Does Unicondylar Knee Arthroplasty Affect Tibial Bone Strain? A Paired Cadaveric Comparison of Fixed- and Mobile-bearing Designs : Clinical Orthopaedics and Related Research®

Secondary Logo

Journal Logo


Does Unicondylar Knee Arthroplasty Affect Tibial Bone Strain? A Paired Cadaveric Comparison of Fixed- and Mobile-bearing Designs

Peersman, Geert MD, PhD; Taylan, Orcun MSc; Slane, Joshua PhD; Vanthienen, Ben MSc; Verhaegen, Jeroen MD; Anthonissen, Lyne MD; van Lenthe, G. Harry PhD; Heyse, Thomas MD, PhD; Scheys, Lennart PhD

Author Information
Clinical Orthopaedics and Related Research 478(9):p 1990-2000, September 2020. | DOI: 10.1097/CORR.0000000000001169



Unicondylar knee arthroplasty (UKA) has been advocated as an alternative to TKA when osteoarthritis is limited to a single knee compartment [9, 18, 27, 49], owing to advantages such as more functional anatomy, and improved postoperative kinematics [9, 13, 23, 27, 31, 49].

The mobile-bearing UKA (UKAMB) implant, developed as an alternative to the traditional fixed-bearing UKA (UKAFB), allows the polyethylene insert to rotate and translate on the metal tibial tray [5, 6, 22, 36, 44, 47]. However, conflicting findings have been reported on potential differences between these designs in terms of survivorship and mechanical performance [2, 5, 6, 18, 44]. The rate of revision surgery after UKAFB is reportedly higher than that for UKAMB [26], which may be linked to abnormally high bone tibial strain as it is the most important factor for osteogenic adaptive response [32, 35, 40, 41, 48].

Unexplained postoperative pain in medial UKA, particularly in the medial side of the anterior proximal tibia, accounts for 23% of all UKA revisions [36, 40]. Data from UK and Australian registries suggests pain as the primary reason for revision UKA in 40% and 10% of the cases, respectively. Experimental studies have suggested abnormal distribution of cortical bone strain as a possible cause of postoperative pain [41, 42].

Few experimental studies have investigated bone strain in vivo owing to the invasiveness of the measurement protocol, lack of experimental control, and consequent ethical issues [4, 29]. Alternatively, cadaver-based studies reporting tibial bone strains are usually conducted under static nonphysiological loading conditions in isolated bone structures, thereby neglecting soft-tissue forces and the effect of UKA-induced kinematic changes [1, 11, 14, 41, 42, 43].

To the best of our knowledge, no cadaver-based study has compared strain in the tibial cortex after UKAMB and UKAFB to the native condition, especially for dynamic physiological joint loading. Moreover, the effect of implant design on increased postoperative strain remains unclear. Therefore, in this cadaver-based study we: (1) performed a cadaver-based kinematic analysis on paired cadaveric specimens before and after mobile-bearing and fixed-bearing UKA; and (2) simultaneously characterized the strain distribution in the anterior and posterior proximal tibia during squatting.

Materials and Methods

Paired lower limbs from five fresh-frozen cadavers were disarticulated at the hip (Table 1). All specimens were screened for trauma sequelae, implant material and severe bone deformities before inclusion in the study. Ethical approval was obtained from the regional ethical committee (NH019 2015-11-03). During testing, all specimens were kept moist using a phosphate-buffered saline solution.

Table 1.:
Anthropometric data of the donor knee specimens used in this study


We affixed 5-mm bicortical bone pins to the tibia and femur at a distance of 14 cm and 17 cm from the joint line, respectively, to rigidly mount motion tracking markers (Appendix Section 1, Supplemental Digital Content 1, CT (slice thickness: 0.6 mm; Siemens Definition Flash, Siemens, Erlangen, Germany) was performed. To minimize inter-rater variability [45], a single researcher (OT) generated three dimensional (3-D) models of the knee and identified anatomical landmarks (Appendix Section 2, Supplemental Digital Content 1, required to define specimen-specific joint coordinate systems [16] (Mimics 20.0, Materialise, Leuven, Belgium) according to the Grood and Suntay convention [16, 19].

Preparation of the Specimens and Sensor Positioning

Specimens were thawed 24 hours before testing. The soft tissue surrounding the knee was removed while carefully preserving the joint capsule, ligaments, and tendons. The femur, tibia, and fibula were resected using an oscillating saw 320 mm proximally and 280 mm distally from the joint line. The femoral head was preserved for later instrumentation with dummy reference strain gauges (further described below). The lateral and medial hamstrings were sutured, and the quadriceps tendon was affixed to a custom-made clamp. Each bone was embedded in metal containers with a cold-cure acrylic resin (VersoCit2, Struers, Ballerup, Denmark). Care was taken to maintain the physiological tibiofemoral alignment.

To allow for measurement of tibial cortical bone strain, three stacked strain gauge rosettes (diameter: 10 mm, grid length: 3 mm; KFG-3-120-D17-11L1M2S, Kyowa, Tokyo, Japan) [8, 38, 42] were attached (Appendix Section 3, Supplemental Digital Content 1, using a previously described method [15] at predefined anterior and posterior positions on the tibial cortex, 34.4 mm ± 5.28 mm distally from the tibial plateau (Fig. 1A). An anteromedial sensor was placed 18.18 mm ± 5.4 mm medially from the tibial mechanical axis. An anterolateral sensor was placed 10.55 mm ± 6.9 mm medially from the tibial mechanical axis. A posterior sensor was placed 6.77 mm ± 4.3 mm medially from the tibial mechanical axis. Additionally, three dummy sensors were placed on the separate piece of femoral head and connected to the measurement strain gauges using a Wheatstone bridge configuration to prevent environmental factors, such as temperature change, from confounding the measurements (Appendix Section 4, Supplemental Digital Content 1,; Appendix Figure 1, Supplemental Digital Content 2, Sensor outputs were recorded at 2000 Hz and synchronized with the other devices using Labview (National Instruments, Austin, TX, USA).

Fig. 1:
A-C Experimental setup: (A) preparation of the tibial surface and attachment of anteromedial and anterolateral strain gauges (B) dynamic knee simulator replicating loaded squatting (C) representation of strain calculation (ɛ1,2,3 = normal strains from each rosette strain gauge; ɛmax and ɛmin = principal strains).

The position of the strain gauges on the tibia were digitized by two operators (OT, JS) using a digitizing wand tracked by a motion capture system (Vicon, Oxford, UK), and normalized as per the length and proximal mediolateral width of the tibia (Appendix Section 5, Supplemental Digital Content 1, The statistical difference between the sides were computed using Wilcoxon signed rank test (p < 0.05). There were no differences between specimen sides (tibial axis (anteromedial [right = 8.9 ± 0.6% and left = 9.6 ± 0.6%; p = 0.08], anterolateral [right = 8.0 ± 1.0% and left = 9.1 ± 1.3%; p = 0.08] and posterior [right = 9.4 ± 1.5% and left = 9.1 ± 1.7%; p = 0.345] and mediolateral width [anteromedial (right = 46.2 ± 14.3% and left = 43.5 ± 7.4%; p = 0.89], anterolateral [right = 25.0 ± 19.6% and left = 27.7 ± 13.0%; p = 0.50], except for the posterior sensor’s normalized mediolateral position [right = 23.0 ± 9.7% and left = 11.0 ± 7.6%; p = 0.04]).

Test Set-up and Protocols

Specimens were mounted in a previously validated cadaveric knee simulator [46] that applied a dynamic squatting motion to simulate load through a full ROM. Clusters with four spherical retro-reflective markers each were mounted on the bone pin-mounted holders (Fig. 1B). An electromechanical actuator was used to apply dynamic load to the quadriceps, while the hamstrings were loaded at 50 N using constant-force springs [46]. Six cycles of preloading to simulate squatting were applied, with a resting time of 1 minute between cycles to precondition the specimen and minimize hysteresis [17, 39]. During squatting, the quadriceps load was programmed to maintain a constant vertical ankle load of 110 N [19, 21, 33].

A six-camera motion capture system (MX40+, Vicon, Oxford, UK) was used to track marker clusters on the bones. Tibiofemoral kinematics during squatting were analyzed using dedicated motion capture software (Nexus 1.8.5, Vicon, Oxford, UK) and custom-programs in Matlab (R2017b, Mathworks Inc, Natick, MA, USA) [16, 21].


After native knee testing, specimens were implanted with a medial UKA by a single surgeon (GP) following the manufacturer guidelines. UKAMB (right leg: Oxford®, Microplasty; Zimmer Biomet, Warsaw, IN, USA) was performed on all right knees and UKAFB (left leg: Vanguard® M; Zimmer Biomet, Warsaw, IN, USA) on the left using a minimally invasive medial parapatellar approach with controlled under-correction of the overall mechanical alignment [3, 7, 24]. All motion trials were repeated for each specimen as above. In addition, each tibial tray position was digitized by a wand to compare varus and posterior tilt angles [30, 37] between both specimen sides, that is, design groups, (Wilcoxon signed rank test). No differences were found in terms of varus and posterior tilt angles (varus angle (right = 5.9 ± 0.8° and left = 5.1 ± 3°; p = 0.59) and posterior tilt angle (right = 3.1 ± 2.5° and left = 3.45 ± 2°; p = 0.56).

Data Processing

Data obtained from the motion capture system were downsampled and interpolated at intervals of 1° of flexion and within a common range of knee flexion (40°-99°) for all specimens. Kinematics were reported as mean ± SD for each condition (native, UKAFB, and UKAMB). Maximum and minimum principal strains were recorded by the rosette sensors as a function of knee flexion [28]

ɛ1,2,3 express the normal strains from each of the rosette strain gauges, ɛmax and ɛmin express the maximum and minimum principal strains, respectively (Fig. 1C).

Statistical Analysis

All kinematic and strain data were expressed as differences between postoperative and native condition, that is, Post(UKAMB) – PreRight (Native) and Post(UKAFB) – PreLeft (Native). This allowed us to consider repeated measurements comparing pre-UKA and post-UKA conditions and left and right legs without sacrificing statistical power.

A linear mixed model was used to test for differences between implant designs (p < 0.05), using the “nlme” package (R-Studio 1.0.136, Boston, MA, USA), with implant side as a function of knee flexion angle (fixed effect) and the donor as the second repeated measurement (random effect) (Data ∼ implant type*flexion angle, random=∼1|Donor). For both designs the difference in peak strain with respect to native over the flexion range was analyzed in terms of effect size (Cohen’s d).


Kinematics During Squatting

The UKAMB and UKAFB designs were not different (Appendix Table 1, Supplemental Digital Content 3, in terms of valgus orientation throughout the range of flexion (Table 2). With respect to the native condition, UKAMB and UKAFB both demonstrated a shift towards increased valgus (Fig. 2A). In contrast, UKAFB demonstrated more external rotation than UKAMB; the latter more-closely replicating the native condition between 79 and 99° of flexion (Fig. 2B).

Table 2.:
Statistical analyses for measured parameters as a function of knee flexion angle with respect to the ability of fixed-bearing and mobile-bearing UKA to replicate the native behavior
Fig. 2:
A-F Kinematics of the knee in the native condition (black) and after unicondylar knee arthroplasty using fixed-bearing implants (green) and mobile-bearing implants throughout the range of flexion for (A) valgus orientation (B) tibial internal rotation (C) inferosuperior position of the medial femoral condyle center and (D) lateral femoral condyle center (E) AP position of the medial femoral condyle center and (F) lateral femoral condyle center. Data is represented as mean (solid) ± SD (shaded); FB = fixed-bearing; MB = mobile-bearing; FMCC IS = medial femoral condyle center inferiorsuperior; FLCC IS = medial femoral condyle center inferior superior; FMCC AP = medial femoral condyle center AP; FLCC AP = lateral femoral condyle center AP.

In terms of translational kinematics, the UKAMB design more closely replicated the native condition for the inferior-superior position of the medial (Fig. 2C) and lateral (Fig. 2D) femoral condyle centers; however, UKAFB consistently demonstrated a more superior position throughout the flexion range (Appendix Table 2, Supplemental Digital Content 4, This introduced a difference in the inferior-superior position between the UKA designs for the medial (86°-92° and 94°-99°) and lateral femoral condyle (82°-99°). Finally, UKAMB also more closely replicated native behavior in terms of AP translation of both condyles between 65° and 94° and between 55° and 99° of flexion for the medial (Fig. 2E) and lateral (Fig. 2F) femoral condyle centers, respectively.

Native knees demonstrated rollback patterns during squatting in which the medial and lateral femoral condyle centers translated anteriorly and posteriorly on the tibial plateau, respectively (Fig. 3A-C). Associated with the above findings in terms of AP translations; the medial and lateral femoral condyle centers qualitatively demonstrated similar patterns during squatting in the UKAMB condition, whereas the UKAFB condition demonstrated posterior translation for the medial and lateral femoral condyle centers as the flexion angle increased.

Fig. 3:
A-C Mean femoral rollback pattern across the specimens in (A) the native condition and after unicondylar knee arthroplasty using (B) fixed-bearing implants and (C) mobile-bearing implants throughout the range of flexion during squatting. Solid dots on the tibial plateau represent the centers of the medial and lateral femoral condyles. FB = fixed-bearing; MB = mobile-bearing.

Tibial Cortical Bone Strain

Bone strain values consistently increased compared with native condition with both designs in the anteromedial and posterior region (Fig. 4A, 4C), with the increase in UKAMB being larger from 95° and 93° onwards, respectively (Table 2), and an increased peak strain of 92% (effect size 1.5) and 97% (effect size 2.1) (Table 3). However, in the anterolateral region of the medial tibial bone, UKAFB demonstrated consistently increased strain, while UKAMB closely replicated strain values of the native knee in this particular area (Fig. 4B). This region also showed the overall highest maximal principal strain values after UKAFB (1010 µε [± 787] or 79% increase, effect size 1.9) (Table 3) compared with all regions in both the native and UKAMB conditions (Appendix Table 3, Supplemental Digital Content 5,

Fig. 4:
A-C Maximum and minimum principal strains in the native condition (black) and after unicondylar knee arthroplasty using fixed-bearing implants (green) and mobile-bearing implants (red) throughout the range of flexion for (A) anteromedial, (B) anterolateral and (C) posterior sensors attached on the proximal tibial cortex. Data is represented as mean (solid) ± SD (shaded). The red dot on the 3-D tibia model indicates the respective position of the sensor associated with each set of graphs, while a black dot indicates the relative position of other sensors for comparison; FB = fixed-bearing; MB = mobile-bearing.
Table 3.:
Peak maximum and minimum principal strains (mean ± SD) for the native, fixed- and mobile-bearing unicondylar knee arthroplasty

In terms of the minimum principal strain, the anteromedial region showed no difference with numbers available between designs throughout the range of knee flexion (Table 2), associated with small effect sizes (Table 3). In contrast, UKAMB and UKAFB showed differences in the minimal anterolateral tibial strain between 97° and 99° of flexion (Table 2), with the UKAFB design demonstrating consistently increased and the overall highest strain (effect size 0.8) (Table 3), similar to the maximal principal strain. The UKAMB values closely replicated those of the native strain (Fig. 4), resulting in an effect size of 0 (Table 3).


It has been suggested that UKAMB replicates native tibiofemoral kinematics better than UKAFB [10, 19-21, 33]. Although previous studies have compared the kinematics of the native knee with those of UKAMB or UKAFB designs individually, this study is, to our knowledge, the first cadaver-based study directly comparing UKAMB and UKAFB using matched pairs and integrated assessment of tibial bone strain. Although the UKA design-specific kinematic changes compared with native knees confirmed previous findings in the literature, we found that both UKAMB and UKAFB lead to an increase in proximal tibial bone strain compared with the native knee. It was only in the anterolateral region that we found that strain was affected by the UKA design with UKAMB demonstrating reduced strain.


This study had several limitations. First, there are inherent limitations to cadaver studies, including limited and high-cost access to specimens, leading to typical sample sizes of 5 to 10 [21]. Despite careful selection and preservation, the cadaver bone used in this study may have behaved differently from the targeted clinical population because it cannot remodel with time to potentially mitigate increased initial bone strains. Nevertheless, for the native condition, a peak maximum anteromedial principal strain of 311 µε ± 190 µε and a peak maximum anterolateral strain of 563 µε ± 234 µε anterolaterally were found during squatting which compare well to the results of prior in vivo studies; Lanyon et al. [29] reported principal strains of 850 µε during running and 400 µε during walking. Additionally, our findings only apply to one motor task, and other tasks may be associated with different strains. As such, Burr et al. [4] focused on more demanding motor tasks and noticed a dramatic increase, especially during zigzag running wherein maximal compressive strains of 1226 µε ± 168 µε and tensile strains of 743 µε ± 77 µε were recorded; these were higher than the native strain ranges in our study. Measurements were subject to the exact location of strain sensors. However, given the absence of systematic differences between both design groups, we do not expect them to have influenced our design-related findings. Nevertheless, we plan to further investigate the full-field proximal tibial strain and the effect of other possible contributing factors, such as ligament tensioning, motor tasks and implant distortion, through a finite element analysis based on the data of this study. Additionally, for safety and inter-specimen variations in tibiofemoral alignment, squatting motions were limited to 40° to 99° of flexion, with a relatively low vertical load of 110 N. Scott et al. [41] performed a digital image correlation-based strain analysis of medium composite tibial sawbones to investigate different UKA designs under a higher, but static vertical load of 2500 N directly applied to the medial tibial compartment. They reported that the fixed-bearing design generated lower maximal vertical strain in the medial aspect of the tibia (1301 µε ± 328 µε) than the mobile-bearing design (1662 µε ± 32 µε) [41], which is consistent with our findings (anteromedial sensor) (Fig. 4A). The findings of the current study were conducted on a single UKAFB and UKAMB design and may not be applicable to other UKA of other manufacturers and designs. Lastly, many of the observed differences remain small and it remains unclear to which extent they are clinically important.


In terms of kinematics, this study confirms the results of previous studies evaluating these UKA designs individually; there were design-specific changes in terms of kinematics [20, 21]. We found that UKAMB replicates native tibiofemoral kinematics better than UKAFB. More specifically, both UKAMB and UKAFB demonstrated increased valgus as compared with the native condition (Fig. 2A), which has been associated with stiffness mismatch induced by these implants on the medial side [21, 24]. However, UKAMB allowed for better preservation of tibial rotation and better AP stability of the medial femoral condyle through the flexion cycle than UKAFB. UKAMB also more closely approximated the native inferior-superior translation than UKAFB, although with both designs, the medial femoral condyle center had a more superior position throughout the range of flexion. Additionally, although tibiae moved into more external rotation with the UKAFB, UKAMB preserved internal rotation of the tibia and the associated screw home mechanism [50] toward full extension. It has been suggested that these differences are because of increased conformity of the UKAMB design, which mimics the concavity of the anatomy of a native knee [19, 33]. In most UKAFB designs, including the one tested here, the concave anatomy is typically replaced with a flat polyethylene articulating surface, resulting in less AP constraint. Similarly, there were differences between the UKAFB and UKAMB designs; with UKAFB, the lateral femoral condyle paradoxically slid anteriorly, while it remained relatively stable with the UKAMB.

Tibial Cortical Bone Strain

Both implants demonstrated increased bone strain in the posterior and anteromedial tibial bone in deep flexion in comparison with the native knee. The biggest difference between both designs was noted in the anterolateral part of the medial compartment where UKAMB demonstrated bone strain close to the native knee, in contrast with a 79% increase in peak strain in UKAFB. Although there were differences in deep flexion between UKAMB and UKAFB in terms of the peak strain at the far medial side of the anterior tibia and in the posterior region, the UKAs demonstrated similar patterns and had only small differences in strain magnitude.

We observed no clear correlation between tibiofemoral kinematics and tibial strain for the tested UKA designs, based on the design and results of this study. With the numbers available, we found no obvious relationship between kinematics and strain behavior of these designs. One might expect that the consistently more posterior position of the medial femoral condyle center on tibia in UKAFB would lead to loads being transferred through the posterior aspect of the tibia and consequently lead to lower compressive strains on the medial tibia’s anterior side. This is the opposite of our most pronounced finding in terms of strain; that is, increased strain on the anterolateral aspect of the medial tibia in UKAFB. Instead, the observed strain differences might be associated with the orientation and magnitude of the quadriceps muscle force that is transferred to the tibia through the patellar tendon [12, 25]. Although the knee goes into flexion during squatting, the force the patellar tendon exerts on the tibia increases, along with a decreasing angle between the patellar tendon and the tibial axis in the sagittal plane (Fig. 5) [25]. Because of this decreased angle, the vertical component of the force vector increases, which further increases compressive strains on the tibia. Pegg et al. [34] reported that after UKA, strains around the patellar tendon’s insertion were increased, and these authors highlighted the impact of muscle forces on tibial strain. Regarding our kinematic findings, the more posterior position of the medial femoral condyle center in UKAFB might have been associated with a further increased vertical component of the patellar tendon’s force vector, perhaps explaining the increased strain at the anterolateral side of the medial tibia. However, because we unfortunately did not measure the position of the patella in our study, no other supporting data are available.

Fig. 5:
Schematic of the patellar tendon force, quadriceps tendon force and medial femoral condyle center in (A) full extension and (B) flexion at 90° after unicondylar knee arthroplasty using fixed-bearing implants (green) and mobile-bearing implants (red); PT = patellar tendon; QT = quadriceps tendon; FMCC = medial femoral condyle center; FB = fixed-bearing; MB = mobile-bearing.

In summary, in this in vitro cadaver study, both UKAMB and UKAFB led to an increase in bone strain in comparison with the native knee. However, in the anterolateral region of the medial tibial plateau, proximal tibial bone strain was lower after UKAMB than after UKAFB. As one of the most important factors for osteogenic adaptive response, increased bone strain after UKA has been suggested as a possible cause of unexplained pain [41, 48]. Further clinical research may discern whether the observed differences in cortical strain after UKA is associated with unexplained pain in patients and whether the observed differences in cortical bone strain between mobile-bearing and fixed unicondylar designs results in a further difference in unexplained pain.


We thank Prof. Geert Molenberghs PhD, Prof. Ben Van Calster PhD, and Lore Hermans MSc, for their support with statistics.


1. Al Nazer R, Lanovaz J, Kawalilak C, Johnston JD, Kontulainen S. Direct in vivo strain measurements in human bone — A systematic literature review. J Biomech. 2012;45:27-40.
2. Aleto TJ, Berend ME, Ritter MA, Faris PM, Meneghini RM. Early failure of unicompartmental knee arthroplasty leading to revision. J Arthroplasty. 2008;23:159-163.
3. Borus T, Thornhill T. Unicompartmental knee arthroplasty. J Am Acad Orthop Surg. 2008;16:9-18.
4. Burr DB, Milgrom C, Fyhrie D, Forwood M, Nyska M, Finestone A, Hoshaw S, Saiag E, Simkin A. In vivo measurement of human tibial strains during vigorous activity. Bone. 1996;18:405-410.
5. Callaghan JJ, Insall JN, Greenwald AS, Dennis DA, Komistek RD, Murray DW, Bourne RB, Rorabeck CH, Dorr LD. Mobile-bearing knee replacement: concepts and results. Instr Course Lect. 2001;50:431-449.
6. Capella M, Dolfin M, Saccia F. Mobile bearing and fixed bearing total knee arthroplasty. Ann Transl Med. 2016;4:1-9.
7. Chang TW, Yang CT, Liu YL, Chen WC, Lin KJ, Lai YS, Huang CH, Lu YC, Cheng CK. Biomechanical evaluation of proximal tibial behavior following unicondylar knee arthroplasty: modified resected surface with corresponding surgical technique. Med Eng Phys. 2011;33:1175-1182.
8. Completo A, Fonseca A, Semões JA. Finite element and experimental cortex strains of the intact and implanted tibia. J Biomech Eng. 2007;129:791-797.
9. Confalonieri N, Manzotti A, Pullen C. Comparison of a mobile with a fixed tibial bearing unicompartimental knee prosthesis: a prospective randomized trial using a dedicated outcome score. Knee. 2004;11:357-362.
10. Ettinger M, Zoch JM, Becher C, Hurschler C, Stukenborg-Colsman C, Claassen L, Ostermeier S, Calliess T. In vitro kinematics of fixed versus mobile bearing in unicondylar knee arthroplasty. Arch Orthop Trauma Surg. 2015;135:871-877.
11. Finlay JB, Bourne RB, McLean J. A technique for the in vitro measurement of principal strains in the human tibia. J Biomech. 1982;15:723-729.
12. Fyhrie DP, Milgrom C, Hoshaw SJ, Simkin A, Dar S, Drumb D, Burr DB. Effect of fatiguing exercise on longitudinal bone strain as related to stress fracture in humans. Ann Biomed Eng. 1998;26:660-665.
13. Gleeson RE, Evans R, Ackroyd CE, Webb J, Newman JH. Fixed or mobile bearing unicompartmental knee replacement? — A comparative cohort study. Knee. 2004;11:379-384.
14. Grassi L, Isaksson H. Extracting accurate strain measurements in bone mechanics: A critical review of current methods. J Mech Behav Biomed Mater. 2015;50:43-54.
15. Gray HA, Taddei F, Zavatsky AB, Cristofolini L, Gill HS. Experimental validation of a finite element model of a human cadaveric tibia. J Biomech Eng. 2008;130:1-9.
16. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J Biomech Eng. 1983;105:136-144.
17. Hannah RL, Reed SE. Strain gage users’ handbook: 1st ed, London, UK: Springer; 1992.
18. Hawi N, Plutat J, Kendoff D, Suero EM, Cross MB, Gehrke T, Citak M. Midterm results after unicompartmental knee replacement with all-polyethylene tibial component: a single surgeon experience. Arch Orthop Trauma Surg. 2016;136:1303-1307.
19. Heyse TJ, El-Zayat BF, De Corte R, Chevalier Y, Scheys L, Innocenti B, Fuchs-Winkelmann S, Labey L. UKA closely preserves natural knee kinematics in vitro. Knee Surg Sports Traumatol Arthrosc. 2014;22:1902-1910.
20. Heyse TJ, El-Zayat BF, De Corte R, Scheys L, Chevalier Y, Fuchs-Winkelmann S, Labey L. Balancing UKA: Overstuffing leads to high medial collateral ligament strains. Knee Surg Relat Res. 2016;24:3218-3228.
21. Heyse TJ, Slane J, Peersman G, Dworschak P, Fuchs-Winkelmann S, Scheys L. Balancing mobile-bearing unicondylar knee arthroplasty in vitro. Knee Surg Sports Traumatol Arthrosc. 2016;25:3733-3740.
22. Huang CH, Liau JJ, Chen CK. Fixed or mobile-bearing total knee arthroplasty. J Orthop Surg. 2007;2:1-8.
23. Iacono F, Raspugli GF, Akkawi I, Bruni D, Filardo G, Budeyri A, Bragonzoni L, Lo Presti M, Bonanzinga T, Marcacci M. Unicompartmental knee arthroplasty in patients over 75 years: a definitive solution? Arch Orthop Trauma Surg. 2016;136:117-123.
24. Innocenti B, Bilgen OF, Labey L, van Lenthe GH, Vander Sloten J, F C. Load sharing and ligament strains in balanced, overstuffed and understuffed UKA — a validated finite element analysis. J Arthroplasty. 2014;29:1491-1498.
25. Kernozek T, Torry M, Shelburne K, Durall CJ, Willson J. From the gait laboratory to the rehabilitation clinic: translation of motion analysis and modeling data to interventions that impact anterior cruciate ligament loads in gait and drop landing. Crit Rev Biomed Eng. 2013;41:243-258.
26. Ko YB, Gujarathi MR, Oh KJ. Outcome of unicompartmental knee arthroplasty: a systematic review of comparative studies between fixed and mobile bearings focusing on complications. Knee Surg Relat Res. 2015;27:141-148.
27. Kwon OR, Kang KT, Son J, Kwon SK, Jo SB, Suh DS, Choi YJ, Kim HJ, Koh YG. Biomechanical comparison of fixed- and mobile-bearing for unicomparmental knee arthroplasty using finite element analysis. J Orthop Res. 2014;32:338-345.
28. Rosette Kyowa. analysis of triaxial stacked KFGS gauges. 2019 Kyowa technical notes. Available at: Accessed January 22, 2019.
29. Lanyon LE, Hampson WG, Goodship AE, Shah JS. Bone deformation recorded in vivo from strain gauges attached to the human tibial shaft. Acta Orthop Scand. 1975;46:256-268.
30. Lee SY, Bae JH, Kim JG, Jang KM, Shon WY, Kim KW, Lim HC. The influence of surgical factors on dislocation of the meniscal bearing after Oxford medial unicompartmental knee replacement — a case-control study. Bone Joint J. 2014;96:914-922.
31. Marya SKS, Thukral R. Unicompartmental knee arthroplasty for tricompartment osteoarthritis in octogenarians. Indian J Orthop. 2009;43:361-366.
32. Nagaraja MP, Jo H. The role of mechanical stimulation in recovery of bone loss—high versus low magnitude and frequency of force. Life (Basel). 2014;4:117-130.
33. Peersman G, Slane J, Vuylsteke P, Fuchs-Winkelmann S, Dworschak P, Heyse T, Scheys L. Kinematics of mobile-bearing unicompartmental knee arthroplasty compared to native: results from an in vitro study. Arch Orthop Trauma Surg. 2017;137:1557-1563.
34. Pegg EC, Walter J, Mellon SJ, Pandit HG, Murray DW, D'Lima DD, Fregly BJ, Gill HS. Evaluation of factors affecting tibial bone strain after unicompartmental knee replacement. J Orthop Res. 2013;31:821-828.
35. Peterman MM, Hamel AJ, Cavanagh PR, Piazza SJ, Sharkey NA. In vitro modeling of human tibial strains during exercise in micro-gravity. J Biomech. 2001;34:693-698.
36. Price AJ, Svard U. A second decade lifetable survival analysis of the Oxford unicompartmental knee arthroplasty. Clin Orthop Relat Res. 2011;469:174-179.
37. Sarmah SS, Patel S, Hossain FS, Haddad FS. The radiological assessment of total and unicompartmental knee replacements. Bone Joint J.. 2012;94:1321-1329.
38. Schileo E, Taddei F, Malandrino A, Cristofolini L, Viceconti M. Subject-specific finite element models can accurately predict strain levels in long bones. J Biomech. 2007;40:2982-2989.
39. Sciammarella CA, Schiammarella FM. Experimental Mechanics of Solids. 1st ed. Chichester, UK, John Wiley & Sons L. 2012:63-64.
40. Scott CEH, Eaton MJ, Nutton RW, Wade FA, Evans SL, Pankaj P. Metal-backed versus all-polyethylene unicompartmental knee arthroplasty. Proximal tibial strain in an experimentally validated finite element model. Bone Joint Res. 2017;6:22-30.
41. Scott CEH, Eaton MJ, Nutton RW, Wade FA, Pankaj P, Evans SL. Proximal tibial strain in medial unicompartmental knee replacements — A biomechanical study of implant design. Bone Joint J. 2013; 95:1339-1347.
42. Simpson DJ, Price AJ, Gulati A, Murray DW, Gill HS. Elevated proximal tibial strains following unicompartmental knee replacement — a possible cause of pain. Med Eng Phys. 2009;31:752-757.
43. Small SR, Berend ME, Rogge RD, Archer DB, Kingman AL, Ritter MA. Tibial loading after UKA: evaluation of tibial slope, resection depth, medial shift and component rotation. J Arthroplasty. 2013;28:179-183.
44. Tjørnild M, Søballe K, Hansen PM, Holm C, Stilling M. Mobile- vs. fixed-bearing total knee replacement. A randomized radiostereometric and bone mineral density study. Acta Orthop. 2015;86:208-214.
45. Victor J, Van Doninck D, Labey L, Innocenti B, Parizel PM, Bellemans J. How precise can bony landmarks be determined on a CT scan of the knee? Knee. 2009;16:358-365.
46. Victor J, Van Glabbeek F, Vander Sloten J, Parizel PM, Somville J, Bellemans J. An experimental model for kinematic analysis of the knee. J Bone Joint Surg Am. 2009;91:150-163.
47. Walker T, Heinemann P, Bruckner T, Streit MR, Kinkel S, Gotterbarm T. The influence of different sets of surgical instrumentation in Oxford UKA on bearing size and component position. Arch Orthop Trauma Surg. 2017;137:895-902.
48. Yang PF, Brüggemann GP, Rittweger J. What do we currently know from in vivo bone strain measurements in humans? J Musculoskelet Neuronal Interact. 2011;11:8-20.
49. Zanasi S. Innovations in total knee replacement: new trends in operative treatment and changes in peri-operative management. Eur Orthop Traumatol. 2011;2:21-31.
50. Zhang LK, Wang XM, Niu YZ, Liu HX, Wang F. Relationship between patellar tracking and the "screw-home" mechanism of tibiofemoral joint. Orthop Surg. 2016;8:490-495.

Supplemental Digital Content

© 2020 by the Association of Bone and Joint Surgeons