Unicompartmental Knee Arthroplasty
N.P. Kort, MD, PhD
N.P. Kort, J.J.A.M. van Raay, B.J.B. Thomassen
Use of an intramedullary rod is advised for the alignment of the femoral
component of an Oxford phase-III prosthesis. There are users moving towards
extra-medullary alignment, which is merely an indicator of frustration with
accuracy of intramedullary alignment. The results of our study with ten
cadaver femora demonstrate that the use of the short and long intramedullary
femoral rod may result in excessive flexion alignment error of the femoral
component. Understanding of the extramedullary alignment possibility and
experience with the visual alignment of the femoral drill guide is essential
toward minimizing potential errors in the alignment of the femoral
In carefully selected patients the results of the Oxford unicompartmental
arthroplasty have shown to be as good as those for total knee arthroplasty 1.
The principle of the Oxford unicompartmental knee 2 is that a
polyethylene bearing, concavely spherical above and flat below, can
maintain perfect congruity between the metal femoral condyle and the metal
tibia plateau while allowing them complete freedom to rotate and slide upon
each other. A spheric femoral component articulating with a congruous
meniscal bearing provides a large area of contact in all positions. Small
errors of alignment of the femoral component do not necessarily result in
loosening3 or loss of congruency. In the manual of surgical technique of the
Oxford unicompartmental knee, the allowed alignment variation is 10
degrees varus or valgus in the coronal plane and 5 degrees flexion or
extension in the saggital plane for the femoral component.
It is possible to implant this prosthesis using a minimally invasive
approach without everting the patella and thus avoiding damage to the
synovial reflections of the suprapatellar pouch 4. The femoral drill guide
should ensure proper placement of the femoral component. The guide is
visually aligned parallel to the long axis of the tibia and the fin on its side
parallel to the intramedullary femoral rod in the coronal and saggital planes.
The femoral drill guide is not fixed to the intramedullary rod.
Currently there is an intense debate going on about the Oxford
intramedullary rod both because of its usefulness and, as a secondary
consideration, its length 5. There is now a small but informed opinion group
moving towards extra-medullary alignment, which is merely an indicator of
frustration with accuracy of intramedullary alignment.
The accuracy of the intramedullary alignment of the femoral component
will be enlightened in this manuscript.
There are two different intramedullary rods on the market: the short thin rod
(200 mm by 4 mm) and the long thin rod (300 mm by 4 mm). The two thin
intramedullary rods (used with the Oxford phase III implant) are used in this
experiment. Digital X-rays were taken of the rods in ten cadaver femora, all
dry specimens of the left leg. Factors such as age, sexe and disease history
were not available for the dry specimens. Variables investigated were rod
length and possible positions of the rod compared to the mechanical axis on
the Anterior-posterior X-ray’s (coronal view) and to the anatomical axis on
the lateral X-ray’s (saggital view) in ten different left cadaver femurs. The
angles were measured in two separate sessions by four orthopaedic
surgeons. The manufacturer advises placing the intramedullary femoral rod
in an anteromedial entry point (1 cm anterior to the anteromedial corner of the intercondylar notch). All rods were positioned in the anteromedial entry
point and a varus/valgus stress was applied tot the distal tip in order to push
the proximal tip against the medial or lateral cortices respectively, while
taking the Anterior-posterior X-ray’s. For the lateral X-ray’s, proximal tips of
the rods were pushed against the ventral and dorsal cortices. All the X-rays
were taken in the same digital manner. A metal head is projected to facilitate
accurate measurements on the digital x-ray.
The different angles of the intramedullary rod compared to the mechanical
axis in the coronal view and the anatomical axis in the saggital view are
measured by each surgeon twice. With the advised entry point location of
the rod in the femoral intramedullary canal (anteromedial) insertion of the
rod in our cadaver study was easy.
Fig 1 shows the measurements for femur 2.
Average intra observer correlation is 0.954 with a Cronbach alpha of
0.976. The average inter observer correlation is 0.969 with a Cronbach
alpha of 0.992.
For the long thin rod there was only one position possible in the
intramedullary canal in all ten femora, an average of 6.5 degrees (sd 0.9)
in the coronal view compared to the mechanical axis and 4.4 degrees (sd 1.6)
in the saggital view compared to the anatomical axis. The long rod is fixed in
the femoral canal without possible movement in medial, lateral, dorsal or
Varus or valgus stress (Fig 2) on the short rod in the other femora causes an
average angle of 6.3 (sd 1.3) to 5.9 (sd 1.0) degrees in the coronal view. In the
saggital view, pushing the proximal tip of the short rod to the ventral or dorsal
cortex (Fig 3) gives respectively a average angle of 4.3 (sd 2.0) and 4.0 (sd 1.9)
compared to the anatomical axis. For the femora one, eight and nine there was
no intramedullary movement possible of the short rod in the four direction
while situated in the canal. With the first introduction of the short rod the
average angle in coronal view was 5.9 degrees (sd 1.4) and in the saggital view
4.0 degrees (sd 1.0). This is within the limits of the other seven femora.
In the saggital plane the long thin rod follow the femoral canal in an average
of 4.4 (sd 1.6) degrees of flexion. For the short rod the average angle is 4.3
degrees (sd 2.0) of flexion when the rod is pushed in flexion. The maximum
potential error of both the short and long intramedullary rod in the saggital
plane is more than 5 degrees in the saggital plane in femur two, four and nine.
An intramedullary rod is advised for the alignment of the femoral component
with the Oxford phase-III implant. With the minimally invasive technique,
the exposure is limited, so exact positioning of the prosthesis components is
more difficult and alignment error may result. Proper alignment does have its
effect on the wear rate of the mobile bearing. Bearings with signs of
impingement, due to the misalignment of the components have a maximum
wear rate of 0.08 mm per year. Those bearings showing no signs of
impingement have a mean wear rate of 0.01 mm per year 6.
With the Oxford phase-III unicompartmental knee arthroplasty, the femoral drill
guide should be visually aligned parallel to the long axis of the tibia, in the
middle of the condyle and parallel to the intramedullary femoral rod in the
coronal and saggital planes. The femoral drill guide is not fixed to the
intramedullary rod, inducing an uncertainty factor in the positioning of the
femoral drill guide, which results in an uncertain final position of the femoral
component. Intramedullary alignment is in our clinical experience not always
easy, the recommended anteromedial insertion site of the rod is not ideal for all
patients. Interference of the rod with the medial cortex during insertion may
prevent further insertion or cause valgus alignment of the rod compared with the
anatomical axis. Flexion of the knee, to make the upper surface of the femoral drill
guide lie parallel with the intramedullary rod, may be difficult because of
impingement of the patella with the IM rod in the clinical setting. The alternative is
the extramedullary alignment where the rod is fixated to the femoral drill guide. The
rod should be visually aligned to the femur in the clinical setting in the coronal and
saggital plane and the femoral drill guide should also be positioned in the middle of
the condyle and parallel with the long axis of the tibia.
With the extramedullary alignment the rod is fixed to the femoral drill guide (de drill
guide is already prepared for extramedullary alignment by themanufacturer); the rod
is now visually aligned parallel to the femur in the
coronal (Fig 4) and in the saggital plane the rod is pointed towards the head
of the femur (Fig 5). The position of the drill guide is completely dependent
on the expertise of the surgeon in the clinical setting. No special study is
possible with the cadaver femur for this extramedullary alignment as the rod
and the femoral drill guide are not fixated to the femur.
The limitation of this analysis is that we only used ten dry femur
specimens. The average canal width of our femora 161 mm from the distal
femoral surface is 20 mm (range 14-24 mm) in the coronal plane and 18 mm
(range 12-21 mm) in the saggital plane. The mean canal diameter measured
the half length of our femur, 13 mm (range 9-16 mm). This is more
comparable with the average measurements of Ma 5 with forty-five cadaver
femora and less than the average dimensions of Novotny 7 with twenty
cadaver femora. We can conclude that our cadaver femora were not
The average intraobserver correlation of 0.954 and interobserver
correlation of 0.969 show good intraobserver agreement and interobserver
agreement. The high Cornbach’s alpha measured is due to the high amount
of homogene items.
With the 7-degree fin on the femoral drill guide there will be an average
of less than 2 degrees of valgus alignment of the femoral component
compared to the mechanical axis in the coronal plane with the long rod and
the short rod. This is similar to the findings of Novotny et al. 7 less than Ma
et al 5, and does not exceed the allowed alignment error of 10 degrees varus
In this study, the maximum potential error of both the short and long
intramedullary rod in the saggital plane exceeds the allowed alignment error
of 5 degrees in the saggital plane in femur two, four and nine.
The results of our study demonstrated that the short and long
intramedullary femoral rod used for the phase III Oxford unicompartmental
knee arthroplasty may result in an excessive flexion alignment error of the
The extramedullary alignment for the Oxford Phase 3 Unicompartmental
knee arthroplasty is not mentioned in the current literature but now a small
but informed opinion group is moving toward extra-medullary alignment.
Understanding of both alignment possibilities and experience with the
visual alignment of the femoral drill guide are crucial toward minimizing
potential errors in alignment of the femoral component.
To rule out the uncertainties of the ‘on the eye’ positioning of the femoral
drill guide in the clinical setting, the current instrumentation should be
modified; will fixation of the femoral drill guide to the intramedullary rod be
possible or is computer navigation the solution?
Without these modifications the visual alignment is the most uncertain
factor. Experience of the surgeon will be the key issue. A clinical trial should
determine the place of the extramedullary alignment compared to the
intramedullary alignment advised by the manufacturer.
1. Murray DW, Goodfellow JW, O'Connor JJ. The Oxford medial unicompartmental
arthroplasty: a ten-year survival study. J Bone Joint Surg Br 1998;80:983-9.
2. Goodfellow JW, Tibrewal SB, Sherman KP, O'Connor JJ. Unicompartmental Oxford
Meniscal knee arthroplasty. J Arthroplasty 1987;2:1-9.
3. Shakespeare D, Ledger M, Kinzel V. Accuracy of implantation of components in the
Oxford knee using the minimally invasive approach. Knee 2005;12:405-9.
4. Price AJ, Webb J, Topf H, Dodd CA, Goodfellow JW, Murray DW. Rapid recovery
after oxford unicompartmental arthroplasty through a short incision. J Arthroplasty
5. Ma B, Long W, Rudan JF, Ellis RE. Three-dimensional analysis of alignment error in
using femoral intramedullary guides in unicompartmental knee arthroplasty. J
6. Psychoyios V, Crawford RW, O'Connor JJ, Murray DW. Wear of congruent meniscal
bearings in unicompartmental knee arthroplasty: a retrieval study of 16 specimens.
J Bone Joint Surg Br 1998;80:976-82.
7. Novotny J, Gonzalez MH, Amirouche FM, YC L. Geometric analysis of potential
error in using femoral intramedullary guides in total knee arthroplasty. J Arthroplasty
The different angles (degrees) of femur 2 compared to the mechanical axis for the coronal view
and the anatomical axis for the saggital view. There are 6 X-ray’s of femur 2. Each direction
shows the 2 measured angles per orthopedic surgeon.
The long rod in AP (coronal) view
The long rod in (saggital) view
By rotating the femoral drill guide the
extramedullary rod is pointed towards the
head of the femur when viewed from above
By adjusting the degree of flexion of the
knee, the extramedullary rod is made to lie
parallel with the femur when viewed from