A Comparison of Wear Patterns on Retrieved and Simulator-Tested Total Knee Replacements

Abstract

Aseptic implant loosening is the most common reason for revision surgery after total knee replacement. This is associated with adverse biological reactions to wear debris from the articulating implant components. To predict the amount of wear debris generated in situ, standard wear testing of total knee replacement (TKR) is carried out before its clinical use. However, wear data reported on retrievals of total knee replacement (TKR) revealed significant discrepancies compared with standard wear simulator studies. Therefore, the aim of the present study was to compare the wear patterns on identical posterior-cruciate-retaining TKR designs by analyzing retrieved and experimentally tested implants. The identification and classification of wear patterns were performed using 21 retrieved ultra-high-molecular-weight-polyethylene (UHMW-PE) inserts and four sets of inserts of identical design and material tested in a knee wear simulator. These four sets had undergone different worst-case conditions and a standard test in a wear simulator according to ISO 14243-1. Macroscopic and microscopic examinations of the polyethylene inserts were performed, including the determination of seven modes of wear that correspond to specific wear patterns, the calculation of wear areas, and the classification of the damage over the whole articulating area. Retrieved and standard wear simulator-tested UHMW-PE inserts showed significant differences in wear area and patterns. The total wear areas and the damage score were significantly larger on the retrievals (52.3% versus 23.9%, 32.7 versus 22.7). Furthermore, the range of wear patterns found on the retrievals was not reproducible in the simulator-tested inserts. However, good correspondence was found with the simulator-tested polyethylene inserts under worst-case conditions (third body wear), i.e., deep wear areas could be replicated according to the in vivo situation compared with other wear test scenarios. Based on the findings presented here, standard simulator testing can be used to directly compare different TKR designs but is limited in the prediction of their in situ wear. Preclinical wear testing may be adjusted by worst-case conditions to improve the prediction of in situ performance of total knee implants in the future.

Keywords: retrieval analysis; total knee replacement; wear pattern; wear simulator study.

Figure 1

Schematic of the proximal surface of a right insert with partitioning in anterior (A) and posterior (P) portions, as well as medial (M) and lateral (L) compartments on each side.

Figure 2

Illustration of wear patterns found on the retrieved UHMW-PE inserts. (a) Burnishing: polishing of the surface → surface appears smoother than manufacturing marks; (b) Pitting: depressions in articulating surface with irregular shape (2–3 mm across and 1–2 mm deep); (c) Abrasion: roughening of the surface, which results in a shredded or tufted appearance → surface appears rougher than manufacturing marks; (d) Striated wear pattern: described in a study [9] as regularly spaced striations in an anterior-posterior direction; (e) Rim-runner [24]: propagation of the worn areas on the elevated, rounded rim of the insert, with a localized form of creeping; (f) Third-body wear: parallel scratches causing thick scars of material removal (sometimes distinct holes from temporarily embedded debris at one end); (g) Deformation/Deep wear: distinct material accumulation on or around articulating areas; (h) Scratching: indented lines, generally in antero-posterior direction with only slight material removal.

Figure 3

Distribution of percentual total wear areas. Significances, marked with asterisks (one asterisk *: p < 0.05 and two asterisks **: p < 0.01), are shown to facilitate the comparison of retrievals with simulator-tested groups.

Figure 4

Distribution of percentual deep wear areas. Significances, marked with asterisks (one asterisk *: p < 0.05), are shown to facilitate the comparison of retrievals with simulator-tested groups.

Figure 5

Dimensions of anterior–posterior elongation of wear areas of retrieved and simulator-tested UHMW-PE inserts. Significances, marked with asterisks (one asterisk *: p < 0.05 and two asterisks **: p < 0.01), are shown to facilitate the comparison of retrievals with the simulator-tested groups.

Figure 6

Dimensions of medial–lateral elongation of wear areas of retrieved and simulator-tested UHMW-PE inserts. Significances, marked with asterisks (one asterisk *: p < 0.05 and two asterisks **: p < 0.01), are shown to facilitate the comparison of retrievals with the simulator-tested groups.

Figure 7

The ratio of anterior part of ap-elongation and posterior part of ap-elongation of retrieved and simulator-tested UHMW-PE inserts for total and deep wear areas. Significances, marked with asterisks (two asterisks **: p < 0.01), are shown to facilitate the comparison of retrievals with the simulator-tested groups.

Figure 8

Distribution of each wear pattern on the damage score for retrieved and simulator-tested inserts. Significances, marked with asterisks (one asterisk *: p < 0.05 and two asterisks **: p < 0.01), are shown to facilitate the comparison of retrievals with the simulator-tested groups. (Maximum score for each wear pattern was 8, Maximum total damage score was 64).