By: 1 August 2008


Anatomic reconstruction of humeral retroversion in complex proximal humeral fractures had been considered difficult to determine primarily because of the involvement of the humeral head articular surface and the tuberosities which disrupts the important landmarks which are commonly used to make this determination. When hemiarthroplasty is performed, a retroversion in the range of 20 to 40 degrees is chosen but anatomic studies have clearly indicated that humeral retroversion varies significantly from one individual to another (range 5 to 60 degrees). Therefore, in proximal humeral fractures, the position of retroversion chosen is quite unlikely to reproduce the anatomy of the specific humerus being treated because of the absence of intact anatomic landmarks.

Recent studies1 showed that the functional consequences of excessive retroversion of a humeral prosthesis in the treatment of proximal humeral fractures has been underestimated. Correct positioning of the greater tuberosity in the horizontal plane is not possible if the prosthesis is implanted with excessive retroversion. In this situation, additional internal rotation of the arm produces increases stress on the greater tuberosity with the potential to result in migration, detachment and nonunion and a poor functional outcome. To reduce the incidence of such complications, new tools and techniques are necessary to facilitate a more reliable tuberosity reconstruction.

The purpose of this study was:1 to identify the three-dimensional shape of the bicipital groove from proximal to distal and2 to determine if it could be used to identify anatomic humeral retroversion in the presence of a proximal humerus fracture.

Materials and Methods

The anatomy laboratory at the Bordeaux School of Medicine (France) provided 49 dried humeri; A three-dimensional coordinate measuring machine was used to digitize the proximal humeral geometry relative to defined reference axes. The anterior and lateral offsets of the bicipital groove (BG) and humeral head retroversion (HHR) were calculated relative to these axes using prepackaged CMM software.

The reference axes were created by marking points along the anatomic neck to establish a plane and at the junction of the humeral shaft and proximal metaphysis to define the proximal portion of the humeral cylinder. The humeral head equatorial plane (HHEP) is defined as a plane perpendicular to the anatomic neck plane that passes through the inferior and superior points of the articular surface. The CMM identified a plurality of points on the articular surface, along the HHEP. These points were bisecting the head in the medial-lateral direction. This method was derived from Hertel et al.6. Finally, 30 points were identified with the CMM between the superior and inferior edges of the proximal humeral cylinder. The axis of this cylinder approximated the intramedullary (IM) axis.

As depicted in Figure 1, the first axis corresponds to the IM axis of the humerus. The second axis is perpendicular to the first and is oriented parallel to the HHEP. The third axis is self-defined by an orthonormal referential. The origin of these axes is located at the intersection of the IM axis and a line perpendicular to the plane of the anatomic neck at the center of the humeral head.

Similar to the technique to describe the location of the humeral head, linear dimensions were used to quantify the position of the BG relative to the IM axis. The anterior offset of the bicipital groove (BGAO) was defined as the shortest distance in the transverse plane between the BG and the IM axis along the anterior-posterior axis (i.e. Z axis). Similarly, the lateral offset of the bicipital groove (BGLO) was defined as the shortest distance in the transverse plane between the BG and the IM axis along the medial-lateral axis (i.e. Y axis). The vertical distance (i.e. X axis) was used to determine the length of the BG.

Figure 2: Definition of sites for bicipital groove palpation using the ruby probe of a CMM

The location of the BG was digitized with a CMM at 4 levels: from its proximal extent (H1) to its distal extent (H4) and at two symmetric points (H2 and H3) in between. The distal portion was defined as the level where the BG becomes nearly flat and its center difficult to define. This level is approximately 10 mm below the most inferior aspect of the articular surface. Since the location of the distal slice (H4) varies between bones, the authors introduced a fifth level (Ha), which is calculated from the measured data to be exactly at the level of the most inferior aspect of the articular surface. This level is assumed to be the location of the surgical neck, a probable location of fracture.

Then the data was used to measure HHR relative to two different anatomic landmarks: 1) the epicondylar axis (HHREP) and 2) the bicipital groove (HHRBG). HHR relative to the epicondylar axis (HHREP anatomic) was defined as the angle between the epicondylar axis and the HH4. HHR relative to the bicipital groove (HHRBG anatomic) was defined as the angle between the plane passing through both the IM axis and the center of the BG at the level Ha and the HHEP. HHRBG anatomic was calculated at the level of Ha using the following equation:

HHRBG anatomic = Arctan (BGAOa / BGLOa) Equation 1

In order to apply the results of the anatomic study to design a novel fracture prosthesis, the relationship between the BG shape and IM canal diameter needed to be established. To this end, AP and ML radiographs were obtained. An orthopaedic surgeon (PHF) templated and sorted the radiographs based upon the optimal fit of three specific diameter stems. The stem diameters utilized were: 7 mm, 9.5 mm, and 12 mm; each prosthesis had a lateral fin offset in the anterior direction by the mean value of BGAO.

In order to evaluate the reliability of the prosthesis to reconstruct HHR, a computational analysis was performed. This analysis calculated the angular error associated with using the mean retroversion value (relative to each landmark) to predict the anatomic HHR for each humerus measured in this study. Using Equation 2;