SEARCH

SEARCH BY CITATION

Keywords:

  • tibiofemoral joint;
  • kinematics;
  • osteoarthritis;
  • total knee arthroplasty;
  • weight-bearing flexion

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Total knee arthroplasty (TKA) is a widely accepted surgical procedure for the treatment of patients with end-stage osteoarthritis (OA). However, the function of the knee is not always fully recovered after TKA. We used a dual fluoroscopic imaging system to evaluate the in vivo kinematics of the knee with medial compartment OA before and after a posterior cruciate ligament-retaining TKA (PCR-TKA) during weight-bearing knee flexion, and compared the results to those of normal knees. The OA knees displayed similar internal/external tibial rotation to normal knees. However, the OA knees had less overall posterior femoral translation relative to the tibia between 0° and 105° flexion and more varus knee rotation between 0° and 45° flexion, than in the normal knees. Additionally, in the OA knees the femur was located more medially than in the normal knees, particularly between 30° and 60° flexion. After PCR-TKA, the knee kinematics were not restored to normal. The overall internal tibial rotation and posterior femoral translation between 0° and 105° knee flexion were dramatically reduced. Additionally, PCR-TKA introduced an abnormal anterior femoral translation during early knee flexion, and the femur was located lateral to the tibia throughout weight-bearing flexion. The data help understand the biomechanical functions of the knee with medial compartment OA before and after contemporary PCR-TKA. They may also be useful for improvement of future prostheses designs and surgical techniques in treatment of knees with end-stage OA. © 2010 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 29:40–46, 2011

Posterior cruciate ligament retaining total knee arthroplasty (PCR-TKA) is an accepted treatment for end-stage knee osteoarthritis. Though excellent long-term survivorship and pain relief after PCR-TKA were reported,1–3 knee function is not always fully recovered. The loss of deep flexion,4 patellofemoral joint complications,5 and polyethylene wear6 remain the most common postoperative problems. Numerous in vitro and in vivo studies reported abnormal knee kinematics after PCR-TKA, including paradoxical anterior femoral motion and decreased internal tibial rotation during knee flexion.7–13

Evidence is emerging that better kinematic patterns after TKA may help patients in their functional performance.4, 14–16 However, the ability of contemporary TKAs to restore kinematics towards normal is still not fully understood.17, 18 Many in vitro studies compared knee kinematics before and after TKA, however, the knees were not at the end-stage of OA.19–21 Most in vivo studies assessed the kinematics of either OA knees18, 22, 23 or TKA knees.7, 9 Recently, one in vivo study evaluated the AP femoral translation and axial rotation of the knee before and after TKA under weight-bearing knee flexion.13 A few intraoperative compared passive knee flexion kinematics before and after TKA using surgical navigation systems.17, 24, 25 However, no study has quantitatively compared the 3D kinematics of normal knees, OA knees, and PCR-TKA knees during weight-bearing activity.

We compared in vivo knee kinematics before and after PCR-TKA in patients with medial compartment OA, and compared the data to the kinematics of a group of healthy control subjects using consistent coordinate systems. The goal was to determine how medial compartment OA affects kinematics during weight-bearing knee flexion and if contemporary PCR-TKA can restore the kinematics towards normal. We hypothesized that the undesirable knee kinematics following PCR-TKA are more due to the surgery than the long standing disease.

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Subject Demographics

Eleven patients with advanced medial compartment knee OA were recruited (age 64 ± 7 years, range 51–73; height 68.3 ± 3.9 in., range 61–73; weight 208.2 ± 31.9 lbs, range 155–265; 6 left and 5 right knees; Table 1). All subjects had patellofemoral disease, and 5 had ACL deficiency. Hence they were not candidates for unicompartmental knee arthroplasty. The medial menisci were severely degenerated in the 5 ACL-deficient knees. The lateral menisci and PCLs were intact in all the knees. The preoperative Knee Society Score was 55.3 ± 12.9 (range, 39–70), and the Knee Society Function Score was 50.0 ± 19.6 (range, 25–75).26 The preoperative average weight-bearing range of motion (ROM) was 102.8 ± 22.3° (1.8 ± 5.2°/104.6 ± 20.4°), and the average varus deformity was 1.9 ± 2.8° (range, 1.0–7.8°). Patients with post-traumatic arthritis, rheumatoid arthritis, or valgus knees were excluded.

Table 1. Demographics for All 11 OA Patients
PTAge/SexKL GradePre-Operative ROM (Ext/Flex) (°)Pre-Operative KSS (KS/FS)Pre-Operative ACL/PCL StatusPost-Operative ROM (Ext/Flex) (°)Post-Operative KSS (KS/FS)
  1. PT, patient; KL grade, Kellgren–Lawrence grade; ROM, range of motion; Ext, extension; Flex, flexion; KSS, Knee Society score; KS, knee score; FS, function score; ACL, anterior cruciate ligament; PCL, posterior cruciate ligament.

167/MIV12/11340/25Ruptured/intact−3/11198/90
256/MIV−2/11361/75Ruptured/intact−3/10395/90
365/MIV1/10964/60Intact/intact0/10192/90
473/MIII1/8040/35Intact/intact6/11090/85
564/MIV−6/12760/45Ruptured/intact−8/10385/95
672/MIII5/13070/70Intact/intact7/10993/90
751/MIV−1/10157/70Ruptured/intact−2/10997/90
862/FIV3/6539/25Intact/intact−3/5860/50
973/FIII0/10167/60Intact/intact−14/91100/100
1061/FIII−3/12370/60Intact/intact−1/11397/95
1161/FIV9/8840/25Ruptured/intact5/9490/70

Twenty-two healthy control subjects (10 females and 12 males; 12 left and 10 right knees; age 31.4 ± 9.3 years, range 19–51; height 67.9 ± 4.1 in., range 61–75; weight 166.9 ± 31 lbs, range 110–215) from a group of 24 subjects studied in a previous investigation were included as normal references.27 The current work does not include all 24 subjects because the current study has been ongoing for a longer duration, and the control group had been fixed. The OA patients were recruited from the practice of a single surgeon. Prior to the study, Institutional Review Board approval and informed patient consent were obtained for all subjects.

Preoperative MRI scans of each OA knee were obtained using a 3.0 Tesla magnet (Siemens, Malvern, PA) using a fat suppressed 3D spoiled gradient-recalled sequence. Sagittal plane image slices (1 mm spacing, resolution of 512 × 512 pixels, field of view 180 mm × 180 mm) were segmented using 3D modeling software (Rhinoceros®, Robert McNeel and Assoc, Seattle, WA) to construct 3D models, including the tibia and femur.27

The OA patients then performed a weight-bearing single leg lunge from full extension to maximal flexion in a dual fluoroscopic imaging system (Fig. 1a). During the lunge, the subjects positioned the knee in the field of view of the two fluoroscopes with their feet and torsos oriented towards the intersection of the two image intensifiers.27 The subjects were instructed to support their body weight on the leg being studied, but were allowed to use the contralateral leg and handrails for balance. As the subjects slowly bent their knees, the fluoroscopes captured knee positions at every 15° increment. At each target flexion angle, the subjects held their knee position for about 2 sec while images were captured. All the OA knees received a high-flexion PCR-TKA (NexGen CR-Flex, Zimmer, Warsaw, IN). The surgeries were performed by the same surgeon following standard procedures. Briefly, femoral cuts were performed first with the aid of an intramedullary guide to determine varus-valgus alignment, and the epicondylar axis was referenced for rotational alignment. The tibial cut was made with a 7° posterior slope using an extramedullary system. The tibial crest and the center of the tibial plateau were used as reference points. The tibial component was aligned with the junction of the medial and middle thirds of the tibial tubercle. Extension-flexion gaps were always well balanced without PCL recession.

thumbnail image

Figure 1. (a) Schema of a patient performing the single leg lunge inside the dual fluoroscopic imaging system. (b) The virtual environment used to reproduce the subject's joint kinematics.

Download figure to PowerPoint

All patients returned for postoperative tests after surgery (average time 8 ± 2.5 months; range, 7–15 months). All patients were clinically successful with no ligamentous laxity or pain. Postoperative TKA knee kinematics were obtained by matching the pre-operative bone models to the fluoroscopic images using the remaining regions of the femur and tibia. Thus, the pre- and post-TKA kinematics could be studied using the same bone embedded coordinate system.

In Vivo Kinematics

The fluoroscopic images were imported into Rhino software to establish a virtual fluoroscopic setup (Fig. 1b). The set of images were placed at their respective positions in the virtual dual fluoroscopic imaging system.27 Next, the 3D MR image-based bone models were imported and individually manipulated until they matched their projections on the fluoroscopic images captured during the weight-bearing activity (Fig. 1b). Thus, in vivo knee motion was represented by a series of 3D models at different flexion angles.28 For kinematics measurements after PCR-TKA, the preoperative models were used for matching.

Six-DOF kinematics were obtained directly from the series of 3D knee models, using the coordinate systems embedded in the distal femur and proximal tibia (Fig. 2).27, 29 In the sagittal plane, the tibial long axis passed through the middle of the tibial spines and was adjusted to be parallel to the posterior wall of the tibial shaft. In the coronal plane it was angled equally with respect to the medial and lateral edges of the shaft. An orthogonal coordinate system was placed on the tibia with the mediolateral (ML) axis obtained by projecting a line passing through the centers of the plateaus onto a plane perpendicular to the long axis. The center of the each plateau was defined as the centroid of the closed curve formed by tracing the edges of the plateau. The midpoint of the tibial ML axis was defined as the origin of the tibial coordinate system. The proximal-distal axis of the tibia was parallel to the tibial long axis. The tibial AP axis was defined to be perpendicular to the tibial ML and proximal-distal axes.

thumbnail image

Figure 2. The kinematics of the OA, TKA, and normal knees were analyzed under the consistent coordinate systems embedded in the bones.

Download figure to PowerPoint

For the femur, two axes were defined, the transepicondylar axis and the femoral long axis. The transepicondylar axis was defined as a line joining the most prominent points on the condyles. The femoral long axis was a line passing through the midpoint of the transepicondylar axis. In the sagittal plane, the femoral long axis was placed where the angle formed by the long axis and the anterior edge of the femoral shaft was equal to the angle formed by the long axis and the posterior edge of the shaft. Similarly, the femoral long axis was angled equally with respect to the medial and lateral edges of the femoral shaft in the coronal plane.

Knee flexion was defined as the angle between the femoral and tibial long axes projected onto the tibial sagittal plane. Internal-external (IE) tibial rotation was defined as the angle between the transepicondylar axis and the tibial ML axis, projected onto the transverse plane of the tibia.27, 29 Varus-valgus (VV) rotation was defined as the angle between the transepicondylar axis and the tibial ML axis, projected onto the coronal plane of the tibia. Flexion, internal tibial rotation, and varus rotation were defined to be positive. The midpoint of the transepicondylar axis was used to measure femoral translations in the AP and ML directions with respect to the tibial coordinate system. Anterior and medial femoral translations were defined to be positive. Additionally, we found a lateral shift of the tibial component, and the distance between the medial edge of the tibial bone surface and the medial edge of the component was measured (the lateral shift of the component) (Fig. 3). The distance between the midpoint of the femoral component and the midpoint of the femoral bone surface was also measured.

thumbnail image

Figure 3. AP radiograph and 3D models of a TKA knee. The tibial component is lateralized.

Download figure to PowerPoint

Statistical Analysis

Kinematics were compared at exact 15° flexion increments. The measured raw data were interpolated to calculate kinematics at exact angles. Two-way ANOVA followed by post hoc Newman–Keuls test was used to detect differences in AP, ML, IE, and VV motions among the groups (OA, PCR-TKA, and normal). p < 0.05 was considered significant. We used the Student's t-test to compare overall axial tibial rotation and AP femoral translation among groups. Best-fit line representing the relationship between ranges of motion (ROM) of the OA knees before and after PCR-TKA was calculated using least-squares regression.

RESULTS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

After PCR-TKA, the Knee Society Knee Score averaged 90.6 ± 11.0 (range, 60–100), and the Knee Society Function Score averaged 85.9 ± 14.1 (range, 70–100)26 (Table 1). Average weight-bearing ROM was 101.7 ± 5.1° (−1.5 ± 6.2°/100.2 ± 15.5°) (Table 1). The preoperative ROM was positively correlated to postoperative ROM (r2 = 0.55, p < 0.01) (Fig. 4).

thumbnail image

Figure 4. The line represents the average postoperative knee ROM, which was positively correlated with preoperative ROM (r2 = 0.55, p < 0.05).

Download figure to PowerPoint

At full extension, the femur of OA knees was located more posteriorly than in normal knees (3.3 ± 4.5 mm vs. 0.2 ± 2.4 mm, p < 0.01) (Fig. 5a). During flexion, the femurs of both OA and normal knees translated posteriorly, but the translation of OA knees (12.6 ± 4.2 mm) was significantly smaller than that of normals (18.1 ± 2.5 mm, p < 0.01). After PCR-TKA, the femur was located significantly more posteriorly (7.4 ± 3.3 mm) compared to either OA or normal knees. During flexion, PCR-TKA knees showed anterior femoral translation until about 40° flexion; thereafter the femur moved posteriorly. The posterior translation of PCR-TKAknees(6.0 ± 2.8 mm) was significantly smaller than either OA (p < 0.01) or normal knees (p < 0.01).

thumbnail image

Figure 5. AP (a) and ML (b) femoral translation during weight-bearing flexion in OA, TKA, and normal knees. *Pre versus post; pre versus normal; post versus normal, p < 0.05. [Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com]

Download figure to PowerPoint

In the ML direction, the femurs of OA and normal knees maintained a relatively constantly medial position relative to the tibia throughout flexion (Fig. 5b). In OA knees, the femur was located more medial than in normal knees between 30° and 60° flexion (p < 0.05). Conversely, after PCR-TKA, the femur was located lateral to the tibia throughout flexion. The average location was about 9 mm lateral to that in the OA knees, and about 6 mm lateral to that in the normal knees. The distance between the medial edge of the tibial component and the medial edge of the tibial bone surface was 6.0 ± 2.8 mm. The femoral component was basically placed in the neutral position relative to the distal femoral resection surface (0.7 ± 1.3 mm).

The magnitude of internal tibial rotation during early flexion (from full extension to 30°) was similar between the OA (5.2 ± 5.1°) and normal knees (7.2 ± 4.6°, p > 0.05) (Fig. 6a). The overall internal tibial rotations of the OA and normal knees were also similar (8.2 ± 5.7° for the OA and 10.8 ± 4.6° for the normal knees, p > 0.05). After PCR-TKA, the internal tibial rotation was −0.1 ± 1.9° from full extension to 30°. The overall internal tibial rotation decreased to 3.4 ± 6.0°, which was significantly smaller than those of the OA (p < 0.01) and normal knees (p < 0.01).

thumbnail image

Figure 6. IE tibial (a) and VV knee (b) rotation during weight-bearing flexion in OA, TKA, and normal knees. *Pre versus post; pre versus normal; post versus normal, p < 0.05. [Color figure can be viewed in the online issue, which is available at http://www.interscience.wiley.com]

Download figure to PowerPoint

For VV rotation, normal knees were in neutral position at full extension, while OA knees showed 1.9 ± 2.8° of varus rotation (Fig. 6b). During flexion, the VV rotation of normal knees varied from neutral to slightly varus. The varus rotation of the OA knees decreased at greater flexion angles. Following PCR-TKA, the knees had 3.1 ± 2.4° of valgus rotation at full extension, but it gradually changed to slightly varus rotation with further flexion.

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

This study evaluated knee kinematics with medial compartment OA before and after PCR-TKA, and compared them to those of normal knees. The data indicated that OA knees were more similar to normal knees in kinematics between 0° and 105° flexion than the knees post PCR-TKA. This supports our hypothesis that undesirable knee kinematics following PCR-TKA are more due to the surgery than the long standing disease. Our results also confirm a positive correlation between pre- and postoperative ROMs.4

Physiological posterior femoral translation is believed to help clear the posterior aspect of the knee and thus enhance flexion.14 We found significant differences in posterior translation from full extension to maximal flexion between normal, OA, and PCR-TKA knees. Other studies also reported less posterior translation in OA compared to normal knees.13, 18, 22 In PCR-TKA knees, the reduction of posterior femoral translation was been reported by many studies.13, 24, 25 For example, Kitagawa et al. observed posterior translation of the lateral condyle from full extension to maximum flexion reduced from 9 ± 1 mm for OA knees to 5 ± 0 mm for TKA knees. The medial femoral condyle showed little posterior motion before and after TKA.13 A systematic literature review showed that posterior femoral translation from full extension to 90° flexion after TKA was always <10 mm in both in vitro and in vivo studies, but was 5.1–16.7 mm (in vivo) and 10–16.8 mm (in vitro) for normal knees.8, 19, 30 Considering the similar ROM and the different magnitudes of posterior femoral translation of the OA knees before and after PCR-TKA in our study, a quantitative correlation between posterior femoral translation and ROM of the knee is not demonstrated in this group of patients.

We also observed a paradoxical femoral translation during early flexion in the PCR-TKA knees. This phenomenon was noted in the literature10–12, 25 and might be partially attributed to ACL resection in PCR-TKA knees,14 since it is also observed in ACL deficient knees.29, 31 At low flexion angles, the quadriceps pulls the tibia anteriorly, which is resisted by the ACL. ACL resection could lead to an excessive posterior femoral position in early flexion. This kinematic behavior is not observed in posterior stabilized knees because it is prevented by the cam-post mechanism.14, 32 The paradoxical motion has potential negative consequences, such as limited flexion, decreased quadriceps efficiency, and accelerated polyethylene wear.4, 31

In the ML direction, our data showed that while the femurs of the OA and normal knees were located medial to the tibia throughout flexion, the femur of the PCR-TKA knees was located lateral (about 6 mm) to the tibia. An in vitro study also showed that the femur after PCR-TKA was located lateral to its position in the intact knee (about 2 mm).33 Differences between the in vivo and in vitro studies may be attributed to different loading conditions.

The lateral femoral shift in the PCR-TKA knees may be due to component design and surgical technique. The symmetry of the tibial component does not match the resected tibial plateau.34 In general, lateral placement of the tibial component is preferable since medial overhang can cause painful irritation of the MCL.35 The lateralized tibial component may affect the ML femoral position due to the ML constraint of the NexGen design. Previous in vitro studies demonstrated that ACL provides resistance to lateral translation.36, 37 In ACL deficient knees, a lateral shift of the femur occurs relative to normal knees.36, 37 The constraining function of the ACL was lost in our PCR-TKAs, which might have caused the lateral femoral translation.

TKAs function within relatively normal anatomic and physiologic boundaries.14 Therefore, the laterally shifted femur may trend to move back “home” due to ligament and muscle forces. Future studies should examine the long-term effect of the lateralization of the tibial component on biomechanical outcomes of PCR-TKAs.

In the present study, the OA knees showed internal tibial rotation from full extension to 30° flexion, similar to normal knees, referred to as “screw-home” motion.8, 38 However, other studies reported different tibial rotation patterns in OA knees.2, 17 Siston et al.17 found internal rotation of 4.9 ± 4.1° between 10 and 90° flexion, significantly less than the 10.1 ± 4.2° of normal cadaveric knees in the same flexion range. Hamai et al.22 observed tibial rotation ranging from 4 ± 2° at 20° flexion to 15 ± 2° at 100° flexion in advanced OA knees during squatting. Kitagawa et al.13 reported 13 ± 6° internal rotation of OA knees from full extension to maximum flexion of 125°, which compared to 8.2 ± 5.7° of rotation for OA knees in our study. The differences might be due to variations in activities, choice of coordinate systems, and preoperative pathological changes.13, 18, 22, 23, 39

After PCR-TKAs, we did not observe the characteristic internal tibial rotation between full extension and 30° flexion seen in normal knees. The overall rotation between full extension and maximum flexion also decreased substantially relative to preoperative values for the OA knees. This is consistent with other studies reporting either a substantial reduction or a loss of internal tibial rotation after PCR-TKA.7–9, 13 The internal tibial rotation in early flexion is attributed to the function of the ACL and the asymmetry of the tibial plateaus.40 Therefore, ACL resection and loss of medial-lateral asymmetry following PCR-TKAs may change tibial rotation.40

Though the starting positions at extension were different in the three groups (varus for OA, neutral for normal, and valgus for PCR-TKA), they all changed to slightly varus at deep flexion, consistent with other studies.17, 41 As VV rotation directly affects the contact force distribution among the medial and lateral compartments, future studies should compare the effects of medial compartment OA and PCR-TKAs on force distribution in the joint.

Our study was limited by the relatively small number of patients. However, knee kinematics before and after the PCR-TKA were measured, which increased the statistical power, allowing us to detect differences between the two conditions. Nonetheless, we did not have sufficient numbers to delineate the effect of ACL status. Our preliminary analysis showed that the differences between the ACL deficient and ACL intact OA knees were subtle compared to the differences between OA and TKA knees. Another limitation is that we used one posterior cruciate-retaining design. Future work will involve other designs. We also only examined knee function during a quasi-static single leg lunge, which does not represent kinematics during a dynamic activity. Future studies should examine functional activities such as gait and stair ascent and descent. Finally, we only examined effect of ML tibial component position on ML location of the femur post-TKA. Other factors such as AP and rotational placement of the components were not investigated.

In conclusion, our results showed that the OA knees were similar to the normal knees in terms of IE tibial rotation during weight-bearing flexion between 0 and 105° flexion. However, compared to normal knees, OA knees had reduced overall posterior femoral translation. Also, the OA knees showed more varus orientation in early flexion (<45°), and the femur in the OA knee was located more medial to the tibia than in the normal knees, particularly from 30 to 60° flexion. Following PCR-TKA, knee kinematics were not restored to normal. Compared to both OA and normal knees, the PCR-TKA had significantly lower amounts of internal tibial rotation and posterior femoral translation. The PCR-TKA also introduced an abnormal anterior femoral translation during flexion between 0 and 30° flexion. Also, while the femur was located medially in OA and normal knees, in the PCR-TKA knees the femur was located lateral to the tibia. Our data showed that changes in knee kinematics following contemporary PCR-TKA are attributable more to the surgery the preoperative disease state. Therefore, continued improvements in implant designs and surgical technique may better restore normal knee biomechanics.

Acknowledgements

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

This study was supported by an AEF grant of Department of Orthopaedics, MGH. The authors acknowledge support of the Program for Shanghai Key Laboratory of Orthopaedic Implant (08DZ2230330) and the National Natural Science Foundation of China (30901517).

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES