- Email: [email protected]
The interobserver reliability in diagnosing osseous lesions after first-time anterior shoulder dislocation comparing plain radiographs with computed tomography scans
J Shoulder Elbow Surg (2013) 22, 1507-1513
www.elsevier.com/locate/ymse
The interobserver reliability in diagnosing osseous lesions after first-time anterior shoulder dislocation comparing plain radiographs with computed tomography scans Alexander Auffarth, MDa,*, Michael Mayer, MDa, Barbara Kofler, MDb, Wolfgang Hitzl, PhDc, Robert Bogner, MDa, Philipp Moroder, MDa, Gundobert Korn, MDa, Heiko Koller, MDa, Herbert Resch, MDa a
Department of Traumatology and Sports Injuries, Paracelsus Medical University, Salzburg, Austria Department of Radiology, Paracelsus Medical University, Salzburg, Austria c Department of Biostatistics, Paracelsus Medical University, Salzburg, Austria b
Background: Recurrence after first-time traumatic anterior shoulder dislocation is frequent. The prevalence of glenoid bone loss ranges from 41% after a first-time dislocation to 86% with recurrent dislocation. Postoperative recurrence can occur in up to 10% of cases. Thus, misdiagnosis of bony glenoid rim lesions has been assumed a major cause for failure. We evaluated the interobserver reliability of radiologic diagnoses after first-time traumatic shoulder dislocation based on radiographs and computed tomography (CT) images. Methods: Digital radiographs before and after reduction and CT images after reduction of 20 patients with a first-time shoulder dislocation were assessed by 6 observers. It was recorded whether they diagnosed a lesion at the greater tuberosity, a lesion at the glenoid rim, a Hill-Sachs lesion, or any other skeletal pathology. The average agreement among the investigators was evaluated, and radiographic diagnoses were compared with those based on CT images. Results: Of the 10 cases that presented with a glenoid rim fracture, each investigator had overlooked at least 1 fracture (range, 1-4) on the radiographs. No investigator had diagnosed all 8 Hill-Sachs lesions on the presented images. The average overall agreement among the investigators and corresponding sensitivity and specificity were calculated. Agreement of diagnoses based on radiographs with those based on CT images was lowest for glenoid rim fractures and Hill-Sachs lesions. Conclusion: Radiographs seem inferior to CT scans for assessing osseous lesions especially at the glenoid rim. We suggest performing a CT scan of the shoulder after primary dislocation to apply the correct treatment early and potentially avoid further dislocations.
Ethics approval: The ethics board of the province of Salzburg, Austria, stated that no ethical approval was required (file No. 415-EP/73/110-2012).
*Reprint requests: Alexander Auffarth, MD, Department of Traumatology and Sports Injuries, Paracelsus Medical University, Muellner Hauptstrasse 48, A-5020 Salzburg, Austria. E-mail address: [email protected] (A. Auffarth).
1058-2746/$ - see front matter Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2013.04.020
1508
A. Auffarth et al.
Level of evidence: Level III, Nonconsecutive Patients, Diagnostic Study. Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Anterior shoulder dislocation; osseous lesions; interobserver reliability; comparison of radiographs and CT scans
The prevalence of a traumatic anterior shoulder dislocation has been estimated at 2% over a lifetime in the general population.17 In such cases with first-time traumatic anterior shoulder dislocations, the initial therapeutic approach will dominantly be the reduction of the joint with a technique preferred by the physician and immobilization for about 3 weeks followed by physical therapy.5,17 Yet, recurrent instability after first-time traumatic anterior shoulder dislocation is frequently observed. Recently, a systematic literature review reported recurrence rates of 58% for conservative treatment and 10% after surgical repair at 3 to 10 years’ follow-up.5 In our experience, often the physician’s attention is not further drawn to the case until the patient may return with recurrent instability. We presume that the reasons for failure of conservative treatment may basically be the same as those for surgery, the only difference being the recurrence rate. Thus, searching the literature for articles on failure of conservatively treated post-traumatic shoulder dislocations is dispensable. For failed surgical stabilization, though, several authors have sought to identify risk factors for such recurrences. Searching for explanations for the failure of arthroscopic stabilizations, for example, Tauber et al27 found that in 47% of cases with postoperative recurrences, significant glenoid rim defects had not been taken into account at the first stabilizing procedure. Boileau et al4 also found that postoperative recurrence was significantly related to the presence of a bone defect on the glenoid side or the humeral head. In our department, patients who have elected to undergo a bone block procedure because of a significant osseous glenoid rim defect have often been previously operated on elsewhere. Still, in about half of the cases, the required bone block procedure will be the first-time surgical intervention. This led to the following question: How it could have happened that glenoid rim fragments have obviously not been assessed in such patients ending up requiring reconstructive surgery? The final defect requiring such bony augmentation may in part be caused by compression through recurrent dislocations. We presume, though, that the origin of such large defects may be an initially undiagnosed glenoid rim fracture with a fragment left unreduced. Because patients scheduled for a stabilization procedure have often sustained multiple recurrences, it may be hard to define which event could have been the event that caused the mentioned fracture. Therefore, with this investigation, we sought to investigate the incidence and especially the interobserver reliability of diagnosing
osseous lesions after a first-time traumatic shoulder dislocation.
Methods During a 2-year period at a level 1 trauma center, all patients admitted with a first-time traumatic anterior shoulder dislocation underwent standard radiographs taken before and after reduction by a physician. In the dislocated shoulder, these were a true anteroposterior (AP) view in neutral rotation and an axillary view, if tolerated by the patient, or a scapular-Y view instead. After reduction, radiographs in the AP view in internal rotation and a scapular-Y view were obtained. In addition, a computed tomography (CT) scan of the reduced shoulder joint with images reconstructed in the axial and sagittal planes was obtained. Radiographs were taken on a digital radiography system (Vertix 3D-III unit; Siemens, Erlangen, Germany) and stored (PACS Magic View VC 42, release A; Siemens). Images were processed with a commercial DICOM (Digital Imaging and Communications in Medicine) viewing program (Escape Medical Viewer V3; Escape, Thessaloniki, Greece). Shoulder CT scans were performed on a 16-row helical CT scanner (Somatom Volume Zoom; Siemens) with the following image parameters: 1mm collimation, 3-mm slice thickness, 140 kV, and 250 mAs. Multiplanar reconstructions were performed in the coronal and sagittal planes at a slice thickness of 2 mm. Despite a standard protocol determining how to proceed in patients admitted with a first-time traumatic anterior shoulder dislocation, in particular, images in 2 planes before reduction had not always been taken. Hence, of the patients with complete sets of radiographs and the following CT scan, 20 with and without osseous lesions about the shoulder joint were selected for the investigation. The author who selected the cases did not participate in the investigation. Images were prepared in a presentation to be observed by 6 investigators (PowerPoint; Microsoft, Redmond, WA, USA). Creating the presentation, original DICOM data with loss-less compression were used. All images were processed with a commercial software program (Adobe Photoshop; Adobe, San Jose, CA, USA). Digital radiographs and CT scans were scaled down with the field of interest placed on the shoulder joint and the glenoid, respectively. No computerized effects were used to influence image quality such as brightness or contrast. The radiographic quality shown to the observers equaled that of a beamer or commercially available DICOM viewer software. In the final presentation, radiographs in 2 planes each, before and after reduction of the shoulder joint, had to be judged. Next, the observer had to state whether he or she would see an indication for a CT scan. Finally, 3 images each of the axial and coronal reconstructions of the CT scan best presenting the area of interest at the glenoid and humeral head were presented to the observers.
Interobserver reliability diagnosing glenoid rim lesions This presentation was shown to 6 investigators: 5 orthopaedic surgeons and 1 radiologist. The orthopaedic surgeons’ experience with interpretation of skeletal radiologic images averaged 4.4 years: 2 years for two of them and 3, 7, and 8 years for the remaining three. The radiologist had had 15 years’ experience with skeletal radiology. The presentation was shown by projection with a high-resolution beamer (1.5 megapixels, Canon Xeed SX60; Canon, Tokyo, Japan). On a worksheet, the observers had to record whether (or not) they diagnosed an osseous lesion at the greater tuberosity, an osseous lesion at the glenoid rim, a HillSachs lesion at the humeral head, or any other skeletal pathology on 1 of the presented radiographs and CT images. For the statistical analysis, radiographs before reduction were compared with the CT scans because after reduction of the shoulder joint, an eventual osseous fragment of the glenoid rim could be back in place by reduction of the humeral head (Fig. 1).
1509
Figure 1 Example of presented radiograph before reduction and CT scan after reduction of shoulder joint.
Statistical methods Within each observed region that was diagnosed with CT scans and radiographs, the overall agreement among 6 raters was computed within each single patient. Overall agreement was given if all 6 raters ended up with the same results simultaneously. If at least 1 rater disagreed with another rater, no overall agreement was given. Hence, the minimal possible number for overall agreement is 3, which corresponds to 50% agreement, and the maximum number is 6, which corresponds to 100% agreement. In the next step, the means and standard deviations of the overall agreement for 20 patients were computed. For the comparisons between CT and radiography within each observed region, a single patient was selected and the corresponding overall agreement was computed. This was performed for each rater separately. In the next step, the average of 6 raters was computed and allocated to each patient. In the last step, the corresponding means and standard deviations for 20 patients were computed. Sensitivity and specificity values of each rater were illustrated by use of receiver operating characteristic curves for each region. Finally, the average sensitivities and specificities, together with corresponding 95% confidence intervals, were computed. All analyses and computations were performed with Statistica (version 10.0; StatSoft, Tulsa, OK, USA) and Mathematica (version 7.0; Wolfram Research, Hanborough, UK).
Results Of the 20 selected cases presented to the investigators, 10 (50%) had a glenoid rim fracture and 8 (40%) had an impression at the humeral head (Hill-Sachs lesion). A fracture of the greater tuberosity was present in 3 of the cases (15%). One patient presented with a cranially migrated humerus due to an additional rotator cuff tear, which was identified as ‘‘other’’ by 4 investigators. For glenoids with a fracture, the glenoid index according to Chuang et al11 was calculated. The mean index was 0.74, ranging from 0.46 to 0.94. On the 4 presented radiographs of each of the 20 patients, each investigator had overlooked at least 1 glenoid rim fracture (range, 1-4), which the investigator
later identified on the CT images. One investigator rated an avulsion at the glenoid rim as an intact glenoid on the CT images. The cases in which a bony Bankart fragment had been overlooked were not always the same cases, which could have led to the assumption that these may have been the cases with the smallest lesions. Eight of the 10 glenoids with a rim fracture were judged as intact on the presented radiographic images by at least 1 investigator. The glenoid index of these 8 glenoids ranged from 0.46 to 0.94. All of the correctly diagnosed Hill-Sachs lesions were identified on the radiographic images but were not always correctly diagnosed on the presented CT images. No investigator had diagnosed all 8 Hill-Sachs lesions on the presented images. Rates of correct identification of these lesions ranged from 3 to 7. Each investigator had correctly diagnosed all 3 fractures of the greater tuberosity. They were always diagnosed on the AP radiographic images before reduction and on the presented coronal CT images. The trans-scapular and axillary views before and after reduction, as well as the AP view after reduction, were not rated as to picture these lesions as consistently. After the 4 presented conventional radiographs had been judged, the need for an additional CT scan of the injured shoulder joint was rated as ‘‘yes’’ in 78% of cases. This showed that each investigator had not seen a need for a CT scan in at least 1 case (up to 4 cases) that then turned out to show an osseous glenoid rim lesion on the CT images. On the statistical analysis, the average agreement of diagnosis of the 6 investigators regarding the 4 observed regions (greater tuberosity, glenoid, Hill-Sachs lesion, and other) based on CT and radiography before reduction AP views was first compared (Table I). Regarding the diagnoses based on the radiographic images, agreement among the raters was lowest for glenoid rim fractures and Hill-Sachs lesions. Comparing the diagnoses based on the CT images, we found that agreement among the raters was lowest for Hill-Sachs lesions. Next, the overall agreement of diagnoses based on radiographic images with diagnoses based on the CT scans
1510
A. Auffarth et al.
Table I Comparison of average agreement of 6 raters for different values of CT scans and radiographs before reduction AP view Variable
Average agreement SD among 6 raters
Radiograph before reduction AP view Other 97% Greater tuberosity 96% Hill-Sachs lesion 83% Glenoid 79% Indication for CT: Yes 78% CT Other 98% Greater tuberosity 98% Glenoid 93% Hill-Sachs lesion 72%
10% 7% 15% 17% 16% 7% 6% 13% 16%
was calculated (Table II). This, for example, showed that the diagnosis of an intact glenoid based on the radiographic images was wrong in 26% of cases. Again, this comparison showed that agreement of radiographic and CT diagnoses was lowest for Hill-Sachs lesions and glenoid rim fractures (Table II). In addition, the sensitivity and specificity of the individual investigators were calculated (Fig. 2). As with the other calculations, the sensitivity for correct diagnosis of Hill-Sachs lesions was lowest, followed by that of glenoid rim fractures (Table III).
Discussion Not only patients with recurrent post-traumatic shoulder instability scheduled for revision surgery but also patients with first-time stabilization often present with large glenoid bone defects. Griffith et al,14 for example, investigating the prevalence of glenoid bone loss in anterior shoulder dislocation, described such glenoid bone loss in 41% of patients with a first-time unidirectional dislocation. In cases with recurrent dislocation, bone loss was described even in 86% of cases. For revision cases, Shah et al25 stated that failures might persist and could be attributed to poor preoperative as well as intraoperative judgment. This statement was also underlined by other authors looking for reasons for recurrence of instability after surgical shoulder stabilization.9,27 In this context, Sasyniuk et al24 found that the inter-rater reliability of 40% for orthopaedic shoulder surgeons’ arthroscopic assessment of intra-articular anatomy in patients with anterior shoulder instability was poor when rating the glenoid and the anterior inferior glenohumeral ligament. This might in part explain the origin of recurrences after surgical stabilization but also underlines the need for detailed preoperative radiologic examination. Therefore, it was our attempt to analyze whether misdiagnosis may already be a problem at the initial
Table II Comparison of CT versus radiography agreement based on 6 raters CT vs radiography before reduction AP view
Average agreement among 6 raters
SD
Other Greater tuberosity Glenoid Hill-Sachs lesion
97% 95% 74% 65%
10% 10% 24% 22%
treatment of these patients, which would be their initial hospital visit with the first-time traumatic shoulder dislocation. In this regard, the purpose of this investigation was to evaluate the reliability of the diagnosis of bony lesions about the shoulder joint at a first-time shoulder dislocation on conventional radiographs, which was controlled by a CT scan. For patients with manifest shoulder instability, Bushnell et al9 reported that even with a true AP view, an internal and external rotation view, a scapular-Y view, an axillary lateral view, a Stryker notch view, a West Point view, and an apical oblique (Garth) view, the sensitivity for detection of significant glenoid bone loss, which was defined as at least involving 25% of the articular surface, was just 50%. Apart from their limited sensitivity in detecting bony lesions, most of these views are usually just obtained to detect glenoid rim lesions in recurrent instability. Apart from this, in the setting of an acute first-time or recurrent shoulder dislocation, most of these images are difficult to obtain because of the required positioning of the patient’s arm. For this reason, the standard images in such cases are an AP view and, if possible, an axial view or a transscapular radiograph instead. In addition, the patient with a first-time shoulder dislocation might not always be directly transferred to a clinic specializing in shoulder surgery but will be treated in the nearest hospital, so training of physicians and radiologists in detecting such osseous lesions on common radiographs first treating the patient may vary. Moreover, Bigliani et al3 and Provencher et al22 stated that glenoid rim lesions appearing either as an acute fracture or a bony erosion could often be missed or inaccurately quantified on routine radiographs. Thus, we assumed that these standard images would be just as insufficient for detecting osseous lesions especially at the glenoid rim as the images described earlier by Bushnell et al.9 Salomonsson et al23 stated that a bony Bankart lesion would be a prognostic factor indicating a good functional result and a stable shoulder joint after a primary dislocation. We presume that this would only be the case when such a bony lesion is correctly diagnosed and thus adequately treated. In their investigation, they also reported that only 60% of glenoid rim fractures were detected on plain radiographs. Identification of such defects is essential, though, because the presence of such pathology has to be
Interobserver reliability diagnosing glenoid rim lesions
1511
Figure 2 Sensitivity and specificity of individual investigators judging 4 regions. Glen, Glenoid; Hill-S, Hill-Sachs lesion; Tub, greater tuberosity.
Table III Mean specificities and sensitivities of 6 raters with 95% confidence intervals Variable Glenoid Specificity Sensitivity Greater tuberosity Specificity Sensitivity Hill-Sachs lesion Specificity Sensitivity Other Specificity Sensitivity
Mean
95% Confidence interval
80% 67%
65%-96% 60%-73%
95% 94%
91%-100% 80%-100%
87% 20%
71%-100% 0%-53%
98% 50%
93%-100% NA
NA, not applicable.
taken into consideration when one is planning the patient’s treatment. The dislocation resistance of a once-dislocated shoulder joint is, according to Gerber and Nyffeler,12 significantly influenced by the presence or absence of a glenoid rim fracture or defect. Size15 and location28 of this pathology determine its individual effect on stability of the shoulder joint. When one takes these factors into account, surgical intervention, either by soft-tissue repair or osteosynthesis in the acute case, has to be considered a therapeutic option and may prevent further dislocations or even the later need for a reconstructive bone block procedure at the glenoid rim. Among the undiagnosed cases, the 4 glenoids with the largest fragments had a glenoid index of
0.46, 0.55, 0.61, and 0.70, respectively. In this context, Bushnell et al8 gave a nice overview citing various authors’ recommendations for bony reconstruction surgery for recurrent instability with a glenoid bone defect. These ranged from a 20% to 30% defect evaluated in biomechanical or clinical trials. Because the mentioned investigations applied different methods to evaluate the defect sizes,1-3,6,7,10,13,16,18,19,21,26,27 which value should generally be recommended as a cutoff remains unanswered so far. The 4 previously mentioned cases would indicate acute 54%, 45%, 39%, and 30% defects, which may not cause much discussion on which treatment to choose. The remaining 4 overlooked cases had a glenoid index ranging from 0.80 to 0.94. Defining the adequate treatment for these cases may lead to more controversy. Still, according to our clinical experience, even fragments smaller than the often attributed 21%16 and 25%1,3,6,7,18,21,26 should be reduced and internally fixed either arthroscopically or by an open procedure to reconstruct the glenoid cavity. In this context, we agree with Ochoa and Burkhart,20 who stated that it would be imperative to achieve an anatomic repair to restore normal biomechanics and the concavity-compression effect. On the long course, separation of such smaller fragments may also be the origin of greater defects constantly enlarged by erosion or compression because of recurring dislocations. Like us, other authors have also stated that radiography may misrepresent osseous lesions in dislocated shoulder joints.8,9 To our knowledge, this is the first investigation, though, to evaluate the interobserver reliability of diagnoses regarding osseous integrity after primary traumatic
1512 shoulder dislocation based on conventional radiographic images. The assumption that conventional radiographs alone would be insufficient for the reliable diagnosis of glenoid rim fractures in patients with a first-time traumatic shoulder dislocation was conclusively demonstrated. This is the reason we suggest performing a CT scan of the reduced shoulder after a first-time dislocation to not initially overlook a lesion that might be the source of recurrent instability.
Conclusion An osseous glenoid rim lesion after a traumatic shoulder dislocation left untreated can cause the onset of recurrent shoulder instability. According to the literature and our investigation, with plain radiographs alone, even osseous fragments that would need refixation are likely to be overlooked. Therefore, we suggest performing a CT scan of the shoulder after a first-time traumatic shoulder dislocation so that the correct treatment can be applied at the correct time.
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. Auffarth A, Schauer J, Matis N, Kofler B, Hitzl W, Resch H. The J-bone graft for anatomical glenoid reconstruction in recurrent posttraumatic anterior shoulder dislocation. Am J Sports Med 2008;36: 638-47. http://dx.doi.org/10.1177/0363546507309672 2. Bahk M, Keyurapan E, Tasaki A, Sauers EL, McFarland EG. Laxity testing of the shoulder: a review. Am J Sports Med 2007;35:131-44. http://dx.doi.org/10.1177/0363546506294570 3. Bigliani LU, Newton PM, Steinmann SP, Connor PM, McLlveen SJ. Glenoid rim lesions associated with recurrent anterior dislocation of the shoulder. Am J Sports Med 1998;26:41-5. 4. 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 5. Brophy RH, Marx RG. The treatment of traumatic anterior instability of the shoulder: nonoperative and surgical treatment. Arthroscopy 2009;25:298-304. http://dx.doi.org/10.1016/j.arthro.2008.12.007 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.
A. Auffarth et al. 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. Bushnell BD, Creighton RA, Herring MM. Bony instability of the shoulder. Arthroscopy 2008;24:1061-73. http://dx.doi.org/10.1016/j. arthro.2008.05.015 9. Bushnell BD, Creighton RA, Herring MM. The bony apprehension test for instability of the shoulder: a prospective pilot analysis. Arthroscopy 2008;24:974-82. http://dx.doi.org/10.1016/j.arthro.2008. 07.019 10. Chen AL, Hunt SA, Hawkins RJ, Zuckerman JD. Management of bone loss associated with recurrent anterior glenohumeral instability. Am J Sports Med 2005;33:912-25. http://dx.doi.org/10.1177/ 0363546505277074 11. 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 12. 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 13. Greis PE, Scuderi MG, Mohr A, Bachus KN, Burks RT. Glenohumeral articular contact areas and pressures following labral and osseous injury to the anteroinferior quadrant of the glenoid. J Shoulder Elbow Surg 2002;11:442-51. http://dx.doi.org/10.1067/mse.2002.124526 14. Griffith JF, Antonio GE, Yung PS, Wong EM, Yu AB, Ahuja AT, et al. Prevalence, pattern, and spectrum of glenoid bone loss in anterior shoulder dislocation: CT analysis of 218 patients. AJR Am J Roentgenol 2008;190:1247-54. http://dx.doi.org/10.2214/AJR.07.3009 15. Itoi E, Lee SB, Amrami KK, Wenger DE, An KN. Quantitative assessment of classic anteroinferior bony Bankart lesions by radiography and computed tomography. Am J Sports Med 2003;31:112-8. 16. 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. 17. Kirkley A, Werstine R, Ratjek A, Griffin S. Prospective randomized clinical trial comparing the effectiveness of immediate arthroscopic stabilization versus immobilization and rehabilitation in first traumatic anterior dislocations of the shoulder: long-term evaluation. Arthroscopy 2005;21:55-63. http://dx.doi.org/10.1016/j.arthro.2004.09.018 18. Lo IK, Parten PM, Burkhart SS. The inverted pear glenoid: an indicator of significant glenoid bone loss. Arthroscopy 2004;20:169-74. http://dx.doi.org/10.1016/j.arthro.2003.11.036 19. Mologne TS, Provencher MT, Menzel KA, Vachon TA, Dewing CB. Arthroscopic stabilization in patients with an inverted pear glenoid: results in patients with bone loss of the anterior glenoid. Am J Sports Med 2007;35:1276-83. http://dx.doi.org/10.1177/0363546507300262 20. Ochoa E Jr, Burkhart SS. Glenohumeral bone defects in the treatment of anterior shoulder instability. Instr Course Lect 2009;58:323-36. 21. Porcellini G, Campi F, Paladini P. Arthroscopic approach to acute bony Bankart lesion. Arthroscopy 2002;18:764-9. http://dx.doi.org/10. 1053/jars.2002.35266 22. Provencher MT, Bhatia S, Ghodadra NS, Grumet RC, Bach BR Jr, Dewing CB, et al. Recurrent shoulder instability: current concepts for evaluation and management of glenoid bone loss. J Bone Joint Surg Am 2010;92(Suppl 2):133-51. http://dx.doi.org/10.2106/JBJS.J.00906 23. Salomonsson B, von Heine A, Dahlborn M, Abbaszadegan H, Ahlstrom S, Dalen N, et al. Bony Bankart is a positive predictive factor after primary shoulder dislocation. Knee Surg Sports Traumatol Arthrosc 2010;18:1425-31. http://dx.doi.org/10.1007/s00167-0090998-3 24. Sasyniuk TM, Mohtadi NG, Hollinshead RM, Russell ML, Fick GH. The inter-rater reliability of shoulder arthroscopy. Arthroscopy 2007; 23:971-7. http://dx.doi.org/10.1016/j.arthro.2007.03.005 25. Shah AS, Karadsheh MS, Sekiya JK. Failure of operative treatment for glenohumeral instability: etiology and management. Arthroscopy 2011;27:681-94. http://dx.doi.org/10.1016/j.arthro.2010.11.057
Interobserver reliability diagnosing glenoid rim lesions 26. Sugaya H, Moriishi J, Kanisawa I, Tsuchiya A. Arthroscopic osseous Bankart repair for chronic recurrent traumatic anterior glenohumeral instability. J Bone Joint Surg Am 2005;87:1752-60. http://dx.doi.org/ 10.2106/JBJS.D.02204 27. Tauber M, Resch H, Forstner R, Raffl M, Schauer J. Reasons for failure after surgical repair of anterior shoulder instability.
1513 J Shoulder Elbow Surg 2004;13:279-85. http://dx.doi.org/10.1016/ S1058274604000254 28. 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
The interobserver reliability in diagnosing osseous lesions after first-time anterior shoulder dislocation comparing plain radiographs with computed tomography scans Alexander Auffarth, MDa,*, Michael Mayer, MDa, Barbara Kofler, MDb, Wolfgang Hitzl, PhDc, Robert Bogner, MDa, Philipp Moroder, MDa, Gundobert Korn, MDa, Heiko Koller, MDa, Herbert Resch, MDa a
Department of Traumatology and Sports Injuries, Paracelsus Medical University, Salzburg, Austria Department of Radiology, Paracelsus Medical University, Salzburg, Austria c Department of Biostatistics, Paracelsus Medical University, Salzburg, Austria b
Background: Recurrence after first-time traumatic anterior shoulder dislocation is frequent. The prevalence of glenoid bone loss ranges from 41% after a first-time dislocation to 86% with recurrent dislocation. Postoperative recurrence can occur in up to 10% of cases. Thus, misdiagnosis of bony glenoid rim lesions has been assumed a major cause for failure. We evaluated the interobserver reliability of radiologic diagnoses after first-time traumatic shoulder dislocation based on radiographs and computed tomography (CT) images. Methods: Digital radiographs before and after reduction and CT images after reduction of 20 patients with a first-time shoulder dislocation were assessed by 6 observers. It was recorded whether they diagnosed a lesion at the greater tuberosity, a lesion at the glenoid rim, a Hill-Sachs lesion, or any other skeletal pathology. The average agreement among the investigators was evaluated, and radiographic diagnoses were compared with those based on CT images. Results: Of the 10 cases that presented with a glenoid rim fracture, each investigator had overlooked at least 1 fracture (range, 1-4) on the radiographs. No investigator had diagnosed all 8 Hill-Sachs lesions on the presented images. The average overall agreement among the investigators and corresponding sensitivity and specificity were calculated. Agreement of diagnoses based on radiographs with those based on CT images was lowest for glenoid rim fractures and Hill-Sachs lesions. Conclusion: Radiographs seem inferior to CT scans for assessing osseous lesions especially at the glenoid rim. We suggest performing a CT scan of the shoulder after primary dislocation to apply the correct treatment early and potentially avoid further dislocations.
Ethics approval: The ethics board of the province of Salzburg, Austria, stated that no ethical approval was required (file No. 415-EP/73/110-2012).
*Reprint requests: Alexander Auffarth, MD, Department of Traumatology and Sports Injuries, Paracelsus Medical University, Muellner Hauptstrasse 48, A-5020 Salzburg, Austria. E-mail address: [email protected] (A. Auffarth).
1058-2746/$ - see front matter Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2013.04.020
1508
A. Auffarth et al.
Level of evidence: Level III, Nonconsecutive Patients, Diagnostic Study. Ó 2013 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Anterior shoulder dislocation; osseous lesions; interobserver reliability; comparison of radiographs and CT scans
The prevalence of a traumatic anterior shoulder dislocation has been estimated at 2% over a lifetime in the general population.17 In such cases with first-time traumatic anterior shoulder dislocations, the initial therapeutic approach will dominantly be the reduction of the joint with a technique preferred by the physician and immobilization for about 3 weeks followed by physical therapy.5,17 Yet, recurrent instability after first-time traumatic anterior shoulder dislocation is frequently observed. Recently, a systematic literature review reported recurrence rates of 58% for conservative treatment and 10% after surgical repair at 3 to 10 years’ follow-up.5 In our experience, often the physician’s attention is not further drawn to the case until the patient may return with recurrent instability. We presume that the reasons for failure of conservative treatment may basically be the same as those for surgery, the only difference being the recurrence rate. Thus, searching the literature for articles on failure of conservatively treated post-traumatic shoulder dislocations is dispensable. For failed surgical stabilization, though, several authors have sought to identify risk factors for such recurrences. Searching for explanations for the failure of arthroscopic stabilizations, for example, Tauber et al27 found that in 47% of cases with postoperative recurrences, significant glenoid rim defects had not been taken into account at the first stabilizing procedure. Boileau et al4 also found that postoperative recurrence was significantly related to the presence of a bone defect on the glenoid side or the humeral head. In our department, patients who have elected to undergo a bone block procedure because of a significant osseous glenoid rim defect have often been previously operated on elsewhere. Still, in about half of the cases, the required bone block procedure will be the first-time surgical intervention. This led to the following question: How it could have happened that glenoid rim fragments have obviously not been assessed in such patients ending up requiring reconstructive surgery? The final defect requiring such bony augmentation may in part be caused by compression through recurrent dislocations. We presume, though, that the origin of such large defects may be an initially undiagnosed glenoid rim fracture with a fragment left unreduced. Because patients scheduled for a stabilization procedure have often sustained multiple recurrences, it may be hard to define which event could have been the event that caused the mentioned fracture. Therefore, with this investigation, we sought to investigate the incidence and especially the interobserver reliability of diagnosing
osseous lesions after a first-time traumatic shoulder dislocation.
Methods During a 2-year period at a level 1 trauma center, all patients admitted with a first-time traumatic anterior shoulder dislocation underwent standard radiographs taken before and after reduction by a physician. In the dislocated shoulder, these were a true anteroposterior (AP) view in neutral rotation and an axillary view, if tolerated by the patient, or a scapular-Y view instead. After reduction, radiographs in the AP view in internal rotation and a scapular-Y view were obtained. In addition, a computed tomography (CT) scan of the reduced shoulder joint with images reconstructed in the axial and sagittal planes was obtained. Radiographs were taken on a digital radiography system (Vertix 3D-III unit; Siemens, Erlangen, Germany) and stored (PACS Magic View VC 42, release A; Siemens). Images were processed with a commercial DICOM (Digital Imaging and Communications in Medicine) viewing program (Escape Medical Viewer V3; Escape, Thessaloniki, Greece). Shoulder CT scans were performed on a 16-row helical CT scanner (Somatom Volume Zoom; Siemens) with the following image parameters: 1mm collimation, 3-mm slice thickness, 140 kV, and 250 mAs. Multiplanar reconstructions were performed in the coronal and sagittal planes at a slice thickness of 2 mm. Despite a standard protocol determining how to proceed in patients admitted with a first-time traumatic anterior shoulder dislocation, in particular, images in 2 planes before reduction had not always been taken. Hence, of the patients with complete sets of radiographs and the following CT scan, 20 with and without osseous lesions about the shoulder joint were selected for the investigation. The author who selected the cases did not participate in the investigation. Images were prepared in a presentation to be observed by 6 investigators (PowerPoint; Microsoft, Redmond, WA, USA). Creating the presentation, original DICOM data with loss-less compression were used. All images were processed with a commercial software program (Adobe Photoshop; Adobe, San Jose, CA, USA). Digital radiographs and CT scans were scaled down with the field of interest placed on the shoulder joint and the glenoid, respectively. No computerized effects were used to influence image quality such as brightness or contrast. The radiographic quality shown to the observers equaled that of a beamer or commercially available DICOM viewer software. In the final presentation, radiographs in 2 planes each, before and after reduction of the shoulder joint, had to be judged. Next, the observer had to state whether he or she would see an indication for a CT scan. Finally, 3 images each of the axial and coronal reconstructions of the CT scan best presenting the area of interest at the glenoid and humeral head were presented to the observers.
Interobserver reliability diagnosing glenoid rim lesions This presentation was shown to 6 investigators: 5 orthopaedic surgeons and 1 radiologist. The orthopaedic surgeons’ experience with interpretation of skeletal radiologic images averaged 4.4 years: 2 years for two of them and 3, 7, and 8 years for the remaining three. The radiologist had had 15 years’ experience with skeletal radiology. The presentation was shown by projection with a high-resolution beamer (1.5 megapixels, Canon Xeed SX60; Canon, Tokyo, Japan). On a worksheet, the observers had to record whether (or not) they diagnosed an osseous lesion at the greater tuberosity, an osseous lesion at the glenoid rim, a HillSachs lesion at the humeral head, or any other skeletal pathology on 1 of the presented radiographs and CT images. For the statistical analysis, radiographs before reduction were compared with the CT scans because after reduction of the shoulder joint, an eventual osseous fragment of the glenoid rim could be back in place by reduction of the humeral head (Fig. 1).
1509
Figure 1 Example of presented radiograph before reduction and CT scan after reduction of shoulder joint.
Statistical methods Within each observed region that was diagnosed with CT scans and radiographs, the overall agreement among 6 raters was computed within each single patient. Overall agreement was given if all 6 raters ended up with the same results simultaneously. If at least 1 rater disagreed with another rater, no overall agreement was given. Hence, the minimal possible number for overall agreement is 3, which corresponds to 50% agreement, and the maximum number is 6, which corresponds to 100% agreement. In the next step, the means and standard deviations of the overall agreement for 20 patients were computed. For the comparisons between CT and radiography within each observed region, a single patient was selected and the corresponding overall agreement was computed. This was performed for each rater separately. In the next step, the average of 6 raters was computed and allocated to each patient. In the last step, the corresponding means and standard deviations for 20 patients were computed. Sensitivity and specificity values of each rater were illustrated by use of receiver operating characteristic curves for each region. Finally, the average sensitivities and specificities, together with corresponding 95% confidence intervals, were computed. All analyses and computations were performed with Statistica (version 10.0; StatSoft, Tulsa, OK, USA) and Mathematica (version 7.0; Wolfram Research, Hanborough, UK).
Results Of the 20 selected cases presented to the investigators, 10 (50%) had a glenoid rim fracture and 8 (40%) had an impression at the humeral head (Hill-Sachs lesion). A fracture of the greater tuberosity was present in 3 of the cases (15%). One patient presented with a cranially migrated humerus due to an additional rotator cuff tear, which was identified as ‘‘other’’ by 4 investigators. For glenoids with a fracture, the glenoid index according to Chuang et al11 was calculated. The mean index was 0.74, ranging from 0.46 to 0.94. On the 4 presented radiographs of each of the 20 patients, each investigator had overlooked at least 1 glenoid rim fracture (range, 1-4), which the investigator
later identified on the CT images. One investigator rated an avulsion at the glenoid rim as an intact glenoid on the CT images. The cases in which a bony Bankart fragment had been overlooked were not always the same cases, which could have led to the assumption that these may have been the cases with the smallest lesions. Eight of the 10 glenoids with a rim fracture were judged as intact on the presented radiographic images by at least 1 investigator. The glenoid index of these 8 glenoids ranged from 0.46 to 0.94. All of the correctly diagnosed Hill-Sachs lesions were identified on the radiographic images but were not always correctly diagnosed on the presented CT images. No investigator had diagnosed all 8 Hill-Sachs lesions on the presented images. Rates of correct identification of these lesions ranged from 3 to 7. Each investigator had correctly diagnosed all 3 fractures of the greater tuberosity. They were always diagnosed on the AP radiographic images before reduction and on the presented coronal CT images. The trans-scapular and axillary views before and after reduction, as well as the AP view after reduction, were not rated as to picture these lesions as consistently. After the 4 presented conventional radiographs had been judged, the need for an additional CT scan of the injured shoulder joint was rated as ‘‘yes’’ in 78% of cases. This showed that each investigator had not seen a need for a CT scan in at least 1 case (up to 4 cases) that then turned out to show an osseous glenoid rim lesion on the CT images. On the statistical analysis, the average agreement of diagnosis of the 6 investigators regarding the 4 observed regions (greater tuberosity, glenoid, Hill-Sachs lesion, and other) based on CT and radiography before reduction AP views was first compared (Table I). Regarding the diagnoses based on the radiographic images, agreement among the raters was lowest for glenoid rim fractures and Hill-Sachs lesions. Comparing the diagnoses based on the CT images, we found that agreement among the raters was lowest for Hill-Sachs lesions. Next, the overall agreement of diagnoses based on radiographic images with diagnoses based on the CT scans
1510
A. Auffarth et al.
Table I Comparison of average agreement of 6 raters for different values of CT scans and radiographs before reduction AP view Variable
Average agreement SD among 6 raters
Radiograph before reduction AP view Other 97% Greater tuberosity 96% Hill-Sachs lesion 83% Glenoid 79% Indication for CT: Yes 78% CT Other 98% Greater tuberosity 98% Glenoid 93% Hill-Sachs lesion 72%
10% 7% 15% 17% 16% 7% 6% 13% 16%
was calculated (Table II). This, for example, showed that the diagnosis of an intact glenoid based on the radiographic images was wrong in 26% of cases. Again, this comparison showed that agreement of radiographic and CT diagnoses was lowest for Hill-Sachs lesions and glenoid rim fractures (Table II). In addition, the sensitivity and specificity of the individual investigators were calculated (Fig. 2). As with the other calculations, the sensitivity for correct diagnosis of Hill-Sachs lesions was lowest, followed by that of glenoid rim fractures (Table III).
Discussion Not only patients with recurrent post-traumatic shoulder instability scheduled for revision surgery but also patients with first-time stabilization often present with large glenoid bone defects. Griffith et al,14 for example, investigating the prevalence of glenoid bone loss in anterior shoulder dislocation, described such glenoid bone loss in 41% of patients with a first-time unidirectional dislocation. In cases with recurrent dislocation, bone loss was described even in 86% of cases. For revision cases, Shah et al25 stated that failures might persist and could be attributed to poor preoperative as well as intraoperative judgment. This statement was also underlined by other authors looking for reasons for recurrence of instability after surgical shoulder stabilization.9,27 In this context, Sasyniuk et al24 found that the inter-rater reliability of 40% for orthopaedic shoulder surgeons’ arthroscopic assessment of intra-articular anatomy in patients with anterior shoulder instability was poor when rating the glenoid and the anterior inferior glenohumeral ligament. This might in part explain the origin of recurrences after surgical stabilization but also underlines the need for detailed preoperative radiologic examination. Therefore, it was our attempt to analyze whether misdiagnosis may already be a problem at the initial
Table II Comparison of CT versus radiography agreement based on 6 raters CT vs radiography before reduction AP view
Average agreement among 6 raters
SD
Other Greater tuberosity Glenoid Hill-Sachs lesion
97% 95% 74% 65%
10% 10% 24% 22%
treatment of these patients, which would be their initial hospital visit with the first-time traumatic shoulder dislocation. In this regard, the purpose of this investigation was to evaluate the reliability of the diagnosis of bony lesions about the shoulder joint at a first-time shoulder dislocation on conventional radiographs, which was controlled by a CT scan. For patients with manifest shoulder instability, Bushnell et al9 reported that even with a true AP view, an internal and external rotation view, a scapular-Y view, an axillary lateral view, a Stryker notch view, a West Point view, and an apical oblique (Garth) view, the sensitivity for detection of significant glenoid bone loss, which was defined as at least involving 25% of the articular surface, was just 50%. Apart from their limited sensitivity in detecting bony lesions, most of these views are usually just obtained to detect glenoid rim lesions in recurrent instability. Apart from this, in the setting of an acute first-time or recurrent shoulder dislocation, most of these images are difficult to obtain because of the required positioning of the patient’s arm. For this reason, the standard images in such cases are an AP view and, if possible, an axial view or a transscapular radiograph instead. In addition, the patient with a first-time shoulder dislocation might not always be directly transferred to a clinic specializing in shoulder surgery but will be treated in the nearest hospital, so training of physicians and radiologists in detecting such osseous lesions on common radiographs first treating the patient may vary. Moreover, Bigliani et al3 and Provencher et al22 stated that glenoid rim lesions appearing either as an acute fracture or a bony erosion could often be missed or inaccurately quantified on routine radiographs. Thus, we assumed that these standard images would be just as insufficient for detecting osseous lesions especially at the glenoid rim as the images described earlier by Bushnell et al.9 Salomonsson et al23 stated that a bony Bankart lesion would be a prognostic factor indicating a good functional result and a stable shoulder joint after a primary dislocation. We presume that this would only be the case when such a bony lesion is correctly diagnosed and thus adequately treated. In their investigation, they also reported that only 60% of glenoid rim fractures were detected on plain radiographs. Identification of such defects is essential, though, because the presence of such pathology has to be
Interobserver reliability diagnosing glenoid rim lesions
1511
Figure 2 Sensitivity and specificity of individual investigators judging 4 regions. Glen, Glenoid; Hill-S, Hill-Sachs lesion; Tub, greater tuberosity.
Table III Mean specificities and sensitivities of 6 raters with 95% confidence intervals Variable Glenoid Specificity Sensitivity Greater tuberosity Specificity Sensitivity Hill-Sachs lesion Specificity Sensitivity Other Specificity Sensitivity
Mean
95% Confidence interval
80% 67%
65%-96% 60%-73%
95% 94%
91%-100% 80%-100%
87% 20%
71%-100% 0%-53%
98% 50%
93%-100% NA
NA, not applicable.
taken into consideration when one is planning the patient’s treatment. The dislocation resistance of a once-dislocated shoulder joint is, according to Gerber and Nyffeler,12 significantly influenced by the presence or absence of a glenoid rim fracture or defect. Size15 and location28 of this pathology determine its individual effect on stability of the shoulder joint. When one takes these factors into account, surgical intervention, either by soft-tissue repair or osteosynthesis in the acute case, has to be considered a therapeutic option and may prevent further dislocations or even the later need for a reconstructive bone block procedure at the glenoid rim. Among the undiagnosed cases, the 4 glenoids with the largest fragments had a glenoid index of
0.46, 0.55, 0.61, and 0.70, respectively. In this context, Bushnell et al8 gave a nice overview citing various authors’ recommendations for bony reconstruction surgery for recurrent instability with a glenoid bone defect. These ranged from a 20% to 30% defect evaluated in biomechanical or clinical trials. Because the mentioned investigations applied different methods to evaluate the defect sizes,1-3,6,7,10,13,16,18,19,21,26,27 which value should generally be recommended as a cutoff remains unanswered so far. The 4 previously mentioned cases would indicate acute 54%, 45%, 39%, and 30% defects, which may not cause much discussion on which treatment to choose. The remaining 4 overlooked cases had a glenoid index ranging from 0.80 to 0.94. Defining the adequate treatment for these cases may lead to more controversy. Still, according to our clinical experience, even fragments smaller than the often attributed 21%16 and 25%1,3,6,7,18,21,26 should be reduced and internally fixed either arthroscopically or by an open procedure to reconstruct the glenoid cavity. In this context, we agree with Ochoa and Burkhart,20 who stated that it would be imperative to achieve an anatomic repair to restore normal biomechanics and the concavity-compression effect. On the long course, separation of such smaller fragments may also be the origin of greater defects constantly enlarged by erosion or compression because of recurring dislocations. Like us, other authors have also stated that radiography may misrepresent osseous lesions in dislocated shoulder joints.8,9 To our knowledge, this is the first investigation, though, to evaluate the interobserver reliability of diagnoses regarding osseous integrity after primary traumatic
1512 shoulder dislocation based on conventional radiographic images. The assumption that conventional radiographs alone would be insufficient for the reliable diagnosis of glenoid rim fractures in patients with a first-time traumatic shoulder dislocation was conclusively demonstrated. This is the reason we suggest performing a CT scan of the reduced shoulder after a first-time dislocation to not initially overlook a lesion that might be the source of recurrent instability.
Conclusion An osseous glenoid rim lesion after a traumatic shoulder dislocation left untreated can cause the onset of recurrent shoulder instability. According to the literature and our investigation, with plain radiographs alone, even osseous fragments that would need refixation are likely to be overlooked. Therefore, we suggest performing a CT scan of the shoulder after a first-time traumatic shoulder dislocation so that the correct treatment can be applied at the correct time.
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. Auffarth A, Schauer J, Matis N, Kofler B, Hitzl W, Resch H. The J-bone graft for anatomical glenoid reconstruction in recurrent posttraumatic anterior shoulder dislocation. Am J Sports Med 2008;36: 638-47. http://dx.doi.org/10.1177/0363546507309672 2. Bahk M, Keyurapan E, Tasaki A, Sauers EL, McFarland EG. Laxity testing of the shoulder: a review. Am J Sports Med 2007;35:131-44. http://dx.doi.org/10.1177/0363546506294570 3. Bigliani LU, Newton PM, Steinmann SP, Connor PM, McLlveen SJ. Glenoid rim lesions associated with recurrent anterior dislocation of the shoulder. Am J Sports Med 1998;26:41-5. 4. 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 5. Brophy RH, Marx RG. The treatment of traumatic anterior instability of the shoulder: nonoperative and surgical treatment. Arthroscopy 2009;25:298-304. http://dx.doi.org/10.1016/j.arthro.2008.12.007 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.
A. Auffarth et al. 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. Bushnell BD, Creighton RA, Herring MM. Bony instability of the shoulder. Arthroscopy 2008;24:1061-73. http://dx.doi.org/10.1016/j. arthro.2008.05.015 9. Bushnell BD, Creighton RA, Herring MM. The bony apprehension test for instability of the shoulder: a prospective pilot analysis. Arthroscopy 2008;24:974-82. http://dx.doi.org/10.1016/j.arthro.2008. 07.019 10. Chen AL, Hunt SA, Hawkins RJ, Zuckerman JD. Management of bone loss associated with recurrent anterior glenohumeral instability. Am J Sports Med 2005;33:912-25. http://dx.doi.org/10.1177/ 0363546505277074 11. 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 12. 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 13. Greis PE, Scuderi MG, Mohr A, Bachus KN, Burks RT. Glenohumeral articular contact areas and pressures following labral and osseous injury to the anteroinferior quadrant of the glenoid. J Shoulder Elbow Surg 2002;11:442-51. http://dx.doi.org/10.1067/mse.2002.124526 14. Griffith JF, Antonio GE, Yung PS, Wong EM, Yu AB, Ahuja AT, et al. Prevalence, pattern, and spectrum of glenoid bone loss in anterior shoulder dislocation: CT analysis of 218 patients. AJR Am J Roentgenol 2008;190:1247-54. http://dx.doi.org/10.2214/AJR.07.3009 15. Itoi E, Lee SB, Amrami KK, Wenger DE, An KN. Quantitative assessment of classic anteroinferior bony Bankart lesions by radiography and computed tomography. Am J Sports Med 2003;31:112-8. 16. 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. 17. Kirkley A, Werstine R, Ratjek A, Griffin S. Prospective randomized clinical trial comparing the effectiveness of immediate arthroscopic stabilization versus immobilization and rehabilitation in first traumatic anterior dislocations of the shoulder: long-term evaluation. Arthroscopy 2005;21:55-63. http://dx.doi.org/10.1016/j.arthro.2004.09.018 18. Lo IK, Parten PM, Burkhart SS. The inverted pear glenoid: an indicator of significant glenoid bone loss. Arthroscopy 2004;20:169-74. http://dx.doi.org/10.1016/j.arthro.2003.11.036 19. Mologne TS, Provencher MT, Menzel KA, Vachon TA, Dewing CB. Arthroscopic stabilization in patients with an inverted pear glenoid: results in patients with bone loss of the anterior glenoid. Am J Sports Med 2007;35:1276-83. http://dx.doi.org/10.1177/0363546507300262 20. Ochoa E Jr, Burkhart SS. Glenohumeral bone defects in the treatment of anterior shoulder instability. Instr Course Lect 2009;58:323-36. 21. Porcellini G, Campi F, Paladini P. Arthroscopic approach to acute bony Bankart lesion. Arthroscopy 2002;18:764-9. http://dx.doi.org/10. 1053/jars.2002.35266 22. Provencher MT, Bhatia S, Ghodadra NS, Grumet RC, Bach BR Jr, Dewing CB, et al. Recurrent shoulder instability: current concepts for evaluation and management of glenoid bone loss. J Bone Joint Surg Am 2010;92(Suppl 2):133-51. http://dx.doi.org/10.2106/JBJS.J.00906 23. Salomonsson B, von Heine A, Dahlborn M, Abbaszadegan H, Ahlstrom S, Dalen N, et al. Bony Bankart is a positive predictive factor after primary shoulder dislocation. Knee Surg Sports Traumatol Arthrosc 2010;18:1425-31. http://dx.doi.org/10.1007/s00167-0090998-3 24. Sasyniuk TM, Mohtadi NG, Hollinshead RM, Russell ML, Fick GH. The inter-rater reliability of shoulder arthroscopy. Arthroscopy 2007; 23:971-7. http://dx.doi.org/10.1016/j.arthro.2007.03.005 25. Shah AS, Karadsheh MS, Sekiya JK. Failure of operative treatment for glenohumeral instability: etiology and management. Arthroscopy 2011;27:681-94. http://dx.doi.org/10.1016/j.arthro.2010.11.057
Interobserver reliability diagnosing glenoid rim lesions 26. Sugaya H, Moriishi J, Kanisawa I, Tsuchiya A. Arthroscopic osseous Bankart repair for chronic recurrent traumatic anterior glenohumeral instability. J Bone Joint Surg Am 2005;87:1752-60. http://dx.doi.org/ 10.2106/JBJS.D.02204 27. Tauber M, Resch H, Forstner R, Raffl M, Schauer J. Reasons for failure after surgical repair of anterior shoulder instability.
1513 J Shoulder Elbow Surg 2004;13:279-85. http://dx.doi.org/10.1016/ S1058274604000254 28. 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