During the past decade the number of shoulder arthroplasties performed around the world has increased greatly, with approximately 10,000 performed per year. Cofield's survivorship analysis of unconstrained total shoulder replacements projected 96% survivorship at 2 years, 92% at 5 years, and 88% at 10 years (1). Based on the parallel experience with hip and knee arthroplasty, with increased numbers of cases, one can expect increased numbers of complications that often require revision surgery (1–8).
Arthroplasty failures can be categorized based on soft tissue or bony deformities, component malpositioning, component loosening, and mechanical failure such as component dissociation and joint instabilities (Fig. 1). Functional improvement is directly related to the condition of the soft tissues, which in many cases are severely compromised. Scar tissue, musculotendinous contractures and deficiencies, nerve injury, and bone loss are challenging to the revision surgeon. Successful, stable outcomes of revision shoulder arthroplasty depend on recognizing and addressing these problematic issues and for this reason a variety of surgical techniques may be necessary to treat a specific underlying pathology.
It is the purpose of this report to outline the indications, preoperative assessment and technical considerations necessary to carry out a comprehensive total shoulder revision.
INDICATIONS AND CONTRAINDICATIONS
The primary goal of revision surgery is pain relief. Often severe bone and soft tissue deficiencies preclude surgical restoration of full range of motion and function. The individual surgeon's ability to recognize and appropriately address the degree of bone loss and soft tissue contracture determines the outcome. In some cases of severe soft tissue and bony deformity acceptable goals include relief of pain and limited restoration of function.
Patients with severe pain, restricted range of motion, and poor function after primary shoulder arthroplasty, especially in cases with initially good results that deteriorated over time, should be considered for revision. The procedure should only be considered after a thorough history, physical examination, and appropriate laboratory studies and imaging have been carried out to delineate the possible causes of failure. Because the causes are often multifactorial, meticulous preoperative evaluation is paramount for success of revision arthroplasty.
Revision shoulder arthroplasty is contraindicated in patients whose medical status puts them at risk in the prospect of such complex surgery. Profound neurologic disorders, psychological impairment, and inability to undergo extensive rehabilitation program because of physical or emotional issues are also contraindications.
A careful history is critical in establishing causes for failure after shoulder arthroplasty. The primary diagnosis and any surgery that preceded the index arthroplasty are very important in establishing a cause of implant failure. For example, failure after total shoulder arthroplasty may be the result of component loosening, instability, rotator cuff insufficiency, or mechanical failure(1,2,4,5,9–18). However, failures in hemiarthroplasty for fractures have been most commonly reported to be related to tuberosity fixation failure, component malposition, or secondary glenoid arthrosis (16–25) (Figs. 2 and 3).
Postoperative rehabilitation programs and any new trauma must also be documented. A history of fever, chills, erythema, or warmth at the incision are indicators of possible infection. Infection should be considered and ruled out in all cases of failed arthroplasty. In suspicious cases aspiration should be carried out before revision surgery. Night pain and weakness may indicate soft tissue or rotator cuff failure and a history of delayed onset of pain and loss of motion with crepitus or clicking may be highly suggestive of an acute glenoid loosening. A complete history and physical examination evaluating the patient's generalized well-being must be considered before any further surgical intervention.
A thorough examination of the shoulder, upper extremity, cervical spine, and ipsilateral thorax is mandatory. Referral patterns for cervical spine and even pulmonary findings must be ruled out as a cause of pain. Range of motion, strength, and stability are carefully evaluated and soft tissue contractures and scars are noted.
The stability of the joint must be carefully assessed. This involves anterior, posterior, and inferior translation of the humerus on the glenoid and a careful assessment of any gross instability. Clicking, whether painful or not, during translation of the glenohumeral joint may indicate overt loosening of the glenoid component. Rotator cuff strength and impingement signs are assessed to determine rotator cuff function. Biceps tests are also performed to document any pathology associated with the long head of the biceps tendon (26).
Finally, a careful neurovascular evaluation is mandatory to document any loss in either motor or sensory function or vascular injury before any further surgical intervention. If necessary, appropriate neurologic studies should be considered.
Evaluation of serial plain radiographs, true anteroposterior, axillary, and outlet views can be helpful in assessing component malposition, instability, and component loosening. Plain radiographs may also be used to evaluate bone loss and to assess the subacromial space for outlet impingement secondary to bone spurs of the acromion or acromioclavicular joint.
With the use of serial plain radiographs the diagnosis of prosthetic loosening is confirmed when a shift in radiographic position is identified. A radiolucent line of 2 mm or more that completely surrounds the implant is also suspicious for component loosening. However, a number of authors have shown that radiolucency does not always correlate with clinical loosening (1,6,16,20,27).
Axillary views are helpful in documenting glenoid bone loss, deformity, and excessive component version. It can also be helpful in the diagnosis of glenohumeral instability. Computed tomography scans with implant subtraction can be a further help when an axillary view is ambiguous.
Careful evaluation of component position, bone quality, cement mantel, and concomitant ipsilateral elbow replacement are also critically important in the preoperative planning.
Diagnostic arthroscopy can be helpful in identifying and treating patients with rotator cuff insufficiency, outlet impingement, acromioclavicular joint problems, loose bodies, and even gross glenoid component loosening (28,29). The arthroscope should be directed away from the humeral head component to avoid the “mirror effect.” A probe through an anterior portal is used to test for gross glenoid stability (Fig. 4). Biceps lesions can be treated with arthroscopic release or tenodesis (26). Finally, arthroscopy can assist rotator cuff repair performed with a mini-open technique.
Because most patients who undergo a revision procedure are older, a complete work-up should be considered to rule out underlying metabolic bone disease. Erythrocyte sedimentation rate, complete white blood cell count with differential, and C-reactive protein should be evaluated to help rule out sepsis. If sepsis is suspected, a diagnostic arthrogram and aspiration may confirm infection and obtain the offending organism. An arthrogram may also confirm aseptic loosening if dye leaks around the implant at the bone cement interface; however, arthrographic findings may be negative in the face of loosening.
Radionuclear scanning, including three-phase bone scans and indium-labeled leukocyte scans, may also indicate indolent sepsis. If nerve injury is suspected, a complete electromyographic study, including nerve conduction studies, is mandatory to document existing neurologic lesions before attempting any revision surgery.
Great care should be taken to ensure proper patient positioning at the time of revision surgery. The patient should be placed in the semi-Fowler's position with the feet elevated. The patient should be placed laterally on the table with proper restraint to prevent dangerous movement during traction of the arm. Lateral positioning should allow room for the arm to be extended fully off the table toward the floor.
At exposure it is critical to utilize old incisions if possible. Subcutaneous contractures should be released at this time. An extended anterior incision from above the clavicle over the coracoid process and distally following the anterior portion of the deltoid is carried out. In some cases in which there is a gross posterior displacement of the glenoid component, a posterior approach should be considered.
Great care should be used in handling the deltoid muscle. Preservation of deltoid function is critical for successful outcome. Preservation of the deltoid origin when possible is the best choice. In cases when the deltoid is detached for repair of a large rotator cuff tear or for other reasons, a stable reattachment through bone is mandatory.
Once the deltopectoral interval has been exposed, a complete release and excision of the subdeltoid and subacromial adhesions should be carried out. In cases in which the rotator cuff is intact, or if a repairable tear is present, release of the coracoacromial ligament or acromioplasty, or both, can be considered for better exposure. For patients with overt rotator cuff insufficiency or irreparable rotator cuff disease, it is best to leave the coracoacromial ligament to prevent further anterosuperior instability secondary to the loss of the coracoacromial arch (4,8,28,30,31). At this time care should be used in re-establishing the plane deep to the conjoint tendon to expose the subscapularis tendon.
After all the subdeltoid adhesions have been removed, the shoulder should again be examined for range of motion with the patient under anesthesia to assess external rotation. If there is less than 30° of external rotation, one can consider one of many options for lengthening the subscapularis tendon and carrying out any appropriate capsular releases. This must be planned before closure so that enough tendon length remains to carry out appropriate soft tissue balance and repair.
Soft Tissue Contractures and Deformities
Recreation of the optimal soft tissue balance is critical for restoration of function after a primary or revision shoulder arthroplasty. Soft tissue balance depends on proper release of the contractures, appropriate lengthening of tendons, and repair or reconstruction of the rotator cuff when possible. At the same time the origin and function of the deltoid must be maintained.
Preoperative assessment and examination under anesthesia give the surgeon important information about the quality and function of the soft tissues. In cases of severe contracture the releases begin with lysis and excision of the subacromial, subdeltoid, and subcoracoid adhesions. Reestablishing the space below the conjoint tendon is critically important for subscapularis lengthening and exposure. Often an incision along the lateral border of the conjoint tendon will facilitate clearing a space to visualize the subscapularis tendon. Great care must be taken to avoid injury to the musculocutaneous nerve deep to the conjoint tendon and the axillary nerve inferior medially to the subscapularis tendon. In many cases exploration and retraction of the axillary nerve facilitate further exposure.
At this time, if external rotation is still less than 30°, a subscapularis tendon lengthening should be considered. The subscapularis tendon is often quite attenuated. It should therefore be skeletonized off the lesser tuberosity to gain length and later reattached to the medial surface of the proximal humeral osteotomy site through drill holes. In this way up to 2.5 cm of length can be obtained, effectively increasing rotation by 25° to 30° or more (Fig. 5).
Once the subscapularis has been released, the tendon can be lengthened further by a 360° release from the labrum and lysis of adhesions in the rotator interval, subcoracoid space, and at the inferior surface of the subscapularis tendon (13) (Fig. 6). When additional length is necessary, the subscapularis may be elevated from the glenoid fossa medially, creating a muscle slide.
Capsular contractures must be released both for soft tissue balance and for proper exposure of the glenoid. Anteriorly the capsulolabral contractures and postarthroplasty fibrous scars are released from the glenoid rim and excised from below the insertion of the long head of the biceps if it remains to the 6 o'clock position. Care must be taken inferiorly to prevent injury to the axillary nerve, which may be caught in the scar. In these cases it is best to expose the nerve before release. The posterior capsule can be similarly released and the labrum excised from the 12 o'clock position to approximately the 8 o'clock position (right shoulder). The inferior capsule is best released off the shaft of the humerus subperiosteally to avoid any injury to the axillary nerve, which in some cases may be very difficult to find. The inferior capsule may also be released directly by exposing the capsule below the subscapularis tendon and incising it anteriorly and inferiorly, continuing posteriorly to release the posterior inferior glenohumeral ligaments or band. All this is done while retracting the exposed axillary nerve. This will allow postoperative restoration of arm elevation. In some cases motion loss may be due to previous use of too large a humeral head component. In these cases down-sizing the humeral component will help increase range of motion.
When possible, restoration of rotator cuff function and integrity is critical for optimal results. Repairable rotator cuff tears should be evaluated, mobilized, and reattached securely to the greater tuberosity using appropriate fixation techniques. In some cases of severe attenuation of the tendon or cases of irreparable rotator cuff tears, acromioplasty should be avoided. Preservation of the coracoacromial arch is critical to prevent subsequent anterior and superior migration of the humeral head. It is best to avoid glenoid replacement in most of these patients in order to avoid the possibility of the eccentric glenoid loading (inferior to superior) and subsequent glenoid component loosening (9,28,30,32). In cases in which there is good anterior and posterior stability total shoulder arthroplasty can still be considered as long as the humeral head remains stable and well centered.
Glenoid Component Loosening and Revision
Bone defects due to primary eccentric wear or secondary to bone loss after glenoid loosening are challenging. Reasonable bone stock is mandatory for proper glenoid replacement. Bone defects can be central, peripheral, or a complex combination of both. After a hemiarthroplasty, eccentric wear of the glenoid requires neutralizing of the version of the glenoid by a number of different methods. Concentric reaming, excision of the high anterior or posterior side, and in cases of severe bone loss, a bone graft can be considered (Fig. 7) (11,16,33–35).
Bell et al. (33) recently reported on a novel technique for dealing with glenoid bone loss: the anterior or posterior glenoid is bone grafted using a levering out of the scapular cortex and a placement of a wedge-shaped graft with a screw fixation (Fig. 8). Other authors have discussed the use of bone grafts in cases of greater than 15° of glenoid retroversion (Fig. 9)(11,28,35). In either case, the glenoid vault must be able to accept the peg fins from the glenoid component. In cases of severe medial bone loss, glenoid fixation may be impossible and hemiarthroplasty should be carried out. Similarly, when the complex bone loss is too severe, hemiarthroplasty may be a more prudent choice.
Severe central bone loss from the glenoid may be treated with a staged procedure. The central defect may be bone grafted and a hemiarthroplasty may be carried out in the hope that the bone will be incorporated and allow for later glenoid reconstruction (Fig. 10). In our experience, most patients have enough pain relief from the hemiarthroplasty that they usually do not come back for glenoid component revision (16,31).
Abnormal glenoid component retroversion may lead to posterior instability (Fig. 1), necessitating revision. The glenoid component must be removed and the bony deformity and glenoid bone stock must be corrected as described previously. In most cases of glenoid malposition there is also an element of soft tissue contracture that must be corrected in order to properly reconstruct the unstable shoulder arthroplasty.
In cases of glenoid revisions, cemented components that are placed in an anatomic position remain the best option. Porous-coated glenoid components are available and may be considered if bone stock allows. The glenoid should be seated flush to the concentric surface of the glenoid. Polymethylmethacrylate cement should never be used to build up a peripheral or central defect, because cement breakage and subsequent loosening are possible.
Humeral Component Revision
Humeral component loosening is rarely a cause of arthroplasty failure; component malposition is much more common (Fig. 11). Component malposition is frequently the cause of tuberosity failures after hemiarthroplasty for trauma (19,21,23,24). A prosthesis positioned too proud can cause impingement. Humeral components placed too low lead to loss of the myofascial tension and deltoid and rotator cuff failure as well as secondary impingement from the greater tuberosity. Extreme varus malposition creates a loss of humeral offset and tuberosity overhang, leading to further rotator cuff insufficiency and failure. Less commonly, abnormal humeral version may also contribute to instability.
Proximal humeral bone loss, after acute or chronic traumatic injuries (22,23), often leads to placement of the humeral component at an improper height, which leads to instability. Placement of the humeral component too low is the most common cause of inferior instability. When it is cemented too high, increased tension in the myofascial sleeve of the deltoid or rotator cuff leads to a painful impingement or superior instability. These situations usually require revision of the humeral component and repositioning to a more anatomic length. Preoperative scanograms are often necessary to plan and assess the anatomy for such revisions.
Poorly positioned one-piece humeral components require removal at the time of revision. In cases of malpositioned or malsized second-and third-generation modular components, the humeral stem component may remain and a standard or custom revision modular head component may be used if available to correct version or height (36) (Fig. 9). In cases of a well-cemented or biologically fixed humeral stem, component removal can be exceedingly difficult and in some cases dangerous. The very thin posterior humeral cortex is extremely susceptible to fracture and perforation during removal.
Humeral stem revision requires appropriate instrumentation including high-speed, low-torque drills, appropriate cement-removing chisels, and ultrasonic instruments for removal of cement. Even with these instruments meticulous exposure and very careful techniques are required to avoid perforation or fracture of the humeral shaft. Adequate exposure and freedom of the humerus are mandatory and require a complete elevation of the fibrous soft tissue circumferentially from the osteotomy site and proximal humerus. This dissection is carried anteriorly to approximately half way up the osteotomy and proximal humerus site posteriorly (Fig. 12). Once free, the proximal humerus should be delivered into the field by extending, externally rotating, and adducting the arm. Even with this exposure, in some cases a redesigned trough, slot, or hinged cortical window may be necessary to remove a humeral component without sustaining severe bone trauma (Fig. 13). In cases of nonmodular humeral components, flexible osteotomes or chisels are necessary to break through cement or bony in-growth. At this point instrumentation that allows for disimpaction of the humeral component is used carefully. If the surgeon suspects that the fixation is too great, the slot or cortical window described above should be considered.
Once the component has been removed, the revision stem should be placed in the proper height and version using appropriate fixation techniques. When a biologic in-growth component was previously used, it may be considered again if there is adequate metaphyseal bone stock. In other cases the appropriate cement technique should be carried out.
When there has been a significant proximal humeral bone loss, primary or custom humeral prosthetic components can be used in conjunction with autogenous bone grafts or allografts to create a stable humeral construct.
Techniques for Revision Surgery of Glenohumeral Instability
Cofield (1,27) and Wirth(18,37) have described instability as the most common complication requiring revision after shoulder arthroplasty (14). Instability after arthroplasty requires a careful and thorough assessment of the direction and probable causes prior to any revision surgery. Soft tissue insufficiency, contracture, component malsizing, and component malposition are the most common causes for such instability. Usually a combination of these problems causes the overt problem. Nerve injury and loose bodies such as a grossly loose glenoid component are less common causes of instability after shoulder arthroplasty (Fig. 5). The surgical repair or revision requires removal of a malpositioned component (as described in the previous sections) with proper repositioning and fixation, as well as appropriate soft tissue repair and balancing. In some cases soft tissue reconstruction using allografts may be necessary.
Anterior and posterior unidirectional instability is usually the result of soft tissue contracture in the opposite direction of the instability or soft tissue failure in the direction of instability (i.e., subscapularis ruptures in acute anterior instability).
Among the most devastating instability patterns after shoulder arthroplasty is a fixed anterosuperior instability through the coracoacromial arch in cases in which arthroplasty has been carried out in the face of massive rotator cuff insufficiency and a previous acromioplasty or coracoacromial ligament resection. These patients can become severely disabled by pain and lack of function and this instability pattern is extremely difficult to treat. Recently, we utilized an Achilles tendon allograft with fixation of the calcaneal bone portion to the coracoid process using a 4.5-mm cancellous screw and washer. The thickened Achilles tendon is then fixed to the full undersurface of the acromion, which has been decorticated using a series of horizontal mattress sutures through bone (Fig. 14). In some cases this procedure was carried out with custom modular humeral head components that produced more retroversion of the humeral head, thereby allowing more stability. In three of four cases fair limited-goal outcomes were achieved.
Anterior instability is usually the result of a soft tissue failure or closure of the subscapularis tendon. This can be caused by traumatic injury or poor repair and in some cases by the use of too large a humeral component. If the glenoid component malposition is not identified on plain radiographs, an arthrogram or magnetic resonance imaging may be necessary to identify a subscapularis rupture. If acute, the subscapularis may be repairable primarily using sutures through bone holes or suture anchors. In some cases, releases that have been described above may be necessary to lengthen the muscle tendon unit effectively for proper fixation. In cases of chronic subscapularis insufficiency, we have described the use of an Achilles tendon allograft to reconstruct the anterior soft tissue insufficiency (14). An Achilles tendon allograft fixed to the glenoid surface by way of a corticocancellous screw and then fixed to the humerus by way of screws or suture anchors has been described previously (14). This technique creates an anterior checkrein and tenodesis effect to prevent anterior translation (Fig. 15).
Posterior instability is most commonly associated with increased humeral head retroversion, posterior glenoid erosion (hemiarthroplasty), or excessive glenoid component retroversion. Posterior capsular distension is a common finding in most cases of posterior instability. In most cases revision of the glenoid component, posterior capsular imbrication, and a decreased retroversion of the humeral head individually or together will usually satisfactorily correct this problem. Component malposition is addressed as described in the above section on component revisions. The goal should be combined retroversion of the humeral head and the glenoid component of between 30° and 40°. Posterior capsule plication can be carried out after the component revision has been done and before the real humeral head component has been placed. A longitudinal incision is made in the capsule and a vertical row of sutures is placed to tighten the capsule. The humeral head component is then impacted into place and the sutures plicating the capsule are tightened in such a way as to avoid excessive tightening of the posterior capsule, yet allowing better soft tissue balance.
The multifactorial nature of total shoulder revision is manifested in the postoperative patient care. Appropriate immobilization or bracing is necessary to protect any soft tissue reconstructions or bone graft techniques that have been carried out.
The postoperative rehabilitation program begins at the time of surgery. The surgeon must establish safe range of motion criteria during surgery that can be allowed in the early rehabilitation period. These safe zones of motion can then be used in an early passive range of motion program directed by the physician and the physical therapist. If soft tissue reconstruction has been carried out, active assistive and active motion is delayed in the postoperative period until these tissues have healed appropriately. In general, early passive range of motion in a protected zone is allowed and encouraged. Once soft tissue or bone healing has been achieved, active motion can be instituted. Stretching and other strengthening exercises are then begun at the appropriate time.
The postoperative rehabilitation program is extremely important for successful revision arthroplasty. Patient dedication to this surgeon-directed program is critical. If a patient is unwilling to undergo this intense postoperative therapy, serious consideration should be given as to whether revision surgery is appropriate.
Although revision rates have been reported up to 13% in unconstrained total shoulder arthroplasties and Cofield has reported an expected failure in survivorship of up to 12% at 10 years, there is little in the written literature on the results of revision arthroplasty (1,4,6,7,38). Most studies appear as annotations or abstracts of meetings and are general results of a number of revisions that have been carried out for multiple causes of arthroplasty failure. Comparison of these revisions is difficult, and owing to the retrospective nature, somewhat misleading.
In a compilation of six series including 150 shoulders Peterson noted that 60% of the patients had satisfactory results; however, 28% had unsatisfactory outcomes (6,7). Our own series included in this compilation evaluated 13 patients undergoing revision from multiple diagnoses (11); 8 patients had satisfactory outcomes and 5 patients had unsatisfactory outcomes after revision surgery.
In general, patients undergoing revision for aseptic glenoid loosening or humeral arthrosis after primary hemiarthroplasty had the best results. Patients requiring revision for soft tissue deficiencies, instabilities, and severe bone loss had the poorest outcomes. In cases of hemiarthroplasty for trauma, those patients with tuberosity failure had the poorest results.
Rodowsky et al. (16) evaluated glenoid component failure in 25 patients and noted satisfactory results in two thirds and unsatisfactory results in the other one third. There was no statistical difference in satisfactory results in patients with glenoid revision or glenoid resection. More recently, Connor et al. (5) evaluated 57 patients with multiple causes of shoulder arthroplasty failure requiring revisions. Forty-four patients had satisfactory results. Most failures were due to instability or bone tuberosity failure after hemiarthroplasty for trauma. As in other series, their patients with aseptic loosening fared the best. Of note in their study, removal of well fixed humeral components led to complications or poor results in six of eight patients. Sonnaband and Mohammed (8) recently reported on 30 patients undergoing component revision for failed shoulder arthroplasty. Their study concluded a trend for satisfactory pain relief but modest functional gains in all cases of component malsizing or malposition. Finally, Jensen and Rockwood (38) reported on glenoid excision in cases of aseptic loosening in 10 cases with significant pain relief and increase of shoulder motion and a dramatic improvement in the ability to perform activities of daily living. As the number of revision surgeries increases, more reports of long-term results and survivorship will give us a better understanding of the true effectiveness of revision procedures in shoulder arthroplasty.
Revision total shoulder arthroplasty remains one of the most difficult reconstructive procedures about the shoulder even in the hands of an experienced shoulder surgeon. The multifactorial etiologies, which make revision necessary, are often difficult to document completely in a preoperative evaluation. For this reason the surgeon must be familiar with multiple surgical maneuvers and techniques that are necessary to address the pathologic anatomy and changes that are encountered at the time of surgery. This demands a careful preoperative assessment including evaluation of the soft tissue, bone stock, and component position. Sepsis must be ruled out before any revision. Successful revision surgery depends on proper exposure, release, and soft tissue reconstruction, and appropriate component placement and fixation. Appropriate equipment for exposure and component removal is mandatory. The surgeon must have all the necessary primary and custom implants that may be necessary during the procedure.
The surgical procedure requires meticulous attention to surgical detail with a special emphasis on preserving deltoid muscle function. A careful individually defined rehabilitation program is also mandatory for success. With these guidelines successful outcomes in revision shoulder arthroplasty have been achieved in more than two thirds of cases.
1. Cofield RH. Revision procedures in shoulder arthroplasty. In: Morrey BF, An KN, Cabanela ME, et al., eds. Reconstructive surgery of the joints
. 2nd edn. New York: Churchill Livingstone, 1996:789.
2. Boyd AD, Thomas WH, Sledge CB, et al. Failed shoulder arthroplasty. Orthop Trans 1990; 14:255.
3. Brenner BC, Ferlic DC, Clayton ML, et al. Survivorship of unconstrained total shoulder arthroplasty. J Bone Joint Surg [Am] 1989; 71:1289–96.
4. Cofield RH, Edgerton BC. Total shoulder arthroplasty complications and revision surgery. Instruct Course Lect 1990; 39:449–62.
5. Connor PM, Levine WN, Arroyo JS, et al. The surgical management of failed shoulder arthroplasty [abstract]. In: Programs and Abstracts of the 13th
Open Meeting of the American Shoulder and Elbow Society, San Francisco, CA, 1997, p27.
6. Petersen SA, Hawkins RJ. Revision of failed total shoulder arthroplasty. Orthop Clin North Am 1998; 29:7.
7. Petersen SA, Hawkins RJ. Revision shoulder arthroplasty. In: Friedman RJ, ed. Arthroplasty of the shoulder. New York: Thieme, 1994:234–41.
8. Sonnaband D, Mohammed KD. Revision shoulder arthroplasty. J Shoulder Elbow Surg 1999; 8:553.
9. Bonutti PM, Hawkins RJ. Component loosening in unconstrained shoulder arthroplasty. Semin Arthroplast 1990; 1:124.
10. Bonutti PM, Hawkins RJ. Revision hemiarthroplasty of shoulder arthroplasty. Orthop Trans 1991; 14:598.
11. Caldwell GL, Dines D, Warren RF, et al. Revision shoulder arthroplasty. Orthop Trans 1993; 17:140.
12. Figgie HE, Inglis AE, Goldberg VM, et al. An analysis of factors affecting the long-term results of total shoulder arthroplasty in inflammatory arthritis. J Arthroplast 1988; 3:123–30.
13. Matsen III, FA Rockwood Jr, Ca Worth MA, Lippitt SA. Glenohumeral arthritis and its management. In: Rockwood CA, Matsen III, FA eds. The shoulder
. 2nd edn. Philadelphia: WB Saunders, 1998:840–942.
14. Moeckel BH, Altchek DW, Warren RF, et al. Instability of the shoulder after arthroplasty. J Bone Joint Surg [Am] 1993; 75:492–7.
15. Neer II, CS Kirby RM. Revision of the humeral head and total shoulder arthroplasties. Clin Orthop 1982; 170:189–95.
16. Rodosky MN, Weinstein DM, Pollock RG, et al. On the rarity of glenoid component failure. J Shoulder Elbow Surg
17. Sperling JW, Cofield RH. Revision total shoulder arthroplasty for glenoarthritis. Orthop Trans 1996; 20:924.
18. Wirth MA, Seltzer DG, Senes HR, et al. An analysis of failed humeral head and total shoulder arthroplasty. J Elbow Shoulder Surg
19. Bigliani LU, Flatow EL, McCluskey GM, et al. Failed prosthetic replacement for displaced proximal humerus fractures. Orthop Trans 1991; 15:747.
20. Bonutti PM, Hawkins RJ. Fracture of the humeral shaft associated with total replacement arthroplasty of the shoulder. A case report. J Bone Joint Surg [Am] 1992; 74:617–8.
21. Boileau P, Walch G, Krishnan SG, et al. Tuberosity osteosynthesis and hemiarthroplasty for 4-part fractures of the proximal humerus. Techn Shoulder Elbow Surg 2000; 1 (2):96–109.
22. Coleman S, Dines DM, Warren RF. Modular shoulder hemiarthroplasty for acute and chronic complex fractures of the proximal humerus. Orthopaedic Special Edition 2000; 6:39–42.
23. Dines DM, Warren RF. Modular shoulder hemiarthroplasty for acute fractures: surgical considerations. Clin Orthop Relat Res 1994; 307:18–26.
24. Dines DM, Warren RF, Altchek DW, et al. Posttraumatic changes of the proximal humerus: malunion, non-union, and osteonecrosis: treatment with modular hemiarthroplasty or total shoulder arthroplasty. J Shoulder Elbow Surg 1993; 2:11–21.
25. Norris TR, McElheny E. The role of long-stem humeral head prosthesis in the treatment of complex humeral fractures and in revision arthroplasty. Semin Arthroplast
26. Dines DM. Lesions of the long head of the biceps after shoulder arthroplasty [abstract]. In:Abstracts of American Shoulder and Elbow Society Meeting (Closed)
Philadelphia, PA, November 14, 1999.
27. Kelleher IM, Cofield RH, Becker DA, et al. Fluoroscopically positioned radiographs of total shoulder arthroplasty. J Shoulder Elbow Surg 1993; 1:307.
28. Bonutti PM, Hawkins RJ, Sademmi S. Arthroscopic assessment of glenoid component loosening after total shoulder arthroplasty. Arthroscopy 1993; 9:272–5.
29. Hersch JG, Dines DM. Arthroscopy for failed shoulder arthroplasty. Arthroscopy 2000; 16:606–612.
30. Franklin JL, Barrett WP, Jacksins SE, et al. Glenoid loosening in total shoulder arthroplasty. J Arthroplasty 1988; 3:39–46.
31. Moeckel BH, Warren RF, Dines DM, et al. The unstable shoulder arthroplasty. In: Freedman RJ, ed. Arthroplasty of the shoulder. New York: Thieme, 1994:254–63.
32. Arntz CT, Matsen III, FA Jackins S. Surgical management of complex irreparable rotator cuff deficiency. J Arthroplast 1991; 6:363–6.
33. Bell RH, Noble JS. The management of significant glenoid deficiency in total shoulder arthroplasty. J Shoulder Elbow Surg 2000; 9:248–56.
34. Hill JM, Norris TR. Long term results of bone grafting for glenoid deficiency in total shoulder arthroplasty. J Shoulder Elbow Surg 1997; 6:176–81.
35. Neer II, CS Morrison DS. Glenoid bone grafting in total shoulder arthroplasty. J Bone Joint Surg [Am] 1988; 70:1154–9.
36. Schaffer BS, Giordano CP, Zuckerman JD. Revision of loose glenoid component facilitated by a modular humeral component: a technical note. J Arthroplast
37. Wirth MA, Rockwood Jr. CA Glenohumeral instability following shoulder arthroplasty. Orthop Trans 1995; 19:459.
38. Jensen K, Rockwood CA Jr. Salvage shoulder surgery following glenoid component failure [abstract]. J Elbow Shoulder Surg