CORR Insights®: Radiostereometric Analysis Permits In Vivo Measurement of Very Small Levels of Wear in TKA

Where Are We Now?

Measuring polyethylene wear is challenging because of the complex geometry and kinematics of knee implants [5]. While several investigators have demonstrated that simple measurements of the minimum distance between the tibial baseplate and femoral component can be used to quantify linear wear [1, 2], the radiographic images require fluoroscopic guidance or carefully controlled patient positioning to ensure that the superior surface of the tibial baseplate is aligned with the radiograph’s beam [13].

In the current study, Teeter and colleagues [10] describe their experience using model-based radiostereometric analysis (RSA) to characterize TKA insert wear. Unlike traditional marker-based RSA techniques that rely on small spherical beads (with simple, precise geometries), they use the three-dimensional (3-D) geometries of the implant components to reconstruct implant position [4, 11] and measure linear penetration of the femur into the polyethylene at different knee flexion angles. While these measurements include both polyethylene material removal and deformation (creep/cold flow), they enable a quantitative assessment of medial and lateral compartment penetration without having to insert beads in the patient or implant. As opposed to measurements based on a single knee angle, characterizing penetration over a patient’s ROM is particularly useful since the patterns of knee wear can vary considerably among patients, with some retrieved components demonstrating maximum wear in the central portions of the condyles while others show wear that is more pronounced near the perimeter of the insert.

In their paper, Teeter and colleagues apply their technique to 49 well-functioning knees at median 12-year followup to evaluate in vivo wear rates (calculated by dividing penetration by in vivo time) and explore how demographic factors and postoperative limb alignment are associated with wear [10]. Some of their findings, such as increased median wear rates in the medial compartment with increasing varus, confirm prevailing knowledge. Other findings, including the absence of a relationship between patient age and linear wear rates, challenge conventional wisdom.

In contrast to knee inserts sterilized with gamma irradiation that have generally demonstrated mean linear wear rates in the range of 0.10 to 0.15 mm/year [1, 6], Teeter and colleagues report median values ranging from 0.025 to 0.05 mm/year for compression-molded inserts terminally sterilized with ethylene oxide. Since these components were never exposed to gamma irradiation, they should not have free radicals or any polyethylene crosslinking. In view of the potential for long-term complications associated with wear, the rates reported by the study authors are reassuringly low and suggest that the new generation of crosslinked polyethylene for knees may only provide a modest reduction in wear (since the wear rates for noncrosslinked polyethylene without free radicals already appear to be close to zero). As a consequence, it may take a decade or longer to see any clinical benefit from crosslinked polyethylene in knees.

Where Do We Need To Go?

To limit the number of confounding factors, Teeter and colleagues [10] confined their study to the Genesis II^TM (Smith & Nephew, Memphis, TN, USA), but applying the model-based RSA techniques to other implant designs and contemporary polyethylenes would provide complementary data to corroborate their findings. Generating larger datasets would also be useful to further validate their methods. Model-based RSA techniques have measurement uncertainties on the order of 0.1 to 0.2 mm [7, 14]. At 10-year followup, these measurement uncertainties could account for 0.01 to 0.02 mm/year of observed wear. While these uncertainties are relatively low, they are not inconsequential in view of the median linear wear rates reported by in the current study. Additionally, manufacturing tolerances can contribute to variations in initial polyethylene insert thickness which adds another component of measurement uncertainty [8]. While Teeter and colleagues used radiographic images from a single time point, immediate postoperative images could serve as a valuable baseline for future studies. In the absence of immediate postoperative images, serial radiographic images obtained at different time points can be used to evaluate the temporal pattern of polyethylene wear. Even in the presence of measurement uncertainty at each followup evaluation, fitting a line or curve to multiple measurements obtained at different intervals can reduce the uncertainty associated with the wear rate and facilitate analysis of how wear varies over time [1]. Examining how serial wear measurements deviate from an underlying trend can also provide a practical assessment of clinical measurement uncertainty that would complement the laboratory work that has been done to validate model-based RSA techniques in carefully controlled environments [7, 14]. Additional validation could also be undertaken by imaging a patient just prior to revision and then analyzing the revised insert after surgery to confirm the radiographic measurements.

In theory, linear measurements over the ROM could be combined with 3-D implant geometries to evaluate volumetric wear. While volumetric wear measurements have been made using microcomputed tomography (Fig. 1), it remains to be seen if the resolution of clinical radiographs combined with implant manufacturing tolerances and the measurement uncertainties associated with contemporary wear analysis techniques will allow accurate volumetric wear assessments [12]. As a consequence, robust validation of volumetric wear accuracy would be essential since linear measurement uncertainties on the order of 0.1 to 0.2 mm could contribute to volumetric uncertainties of 10 to 20 cubic millimeters for each square centimeter of contact area. Even without the ability to accurately determine volumetric wear, component position data generated over patients’ ROM would enable evaluation of implant kinematics and this may prove to be a valuable direction for future research [9].

Fig. 1.

A 3-D reconstruction derived from microcomputed tomography images can be used to evaluate the volumetric wear of polyethylene knee inserts. The linear displacement data obtained using a patient’s ROM with model-based RSA could also be used to compute volumetric wear. Although this method still needs thorough validation and could not capture polyethylene deformation or focal damage (like the pitting demonstrated on the worn insert), it would allow multiple linear measurements derived from ROM radiographs to be summarized on a single image. (Published with permission from Robert H. Hopper Jr. PhD on behalf of the Anderson Orthopaedic Research Institute).

How Do We Get There?

In the immediate future, larger datasets derived from different types of implant systems would enable robust multivariate analyses to discern how multiple factors interact to influence wear and which are most important. Although clinical radiographic systems that can simultaneously obtain radiographic images in different planes are now commercially available, the need for biplanar images to make accurate measurements coupled with calibration requirements and the time-intensive analytical methods associated with RSA likely mean that the measurement of knee wear will remain a research pursuit rather than a practical clinical tool for the immediate future. Nevertheless, if libraries of implant geometries become available and model-based RSA software continues to evolve, measurement of knee wear from clinical radiographs may become more widespread in the same way that hip implant wear has been measured using commercially-available computer-assisted techniques [3]. Yet even as it becomes technologically feasible to measure knee wear from clinical radiographs, the relatively low wear rates associated with contemporary implants may render subtle differences of little interest to clinicians. Although the importance of measuring knee wear may diminish, implant kinematics will likely remain relevant as the orthopaedic community attempts to determine how specific implant designs and surgical techniques influence patient outcome. While only a few centers are currently working to accurately quantify knee wear, the progress made to date has created a new set of tools to better quantify patient outcome and offer additional insight into what contributes to a well-functioning knee replacement over the long term. While it is not unusual for hip replacement patients to forget they have an artificial joint a few years after surgery, the same is not true for knee replacement. By combining implant wear and kinematics with patient-reported outcomes, functional assessments and radiographic evaluation of component loosening and osteolysis, it will be possible to determine how specific implant designs, polyethylene manufacturing processes, surgical techniques and patient characteristics are related to outcome. Teeter and colleagues should be commended for their ongoing efforts to help realize this goal.