Arch Orthop Trauma Surg (2015) 135:697–701
DOI 10.1007/s00402-015-2157-2
KNEE ARTHROPLASTY
Improved tibial component rotation in TKA using patient-specific
instrumentation
Thomas J. Heyse • Carsten O. Tibesku
Received: 5 December 2014 / Published online: 1 April 2015
Ó Springer-Verlag Berlin Heidelberg 2015
Abstract
Introduction Patient-specific instrumentation (PSI) was
introduced in an attempt to reduce positional outliers of
components in total knee arthroplasty (TKA). It was hy-
pothesized that PSI could help with the positioning of tibial
components in optimal rotational alignment.
Methods A magnetic resonance imaging (MRI) analysis
of 58 patients following TKA was conducted. Of these, 30
operations were performed using PSI and 28 using con-
ventional instrumentation. The rotation of the tibial com-
ponents was determined in MRI using three different
reference lines: a tangent to the dorsal tibial condyles, the
tibial epicondylar line, and the tibial tubercle. Deviations
[9° were considered outliers. Also internal rotation [1°
was considered an outlier. Data were analyzed statistically
for positional outliers using the Chi-squared test.
Results There was excellent inter- and intraobserver re-
liability with low standard deviations for the determination
of tibial component rotation using the tangent to the dorsal
condyles and the tibial epicondylar line as reference. Using
the dorsal tangent as reference, there were eight compo-
nents in excessive external rotation (28.6 %) and one
component being in relative internal rotation (5.4°) in the
conventional group, while there were two components in
excessive external rotation in the PSI group (6.7 %). Using
T. J. Heyse
Department of Orthopedics and Rheumatology,
University Hospital Marburg, Baldingerstrasse,
35043 Marburg, Germany
e-mail: heyse@med.uni-marburg.de; t.heyse@gmx.de
C. O. Tibesku (&)
Sporthopaedicum Straubing, Bahnhofplatz 27,
94315 Straubing, Germany
e-mail: carsten@tibesku.de
the tibial epicondyles as reference, there were seven com-
ponents in excessive external rotation (21.4 %) and one
component being in relative internal rotation (4.4°) in the
conventional group; while there were two components in
excessive external rotation in the PSI group (6.7 %). These
differences were statistically significant (p \ 0.05). Mea-
surements based on the tibial tubercle showed poor repro-
ducibility in terms of intra- and interobserver reliability and
was of little use in the context of the research question.
Discussion and conclusion In this setup, PSI was effec-
tive in significantly reducing outliers of optimal rotational
tibial component alignment during TKA. Anatomy of the
proximal tibia does not deliver clear landmarks that are
prominent and consistent. This makes both, MRI analysis
as well as cutting jig production and intraoperative place-
ment a challenge.
Keywords Patient-matched cutting jigs
Tibial
component rotation
TKA
PSI
MRI
Introduction
Rotational implant alignment is important for long-term
success of total knee arthroplasty (TKA). It has been re-
ported that more than half of the painful TKA have internal
rotational errors in the alignment of components [1]. In
another study, all included TKA revised for stiffness had
components positioned in excessive internal rotation [2].
Especially for the femoral component, it has been shown
that when internally rotated, this may provoke (anterior)
knee pain, synovitis and patellofemoral complications such
as subluxation, dislocation and eccentric wear [3–5].
Symptoms were reported to be reversible by revision of
malrotated components [6, 7].
123
698
Also for the tibial component, it was shown that internal
rotation is related to clinical complaints such as postop-
erative knee stiffness [2] and pain [1, 8]. In one study, it
was estimated that at least 4.6 % of their TKA had been
implanted with significant internal rotational tibial com-
ponent errors [1]. In the same study, excessive external
rotation of the tibial component was not found to be as-
sociated with pain. The amount of internal rotation of the
tibial component required to produce clinical symptoms
seems to be unknown but Nicoll et al. [1] suggested nine
degrees of internal rotation as cut off level.
It was proposed that small degrees of internal rotation
would result in patellar maltracking and pain, whereas
large amounts could add to the development of postop-
erative stiffness being too painful or impossible for patients
to bend their knee well [2].
The incidence of excessive internal rotation of the tibial
component may be underestimated and underreported. In
one study, it was more than double that of excessive in-
ternal rotation of the femoral component and the size of
internal rotational errors of the tibial component was much
greater than that of the femoral component [1].
Berger et al. [3, 4] proposed the concept of an additive
effect of internal malrotational alignment of both tibial and
femoral components. Barrack et al. [8] found that patients
with combined component internal rotation were more than
five times as likely to experience anterior knee pain after
TKA compared to those with combined component exter-
nal rotation. An internal rotational mismatch of the tibial
component greater than 11° in relation to the femoral
component has also been described to be associated with a
higher likelihood of pain following TKA [1].
Patient-specific instrumentation (PSI) was introduced in
an attempt to reduce positional outliers of components in
total knee arthroplasty (TKA). In an earlier study, PSI was
shown to reduce femoral component rotation outliers in
comparison with the conventional technique [9]. No such
advantage for PSI was found in another recent publication
[10]. There is a definite paucity on data analyzing the re-
liability of PSI in reduction of rotational positional outliers
of components in TKA.
Magnetic resonance imaging (MRI) has been shown to be
an effective tool for analysis of rotational position of TKA
components [9, 11, 14]. It was hypothesized that PSI could
help with the positioning of tibial components in optimal
rotational alignment as assessed in postoperative MRI.
Materials and methods
The study was performed in analogy to earlier publications
that were partially based on the same patient cohort [9, 12]. A
retrospective analysis of MRI of the knee following TKA was
123
Arch Orthop Trauma Surg (2015) 135:697–701
performed. MRI had been performed at average
6.1 ± 6.9 months following TKA. Fifty-eight patients (30
women and 28 men) with an average age of 65.3 ± 8.2 years
at index TKA were included (26 left, 32 right knees). Mean
age in the PSI group was 66.5 ± 8.2 years (12 men, 18
women) and in the conventional group, it was 64.3 ± 9 years
(16 men, 12 women). All patients underwent cemented pos-
terior stabilized (PS) TKA (Genesis IITM, Smith and Nephew,
Memphis, TN, USA) for degenerative joint disease. The
modular tibial components consist of a PE insert and a tita-
nium/aluminum/vanadium (TiAlV) alloy base plate. Of the
included patients, 30 operations were performed using PSI
and 28 using conventional instrumentation.
The patients were informed that radiological data of
their cases were to be submitted for publication. The study
follows the principles set forth in the Declaration of Hel-
sinki and received institutional research board approval.
Postoperative MRI was performed as described in an
earlier publication [13]. Patients were placed supine on the
scanning table with the extremity in relative extension and
slight external rotation. Coronal, sagittal, and axial images
were obtained in a standardized fashion. All imaging was
performed in a 1.5 T superconducting magnet (Magnetom
Espree, Siemens Medical Systems, Erlangen, Germany).
The knee was placed in a transmit/receive extremity coil (CP
Extremity, Siemens Medical Systems, Erlangen, Germany).
Axial fast spin echo images were obtained with a TR/TE of
4000–5000/34, 12–20 echo train length, 3 mm slice thick-
ness with no interslice gap, 62.5–100 kHz bandwidth over
the entire frequency range, 512 9 288–320 matrix,
20–22 cm field of view at 3 excitations. These were used for
measurements as mentioned below.
Three different methods were applied to measure tibial
component rotation using a tangent to the tibial keel as a
reference (Fig. 1) [12]: a tangent to the dorsal tibial con-
dyles (Fig. 1), the tibial transepicondylar axis as proposed
by Bonnin et al. [14] for both (Fig. 2), and the tibial tu-
bercle (Fig. 3) following a technique proposed for CT
scans by Berger et al. [4]. For the latter, neutral rotation of
the tibial component is defined to be 18° of internal rota-
tion from the center of the tip of the tuberosity in corre-
spondence to the rotation of the native knee.
Readers were free to choose the axial level for their
measurements as it was felt, that this best simulates the
clinical situation. The Synedra ViewPersonal 3, Version
3.4.0.2 (Munich, Germany) was used for image processing.
The PACS software was mediDOK, Dossenheim, Germany.
Statistical analysis
Statistical analysis was performed as described in an earlier
publication [9]. Continuous variables were shown as mean
and standard deviation. Categorical data were given in
Arch Orthop Trauma Surg (2015) 135:697–701
Fig. 1 Analysis of tibial component rotation using the dorsal tangent
to the tibial epicondyles as reference in a right knee: the component
was measured to be in 5.7° of external rotation
Fig. 2 Analysis of tibial component rotation using the transepi-
condylar as reference in a left knee: the component was measured to
be in 3.9° of external rotation
absolute figures. The MRI studies were evaluated two
times independently by the two authors with an interval of
4 weeks. The authors were blinded to the clinical infor-
mation. The mean values of their results for measurements
of tibial component rotation were used as the basis for
assessment of intra- and interobserver reliability by cal-
culation of intra-class correlation coefficients (ICC). They
were calculated by one-way analysis of variance [ICC
699
(1, 1)] and two-way random effects analysis of variance
with random observers [ICC (2, 1)], respectively. Values
[0.75 are considered to be significant. This method had
been proven to be effective in an earlier study, especially
for femoral components made of zirconium/niobium alloy
[14]. A Chi-squared test was performed to analyze differ-
ences in positional outliers between groups. For analysis of
data, p \ 0.05 was considered as statistically significant.
Statistical analysis was supported by using Microsoft
Excel (Microsoft Corporation, Seattle, USA) and IBM
SPSS Statistics 18 (PASW 18, SPSS Inc., Chicago, IL,
USA).
Results
There was excellent inter- and intraobserver reliability with
low standard deviations for the determination of tibial
component rotation using the tangent to the dorsal
condyles (intraobserver reliability 0.967, mean difference
-0.6° ± 1.8°, and interobserver reliability 0.956, mean
difference -0.3° ± 1.7°) and the tibial epicondylar line
(intraobserver reliability 0.946, mean difference
0.0° ± 1.8°, and interobserver reliability 0.958, mean dif-
ference -0.2° ± 1.6°) as reference.
Using the dorsal tangent as reference, there were eight
components in excessive external rotation (28.6 %) and
one component being in relative internal rotation (5.4°) in
the conventional group; while there were two components
in excessive external rotation in the PSI group (6.7 %) and
none in relative internal rotation (p \ 0.01).
Using the tibial epicondyles as reference, there were
seven components in excessive external rotation (21.4 %)
and one component being in relative internal rotation (4.4°)
in the conventional group; while there were two compo-
nents in excessive external rotation in the PSI group
(6.7 %) and none in internal rotation. These differences
were statistically significant (p \ 0.05).
Measurements based on the tibial tubercle showed poor
reproducibility in terms intra- and interobserver reliability
(ICC 0.665 and 0.566) and was of little use in the context
of the research question.
Discussion
In this setup, PSI was effective in significantly reducing
outliers of optimal rotational tibial component alignment
during TKA and the initial hypothesis was proven correct.
Of note is that, in both group, almost all outliers were in
excessive external rotation, which may have less detri-
mental impact on the function following TKA than internal
rotation.
123
700
Fig. 3 Analysis of tibial component rotation using the tip of the tibial
tuberosity (a) as reference in the same left knee as Fig. 2: a line
between the tip of and the geometric center of the tibial plateau as
As reported before, there are several limitations to
studies that assess component alignment in MRI [9]: ret-
rospective radiological data are presented with no corre-
lation with clinical symptoms. The dorsal tangent to the
tibial epicondyles and the tibial epicondylar line have never
been used to correlate clinical symptoms with rotational
component alignment. Measurements of angles in MRI will
always remain somewhat subjective and dependent on
observers and technique, which was tried to be overcome
by determining inter- and intraobserver reliability. It should
also be mentioned that optimal tibial component rotation in
TKA not only depends on defining the landmarks and re-
spective positioning of the tibial tray. The surgeon also
needs to consider optimal bone coverage to avoid compo-
nent overhang, especially when using the conventional
technique. It should be considered that data originate from
a high volume practice from a single experienced surgeon.
The effect of PSI may in fact prove to be more pronounced
in the hand of the less experienced surgeon in a low volume
practice. The patient numbers included into this study re-
main relatively small as the performance of postoperative
MRI is an economic and logistic challenge. However, the
included numbers seem to be large enough to prove the
initially formulated hypothesis.
There are three other publications on this subject. Par-
ratte et al. [10] were not able to show any advantages for
PSI in terms of rotational component outlier reduction. In
their study, they reported on 40 patients randomized to
receive TKA (NexGen LPS-Flex mobile, Zimmer, War-
saw, IN, USA) via conventional or PSI technique. Tibial
component rotation as measured in pre- and postoperative
CT scans was determined using the tibial tuberosity as
reference. They stated that there were no significant
123
Arch Orthop Trauma Surg (2015) 135:697–701
determined by the center of the overlying circle is drawn (b). The
component is calculated to be in 0.1° of external rotation (-90°) and
after adjustment (-18°) as proposed by Berger et al. [3, 4]
differences in mean tibial rotation, while 35 % of the
conventionally implanted tibias vs. 20 % of the PSI tibias
were implanted in internal rotation. They concluded that
PSI was not helpful. Tests for inter- and intraobserver re-
liability of rotational measurements were not made.
Silva et al. [15] reported on a prospectively randomized
study with 22 patients receiving conventional instrumen-
tation and 23 patients receiving patient-specific instru-
mentation (Signature, Biomet, Warsaw, IN, USA). Tibial
component rotation was analyzed on computed tomogra-
phy using the tibial tuberosity as landmark as proposed by
Berger [3, 4]. In their conventional group, there were two
tibial components in[5° of internal rotation versus none in
their PSI group. They concluded that there is a smaller
chance of internal malrotation of the tibial component with
PSI, with less dispersion and amplitude of the tibial com-
ponent rotation around the neutral position. Intra- (0.934
and 0.988) and interobserver reliability (0.730 and 0.883)
was tested based on five patients without giving standard
deviations for those measurements.
The same group reported on a comparison of CT-guided
and MRI-guided PSI instrumentation in TKA and the effect
on component rotation. In their CT-guided group, there
were three of twenty-one tibial components in [5° of in-
ternal rotation versus none of 23 in their MRI-guided PSI
group. It was concluded that MRI may be more accurate
than CT using the Signature system (Biomet, Warsaw, IN,
USA) when planning the cutting jigs for TKA, with fewer
patients with malrotation of the tibial component. Intra-
(0.934 and 0.988) and interobserver reliability (0.730 and
0.883) was tested based on five patients without giving
standard deviations for those measurements and were the
same as in their earlier report [16].
Arch Orthop Trauma Surg (2015) 135:697–701
Differences between findings in the current publication
and the three quoted ones as mentioned above may be
explicable as different PSI systems, manufacturers, imag-
ing and design protocols as well as surgical techniques
were applied. Determining tibial component rotation in
TKA is also challenging as the anatomy of the proximal
tibia does not deliver constant prominent landmarks that
would be easy to define. This makes both CT or MRI
analysis as well as cutting jig production and intraoperative
placement a challenge. The sole use of the tibial tuberosity
as landmark with its poor reliability may be a limitation to
the otherwise excellent studies. Of note, tests for reliability
of measurements were not performed or only on a small
fragment of the included patients. Standard deviations be-
tween measurements were not reported at all, although
importants to give an idea of the mean error between
measurements and the effect on outlier analysis. Espe-
cially, the lack of accuracy of tibial component rotation
measurements based on the tibial tuberosity has to be
considered carefully when comparing data of different
authors. We believe it to be crucial to apply measurements
using different anatomic landmarks.
The results of PSI in reduction of rotational component
outliers seem to be promising. Especially, when consider-
ing that other techniques designed to increase accuracy of
component placement such as navigation failed to deliver
strong evidence to allow for better component rotation in
comparison with conventional techniques [17].
In this setup, PSI was effective in significantly reducing
outliers of optimal rotational tibial component alignment
during TKA. Future studies will be needed to show if this is
of clinical relevance.
References
1. Nicoll D, Rowley DI (2010) Internal rotational error of the tibial
component is a major cause of pain after total knee replacement.
J Bone Joint Surg Br 92(9):1238–1244
2. Bedard M, Vince KG, Redfern J, Collen SR (2011) Internal ro-
tation of the tibial component is frequent in stiff total knee
arthroplasty. Clin Orthop Relat Res 469(8):2346–2355
701
3. Berger RA, Rubash HE, Seel MJ, Thompson WH, Crossett LS
(1993) Determining the rotational alignment of the femoral
component in total knee arthroplasty using the epicondylar axis.
Clin Orthop Relat Res 286:40–47
4. Berger RA, Crossett LS, Jacobs JJ, Rubash HE (1998) Malrota-
tion causing patellofemoral complications after total knee
arthroplasty. Clin Orthop Relat Res 356:144–153
5. Murakami AM, Hash TW, Hepinstall MS, Lyman S, Nestor BJ,
Potter HG (2012) MRI evaluation of rotational alignment and
synovitis in patients with pain after total knee replacement.
J Bone Joint Surg Br 94(9):1209–1215
6. Sternheim A, Lochab J, Drexler M, Kuzyk P, Safir O, Gross A
et al (2012) The benefit of revision knee arthroplasty for com-
ponent malrotation after primary total knee replacement. Int
Orthop 36(12):2473–2478
7. Hofmann S, Romero J, Roth-Schiffl E, Albrecht T (2003) Rota-
tional malalignment of the components may cause chronic pain or
early failure in total knee arthroplasty. Orthopade 32(6):469–476
8. Barrack RL, Schrader T, Bertot AJ, Wolfe MW, Myers L (2001)
Component rotation and anterior knee pain after total knee
arthroplasty. Clin Orthop Relat Res 392:46–55
9. Heyse TJ, Tibesku CO (2014) Improved femoral component ro-
tation in TKA using patient-specific instrumentation. Knee
21(1):268–271
10. Parratte S, Blanc G, Boussemart T, Ollivier M, Le Corroller T,
Argenson JN (2013) Rotation in total knee arthroplasty: no dif-
ference between patient-specific and conventional instrumenta-
tion. Knee Surg Sports Traumatol Arthrosc 21(10):2213–2219
11. Heyse TJ, Figiel J, Hahnlein U, Schmitt J, Timmesfeld N, Fuchs-
Winkelmann S et al (2013) MRI after unicondylar knee arthro-
plasty: rotational alignment of components. Arch Orthop Trauma
Surg 133(11):1579–1586
12. Heyse TJ, Stiehl JB, Tibesku CO (2015) Measuring tibial com-
ponent rotation of TKA in MRI: what is reproducible? Knee
[Epub ahead of print]
13. Heyse TJ, le Chong R, Davis J, Boettner F, Haas SB, Potter HG
(2012) MRI analysis for rotation of total knee components. Knee
19(5):571–575
14. Bonnin MP, Saffarini M, Mercier PE, Laurent JR, Carrillon Y
(2011) Is the anterior tibial tuberosity a reliable rotational land-
mark for the tibial component in total knee arthroplasy?
J Arthroplast 26(2):260–267
15. Silva A, Sampaio R, Pinto E (2014) Patient-specific instrumen-
tation improves tibial component rotation in TKA. Knee Surg
Sports Traumatol Arthrosc 22(3):636–642
16. Silva A, Pinto E, Sampaio R (2014) Rotational alignment in
patient-specific instrumentation in TKA: MRI or CT? Knee Surg
Sports Traumatol Arthrosc [Epub ahead of print]
17. Meijer MF, Reininga IH, Boerboom AL, Bulstra SK, Stevens M
(2014) Does imageless computer-assisted TKA lead to improved
rotational alignment or fewer outliers? A systematic review. Clin
Orthop Relat Res 472(10):3124–3133
123