If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Glenoid deficiency and erosion (excessive retroversion/inclination) must be corrected in reverse shoulder arthroplasty (RSA) to avoid prosthetic notching or instability and to maximize function, range of motion, and prosthesis longevity. This study reports the results of RSA with an angled, autologous glenoid graft harvested from the humerus (angled BIO-RSA).
A trapezoidal bone graft, harvested from the humeral head and fixed with a long-post baseplate and screws, was used to compensate for residual glenoid bone loss/erosion. For simple to moderate (<25°) glenoid defects, standardized instrumentation combined with some eccentric reaming (<15°) was used to reconstruct the glenoid and obtain neutral implant alignment. For severe (>25°) and complex (multiplanar) glenoid bone defects, patient-specific grafts and guides were used after 3-dimensional planning. Patients were reviewed with minimum 2 years of follow-up. Mean follow-up was 36 months (range, 24-81 months). Preoperative and postoperative measurements of inclination and version were performed in the plane of the scapula on computed tomography images.
The study included 54 patients (41 women, 13 men; mean 73 years old). Fifteen patients had combined vertical and horizontal glenoid bone deficiency. Among E2/E3 glenoids, inclination improved from 37° (range, 14° to 84°) to 10.2° (range −28° to 36°, P < .001). Among B2/C glenoids, retroversion improved from −21° (range, −49° to 0°) to −10.6° (−32° to 4°, P = .06). Complete radiographic incorporation of the graft occurred in 94% (51 of 54). Complications included infection in 1 and clinical aseptic baseplate loosening in 2. Mild notching occurred in 25% (13 of 51) of patients. Constant-Murley and Subjective Shoulder Value assessments increased from 31 to 68 and from 30% to 83%, respectively (P < .001).
Angled BIO-RSA predictably corrects glenoid deficiency, including severe (>25°) multiplanar deformity. Graft incorporation is predictable. Advantages of using an autograftharvested in situ include bone stock augmentation, lateralization, low donor-site morbidity, low relative cost, and flexibility needed to simultaneously correct posterior and superior glenoid defects.
Failure to correct superior bone loss (ie, Favard type E2, E3) can lead to superior tilt of the baseplate, increased scapular impingement, instability, inferior scapular notching, and medial polyethylene wear.
Biomechanically, superior tilt increases tensile baseplate forces during deltoid contraction and can lead to early loosening.
In osteoarthritis with severe retroversion and biconcavity (ie, Walch type B2) or excessive hypoplastic glenoid retroversion (type C), failure to correct posterior bone loss can lead to retroversion of the baseplate, reduced external rotation, posterior scapular notching, and posteromedial polyethylene wear.
In addition, failure to restore the glenoid bone stock during RSA leads to excessive medialization of the center of rotation. Secondary medialization of the humerus can potentially decrease motion and flexion strength, decrease deltoid wrapping, increase the propensity for instability, and fail to restore the rounded cosmesis of the shoulder.
Historically, options to address glenoid bone defects combined eccentric reaming with glenoid bone grafting with an autograft iliac crest bone graft (ICBG) or allograft or augmented glenoid baseplates.
The humeral head autograft may be symmetrical (BIO-RSA) or asymmetrical (angled BIO-RSA), depending on the presence, amount, and orientation of glenoid deficiency. In the latter technique, purpose-designed instrumentation is used to harvest the graft from the humeral head such that it is trapezoidal to match the glenoid bone defect. Potential advantages of this technique include flexibility to reconstruct multiplanar deformity, restoration of glenoid bone stock, and ability to lateralize the center of rotation.
The present study reports the results of angled bony-increased offset-reverse shoulder arthroplasty (angled BIO-RSA) in addressing glenoid bone deficiency. We sought to determine (1) the amount of glenoid deficiency that can be corrected, (2) the rate of graft incorporation, (3) the rate of scapular notching, and (4) the functional outcomes. We hypothesized that the asymmetrical (trapezoidal) autologous bone graft harvested from the humeral head (1) would incorporate to the native scapula, (2) would restore the glenoid bone stock and allow correct baseplate alignment, and (3) would lateralize the center of rotation and provide good functional results.
Materials and methods
This was a retrospective single-center study including a consecutive series of patients that underwent a primary RSA and an angled, autologous humeral head bone graft to address significant glenoid bone loss (Walch A2, B2, C or Favard E2, E3, E4). Exclusion criteria were patients who underwent a revision RSA and those who underwent primary RSA and glenoid reconstruction with other sources of bone graft, such as ICBG or allograft.
Between 2006 and 2013, 93 patients with severe glenoid erosion or deficiency underwent glenoid bone graft reconstruction during primary RSA implantation. After excluding patients who received an allograft or an ICBG, 63 patients remained who were treated with an angled BIO-RSA. Two patients died, and 7 were lost to follow-up before 2 years, leaving 54 patients (89% of those eligible) who were included in the study. Mean follow-up was 36 months (range, 24-81 months). All patients signed written, informed consent before the procedure.
Angled BIO-RSA concept
The concept of angled BIO-RSA is to use an autologous, trapezoidal bone graft harvested from the humeral head to restore the glenoid bone stock and lateralize the center of rotation of the RSA. The technique provides the flexibility to reconstruct multiplanar deformity (ie, to correct baseplate version and inclination; Fig. 1). Although this is a retrospective study and no prospective protocol was applied, during the study period, we generally used this technique for deformities exceeding 10° to 15°, and smaller deformities were generally corrected with eccentric reaming alone.
Two-dimensional preoperative templating
Our experience has been that preoperative templating is critical to determine the graft thickness necessary to restore a normal or nearly normal inclination and retroversion. Computed tomography (CT) scans of the shoulder were obtained in all patients scheduled for an RSA. We used OsiriX 5.6 software (Pixmeo, Geneva, Switzerland) for 2-dimensional (2D) preoperative templating. This software provided simple-to-use tools for making linear and angular measurements. The modified Friedman method was used to measure the 2D glenoid version.
is the angle between the line of the glenoid fossa (where an anatomic total shoulder arthroplasty glenoid component would be implanted) and the floor of the supraspinatus fossa. This angle represented the global glenoid inclination. Because an appropriately placed baseplate only occupies the inferior portion of the glenoid and because the glenoid has a curvature, the global glenoid inclination provided by the β angle systematically underestimates the focal inclination at the level of the baseplate. Our experience has been that preoperative planning using the β angle leads to baseplate placement with insufficient correction of superior tilt (Fig. 2, A).
By contrast, the RSA angle is defined as the angle between the inferior glenoid fossa (where a reverse total shoulder arthroplasty baseplate would be implanted) and the perpendicular to the floor of the supraspinatus fossa. Development of the RSA angle provided a method to describe the local, inferior glenoid inclination, which is the inclination relevant to implantation of a baseplate in RSA (Fig. 2, B). With increased graft thickness for lateralization, the graft becomes trapezoidal (Fig. 2, C).
3D preoperative templating
In case of severe (>25°) or multiplanar deformity or erosion, our preference is to use 3D preoperative planning (Glenosys software; Imascap, Brest, France). Standard preoperative CT images of the patient's scapula allowed for creation of a 3D virtual model of the patient's glenoid and accurate and reproducible measurements of version, inclination, bone loss, and bone graft dimensions.
Using algorithms specific to the manufacturer, the virtual model was used to template placement of the guidewire and central peg and to predetermine optimal baseplate positioning (Fig. 3). A patient-specific guide was then fabricated to direct the insertion point and orientation of the central guidewire.
All surgical procedures were performed by the senior author (P.B.) or under his direction. The procedure was done with the patient in the beach chair position and under general anesthesia with an interscalene block. The deltopectoral approach was used in all cases. Any remaining subscapularis tendon was detached from the lesser tuberosity and was tagged for reattachment at the end of the procedure, if appropriate. The long head of the biceps tendon was tenodesed in the groove, if still present. The goals of bone grafting were to restore version and inclination to as close to neutral as possible while increasing glenoid bone stock and lateralizing the joint line.
For simple to moderate (<25°) glenoid defects, standardized instrumentation combined with some eccentric reaming was used to reconstruct the glenoid and obtain neutral implant alignment (Fig. 4). After exposure and dislocation of the humeral head, an oscillating saw was used to remove a small amount of bone at the summit of the humeral head, providing a flat surface and removing the subchondral plate. A threaded guidewire was then placed perpendicular to this cut and driven to the lateral cortex of the humerus. A cannulated drill was used to bore a central hole with a diameter of 8 mm. A bell saw was passed to the desired depth. An oblique cutting guide was then used to harvest bone graft with a 12° angle (12.5 mm thick) and perform the anatomic neck osteotomy. The graft was removed and inserted over the long peg (25 or 30 mm) of the Aequalis baseplate (Wright-Tornier, Memphis, TN, USA).
In most cases, the inferior glenoid rim was still intact and identifiable and was used as landmark to check guidewire placement. Our aim was to place the guidewire to allow the baseplate to sit flush with the inferior border of the glenoid perpendicular to the line of the floor of the supraspinatus fossa. After insertion of a threaded guidewire into the glenoid vault using the 10° inferiorly angled guide, a cannulated reamer was used to abrade the glenoid until the subchondral plate was reached, which was typically at a depth of approximately 2 to 5 mm. In our experience, no more than 15° of eccentric reaming can be used to correct abnormal retroversion or inclination, and we avoided excessive reaming. Depending on the depth of the deformity, this reaming technique can create a mildly biplanar glenoid face, which can be overcome through compression of the cancellous bone of the graft against the glenoid face.
The central peg was drilled, and small peripheral drill holes were made into the reamed and unreamed portions of the glenoid face using a threaded guidewire to obtain a complete bleeding bone surface. The long-peg (25 mm or 30 mm) baseplate and trapezoidal bone graft were impacted into the center hole, with the bone graft oriented to fill the defect. Baseplate fixation was performed first with nonlocking compression screws in the superior and inferior holes directed parallel to the central peg. Once compressed, convergent locking screws were placed in the anterior and posterior baseplate holes. Finally, the glenosphere was fixed to the baseplate with both a Morse taper and a countersunk screw. Humeral preparation and implantation was performed using the standard surgical technique described for the Aequalis reversed prosthesis (Wright-Tornier).
For more complex glenoid bone defects or severe glenoid erosion (>25°), patient-specific grafts and guides were used. In these cases, the bell saw was extended deeper into the humeral cancellous metaphysis. A 5-mm osteotome was then passed through an anterior cortical window at the level of the surgical neck to free the bone graft distally. The bone graft was then contoured to fit the defect according to the preoperative 3D-CT plan. The patient-specific guide was used to position the guidewire with the desired inclination and retroversion correction. The high sides (anterior or inferior, or both) of the glenoid surface were slightly reamed with a cannulated reamer (<15°). In these cases of extreme glenoid deformity, in which the circular reamer cannot be flush with the glenoid, unreamed areas were abraded with a curette or a burr to stimulate graft healing. A longer-peg (30- or 35-mm) baseplate was used to fixate the patient-specific angled graft to the native glenoid. The final fixation of the graft was achieved using long screws through the baseplate, spanning the graft into the native glenoid. After final fixation of the glenoid, the humeral preparation and implantation was completed using the standard technique.
The rehabilitation protocol used for the angled BIO-RSA was no different from that for a standard RSA.
were measured preoperatively and at each follow-up visit. Range of motion measurements included active forward elevation in the scapular plane, external rotation in adduction, and internal rotation in adduction. Abduction strength was measured with the arm at 90° abduction in the scapular plane using a handheld dynamometer. Patients were asked to estimate the value of their shoulder as a percentage of an entirely normal shoulder preoperatively and at the final follow-up.
All patients underwent radiographic and CT evaluation to compare immediate postoperative and final follow-up images. Each radiograph and CT was examined for (1) bone graft healing (absence of a lucent line observed between the humeral bone graft and native glenoid), (2) bone graft resorption or lysis, and (3) inferior scapular notching, which was graded according to the classification system of Sirveaux et al.
and the RSA angle, respectively. Minimum follow-up was 2 years.
Data normality was analyzed with the d'Agostino-Pearson test, and parametric and nonparametric tests were used as appropriate. Preoperative and postoperative scores and measurements of shoulder mobility were compared. Paired observations were compared using paired Student t tests and Wilcoxon signed rank tests, as appropriate, and unpaired observations were compared using Student t tests and the Mann-Whitney U test, as appropriate. Categoric data were compared using χ2 tests and Fisher exact tests, as appropriate, depending on cell populations. Statistical analysis was performed with MedCalc 11.0 software (MedCalc Software, Mariakerke, Belgium).
Our cohort included 54 arthritic patients (70% female) with a mean age of 73 years (range, 52-85 years) treated with an angled BIO-RSA. Thirty-one patients received an angled BIO-RSA for cuff-tear arthropathy, 13 for primary glenohumeral osteoarthritis, 6 for rheumatoid arthritis, 2 for postinstability arthropathy, and 2 for fracture sequelae. According to the Walch system,
15 shoulders were classified as E2, 21 as E3, and 3 as E4. Combined vertical and horizontal glenoid bone deficiency was present in 15 patients.
Complications and revisions
Glenoid loosening occurred in 3 patients (5%). Each of these occurred in the first 6 months after the operation. These included 1 infected loosening, 1 traumatic glenoid loosening caused by a fall 4 weeks postoperatively, and 1 glenoid loosening caused by uncorrected superior inclination. The latter 2 patients were revised with revision RSA with ICBG with good results. There was no radiographic humeral loosening or failure. No patients suffered a postoperative dislocation or subluxation.
One patient sustained a fracture of scapula spine 3 months after surgery, which was treated conservatively. Another patient sustained a pulmonary embolism treated with curative injection of heparin.
Aside from the 3 patients with glenoid loosening, no other patients showed radiographic signs of baseplate loosening such as lucency around the screws or a change in position of the baseplate.
Complete incorporation was observed in all patients other than the 3 with complications. Radiographs and CT images demonstrated union between the cancellous bone graft and the surface of the native glenoid in 94% (51 of 54) of the patients.
At final follow-up, 13 patients (25%) had grade 1 to 3 scapular notching. No patients had grade 4 inferior scapular notching.
Correction of glenoid deformity
Preoperative and postoperative CT measurements of inclination and version were performed using the multiplanar reconstruction mode to make our measurements in the plane of the scapula. The correction of glenoid orientation was significant in both planes.
The inclination and version deformity corrections achieved with the angled BIO-RSA, as measured on reformatted 2D CT-scans, are summarized in Table I.
Table ICorrection of vertical and horizontal deformity of the glenoid as measured with preoperative and postoperative 2-dimensional reformatted CT scans
In the 15 patients with combined (vertical and horizontal) glenoid bone loss, the angled BIO-RSA technique allowed simultaneous correction of both the posterior and superior defects. The correction achieved in both planes was statistically significant (Table II, Fig. 5).
Table IICorrection of glenoid orientation in patients with combined (coronal and axial) glenoid deficiencies as measured on reformatted 2-dimensional computed tomography scans
Our original surgical option, since 2006, has been to use an angled bone graft harvested from the humeral head (angled BIO-RSA) to restore the glenoid bone stock and obtain correct alignment of the implant with minimal morbidity.
The purpose of the present study was to report the results of angled BIO-RSA in addressing glenoid bone deficiency, including combined (multiplanar) glenoid bone loss. Our 3 hypotheses were verified: (1) the asymmetrical (trapezoidal) autologous bone graft harvested from the humeral head predictably heals and incorporates to the native glenoid, (2) the bone graft restores glenoid bone stock and allows correct baseplate alignment, and (3) the graft allows lateralization of the baseplate and sphere (or at least, avoids its medialization), providing good functional results, comparable to those of RSA in the absence of glenoid deficiency, and avoids complications such as instability and severe scapular notching.
Our first hypothesis is confirmed: despite the advanced age of the patients (73 years), at a mean 3 years of follow-up, 94% of the grafts incorporated to the native scapula. We observed only 3 patients with glenoid loosening: 1 after early trauma, 1 secondary to a technical mistake (persistent superior inclination), and 1 secondary to infection. Our results confirm that even in elderly patients, an autologous bone graft harvested from the humeral head reliably and predictably fuses to the native glenoid in RSA. These results are in agreement with previous reports on the use of autografts in RSA
recently demonstrated a higher rate of graft incorporation in RSA for autografts compared with allografts (86% complete or partial incorporation for autografts vs. 66% for allografts). Furthermore, compared with anatomic shoulder arthroplasty, the RSA presents a favorable environment for glenoid graft incorporation. Immediate graft fixation and compression are obtained by the combination of the long-peg baseplate and screws in the native scapula, and compression forces (after 30° of abduction) are favorable to graft healing and incorporation .
Our second hypothesis is also confirmed: the angled BIO-RSA technique allows for correction of severe glenoid deficiency, including retroversion or inclination of 25° or more, and provides the flexibility needed to reconstruct multiplanar deformity (Table I, Table II). Our study demonstrates that patients with severe Walch B2/C and/or Favard E2/E3 glenoid deformities can reliably undergo RSA with this technique. Simultaneous correction of posterior and superior glenoid defects in patients with combined (multiplanar) glenoid bone loss is undoubtedly one of the main advantages of using a cancellous bone graft. Other advantages of the angled BIO-RSA technique include minimal donor-site morbidity (compared with ICBG
). In comparison, augmented baseplates do not have this flexibility because they allow correction in only 1 plane. Only a patient-specific augmented baseplate could be used to correct a complex multiplanar deformity; however, the process to build such an implant is more complex and expensive. In addition, it does not offer the possibility to reconstruct the glenoid bone stock.
Finally, our third hypothesis is verified: functional outcomes after angled BIO-RSA are good, similar to those with standard RSA without glenoid bone deficiency.
No prosthetic dislocation was observed in the present series, and all parameters of the Constant-Murley score significantly improved (Table III). The angled BIO-RSA not only allows correction of severe glenoid erosion but also provides lateralization of the glenoid component (or at least, avoids glenoid medialization). This results in great postoperative motion, improved shoulder contour, absence of instability, and a relatively low rate of scapular notching.
Although no severe scapular notching (grade 4) was observed in the present series, we still observed a 25% rate of mild (grade 1 to 3) scapular notching. All cases of scapular notching were observed with the use of the small (36-mm) sphere and large (29-mm) baseplate (the small [25-mm] baseplate was not available at the time of this series). These results confirm that lateralization alone is not sufficient to avoid scapular notching and that a minimum of 5 mm of inferior overhang is also needed.
Nowadays, we constantly achieve this 5 mm of inferior overhang (in addition to glenoid lateralization) by matching the baseplate and sphere sizes: we use a 25-mm baseplate with the 36-mm glenosphere and the 29-mm baseplate with the 42-mm sphere. Since the recent release of the 25-mm baseplate, we have observed a dramatic decrease in the rate of scapular notching. In addition, inferior eccentric sphere and lateralized humeral implants are available today.
In our experience, preoperative planning is the key to success for bone grafting in RSA. We systematically use 2D or 3D CT images, or both, to measure glenoid bone loss and to anticipate baseplate and bone graft positioning.
In the axial plane, we use the corrected Friedman line (or plane of the scapula) for the best glenoid implant alignment: our goal is to implant the baseplate within 10° of neutral version. In the coronal plane, our goal is to implant the baseplate perpendicular to the floor of the supraspinatus fossa (ie, the peg parallel to the line of the supraspinatus fossa). As mentioned above, the RSA angle measures focal inferior glenoid inclination and thus is helpful to accurately determine the size and shape of bone graft needed to compensate for superior inclination in RSA (Fig. 2).
For simple (one plane) and moderate (<25°) glenoid bone loss, we used standardized instrumentation combined with slight eccentric reaming to achieve glenoid reconstruction and implant alignment. For instance, in a patient with a cuff tear arthritis, to compensate for a 22° glenoid superior bone loss (22° of superior inclination), we achieve 10° of correction using inferior eccentric reaming, whereas the additional correction is achieved through a standardized 12° angled bone graft (Fig. 4). Fixation of the graft is of paramount importance: a long-pegged baseplate (≥25 mm) is used to ensure that the peg contacts a minimum of 10 mm of native glenoid for immediate secure fixation and early ingrowth. Long screws (35 to 45 mm) are also necessary to fixate the baseplate to both the graft and the scapula. For more severe glenoid erosion (>25°) or complex (multiplanar) glenoid bone defects, 3D preoperative planning with CT scans and dedicated planning software (Glenosys) is useful to quantify the glenoid erosion, to plan the depth of reaming, the thickness and shape of the graft, and the position of the baseplate (Fig. 3). In these cases, a longer-pegged glenoid component (30- or 35-mm) and longer screws may be necessary to obtain graft fixation and incorporation.
This study has several limitations. This is a retrospective, single-center study without a control group, and only midterm results are reported. In addition, only a single implant was used, and this technique may not be as successful with other implants. We included a variety of indications within our cohort. This increases generalizability but may limit internal validity.
Bone graft resorption may occur with longer follow-up. RSA, however, provides rigid bone graft fixation under compressive forces with the help of the baseplate and screws, which makes us believe that there is low risk of further resorption or lysis of the autograft with extended follow-up. Most patients with clinical loosening after RSA experience this complication within the first 2 years postoperatively.
Based on our extensive experience with the standard BIO-RSA (starting in 2006), our opinion is that significant resorption or lysis will not occur with extended follow-up. A limit of this technique is that it cannot be used for revision of a previous arthroplasty or when there is necrosis or absence of the humeral head; in this case, we use allograft, ICBG, or an augmented baseplate.
Strengths of our study are that, to our knowledge, this is the largest reported series of patients with humeral autograft used to reconstruct significant glenoid deficiencies during RSA with a minimum of 2 years of follow-up. Furthermore, the methods used to assess graft incorporation and complications are robust: radiographs and CT images were both used to evaluate preoperative bone loss, postoperative graft incorporation, glenoid loosening, and scapular notching.
The angled BIO-RSA is a viable and simple solution to address severe glenoid bone loss or erosion during RSA. The humeral head is an excellent source of autograft: the graft constantly and predictably incorporates. The asymmetrical (trapezoidal) autologous bone graft, harvested from the humeral head, allows restoration of glenoid bone stock, and correction of baseplate alignment with minimal morbidity. The angled BIO-RSA technique provides the flexibility needed to reconstruct multiplanar deformity. In addition, the graft allows lateralization of the baseplate and sphere, providing good functional results, comparable to those of RSA in the absence of glenoid deficiency. Other advantages include the lack of donor-site morbidity (compared to ICBG), the absence risk of disease transmission (compared to allograft), and low cost (compared to augmented baseplate or allograft). The patient is his or her own donor, and the graft is located in situ.
Pascal Boileau and Gilles Walch receive royalties from Wright, are involved in the development of the software (Glenosys) used in the present study, and are shareholders of Imascap. All of the other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
Does bony increased-offset reverse shoulder arthroplasty decrease scapular notching?.