Glenoid bone loss when present with primary osteoarthritis is usually posterior and with rotator cuff deficiency is superior and medial. When primary arthroplasty is performed, use of the humeral head as a source of autogenous graft material can provide restoration of the fossa volume and version and inclination. When cavitary glenoid bone loss exists after failure of a glenoid component allograft or autograft, cancellous bone graft can be helpful if the defect is contained within the vault and the rim is intact. In cases of a noncontained glenoid defect, uncontained structural grafts are preferred.
Generous exposure of the glenoid is needed to use bone grafts for glenoid bone deficiency.
When structural grafts are used with a resurfacing polyethylene component, additional screw fixation of the graft is needed. A metal-backed glenoid with screw fixation may be sufficient in many cases to fix the graft material without the use of additional screws for the graft.
Use of intraoperative templates for graft size and preparation can facilitate fit and fill of the graft. These are outlined in this chapter
Glenoid bone loss is a late consequence of severe arthritis or a result of a failed glenoid component. The pattern of bone loss often follows a pattern typical of the specific disease process. Posterior glenoid bone loss occurs in advanced osteoarthritis, superior and central bone loss with rotator cuff deficient arthritis, and cavitary bone loss with a loose glenoid component. The severity of the bone loss is variable but when significant will require modification of the standard surgical technique and postoperative rehabilitation and can decrease the clinical outcome.
Walch et al. have classified glenoid bone loss associated with osteoarthritis based upon the extent of glenoid retroversion, glenoid erosion, and the position of the humeral head in relation to the center of the glenoid fossa ( Fig. 10-1 ). Within this classification there are three main types of glenoid wear: A, B, and C. In type A, bone loss is concentric and the humeral head remains centered. Type A is further subdivided into minor bone loss (A1) and major bone loss (A2). For type B, the bone loss is eccentrically posterior with the humeral head subluxation. Subluxation is defined as translation of the humeral head in the glenoid fossa greater than 5% of the overall diameter of the humeral head. Type B is also further subdivided into minor bone loss (B1) and major bone loss (B2). Furthermore, type B2 represents the classic “bi-concave” glenoid wear pattern seen in some advanced stages of glenohumeral osteoarthritis. Type C is defined by glenoid retroversion greater than 25 degrees, regardless of humeral head position, as is often but not exclusively associated with congenital hypoplasia of the glenoid.
Walch analyzed the preoperative computed tomography (CT) scans of 113 patients with primary osteoarthritis. Two observers classified glenoid morphology into one the three major categories (A, B, or C) with overall good interrater agreement. Although they did not correlate their classification with clinical outcomes, they suggested that an awareness of different glenoid pathologic morphologies is important for surgical planning.
The relevance of understanding glenoid morphology as well as any tendency of humeral head subluxation has been demonstrated in several clinical series. In a review of 31 patients, Levine et al. demonstrated greater than a 2.5-fold rate of failure or poor outcomes in those shoulders whose eccentric glenoid wear was left uncorrected at the time of humeral replacement for primary arthritis. Others have demonstrated that the prosthetic glenoid component, if implanted in abnormal or pathologic glenoid version, experiences adverse loading resulting in high shear forces, increased rim loading, prosthetic loosening, and ultimately higher rates of revision surgery. Iannotti and Norris demonstrated in a review of 29 patients undergoing total shoulder arthroplasty (TSA) with moderate or severe glenoid erosions that clinical outcomes could be improved in those whose glenoid version was corrected toward physiologic with appropriate glenoid reaming.
Rotator Cuff Tear Arthroplasty
Rotator cuff tear arthropathy (RCTA) represents a different pathologic process with different forces across the glenoid surface resulting in patterns of glenoid wear unlike those seen in standard glenohumeral osteoarthritis. Distinct from osteoarthritis, in which the rotator cuff is typically intact, RCTA is associated with superior migration of the humeral head within the glenoid fossa. When erosions do occur, they are more often in the superior portion of the glenoid surface. Favard et al. originally defined four types of glenoid erosions associated with RCTA (E0, E1, E2, and E3) ( Fig. 10-2 ). E0 represents superior humeral head migration without glenoid erosions, whereas E1 demonstrates central and concentric glenoid bone loss. In type E2 the superior portion of the glenoid is eroded, and in type E3 the predominately superior glenoid erosions extends into the inferior part of the glenoid. Oudet et al. reported the presence of superior glenoid wear correlated with lower functional outcomes in a series of 66 unconstrained shoulder arthroplasties. In their review of 80 patients undergoing total shoulder arthroplasty using a reverse-type prosthesis for RCTA, Sirveaux et al. demonstrated that the extent of scapular notching (i.e., erosion of the glenoid neck secondary to impingement with the humeral component during adduction) was associated with greater degrees of superior glenoid erosions (i.e., types E2 and E3). Furthermore, the presence of notching negatively affected the Constant-Murley scores when it involved the inferior fixation screw of the glenoid baseplate. The authors attributed this relationship with postoperative notching to the tendency of the surgeon to implant the glenoid component with relative superior tilt in the presence of superior glenoid erosions. The results of bone grafting superior glenoid defects in cases of RCTA has been recently addressed. These early results suggest that reliable pain relief can be achieved, whereas functional improvements are less predictable.
Failed Total Shoulder Arthroplasty
An entirely different type of glenoid bone loss can be encountered in situations of a failed total shoulder arthroplasty. Loosening of a glenoid component (keeled or pegged) may result in varying degrees glenoid vault bone loss, occasionally leaving the entire cancellous vault devoid of bone. Further involvement of the cortical vault walls can result in truly massive defects, leaving limited reconstructive options.
A classification for glenoid defects in 48 patients seen at the time of revision arthroplasty was presented by Antuna et al. Defects were divided into central and peripheral and combined with further subdivision into mild, moderate, and severe ( Fig. 10-3 ). In this series the difficulties in managing defects involving both the glenoid vault and walls were outlined. In these extreme cases, bone grafting may be the only means by which subsequent glenoid reimplantation can be achieved. In their series, 66% of the patients with glenoid defects treated with impacted allograft bone demonstrated satisfactory pain relief; however, the extent of graft resorption was not reported.
Williams provided a classification of bone loss based upon three structural components of the glenoid vault: the subchondral bone of the fossa (S), the vault rim (R), and the vault wall (W) ( Fig. 10-4 ). Glenoid defects with an intact rim and wall are contained defects, and uncontained defects have a deficiency of the rim and or vault wall. Given wide variations in glenoid bone deficiency in this classification, each aspect of the vault (S, R, and W) is described by the amount and location of the deficiency. Glenoid components that derive their stability from the subchondral bone should have approximately 40% or greater surface area to seat the back side of the component. The intactness of the vault rim and vault wall influences the ability to use either nonstructural (cancellous bone chips) or structural bone (femoral head or cortical cancellous bulk iliac crest graft) to fill the defect.
Uncontained surface defects as seen in primary arthroplasty for osteoarthritis or rotator cuff tear arthroplasty require structural bone grafts. These defects most commonly occur in patients with severe osteoarthritis or chronic dislocations or after glenoid fossa fractures with posttraumatic arthritis. The humeral head is an excellent source of graft for most patients undergoing primary arthroplasty. Bicortical or tricortical iliac crest grafts are ideal for revision cases. Allograft can be used but has greater likelihood of incomplete bone incorporation and bone graft collapse.
Uncontained cavitary defects are often seen after failed total shoulder arthroplasty. A structural graft is needed when there is insufficient support from the vault rim or walls for the humeral head to articulate. With uncontained cavitary defects, structural graft can be obtained from the iliac crest if the defect is smaller is size. In many cases the defect is large and an allograft is required to fill the defect. In most cases with massive uncontained cavitary defects the best allograft is a frozen femoral head. In cases in which there is a contained defect (intact glenoid vault wall and rim), cancellous bone grafts packed into the defect are sufficient. Cancellous bone can be obtained from the iliac crest or from an allograft source. In cases in which there is at least 40% of the native subchondral surface to place another glenoid component, the authors will, when indicated, place another glenoid component at the time of revision surgery. In cases in which there is a large irreparable rotator cuff deficiency, a reverse arthroplasty is performed; when the cuff is intact or reparable (one tendon defect), an anatomic total shoulder is performed. When there is less than this amount of native glenoid surface, the authors perform a grafting with a hemiarthroplasty. If the patient has a deficient cuff, a hemiarthroplasty with a reverse stem (Delta X-tend with CTA head, DePuy, Warsaw, Ind.) is preferred. When the cuff is intact, an anatomic hemiarthroplasty is placed. If the patient has persistent pain and requires revision to place the glenoid component, we wait at least 9 months and preferably 12 months before the second-stage revision is performed. In this time frame the graft has incorporated sufficiently to allow for secure placement of the reverse or anatomic glenoid components.
Some surgeons have used a soft tissue graft (fascia lata) over the morcellized graft material as a means to contain the graft. We have used this technique on occasion but have not found it useful. When there is a contained defect and an intact rim, the cancellous graft is contained by the articulating humeral prosthetic. There are also the concerns of difficult healing and graft incorporation when adding an avascular allograft material (facia lata) over another avascular allograft material (cancellous bone). When there is an uncontained defect, we prefer to use structural grafts.
Measurement of glenoid bone defects and preoperative planning have traditionally relied upon anterior posterior and axillary radiographs and two-dimensional (2D) computed tomography. Newer techniques have utilized three-dimensional (3D) computed tomography to better define the location and severity of glenoid bone loss. An interactive preoperative computer surgical simulator has been developed by our research group at the Cleveland Clinic. The use of a vault implant that reflects the size and shape of the typical normal glenoid vault can be implanted into the abnormal glenoid vault in the computer simulator to determine the correct version and degree and location of the glenoid bone defect. After the glenoid vault defect is defined by this method, the surgical steps (bone graft and reaming) to best correct the deficiency can be planned using the software. Options for management of uncontained fossa defects include reaming the high side to achieve anatomic glenoid version, altering glenoid version from its anatomic position, use of a bone graft, or use of an augmented glenoid component. In many cases a combination of methods are used to preserve glenoid bone and at the same time correction of abnormal version.
Our clinical experience has demonstrated that most defects of less than 1 cm or an increase in glenoid retroversion of 10 degrees or less (15 to 20 degrees of measured retroversion) can be managed by reaming the high side. Greater amounts of increased glenoid retroversion can not in most cases be managed by asymmetric reaming because of excessive loss of bone and inability to place a proper sized or adequately fixed glenoid component. In these cases we currently prefer to use a bone graft. In the future if a augmented glenoid component were to be available, this might allow for conservation of bone and normalization of glenoid version.
In most cases, use of a glenoid bone graft to reconstruct a glenoid defect can be technically challenging. Having a method for dealing with these defects provides a foundation to deal with each individual case. Although the types of bone loss fall into typical patterns and the techniques to manage them can be put into a typical set of steps, each patient will have their own specific pathology and patient factors that may alter the specific methods described in this chapter to manage a specific patient’s glenoid bone loss. There is very little written about the surgical methods for glenoid grafting in either textbooks or the peer review literature, and no specific methods have been shown to be superior over another. In this chapter the authors provide case examples with detailed figure legends to describe some commonly used surgical techniques that have been shown to be successful for typical patterns of glenoid bone loss. The authors offer these cases to illustrate our current and preferred method to manage glenoid bone deficiency.