Accuracy of the glenohumeral subluxation index in nonpathologic shoulders.

Abstract

Correction of posterior humeral subluxation, measured by the humeral subluxation index (HSI) according to Walch, is necessary in total shoulder arthroplasty to prevent early loosening. The 3-dimensional (3D) measurement of the shoulder is becoming well accepted and common practice as it overcomes positional errors to which 2-dimensional (2D) glenohumeral measurements are prone. The first objective was to describe the HSI in a nonpathologic population with the 2D HSI according to Walch and a newly described 3D HSI method. The second objective was to compare both measuring methods with each other. In 151 nonpathologic shoulders, the 2D HSI was measured on the midaxial computed tomography scan cut of the scapula. The 3D HSI, based on the native glenoid plane, was defined as [formula in text], in which X is the projection of the center of the humeral head to the anteroposterior axis of the glenoid fossa and R is the radius of the humeral head. Both measuring methods were compared with each other. Correlation was determined. Interobserver and intraobserver reliability of the 3D HSI was measured. The mean 3D HSI (51.5% ± 2.7%) was significantly (P < .001) more posterior than the mean 2D HSI (48.7% ± 5.2%), with a mean difference of 2.9% ± 5.6%. No correlation was found between the 2D and 3D HSI. The interobserver and intraobserver reliability was excellent. The 2D HSI seems to underestimate the humeral subluxation compared with a 3D reliable equivalent.