Medial elbow exposure: an anatomic comparison of 5 approaches

Purpose: Several surgical approaches to the medial elbow are described; however, it remains unclear which exposure provides the optimal view of relevant medial elbow structures. The purpose of this anatomic study was to determine the visible surface area of the coronoid process, distal humerus, and radial head through 5 approaches to the medial elbow.

Methods: Eight fresh-frozen cadaveric upper extremity specimens were dissected. Five surgical approaches were performed on each specimen. The Smith muscle-splitting approach to the ulnar collateral ligament was performed first (Smith), followed by the Hotchkiss medial "Over the top" approach (Hotchkiss), the extended medial elbow approach (EMEA), the flexor carpi ulnaris splitting approach (FCU-Split), and the Taylor and Scham approach (T&S). Bony visualization was determined using laser surface scanning (Artec Space Spider; Artec 3D). The scans were segmented using commercially available digital software (Geomagic Wrap; 3D Systems Corporation), and the surface area visualized was determined. A descriptive analysis of the joint areas visible using the medial collateral ligament (MCL) as a clinical landmark was performed.

Results: The EMEA visualized the highest proportion of the total elbow joint from the medial side showing 13.9 ± 6.0 cm², or 15% ± 4% of the joint. It also provided the best visualization of the coronoid (3.2 ± 1.7 cm² of surface area, or 26% ± 9%) and distal humerus (9.9 ± 4.3 cm², or 15% ± 4%). The Hotchkiss approach was best at visualizing the radial head (0.8 ± 0.3 cm², or 7% ± 3%). The EMEA, Hotchkiss, and Smith approaches showed primarily the anterior bundle of the MCL, its insertion, and the regions anterior to it, whereas the FCU-Split showed the anterior bundle of the MCL and regions both anterior and posterior to it. The T&S showed primarily the areas posterior to the anterior bundle of the MCL; the anterior regions were not visible. The FCU-Split and the T&S allowed visualization of the posterior bundle of the MCL. The intraclass correlation coefficients (ICCs) for intraobserver reliability were 0.997, 0.992, and 0.974 for the test distal humerus, test coronoid, and test radial head, respectively. The ICCs for interobserver reliability were 0.915 for the test distal humerus, 0.66 for the coronoid, and 0.583 for the radial head.

Conclusion: The EMEA provides the most visualization of the coronoid and distal humerus, whereas the Hotchkiss showed the most radial head. However, these approaches mainly visualize structures anterior to the MCL. If exposure of structures posterior to the MCL is required, the FCU-Split and T&S approaches are more appropriate.