In order to compare the two femoral prostheses, we implanted them into synthetic (and later paired cadaveric) femurs. These femurs were then subjected to torsional mechanical testing.
This study compared a successful uncemented long-stem design with a shorter one. From shoulder to tip, the longer stems measured 152 mm, while the shorter ones were 100 mm. Besides the apparent difference in length, the shorter stem had a wider proximal section, and was reduced laterally to make insertion easier and minimise the risk of fracture of the greater trochanter. Both stems were fully hydroxyapatite-coated with 12/14 neck tapers and collars to prevent implantation past the required depth (Fig. 1). These stems also required different femoral preparations. The short-stem rasps were designed to be more bone-sparing by impacting loose bone, while the longer stem rasp was designed for more bone extraction.
Left-sided “medium”-sized synthetic composite femoral bones (Sawbone Model Number 1121; Sawbones Europe AB, Sweden) were used for their consistency of geometry and to aid a repeatable and controllable methodology (Fig. 2). These bones were dual density, with a foam polyurethane cortical shell. Bones from the same batch were used to avoid any inter-batch variation in mechanical properties. One synthetic bone was scanned by computed tomography (CT) to generate a digital three-dimensional (3D) model, which was later used for planning and validation of correct implant positioning.
A 3D surgical plan was made by one of the authors (S.C.) using the CT scan data from the synthetic bone, and 3D data files of the implants. Ideal positioning for each implant was determined based on alignment of the implant neck and head within the original bone (Fig. 1). From this data, the optimal position for the neck osteotomy and box chiselling entry point could be determined and planned.
Two 3D cutting guides (Embody, UK)—one for each femoral stem—were produced to ensure accuracy and repeatability of our osteotomy cuts and our box chiselling. These guides are designed to precisely match the surface anatomy of the bone (Fig. 2).
Use of these guides ensured that cutting and box chiselling of bone was restricted to areas pre-defined by the 3D planning. Subsequent reaming and rasping thus began in the correct location and planes.
We began by pinning the cutting guide to the specimen. The specimen-matched guide then directed the neck osteotomy and box chiselling of the femoral shaft (Fig. 2). Each bone was sequentially reamed and rasped according to the manufacturer’s instructions.
An experienced surgeon (J.P.C.) used standard intra-operative techniques to determine the appropriate implant size. A size 11 was used for the long, and a size 12 for the short stem. The prostheses were then inserted until seated.
The distal 18 cm of each femur were sawn off, and the implanted proximal femurs were potted in polymethylmethacrylate (PMMA) bone cement (within a metal cylinder). The cement was fixed to the cylinder with three screws to prevent rotation and left for 30 min to cure.
The metal cylinder was mounted to the base of a servohydraulic testing machine (Instron 8874 Biaxial Testing System; Instron Corporation, MA, USA) using a bespoke adjustable vice. The potted bone was aligned such that the plane of the femoral stem was vertical, and directly underneath the centre of the servohydraulic crosshead. A 6-mm hex key was attached to the crosshead and lowered into the 6-mm hex hole in the implant (this hole is aligned with the centre of the distal femoral stem). This allowed the stem to be rotated about its central axis (Fig. 3).
Throughout the testing a small constant vertical load of 10 N was applied, to counteract any vertical loosening, and to ensure engagement of the hex key in the implant hex hole. Before each test, the Instron crosshead was manually positioned in a neutral position, fully engaged with the implant but with no vertical or rotational force.
To test resistance to femoral fracture, the implant was rotated clockwise through 90° in 1 s. This testing protocol has been described previously , and is designed to simulate peri-prosthetic fracture due to internal rotation on a planted foot, as might occur during stumbling.
Torque, rotation, vertical load and vertical position data were sampled 50 times per second throughout the testing protocols, and were exported to a data spreadsheet file (Microsoft Excel; Microsoft Corporation, WA, USA).
Following ethical approval, a single pair of cadaveric femurs were extracted from an embalmed cadaver donated to the Human Anatomy Unit (Charing Cross Hospital, London, UK). The cadaver had been embalmed with a mixture of formaldehyde, phenol, polyethylene glycol and alcohol, which has been shown not to significantly affect the stiffness of bone .
An experienced surgeon used a posterior approach and standard intra-operative techniques to implant and size the short and long femoral stems. The femurs were then carefully dissected from the cadavers and stripped of soft tissues.
The implanted femurs underwent the same experimental setup as the synthetic bones. Testing was in a clockwise direction on the left, and anticlockwise on the right femur to ensure both hips were torqued in internal rotation. The data was analysed using SPSS (IBM SPSS Statistics, version 20) using a Mann–Whitney U test as data was not found to be parametric.