Individualizing glenoid lateralization to optimize range of motion after reverse shoulder arthroplasty with the Arthrex Univers reverse: a virtual assessment of the influence of scapular neck length using a statistical shape model.

Abstract

Preoperative scapular neck length (SNL) varies widely due to anatomical factors and pathologic glenoid wear. A short SNL can lead to early impingement, decreased range of motion (ROM), and increased rates of scapular notching after reverse shoulder arthroplasty (rTSA). Glenoid-sided implant lateralization can avoid these issues, but it is unclear how much is necessary for varying SNLs. The objective was to use a statistical shape model, varying SNL, to evaluate impingement-free rotational ROM across different glenoid positions and lateralizations. A total of 100 scans were randomly chosen from a clinical database of over 10,000 shoulder computed tomography scans uploaded for shoulder arthroplasty preoperative planning. These 100 scans were utilized to create and validate a statistical shape model. Modes corresponding to scapular size and scapular neck angle were identified as producing the greatest variance in SNL, from which the mean SNL and 2 standard deviations (positive and negative) were evaluated for the study. For each scapula, a single glenosphere diameter (33-42 mm) was selected by a consensus of 3 surgeons. A 135° NSA inlay humeral prosthesis was utilized for all simulations (Arthrex Univers rTSA). Impingement-free rotational ROM was then assessed for each of the 5 scapulae for the following virtual implantation variables: lateralization (0-12 mm), inferior glenosphere eccentricity (0-2.5 mm), and posterior glenosphere eccentricity (0-2.5 mm). Both SNL and glenoid lateralization had significant contributions to rotational ROM. Progressive glenoid lateralization improved external rotation at 0° of abduction (ER0), external rotation at 60° of abduction (ER60), and internal rotation at 60° (IR60) across all 5 standard deviations of SNL. IR60 was greatest for the longest SNL, and the value of progressive glenoid lateralization was the greatest for improving IR60 for the shortest SNL, where 6 mm of lateralization achieved maximum IR60. ER0 was greatest for the shortest SNL, and the value of glenoid lateralization for ER0 was not seen for the longest SNL until >4 mm of lateralization. ER60 was largely not influenced by lateralization or SNL, where 2-4 mm of lateralization maximized sufficiently. Both SNL and glenoid implant lateralization influence rotational ROM after rTSA with a 135° NSA and should be considered as covariates when determining appropriate implant selection and positioning. For smaller SNL, 6 mm of implant lateralization is needed to maximize impingement-free internal rotation. For larger SNL, a minimum 6 mm of implant lateralization is needed to maximize impingement-free external rotation.