- Email: [email protected]
Reliability and validity assessment of a glenoid bone loss measurement using the Bernageau profile view in chronic anterior shoulder instability
J Shoulder Elbow Surg (2013) 22, 1193-1198
www.elsevier.com/locate/ymse
Reliability and validity assessment of a glenoid bone loss measurement using the Bernageau profile view in chronic anterior shoulder instability Erwan Pansard, MDa,b,*, Shahnaz Klouche, MDa, Nicolas Billot, MDa,b, Benoit Rousselin, MDa,b, Tobias Maximilien Kraus, MDa, Thomas Bauer, MDa,b, Philippe Hardy, MD, PhDa,b a b
H^ opitaux Universitaires Paris Ile-de-France Ouest, AP-HP, F-92100 Boulogne-Billancourt, France Universite de Versailles Saint-Quentin, F-78035 Versailles, France Background: One of the identified risk factors for anterior shoulder instability is bone loss on the anteriorinferior glenoid rim. The aim of our study was to assess intraobserver and interobserver reproducibility of the Bernageau view to estimate glenoid bone loss and validate this radiographic method with computed tomography (CT) scan. The second objective was to find correlation between Bernageau and arthroscopic bone loss measurements. Materials and methods: Twenty patients were included retrospectively. Two independent observers evaluated glenoid bone loss with the ratio between glenoid joint surface diameters of the pathologic and healthy shoulders on Bernageau views. Results were compared with CT (gold standard) and arthroscopic measurements. Validity and reliability of Bernageau measurement were assessed with Spearman correlation coefficients (r) and intraclass correlation coefficients (r). Results: The interobserver and intraobserver reliability and the validity of Bernageau measurement compared with the reference test, the CT scan, were all excellent, with a Spearman r between 0.56 (P ¼ .0002) and 0.95 (P < .00001) and an intraclass correlation coefficient between 0.82 (P ¼ .0007) and 0.97 (P < .105). There was no correlation with arthroscopic evaluation. Conclusion: The glenoid bone defect measurement on the Bernageau profile view is a valid and reliable method. Furthermore, it is easy to use in current clinical practice. Surgeons can therefore consider it as a tool for preoperative planning, and its use could decrease CT scan indications. Level of evidence: Level III, Nonconsecutive Patients, Diagnostic Study. Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Shoulder; instability; radiography; glenoid rim; Bernageau; bone loss
The study was approved by the institutional review board of Comite de Protection des Personnes Ile de France VIII, H^opital Ambroise Pare, Boulogne, France. It was a non-interventional study following our usual routine workup and surgical decision making for chronic shoulder instability. In France, patient consent is not required for an anonymous retrospective data analysis. We confirm that the institutional review board specifically waived the need for consent for this study.
*Reprint requests: Erwan Pansard, MD, Orthopedic Surgery Department, Ambroise Pare Hospital, 9 Avenue Charles-de-Gaulle, F-92100 Boulogne, France. E-mail address: [email protected] (E. Pansard).
1058-2746/$ - see front matter Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2012.12.032
1194 Anterior shoulder instability is a significant problem that requires surgical treatment for patients at risk for recurrent dislocation16,23 to reduce instability episodes and, at long term, the incidence of arthritis.8 Deficiencies of the anterior glenoid rim and capsule are poor prognostic factors for the Bankart repair on long-term follow-up.23,28 As a result of unsatisfactory outcomes after soft-tissue repairs5 and findings in biomechanical studies,18,29 bony augmentation procedures are recommended for patients with bone loss of greater than 20% to 25% of the glenoid surface area.6 Therefore, in clinical practice, the difficulty is to determine the amount of bone loss to decide between the different operative treatment options.9 The Bernageau glenoid profile view of the affected shoulder with the contralateral comparison view is a reproducible technique to identify glenoid bone loss with sensitivity and specificity higher than 90%.1,12 However, this technique has been given up since the introduction of 3-dimensional (3D) computed tomography (CT).9 Today, when bone loss is suspected, either CT or magnetic resonance imaging is considered to more accurately quantify bone loss. Sugaya et al27 proposed a preoperative quantification method using 3D-CT scanning. They defined the inferior part of the glenoid as a circle and correlated the size of the defect to the missing part of that circle, which was confirmed by Huysmans et al.17 Burkhart et al7 described an arthroscopic method to confirm preoperative findings using the glenoid bare spot as a reference.9 The distance from the posterior edge of the glenoid to the bare spot is compared with the measured distance from the bare spot to the anterior glenoid margin.7 Although the Bernageau glenoid profile view is used more rarely now since the introduction of 3D-CT,9 the risk of patient irradiation is higher with a scanner. We conducted a retrospective study to assess Bernageau measurements and to compare them with the 2 other techniques cited earlier. Our hypothesis is that with the Bernageau view of the affected shoulder and contralateral comparison, it is possible to quantify glenoid bone loss in patients with anterior shoulder instability, which might be an easier and more quickly available tool.
Materials and methods This retrospective 1-center study was conducted between November 2006 and April 2012. The criteria for inclusion were isolated anterior shoulder instability without previous surgery, requiring arthroscopic management, and provision of a radiology file containing the bilateral Bernageau glenoid profile view and a CT scan with 3D reconstruction. The exclusion criteria were concomitant rotator cuff lesions, bilateral instability, previous surgery of the affected shoulder, and multidirectional instability. High-risk sports, shoulder-specific activities, ligamentous hyperlaxity, or bony Bankart lesions were not exclusion criteria. The diagnosis of dynamic shoulder instability was confirmed preoperatively by physical examination.14 An apprehension sign and
E. Pansard et al.
Figure 1
Bernageau glenoid profile view.
painful subluxation with symptoms of apprehension were used to distinguish the direction of instability. The main aim of this study was to assess intraobserver and interobserver reproducibility of the Bernageau view used to estimate glenoid bone loss and validate this radiographic method with CT scan considered the gold standard. The second objective was to find correlation between Bernageau and arthroscopic bone loss measurements. Bilateral Bernageau glenoid profile views were obtained with the patient in an upright posture and the arm in abduction (Fig. 1).1 An exact glenoid profile requires that the line projecting the anterior edge of the superior part of the cavity is continuous with the anterior line of the scapula (Figs. 2 and 3). From this view, the ratio between the glenoid joint surface diameters of the pathologic shoulder (D1) and healthy shoulder (D2) (D1/D2 ratio) was calculated to estimate glenoid bone loss according to the following formula (Fig. 4): [1 – (D1/D2)] 100. Two orthopaedic trauma surgeons with similar experience independently viewed the Bernageau profile radiographs on 2 separate occasions at a mean of 4 weeks apart. For the second evaluation, the radiographs were blinded and viewed in a different sequence. Each surgeon was blinded to the other’s measurements. The radiographs were anonymized. Both observers were blinded to the clinical and CT scan findings. An independent biostatistician ensured the standardization of the study protocol. Second, the preoperative CT scan evaluation with 3D-CT reconstruction was performed to evaluate glenoid bone loss according to the method of Sugaya et al.27 On the en face 3dimensionally reconstructed CT image, the inferior portion of the glenoid contour can be approximated to a true circle. The size of the osseous defect, as a percentage of the glenoid rim, was calculated as the ratio of the area of glenoid bone loss, which is traced manually, to the area of an assumed outer fitting circle based on the inferior portion of the glenoid. This was determined as B/A 100% (where B is the area of glenoid bone loss and A is the area of an assumed outer fitting circle based on the inferior portion of the glenoid). The assessment was performed with an open access version of OsiriX image processing software (version 3.8; Pixmeo, Geneva, Switzerland). One observer supported by a well-experienced radiologist in musculoskeletal imaging
Glenoid bone loss and Bernageau view
1195
Figure 3 view.
Radiologic criteria for exact Bernageau glenoid profile
coefficient (ICC) (r) were used to assess the intraobserver and interobserver reliability of the Bernageau measurement and its validity (ie, its accuracy) compared with CT scan measurements. The strength of the relationship was classified as strong (r > 0.5), medium (0.3 < r < 0.5), small (0.1 < r < 0.3), or none (r < 0.1).10 The strength of the agreement was classified as excellent (r > 0.75), good (0.40 < r < 0.75), or poor (r < 0.40).13 All correlations were tested for statistical significance by use of the P value. A P value of less than .05 was considered to be significant.
Results Figure 2
Criteria for exact Bernageau glenoid profile view.
acquired each dataset. The observer and radiologist were blinded to the results of the other measurements. Third, multi-operator perioperative measurements were performed with patients under general anesthesia in the lateral decubitus position under double traction. Operators were blinded to the previous results and were independent. Measurements were made by use of the bare spot after tissue release. Two approaches were used, anterior for the arthroscope and posterior for the probe.7 The bare spot was located. A graduated probe with 5-mm calibrated marks was placed through the posterior portal across the glenoid so that its tip rested on the bare spot. The distance from the center of the bare spot to the posterior glenoid rim was then measured. The probe was next used to measure the distance from the anterior glenoid rim to the center of the bare spot. The ratio between the measured inferior glenoid diameter and the theoretical inferior glenoid diameter was calculated to estimate glenoid bone loss.
Statistical analysis Statistical analyses were performed with Stata/IC (version 10.0; StataCorp, College Station, TX, USA). The Spearman rank correlation coefficient (r) and the intraclass correlation
The study included 20 patients (12 men and 8 women). Patients had a mean age of 29.5 8.8 years, and 12 right and 8 left shoulders were included. The mean measurement of glenoid bone loss was 9.0% 5.9% on 3D-CT, 8.5% 4.5% for the D1/D2 ratio, and 7.9% 7.2% with the bare spot. For the intraobserver reliability of the Bernageau measurement, the test-retest correlation was strong and significant, with a Spearman r equal to 0.84 and 0.95 (P < .105). Physicians’ agreement between their initial scoring and rerating of the same patient showed excellent ICCs of 0.92 and 0.97 (P < .105) (Table I). The interobserver reliability of the Bernageau measurement was strong (Spearman r between 0.76 [P ¼ .0002] and 0.87 [P < .105]). The agreement among physicians’ ratings of patients had excellent ICCs between 0.89 and 0.94 (P < .105) (Table II). For the validity of the Bernageau measurement compared with the reference test (ie, CT scan), the correlation was strong (Spearman r between 0.56 [P ¼ .0002] and 0.8 [P < .105]) and the ICCs were excellent, between 0.82 (P ¼ .0007) and 0.86 (P ¼ .0001) (Table III). Compared with the arthroscopic evaluation, the correlation was not significant because the Spearman coefficient varied between 0.4 (P ¼ .08) and 0.6 (P ¼ .006).
1196
E. Pansard et al.
Figure 4 Bilateral Bernageau glenoid profile view with glenoid joint surface diameter measurements of pathologic shoulder (D1) and healthy shoulder (D2) used to calculate glenoid bone loss as follows: [1 – (D1/D2)] 100.
Table I
Results of glenoid bone loss estimations by 2 independent observers based on D1/D2 ratio and intraobserver agreement Evaluation 1 (%)
Evaluation 2 (%)
Spearman
ICC
Intraobserver agreement Observer 1 Observer 2
8.5 4.3 8.3 5.2
8.1 4.6 9.1 4.9
0.84 (strong) 0.95 (strong)
Table II Interobserver agreement for D1/D2 ratio among different observers’ means Observer 1 Evaluation 1 Observer 2 Evaluation 1 r ¼ 0.81 (strong), P < .105 r ¼ 0.90 (excellent), P < .105 Evaluation 2 r ¼ 0.87 (strong), P < .105 r ¼ 0.94 (excellent), P < .105
Evaluation 2 r ¼ 0.76 (strong), P ¼ .0002 r ¼ 0.89 (excellent), P < .105 r ¼ 0.86 (strong), P < .105 r ¼ 0.92 (excellent), P < .105
Discussion The most important finding of this study was that the glenoid bone defect measurement on the Bernageau profile view was a valid and reliable method compared with CT scan measurement. Anterior bone loss of the glenoid superior to 25% can lead to a high risk of failure of the Bankart arthroscopic
P value 5
<10 <105
Intraobserver agreement
P value
0.92 (excellent) 0.97 (excellent)
<105 <105
procedure.3,6 This bone loss decreases the glenoid’s arc length and reduces its concavity,2,5,6,18 so for such patients, bony augmentation procedures are recommended. These bony lesions must be diagnosed and quantified for the selection of the surgical technique, and preoperative imaging usually includes magnetic resonance imaging or CT,19 which is the most accurate radiologic evaluation method for the cortical glenoid rim.25 Burkhart et al6,7 chose the surgical procedure after arthroscopic measurement using the bare spot as a marker for the center of the inferior glenoid. In the study of Chuang et al,9 3D-CT scan accurately predicted the requirement for a bone grafting procedure for 96% of patients. The circle method described by Sugaya et al27 is based on the observation that the inferior part of the glenoid has the shape of a true circle, which can be drawn on the sagittal en face view of the glenoid by use of 3D-CT images. Huysmans et al17 validated the method in a cadaveric study, showing very good intraobserver and interobserver reproducibility because they found similar results. Griffith et al15 proposed to compare the maximum glenoid width with the contralateral side, and Chuang et al9 calculated a glenoid index. These methods required CT
Glenoid bone loss and Bernageau view Table III
1197
Validity assessment of Bernageau profile compared with 3D-CT Observer 1
Spearman correlation coefficient ICC
Observer 2
Evaluation 1
Evaluation 2
Evaluation 1
Evaluation 2
r ¼ 0.64 (strong), P < .105 r ¼ 0.86 (excellent), P ¼ .0001
r ¼ 0.62 (strong), P ¼ .0049 r ¼ 0.84 (excellent), P ¼ .0003
r ¼ 0.80 (strong), P < 105 r ¼ 0.88 (excellent), P < 105
r ¼ 0.56 (strong), P ¼ .0002 r ¼ 0.82 (excellent), P ¼ .0007
scans of the contralateral shoulder and increased the patient’s irradiation. Magarelli et al22 used the circle method on 2-dimensional CT images to avoid bias from 3D-CT reconstructions9 but have not used this method without comparison to the healthy shoulder. Our 3D-CT results showed glenoid bone loss in 100% of our patients. This result is higher than that of Sugaya et al (90%) but can be explained by our small number of patients. Moreover, Sugaya et al found 40% of changes to the anterior glenoid rim to be consistent with compression. Some authors have used the bare-spot arthroscopic measurement technique described by Burkhart et al7 as a reference technique.9,15,21 This method, which considers the bare spot as the center of the inferior glenoid rim,7 is well accepted because compressive bone loss eliminates the possibility of direct measurement of glenoid bone loss due to disappearance of the bone fragment, which is compressed and not viewable anymore. Kralinger et al20 rejected this hypothesis in their anatomic study, but the mean age of their specimens was 81.6 years, which can introduce measurement artifact.9 Moreover, at this age, the bare spot may be difficult to find,11 whereas instability problems most frequently concern young, healthy and athletic patients.4,9 Corresponding to the view of Huysmans et al,17 we believe that the bare spot is a good reference point to judge bone loss perioperatively. However, a defect margin at a 45 angle to the longitudinal axis of the glenoid can lead to incorrect measurements with this technique. It may be better to evaluate bone loss parallel to the longitudinal glenoid axis.23,24 In 1976, Bernageau et al1 recommended a comparative ‘‘glenoid profile’’ view to assess bone defects. This view provides a glenoid profile with an image of the glenoid containing the sharp outline of a triangle corresponding to the anteroinferior part of the glenoid. Edwards et al12 proposed a classification into 3 groups (fracture, ‘‘cliff’’ sign, and ‘‘blunted angle’’ sign) but did not quantify glenoid bone loss. Because radiographs remain the first examination in most diagnostic algorithms,19 the Bernageau view should be considered before one performs further radiologic examinations. However, a limitation of this technique is that the contralateral shoulder needs to be included in the clinical and radiologic examinations. Furthermore, this shoulder must be healthy,12 and an inferior glenoid fracture fragment is not visible on the Bernageau view. However, because this view is usually sampled
with other radiographs, such as the anteroposterior view, it is unlikely to oversee a bony Bankart lesion.12 Our results including the 2 measurements of observers 1 and 2 found an incidence of glenoid bone loss of 98%, showing a higher prevalence than other studies, such as those of Bernageau et al (90%) or Edwards et al (90%). Mean glenoid bone loss was 8.9% (range, 4.7%-19.9%). Sommaire et al,26 using the same measurement, found a mean of less than 5% in 77 patients. The difference between the CT scan results (100% incidence of glenoid bone loss) and Bernageau view results (98% incidence of glenoid bone loss) could probably be explained by the better assessment of the glenoid’s inferior width with the CT scan. Similar to our 3D-CT scan results, these high results could be explained by the number of patients included in our study. Statistical results of glenoid bone defect measurement on the Bernageau profile view showed excellent intraobserver and interobserver reliability, and compared with CT scan, the reliability was also excellent. Our study had several limitations. First, it was a retrospective work with a small number of patients. Second, there was no correlation found between Bernageau view and arthroscopic evaluation of glenoid bone loss, maybe because 2 experimental shoulder surgeons described this technique6 and our measurements were performed by several surgeons with different experience levels (advanced to expert shoulder surgeons). Burkhart et al7 used a 3-mm probe with an anterosuperior viewing portal. We used a 5-mm probe with an anterior viewing portal. Third, we used an en face view for the 3D-CT measurements. The acquisition of this view is operator dependent and could introduce human error.9 However, 3D-CT reconstruction provides a far better assessment of the glenoid’s concavity and particularly the inferior width than a Bernageau view. Finally, radiologic measurements were performed manually, which could have caused bias.
Conclusion The glenoid bone defect measurement on the Bernageau profile view is a valid and reliable method. Furthermore, it is easy to use in current clinical practice. Surgeons can therefore consider it as a tool for preoperative planning, and its use could decrease CT scan indications.
1198
Acknowledgments We thank Dr Mark Ferguson, Rosebank Clinic, Center for Sports Medicine and Orthopaedics, Johannesburg, South Africa, for his assistance reviewing the manuscript.
Disclaimer The 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.
References 1. Bernageau J, Patte D, Debeyre J, Ferrane J. Value of the glenoid profil in recurrent luxations of the shoulder [in French]. Rev Chir Orthop Reparatrice Appar Mot 1976;62:142-7. 2. Bigliani LU, Newton PM, Steinmann SP, Connor PM, McLiveen SJ. Glenoid rim lesions associated with recurrent anterior dislocation of the shoulder. Am J Sports Med 1998;26:41-5. 3. Boileau P, Villalba M, Hery JY, Balg F, Ahrens P, Neyton L. Risk factors for recurrence of shoulder instability after arthroscopic Bankart repair. J Bone Joint Surg Am 2006;88:1755-63. http://dx.doi.org/10. 2106/JBJS.E.00817 4. Burkhart SS. The bare spot of the glenoid. Arthroscopy 2007;23:449. author reply 49-50. http://dx.doi.org/10.1016/j.arthro.2006.12.018 5. Burkhart SS, Danaceau SM. Articular arc length mismatch as a cause of failed Bankart repair. Arthroscopy 2000;16:740-4. 6. Burkhart SS, De Beer JF. Traumatic glenohumeral bone defects and their relationship to failure of arthroscopic Bankart repairs: significance of the inverted-pear glenoid and the humeral engaging HillSachs lesion. Arthroscopy 2000;16:677-94. 7. Burkhart SS, Debeer JF, Tehrany AM, Parten PM. Quantifying glenoid bone loss arthroscopically in shoulder instability. Arthroscopy 2002; 18:488-91. http://dx.doi.org/10.1053/jars.2002.32212 8. Cheung EV, Sperling JW, Hattrup SJ, Cofield RH. Long-term outcome of anterior stabilization of the shoulder. J Shoulder Elbow Surg 2008; 17:265-70. http://dx.doi.org/10.1016/j.jse.2007.06.005 9. Chuang TY, Adams CR, Burkhart SS. Use of preoperative threedimensional computed tomography to quantify glenoid bone loss in shoulder instability. Arthroscopy 2008;24:376-82. http://dx.doi.org/10. 1016/j.arthro.2007.10.008 10. Cohen J. Statistical power analysis for the behavioral sciences. Hillsdale (NJ); Lawrence Erlbaum Associates; 1988. 11. De Wilde LF, Berghs BM, Audenaert E, Sys G, Van Maele GO, Barbaix E. About the variability of the shape of the glenoid cavity. Surg Radiol Anat 2004;26:54-9. http://dx.doi.org/10.1007/s00276003-0167-1 12. Edwards TB, Boulahia A, Walch G. Radiographic analysis of bone defects in chronic anterior shoulder instability. Arthroscopy 2003;19: 732-9. http://dx.doi.org/10.1016/S0749-8063(03)00684-4 13. Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. Hoboken (NJ); John Wiley; 2003.
E. Pansard et al. 14. Gerber C, Nyffeler RW. Classification of glenohumeral joint instability. Clin Orthop Relat Res 2002;400:65-76. http://dx.doi.org/10. 1097/00003086-200207000-00009 15. Griffith JF, Antonio GE, Tong CW, Ming CK. Anterior shoulder dislocation: quantification of glenoid bone loss with CT. AJR Am J Roentgenol 2003;180:1423-30. 16. Hovelius LK, Sandstrom BC, Rosmark DL, Saebo M, Sundgren KH, Malmqvist BG. Long-term results with the Bankart and BristowLatarjet procedures: recurrent shoulder instability and arthropathy. J Shoulder Elbow Surg 2001;10:445-52. 17. Huysmans PE, Haen PS, Kidd M, Dhert WJ, Willems JW. The shape of the inferior part of the glenoid: a cadaveric study. J Shoulder Elbow Surg 2006;15:759-63. http://dx.doi.org/10.1016/j.jse.2005.09.001 18. Itoi E, Lee SB, Berglund LJ, Berge LL, An KN. The effect of a glenoid defect on anteroinferior stability of the shoulder after Bankart repair: a cadaveric study. J Bone Joint Surg Am 2000;82:35-46. 19. Jankauskas L, Rudiger HA, Pfirrmann CW, Jost B, Gerber C. Loss of the sclerotic line of the glenoid on anteroposterior radiographs of the shoulder: a diagnostic sign for an osseous defect of the anterior glenoid rim. J Shoulder Elbow Surg 2010;19:151-6. http://dx.doi.org/10. 1016/j.jse.2009.04.013 20. Kralinger F, Aigner F, Longato S, Rieger M, Wambacher M. Is the bare spot a consistent landmark for shoulder arthroscopy? A study of 20 embalmed glenoids with 3-dimensional computed tomographic reconstruction. Arthroscopy 2006;22:428-32. http://dx.doi.org/10. 1016/j.arthro.2005.12.006 21. Larrain MV, Montenegro HJ, Mauas DM, Collazo CC, Pavon F. Arthroscopic management of traumatic anterior shoulder instability in collision athletes: analysis of 204 cases with a 4- to 9-year follow-up and results with the suture anchor technique. Arthroscopy 2006;22: 1283-9. http://dx.doi.org/10.1016/j.arthro.2006.07.052 22. Magarelli N, Milano G, Baudi P, Santagada DA, Righi P, Spina V, et al. Comparison between 2D and 3D computed tomography evaluation of glenoid bone defect in unilateral anterior gleno-humeral instability. Radiol Med 2012;117:102-11. http://dx.doi.org/10.1007/ s11547-011-0712-7 23. Nofsinger C, Browning B, Burkhart SS, Pedowitz RA. Objective preoperative measurement of anterior glenoid bone loss: a pilot study of a computer-based method using unilateral 3-dimensional computed tomography. Arthroscopy 2011;27:322-9. http://dx.doi.org/10.1016/j. arthro.2010.09.007 24. Provencher MT, Detterline AJ, Ghodadra N, Romeo AA, Bach BR Jr, Cole BJ, et al. Measurement of glenoid bone loss: a comparison of measurement error between 45 degrees and 0 degrees bone loss models and with different posterior arthroscopy portal locations. Am J Sports Med 2008;36:1132-8. http://dx.doi.org/10.1177/0363546508316041 25. Roger B, Skaf A, Hooper AW, Lektrakul N, Yeh L, Resnick D. Imaging findings in the dominant shoulder of throwing athletes: comparison of radiography, arthrography, CT arthrography, and MR arthrography with arthroscopic correlation. AJR Am J Roentgenol 1999;172:1371-80. 26. Sommaire C, Penz C, Clavert P, Klouche S, Hardy P, Kempf JF. Recurrence after arthroscopic Bankart repair: Is quantitative radiological analysis of bone loss of any predictive value? Orthop Traumatol Surg Res 2012;98:514-9. http://dx.doi.org/10.1016/j.otsr.2012.03.015 27. Sugaya H, Moriishi J, Dohi M, Kon Y, Tsuchiya A. Glenoid rim morphology in recurrent anterior glenohumeral instability. J Bone Joint Surg Am 2003;85:878-84. 28. Thomazeau H, Courage O, Barth J, Pelegri C, Charousset C, Lespagnol F, et al. Can we improve the indication for Bankart arthroscopic repair? A preliminary clinical study using the ISIS score. Orthop Traumatol Surg Res 2010;96:S77-83. http://dx.doi.org/10. 1016/j.otsr.2010.09.007 29. Yamamoto N, Itoi E, Abe H, Kikuchi K, Seki N, Minagawa H, et al. Effect of an anterior glenoid defect on anterior shoulder stability: a cadaveric study. Am J Sports Med 2009;37:949-54. http://dx.doi.org/ 10.1177/0363546508330139
www.elsevier.com/locate/ymse
Reliability and validity assessment of a glenoid bone loss measurement using the Bernageau profile view in chronic anterior shoulder instability Erwan Pansard, MDa,b,*, Shahnaz Klouche, MDa, Nicolas Billot, MDa,b, Benoit Rousselin, MDa,b, Tobias Maximilien Kraus, MDa, Thomas Bauer, MDa,b, Philippe Hardy, MD, PhDa,b a b
H^ opitaux Universitaires Paris Ile-de-France Ouest, AP-HP, F-92100 Boulogne-Billancourt, France Universite de Versailles Saint-Quentin, F-78035 Versailles, France Background: One of the identified risk factors for anterior shoulder instability is bone loss on the anteriorinferior glenoid rim. The aim of our study was to assess intraobserver and interobserver reproducibility of the Bernageau view to estimate glenoid bone loss and validate this radiographic method with computed tomography (CT) scan. The second objective was to find correlation between Bernageau and arthroscopic bone loss measurements. Materials and methods: Twenty patients were included retrospectively. Two independent observers evaluated glenoid bone loss with the ratio between glenoid joint surface diameters of the pathologic and healthy shoulders on Bernageau views. Results were compared with CT (gold standard) and arthroscopic measurements. Validity and reliability of Bernageau measurement were assessed with Spearman correlation coefficients (r) and intraclass correlation coefficients (r). Results: The interobserver and intraobserver reliability and the validity of Bernageau measurement compared with the reference test, the CT scan, were all excellent, with a Spearman r between 0.56 (P ¼ .0002) and 0.95 (P < .00001) and an intraclass correlation coefficient between 0.82 (P ¼ .0007) and 0.97 (P < .105). There was no correlation with arthroscopic evaluation. Conclusion: The glenoid bone defect measurement on the Bernageau profile view is a valid and reliable method. Furthermore, it is easy to use in current clinical practice. Surgeons can therefore consider it as a tool for preoperative planning, and its use could decrease CT scan indications. Level of evidence: Level III, Nonconsecutive Patients, Diagnostic Study. Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Shoulder; instability; radiography; glenoid rim; Bernageau; bone loss
The study was approved by the institutional review board of Comite de Protection des Personnes Ile de France VIII, H^opital Ambroise Pare, Boulogne, France. It was a non-interventional study following our usual routine workup and surgical decision making for chronic shoulder instability. In France, patient consent is not required for an anonymous retrospective data analysis. We confirm that the institutional review board specifically waived the need for consent for this study.
*Reprint requests: Erwan Pansard, MD, Orthopedic Surgery Department, Ambroise Pare Hospital, 9 Avenue Charles-de-Gaulle, F-92100 Boulogne, France. E-mail address: [email protected] (E. Pansard).
1058-2746/$ - see front matter Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2012.12.032
1194 Anterior shoulder instability is a significant problem that requires surgical treatment for patients at risk for recurrent dislocation16,23 to reduce instability episodes and, at long term, the incidence of arthritis.8 Deficiencies of the anterior glenoid rim and capsule are poor prognostic factors for the Bankart repair on long-term follow-up.23,28 As a result of unsatisfactory outcomes after soft-tissue repairs5 and findings in biomechanical studies,18,29 bony augmentation procedures are recommended for patients with bone loss of greater than 20% to 25% of the glenoid surface area.6 Therefore, in clinical practice, the difficulty is to determine the amount of bone loss to decide between the different operative treatment options.9 The Bernageau glenoid profile view of the affected shoulder with the contralateral comparison view is a reproducible technique to identify glenoid bone loss with sensitivity and specificity higher than 90%.1,12 However, this technique has been given up since the introduction of 3-dimensional (3D) computed tomography (CT).9 Today, when bone loss is suspected, either CT or magnetic resonance imaging is considered to more accurately quantify bone loss. Sugaya et al27 proposed a preoperative quantification method using 3D-CT scanning. They defined the inferior part of the glenoid as a circle and correlated the size of the defect to the missing part of that circle, which was confirmed by Huysmans et al.17 Burkhart et al7 described an arthroscopic method to confirm preoperative findings using the glenoid bare spot as a reference.9 The distance from the posterior edge of the glenoid to the bare spot is compared with the measured distance from the bare spot to the anterior glenoid margin.7 Although the Bernageau glenoid profile view is used more rarely now since the introduction of 3D-CT,9 the risk of patient irradiation is higher with a scanner. We conducted a retrospective study to assess Bernageau measurements and to compare them with the 2 other techniques cited earlier. Our hypothesis is that with the Bernageau view of the affected shoulder and contralateral comparison, it is possible to quantify glenoid bone loss in patients with anterior shoulder instability, which might be an easier and more quickly available tool.
Materials and methods This retrospective 1-center study was conducted between November 2006 and April 2012. The criteria for inclusion were isolated anterior shoulder instability without previous surgery, requiring arthroscopic management, and provision of a radiology file containing the bilateral Bernageau glenoid profile view and a CT scan with 3D reconstruction. The exclusion criteria were concomitant rotator cuff lesions, bilateral instability, previous surgery of the affected shoulder, and multidirectional instability. High-risk sports, shoulder-specific activities, ligamentous hyperlaxity, or bony Bankart lesions were not exclusion criteria. The diagnosis of dynamic shoulder instability was confirmed preoperatively by physical examination.14 An apprehension sign and
E. Pansard et al.
Figure 1
Bernageau glenoid profile view.
painful subluxation with symptoms of apprehension were used to distinguish the direction of instability. The main aim of this study was to assess intraobserver and interobserver reproducibility of the Bernageau view used to estimate glenoid bone loss and validate this radiographic method with CT scan considered the gold standard. The second objective was to find correlation between Bernageau and arthroscopic bone loss measurements. Bilateral Bernageau glenoid profile views were obtained with the patient in an upright posture and the arm in abduction (Fig. 1).1 An exact glenoid profile requires that the line projecting the anterior edge of the superior part of the cavity is continuous with the anterior line of the scapula (Figs. 2 and 3). From this view, the ratio between the glenoid joint surface diameters of the pathologic shoulder (D1) and healthy shoulder (D2) (D1/D2 ratio) was calculated to estimate glenoid bone loss according to the following formula (Fig. 4): [1 – (D1/D2)] 100. Two orthopaedic trauma surgeons with similar experience independently viewed the Bernageau profile radiographs on 2 separate occasions at a mean of 4 weeks apart. For the second evaluation, the radiographs were blinded and viewed in a different sequence. Each surgeon was blinded to the other’s measurements. The radiographs were anonymized. Both observers were blinded to the clinical and CT scan findings. An independent biostatistician ensured the standardization of the study protocol. Second, the preoperative CT scan evaluation with 3D-CT reconstruction was performed to evaluate glenoid bone loss according to the method of Sugaya et al.27 On the en face 3dimensionally reconstructed CT image, the inferior portion of the glenoid contour can be approximated to a true circle. The size of the osseous defect, as a percentage of the glenoid rim, was calculated as the ratio of the area of glenoid bone loss, which is traced manually, to the area of an assumed outer fitting circle based on the inferior portion of the glenoid. This was determined as B/A 100% (where B is the area of glenoid bone loss and A is the area of an assumed outer fitting circle based on the inferior portion of the glenoid). The assessment was performed with an open access version of OsiriX image processing software (version 3.8; Pixmeo, Geneva, Switzerland). One observer supported by a well-experienced radiologist in musculoskeletal imaging
Glenoid bone loss and Bernageau view
1195
Figure 3 view.
Radiologic criteria for exact Bernageau glenoid profile
coefficient (ICC) (r) were used to assess the intraobserver and interobserver reliability of the Bernageau measurement and its validity (ie, its accuracy) compared with CT scan measurements. The strength of the relationship was classified as strong (r > 0.5), medium (0.3 < r < 0.5), small (0.1 < r < 0.3), or none (r < 0.1).10 The strength of the agreement was classified as excellent (r > 0.75), good (0.40 < r < 0.75), or poor (r < 0.40).13 All correlations were tested for statistical significance by use of the P value. A P value of less than .05 was considered to be significant.
Results Figure 2
Criteria for exact Bernageau glenoid profile view.
acquired each dataset. The observer and radiologist were blinded to the results of the other measurements. Third, multi-operator perioperative measurements were performed with patients under general anesthesia in the lateral decubitus position under double traction. Operators were blinded to the previous results and were independent. Measurements were made by use of the bare spot after tissue release. Two approaches were used, anterior for the arthroscope and posterior for the probe.7 The bare spot was located. A graduated probe with 5-mm calibrated marks was placed through the posterior portal across the glenoid so that its tip rested on the bare spot. The distance from the center of the bare spot to the posterior glenoid rim was then measured. The probe was next used to measure the distance from the anterior glenoid rim to the center of the bare spot. The ratio between the measured inferior glenoid diameter and the theoretical inferior glenoid diameter was calculated to estimate glenoid bone loss.
Statistical analysis Statistical analyses were performed with Stata/IC (version 10.0; StataCorp, College Station, TX, USA). The Spearman rank correlation coefficient (r) and the intraclass correlation
The study included 20 patients (12 men and 8 women). Patients had a mean age of 29.5 8.8 years, and 12 right and 8 left shoulders were included. The mean measurement of glenoid bone loss was 9.0% 5.9% on 3D-CT, 8.5% 4.5% for the D1/D2 ratio, and 7.9% 7.2% with the bare spot. For the intraobserver reliability of the Bernageau measurement, the test-retest correlation was strong and significant, with a Spearman r equal to 0.84 and 0.95 (P < .105). Physicians’ agreement between their initial scoring and rerating of the same patient showed excellent ICCs of 0.92 and 0.97 (P < .105) (Table I). The interobserver reliability of the Bernageau measurement was strong (Spearman r between 0.76 [P ¼ .0002] and 0.87 [P < .105]). The agreement among physicians’ ratings of patients had excellent ICCs between 0.89 and 0.94 (P < .105) (Table II). For the validity of the Bernageau measurement compared with the reference test (ie, CT scan), the correlation was strong (Spearman r between 0.56 [P ¼ .0002] and 0.8 [P < .105]) and the ICCs were excellent, between 0.82 (P ¼ .0007) and 0.86 (P ¼ .0001) (Table III). Compared with the arthroscopic evaluation, the correlation was not significant because the Spearman coefficient varied between 0.4 (P ¼ .08) and 0.6 (P ¼ .006).
1196
E. Pansard et al.
Figure 4 Bilateral Bernageau glenoid profile view with glenoid joint surface diameter measurements of pathologic shoulder (D1) and healthy shoulder (D2) used to calculate glenoid bone loss as follows: [1 – (D1/D2)] 100.
Table I
Results of glenoid bone loss estimations by 2 independent observers based on D1/D2 ratio and intraobserver agreement Evaluation 1 (%)
Evaluation 2 (%)
Spearman
ICC
Intraobserver agreement Observer 1 Observer 2
8.5 4.3 8.3 5.2
8.1 4.6 9.1 4.9
0.84 (strong) 0.95 (strong)
Table II Interobserver agreement for D1/D2 ratio among different observers’ means Observer 1 Evaluation 1 Observer 2 Evaluation 1 r ¼ 0.81 (strong), P < .105 r ¼ 0.90 (excellent), P < .105 Evaluation 2 r ¼ 0.87 (strong), P < .105 r ¼ 0.94 (excellent), P < .105
Evaluation 2 r ¼ 0.76 (strong), P ¼ .0002 r ¼ 0.89 (excellent), P < .105 r ¼ 0.86 (strong), P < .105 r ¼ 0.92 (excellent), P < .105
Discussion The most important finding of this study was that the glenoid bone defect measurement on the Bernageau profile view was a valid and reliable method compared with CT scan measurement. Anterior bone loss of the glenoid superior to 25% can lead to a high risk of failure of the Bankart arthroscopic
P value 5
<10 <105
Intraobserver agreement
P value
0.92 (excellent) 0.97 (excellent)
<105 <105
procedure.3,6 This bone loss decreases the glenoid’s arc length and reduces its concavity,2,5,6,18 so for such patients, bony augmentation procedures are recommended. These bony lesions must be diagnosed and quantified for the selection of the surgical technique, and preoperative imaging usually includes magnetic resonance imaging or CT,19 which is the most accurate radiologic evaluation method for the cortical glenoid rim.25 Burkhart et al6,7 chose the surgical procedure after arthroscopic measurement using the bare spot as a marker for the center of the inferior glenoid. In the study of Chuang et al,9 3D-CT scan accurately predicted the requirement for a bone grafting procedure for 96% of patients. The circle method described by Sugaya et al27 is based on the observation that the inferior part of the glenoid has the shape of a true circle, which can be drawn on the sagittal en face view of the glenoid by use of 3D-CT images. Huysmans et al17 validated the method in a cadaveric study, showing very good intraobserver and interobserver reproducibility because they found similar results. Griffith et al15 proposed to compare the maximum glenoid width with the contralateral side, and Chuang et al9 calculated a glenoid index. These methods required CT
Glenoid bone loss and Bernageau view Table III
1197
Validity assessment of Bernageau profile compared with 3D-CT Observer 1
Spearman correlation coefficient ICC
Observer 2
Evaluation 1
Evaluation 2
Evaluation 1
Evaluation 2
r ¼ 0.64 (strong), P < .105 r ¼ 0.86 (excellent), P ¼ .0001
r ¼ 0.62 (strong), P ¼ .0049 r ¼ 0.84 (excellent), P ¼ .0003
r ¼ 0.80 (strong), P < 105 r ¼ 0.88 (excellent), P < 105
r ¼ 0.56 (strong), P ¼ .0002 r ¼ 0.82 (excellent), P ¼ .0007
scans of the contralateral shoulder and increased the patient’s irradiation. Magarelli et al22 used the circle method on 2-dimensional CT images to avoid bias from 3D-CT reconstructions9 but have not used this method without comparison to the healthy shoulder. Our 3D-CT results showed glenoid bone loss in 100% of our patients. This result is higher than that of Sugaya et al (90%) but can be explained by our small number of patients. Moreover, Sugaya et al found 40% of changes to the anterior glenoid rim to be consistent with compression. Some authors have used the bare-spot arthroscopic measurement technique described by Burkhart et al7 as a reference technique.9,15,21 This method, which considers the bare spot as the center of the inferior glenoid rim,7 is well accepted because compressive bone loss eliminates the possibility of direct measurement of glenoid bone loss due to disappearance of the bone fragment, which is compressed and not viewable anymore. Kralinger et al20 rejected this hypothesis in their anatomic study, but the mean age of their specimens was 81.6 years, which can introduce measurement artifact.9 Moreover, at this age, the bare spot may be difficult to find,11 whereas instability problems most frequently concern young, healthy and athletic patients.4,9 Corresponding to the view of Huysmans et al,17 we believe that the bare spot is a good reference point to judge bone loss perioperatively. However, a defect margin at a 45 angle to the longitudinal axis of the glenoid can lead to incorrect measurements with this technique. It may be better to evaluate bone loss parallel to the longitudinal glenoid axis.23,24 In 1976, Bernageau et al1 recommended a comparative ‘‘glenoid profile’’ view to assess bone defects. This view provides a glenoid profile with an image of the glenoid containing the sharp outline of a triangle corresponding to the anteroinferior part of the glenoid. Edwards et al12 proposed a classification into 3 groups (fracture, ‘‘cliff’’ sign, and ‘‘blunted angle’’ sign) but did not quantify glenoid bone loss. Because radiographs remain the first examination in most diagnostic algorithms,19 the Bernageau view should be considered before one performs further radiologic examinations. However, a limitation of this technique is that the contralateral shoulder needs to be included in the clinical and radiologic examinations. Furthermore, this shoulder must be healthy,12 and an inferior glenoid fracture fragment is not visible on the Bernageau view. However, because this view is usually sampled
with other radiographs, such as the anteroposterior view, it is unlikely to oversee a bony Bankart lesion.12 Our results including the 2 measurements of observers 1 and 2 found an incidence of glenoid bone loss of 98%, showing a higher prevalence than other studies, such as those of Bernageau et al (90%) or Edwards et al (90%). Mean glenoid bone loss was 8.9% (range, 4.7%-19.9%). Sommaire et al,26 using the same measurement, found a mean of less than 5% in 77 patients. The difference between the CT scan results (100% incidence of glenoid bone loss) and Bernageau view results (98% incidence of glenoid bone loss) could probably be explained by the better assessment of the glenoid’s inferior width with the CT scan. Similar to our 3D-CT scan results, these high results could be explained by the number of patients included in our study. Statistical results of glenoid bone defect measurement on the Bernageau profile view showed excellent intraobserver and interobserver reliability, and compared with CT scan, the reliability was also excellent. Our study had several limitations. First, it was a retrospective work with a small number of patients. Second, there was no correlation found between Bernageau view and arthroscopic evaluation of glenoid bone loss, maybe because 2 experimental shoulder surgeons described this technique6 and our measurements were performed by several surgeons with different experience levels (advanced to expert shoulder surgeons). Burkhart et al7 used a 3-mm probe with an anterosuperior viewing portal. We used a 5-mm probe with an anterior viewing portal. Third, we used an en face view for the 3D-CT measurements. The acquisition of this view is operator dependent and could introduce human error.9 However, 3D-CT reconstruction provides a far better assessment of the glenoid’s concavity and particularly the inferior width than a Bernageau view. Finally, radiologic measurements were performed manually, which could have caused bias.
Conclusion The glenoid bone defect measurement on the Bernageau profile view is a valid and reliable method. Furthermore, it is easy to use in current clinical practice. Surgeons can therefore consider it as a tool for preoperative planning, and its use could decrease CT scan indications.
1198
Acknowledgments We thank Dr Mark Ferguson, Rosebank Clinic, Center for Sports Medicine and Orthopaedics, Johannesburg, South Africa, for his assistance reviewing the manuscript.
Disclaimer The 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.
References 1. Bernageau J, Patte D, Debeyre J, Ferrane J. Value of the glenoid profil in recurrent luxations of the shoulder [in French]. Rev Chir Orthop Reparatrice Appar Mot 1976;62:142-7. 2. Bigliani LU, Newton PM, Steinmann SP, Connor PM, McLiveen SJ. Glenoid rim lesions associated with recurrent anterior dislocation of the shoulder. Am J Sports Med 1998;26:41-5. 3. Boileau P, Villalba M, Hery JY, Balg F, Ahrens P, Neyton L. Risk factors for recurrence of shoulder instability after arthroscopic Bankart repair. J Bone Joint Surg Am 2006;88:1755-63. http://dx.doi.org/10. 2106/JBJS.E.00817 4. Burkhart SS. The bare spot of the glenoid. Arthroscopy 2007;23:449. author reply 49-50. http://dx.doi.org/10.1016/j.arthro.2006.12.018 5. Burkhart SS, Danaceau SM. Articular arc length mismatch as a cause of failed Bankart repair. Arthroscopy 2000;16:740-4. 6. Burkhart SS, De Beer JF. Traumatic glenohumeral bone defects and their relationship to failure of arthroscopic Bankart repairs: significance of the inverted-pear glenoid and the humeral engaging HillSachs lesion. Arthroscopy 2000;16:677-94. 7. Burkhart SS, Debeer JF, Tehrany AM, Parten PM. Quantifying glenoid bone loss arthroscopically in shoulder instability. Arthroscopy 2002; 18:488-91. http://dx.doi.org/10.1053/jars.2002.32212 8. Cheung EV, Sperling JW, Hattrup SJ, Cofield RH. Long-term outcome of anterior stabilization of the shoulder. J Shoulder Elbow Surg 2008; 17:265-70. http://dx.doi.org/10.1016/j.jse.2007.06.005 9. Chuang TY, Adams CR, Burkhart SS. Use of preoperative threedimensional computed tomography to quantify glenoid bone loss in shoulder instability. Arthroscopy 2008;24:376-82. http://dx.doi.org/10. 1016/j.arthro.2007.10.008 10. Cohen J. Statistical power analysis for the behavioral sciences. Hillsdale (NJ); Lawrence Erlbaum Associates; 1988. 11. De Wilde LF, Berghs BM, Audenaert E, Sys G, Van Maele GO, Barbaix E. About the variability of the shape of the glenoid cavity. Surg Radiol Anat 2004;26:54-9. http://dx.doi.org/10.1007/s00276003-0167-1 12. Edwards TB, Boulahia A, Walch G. Radiographic analysis of bone defects in chronic anterior shoulder instability. Arthroscopy 2003;19: 732-9. http://dx.doi.org/10.1016/S0749-8063(03)00684-4 13. Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. Hoboken (NJ); John Wiley; 2003.
E. Pansard et al. 14. Gerber C, Nyffeler RW. Classification of glenohumeral joint instability. Clin Orthop Relat Res 2002;400:65-76. http://dx.doi.org/10. 1097/00003086-200207000-00009 15. Griffith JF, Antonio GE, Tong CW, Ming CK. Anterior shoulder dislocation: quantification of glenoid bone loss with CT. AJR Am J Roentgenol 2003;180:1423-30. 16. Hovelius LK, Sandstrom BC, Rosmark DL, Saebo M, Sundgren KH, Malmqvist BG. Long-term results with the Bankart and BristowLatarjet procedures: recurrent shoulder instability and arthropathy. J Shoulder Elbow Surg 2001;10:445-52. 17. Huysmans PE, Haen PS, Kidd M, Dhert WJ, Willems JW. The shape of the inferior part of the glenoid: a cadaveric study. J Shoulder Elbow Surg 2006;15:759-63. http://dx.doi.org/10.1016/j.jse.2005.09.001 18. Itoi E, Lee SB, Berglund LJ, Berge LL, An KN. The effect of a glenoid defect on anteroinferior stability of the shoulder after Bankart repair: a cadaveric study. J Bone Joint Surg Am 2000;82:35-46. 19. Jankauskas L, Rudiger HA, Pfirrmann CW, Jost B, Gerber C. Loss of the sclerotic line of the glenoid on anteroposterior radiographs of the shoulder: a diagnostic sign for an osseous defect of the anterior glenoid rim. J Shoulder Elbow Surg 2010;19:151-6. http://dx.doi.org/10. 1016/j.jse.2009.04.013 20. Kralinger F, Aigner F, Longato S, Rieger M, Wambacher M. Is the bare spot a consistent landmark for shoulder arthroscopy? A study of 20 embalmed glenoids with 3-dimensional computed tomographic reconstruction. Arthroscopy 2006;22:428-32. http://dx.doi.org/10. 1016/j.arthro.2005.12.006 21. Larrain MV, Montenegro HJ, Mauas DM, Collazo CC, Pavon F. Arthroscopic management of traumatic anterior shoulder instability in collision athletes: analysis of 204 cases with a 4- to 9-year follow-up and results with the suture anchor technique. Arthroscopy 2006;22: 1283-9. http://dx.doi.org/10.1016/j.arthro.2006.07.052 22. Magarelli N, Milano G, Baudi P, Santagada DA, Righi P, Spina V, et al. Comparison between 2D and 3D computed tomography evaluation of glenoid bone defect in unilateral anterior gleno-humeral instability. Radiol Med 2012;117:102-11. http://dx.doi.org/10.1007/ s11547-011-0712-7 23. Nofsinger C, Browning B, Burkhart SS, Pedowitz RA. Objective preoperative measurement of anterior glenoid bone loss: a pilot study of a computer-based method using unilateral 3-dimensional computed tomography. Arthroscopy 2011;27:322-9. http://dx.doi.org/10.1016/j. arthro.2010.09.007 24. Provencher MT, Detterline AJ, Ghodadra N, Romeo AA, Bach BR Jr, Cole BJ, et al. Measurement of glenoid bone loss: a comparison of measurement error between 45 degrees and 0 degrees bone loss models and with different posterior arthroscopy portal locations. Am J Sports Med 2008;36:1132-8. http://dx.doi.org/10.1177/0363546508316041 25. Roger B, Skaf A, Hooper AW, Lektrakul N, Yeh L, Resnick D. Imaging findings in the dominant shoulder of throwing athletes: comparison of radiography, arthrography, CT arthrography, and MR arthrography with arthroscopic correlation. AJR Am J Roentgenol 1999;172:1371-80. 26. Sommaire C, Penz C, Clavert P, Klouche S, Hardy P, Kempf JF. Recurrence after arthroscopic Bankart repair: Is quantitative radiological analysis of bone loss of any predictive value? Orthop Traumatol Surg Res 2012;98:514-9. http://dx.doi.org/10.1016/j.otsr.2012.03.015 27. Sugaya H, Moriishi J, Dohi M, Kon Y, Tsuchiya A. Glenoid rim morphology in recurrent anterior glenohumeral instability. J Bone Joint Surg Am 2003;85:878-84. 28. Thomazeau H, Courage O, Barth J, Pelegri C, Charousset C, Lespagnol F, et al. Can we improve the indication for Bankart arthroscopic repair? A preliminary clinical study using the ISIS score. Orthop Traumatol Surg Res 2010;96:S77-83. http://dx.doi.org/10. 1016/j.otsr.2010.09.007 29. Yamamoto N, Itoi E, Abe H, Kikuchi K, Seki N, Minagawa H, et al. Effect of an anterior glenoid defect on anterior shoulder stability: a cadaveric study. Am J Sports Med 2009;37:949-54. http://dx.doi.org/ 10.1177/0363546508330139