Method for the location of primary wear scars from retrieved metal on metal hip replacements

Background Retrieved metal-on-metal acetabular cups are valuable resources in investigating the wear behaviour of failed hip implants, but adequate methods to do so are lacking. To further contribute to addressing this issue, we developed a method to detect the in vivo location of the primary wear scar of an explanted cup. Methods We proposed a new method in which thirteen patients with failed metal hip resurfacings were recruited, and their acetabular components retrieved. A 3D wear map was generated and the precise location of the primary wear scar in each cup was identified using a coordinate measuring machine. This wear scar location was noted in relation to the features on the acetabular cup. Having identified the location of the wear scar, this 3D positional map was co-registered to the implant on the patient’s pelvic 3D CT scan. Results Using our proposed technique, we were able to demonstrate that the in vivo position of the primary wear scar in explanted metal acetabular cups can be variable. Conclusions This method has utilised existing techniques to better understand the three-dimensional properties of wear behaviour, and may be a method which can be used in further studies to investigate variables that affect the position of the primary wear scar.


Background
Metal-on-metal (MoM) hip implants proved to be popular, but had a high failure rate, and lead to a dangerous level of metal ions into the blood stream [1][2][3][4]. With patient factors a very important consideration, the pattern of failure due to biomechanical implications as a result of the patient's anatomy must be examined [5,6]. Pin pointing the in vivo location of the primary wear scar (WS 1 ) may allow us to correlate a number of factors to help ascertain the tribology behind qualitative patterns of wear.
Limited work has been previously done to identify the location of the WS 1 in explanted acetabular components of a MoM hip replacement with its in vivo location. Published work exploring the significance of wear scar location has so far been limited to an in vitro study [7]. In this study, Angadji et al. used schematic drawings to map two-dimensional acetabular wear scar locations. However as this study was done using a hip joint simulator, there was no method that described the process of co registering the WS 1 with its in vivo location.
This limited previous work highlights the methodology is missing in this discipline, and our study aimed to address this gap in both technique and knowledge of in vivo wear scar location. Ours was a study aimed to develop a novel technique which allowed the mapping of a primary wear scar with its 3D pre-explantation in vivo location.

Methods
As the purpose of this paper was to devise a novel technique that could use explanted acetabular cups to map the WS 1 with its pre-explanation in vivo location, the method consisted of two main stages: retrieval analysis and 3D CT co-registration.

Patients
CT scans for thirteen patients (Table 1) who had undergone a revision of their hip resurfacing that had subsequently failed was gathered. The two inclusion criteria were that the acetabular component needed to be available for physical analysis and that pre-revision CT scans of the patients were required. The mean age of the patients at the time of failure was 51.8 years and the mean period of between implant insertion and failure was 47.5 months. Twelve patients had been fitted with a Birmingham Hip Replacement (BHR) system. One patient had a Cormet system. Four patients had an implant on their left hip. The remaining nine patients had an implant on their right hip. One patient had a bilateral BHR but only the right hip was used for the purposes of this study. The internal diameter ranged from 42-54mm (Table 1).
Informed written consent for this study was obtained from all patients.

Retrieval analysis
To determine the location of the wear scar, each of the bearing surfaces of the cup were measured using a Zeiss Prismo (Carl Zeiss Ltd, Rugby, UK) coordinate measuring machine (CMM). Previously published protocols [8] were used to take up to 300 000 unique data points on the surface by translating a 2mm ruby stylus along 400 polar scan lines. An iterative least square fitting method was used to analyse the raw data, allowing a 3D wear map of the cup under study to be generated (Fig. 1). An unworn cup should have had a uniform depth throughout the cup, and any areas of increased depth reflected locations of wear patches. The WS 1 could be visualised according to their location according to concentric zone (Fig. 2). The location of WS 1 was also noted in terms of degrees clockwise from the acetabular features (Fig. 3) which were used as markers for rotational identification.

3D CT and co-registration
A low dose 3DCT protocol was followed [9], generating an image of the pelvis. The in vivo location of WS 1 could be co-registered on this image as the location of the WS 1 was known in relation to the acetabular cup rim and also in relation to the acetabular features. First, the acetabular features were located on the 3D CT image. Next, the angle in degrees between these fins and the WS 1 was measured using the appropriate measuring tools. Finally, the vertical distance of the WS 1 from the acetabular cup rim was measured. This gave us the epicentre of the WS 1 and hence, the location of the WS 1 of an explanted acetabular cup could be co-registered with its in vivo location (Fig. 4).
Acetabular fins are intended add to the component's stability [10], but in this case they could be used as a

Results
Using the method which was discussed the WS 1 was identified for all 13 components. The out-of-roundness machine was able to take circular measurements at 0.1°i ntervals around the acetabular cup rim and also in 0.5mm increments down from the acetabular cup rim. The location of the acetabular features was also noted successfully in all cases, allowing the co-registration of the in vivo location of the WS 1 to be visualised. A diagram was then produced showing a 3D representation the location of each of the thirteen WS 1 on a single acetabular cup. This allowed us to appreciate the range of their locations, with reference to the anterior pelvic plane, according to quadrants (Fig. 5).

Discussion
The distribution of the location of the in vivo WS 1 can be reported as follows: six were in the upper outer quadrant; three were in the lower outer quadrant; one was in the upper inner quadrant; and three were in the lower inner quadrant. These results clearly show that the location of the in vivo WS 1 can be very variable, which does not fit with the current theory of edge loading [6,7]. Further work can be done to correlate a range of factors with the location of WS 1 , and as such, the method devised here may be of huge importance. We did not encounter any components where this technique could not be used. This is a novel method because it allows us to use an explanted component and work retrospectively to coregister the WS 1 from this explanted component with what would have been its in vivo location. Mapping of wear scars of large numbers of failed components, and hence the application of our technique, may help us better investigate the qualitative three-dimensional properties of wear behaviour and correlate the location of wear with a number of variables. Knowledge of in vivo wear The clinical value of the information obtained from this method will be of interest to surgeons who can suggest a more accurate patient-based prediction of implant longevity pre-revision surgery. Furthermore, it will be of use in the context of explaining wear behaviour of hip implants because the effect and importance of many variables are currently unknown. This is because of the limited investigations gathered out with respect to the qualitative properties of hip wear, such as primary wear scar location.
The mean period of implant failure is just under four years for our patient sample. This short length of time between implantation and failure may be explained by a selection bias, as we did not include patients whose implants had not failed, even though the implant may have had a substantial wear scar. Hence, an interesting aspect that has been unexplored in this study is whether there is a correlation between the period of implant insertion and the magnitude of the wear scar, This was one limitation of the study and in future, a method may be devised that allows us to visualise the in vivo primary wear scar pre-failure.

Conclusions
We have developed a method which can be used as part of future mechanical and tribological studies looking into the failure of MoM hip implants. Future studies should investigate variables that affect position of the WS 1 , and our technique may be integral to these studies. Examples of variables could include component position, pelvic tilt, gender and joint reaction force. In doing so, we may be able to better understand the implant, surgical and patient factors that lead to implant failure and use this knowledge to predict the suitability of patients to certain joint replacement procedures and its prognosis.