15.1 Introduction

An expected consequence of shoulder arthroplasty surgery is that complications may arise requiring revision surgery. Revision surgery for a failed shoulder arthroplasty can be unforgiving. The introduction of the reverse shoulder prosthesis (RSP) has improved the ability to manage much of the complex pathology seen in the setting of failed shoulder arthroplasty with reproducible success. By not depending on the rotator cuff for stability and function, RSP has been shown to be able to manage complex issues including rotator cuff deficiency, prosthetic instability, glenoid humerus bone loss, and varying degrees of proximal humeral bone loss. Over the course of this chapter, we will outline the clinical scenarios in which we employ an RSP in a revision setting and detail our approach to dealing with the various challenging problems encountered.

15.2 Approach to Revision Shoulder Arthroplasty

The decision to use an RSP in the setting of a revision arthroplasty is based primarily on how to best manage the specific pathology presented by each case and ultimately provide the patient with a satisfactorily functional shoulder. Shoulder arthroplasty failures that require revision to reverse shoulder arthroplasty (RSA) are associated with several common pathological conditions that must be addressed.

15.2.1 Rotator Cuff Insufficiency

One of the most common reasons for shoulder arthroplasty failure includes failed subscapularis repairs from the index primary arthroplasty, as well as subsequent tears of the rotator cuff. Subscapularis failures can relate to overstuffing of the total shoulder arthroplasty (► Fig. 15.1), poor tendon quality following multiple previous open surgeries, aggressive postoperative rehabilitation, or failure to adhere to postoperative restrictions. Patients with subscapularis insufficiency typically have excessive passive external rotation, a loss of internal rotation strength, and/or a positive belly-press or lift-off tests. Radiographs often reveal anterior (► Fig. 15.2) or anterosuperior subluxation (► Fig. 15.3). Functional results, however, may include a spectrum of functional abilities from dynamic pseudoparesis to full elevation. Only those patients with poor functional results and/or excessive pain relate a desire for a revision surgery.

The incidence of late rotator cuff dysfunction has been estimated to occur at rates as high as 45% at 10 years.1 A deficient rotator cuff is associated with a loss of strength and function, as well as radiographic subluxation. Rotator cuff deficiency has also been shown to be associated with glenoid component loosening after TSA related to the rocking-horse effect of rotator cuff dysfunction.2 Advanced imaging with metal suppression can often identify post-arthroplasty rotator cuff tears (► Fig. 15.4).

15.2.2 Glenoid Bone Loss

Glenoid bone loss is commonly present when a revision to an RSP is considered. Patients with failed total shoulders commonly have aseptic glenoid loosening with varying degrees of glenoid bone loss related to glenoid component micromotion3 (► Fig. 15.5). Patients with failed hemiarthroplasty for cuff tear arthropathy and glenohumeral arthritis typically have excessive glenoid wear patterns4 that must be identified. The pattern of bone loss must be recognized, as it will guide the proper placement of the glenosphere baseplate during revision to RSP, and will help define the need for glenoid bone grafting. Computed tomographic (CT) scans have become critical in recognizing glenoid wear patterns and areas of glenoid deficiencies (► Fig. 15.3).

15.2.4 Instability

Instability following shoulder arthroplasty has been reported as the most common reason for early revision.5 Instability following anatomic TSA and hemiarthroplasty typically relates to soft-tissue deficiencies, most commonly related to the subscapularis (► Fig. 15.2). However, component malposition and preoperative glenoid morphology may also play a significant role (► Fig. 15.6). In the setting of such instability, revision to another anatomic shoulder arthroplasty has been associated with high failure rates; thus, conversion to RSA has become popular, given the favorable short-term results.6 In evaluating the unstable hemiarthroplasty or TSA, one must focus attention on rotator cuff deficiency and preoperative glenoid morphology (i.e., excessive retroversion or anteversion), and intraoperatively assess glenoid and humeral component position. The height and version of the humeral stem will greatly influence the need to remove and revise the humeral stem, even in cases where convertibility is possible.

15.3 Failed Hemiarthroplasty or Total Shoulder Arthroplasty

Revision of a failed hemiarthroplasty or anatomic TSA is typically performed in the setting of pain and/or unacceptable decreased function caused by a torn rotator cuff, instability, implant loosening, or progressive glenoid wear. The presence of a deficient rotator cuff and instability are assessed preoperatively by means of a physical exam looking for pseudoparalysis, gross rotator cuff weakness, and dynamic or static subluxation. Radiographs demonstrating superior humeral head migration, eccentric wear pattern, or joint subluxation support this. Similarly, glenoid wear and bone loss can be assessed using standard shoulder radiographs and CT scans with both two- and three-dimensional reconstructions. Radiographs are evaluated for the types of implants in place, how well the humeral stem is fixed, the quality of bone, and the position and version of the components. Having previous operative reports facilitates this assessment. If a humeral stem is in good alignment and well fixed, thought is given to retaining the stem and converting it to a reverse shoulder construct if the system allows, often hybridizing it with our choice of glenosphere where possible (► Fig. 15.7). This avoids the risk for significant bone loss or fracture involved with removing the stem.

Intraoperatively, a standard deltopectoral approach is utilized. A vital part of any revision surgery is obtaining adequate exposure. To facilitate this, care is taken to make appropriate releases of the sub-deltoid, sub-acromial, and sub-coracoid spaces, which tend to have a generous amount of scar. Likewise, the pectoralis major tends to be adhered to the conjoined tendon and needs to be released especially in cases where stiffness is an issue. Owing to the previous surgery and scar formation, time is taken to ensure appropriate identification of the axillary nerve when the sub-coracoid space is being developed. While the nerve is not necessarily dissected out, the sub-coracoid space is developed slowly and the nerve identified by palpation and its continuity confirmed using the tug test.13 Once the subcoracoid space has been developed and the axillary nerve identified, the subscapularis can be released. When present, the subscapularis tendon is peeled off of the lesser tuberosity and tagged for later repair. The capsule is released subperiosteally off the humerus from anterior to posterior staying on bone to protect the axillary nerve.

Once exposed, the humeral head can be dislocated and removed. In cases where the humeral stem can be converted to an RSP, it is paramount to assess the height of the stem and the version in which the stem was implanted. One must resist the temptation to utilize the platform stem and implant an RSP in excessive tension. If the humeral stem was implanted properly and proper soft-tissue tension can be established, then utilization of the humeral stem for the RSP can be performed successfully.14,15,16

When removal of the humeral stem is deemed necessary soft tissue, bone, and/or cement surrounding the exposed portion of the humeral stem is removed. A pencil-tip, fluted, high-speed burr is used to circumferentially break up the cement mantle or bone ingrowth proximally. Flexible osteotomes can be used to disrupt the more distal points of fixation. Knowledge of the textured or porous surface of the stem is necessary to identify how far down the stem the fixation must be disrupted. All bone or cement must be removed from around any fins to prevent humeral fractures during extraction. Once the stem is sufficiently freed up, a carbide tip punch can be used to disimpact the stem by placing the punch just under the neck of the stem.

With the stem removed, the focus is directed to the glenoid. The proximal humerus is gently retracted out of the way, posteriorly, to expose the glenoid. The subscapularis is mobilized by releasing along its superior border along the rotator cuff interval, releasing adhesions above the axillary nerve within the sub-coracoid space, and freeing up adhesions along the anterior glenoid wall. Loose glenoids can easily be removed. However, well-fixed glenoids require a careful removal, which can be implant specific. For a cemented pegged or keeled glenoid, a small osteotome inserted parallel to the glenoid surface disrupts the cement-bone interface. If sufficiently freed by doing this, the implant can be removed; otherwise, the pegs or keel can be amputated flush with the glenoid bone using the osteotome parallel to the glenoid surface. Hybrid glenoids with metal ingrowth require sequential removal of bone surrounding the metallic ingrowth interface, which can often be done using implant-specific instrumentation.

Once all implants are removed, an intraoperative assessment of bone loss is made. Starting with the glenoid, a 360-degree periglenoid capsular release is performed. In cases of glenoid bone loss, we feel that optimal baseplate fixation can be achieved by using an implant with a baseplate fixed by means ofa central screw. A drill path is created along a path ofglenoid bone where optimal fixation can be achieved. In cases of sufficient bone, this will be along the anatomic center line; however, in cases of glenoid bone loss, the alternative center line is utilized along the path where the base of the coracoid meets the scapular spine (► Fig. 15.8).17,18 The goal is to achieve optimal fixation with bicortical purchase of the central screw. With the pilot hole drilled and tap in place, the glenoid surface is reamed using a starting reamer and the surface assessed to determine the area of bone available for baseplate support and the amount of bone loss that may be in need of bone grafting (reamed areas represent bone available for baseplate support while areas untouched by the reamer represent deficient areas). Previous work has demonstrated that a baseplate requires at least 50% of bone support to avoid a compromising amount of micromotion that limits bone ingrowth (150 μm).19 In cases of contained defects such as those seen with failed cemented glenoid components, bone graft from the reaming can be used to fill the central contained defects. In cases of peripheral defects or significant wear patterns, a femoral head allograft is typically used as a source of bone graft and is fashioned to fill the defect. The graft is placed within the defect and secured in place with peripheral k-wires taking care in k-wire placement to avoid the path of the reamer. With the graft in place, the glenoid surface and graft are reamed together with the standard reamer and the baseplate with a porous bone ingrowth surface, and central compression screw is inserted to achieve maximum compression on the native glenoid and the glenoid graft. Peripheral locked screws are placed to neutralize micromotion.

Glenosphere selection will take into account the fixation achieved with the baseplate, the diameter of the glenoid, the overall volume of the shoulder, and any bone deficiencies. Goals of glenosphere selection are to maximize the glenoid bone and bone graft coverage, compress the bone graft, and create optimal stability and soft-tissue tension by lateralizing and distalizing appropriately. A glenosphere which has a partially hooded augment is often utilized to compress the secured graft onto the native glenoid and create a load-sharing effect as the periphery of the glenosphere becomes in direct contact with the glenoid20 (► Fig. 15.5).

On the humeral side, a monobloc stem is typically selected that allows for a metaphyseal press-fit where possible. When a press-fit cannot be achieved, the stem is cemented. In cases where a previous well-fixed cement mantle is present, the stem can be cemented within the previous cement mantle (► Fig. 15.5). Prior to inserting the final stem, the shoulder is reduced with a humeral trial in place. The joint is assessed for appropriate stability and motion. When present, instability may be due to insufficient soft-tissue tension, soft-tissue impingement, or bone impingement.

A joint that is easily dislocated with an anterior drawer with the shoulder adducted and internally rotated, and has significant diastasis with a sulcus test, is typically under-tensioned. At this stage, a thicker humeral trial is selected often with additional constraint. Careful inspection for impingement points can be done during the trial. In cases where impingement occurs between the greater tuberosity and the acromion, a tuberoplasty can be performed by burring additional areas of the greater tuberosity. Occasionally, use of a deeper socket may allow for clearance of the greater tuberosity as the shoulder is abducted, but care must be taken not to compromise soft-tissue tensioning if this step is taken. If impingement occurs at the superior glenoid or around the glenoid bone graft, this can be burred as well. Finally, soft-tissue impingement may be caused by capsular tissue, or remnant labral or rotator cuff (infraspinatus/teres minor) tissue. Previous periglenoid labral excision and capsulotomy or capsulectomy as needed will, in most cases, avoid this problem. Once appropriate stability and acceptable ROM is attained, the humerus is prepared for final implant insertion and bone tunnels drilled for subscapularis repair. The final implants are inserted and the subscapularis is repaired to decrease dead space and improve internal rotation strength.