Computer-Assisted High Tibial and Double-Level Osteotomies for Genu Varum Deformity

We have utilized computer navigation to perform osteotomies around the knee. The goal of this article is to present our rationale, indications, and surgical technique of computer-assisted high tibial and double-level osteotomies for genu varum deformity. The results are based on 2 studies: (1) a comparative cohort study of computer-assisted versus conventional high tibial osteotomy, showing a 96% reproducibility in achieving a mechanical axis of 184 ± 2 degrees in the computer-navigated group versus 71% in the conventional osteotomy group (P < 0.0015); (2) a prospective study on double-level osteotomy showing that the preoperative goal (182 ± 2 degrees) has been achieved in 91% of the cases.

O steotomy around the knee has been used in the treatment of varus and valgus malalignment for more than 50 years. 1Y3 Currently, the most commonly used osteotomies are high tibial osteotomy (HTO), a valgus tibial osteotomy for genu varum, and low femoral osteotomy, a varus femoral osteotomy for genu valgum. The HTO gives the best results when the procedure is well indicated, and the planned correction is reproduced. 4Y10 Normally, one plans an overcorrection into valgus of 3 to 6 degrees, 5,7,9,11 giving a mechanical axis (center femoral head through center of tibial spines to center of the ankle) of 183 to 186 degrees. Making the initial osteotomy and subsequently fixing it are not demanding. However, producing the required correction continues to be the complicated step regardless of vigorous attempts to plan the procedure preoperatively. 11Y13 The methods of controlling this step intraoperatively are nonexistent for some surgeons and basic or archaic for others. Because of the uncertainty of the result and the significant development of total knee arthroplasty (TKA), some surgeons have abandoned the osteotomy to avoid the problems of undercorrection and the cosmetic, functional, and medicolegal consequences of overcorrection.
In addition, the HTO exposes the patient to the serious consequence of altered joint-line obliquity ( Fig. 1). This complication has only been mentioned a few times in the current literature 14 despite the concerns that it poses during TKA. This alteration of joint-line obliquity is all the more commonly seen and as important as the varus, whether the varus originates from the femur (congenitally curved femur or distal femoral varus) or from tibia and femur. The hypercorrection required for a good result (3Y6 degrees) often makes the joint-line obliquity even worse and when beyond 10 degrees may require another osteotomy during the TKA to correct the malunion and avoid a severe ligamentous imbalance. 15 We have always been interested in double-level osteotomies (DLOs) to avoid this problem, but the difficulties in this procedure and in the correction of the leg axis have been a brake on the expansion of this technique.
Drawing on our experience with computer-guided TKA 16Y18 and the precision attainable, it would appear that osteotomy around the knee is an excellent application of this technology. We used the OrthoPilot (B.

CONTRAINDICATIONS FOR HTO AND DLO
The best indication for osteotomy is a ''young'' (G60Y65 years) and nonsedentary patient with a low arthrosis grade (stage 1, 2, or 3 of 5) according to Ahlbäck's modified criteria. 19 High tibial osteotomy is indicated in case of varus deformity of the proximal end of the tibia. It is contraindicated when the tibial mechanical axis is at 90 degrees or in valgus (very rare). The overcorrection needed to have a good clinical result will lead to oblique joint line.
Double-level osteotomy is indicated in case of severe genu varum (910 degrees) when femoral mechanical axis (FMA) is less than 90 degrees and when there is a proximal varus deformity of the tibia. It is contraindicated when FMA is in valgus. In our experience (unpublished data of a prospective study of 89 cases), it is the case in 52.8% of genu varum deformities (91Y93 degrees in 43.8% of the cases and 993 degrees in 9%).

| PREOPERATIVE PLANNING
The OrthoPilot is a kinematic nonYimage-based model, and it is not necessary to have a preoperative computed tomography scan. We need only standard radiograph (anteroposterior standing and Rosenberg view, lateral and femoropatellar view) and, most importantly, longleg radiograph. It is very important to measure the femorotibial mechanical axis, FMA, and tibial mechanical axis and to plan the overcorrection as a conventional technique.

Computer-Navigated Opening Wedge HTO
The software is a derivative of the one used for TKA, which has been fully described elsewhere. 16Y18,20,21 The same principle of real-time acquisition of the rotation center of the hip, knee, and ankle centers and of the anatomical landmarks at the level of the knee joint line and ankle is applied. It allows the mechanical axis of the lower limb to be shown dynamically on the computer screen (ie, the axis of the lower limb to be seen both preosteotomy and postosteotomy and to check if the preplanned correction has been established).
A sterile tourniquet is placed at the root of the thigh, and the procedure follows in this sequence: the rigid body markers are fixed percutaneously at the level of the distal femur and proximal tibia, allowing acquisition of the centers of the hip, knee, and ankle (Fig. 2); the lower limb mechanical axis then appears on the screen and can be compared with the preoperative radiological goniometry.
A 5-to 6-cm-long incision is made on the medial upper end of the tibia just at the level of the anterior tuberosity of the tibia. The pes anserinus is incised just above the gracilis tendon, and a retractor is placed against the posteromedial corner of the tibia. Then, the superficial medial collateral ligament is released from its tibial insertion to allow an adequate opening of the osteotomy. The HTO is performed 3 cm below the level   of the medial joint line, the level confirmed by placing an intra-articular needle. The osteotomy is directed at the fibula head, keeping the saw as horizontal as possible to avoid fracturing the lateral tibial plateau.
With the aid of 2 Pauwels osteotomes inserted along the track of the saw cut, the tibia is placed into valgus. These are then replaced by a metal spacer that is inherently stable and allows the amount of correction to be calmly checked (Fig. 3). If there was 8 degrees of varus, one would try a 10-to 11-mm spacer and make sure that an appropriate hypercorrection is produced real time on the computer screen. If this is insufficient, we try a thicker spacer, and the reverse if the correction is too great. The metallic spacer is then replaced with a bioabsorbable tricalcium phosphate wedge (BioSorb; B. Braun Aesculap, Boulogne, France) of the desired thickness, and the intervention is completed by plating the proximal tibia (Fig. 4). Then, the accuracy of the osteosynthesis is checked with an image intensifier (Fig. 5), and the wound is closed.

Computer-Assisted DLO
A sterile tourniquet is placed at the root of the thigh, and the first stage is essentially the same as for that of an HTO: percutaneous insertion of the rigid body markers (high enough not to hamper the femoral osteotomy and low enough on the other level to avoid interfering with the tibial osteotomy), followed by kinematic acquisition of the hip center, middle of the knee, and tibiotarsal joints to find the mechanical axis of the lower limb.
The second stage consists of making the femoral closing osteotomy in the distal femur (in general, a 5-to 6-degree alteration is made, although sometimes more in congenital femoral varus to give 3 or 4 degrees of    valgus in the femur) 22,23 and fixing it in position with a T-plate (AO/Synthes, Etupes, France). A closing wedge osteotomy is indicated to avoid too much lengthening of the lower leg. A lateral approach with elevation of the vastus lateralis is chosen, the lateral arthrotomy allowing to locate the tip of the trochlea (Fig. 6). The track of the osteotomy lies above the trochlea and is directed obliquely from above laterally to below on the medial femoral cortex. A wedge of bone is then excised from the distal femur with a 4-to 5-mm lateral base, corresponding to a 5-to 6-degree correction (Fig. 7). The osteotomy is fixed with the T-plate after placing the femur into valgus manually (Fig. 8). Once this stage is reached, the mechanical axis is rechecked so that the required correction at the level of the tibia can be calculated to achieve the preoperative objectives. Then, the wound is closed on a drain.
The last stage is to perform the HTO exactly in the fashion described above. The definitive axis is then displayed on the computer screen, and the osteosynthesis is checked with an image intensifier (Fig. 9).

| POSTOPERATIVE MANAGEMENT
The patient can stand up the day after the operation and walk with 2 crutches and partial weight bearing for 45 days (for HTO) and from 60 to 75 days (for DLO). Full range of motion is quickly regained after HTO. Doublelevel osteotomy requires more rehabilitation because of the distal femoral osteotomy, but we have never seen problems with stiffness in our experience.

Computer-Assisted HTO
Two years ago, we published a case-control study comparing computer-assisted HTO with similar cases performed without computer assistance. 24 Two groups were identified: group A represented 28 computernavigated osteotomies performed between March 2001 and April 2002, and group B represented 28 conventional osteotomies randomly identified from 140 osteotomies performed between January 1997 and December 2000. The 2 groups were matched for age (group A: mean, 54 years; range, 35Y71 years; group B: mean, 55 years; range, 27Y70 years), sex, the degree of osteoarthritis according to Ahlbäck's modified criteria (10 patients in stage 1, with G50% joint narrowing; 32 patients in stage 2, with 50%Y100% narrowing; 14 patients in stage 3, with complete loss of joint space, without bone loss or instability), and similar degrees of varus (group A: mean T SD, 173.3 T 3.8 degrees; range, 169Y178 degrees; group B: mean T SD, 172.79 T 3.18 degrees; range, 168Y178 degrees). The mechanical axis of the lower limbs was measured on full-length radiographs according to the protocol of Ramadier et al. 25 In both groups, a preoperative plan was made on the radiographs consisting of drawings of the axis, hypercorrection required, and opening wedge needed to give a 2-to 6-degree valgus or 182-to 186-degree mechanical axis (femorotibial mechanical axis angle).  The preoperative control of correction was checked differently in the 2 groups. In group B, a cardiac electrode was placed on the skin in position of the center of the femoral head. Under image intensifier guidance, a suture was then set from this electrode along the length of the lower limb to the medial malleolus, passing through the center of the knee. In group A, the correction was verified using the OrthoPilot. All patients had a follow-up radiograph at 3 months with the same protocol as preoperative for checking the goniometry.
The mechanical axis was checked with the OrthoPilot preoperatively and seen to concur well with the radiograph calculations (mean, 173.3 degrees; range, 169Y178 degrees, cf mean, 172.9 degrees; range, 169Y180 degrees).
In group A, the postoperative goniometry gave a mean T SD of 183.5 T 1.26 degrees with a median of 184 degrees and a range of 180 to 186 degrees. In group B, the mean T SD was 184 T 2.28 degrees, with a median of 184 degrees and a range of 181 to 189 degrees. The preoperative aim was attained in 27 (96%) of 28 cases in group A, compared with 20 (71%) of 28 in group B. This difference was statistically significant (Fisher exact test, P = 0.0248 G 0.05; Student t test, P = 0.0015), in favor of the computer-navigated cases. Not only was the objective achieved, but also a wide dispersion of results was avoided (especially in overcorrection).

Computer-Assisted DLO
Between March 2001 and April 2005, 11 double osteotomies (7%) were performed of a total of 158 osteotomies for genu varum at the knee. The series consisted of 4 female and 7 male patients with an average age of 48.5 years (range, 20Y62 years). One patient had a gross cosmetic deformity with no evidence of osteoarthritis; the remaining 10 patients had evident osteoarthritis. According to the modified Ahlbäck criteria, 19 6 patients had stage 3 osteoarthritis (complete loss of joint space without bone loss or ligament damage); 3 patients, stage 4 (bone loss without lateral joint-line opening); and 1 patient, stage 5 (bone loss with opening of the lateral joint space with or without posterolateral subluxation). The average radiological preoperative varus was 167.5 T 2.16 degrees (range, 164Y170 degrees), being notably higher than that of the HTO groups.
The objective was to obtain a mechanical axis of 182 T 2 degrees, a correction slightly less than for the HTO (184 T 2 degrees). All patients had a follow-up radiograph at 3 months according to the protocol of Ramadier et al. 25 We did not have any complication. The average preoperative mechanical axis before osteotomy was 168.1 T 1.1 degrees (range, 164Y170 degrees), an angle comparable with the preoperative radiograph calculation. After the osteotomy, the mechanical axis checked on the computer measured 182.7 T 1.1 degrees (range, 182Y184 degrees). At 3 months' follow-up, the radiograph evaluation gave an average axis of 180.8 T 1.6 degrees (range, 177Y182 degrees). The preoperative objective had been attained in all but 1 case (91%). No radiographs showed an abnormal oblique joint line.
All the osteotomies consolidated in 75 days, after which the patients were allowed full weight bearing with the aid of single walking stick. Despite having to perform an arthrotomy for the femoral intervention, there was no problem regaining knee flexion, with knee mobility 75 days postoperatively equal to that noted preoperatively.

| CONCLUSION
The computer-navigated HTO for osteoarthritis secondary to genu varum is a reliable and reproducible technique that has permitted us to achieve our preoperative aims in 96% of cases. The characteristics of the OrthoPilot (kinematic rather than image-based) seem to be especially well adapted in producing the osteotomy without arthrotomy in the majority of cases. The development of such techniques should regain the merit of the osteotomy for the treatment of unicompartmental osteoarthritis.
Regarding the computer-assisted double osteotomy in the treatment of excessive genu varum (tibia and femur), it has also proved to be reliable, precise, and reproducible. Using computer navigation simplifies a generally difficult procedure, enabling the surgeon to meet the preoperatively defined objectives. The development of such techniques is important to avoid altered joint-line mechanics, one of the main sources of difficulties when performing a TKA.