Question: In patients having knee arthroplasty, is computer-assisted
knee arthroplasty (CAKA) more effective than a conventional jig-based
technique?
Design: Randomized (allocation concealed), blinded (outcome assessor
and statistician), controlled trial with 6-week follow-up.
Setting: A hospital in Perth, Western Australia, Australia.
Patients: 71 patients who were scheduled for total knee
arthroplasty, were medically fit for surgery, and had no active infection or
malignant disease. Follow-up was 100%.
Intervention: Patients were allocated to CAKA (n = 34) or
conventional jig-based knee arthroplasty (JBKA) (n = 36). CAKA was done using
the Stryker Knee Navigation system (Stryker, Kalamazoo, Michigan), which makes
use of anatomical mapping of the knee and kinematic analysis of the limb to
make a working model of the knee, and an infrared camera to track and relate
the position of a pointer or attachment to the mapped anatomy shown in real
time on a computer screen. JBKA was done with use of intramedullary
instrumentation for the femur and extramedullary instrumentation for the
tibia.
Main outcome measures: Blood loss and complications after surgery,
alignment of femoral and tibial components at the time of hospital discharge,
and long-leg radiographs at 6 weeks.
Main results: Mean blood loss was less in the CAKA group than in the
JBKA group (252 vs 446 mL, p = 0.001). CAKA and JBKA groups did not differ
with regard to the following complications, from which all patients recovered:
deep venous thrombosis (1 vs 2), superficial infection (1 vs 2), stiff knee (1
vs 0), acute postoperative confusional state (1 vs 10), and transient ischemia
(0 vs 1). The CAKA group showed greater improvement than did the JBKA group
with regard to femoral varus/valgus alignment (range, 6° vs 8°; p =
0.032)*, femoral
rotation (range, 4° vs 5°; p =
0.001)*, tibial
varus/valgus alignment (range, 5° vs 6°; p =
0.047)*, tibial
posterior slope (range, 7° vs 9°; p =
0.001)*, tibial
rotation (range, 14° vs 18°; p =
0.011)*, and
femorotibial mismatch (range, 11° vs 21°; p =
0.037)*. The groups
did not differ with regard to flexion of the femoral component (p = 0.425). At
6 weeks, the CAKA group showed greater improvement in standing alignment (p =
0.004). The CAKA procedure took longer to perform than did the JBKA procedure
(mean operative time, 80 vs 67 min, p = 0.001).
Information provided by author.
Conclusions: In patients having knee arthroplasty, the CAKA
procedure was associated with better alignment of the femoral and tibial
components and was associated with less blood loss than the JBKA
technique.
This article suggests that, at least in selected surgeons' hands, a
computer-assisted technique for total knee arthroplasty can improve the
accuracy of bone cuts and reduce the number of outliers with less than perfect
limb alignment or component orientation. The methodology of the study, which
includes computed tomographybased analysis of final limb alignment, provides
the reader with confidence that the information presented regarding alignment
is accurate.
Although some may argue that small average improvements in limb alignment
are of questionable clinical relevance, certainly the reduction of outliers
with poor limb alignment has a high likelihood of providing functional benefit
and improved implant durability for those patients.
The authors found that there was less blood loss with the computer-assisted
method, which they attributed in part to avoidance of the use of medullary
canal instrumentation. This finding may be dependent on the specifics of
surgical technique (tourniquet or no tourniquet, implants or techniques that
occlude or do not occlude the instrumented intramedullary canal) and may not
be generalizable.
One of the potential drawbacks of computer-assisted surgery is added
operative time and expense. In this study, the average added time of 13
minutes per procedure was modest. Of note, the authors had learned the method
on cadavers and had performed 12 such surgeries prior to this study; hence, a
time-intensive learning curve was not included in this study.
The degree to which the many different types of computer-assisted
instrumentation will provide benefits or create problems is, as yet, unknown.
As techniques evolve and become more simple and reproducible, it seems likely
that some form of computer assistance ultimately will become an everyday part
of many orthopaedic procedures.
Each surgeon will need to decide at what point in the evolution of this
emerging technology that the potential benefits of computer-assisted
methods—well documented in this important study—justify the costs
and potential risks in an individual practice.