14.3.1 Hemiarthroplasty

Although there are no reports on hemiarthroplasty specific to the B2 glenoid, results of hemiarthroplasty are clearly influenced by glenoid morphology. In an initial report of 31 hemiarthroplasties with a mean follow-up of 29 months, Levine et al reported that when the glenoid was concentric preoperatively a satisfactory outcome was achieved in 86% of cases.10 However, when the glenoid had posterior erosion, satisfactory outcomes were achieved with a hemiarthroplasty in only 63% of cases. The results deteriorated with time, particularly in the setting of glenoid erosion. In a subsequent report, 28 of the 31 hemiarthroplasties were reviewed at a mean of 17 years postoperatively.11 At long-term follow-up, they found that 42% of concentric glenoids and 13% of nonconcentric glenoids had a satisfactory result. As noted previously, Iannotti and Norris6 found that in the setting of posterior glenoid erosion, TSA led to better outcomes compared to hemiarthroplasty alone.

Matsen has popularized the “ream-and-run” procedure in which the glenoid is reamed to create a concentric socket and the humeral head is resurfaced with a hemiarthroplasty; he suggested that this technique could be used to manage a B2 glenoid. Biomechanical investigation has shown that removal of glenoid cartilage and labrum leads to decreased stability and that subsequent reaming alone of the glenoid to a concentric socket helps restore stability to the glenohumeral joint. Gilmer et al recently reported clinical outcomes of 162 ream-and-runs.12 At minimum 2-year follow-up, 124 cases had improvements in their simple shoulder test (SST) that met minimal clinically important difference (MCID) criteria, 22 (14%) required revision, and 16 (10%) did not have MCID improvement. Notably, MCID was not achieved until a mean of 6 months postoperatively and the SST did not maximally improve until 2 years postoperatively. They reported that no patients had postoperative posterior instability despite glenoid morphology. However, among the biconcave glenoids (number not reported), only 23% had MCID improvement in SST; 38% did not have MCID improvement, 14% required revision, and 21% did not have adequate follow-up. Moreover, in an earlier report at the same center with a mean of 2.7-year follow-up, Lynch et al reported that 4 of 34 (12%) ream-and-runs demonstrated progressive medial erosion and 6 (18%) had recurrent posterior glenoid erosion.13 Given that corrective reaming of a B2 glenoid involves violation of the strong subchondral bone, we do not advise this technique. We believe that violation of the subchondral bone should be avoided, as glenoid bone stock is limited and excessive reaming has been correlated with medial implant subsidence in the midterm.

Hemiarthroplasty with biological resurfacing has also been used particularly in young adults, where there is concern for longevity of a polyethylene glenoid implant. A variety of techniques have been proposed, including meniscal allograft, fascia lata, and Achilles allograft. However, the results of these techniques have been mixed and the long-term outcome of these techniques is not well quantified, particularly in the setting of a B2 glenoid.

14.3.2 Total Shoulder Arthroplasty

Multiple studies have demonstrated that TSA is superior to hemiarthroplasty alone in the treatment of primary glenohumeral arthritis. As such TSA has also been the mainstay of treatment of a B2 glenoid for several years. The goal is to correct posterior subluxation by soft-tissue balancing and reduce eccentric loading of the polyethylene component and prevent early loosening secondary to the “rocking horse” phenomenon. Several techniques can be used to address glenoid retroversion and implant a glenoid component, including eccentric glenoid reaming, posterior glenoid bone grafting, and an augmented posterior glenoid.

The most common technique to address retroversion is eccentric reaming in which the glenoid is reamed primarily anteriorly in order to re-create a concentric socket which is then resurfaced with a polyethelene glenoid (► Fig. 14.4). This technique is limited by the amount of preoperative retroversion, the fact that reaming leads to loss of bone stock as the glenoid vault narrows medially, and the impact of reaming subchondral bone. Cadaveric and computer simulation studies have indicated that approximately 15 degrees of retroversion can be successfully corrected without vault penetration.14,15,16 However, this limit relates to the penetration of the keel or pegs through the cortex of the glenoid vault and does not consider the subchondral bone or necessarily correlate with the ability to correct posterior subluxation. Moreover, a recent study by Iannotti et al indicated that even when correction of version is feasible, it is not always technically accomplished, even in the hands of an experienced surgeon.17 The difficulty in corrective reaming was highlighted further in a study by Churchill et al in a consecutive series of TSAs in which B2 glenoids were addressed with corrective reaming.18 Despite reaming, 47% of cases had incomplete glenoid seating at the time of implantation. In all cases, the subchondral bone was violated to obtain correction and in 30% of cases the violation was more than 50%.

Notably, there are devastating potential downsides to progressive medial reaming. As reaming progresses medially, the glenoid vault narrows down, which decreases the amount of bone stock available for glenoid implantation. Significant reaming may therefore result in the implantation of a smaller glenoid component with substantial mismatch between the glenoid and humeral head. Severe medialization of the glenoid may also decrease tension in the rotator cuff which may have functional consequences. Finally, excessive glenoid reaming to correct glenoid version violates the subchondral bone and may increase the risk of medial subsidence as recently demonstrated by Walch et al.19

Gerber et al reported on 23 TSAs in which eccentric reaming was used to address glenoid retroversion and posterior humeral head subluxation.20 Five patients had a B2 glenoid and the remainder had either a B1 or C glenoid. Glenoid retroversion was measured on CT scans using the Friedman’s method which is equivalent to the intermediate glenoid retroversion. The mean retroversion was 18 degrees preoperatively and corrective reaming was performed up to a maximum of 15 degrees. At 42 months postoperatively, they reported that 21 of 23 patients had postoperative resolution of posterior humeral head subluxation. However, postoperative retroversion was still a mean of 9 degrees and ranged from 0 to 25 degrees. These values are concerning, given the risk of early failure when the glenoid is implanted in greater than 10 degrees of retroversion. Finally, the results are difficult to interpret because only 12 of the 23 patients (52%) had primary glenohumeral arthritis with the remaining having mixed diagnoses. Yet, the B2 glenoid was described specifically for primary glenohumeral arthritis which is distinct in development and natural history from pathology such as posttraumatic and postinstability arthritis. Similar to Gerber et al, Habermeyer et al reported that posterior humeral head subluxation could be adequately corrected with eccentric reaming and TSA.21 At a mean follow-up of 2 years (minimum 1 year), they reported that 20 of 24 TSAs were centered. They did not state which of these patients had a B2 glenoid and their measurements of subluxation were based on plain axillary radiographs which are less precise than CT scan evaluation. Both this study and the former study by Gerber et al are also limited by the short-term follow-up.

Recently, Walch et al reported on 92 TSAs performed for B2 glenoids reviewed at a mean of 77 months postoperatively.22 Revision surgery was required in 16% and glenoid loosening was observed in 21% of cases. Preoperative posterior humeral head subluxation of 80% or greater carried an 11% rate of posterior dislocation with TSA. Similarly, when the neoglenoid retroversion was 27 degrees or greater, the risk of glenoid loosening or posterior dislocation was 44% (► Fig. 14.5). Notably, the cases requiring revision for glenoid loosening were performed at a mean of 96 months following the initial surgery and the revisions for posterior dislocation took place at a mean of 30 months following the initial surgery. These timeframes suggest that the follow-up in the aforementioned studies may not have been adequate to detect the midterm failures that occur with the use of anatomic TSA in the setting of a B2 glenoid.

In addition to eccentric glenoid reaming, it is also possible to manage posterior glenoid erosion with bone grafting or a posteriorly augmented glenoid (► Fig. 14.6). Neer and Morrison reported excellent results in 16 of 19 (89%) TSAs at a mean of 4.4 years following autologous bone grafting of the humeral head to the glenoid.23 Similarly, Steinmann and Cofield reported that at a mean of 5 years postoperatively, 23 of 28 (82%) had satisfactory results following concomitant bone grafting and TSA.24 However, three of the glenoids (11%) were considered radiographically loose and the length of follow-up was 7.7 years in the radiographically loose group compared to 5.7 years in the group without loosening, questioning the long-term viability of this approach. Hill and Norris reviewed 17 TSAs at a mean of 70 months postoperative that had undergone concomitant bone grafting to address glenoid erosion.25 Five (29%) of the grafts failed and the results were considered unsatisfactory in eight (47%) cases. Moreover, only five cases were performed for primary osteoarthritis. In the aforementioned B2 series by Walch et al,22 seven patients underwent concomitant glenoid bone grafting (from the humeral head) and anatomic TSA when it was anticipated that anterior reaming to correct retroversion to less than 10 degrees was not possible. Unfortunately, this method was successful only in two patients.

Augmented glenoids, with either a posterior-step or posterior-wedge component, have been proposed as a way to address the B2 deformity. In cases of substantial retroversion or biconcave defects, such augments can actually require less reaming of native bone compared to a standard polyethylene implant. In biomechanical analysis, a polyethylene posterior-step glenoid performs similar to a standard TSA in simulated B2 defects.26 While the concept of preserving the joint is appealing, neither biomechanical nor reaming studies can account for the static instability (posterior subluxation) seen with a B2 glenoid. Rice et al reported a clinical study of 14 TSAs performed with a polyethylene glenoid augmented with a posterior wedge.27 They found that instability, particularly posterior subluxation, persisted despite the novel glenoid and the device has been discontinued. Similarly, Neer et al described the use of a posteriorly augmented polyethelene component that was discontinued.3,23

In addition to addressing the glenoid, strategies for restoring glenohumeral stability in the setting of a B2 glenoid include anterior release, posterior capsulorrhaphy, and altering humeral retroversion. As in all cases of shoulder arthroplasty, a 360-degree release of the subscapularis tendon is important in restoring motion and achieving soft-tissue balance. In the setting of posterior humeral head subluxation, the anterior structures become tight and push the humeral head posteriorly. Therefore, a thorough anterior release may help restore glenohumeral stability. In the corollary to anterior release, a posterior capsular plication can be performed at the time of TSA to tighten the posterior structures. Most studies make only casual mention of posterior capsular plication with unclear guidelines for its use. In the B2 series by Walch et al,22 nine patients had a supplemental posterior capsular plication when intraoperative testing with the arm in neutral revealed posterior static subluxation of more than 50% of the humeral head diameter. These patients were noted to have lower postoperative forward elevation and mobility component of the Constant score. In addition, three patients who underwent revision surgery for postoperative posterior dislocation had a posterior capsular plication and all three were considered failures. Finally, performing the humeral osteotomy in less retroversion may theoretically be used to compensate for posterior subluxation. The utility of these techniques, however, are difficult to quantify given that they are often used together and in conjunction with the aforementioned techniques described to address posterior glenoid erosion. Moreover, it appears that even when correction of the posterior glenoid erosion is possible, the long-term ability to correct posterior instability is not reliable. In other words, posterior humeral head subluxation appears to be the “primum movens” or source, rather than the consequence of a B2 glenoid.