Limited Predictive Value of the Instability Severity Index Score: Evaluation of 217 Consecutive Cases of Recurrent Anterior Shoulder Instability.

Purpose: To review the existing variables and their ability to predict recurrence of shoulder instability as it relates to the Instability Severity Index Score (ISIS), as well as evaluate any other pertinent imaging and patient history variables that may impact risk of recurrent anterior instability after arthroscopic Bankart repair.

Methods: All consecutive patients with recurrent anterior shoulder instability and who had arthroscopic instability repair were identified. Exclusion criteria were prior surgery on the shoulder, posterior or multidirectional instability, instability caused by seizure disorder, or a rotator cuff tear. All ISIS variables were recorded (age <20 years, sport type and level, hyperlaxity, Hill-Sachs on anteroposterior external rotation radiograph, loss of glenoid contour on anteroposterior radiograph), as well as additional variables: (1) number of instability events; (2) total time of instability; (3) glenoid bone loss (GBL) percent; and (4) Hill-Sachs measures (H/L/W/D/Volume). Postoperative outcomes were assessed based on the Western Ontario Shoulder Instability Index (WOSI), Single Assessment Numeric Evaluation (SANE) scores, and American Shoulder and Elbow Surgeons (ASES) scores, and recurrent anterior instability. Regression analysis was used to determine preoperative variables that predicted outcomes and failures.

Results: There were 217 consecutive patients (209 male patients [96.5%], 8 female patients [3.5%]) who met the inclusion criteria and were all treated with a primary arthroscopic shoulder stabilization during a 3.5-year period (2007-2011), with a mean follow-up time of 42 months (range, 26-58). The mean age at first instability event was 23.9 years (range, 16-48 years) and the mean cumulative ISIS score for the overall group was 3.6 (range, 1-6). Outcomes were improved from mean preoperative (WOSI = 1,050/2,100; ASES = 61.0; SANE = 52.5) to postoperative (WOSI = 305/2,100; ASES = 93.5; SANE = 95.5). A total of 11.5% (25/217) of patients had evidence of recurrent instability (subluxation or dislocation). Additionally, all 25 patients who failed postoperatively also had consistently inferior ASES, SANE, and WOSI outcome scores when compared with successfully treated patients. Factors associated with failure were GBL greater than 14.5% (P < .001), total time of instability symptoms greater than 3 months (P = .03), Hill-Sachs volume greater than 1.3 cm (P = .02), contact sports participation (P = .05), and age 20 years or younger (P < .01). There was no correlation in outcomes with Hill-Sachs on presence of glenoid contour loss on radiograph (P = .07), participation sports, or ISIS (mean = 3.4 success vs 3.9 failure, P > .05).

Conclusions: At a mean follow-up of 42 months was an 11.5% failure rate after arthroscopic Bankart stabilization surgery. This study shows no correlation between treatment outcome and the ISIS measure, given a mean score of 3.4 for the overall cohort with little difference identified in those who failed. However, several important parameters previously unidentified were detected including, GBL greater than 14.5%, Hill-Sachs volume greater than 1.3 cm, and duration of instability symptoms (>3 months). The ISIS may need to be redesigned to incorporate variables that more accurately portray the actual risk of failure after arthroscopic stabilization, including quantification of both glenoid and humeral head bone loss.

Level Of Evidence: III (Retrospective Case Series).