KNEE
The anteroposterior axis of the tibia in Korean patients undergoing total knee replacement C. W. Kim, S. S. Seo, J. H. Kim, S. M. Roh, C. R. Lee From Inje University, Busan Paik Hospital, Busan, South Korea
The aim of this study was to find anatomical landmarks for rotational alignment of the tibial component in total knee replacement (TKR) in a CT-based study. Pre-operative CT scanning was performed on 94 South Korean patients (nine men, 85 women, 188 knees) with osteoarthritis of the knee joint prior to TKR. The tibial anteroposterior (AP) axis was defined as a line perpendicular to the femoral surgical transepicondylar axis and passing through the centre of the posterior cruciate ligament (PCL). The angles between the defined tibial AP axis and anatomical landmarks at various levels of the tibia were measured. The mean values of the angles between the defined tibial AP axis and the line connecting the anterior border of the proximal third of the tibia to the centre of the PCL was -0.2° (-17 to 14.1, SD 4.1). This was very close to the defined tibial axis, and remained so regardless of lower limb alignment and the degree of tibial bowing. Therefore, AP axis defined as described, is a reliable anatomical landmark for rotational alignment of tibial components. Cite this article: Bone Joint J 2014; 96-B:1485–90.
C. W. Kim , MD, PhD, Orthopaedic Surgeon, Professor J. H. Kim, MD, Orthopaedic Surgeon, Professor S. M. Roh, MD, Orthopaedic Surgeon C. R. Lee , MD , Orthopedic Surgeon Department of Orthopedic Surgery, Inje University, Busan Paik Hospital, College of Medicine, 633-166 Gaegeimdong, Busan Jin-gu, Busan 614735, South Korea. S. S. Seo, MD, PhD, Orthopaedic Surgeon Department of Orthopedic Surgery, Bumin Hospital, 380-4 Deocheon 1-dong, Buk-gu, Busan 616-819, South Korea. Correspondence should be sent to Dr C. R. Lee; e-mail: [email protected] ©2014 The British Editorial Society of Bone & Joint Surgery doi:10.1302/0301-620X.96B11. 33794 $2.00 Bone Joint J 2014;96-B:1485–90. Received 8 January 2014; Accepted after revision 18 July 2014
In total knee replacement (TKR) the rotational alignment of femoral and tibial components affect the overall function and durability of the implant. Malalignment between components can induce patellofemoral problems, aseptic loosening, polyethylene wear and instability.1-5 Previous studies have tended to focus on the rotational alignment of the femoral component.6-9 Tibial bowing is particularly common in Korean, Chinese and Japanese populations. Although the medial one-third of the tibial tubercle is widely used as an anatomical landmark for rotational alignment of tibial components,10 there have been reports suggesting that it is not appropriate when there are large variations in the proximal tibia among individuals.5,11-13 Alternative landmarks include the posterior condylar line of the tibia,11 the transcondylar line of the tibia,14 and the midsulcus of the tibial spine,15 but these are not easily identified during surgery.16 Although Akagi et al16 recently reported that the line that connects the centre of the posterior cruciate ligament (PCL) to the medial border of the patellar tendon at its insertion is a reliable landmark for determining the anteroposterior (AP) axis of the tibia, landmarks for the rotational alignment of the tibial component requires further consideration. The purpose of this study was to identify, on CT scans, reliable landmarks on the tibiae of elderly patients in a Korean population with
VOL. 96-B, No. 11, NOVEMBER 2014
osteoarthritis (OA) of the knee and to measure femorotibial angles and the degree of medial or lateral bowing of the tibia.
Patients and Methods The study was performed with 94 patients (188 knees) who underwent TKR between January 2012 and October 2013. Patients who had any congenital or pathological malformation or previous fractures of the lower limb were excluded. In addition, to maintain consistency in the CT scanning methodology, patients with flexion contractures ≥ 30° were excluded, as were patients with an unclear CT image of the sulcus on the medial epicondyle. The mean age of the patients was 68.7 years (56 to 80) and there were nine men (18 knees) and 85 women (170 knees). Their mean body mass index (BMI) was 25.8 (16.6 to 40.3). The study was approved by the local ethical committee. In all patients standing AP radiographs of the entire lower limbs were taken. The focus film distance was 180 cm. The patients were positioned standing with patella forward. Femorotibial angles were measured based on the anatomical axes of the femur and the tibia and patients were divided into a varus group (< 0°), a neutral group (0° to 6° valgus) and a valgus group (> 6°). In order to assess tibial bowing, the tibia was divided into four equal parts and the angles between the axis of the proximal one-quarter and the axis of the distal one1485
1486
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
Fig. 2a Fig. 1 Radiograph showing tibial bowing, which was calculated by measuring the angle between the axis of the proximal quarter and the axis of the distal quarter of the tibia.
quarter were measured (Fig. 1). The axis of the proximal one-quarter was defined as the line that connects the midpoints of the medial and lateral tibial spines and the midpoint of the intramedullary canal. The axis of the distal one-quarter was defined as the line that connects the midpoint of the intramedullary canal and the midpoint of the talar dome. Cases where the point of intersection of the extensions of the two axes was orientated laterally were defined as lateral bowing, and the opposite was defined as medial bowing. In order to analyse the relationships between the landmarks and the AP axis of the tibia at various levels measured by CT, the study population was divided into a group that showed ≤ 3° medial or lateral bowing and a group with > 3° medial or lateral bowing. CT (Discovery HD 750, GE Healthcare, Milwaukee, Wisconsin) images were taken with the patient in a supine position and both knees fixed in a neutral position using a frame, while maintaining full extension for as long as possible. The scans were performed at 2 mm intervals from approximately 10 cm above the knee joint line to the ankle joint. Using an Archiving and Communication System (PACS) (Marosis, Infinitt, Seoul, South Korea), two independent orthopaedic surgeons (CRL, SMR) measured the values on two separate occasions for each variable and recorded the mean value. In the CT axial images, the line that connects the sulcus of the medial epicondyle of the femur and the lateral epicondyle was defined as the surgical transepicondylar axis (SEA) (Fig. 2a), and this was projected on to the image at the level on the tibial plateau where the PCL can be identified at the posterior condylar notch. The line that is perpendicular to the SEA and passes through the centre of the PCL is defined as the AP axis of the tibia (Fig. 2b).
Fig. 2b CT axial images showing a) the line connecting the sulcus of the femoral medial epicondyle to the lateral epicondyle which was set up as the surgical transepicondylar axis (SEA) and b) the anteroposterior (AP) axis -– defined as the line that is perpendicular to the SEA and passes through the centre of the posterior cruciate ligament.
The defined AP axis can now be projected on to the image at the level of the patellar tendon–tibial attachment to measure the angle between the defined AP axis and the line connecting the medial one-third of the tibial tubercle to the PCL (angle A) and the angle between the defined AP axis and the line connecting the medial border of the patellar tendon to the PCL (angle B) (Fig. 3a). The position of the medial one-third of the patella tendon was measured instead of the tibial tubercle, as the border of the patella tendon was more easily identifiable on the CT images. The SEA was projected on to the image of the proximal onethird (angle C), midlevel (angle D) and distal one-third (angle E) of the tibia and the angles between the defined AP axis and the line connecting the most prominent anterior border to the PCL at each level (Fig. 3b to 3d). At the level of the ankle, the line connecting the lateral margin of the THE BONE & JOINT JOURNAL
THE ANTEROPOSTERIOR AXIS OF THE TIBIA IN KOREAN PATIENTS UNDERGOING TOTAL KNEE REPLACEMENT
Fig. 3a
Fig. 3c
1487
Fig.3b
Fig. 3d
CT images showing a) the angle between the defined anteroposterior (AP) axis and the line connecting the medial onethird of the tibial tubercle to the posterior cruciate ligament (PCL) at the patellar tendon attachment level (angle A) and the angle between the defined AP axis and the line connecting the medial border of the patellar tendon to the PCL (angle B); b) to d) the surgical epicodylar axis (SEA) was projected onto the images of proximal one-third, midlevel and distal one-third of the tibia. The angles between the defined AP axis and the line connecting the most prominent anterior border to the PCL at each level were measured (angles C, D and E) and e) at the ankle level, the line connecting the lateral margin of the extensor hallucis tendon to the centre of the PCL (angle F) was used as a reference line.
Fig. 3e
extensor hallucis tendon to the centre of the PCL (angle F) was used as a reference (Fig. 3e). Statistical analysis. Statistical analyses were conducted using SAS 9.3 software (SAS Institute Inc., Cary, North Carolina). As the data were not normally distributed the VOL. 96-B, No. 11, NOVEMBER 2014
Kruskal–Wallis test was used to compare angles between the groups classified based on femorotibial angles (varus, neutral and valgus). A Bonferroni correction was used for ex post facto tests (varus vs neutral, varus vs valgus and neutral vs valgus) (Table I). Wilcoxon’s rank sum tests were
1488
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
Table I. Comparison of mean angles (standard deviation (SD) with ranges) between the groups classified based on femorotibial angles Lower limb alignment (femorotibial angle) Total (mean, range, SD) (n = 188)
Varus (mean, range, SD) (n = 46)
A
7.5 (-10.6 to 20.2, SD 5.0)
8.5 (-2.4 to 17.2, SD 4.4)
8.0 (-3.2 to 20.2, SD 4.9)
6.0 (-10.6 to 18.9, SD 5.5)
0.067
B
-1.5 (-16.9 to 9.7, SD 4.3)
-0.5 (-8.3 to 9.6, SD 3.5,)
-1.3 (-13.1 to 9.7, SD 4.0)
-2.3 (-16.9 to 7.6, SD 5.1)
0.475
C
-0.2 (-17.0 to 14.1, SD 4.1)
1.7 (-17.0 to 14.1, SD 4.7)
-0.1 (-14.9 to 12.2, SD 4.2)
-1.8 (-14.3 to 11.2, SD 3.5)
0.025
Angle (°)
Neutral (mean, range, SD) (n = 81)
Valgus (mean, range, SD) (n = 61)
p-value (Kruskal–Wallis test)
p-value (post hoc test)
Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324 D
-8.1 (-40 to 26.3, SD 11.7)
-3.7 (-40 to 25.4, SD 12.4)
-7.9 (-28.2 to 26.3, SD 11.3)
-11.4 (-39.1 to 10.0, SD 10.5)
0.035 Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324
E
-23.1 (-65.9 to 5.3, SD 17.3)
-18.6 (-51.6 to 13.2, SD 17.6)
-21.3 (-46.6 to 10.8, SD 17.4)
-28.4 (-65.9 to 5.3, SD 16.8)
0.022 Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324
F
-49.1 (-80.9 to -11.7, SD 17.6)
-48.7 (-74.9 to -16.0, SD 15.9)
-46.1 (-80.9 to -11.7, SD 18.2)
-53.4 (-77.4 to -13.4, SD 18.9)
0.123
Reference point of each angle; A, medial one-third tibial tubercle; B, medial border of the patellar tendon; C, most prominent anterior border of the tibia at proximal one-third level; E, most prominent anterior border of the tibia at distal one-third level F, lateral margin of the extensor hallucis tendon at ankle level
Table II. Comparison of mean angles (standard deviation (SD)) between the groups classified based on the degree of tibial bowing Angle (°)
0° to 3° (n = 113)
> 3° (n = 75)
p-value (Wilcoxon’s rank sum test)
A B C D E F
7.9 (SD 4.9) –1.3 (SD 4.3) 0.1 (SD 4.6) –6.9 (SD 11.9) –20.1 (SD 17.5) –47.5 (SD 17.4)
6.9 (SD 5.2) –1.7 (SD 4.2) –0.7 (SD 3.3) –9.8 (SD 11.3) –27.5 (SD 16.9) –51.5 (SD 17.9)
0.115 0.495 0.320 0.086 0.012 0.03
Reference point of each angle: A, medial one-third tibial tubercle; B, medial border of the patellar tendon; C, most prominent anterior border of the tibia at proximal one-third level; D, most prominent anterior border of the tibia at mid level; E, most prominent anterior border of the tibia at distal one-third level; F, lateral margin of the extensor hallucis tendon at ankle level
used to compare angles between the groups based on the degree of tibial bowing. A p-value < 0.05 was considered statistically significant.
Results The mean femorotibial angle was neutral (0° to 6° valgus) in 81 knees (mean 3.4° varus, -13.0° to 15.8°), was varus (< 0°) in 46 knees (mean -3.9°, -13.0° to -0.1°) and was valgus (> 6°) in 61 knees (mean 8.8°, 6.1° to 15.8°). As for tibial bowing, 113 cases had medial or lateral bowing within a range of 0° to 3° and 75 cases had bowing > 3° (-6.1 to 7.6). The mean values of the angles between the defined tibial AP axis and the lines connecting the medial one-third of the tibial tuberosity or the medial border of the patellar tendon at the level of the patellar tendon attachment to the centre of the PCL were respectively 7.5° (-10.6° to 20.2°, SD 5.0)
and –1.5° (-16.9° to 9.7°, SD 4.3). The mean values of the angles between the defined tibial AP axis and the lines connecting the most prominent anterior border at the proximal one-third, mid-level, distal one-third or ankle-level of the tibia to the centre of the PCL are shown in Table I. In the group with varus alignment, the mean value of angle B was -0.5° (-8.3° to 9.6°, SD 3.5) and was closest to the defined tibial AP axis, and mean angle C was 1.7° (-17.0° to 14.1°, SD 4.7). In the group with neutral alignment and the group with valgus alignment, the mean angle C was -0.1° (-14.9° to 12.2°, SD 4.2) and -1.8° (-14.3° to 11.2°, SD 3.5) respectively, and was closest to the defined tibial AP axis. In all three groups the anatomical landmarks showed a tendency to rotate internally as their locations moved from proximal to distal levels. In the comparison of measured angles between the groups classified according to THE BONE & JOINT JOURNAL
THE ANTEROPOSTERIOR AXIS OF THE TIBIA IN KOREAN PATIENTS UNDERGOING TOTAL KNEE REPLACEMENT
the degree of tibial bowing, there were no statistical differences between angles A, B, C and D, but there were significant differences between angles E and F (Table II).
Discussion We found that the described perpendicular line connecting the centre of the PCL to the anterior tibial border at the proximal one-third level of the tibia is the closest to its true AP axis and can be used as a reliable landmark for the rotational alignment of the tibial component. In contrast, the line connecting the medial one-third of the tibial tubercle to the PCL was found to be externally rotated by 7.5°. The medial border of the patellar tendon was internally rotated by approximately 1.5° at the level of attachment of the patellar tendon, and this was the next most accurate landmark. However, it was slightly internally rotated compared with the results of a study performed by Akagi et al16 using younger Japanese volunteers, who were from similar in racial origin. This difference is considered to be attributable to the fact that our patients were elderly with OA. In all the groups classified according to femorotibial angle, angle C was the closest to the defined tibial AP axis. The value of angle C was the closest to the tibial AP axis within a mean of -0.1° when there was neutral alignment (0° to 6° valgus), it was externally rotated by a mean of 1.7° in the group with varus alignment (< 0°), and internally rotated by a mean of 1.8° in the group with valgus alignment (> 6°). In a study on a Chinese population, Sun et al5 argued that proximal tibia was externally rotated in patients with varus or valgus deformity compared with those with neutral alignment. However, the measurement levels and reference points they used were different from ours. On reviewing the relationships between the landmarks and the tibial AP axis at various levels of the tibia, it is clear that the landmarks rotate internally as their locations move more distally. In a study conducted in Japanese people with normal knees and Japanese patients with OA, Nagamine et al13 reported that on CT scans, the distal part of the tibia was medially twisted compared with the proximal part in patients with neutral and valgus alignment. Given these study results, it can be expected that if the medial one-third of the tibial tubercle is used as a reference point for the rotational alignment of the tibial component, then the internal rotation of the feet will be worsened and may affect gait. Based on our results in patients with severe varus deformity in the proximal part of the tibia, the reference line for defining the AP axis is located in the distal part of the tibia on the lateral side. There is no widely accepted standard for the measurement and extent of tibial bowing. Although Yau et al17 divided the tibial diaphysis into three equal parts to measure tibial bowing, we chose to divide it into four parts, to better reflect the degree of deformity of the entire tibia. We also divided the data into one group with mild deformity and a more severely deformed group. The perpendicular VOL. 96-B, No. 11, NOVEMBER 2014
1489
line extending from the PCL and intersectiong the tibial anterior border in the proximal one-third of the tibia was shown to be the closest to the tibial AP axis in both groups, and there was no statistically significant difference between them. Accordingly, the rotational alignment of the tibial component can be based on this line, as it is not greatly affected by the degree of tibial bowing. A limitation of this study is that the number of patients was small, however, it is nevertheless relatively large compared with other studies,5,16 and our patients were elderly with OA and were actually undergoing TKR. Other limitations include intra- and inter-observer variability. In addition, there may have been errors in the measurement of the CT images where patients with flexion contractures could not fully extend their knee joints. We do not think that there are absolute landmarks for the rotational alignment of the tibial component, and the most appropriate landmarks should be used depending on the patient’s anatomical morphology as observed before or during surgery, and that the line connecting the anterior border of the proximal third of the tibia to the centre of the PCL can have a role as one of those landmarks. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by D. Rowley and first proof edited by G. Scott.
References 1. Barrack RL, Schrader T, Bertot AJ, et al. Component rotation and anterior knee pain after total knee arthroplasty. Clin Orthop Relat Res 2001;392:46–55. 2. Berger RA, Crossett LS, Jacobs JJ, et al. Malrotation causing patellofemoral complications after total knee arthroplasty. Clin Orthop Relat Res 1998;356:144–153. 3. Lewis P, Rorabeck CH, Bourne RB, et al. Posteromedial tibial polyethylene failure in total knee replacements. Clin Orthop Relat Res 1994;299:11–17. 4. Merkow RL, Soudry M, Insall JN, et al. Patellar dislocation following total knee replacement. J Bone Joint Surg [Am] 1985;67-A:1321–1327. 5. Sun T, Lu H, Hong N, Wu J, Feng C. Bony landmarks and rotational alignment in total knee arthroplasty for Chinese osteoarthritic knees with varus or valugs deformities. J Arthroplasty 2009;24:427–431. 6. Stiehl JB, Abbott BD. Morphology of the transepicondylar axis and its application in primary and revision total knee arthroplasty. J Arthroplasty 1995;10:785–789. 7. Hungerford DS, Krackow KA. Total joint arthroplasty of the knee. Clin Orthop Relat Res 1985;192:23–33. 8. Churchill DL, Incavo SJ, Johnson CC, Beynnon BD. The transepicondylar axis approximates the optimal flexion axis of the knee. Clin Orthop Relat Res 1998;356:111–118. 9. Miller MC, Berger RA, Petrella AJ, et al. Optimizing femoral component rotation in total knee arthroplasty. Clin Orthop Relat Res 2001;392:38–45. 10. Insall JN, Windsor RE, Scott WN, et al. Surgical techniques and instrumentation in total knee arthroplasty. In: Surgery of the Knee. Second ed. New York: Churchill-Livingstone, 1993:739-804. 11. Moreland JR. Mechanisms of failure in total knee arthroplasty. Clin Orthop Relat Res 1988;226:49–64. 12. Howel SM, Chen J, Hull ML. Variability of the location of the tibial tubercle affects the rotational alignment of the tibial component in kinematically aligned total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2013;21:2288–2295. 13. Nagamine R, Miyanishi K, Miura H, et al. Medial torsion of the tibia in Japanese patients with osteoarthritis of the knee. Clin Orthop Relat Res 2003;408:218–224.
1490
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
14. Eckhoff DG, Johnston RJ, Stamm ER, et al. Version of the osteoarthritic knee. J Arthroplasty 1994;9:73–79. 15. Dalury DF. Observations of the proximal tibia in total knee arthroplasty. Clin Orthop Relat Res 2001;389:150–155.
16. Akagi M, Oh M, Nonaka T, et al. An anteroposterior axis of the tibia for total knee arthroplasty. Clin Orthop Relat Res 2004;420:213–219. 17. Yau WP, Chiu KY, Tang WM, Ng TP. Coronal bowing of the femur and tibia in Chinese: its incidence and effects on total knee arthroplasty planning. J Orthop Surg 2007;15:32–36.
THE BONE & JOINT JOURNAL
The anteroposterior axis of the tibia in Korean patients undergoing total knee replacement C. W. Kim, S. S. Seo, J. H. Kim, S. M. Roh, C. R. Lee From Inje University, Busan Paik Hospital, Busan, South Korea
The aim of this study was to find anatomical landmarks for rotational alignment of the tibial component in total knee replacement (TKR) in a CT-based study. Pre-operative CT scanning was performed on 94 South Korean patients (nine men, 85 women, 188 knees) with osteoarthritis of the knee joint prior to TKR. The tibial anteroposterior (AP) axis was defined as a line perpendicular to the femoral surgical transepicondylar axis and passing through the centre of the posterior cruciate ligament (PCL). The angles between the defined tibial AP axis and anatomical landmarks at various levels of the tibia were measured. The mean values of the angles between the defined tibial AP axis and the line connecting the anterior border of the proximal third of the tibia to the centre of the PCL was -0.2° (-17 to 14.1, SD 4.1). This was very close to the defined tibial axis, and remained so regardless of lower limb alignment and the degree of tibial bowing. Therefore, AP axis defined as described, is a reliable anatomical landmark for rotational alignment of tibial components. Cite this article: Bone Joint J 2014; 96-B:1485–90.
C. W. Kim , MD, PhD, Orthopaedic Surgeon, Professor J. H. Kim, MD, Orthopaedic Surgeon, Professor S. M. Roh, MD, Orthopaedic Surgeon C. R. Lee , MD , Orthopedic Surgeon Department of Orthopedic Surgery, Inje University, Busan Paik Hospital, College of Medicine, 633-166 Gaegeimdong, Busan Jin-gu, Busan 614735, South Korea. S. S. Seo, MD, PhD, Orthopaedic Surgeon Department of Orthopedic Surgery, Bumin Hospital, 380-4 Deocheon 1-dong, Buk-gu, Busan 616-819, South Korea. Correspondence should be sent to Dr C. R. Lee; e-mail: [email protected] ©2014 The British Editorial Society of Bone & Joint Surgery doi:10.1302/0301-620X.96B11. 33794 $2.00 Bone Joint J 2014;96-B:1485–90. Received 8 January 2014; Accepted after revision 18 July 2014
In total knee replacement (TKR) the rotational alignment of femoral and tibial components affect the overall function and durability of the implant. Malalignment between components can induce patellofemoral problems, aseptic loosening, polyethylene wear and instability.1-5 Previous studies have tended to focus on the rotational alignment of the femoral component.6-9 Tibial bowing is particularly common in Korean, Chinese and Japanese populations. Although the medial one-third of the tibial tubercle is widely used as an anatomical landmark for rotational alignment of tibial components,10 there have been reports suggesting that it is not appropriate when there are large variations in the proximal tibia among individuals.5,11-13 Alternative landmarks include the posterior condylar line of the tibia,11 the transcondylar line of the tibia,14 and the midsulcus of the tibial spine,15 but these are not easily identified during surgery.16 Although Akagi et al16 recently reported that the line that connects the centre of the posterior cruciate ligament (PCL) to the medial border of the patellar tendon at its insertion is a reliable landmark for determining the anteroposterior (AP) axis of the tibia, landmarks for the rotational alignment of the tibial component requires further consideration. The purpose of this study was to identify, on CT scans, reliable landmarks on the tibiae of elderly patients in a Korean population with
VOL. 96-B, No. 11, NOVEMBER 2014
osteoarthritis (OA) of the knee and to measure femorotibial angles and the degree of medial or lateral bowing of the tibia.
Patients and Methods The study was performed with 94 patients (188 knees) who underwent TKR between January 2012 and October 2013. Patients who had any congenital or pathological malformation or previous fractures of the lower limb were excluded. In addition, to maintain consistency in the CT scanning methodology, patients with flexion contractures ≥ 30° were excluded, as were patients with an unclear CT image of the sulcus on the medial epicondyle. The mean age of the patients was 68.7 years (56 to 80) and there were nine men (18 knees) and 85 women (170 knees). Their mean body mass index (BMI) was 25.8 (16.6 to 40.3). The study was approved by the local ethical committee. In all patients standing AP radiographs of the entire lower limbs were taken. The focus film distance was 180 cm. The patients were positioned standing with patella forward. Femorotibial angles were measured based on the anatomical axes of the femur and the tibia and patients were divided into a varus group (< 0°), a neutral group (0° to 6° valgus) and a valgus group (> 6°). In order to assess tibial bowing, the tibia was divided into four equal parts and the angles between the axis of the proximal one-quarter and the axis of the distal one1485
1486
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
Fig. 2a Fig. 1 Radiograph showing tibial bowing, which was calculated by measuring the angle between the axis of the proximal quarter and the axis of the distal quarter of the tibia.
quarter were measured (Fig. 1). The axis of the proximal one-quarter was defined as the line that connects the midpoints of the medial and lateral tibial spines and the midpoint of the intramedullary canal. The axis of the distal one-quarter was defined as the line that connects the midpoint of the intramedullary canal and the midpoint of the talar dome. Cases where the point of intersection of the extensions of the two axes was orientated laterally were defined as lateral bowing, and the opposite was defined as medial bowing. In order to analyse the relationships between the landmarks and the AP axis of the tibia at various levels measured by CT, the study population was divided into a group that showed ≤ 3° medial or lateral bowing and a group with > 3° medial or lateral bowing. CT (Discovery HD 750, GE Healthcare, Milwaukee, Wisconsin) images were taken with the patient in a supine position and both knees fixed in a neutral position using a frame, while maintaining full extension for as long as possible. The scans were performed at 2 mm intervals from approximately 10 cm above the knee joint line to the ankle joint. Using an Archiving and Communication System (PACS) (Marosis, Infinitt, Seoul, South Korea), two independent orthopaedic surgeons (CRL, SMR) measured the values on two separate occasions for each variable and recorded the mean value. In the CT axial images, the line that connects the sulcus of the medial epicondyle of the femur and the lateral epicondyle was defined as the surgical transepicondylar axis (SEA) (Fig. 2a), and this was projected on to the image at the level on the tibial plateau where the PCL can be identified at the posterior condylar notch. The line that is perpendicular to the SEA and passes through the centre of the PCL is defined as the AP axis of the tibia (Fig. 2b).
Fig. 2b CT axial images showing a) the line connecting the sulcus of the femoral medial epicondyle to the lateral epicondyle which was set up as the surgical transepicondylar axis (SEA) and b) the anteroposterior (AP) axis -– defined as the line that is perpendicular to the SEA and passes through the centre of the posterior cruciate ligament.
The defined AP axis can now be projected on to the image at the level of the patellar tendon–tibial attachment to measure the angle between the defined AP axis and the line connecting the medial one-third of the tibial tubercle to the PCL (angle A) and the angle between the defined AP axis and the line connecting the medial border of the patellar tendon to the PCL (angle B) (Fig. 3a). The position of the medial one-third of the patella tendon was measured instead of the tibial tubercle, as the border of the patella tendon was more easily identifiable on the CT images. The SEA was projected on to the image of the proximal onethird (angle C), midlevel (angle D) and distal one-third (angle E) of the tibia and the angles between the defined AP axis and the line connecting the most prominent anterior border to the PCL at each level (Fig. 3b to 3d). At the level of the ankle, the line connecting the lateral margin of the THE BONE & JOINT JOURNAL
THE ANTEROPOSTERIOR AXIS OF THE TIBIA IN KOREAN PATIENTS UNDERGOING TOTAL KNEE REPLACEMENT
Fig. 3a
Fig. 3c
1487
Fig.3b
Fig. 3d
CT images showing a) the angle between the defined anteroposterior (AP) axis and the line connecting the medial onethird of the tibial tubercle to the posterior cruciate ligament (PCL) at the patellar tendon attachment level (angle A) and the angle between the defined AP axis and the line connecting the medial border of the patellar tendon to the PCL (angle B); b) to d) the surgical epicodylar axis (SEA) was projected onto the images of proximal one-third, midlevel and distal one-third of the tibia. The angles between the defined AP axis and the line connecting the most prominent anterior border to the PCL at each level were measured (angles C, D and E) and e) at the ankle level, the line connecting the lateral margin of the extensor hallucis tendon to the centre of the PCL (angle F) was used as a reference line.
Fig. 3e
extensor hallucis tendon to the centre of the PCL (angle F) was used as a reference (Fig. 3e). Statistical analysis. Statistical analyses were conducted using SAS 9.3 software (SAS Institute Inc., Cary, North Carolina). As the data were not normally distributed the VOL. 96-B, No. 11, NOVEMBER 2014
Kruskal–Wallis test was used to compare angles between the groups classified based on femorotibial angles (varus, neutral and valgus). A Bonferroni correction was used for ex post facto tests (varus vs neutral, varus vs valgus and neutral vs valgus) (Table I). Wilcoxon’s rank sum tests were
1488
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
Table I. Comparison of mean angles (standard deviation (SD) with ranges) between the groups classified based on femorotibial angles Lower limb alignment (femorotibial angle) Total (mean, range, SD) (n = 188)
Varus (mean, range, SD) (n = 46)
A
7.5 (-10.6 to 20.2, SD 5.0)
8.5 (-2.4 to 17.2, SD 4.4)
8.0 (-3.2 to 20.2, SD 4.9)
6.0 (-10.6 to 18.9, SD 5.5)
0.067
B
-1.5 (-16.9 to 9.7, SD 4.3)
-0.5 (-8.3 to 9.6, SD 3.5,)
-1.3 (-13.1 to 9.7, SD 4.0)
-2.3 (-16.9 to 7.6, SD 5.1)
0.475
C
-0.2 (-17.0 to 14.1, SD 4.1)
1.7 (-17.0 to 14.1, SD 4.7)
-0.1 (-14.9 to 12.2, SD 4.2)
-1.8 (-14.3 to 11.2, SD 3.5)
0.025
Angle (°)
Neutral (mean, range, SD) (n = 81)
Valgus (mean, range, SD) (n = 61)
p-value (Kruskal–Wallis test)
p-value (post hoc test)
Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324 D
-8.1 (-40 to 26.3, SD 11.7)
-3.7 (-40 to 25.4, SD 12.4)
-7.9 (-28.2 to 26.3, SD 11.3)
-11.4 (-39.1 to 10.0, SD 10.5)
0.035 Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324
E
-23.1 (-65.9 to 5.3, SD 17.3)
-18.6 (-51.6 to 13.2, SD 17.6)
-21.3 (-46.6 to 10.8, SD 17.4)
-28.4 (-65.9 to 5.3, SD 16.8)
0.022 Varus vs neutral 0.510 Varus vs valgus 0.020 Neutral vs valgus 0.324
F
-49.1 (-80.9 to -11.7, SD 17.6)
-48.7 (-74.9 to -16.0, SD 15.9)
-46.1 (-80.9 to -11.7, SD 18.2)
-53.4 (-77.4 to -13.4, SD 18.9)
0.123
Reference point of each angle; A, medial one-third tibial tubercle; B, medial border of the patellar tendon; C, most prominent anterior border of the tibia at proximal one-third level; E, most prominent anterior border of the tibia at distal one-third level F, lateral margin of the extensor hallucis tendon at ankle level
Table II. Comparison of mean angles (standard deviation (SD)) between the groups classified based on the degree of tibial bowing Angle (°)
0° to 3° (n = 113)
> 3° (n = 75)
p-value (Wilcoxon’s rank sum test)
A B C D E F
7.9 (SD 4.9) –1.3 (SD 4.3) 0.1 (SD 4.6) –6.9 (SD 11.9) –20.1 (SD 17.5) –47.5 (SD 17.4)
6.9 (SD 5.2) –1.7 (SD 4.2) –0.7 (SD 3.3) –9.8 (SD 11.3) –27.5 (SD 16.9) –51.5 (SD 17.9)
0.115 0.495 0.320 0.086 0.012 0.03
Reference point of each angle: A, medial one-third tibial tubercle; B, medial border of the patellar tendon; C, most prominent anterior border of the tibia at proximal one-third level; D, most prominent anterior border of the tibia at mid level; E, most prominent anterior border of the tibia at distal one-third level; F, lateral margin of the extensor hallucis tendon at ankle level
used to compare angles between the groups based on the degree of tibial bowing. A p-value < 0.05 was considered statistically significant.
Results The mean femorotibial angle was neutral (0° to 6° valgus) in 81 knees (mean 3.4° varus, -13.0° to 15.8°), was varus (< 0°) in 46 knees (mean -3.9°, -13.0° to -0.1°) and was valgus (> 6°) in 61 knees (mean 8.8°, 6.1° to 15.8°). As for tibial bowing, 113 cases had medial or lateral bowing within a range of 0° to 3° and 75 cases had bowing > 3° (-6.1 to 7.6). The mean values of the angles between the defined tibial AP axis and the lines connecting the medial one-third of the tibial tuberosity or the medial border of the patellar tendon at the level of the patellar tendon attachment to the centre of the PCL were respectively 7.5° (-10.6° to 20.2°, SD 5.0)
and –1.5° (-16.9° to 9.7°, SD 4.3). The mean values of the angles between the defined tibial AP axis and the lines connecting the most prominent anterior border at the proximal one-third, mid-level, distal one-third or ankle-level of the tibia to the centre of the PCL are shown in Table I. In the group with varus alignment, the mean value of angle B was -0.5° (-8.3° to 9.6°, SD 3.5) and was closest to the defined tibial AP axis, and mean angle C was 1.7° (-17.0° to 14.1°, SD 4.7). In the group with neutral alignment and the group with valgus alignment, the mean angle C was -0.1° (-14.9° to 12.2°, SD 4.2) and -1.8° (-14.3° to 11.2°, SD 3.5) respectively, and was closest to the defined tibial AP axis. In all three groups the anatomical landmarks showed a tendency to rotate internally as their locations moved from proximal to distal levels. In the comparison of measured angles between the groups classified according to THE BONE & JOINT JOURNAL
THE ANTEROPOSTERIOR AXIS OF THE TIBIA IN KOREAN PATIENTS UNDERGOING TOTAL KNEE REPLACEMENT
the degree of tibial bowing, there were no statistical differences between angles A, B, C and D, but there were significant differences between angles E and F (Table II).
Discussion We found that the described perpendicular line connecting the centre of the PCL to the anterior tibial border at the proximal one-third level of the tibia is the closest to its true AP axis and can be used as a reliable landmark for the rotational alignment of the tibial component. In contrast, the line connecting the medial one-third of the tibial tubercle to the PCL was found to be externally rotated by 7.5°. The medial border of the patellar tendon was internally rotated by approximately 1.5° at the level of attachment of the patellar tendon, and this was the next most accurate landmark. However, it was slightly internally rotated compared with the results of a study performed by Akagi et al16 using younger Japanese volunteers, who were from similar in racial origin. This difference is considered to be attributable to the fact that our patients were elderly with OA. In all the groups classified according to femorotibial angle, angle C was the closest to the defined tibial AP axis. The value of angle C was the closest to the tibial AP axis within a mean of -0.1° when there was neutral alignment (0° to 6° valgus), it was externally rotated by a mean of 1.7° in the group with varus alignment (< 0°), and internally rotated by a mean of 1.8° in the group with valgus alignment (> 6°). In a study on a Chinese population, Sun et al5 argued that proximal tibia was externally rotated in patients with varus or valgus deformity compared with those with neutral alignment. However, the measurement levels and reference points they used were different from ours. On reviewing the relationships between the landmarks and the tibial AP axis at various levels of the tibia, it is clear that the landmarks rotate internally as their locations move more distally. In a study conducted in Japanese people with normal knees and Japanese patients with OA, Nagamine et al13 reported that on CT scans, the distal part of the tibia was medially twisted compared with the proximal part in patients with neutral and valgus alignment. Given these study results, it can be expected that if the medial one-third of the tibial tubercle is used as a reference point for the rotational alignment of the tibial component, then the internal rotation of the feet will be worsened and may affect gait. Based on our results in patients with severe varus deformity in the proximal part of the tibia, the reference line for defining the AP axis is located in the distal part of the tibia on the lateral side. There is no widely accepted standard for the measurement and extent of tibial bowing. Although Yau et al17 divided the tibial diaphysis into three equal parts to measure tibial bowing, we chose to divide it into four parts, to better reflect the degree of deformity of the entire tibia. We also divided the data into one group with mild deformity and a more severely deformed group. The perpendicular VOL. 96-B, No. 11, NOVEMBER 2014
1489
line extending from the PCL and intersectiong the tibial anterior border in the proximal one-third of the tibia was shown to be the closest to the tibial AP axis in both groups, and there was no statistically significant difference between them. Accordingly, the rotational alignment of the tibial component can be based on this line, as it is not greatly affected by the degree of tibial bowing. A limitation of this study is that the number of patients was small, however, it is nevertheless relatively large compared with other studies,5,16 and our patients were elderly with OA and were actually undergoing TKR. Other limitations include intra- and inter-observer variability. In addition, there may have been errors in the measurement of the CT images where patients with flexion contractures could not fully extend their knee joints. We do not think that there are absolute landmarks for the rotational alignment of the tibial component, and the most appropriate landmarks should be used depending on the patient’s anatomical morphology as observed before or during surgery, and that the line connecting the anterior border of the proximal third of the tibia to the centre of the PCL can have a role as one of those landmarks. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This article was primary edited by D. Rowley and first proof edited by G. Scott.
References 1. Barrack RL, Schrader T, Bertot AJ, et al. Component rotation and anterior knee pain after total knee arthroplasty. Clin Orthop Relat Res 2001;392:46–55. 2. Berger RA, Crossett LS, Jacobs JJ, et al. Malrotation causing patellofemoral complications after total knee arthroplasty. Clin Orthop Relat Res 1998;356:144–153. 3. Lewis P, Rorabeck CH, Bourne RB, et al. Posteromedial tibial polyethylene failure in total knee replacements. Clin Orthop Relat Res 1994;299:11–17. 4. Merkow RL, Soudry M, Insall JN, et al. Patellar dislocation following total knee replacement. J Bone Joint Surg [Am] 1985;67-A:1321–1327. 5. Sun T, Lu H, Hong N, Wu J, Feng C. Bony landmarks and rotational alignment in total knee arthroplasty for Chinese osteoarthritic knees with varus or valugs deformities. J Arthroplasty 2009;24:427–431. 6. Stiehl JB, Abbott BD. Morphology of the transepicondylar axis and its application in primary and revision total knee arthroplasty. J Arthroplasty 1995;10:785–789. 7. Hungerford DS, Krackow KA. Total joint arthroplasty of the knee. Clin Orthop Relat Res 1985;192:23–33. 8. Churchill DL, Incavo SJ, Johnson CC, Beynnon BD. The transepicondylar axis approximates the optimal flexion axis of the knee. Clin Orthop Relat Res 1998;356:111–118. 9. Miller MC, Berger RA, Petrella AJ, et al. Optimizing femoral component rotation in total knee arthroplasty. Clin Orthop Relat Res 2001;392:38–45. 10. Insall JN, Windsor RE, Scott WN, et al. Surgical techniques and instrumentation in total knee arthroplasty. In: Surgery of the Knee. Second ed. New York: Churchill-Livingstone, 1993:739-804. 11. Moreland JR. Mechanisms of failure in total knee arthroplasty. Clin Orthop Relat Res 1988;226:49–64. 12. Howel SM, Chen J, Hull ML. Variability of the location of the tibial tubercle affects the rotational alignment of the tibial component in kinematically aligned total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2013;21:2288–2295. 13. Nagamine R, Miyanishi K, Miura H, et al. Medial torsion of the tibia in Japanese patients with osteoarthritis of the knee. Clin Orthop Relat Res 2003;408:218–224.
1490
C. W. KIM, S. S. SEO, J. H. KIM, S. M. ROH, C. R. LEE
14. Eckhoff DG, Johnston RJ, Stamm ER, et al. Version of the osteoarthritic knee. J Arthroplasty 1994;9:73–79. 15. Dalury DF. Observations of the proximal tibia in total knee arthroplasty. Clin Orthop Relat Res 2001;389:150–155.
16. Akagi M, Oh M, Nonaka T, et al. An anteroposterior axis of the tibia for total knee arthroplasty. Clin Orthop Relat Res 2004;420:213–219. 17. Yau WP, Chiu KY, Tang WM, Ng TP. Coronal bowing of the femur and tibia in Chinese: its incidence and effects on total knee arthroplasty planning. J Orthop Surg 2007;15:32–36.
THE BONE & JOINT JOURNAL
