Abstract
Introduction
Failure of total knee replacements due to the generation of polyethylene wear debris remains a crucial issue in orthopedics. Unlike the hip, it is difficult to accurately determine knee implant wear rates from retrieved components. Several studies have relied on thickness measurements to estimate penetration, but the complicated geometry of contemporary tibial liners poses a challenge to accurately assess wear. In this study we address the question whether linear penetration can serve as a surrogate measure for volumetric material loss.
Methods
Eighty-one retrieved UHMWPE NexGen cruciate-retaining tibial liners (Zimmer, Warsaw, IN) with an average time in situ of 5.27±2.89 years were included in the study. Metrology data for the surfaces of the tibial liners were obtained with a coordinate measuring machine (OGP, Rochester, NY). Using a laser scanner with two micrometer depth accuracy, at least 400,000 measurement points were taken by investigator #1. Areal thickness changes were mapped for the lateral and medial sides with the help of an autonomous mathematical reconstruction algorithm and volume loss was calculated based on wear scar area and local thickness change. Investigator #2, blinded from these results, measured the minimum thickness of the medial and lateral tibial plateau using a dial indicator with a spherical tip radius of 3mm. Twenty-three short term retrievals (3 to 4 per implant size), removed due to infection and without any signs of wear, served as “unused” reference. Linear penetration was then calculated by subtracting the minimum thickness of each plateau from the average thickness of the reference components.
Results
The autonomous reconstruction algorithm delivered results for all investigated components and wear maps could be generated in 100% of the cases. There was a linear association between volume loss and time in situ for the medial and lateral tibial plateau, respectively (R2=0.23, p<0.001 and R2=0.32, p<0.001; Fig. 1). The calculated wear rates were 6.89±1.33 medially and 6.91±1.11 mm3/year laterally with an average total wear rate of 13.81±2.04 mm3/year. Also linear penetration correlated with time in situ (R2=0.19, p<0.001). The annual linear penetration was 0.03±0.01 mm/yr medially and 0.03±0.01 mm/yr laterally. Linear penetration and volumetric loss correlated significantly for both the medial and lateral sides (R2= 0.46 re. R2= 0.22; p <0.001); however, the specific uncertainty for each component was relatively high with ±60 mm3 medially and ±63 mm3 laterally as suggested by the 95% single point prediction interval of the correlation.
Discussion
There is a reasonable global correlation between linear penetration and volume loss. However, the uncertainty in obtaining a wear volume from an individual penetration value is relatively high. For example a penetration of 0.5 mm on the medial side correlates to a wear volume of 40 to 160 mm3. The large uncertainty might be because the location of the thickness minimum after wear does no longer coincide with the manufactured minimum of this device (Fig. 2). In conclusion, the results suggest that penetration is not a good surrogate measure for estimating wear volume of individual TKR polyethylene components, but might be a useful surrogate for large cohort studies.
