoulder ar- throdesis. Orthopedics 1991;14:849- 855. 19. Groh GI, Williams GR, Jarman RN, Rockwood CA Jr: Treatment of com- plications of shoulder arthrodesis. J Bone Joint Surg Am 1997;79:881- 887. 20. Watson Jones R: Extra-articular ar- throdesis of the shoulder. J Bone Joint Surg 1933;15:862-871. 21. Naranja RJ Jr, Iannotti JP: Surgical op- tions in the treatment of arthritis of the shoulder: Alternatives to prosthetic ar- throplasty. Semin Arthroplasty 1995; 6:204-213. 22. Morgan CD, Casscells CD: Arthroscopic-assisted glenohumeral arthrodesis. Arthroscopy 1992;8:262- 266. 23. Richards RR, Sherman RM, Hudson AR, Waddell JP: Shoulder arthrodesis using a pelvic-reconstruction plate: A report of eleven cases. J Bone Joint Surg Am 1988;70:416-421. 24. Charnley J, Houston JK: Compression arthrodesis of the shoulder. J Bone Joint Surg Br 1964;46:614-620. 25. Beltran JE, Trilla JC, Barjau R: A sim- plified compression arthrodesis of the shoulder. J Bone Joint Surg Am 1975; 57:538-541. 26. Johnson CA, Healy WL, Brooker AF Jr, Krackow KA: External fixation shoul- der arthrodesis. Clin Orthop 1986; 211:219-223. 27. Kocialkowski A, Wallace WA: Shoul- der arthrodesis using an external fix- ator. J Bone Joint Surg Br 1991;73: 180-181. 28. Nagano A, Okinaga S, Ochiai N, Kurokawa T: Shoulder arthrodesis by external fixation. Clin Orthop 1989; 247:97-100. 29. Rouholamin E, Wootton JR, Jamieson AM: Arthrodesis of the shoulder fol- lowing brachial plexus injury. Injury 1991;22:271-274. 30. Rybka V, Raunio P, Vainio K: Arthro- desis of the shoulder in rheum- atoid arthritis: A review of forty-one cases. J Bone Joint Surg Br 1979;61: 155-158. Ori Safran, MD, and Joseph P. Iannotti, MD, PhD Volume 14, Number 3, March 2006 153 . with glenohumeral joint destruction dur- ing the era in which anti- tuberculosis antibiotics were not available. 20 Extra-articular fusions are rarely done as isolated proce- dures today. Intra-articular. with ADLs around the face. In contrast, Cofield and Briggs 16 reported that, of the 65 pa- tients available for follow-up, 70% could lift moderate weights, dress themselves, tend to personal hy- giene,. devic- es. The frequency of these complica- tions is related to the quality and quantity of bone available for fusion as well as to the condition of the soft-tissue envelope. Nonunion In most series,