28 / 242
D. Saragaglia, A. Krayan
28
of this sort could only complicate the operative
procedure and lastly, this would have added
additional cost and a considerable amount of
radiation exposure for the patient. We needed
to have a reference to the mechanical leg axis
throughout the whole operation so that the
cutting guides could be placed perpendicular to
this axis in a frontal and a sagittal plane. The
cutting guides needed to be placed freehandedly
without any centro-medullary or extra-
medullary rods. Finally, the operation was not
supposed to last more than 2 hours (maximum
tourniquet time) and the procedure was to be
accessible to all surgeons, whatever their
computing skills.
The project was assigned to F. Picard, as part of
his Postgraduate Diploma in Medical and
Biological Engineering, and to F. Leitner a
computer scientist who was completing his
training. After 2 years of research, the system
was validated by the implantation of 10 knee
prostheses on 10 cadaver knees, and the results
were published in 1997 [12, 13] in several
national and international publications,
including CAOS, SOFCOT and SOBCOT.
After obtaining consent of the local ethics
committee on December 4, 1996, the first
computer-assisted prosthesis was implanted in
a patient on January 21, 1997 (D. Saragaglia, F.
Picard, T. Lebredonchel). The operation lasted
2 hours and 15 minutes and was uneventful.
A prospective randomized study comparing
this technique to the conventional technique
began in January 1998 and was completed in
March 1999. The results were published in
several national and international meetings and
in a lead article in the French Journal of
Orthopaedic Surgery [1]. In March 1999, the
prototype that we had used in this study evolved
to a final model called Orthopilot™ [B-Braun-
Aesculap, Tuttlingen, Germany]. The software
packages have evolved over the years but the
basic principle has remained the same since the
system was created. To day, in our hands,
computer-assisted TKA takes around one hour
(from 50 to 75 minutes) and is routinely used.
What have I learned
with navigation?
Navigation is probably the best tool to teach
the implantation of TKA. During the
operation, the fellow can follow exactly the
operative procedure and familiarize with the
HKA angle, the femoral mechanical angle
(FMA), the reducibility of the deformity, the
tibial slope, the need to do a release or not or
to give rotation or not to the femoral implant.
All these data are analysed in live and
discussed during surgery. Moreover, for a
young surgeon with little experience, the
learning curve is much less long than for
conventional technique [14].
In my own practice, I learned that I could reach
the goal I wanted to reach preoperatively in an
easier way. In other words, whatever the HKA
angle, included severe varus or valgus
deformity, navigation allows the final HKA
angle to be within 3° residual valgus or varus
deformity in more than 95% of the cases.
Regarding the tibial cut, navigation is not very
different from conventional technique except
that it is more precise than extra, intra or
combined extra and intra medullary guides. For
the distal femoral cut, we have learned that the
medial FMA is not always in valgus and it can
range from 10° of varus in some severe genu
varum deformities, to 10° of valgus or more in
some severe valgus deformities. In these severe
conditions, above all for genu varum, the intra
medullary guides do not allow to give sufficient
valgus in order to put the femoral implant
perpendicular to the mechanical axis of the
lower limb (fig. 1).
I have learned also that, in case of severe
deformity, the varus or the valgus can be
reducible or over reducible avoiding to perform
extensive releases [15]. It is the reason why,
currently, in my hands, extensive release is
around 5% of the cases and I do pie crusting of
the medial collateral ligament (genu varum) or
fascia lata (genu valgum) also in around 10%
of the cases.