Arthroscopic Bankart repair with and without remplissage in longer inferior craniocaudal Hill-Sachs extensions: secondary analysis of a randomized clinical trial.

Abstract

Arthroscopic remplissage with arthroscopic Bankart repair (ABR) may be considered for those with Hill-Sachs (HSL) and Bankart lesions. This treatment has a reported postoperative instability rate between 3.2% and 13.0%. Remplissage has demonstrated improved outcomes in patients with medially extended HSL, but the relevance of inferior craniocaudal Hill-Sachs extension (IC-HSE) in postoperative recurrent instability and the benefits of remplissage have not been thoroughly examined. The main purpose of this study was to compare ABR with and without remplissage with respect to postoperative recurrent instability and clinical outcomes in patients with and without IC-HSE. Data for the present study were collected as part of a previous randomized clinical trial comparing ABR (NO REMP) vs. ABR with remplissage (REMP) in the treatment of anterior shoulder instability with HSL. The original trial was double-blinded at 2 centers involving 6 surgeons. Patients were more than 13 years old with recurrent traumatic anterior shoulder instability with an engaging HSL of any size with an available preoperative magnetic resonance imaging (MRI) or computed tomographic (CT) scan. The extent of the IC-HSE was determined by measuring the lower-edge angle (LEA) in the sagittal-oblique plane. Patients were categorized as having an LEA of greater than or less than 90°. The primary outcome was recurrence of instability defined as at least 1 episode of redislocation after surgery or a minimum of 2 subluxations occurring at least 12 weeks postoperatively. Odds ratios were generated for postoperative recurrent instability based on an LEA less than or greater than 90°, in patients who underwent a Bankart repair with or without remplissage. One hundred two patients (50 NO REMP, 52 REMP) were included in the original trial, of which 95 patients (47 NO REMP, 48 REMP) had preoperative MRI (n = 94) or CT (n = 1) available. In those patients with LEA less than 90°, the odds of recurrent instability in NO REMP (n = 3/17; 17.6%) were 1.21 (95% confidence interval [CI] 0.211, 6.985; P = .828) compared with REMP (n = 3/20; 15.0%). In those with LEA greater than 90°, the odds of recurrent instability in NO REMP (n = 11/30; 36.7%) were 7.53 (95% CI 1.492, 37.978; P = .015) compared with REMP (n = 2/28; 7.1%). In cases with inferior extension of the HSL, where LEA was greater than 90°, remplissage significantly reduced the rate of recurrent instability compared with isolated ABR, whereas the remplissage did not significantly affect the rate of recurrent instability if the LEA was less than 90°.