Background

The different types of glenoid morphology encountered in osteoarthritis were first described and classified by Walch in 1999 [37] and modified further in 2016 [4]. The B2 subtype is defined by an asymmetric posterior bone loss with biconcavity and posterior translation of the humeral head. Management of end-stage osteoarthritis with a B2 glenoid deformity presents a challenging problem in shoulder arthroplasty. The etiology of the Walch B2 shoulder remains unclear but recent evidence has highlighted its complexity. Raniga [27] and colleagues have shown that in Walch Type B glenoids, the humerus has significantly less retrotorsion than non-osteoarthritic shoulders. More recently, the same group also showed that rotator cuff muscles may be responsible for an increasing shear force on the glenoid face with increasing retroversion, possibly exacerbating the glenoid deformity [5].

In 2012 Walch [38] et al. reported a higher failure rate with non-constrained anatomic total shoulder arthroplasty (aTSA) using eccentric reaming and bone grafting in patients with a B2 glenoid compared to patients with a less severe glenoid retroversion deformity such as the A1 glenoid. The authors recommended that clinicians should consider using a reverse arthroplasty with or without bone grafting, which had a more predictable outcome in patients with significant glenoid deformity. Recently, aTSA with a posterior augmented glenoid (PAG) wedge or step has been gaining popularity. Recent publications have reported good clinical and radiological outcomes with short-to-midterm follow-up (24–48 months; [9, 10, 15, 16, 30, 31, 36, 40]).

Aims of the study

The aim of this systematic review is to study the outcomes of aTSA in B2 glenoids both before and after the 2012 publication by Walch et al. and to determine whether there has been a positive change, which may encourage surgeons to still consider aTSA in these complex cases. Furthermore, we hypothesized that aTSA techniques used after 2012 are associated with a lower revision rate.

Materials and methods

Database search

A systematic search was performed according to the Preferred Reporting Items Systematic reviews and Meta-analysis (PRISMA) guidelines ([24]; Fig. 1). The review was conducted in four phases. In the first phase, the relevant articles were searched for TSA studies. PubMed, Embase, OVID Medline, Cochrane, and Google Scholar databases were accessed on 28 October 2022. We searched for terms with the following items (Glenohumeral arthritis) or (Shoulder arthritis) or (Glenohumeral bone loss) or (Glenohumeral wear) or (Glenohumeral biconcavity) or (Glenoid Retroversion) or (B2 Deformity) or (Glenoid wear) or AND (Total Shoulder replacement) or (anatomic shoulder arthroplasty) or (Eccentric reaming) or (Augmented Glenoid) or (Glenoid Bone grafting). The initial search yielded 268 articles. Original studies, case reports, simulation studies, or studies that included patients who underwent reverse shoulder arthroplasty were excluded after the first (120) and second selection phase (104), and 25 publications remained after the final review (Fig. 1).

Fig. 1
figure 1

Flowchart of the study selection procedure

Full size image

Study selection

After the removal of duplicates, the titles and abstracts were screened by two reviewers independently, and potentially eligible studies underwent a full-text review. The titles and abstracts were screened for relevance in the second phase. Publications were included if they met all of the following criteria: primary aTSA; B2 glenoid morphology; postoperative clinical or radiological analysis; and a minimum of 24 months of follow-up. Articles were excluded for the following reasons: non-English text; revision cases; only abstract available; hemiarthroplasty cases; glenoid morphology other than B2. In the last phase of our review, all of the articles that met the inclusion criteria underwent a comprehensive analysis for quality. The main techniques of the anatomic arthroplasties were reviewed. We only selected papers that allowed us to look at B2 subgroups in detail for our series. Once a paper was identified as suitable for inclusion in the review, it was categorized initially according to when the procedures were performed and published. Articles that were included were grouped into two categories, i.e., before 2012 and after 2012 (the year of the Walch paper recommending a change in strategy for B2 glenoid management).

Statistical analysis

Means and standard deviations were calculated for continuous variables. For the analysis of contingency tables, Fisher’s exact test was used. The level of significance was set at p < 0.05. Our aim was to determine whether there was a statistical difference in the reported outcomes (by revision/reoperation; Table 1) between the earlier papers, which led to the Walch recommendation, and those papers based on more recent knowledge, planning, experience, and technology.

Table 1 Statistical analysis of outcomes
Full size table

Results

In total, 779 TSA procedures with B2 glenoids were reviewed with a mean follow-up of 55.6 months (range: 24–110 months). We formed two groups according to the date of publication: before 2012 (group I, n = 171) and after 2012 (group II, n = 608; Tables 2 and 3).

Table 2 Group I: studies performed before 2012
Full size table
Table 3 Group II: studies performed after 2012
Full size table

Eccentric reaming was performed on 114 shoulders in group I with a mean follow-up of 47.67 (SD: 26.95) months and on 332 shoulders in group II with a mean follow-up of 55.37 (SD: 23.13) months. The revision rate decreased significantly from 13.15% to 3.28%, (odds ratio: 3.5; p = 0.0002; CI: 1.49–8.64).

Posterior bone grafting was reported in 43 shoulders in group I with a mean follow-up of 70.30 (SD: 9.48) months versus 36 with a follow-up of 65.33 months (SD: 25.81) in group II. Complications in these groups have not significantly changed over the years (odds ratio: 0.683, p = 0.7245, CI: 0.12–3.25) and remain concerning (Table 4).

Table 4 Summary of complications before 2012 and after 2012
Full size table

The use of the posterior augmented glenoid has increased significantly in the past 10 years. This was reported in 14 shoulders in group I (mean follow-up of 60 months) versus 240 shoulders in group II (mean follow-up of 32.7 months, SD: 7.82). In group I there were no prosthesis-related complications, while only five prosthesis-related complications were found in group II, resulting in revision. A comparison of these two groups showed no statistically significant difference in complication rates (odds ratio: 0.0; p = 1.0; CI: 0.0–15.60).

Outcomes

With respect to the clinical data, a number of different patient-reported outcome measures (PROM) were used including the Constant Score, American Shoulder and Elbow Surgeons (ASES) Score, Penn Shoulder Score, and the Neer rating system. Overall, 11 articles used the Constant Score, eight the ASES, three the Neer rating system, two the Oxford Shoulder Score, one the Western Ontario Osteoarthritis of Shoulder (WOOS) Index and the one the Subjective Shoulder Value (SSV). In each of the papers reviewed, the improvement in these PROMs was demonstrated to be clinically significant. In group I the Constant Score improved from 36.7 to 75.2, while in group II there was an improvement from 38.7 to 77.8: There was no significant difference between the groups. The other scores (WOOS, SSV etc.) were used inconsistently, making comparison of these outcomes difficult.

Radiological parameters were evaluated by most authors using the Lazarus Score [20]. In group I, only one paper used this scoring system, and two used the radiolucent line scoring system according to Molé, compared to group II in which 14 papers used the Lazarus scoring system. This lack of data in group I made meaningful comparison between the groups impossible.

When considering the combined outcomes for both groups I and II (without making any distinction between the operation techniques), there was a statistically significant decrease in revision rates in papers published after 2012 (odds ratio: 3.12, p = 0.0012, CI: 1.56–6.17). Analyzing the subtypes of procedures showed that bone grafting techniques did not result in improved outcomes over the years. By contrast, eccentric reaming has demonstrated improved results in more recent years compared to earlier reports (Table 4). Posterior glenoid prosthetic augmentation has maintained a low revision rate to date, although longer-term data are pending.

Discussion

Management of the Walch B2 shoulder remains controversial. Since the publication by Walch et al. in 2012 [38], there has been an improved understanding of glenoid morphology, which has led to more precise preoperative planning and the introduction of innovative solutions such as posterior augmented glenoid components [10, 16, 31, 33, 36, 40]. Furthermore, we hypothesized that anatomical total joint replacement techniques after 2012 were associated with a lower revision rate.

Eccentric Reaming

Based on our systematic review, it can be concluded that for the past 10 years, outcomes using eccentric reaming have improved and have resulted in a more reliable, precise technique with lower revision rates. Earlier reports using eccentric reaming to correct excessive retroversion of the glenoid showed a high revision rate [38] with early loosening, posterior instability, component malposition, and incomplete correction of pathologic retroversion. A better understanding of the premorbid anatomy, improvement in preoperative planning, and more precise techniques may have contributed to the reduced rate of these complications [3].

Raiss et al. [26] showed high concordance between their preoperative plan and the final implant selection. Their study used the planned glenoid implant size in 96% of cases. However, 42 out of 100 implanted anatomical shoulder prostheses showed B2 deformity. The correction of the B2 glenoid retroversion still has limitations. Because the normal version of the glenoid is approximately 7° retroverted, aiming to correct the version to neutral may lead to overcorrection and unnecessary removal of bone, particularly anterior subchondral bone [6]. Anteversion correction of more than 15° results in excessive glenoid bone removal, violation of the cortical bone, and increased medialization with possible glenoid vault perforation and change of the rotator cuff tension.

Bone grafting

With regard to bone grafting, we found that there was no significant difference between the two groups. Theoretically, bone grafting addresses several issues including bone loss, correction of retroversion, and soft tissue balancing. Furthermore, it provides a solid foundation for glenoid component seating and protects against potential vault perforation. However, clinical and radiological outcome studies have reported varied results [19, 22, 29, 38].

In their paper, Walch [38] and colleagues reported that seven of 92 B2 shoulders (7.6%) required a posterior bone graft and resulted in five cases of complications that affected the implant longevity and functional outcome (Table 2). Recent studies using bone grafts have not been able to improve on the failure rates reported earlier. However, overall, better clinical outcomes have been reported [19, 22, 29]. Nicholson [22] et al. in their study of 15 B2 glenoids showed good clinical and radiological outcomes with 100% graft incorporation without the need for revision surgery. They suggested that the consistent incorporation of the bone graft may be a result of the cortex-to-cortex matched surfaces of the glenoid-to-humeral head graft, the compression of those surfaces by the screw technique that does not allow any cement to extrude between the surfaces, and the normalization of load on the corrected glenoid version. As an alternative, Sabesen [29] et al. recommended preparing the glenoid with a step-cut and felt this provided more stability than a wedge-shape graft with screw fixation.

Posterior glenoid components

Prior to 2012, there was only one publication reporting on the results of posterior augmented glenoid components with no complications [28]. Since then, its use has become more popular especially in younger patients when glenoid retroversion exceeds 15°. There are three alternative designs available: wedge-shaped, stepped, or irregular. When comparing these options, a wedge has a better capacity for version correction, improved loading characteristics [17], and less bone removal [2, 17] than other alternatives [13].

A stepped augmented glenoid requires removal of additional bone to seat the implant and may result in slight medialization of the joint line and consequent rotator cuff musculotendinous unit shortening [40]. There are limitations in using polyethylene augments. Their use in cases of large angular deformity appears to create a significant increase in shear force on the implants [39]. Ko et al. reported a higher failure rate with augments greater than 16° [25]. Using finite element analysis, Hermida et al. [13] found that wedge-shaped implants dissipated compressive stress better and predicted a greater lifespan before cyclic fatigue failure.

A porous metal glenoid augment replicates the capabilities of a wedge-shaped autologous bone graft, but without technical difficulties, which include bone preparation and fixation, or the propensity for graft collapse in the medium to long term [31]. Kevin et al. [18] demonstrated that augmented glenoid components allowed for greater correction in radiographic parameters with significant improvements in glenoid version and a decrease in humeral head subluxation in relation to both the glenoid face and the scapular plane. Sandow [30, 31] and colleagues reported promising results with a trabecular metal implant and a cemented glenoid component with a mean follow-up of 55.6 months without implant-related complications. Wedge-shaped augments achieved satisfactory loading and version correction; however, the long-term outcome remains unknown.

Revision and complication rates

The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR; [1]) reports the prosthesis-related revision rate for mid-head prothesis (humeral head resection with an epiphyseal fixation stem) to be 2.0% in the midterm, which is better than the primary total stemmed shoulder replacement rate of 3.5% at 3 years. By comparison the revision rate for reverse shoulder arthroplasty in patients with a B2 glenoid at 3 years is 2.7% [1]. As the revision rate for rTSA has remained constant over a longer period of time [1], longer follow-up for eccentric reaming and PAG aTSA will be needed in order to better compare the results to the rTSA. Nevertheless, the short- to medium-term results of eccentric reaming and PAG are positive.

The mean follow-up times do not differ significantly for eccentric reaming and bone grafting before and after 2012. For posterior glenoid augmentation a similar conclusion cannot be made due to the lack of observations before 2012. Since the mean follow-up does not differ significantly before and after 2012, the observed improvement in the complication rates (eccentric reaming: 13.15% vs. 3.31%, bone grafting: 9.30% vs. 5.56%) is mostly attributable to the improvement in the surgical options applied.

Limitations

This study has also some limitations. Firstly, completely different types of fixation methods are compared. In most cases stemmed shoulder replacements were used, but in the past few years stemless midhead prostheses have become popular. We did not find a significant difference between the revision rates in the groups based on the AOANJRR reports [1]. Secondly, we have not looked into the revision rate of the different glenoid components used in our review. Furthermore, the heterogeneity of the clinical and radiological parameters used in the examined studies makes a fair analysis difficult.

The strength of our study is in the volume of papers and the number of patients with B2 glenoids who underwent aTSA. To our knowledge, this is the only review to date that has compared the changing results of aTSA in the treatment of B2 shoulders before 2012 and after 2012 with the evolution of newer prostheses and surgical planning.

Practical conclusion

  • The outcomes of anatomic total shoulder arthroplasty for osteoarthritis with B2 glenoids have shown significantly lower prothesis-related complication rates since 2012.

  • Eccentric reaming and posterior augmented glenoid components may now be considered as reliable surgical solutions in managing many patients with a B2 glenoid and an intact rotator cuff.

  • Using modern preoperative planning techniques gives surgeons a better understanding of the glenoid morphology and enables more accurate placement of components, hopefully leading to longer survival.