Influence of marker-selection method in radiostereometric analysis of total knee arthroplasty on tibial baseplate migration patterns: a secondary analysis of a randomized controlled trial with 5-year follow-up

Background and purpose — Different marker-selection methods are applied to represent implant and tibial seg-ments in radiostereometric analysis (RSA) studies of total knee arthroplasty (TKA). Either a consistent set of markers throughout subsequent RSA examinations (“consistent-marker method”) is used or all available markers at each follow-up (“all-marker method”). The aim of this secondary analysis was to compare marker-selection methods on individual and group level TKA migration results. Methods — Data from a randomized RSA study with 72 patients was included. Tibial baseplate migration was evalu-ated at 3 months, 1, 2, and 5 years postoperatively with both marker-selection methods. Additionally, migration was calculated using 5 fictive points, either plotted based on the consistent set of markers or all available markers. Results — Migration could be calculated with both marker-selection methods for 248 examinations. The same prosthesis and bone markers (n = 136), different prosthesis markers (n = 71), different bone markers (n = 21), or different prosthesis and bone markers (n = 20) were used. The mean difference in maximum total point motion (MTPM) between all examinations was 0.02 mm, 95% confidence interval –0.26 to 0.31 mm. 5 implants were classified as continuously migrating with the consistent-marker method versus 6 implants (same 5 plus one additional implant) with the all-marker method. Using fictive points, fewer implants were classified as continuously migrating in both marker-selection methods. Differences between TKA groups in mean MTPM were comparable with both marker-selection methods, also when fictive points were used. Conclusion — Estimated group differences in mean MTPM were similar between marker-selection methods, but individual migration results differed. The latter has implications when classifying implants for estimated risk of future loosening.