A 3D comparison of humeral head retroversion by sex and measurement technique

Accurate humeral head reconstruction during shoulder arthroplasty is partially dependent on correctly estimating and replicating native version. The present study evaluated the effects of sex and measurement technique on three-dimensional (3D) humeral version measurements made using the transepicondylar, forearm and flexion–extension axes. Fifty-two full-arm computed tomography scans were converted to 3D models and geometry extracted to define landmarks and coordinate systems. An anatomic humeral head osteotomy plane was used to measure version relative to the three measurement techniques and compare between sexes. The measurement technique used had a significant affect (p？<？0.001) on the resulting version measurement. The forearm axis technique consistently resulted in higher measured version compared to either the flexion–extension [mean (SD) males 9° (4°), females 13° (5°), p？<？0.001] or the transepicondylar axes [mean (SD) males 8° (4°), females 11° (4°), p？<？0.001]. Version in males was 7° greater than females when referencing either the flexion–extension [p？=？0.029; mean (SD) males 37.7° (11°), females 30.4° (13°)] or transepicondylar axes [p？=？0.045; mean (SD) males 39° (11°), females 32° (12°)]. The choice of measurement technique can affect the humeral version angle. These results are important because measuring version using the epicondyles pre-operatively, and subsequently the forearm intra-operatively, will result in approximately 10° under-retroverted osteotomy. For example, 0° neutral version cut during reverse arthroplasty measured referencing the forearm results in 10° anteverted osteotomy when referencing the distal humerus