Reliability and accuracy of templating the proximal humeral component for shoulder arthroplasty

The purpose of this paper is to examine the intra-observer and inter-observer reliability of preoperative templating to help determine its usefulness for approximating proximal humeral stem and head size in shoulder arthroplasty.

Thirty-one nonconsecutive patients underwent uncemented total shoulder arthroplasty using 1 of 2 commonly used prostheses. Templating was performed by 4 independent observers (2 experienced shoulder surgeons and 2 PGY-3 residents). Two readings were performed with a minimum 2-week interval between templating sessions. Templating was performed in a blinded manner using appropriately magnified templates provided by the prosthesis manufacturer. Postoperative radiographs were analyzed in the same fashion and each observer graded qualitative stem and head size radiographically. Pre-operative templated sizes were compared to actual implant sizes obtained from dictated operative reports.

There was substantial (kappa > 0.7) intraobserver reliability for stem and head size selection for both prostheses. Interobserver reliability was moderate for stem size (kappa = 0.53) and substantial (kappa = 0.66) for head size. Preoperative templating accurately predicted stem (84-95%) and head (44-66%) size within 1 size variation (stem width, head thickness, or diameter) for shoulder replacement surgery. There was no significant difference in accuracy between attendings and residents when templating for stem or head size.

Preoperative templating is a reliable and reasonably accurate step when planning shoulder replacement surgery. Templating for stem size is more accurate than for head size.