Reliability of the ISAKOS Modification to Subclassify Rockwood Type III Acromioclavicular Joint Injuries

BACKGROUND: To improve clinical management, the International Society of Arthroscopy, Knee Surgery and Orthopaedic Sports Medicine (ISAKOS) developed a specific subclassification of Rockwood type III acromioclavicular (AC) injuries: IIIA (no overriding distal clavicle) and IIIB (overriding distal clavicle). PURPOSE/HYPOTHESIS: The study aimed to determine the inter- and intraobserver reliability of the radiographic classification proposed by ISAKOS for AC injuries. It was hypothesized that the strength of agreement for the ISAKOS modification will be substantial to almost perfect. STUDY DESIGN: Cohort study (diagnosis); Level of evidence, 3. METHODS: We evaluated 40 radiographs of all types of AC joint injuries from patients at a single institution. The images were distributed to 6 shoulder and elbow fellowship–trained orthopaedic surgeons, along with standardized assessment questionnaires. The evaluators measured the bilateral coracoclavicular distance and the coracoclavicular ratio and classified the severity of the injuries according to the modified ISAKOS Rockwood classification. Four of the surgeons repeated the evaluation 6 weeks later to calculate intraobserver agreement. The kappa (κ) statistic was calculated for categorical inter- and intraobserver reliability. Intraclass correlation coefficient (ICC) estimates were calculated for the reliability of the coracoclavicular distance measurement. A 2-tailed test was performed to assess statistical significance. RESULTS: Overall interobserver agreement was substantial (κ = 0.637; 95% CI, 0.595-0.680) among the 6 evaluators, and intraobserver agreement was also substantial among the 4 evaluators (κ = 0.616; 95% CI, 0.549-0.638). The interobserver agreement for evaluating types IIIA and IIIB was fair (κ = 0.215; 95% CI, 0.135-0.295) and moderate (κ = 0.473; 95% CI, 0.393-0.553), respectively. The agreement on the coracoclavicular distance measurements was excellent among the 6 evaluators on both the affected side (ICC, 0.982; 95% CI, 0.970-0.990) and the unaffected side (ICC, 0.930; 95% CI, 0.894-0.958). CONCLUSION: Substantial agreement was found when categorizing AC joint injuries using the ISAKOS modification of the Rockwood classification, with excellent reliability demonstrated for the quantitative assessment of vertical displacement of the clavicle. Visual examination was unreliable in differentiating type IIIA injuries from type IIIB injuries.