Keywords total knee arthroplasty - valgus correction angle - distal femur - alignment - computer-assisted
TKAs
The survivorship of total knee arthroplasty (TKA) can be affected by factors such
as accuracy of bone cuts, limb and component alignment, and soft tissue balancing
achieved during surgery.[1 ]
[2 ]
[3 ] Component malalignment can lead to failure of TKA by causing early polyethylene
wear and aseptic loosening; therefore, if the surgeon is aiming for classical mechanical
alignment, he would like to align both femoral and tibial components within ± 3 degrees
of neutral with respect to their mechanical axes to facilitate equal distribution
of forces across the implant postoperatively.[4 ]
[5 ]
[6 ]
[7 ]
Since the femoral mechanical and anatomical axes are not coincident, a distal femoral
cut perpendicular to the coronal femoral mechanical axis is usually achieved by resecting
the distal femur using the distal femoral valgus correction angle (VCA), which is
equal to the angle between the femoral mechanical and anatomical axes.[8 ]
[9 ]
[10 ]
[11 ] This cut can be achieved by taking a fixed VCA of 5 to 7 degrees for all cases (which
is the standard practice by many surgeons), tailoring the distal cut according to
the preoperatively measured VCA, or by using computer-assisted navigation.[12 ]
[13 ]
Several studies have highlighted wide variations in VCA in patients undergoing TKA
where it can range from 2 to 13 degrees and how using a fixed VCA range can lead to
error in distal femoral cut and femoral component malalignment; hence, it is recommended
that the distal femoral cut should be individualized in each patient according to
the measured VCA.[13 ] However, in a meta-analysis of 29 studies, it was found that the femoral component
position is accurately positioned only in 65.9% of knees within 2 degrees of perpendicular
to the mechanical axis in conventional TKAs.[14 ]
We devised a novel three-step technique that involved radiographically measuring the
VCA preoperatively for each knee, using an image intensifier before incision to identify
the femoral head center and using an extramedullary verification tool intraoperatively
to determine the accuracy of the distal femoral resection. This study hypothesized
that the accuracy of distal femoral cut and femoral component placement in the coronal
plane with this three-step technique, incorporated in the conventional technique of
TKA, would be as accurate as computer navigation. To the best of our knowledge, this
is the first study in the literature to compare our novel three-step technique with
the navigation technique.
Methods
This is a retrospective radiographic review, performed between February 2019 and October
2019, of the records of 500 consecutive primary TKAs operated by a single surgeon.
The inclusion criteria were all patients who underwent primary TKA for osteoarthritis
and rheumatoid arthritis and all patients who provided their informed consent for
participation in our research study. We excluded patients who had pre- or postoperative
radiographs of inappropriate quality or patients whose radiographs were unavailable
for evaluation, and patients who were lost to follow-up. The study received institutional
ethics committee approval.
All surgeries were performed at either of two centers: the optimized conventional
technique of TKA in one center and the navigation technique of TKA in the other center.
Hence, patients received either an optimized conventional or navigated TKA depending
on which center was selected by the patient (chosen by patients based on the proximity
of their geographic place of residence to the surgery center). The study was approved
by the ethics committee of the institution. Of the 500 TKAs enrolled during the study
period, 42 TKAs were excluded due to the unavailability of radiographs or inappropriate
quality of pre- or postoperative radiographs. A total of 458 TKAs were available for
radiographic analysis, with 178 TKAs in the optimized conventional group and 280 TKAs
in the computer navigation group ([Table 1 ]). There were 435 varus knees and 23 valgus knees in the study.
Table 1
Comparison of independent variables between the two groups
Independent variables
Navigated group
(n = 280)
Novel conventional group
(n = 178)
Total
(n = 458)
p -Value and significance
Gender
Male
67 (23.9%)
41 (23.0%)
108 (23.6%)
Chi-square test = 0.048
p = 0.826, NS
Female
213 (76.1%)
137 (77.0%)
350 (76.4%)
Side
Left
127 (45.4%)
91 (51.1%)
218 (47.6%)
Chi-square test = 1.451
p = 0.228, NS
Right
153 (54.6%)
87 (48.9%)
240 (52.4%)
Preoperative alignment
(HKA angle in deg)
167.5 ± 7.8 (138.4–206.5)
166.8 ± 11.1 (126.5–213.0)
167.2 ± 9.2 (126.5–213.0)
t = − 0.827
p = 0.409, NS
Type of deformity
Varus
270
165
435
Chi-square test = 3.177
p = 0.075, NS
Valgus
10
13
23
Abbreviations: HKA, hip–knee–ankle; NS, not significant.
Notes: All values are presented as mean ± standard deviation (minimum–maximum) or
number (percentage). p < 0.05 is considered significant.
Full-length, weight-bearing, hip-to-ankle radiographs were obtained before and within
4 weeks after surgery. The postoperative radiographs were all obtained in the same
institution using a standard radiographic protocol (with both knees in full extension,
patellae facing forward, and both feet pointing forward to avoid malrotation of the
limb during radiography). All full-length radiographs were screened by one of the
authors for excessive rotation in the coronal plane which made the radiograph unsuitable
for analysis.[15 ] The excessive rotation was determined on radiographs by the profile of the lesser
trochanter and fibular head and whether the patella was centered or medial/lateral.
Digital images of the radiographs were used for measuring various radiographic parameters
using the Image J image processing and analysis software (version 1.41, U.S. National
Institute of Health). Radiographic parameters measured included pre- and postoperative
coronal limb alignment or hip–knee–ankle (HKA) angle defined as the angle between
the femoral mechanical axis (line joining the center of the femoral head and center
of the knee joint) and the tibial mechanical axis (line joining the center of the
knee joint and center of the ankle joint), and the postoperative femoral component
coronal alignment defined as the medial angle between the femoral mechanical axis
and the line tangential to the prosthetic femoral condyles. All measurements on radiographs
were performed by an independent surgeon who was not associated with any of the surgeries
and was blinded to the type of technique used.
All procedures were performed under a tourniquet which was deflated when the cement
hardened. All knees were approached with a standard medial parapatellar arthrotomy.
All patients received a cemented, posterior-stabilized, fixed-bearing implant (PFC
Sigma, DePuy Inc, Warsaw, IN) with resurfacing of the patella. The same cutting blocks
were used for performing the bony resections in both groups. The aim was to achieve
a limb coronal mechanical axis of 180 ± 3 degrees and position the femoral component
perpendicular to the femoral mechanical axis ± 3 degrees. The degree of soft tissue
release was based on the amount of soft tissue tightness assessed using a manual tensioning
device and spacer block. Medial release in varus knees and lateral release in valgus
knees were performed to achieve rectangular balanced gaps and a fully restored mechanical
axis. We used the balanced gap technique, the tibial cut being performed first, and
the extension gap being balanced before the flexion gap.
In the patients who had TKAs performed with computer navigation (Brain Laboratory,
Munich, Germany), registration was performed in a standard fashion after fixation
of infrared reflecting arrays on the distal femur and proximal tibia. The coronal
mechanical axis of the lower limb was derived by the navigation software using the
center of the femoral head, the center of the intercondylar notch, the center of the
tibial plateau, and the center of the ankle plafond. A verification tool with reflecting
arrays was used to position the cutting blocks and verify and quantify the distal
femoral cut.
In the conventional TKA group, we used a novel three-step method comprising (1) measuring
the VCA for each knee on preoperative full-length, hip-to-ankle radiographs. This
angle was used to individualize the distal femoral resection in the coronal plane
in each patient. Hence, on a full-length radiograph, if the VCA was measured as 8 degrees,
the distal femoral resection guide was set at 8 degrees valgus so as to achieve a
distal femoral cut perpendicular to the femoral mechanical axis. (2) After the patient
was anesthetized, an image intensifier was used to identify the femoral head center,
and a radio-opaque adhesive marker was placed in the groin as a surface marker for
the femoral head center. (3) During surgery, an extramedullary verification tool,
consisting of a long rod which was fixed perpendicular to a flat plate, was used to
determine the accuracy of the distal femoral resection ([Fig. 1 ]) (after the individualized cut had been performed using the distal resection block
placed with an intramedullary rod) with reference to the previously placed femoral
head center radio-opaque marker. Any disparity was corrected by a freehand recut.
Fig. 1 The extramedullary verification tool comprises a base plate (red arrowhead) that
is placed on the resected surface of the distal femoral. The plate is fixed to a vertical
arm (yellow arrowhead) with a sleeve, into which is slid a telescoping rod. The relationship
of the proximal end of the rod to the marker placed in the groin (not shown, presented
by the black dot) is palpated.
Statistical analysis was performed using Microsoft Excel v16.37 and StatPlus2 v 7.1.
Data from the two groups of patients were compared using Student's t -test, Mann–Whitney's U -test, chi-square test, Fisher's exact test, and analysis of variance with Tukey's
post hoc test. A significant difference was taken to be p < 0.05.
Results
[Table 1 ] shows that there is no significant difference for all the four independent variables
(gender, side, preoperative alignment value [HKA], and type of deformity [varus or
valgus]) between navigated and novel conventional groups, implying that all the independent
variables were equally distributed between the two groups.
The mean femoral component coronal alignment was 89.3 degrees ± 2.2 (80.7–98.8 degrees)
in the navigated group and 89.5 degrees ± 2.0 (82.7–94.3 degrees) in the novel conventional
group ([Fig. 2 ]). After carrying out an independent t -test between the two groups, there was no significant difference (p = 0.314).
Fig. 2 Box and whisker plot showing femoral component alignment (degrees) in the novel conventional
technique and with navigation.
On excluding the valgus knees from both groups, the mean femoral component coronal
alignment was 89.2 degrees ± 2.2 (80.7–98.8 degrees) in the navigated group and 89.4 degrees ± 1.9
(82.7–93.8 degrees) in the conventional group. After carrying out an independent t -test between the two groups, there was no significant difference (p = 0.294).
In varus knees, there were 40 outliers outside the ± 3 degrees range in the navigated
group (31 <87 degrees, 9 >93 degrees) and 21 outliers in the conventional group (17
<87 degrees, 4 >93 degrees). There were two outliers among the valgus knees in the
conventional group (one <87 degrees and one >93 degrees), and none in the navigated
group ([Fig. 3 ]). Totally, there were 12.9 and 14.3% outliers with the novel conventional technique
and the navigation, respectively. The number of outliers for femoral component coronal
alignment outside the ± 3 degrees range from a neutral alignment of 90 degrees in
the coronal plane was not significantly different between the two groups (navigation
and novel conventional techniques) when assessed separately both for varus (chi-square
statistic 0.3702; p -value 0.543) and valgus deformities (Fisher's exact test statistic 0.1779, the result
being not significant at p < 0.05).
Fig. 3 Box and whisker plot showing femoral component alignment (degrees) separately for
knees with preoperative varus and valgus alignment in the novel conventional technique
and with navigation.
The mean value of femoral component alignment using the conventional technique in
knees with varus deformity <10 degrees was 88.8 degrees, in knees with varus deformity
10 to 20 degrees was 89.4 degrees, and in those with varus deformity >20 degrees was
90.2 degrees ([Fig. 4 ]). When varus knees in the conventional technique were compared based on their preoperative
HKA angles, there was a significant difference on analysis of variance in the mean
femoral component coronal alignment between knees with varus deformity <10 degrees,
and varus deformity 10 to 20 degrees, and varus deformity >20 degrees (F statistic
4.824, p = 0.009). Tukey's post hoc test identified the group of varus knees <10 degrees being
significantly different from those >20 degrees (Q statistic 4.381, p = 0.006). There was no significant difference between varus knees <10 degrees and
those with varus deformity 10 to 20 degrees (p = 0.251). There was no significant difference between 10 and 20 degrees varus knees
and those with varus knees >20 degrees (p = 0.116).
Fig. 4 Box and whisker plot showing femoral component alignment (degrees) for knees with
preoperative varus alignment grouped into knees with preoperative varus <10, 10 to
20, and >20 degrees in the novel conventional technique.
Using the Mann–Whitney's U -test, there was no significant difference in femoral component alignment between
valgus knees in the navigated and novel conventional groups. (U -value = 43, result not significant at p < 0.05).
Discussion
The main finding of this study was that the mean femoral component coronal alignment
was not significantly different when navigation and optimized conventional groups
were compared and the percentage of outliers for femoral component coronal alignment
outside the ± 3 degrees range from a neutral alignment of 90 degrees in the coronal
plane was not significantly different between the two groups. We found our novel three-step
technique of comparable accuracy as computer navigation in improving the accuracy
of the distal femoral resection. The technique involved using (1) individualized distal
femoral valgus cut according to the measured VCA, (2) markers placed under image intensification
to identify the surface position of the hip joint center, and (3) an extramedullary
verifying tool. The method of measuring VCA on radiographs is important. Key points
to note are as follows: The radiographs must include the entire femur. The limb should
be correctly rotated so that the patella is anterior and in the midline. The profile
of the lesser trochanter should be just visible. The fibular head should not be obscured
by the tibia due to external rotation of the limb.
Also, the distal femoral axis must be a line drawn from the apex of the intercondylar
notch to the femur's midpoint at the level (usually middle third-lower third junction
in a patient with excess femoral bowing) where the intramedullary rod is likely to
engage the canal. In more bowed femora, it may be prudent to use a short rod. The
VCA should not be measured using the classical anatomical femoral axis drawn by joining
the femoral canal's midpoints at the upper one-third and lower one-third of the femur.
Mason et al,[14 ] in a meta-analysis of 29 studies comparing computer-assisted TKAs to conventional
TKAs, reported that 90.4% of computer-assisted TKAs had femoral component alignment
within 2 degrees of perpendicular to the mechanical axis versus 65.9% in conventional
TKAs. However, most of these studies comparing conventional and navigated TKAs used
a fixed VCA range of 5 to 7 degrees and have not individualized VCA during TKA which
may be the reason why the accuracy of the conventional technique in distal femoral
resection was inferior to the navigation technique. Palanisami et al[16 ] in a prospective study reported significant improvement in both femoral component
placement and postoperative alignment when VCA was individualized when compared with
taking a fixed VCA of 5 degrees in patients with moderate and severe varus deformities.
Similarly, Shi et al,[17 ] in a prospective, randomized study, reported that 83.6% of knees with femoral component
alignment within ± 3 degrees of the femoral mechanical axis in the individualized
VCA group compared with 39.4% in knees with fixed VCA group.
The VCA shows wide variation among knees, especially with associated extra-articular
deformities (such as excessive coronal bowing of the femur) or significant varus deformities.[8 ]
[12 ] Using a fixed VCA range of 5 to 7 degrees in such patients can cause significant
error in distal femoral resection and final femoral implant coronal alignment. Yau
et al[12 ] in a study involving Chinese patients reported that at least in 31, 31, or 34% of
the limbs with femoral bowing, a planning error of more than 2 degrees could result
if a routine VCA of 5, 6, or 7 degrees was chosen, respectively. Similarly, Mullaji
et al[8 ] reported that by choosing a routine VCA of 5, 6, or 7 degrees, the planning error
in VCA would be more than 2 degrees in 45.1, 28.2, or 21.1% of limbs, respectively.
An intramedullary rod may be misleading in cases with significant bowing of the femoral
shaft or an extra-articular deformity such as a malunited fracture in the distal femur
which may cause malalignment of the intramedullary guide rod and distal femoral cutting
block. Determining the center of the femur head beforehand using an image intensifier
in conjunction with using a patient-specific VCA further improves the accuracy of
the distal femoral resection, and the extramedullary verification rod affords an opportunity
to confirm and fine-tune the precision of the cut.
The mean value of femoral component alignment using the novel conventional technique
in knees with varus deformity <10 degrees was 88.8 degrees, in knees with varus deformity
10 to 20 degrees° was 89.4 degrees, and in those with varus deformity >20 degrees
was 90.2 degrees. Though analysis of variance showed these three groups to be significantly
different, Tukey's post hoc test identified only the group of varus knees <10 degrees
being significantly different from those >20 degrees. However, though statistically
significant, the difference of a little over 1 degree is clearly not of clinical significance.
One drawback of the optimized conventional technique is a modest increase by a few
minutes in the overall surgical time due to the use of an image intensifier to identify
the femoral head center and the ankle joint. This can be minimized by positioning
the image intensifier while the patient is being anesthetized. Computer navigation
can bypass any femoral extra-articular deformity including excess lateral bowing and
directly align the distal femur cutting block perpendicular to the coronal mechanical
axis of the femur minimizing the chances of a cutting error. Furthermore, most conventional
distal femur cutting guides have a maximum VCA setting of 9 degrees beyond which the
surgeon may have to lateralize the entry point which increases the chances of cutting
error. In these cases, using computer navigation or the described optimized conventional
technique can help improve the accuracy of femoral component placement. However, there
is scope for improvement as we still had 12.9% outliers with femoral component alignment
outside the ± 3 degrees range; that this series included knees with varus even greater
than 20 degrees may be borne in mind.
The strengths of this study are the large number of cases, all measured by a standardized
technique, by a single independent senior orthopaedic surgeon, and that all surgeries
were performed by a single surgeon, using a uniform technique in both groups of patients.
There are some limitations to our study. First, this study has only considered the
femoral component placement in the coronal plane. We have not measured femoral component
placement in the sagittal plane as full-length lateral sonograms are not performed
in our radiology department. Second, since the objective of this study was to compare
the accuracy of the femoral component in the coronal plane on full-length radiographs,
we have not considered clinical follow-up data regarding functional scores or revision
rates. Very often, a compensatory correction may occur deliberately or providentially
in the proximal tibial cut. This may help in improving the HKA angle. Occasionally,
it may compound the error. Hence, we have not presented the HKA and tibial component
values, nor the patient-reported outcome measures. Future follow-up studies regarding
the clinical outcome and implant survival in the outliers between the two groups would
be valuable. However, the authors believe that the novel three-step modification of
the optimized conventional technique described in this study will help surgeons using
the conventional technique during TKA to achieve accurate placement of the femoral
component in knees comparable to navigation. Third, this study involved measuring
various parameters on radiographs which is prone to measurement errors. However, limb
and component alignment measurements on full-length hip-to-ankle radiographs have
been reported to be reliable and reproducible with good intra- and interobserver correlations.[18 ] Fourth, the number of valgus knees was fewer, as valgus deformity is relatively
uncommon in the study region. Fifth, we could not include a third group, using a fixed
VCA, as several studies have already stated the variability associated with the method.
Finally, the study is retrospective.
Conclusion
In conclusion, using the novel three-step technique during conventional TKA to perform
the distal femoral cut can increase the accuracy of femoral component coronal alignment
and achieve positioning comparable to that with computer navigation.
