Tibial Component Undersizing Is Related to High Degrees of Implant Migration Following Cementless Total Knee Arthroplasty: A Study of Radiostereometric Analysis Data for 111 Patients with 2-Year Follow-up : JBJS Open Access

Journal Logo

Scientific Articles

Tibial Component Undersizing Is Related to High Degrees of Implant Migration Following Cementless Total Knee Arthroplasty

A Study of Radiostereometric Analysis Data for 111 Patients with 2-Year Follow-up

Andersen, Mikkel Rathsach MD, PhD; Winther, Nikolaj MD, PhD; Lind, Thomas MD; Schrøder, Henrik MD; Flivik, Gunnar MD, DMSc; Petersen, Michael Mørk MD, PhD, DMSc

Author Information
JBJS Open Access 8(3):e23.00032, July-September 2023. | DOI: 10.2106/JBJS.OA.23.00032



Radiostereometric analysis (RSA) studies have shown that the continuous migration of tibial components is predictive of aseptic loosening following total knee arthroplasty (TKA). In the present study, we investigated whether accurate sizing and placement of tibial components are related to the degree of implant migration as measured with use of RSA.


A total of 111 patients who underwent TKA surgery with a cementless tibial component were followed for a period of 2 years postoperatively, during which implant migration was assessed with use of RSA. RSA was performed within 7 days postoperatively and after 3, 6, 12, and 24 months. Postoperative radiographs were evaluated for component size and placement in the tibia. The evaluations were performed by experienced knee surgeons who were blinded to the migration data and clinical outcomes. A multivariable linear regression analysis was conducted.


Continuous implant migration (i.e., migration occurring between 12 and 24 months postoperatively) had a negative association with tibial component size (coefficient [B], −0.2; 95% confidence interval [CI], −0.33 to −0.08). Subsidence was associated with the absence of posterior cortical bone support (B, −0.7; 95% CI, −1.09 to −0.28), the absence of lateral cortical bone support (B, 0.8; 95% CI, 0.29 to 1.37), frontal-plane varus malalignment (B, 0.6; 95% CI, 0.12 to 1.16), and component undersizing (B, −0.4; 95% CI, −0.06 to −0.68). Posterior tilt was associated only with undersizing (B, 0.6; 95% CI, 0.27 to 1.11).


Undersized cementless tibial components are at a higher risk for poor fixation with continuous migration following TKA. Therefore, a higher risk of aseptic loosening should be expected.

Level of Evidence: 

Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Cementation of tibial components remains the gold standard in total knee arthroplasty (TKA); however, the use of cementless TKA is increasing. Meta-analyses and literature reviews comparing newer cementless TKA implant designs with cemented implants have found similar or equivalent functional outcomes and survivorship between the 2 implant types1–5. The advantages of cementless fixation are the elimination of aseptic loosening as a result of bone resorption at the bone-cement interface and enhanced fixation as a result of bone ingrowth6. Cementless fixation is often chosen in young patients with high physical demands and good bone quality, and newer cementless designs have been shown to be durable and reliable at long-term follow-up7,8.

Radiostereometric analysis (RSA) has proven to be the best available method for predicting future aseptic loosening and late revision following TKA9,10. RSA provides detailed information regarding the migration of inserted arthroplasty components in relation to the surrounding bone. For cementless tibial components, the degree of continuous migration is arguably more relevant than the degree of initial migration11–13. For both cementless and cemented tibial components, the dominant directions of migration are subsidence and posterior rotation in the tibial host bone11,14,15 (Fig. 1). The RSA technique is typically utilized in randomized trials comparing new component designs with an established gold-standard component that has low levels of migration, and the technique has been recommended as an important part of the early evaluation of new orthopaedic prostheses16.

Fig. 1:
A radiograph showing implant migration following cementless TKA. The arrows indicate the predominant migration directions of the cementless tibial component. Curved arrow = posterior tilt, straight arrow = subsidence.

In this exploratory study, we aimed to investigate the relationship between the migration of cementless tibial components and factors related to surgical technique, such as the sizing of the tibial component; the placement of the tibial component, including a lack of cortical bone support (i.e., the bone-shell in the epiphysis) on either side of the tibial component; and varus or valgus angulation. We hypothesized that tibial component undersizing and a lack of cortical bone support on either side of the tibial component would be related to higher degrees of migration and, specifically, to subsidence of the tibial component. We reasoned that undersizing of the cementless tibial component would result in a decreased and altered area of weight transfer in the trabecular bone of the proximal tibia.

Materials and Methods

We included 111 patients who underwent cementless TKA for primary osteoarthritis and who had completed 2-year follow-up with RSA data. All patients included had taken part in 1 of 2 prospective randomized clinical trials; the flowcharts for each trial have been presented in previous publications11,17. Patients received either the Vanguard Porous Plasma Spray (Zimmer Biomet), the Vanguard Regenerex (Zimmer Biomet), or the monoblock or modular version of the NexGen Cruciate Retaining Trabecular Metal Technology (Zimmer Biomet) cementless tibial component. All patients received a cruciate-retaining cementless femoral component and underwent patellar resurfacing with a cemented all-polyethylene patellar component. All operations were performed by experienced knee surgeons at the Clinic for Knee and Hip Replacement, Department of Orthopedics, Herlev-Gentofte Hospital, or at the Hørsholm Knee Clinic. All patients were <70 years old at the time of surgery and had no bone-related diseases other than osteoarthritis. All inclusion and exclusion criteria for the 2 randomized clinical trials were identical.

Postoperative radiographs were assessed by 2 experienced knee surgeons at different institutions who were blinded to the identity of the patient, the operating surgeon, the migration data, and the clinical outcome. The assessors were asked to judge the radiographs according to the following factors: tibial component size (undersized, fitted, or oversized), presence of cortical bone support (anterior, lateral, medial, and posterior support), frontal-plane alignment (varus, neutral, or valgus), and sagittal-plane alignment (less than, equal to, or greater than the planned posterior slope). A lack of cortical bone support was defined as the presence of a gap from the tibial component to the edge of the epiphyseal bone in the proximal tibia and the placement of the tibial component on the cut tibial surface.

Horizontal-plane alignment was not possible to evaluate from the radiographs. Varus and valgus angulation were assessed in relation to the tibial axis with use of anteroposterior radiographs. Sagittal alignment was assessed with use of lateral radiographs with reference to the planned posterior slope, which was 7° for all of the investigated components. Full-length lower-extremity radiographs were not available, and thus hip-knee-ankle angles could not be measured accurately.


The RSA was performed with use of marker-based software (UmRSA version 6.0; RSA Biomedical). We inserted 0.8-mm tantalum markers into the tibial host bone and 1-mm markers into the polyethylene. Markers were inserted using the same technique in similar grids in all patients. RSA radiographs were made at the Rigshospitalet Department of Radiology within 7 days postoperatively and after 3, 6, 12, and 24 months. Patients were positioned in a standardized supine position with the operative knee placed in a plexiglass biplane calibration cage (Calibration Cage 21; Tilly Medical Products). The same physician positioned the patients at each examination. Ceiling-mounted, portable x-ray tubes were positioned perpendicular in the anterior-posterior and medial-lateral planes at a distance of 100 cm from the films, which were placed in portable cassettes. The x-ray settings were 50 kV and 25 mAs for all examinations. All radiographs were approved by the same physician to ensure sufficient quality. The radiographs were digital, with 9 pixels per millimeter. The radiographs were imported into the UmRSA software with use of DICOM Link (version 3.0; RSA Biomedical) with a resolution of 254 dots per inch. To ensure the reliability of the migration results, the condition number cutoff was set at 130 and the mean error cutoff was set at 0.300 mm in accordance with general RSA guidelines18.


The prior studies were approved by the Scientific Ethical Committee of Copenhagen (H-1-2012-033 and H-3-2009-007). After written and oral information were provided, informed consent was obtained from all participants prior to their inclusion in the study. The present study was approved by the Danish Data Protection Agency (ID 01766, GEH-2012-027, 2009-41-3737).

Statistical Analysis

We investigated the relationship between surgical technique factors, such as the sizing and placement of the component, and tibial component migration by performing a multivariable linear regression analysis with maximum total point motion (MTPM), subsidence (negative y-axis translation), or posterior tilt (negative x-axis rotation) as the dependent variable and tibial component size; lateral, medial, anterior, and posterior cortical bone support; and malalignment as the independent variables. We refrained from excluding any variables from the statistical analysis in order to achieve the best-fitting model. The 4 included types of cementless tibial components were not differentiated in the statistical analysis. The unstandardized coefficient of regression (B), p value, coefficient of determination (R2), and 95% confidence interval (CI) were calculated. Significance was set at p < 0.05.

Two experienced knee surgeons who were blinded to migration data and clinical outcomes assessed the size of the component and its placement in the tibia. One primary blinded assessment was utilized for the statistical analysis, and 1 secondary blinded assessment was utilized as a control to assess reproducibility. Interobserver agreement regarding the blinded assessments of the postoperative radiographs was evaluated by calculating the kappa value for each of the factors that the observers were asked to assess.

Source of Funding

No external funding was received for this study.


Over the course of the 24-month follow-up period, the migration pattern of the cementless tibial component for all patients followed the expected migration pattern, with a relatively high initial migration (mean total MTPM after 3 months, 1.07 mm; standard deviation [SD], 0.57 mm). From 6 to 12 months postoperatively, the components stabilized, with an increase in the MTPM from a mean of 1.13 mm (SD, 0.56 mm) to a mean of 1.15 mm (SD, 0.58 mm). There was similarly little migration from 12 to 24 months postoperatively, with a mean MTPM of 1.21 mm (SD, 0.56 mm) after 24 months (Fig. 2-A). The predominant directions of migration were subsidence and posterior tilt, with most of the migration occurring during the first 3 postoperative months (Figs. 2-B and 2-C).

Fig. 2:
Fig. 2-A The mean MTPM and standard deviation for the entire cohort during the 24-month follow-up period. Fig. 2-B The mean subsidence (negative y-axis translation) and standard deviation for the entire cohort during the 24-month follow-up period. Fig. 2-C The mean posterior tilt (negative x-axis rotation) and standard deviation for the entire cohort during the 24-month follow-up period.

The results of the blinded radiographic assessments, including the number of patients assigned to each category by each assessor and the kappa value showing the amount of interrater variation for each factor, are presented in Table I. The kappa values for all assessed factors averaged 0.63 (range, 0.40 to 0.79), with the lowest kappa values found for sagittal malalignment and posterior cortical bone support of the tibial component and the highest kappa values found for tibial component size and anterior cortical bone support.

TABLE I - Interobserver Agreement and the Number of Patients Assigned to Each Category
Factor Assessed No. of Cases Kappa
Primary Assessor Secondary Assessor
Tibial component size 0.79
 Undersized 31 35
 Fitted 69 69
 Oversized 11 7
Posterior cortical bone support 0.52
 Yes 83 67
 No 28 44
Anterior cortical bone support 0.76
 Yes 75 69
 No 36 42
Lateral cortical bone support 0.69
 Yes 100 100
 No 11 11
Medial cortical bone support 0.63
 Yes 93 81
 No 18 30
Frontal malalignment 0.60
 Varus 8 8
 Neutral 103 103
 Valgus 0 0
Sagittal malalignment* 0.40
 Posterior slope 13 22
 Correct slope 74 80
 Anterior slope 24 9
*Posterior and anterior slopes were relative to the intended 7° of the postoperative posterior slope.

The results of the multivariable linear regression analysis of the relationship of each of the surgical technique factors to total migration (i.e., the MTPM from 0 to 24 months postoperatively) and continuous migration (i.e., the MTPM from 12 to 24 months postoperatively) are presented in Table II. Tibial component size was negatively associated with both total migration and continuous migration; that is, a smaller size of the tibial component relative to the host bone was associated with a higher degree of total and continuous migration (Table II).

TABLE II - Results of the Multivariable Stepwise Regression Best-Fitting Model
B 95% CI P Value R2
Total migration (MTPM0-24 months), in mm 67%
 Tibial component size −0.72 −1.19 to −0.24 0.003
 Posterior cortical bone support 1.33 0.70 to 1.95 <0.001
 Anterior cortical bone support 0.18 −0.27 to 0.63 0.428
 Lateral cortical bone support −1.49 −2.31 to 0.67 <0.001
 Medial cortical bone support −0.51 −1.14 to 0.13 0.115
 Frontal malalignment −0.57 −1.40 to 0.27 0.178
 Sagittal malalignment 0.13 −0.27 to 0.53 0.514
Continuous migration (MTPM12-24 months), in mm 44%
 Tibial component size −0.21 −0.33 to −0.08 <0.001
 Posterior cortical bone support 0.05 −0.13 to 0.22 0.602
 Anterior cortical bone support 0.06 −0.06 to 0.18 0.350
 Lateral cortical bone support −0.15 −0.37 to 0.07 0.172
 Medial cortical bone support −0.11 −0.28 to 0.06 0.216
 Frontal malalignment 0.01 −0.21 to 0.23 0.961
 Sagittal malalignment 0.00 −0.10 to 0.11 0.949
Total subsidence (negative y-axis translation), in mm 62%
 Tibial component size −0.37 −0.06 to −0.68 0.021
 Posterior cortical bone support −0.69 −1.09 to −0.28 <0.001
 Anterior cortical bone support −0.13 −0.43 to 0.17 0.376
 Lateral cortical bone support 0.83 0.29 to 1.37 0.003
 Medial cortical bone support 0.37 −0.05 to 0.79 0.081
 Frontal malalignment 0.64 0.12 to 1.16 0.017
 Sagittal malalignment 0.02 −0.25 to 0.28 0.912
Total posterior tilt (negative x-axis rotation), in degrees 36%
 Tibial component size 0.57 0.27 to 1.11 0.007
 Posterior cortical bone support −0.36 −1.16 to 0.44 0.376
 Anterior cortical bone support −0.29 −0.87 to 0.30 0.339
 Lateral cortical bone support −0.00 −1.07 to 1.06 0.994
 Medial cortical bone support 0.16 −0.67 to 0.99 0.699
 Frontal malalignment 0.11 −0.92 to 1.14 0.837
 Sagittal malalignment 0.10 −0.43 to 0.62 0.715

Subsidence (negative y-axis translation) was significantly associated with tibial component size, posterior and lateral cortical bone support, and frontal-plane varus malalignment of the component (Table II). Continuous subsidence (negative y-axis translation from 12 to 24 months postoperatively) was not significantly associated with any variable.

Posterior tilt (negative x-axis rotation) was significantly associated only with tibial component size (Table II). Continuous posterior tilt (negative x-axis rotation from 12 to 24 months postoperatively) was not significantly associated with any variable.

None of the included patients underwent a revision TKA procedure during the 24-month follow-up period.


The expected migration pattern for cementless components is an initial migration within the first 6 months that is relatively high compared with that of cemented tibial components followed by a stabilization period similar to that of cemented components11,12,14. Implants that do not achieve stabilization are considered at risk for aseptic loosening10.

The findings of this study indicate that undersizing of the cementless tibial component results in poorer fixation, a higher degree of continuous migration, and increased subsidence and posterior tilt. Although this study was exploratory in nature, these findings could be of clinical importance. This cohort of 111 patients (111 cementless components) with 2 years of RSA follow-up represents a high-quality data set for investigating the relationship between surgical technique factors and implant migration, which could not have been done ethically and practically in a randomized setting.

In our experience, surgeons tend to undersize the tibial component in cases in which a perfect fit is not possible. This pattern was also evident in our study cohort, in which the 2 blinded observers assessed a total of 31 and 35 of 111 components as having been undersized and a total of 11 and 7 components as having been oversized (Table I). Undersizing the tibial component necessitates placing that component without sufficient cortical bone support on either or all sides of the component. In the cut surface of the tibia, compressive strength is highest in the lateral and medial periphery of the trabecular bone and is lowest in the central part of the trabecular bone. Therefore, in addition to reducing the area of weight distribution, undersizing also moves a greater part of the weight transfer to the weaker central bone19,20. The tendency for surgeons to undersize tibial components may be the result of several previous studies showing that tibial component oversizing and medial overhang are associated with postoperative pain and poor functional outcomes21–23. However, other studies have indicated that overhang is not associated with pain and poor function24,25. Undersizing may be especially undesirable in cementless TKA if the tibia is cut distal to the metaphyseal bone in order to compensate for erosion of the medial or lateral condyle of the tibia, as trabecular bone strength decreases substantially at ≥5 mm below the surface of the joint20. Bone strength is linearly correlated with bone mineral density, and a relationship between low tibial bone mineral density and high degrees of migration was demonstrated in a previous study of ours26. However, further research is needed to account for the relationship between bone mineral density and body mass index in this patient group.

In our experience, an undersized tibial component or a tibial component that fits in the coronal plane but not in the sagittal plane will most often be placed closer to the anterior cortex than to the posterior cortex because that is surgically easier; however, the supporting cortical bone is strongest in the posterior regions19,20.

Newer TKA systems provide a wider selection of component sizes and shapes than older systems, which should improve the possibility for exact component sizing. However, when the ideal tibial component size is unavailable, it is important for surgeons to consider that undersizing the tibial component increases the risk of poor fixation, as demonstrated in the present study.


This was a retrospective exploratory study performed with use of data from separate randomized clinical trials. The study population was insufficiently sized to detect a relationship between tibial component size and clinical or patient-reported outcomes. As such, larger studies are needed to investigate these potential relationships.


1. Nakama GY, Peccin MS, Almeida GJ, Lira Neto OdeA, Queiroz AA, Navarro RD. Cemented, cementless or hybrid fixation options in total knee arthroplasty for osteoarthritis and other non-traumatic diseases. Cochrane Database Syst Rev. 2012 Oct 17;10:CD006193.
2. Mont MA, Pivec R, Issa K, Kapadia BH, Maheshwari A, Harwin SF. Long-term implant survivorship of cementless total knee arthroplasty: a systematic review of the literature and meta-analysis. J Knee Surg. 2014 Oct;27(5):369-76.
3. Newman JM, Sodhi N, Khlopas A, Sultan AA, Chughtai M, Abraham R, Oh J, Molloy RM, Harwin SF, Mont MA. Cementless Total Knee Arthroplasty: A Comprehensive Review of the Literature. Orthopedics. 2018 Sep 1;41(5):263-73.
4. Zhou K, Yu H, Li J, Wang H, Zhou Z, Pei F. No difference in implant survivorship and clinical outcomes between full-cementless and full-cemented fixation in primary total knee arthroplasty: A systematic review and meta-analysis. Int J Surg. 2018 May;53:312-9.
5. Haeberle HS, Salem HS, Ehiorobo JO, Sodhi N, Mont MA. Newer Generation of Cementless Total Knee Arthroplasty: A Systematic Review. Surg Technol Int. 2020 May 28;36:351-9.
6. Gallo J, Goodman SB, Konttinen YT, Wimmer MA, Holinka M. Osteolysis around total knee arthroplasty: a review of pathogenetic mechanisms. Acta Biomater. 2013 Sep;9(9):8046-58.
7. Gibon E, Lewallen DG, Larson DR, Stuart MJ, Pagnano MW, Abdel MP. John N. Insall Award: Randomized Clinical Trial of Cementless Versus Cemented Tibial Components: Durable and Reliable at a Mean 10-Years Follow-Up. J Arthroplasty. 2023 Jun;38(6S):S14-20.
8. Ritter MA, Meneghini RM. Twenty-year survivorship of cementless anatomic graduated component total knee arthroplasty. J Arthroplasty. 2010 Jun;25(4):507-13.
9. Ryd L, Albrektsson BE, Carlsson L, Dansgård F, Herberts P, Lindstrand A, Regnér L, Toksvig-Larsen S. Roentgen stereophotogrammetric analysis as a predictor of mechanical loosening of knee prostheses. J Bone Joint Surg Br. 1995 May;77(3):377-83.
10. Pijls BG, Valstar ER, Nouta KA, Plevier JW, Fiocco M, Middeldorp S, Nelissen RG. Early migration of tibial components is associated with late revision: a systematic review and meta-analysis of 21,000 knee arthroplasties. Acta Orthop. 2012 Dec;83(6):614-24.
11. Andersen MR, Winther N, Lind T, SchrøDer H, Flivik G, Petersen MM. Monoblock versus modular polyethylene insert in uncemented total knee arthroplasty. Acta Orthop. 2016 Dec;87(6):607-14.
12. Henricson A, Nilsson KG. Trabecular metal tibial knee component still stable at 10 years. Acta Orthop. 2016 Oct;87(5):504-10.
13. Wojtowicz R, Henricson A, Nilsson KG, Crnalic S. Uncemented monoblock trabecular metal posterior stabilized high-flex total knee arthroplasty: similar pattern of migration to the cruciate-retaining design - a prospective radiostereometric analysis (RSA) and clinical evaluation of 40 patients (49 knees) 60 years or younger with 9 years’ follow-up. Acta Orthop. 2019 Oct;90(5):460-6.
14. Dunbar MJ, Wilson DA, Hennigar AW, Amirault JD, Gross M, Reardon GP. Fixation of a trabecular metal knee arthroplasty component. A prospective randomized study. J Bone Joint Surg Am. 2009 Jul;91(7):1578-86.
15. Gudnason A, Adalberth G, Nilsson KG, Hailer NP. Tibial component rotation around the transverse axis measured by radiostereometry predicts aseptic loosening better than maximal total point motion. Acta Orthop. 2017 Jun;88(3):282-7.
16. Malchau H. Introducing new technology: a stepwise algorithm. Spine (Phila Pa 1976). 2000 Feb 1;25(3):285.
17. Winther NS, Jensen CL, Jensen CM, Lind T, Schrøder HM, Flivik G, Petersen MM. Comparison of a novel porous titanium construct (Regenerex®) to a well proven porous coated tibial surface in cementless total knee arthroplasty - A prospective randomized RSA study with two-year follow-up. Knee. 2016 Dec;23(6):1002-11.
18. Valstar ER, Gill R, Ryd L, Flivik G, Börlin N, Kärrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop. 2005 Aug;76(4):563-72.
19. Goldstein SA, Wilson DL, Sonstegard DA, Matthews LS. The mechanical properties of human tibial trabecular bone as a function of metaphyseal location. J Biomech. 1983;16(12):965-9.
20. Harada Y, Wevers HW, Cooke TDV. Distribution of bone strength in the proximal tibia. J Arthroplasty. 1988;3(2):167-75.
21. Bonnin MP, Saffarini M, Shepherd D, Bossard N, Dantony E. Oversizing the tibial component in TKAs: incidence, consequences and risk factors. Knee Surg Sports Traumatol Arthrosc. 2016 Aug;24(8):2532-40.
22. Nielsen CS, Nebergall A, Huddleston J, Kallemose T, Malchau H, Troelsen A. Medial Overhang of the Tibial Component Is Associated With Higher Risk of Inferior Knee Injury and Osteoarthritis Outcome Score Pain After Knee Replacement. J Arthroplasty. 2018 May;33(5):1394-8.
23. Simsek ME, Akkaya M, Gursoy S, Isik C, Zahar A, Tarabichi S, Bozkurt M. Posterolateral overhang affects patient quality of life after total knee arthroplasty. Arch Orthop Trauma Surg. 2018 Mar;138(3):409-18.
24. Abram SGF, Marsh AG, Brydone AS, Nicol F, Mohammed A, Spencer SJ. The effect of tibial component sizing on patient reported outcome measures following uncemented total knee replacement. Knee. 2014 Oct;21(5):955-9.
25. Foubert K, Heylen S, Plaeke P, Somville J, Nicolai P. Tibial Component In Total Knee Replacement :The Effect Of Overhang And Sizing On Outcome. Acta Orthop Belg. 2017 Dec;83(4):659-63.
26. Andersen MR, Winther NS, Lind T, Schrøder HM, Flivik G, Petersen MM. Low Preoperative BMD Is Related to High Migration of Tibia Components in Uncemented TKA-92 Patients in a Combined DEXA and RSA Study With 2-Year Follow-Up. J Arthroplasty. 2017 Jul;32(7):2141-6.

Supplemental Digital Content

Copyright © 2023 The Authors. Published by The Journal of Bone and Joint Surgery, Incorporated. All rights reserved.