LIVRE JDG 2010

DISCUSSION

First of all we would like to mention the grea-

test drawback to our study: Inaccuracy due to

magnification effects taking radiographs and

inaccuracy of defining the points of measure-

ments and measurement itself. We think we

could avoid the magnification effect by com-

paring the fibula length pre and postoperative-

ly (Table 2). We did not see big differences

here. As we believe this is the result of stan-

dardised methods and an experienced staff

taking the radiographs.

Calculating the average leg out of our measu-

rements gives us a triangle with the side a

being 36.3 cm and side b being 44.9 cm long.

In a model we can now calculate side c_pre.

Choosing the angle 195° or 165° we can cal-

culate side c_pre being 80.5 cm. Now we can

calculate c_predictive (

now being 180°) the

result is a leg length of 81.2 cm. In this model

calculation of an average leg we will achieve a

leg lengthening of 0.7 cm by correcting the

angel. This gives us an idea of the estimated

lengthening after angle correction of 15°. This

value lies also in the range of the mean leng-

thening after TKA of 0.54 cm calculated out of

our data (Table 1). (Mean(total)c_post81.28cm

– mean(total)c_pre80.7cm=0.54cm.)

By calculating all the differences between the

predicted and postoperative leg lengths we

could define a mean error of 0.55 cm between

predicted leg lengths and measured leg lengths

after TKA. This lies below the mean error of

0.7 cm calculated by Kessler 2 for length

changes in tibial osteotomy with angular cor-

rection. This higher accuracy may be the result

of calculating c_predictive with angle

post-

operative rather than just adding a_pre and

b_pre together. This is of course a result of our

retrospective study in which we used

postop

to calculate the predictive leg length. Applying

our method preoperatively one would have to

choose

180° to calculate the predictive leg

length and so one would receive more often

higher values for the difference between c_pre

and c_predictive. Our study shows, that the

bigger the deformity is the greater leg lengthe-

ning will be predicted. This is also represented

in the mean difference between c_pre and

c_post. In the valgus group this value is 1.2

cm, in the varus group 1.0 cm whereas this

value is only 0.3 cm in the normal group where

nearly no angular correction occurred.

We took the value of up to 0.5 cm difference

between (value just beneath mean error of

0.55 cm) c_post and c_predictiv to say that leg

length was predicted correctly. Values diver-

ging over 0.5 cm we analysed more exactly.

Here we could see that each value diverging

more than 0.5 cm could be satisfyingly explai-

ned by lengthening or shortening of femur

and/or tibia.

Using our method preoperatively will not pre-

vent leg lengthening because correction of the

deformity must stay a central point in TKA7-

11 to achieve good results. But this study

shows the tendency of lengthening above the

predicted length due to angular correction in

the varus and valgus groups, and there again

mainly on the tibial side (Table 4). We think it

could be helpful during operation to know the

estimated leg length. Maybe this method can

be combined with computer assisted surgery

(CAS). In this way the positioning of the

implant, the alignment [12, 13] and the resul-

ting leg length can be simulated. Starting from

the healthy condyle or tibial plateau side as

reference for our joint line we then may

achieve a greater accuracy in positioning our

implant in respect of the joint line [11].

CONCLUSION

We belief, that knowing the predicted leg

length can help to position the implant with

more accuracy. Of course in case of flexum

our method is not applicable but in these cases

a correction factor to calculate the true length

[1] may be helpful. Or the combination with

CAS could calculate the leg length and the

predicted leg length in real time on the opera-

ting table. By simulation of implant positio-

ning in combination of alignment we may so

increase accuracy and reproducibility in TKA.

JOURNÉES LYONNAISES DE CHIRURGIE DU GENOU