Geometric variables in anatomic replacement of the proximal humerus: how much prosthetic geometry is necessary?

BACKGROUND: Contemporary prosthetic designs for replacement of the proximal humerus offer a range of variable geometry aiming for restoration of the anatomy. This study evaluated five fictitious prosthetic scenarios looking to answer: How much prosthetic geometry is needed to achieve a minimum standard of anatomic reconstruction?

METHODS: Five prosthetic scenarios graduated in complexity in the number of inclination angles offered, number of head heights available and the number of offset positions of the head were compared to a database of known proximal humeral anatomy. Inclination ranged from a single inclination of 42 degrees to six inclination angles, head height ranged from a single head thickness to five possible head heights at any of four radii of curvature, and offset ranged from one to four variable positions that allowed for an increasing number of head positions relative to the stem. The five scenarios were compared to the database of normal anatomy using a computer optimization algorithm that aimed to minimize the displacement of the center of rotation and articular surface. For the purposes of this analysis, the minimum standard of reconstruction was defined as one that re-positioned the center of rotation within 4mm maintaining articular surface arc within 30 degrees for 95% of all specimens.

RESULTS: The mean displacement of the center of rotation and articular surface decreased with progressive increase in the geometric complexity of the prosthetic scenarios. However, the best fit observed in the entire analysis occurred in the simplest scenario with a single inclination angle (42 degrees) because the available prosthetic geometry happened to match one of the specimens. The worst fit also occurred in this scenario despite the specimen having a similar inclination angle (39 degrees) to that of the scenario due to differential offsets, which were more challenging for the algorithm than inclination. The most challenging anatomy for the computer algorithm to match was that with extreme offset of the head. The threshold 95% C.I. that was the stated goal of this study was achieved with the geometric complexity provided in the scenario with two inclination angles, three head heights (12 head sizes) and two offset positions of the taper mechanism.

CONCLUSION: In this analysis, even the simplest prosthetic scenario allowed for replication of normal anatomy within a range comparable to some existing third generation prosthetic systems. From a geometric perspective, surgeons should be able to re-approximate normal anatomy with a variety of implant configurations.