Introduction

Total knee replacement (TKR) has become a standard operative procedure. Long term results show approximately 10% aseptic loosening after 10 years [10, 24]. Still, aseptic loosening is one of the major problems connected to TKR. Many studies have been performed to more closely understand the interaction between human tissue, metallic and polyethylene implants [7, 9, 13, 15, 28]. After establishment of conventional total knee replacement, numerous studies have focussed on tribology, metal alloys, polyethylene structure and design of implants to decrease polyethylene wear.

Axial alignment of the limb with restoration of the mechanical axis is a determinant of the outcome. A mechanical axis within a range of ±3° varus/valgus is thought to be associated with a better outcome. However, in previous studies post operative alignment of the limb exceeded a range of ±3° in up to 30% of knees. Petersen and Engh [22] reported the radiological results of 50 primary TKAs performed using conventional technique. In their study, 26% of TKAs failed to achieve an alignment within ±3° varus/valgus. Mahaluxmivala et al. [19] analysed 673 TKAs. They found an alignment of more than ±3° varus/valgus in 25% of cases, independent of the surgeon’s experience. Although various guides for alignment have been designed to improve accuracy, several limitations of these instruments are reported [8, 16, 21, 23]. Errors may be due to variations in the bony anatomy, visual misjudgement by the surgeon or limitations of technique. As malalignment and instability can lead to earlier loosening of prosthesis, computer assisted surgery (CAS) was brought into the field. It allows correct axial alignment and has been introduced and established in TKR over the past years [2, 3, 6, 14, 18, 25, 27]. It is suggested that a better reconstruction of leg alignment and a stable knee over the full range of movement leads to better long term results and higher patients satisfaction. It was therefore the aim of the authors to test the following hypothesis: there are differences for

  1. (a)

    subjective (WOMAC score) and

  2. (b)

    objective (Knee Society score, medio-lateral stability, Biodex™-measurement) criteria between CAS TKR versus freehand TKR after 2 years.

Methods

We performed a retrospective matched-pair analysis with investigation of 50 patients who received primary TKR due to primary and varus osteoarthritis from August to October 2002 (out of 156 within this period) Other aetiologies (rheumatoid arthritis, trauma, valgus deformities) were excluded. Both groups were similar in age, gender, BMI, treated leg, preoperative leg deformity and operation technique (Table 1). Twenty-five patients received a cruciate retaining PFC Sigma® total knee prosthesis (DePuy®, Warzwa, USA) with fixed bearing within a computer assisted implantation (VectorVision®, BrainLAB®, Munich, Germany). Twenty-five patients received a cruciate retaining PFC Sigma® total knee prosthesis (DePuy®, Warzwa, USA) with fixed bearing in the standard freehand conventional technique. No patellar replacement was performed in both groups. All surgeons were experienced using both techniques, no learning curve for computer assisted implantation of TKR was included. Implantation followed after standardised spinal anaesthesia with tourniquet. Median skin incision was followed by median parapatellar arthrotomy. In both groups the tibia cut was perpendicular to the mechanical axis and the femoral component was rotated parallel to the posterior axis. The medial and lateral posterior condyle was marked manually as well in the CAS as the freehand group. In the CAS group it was digitised by the computer and in the freehand group the cutting jig was rotated according parallel to this axis. The only difference between both groups was that the implantation of group B was performed under control of a navigation system, which has an optical tracking unit which detects reflecting marker spheres using an infrared camera. For this procedure two reference arrays with passive marker spheres have to be rigidly attached to the femur and tibia.

Table 1 Demographic parameters of both treatment groups
Full size table

Standardised postoperative procedure followed by full weight bearing on day one after operation with two crutches and daily continuous passive motion and physiotherapy for 3 weeks. Leg alignment was obtained postoperatively by long leg weight bearing radiographs.

Follow-up measurement

Two years after TKR all patients were invited to take part within a clinical follow-up. The Western Ontario and McMaster's Universities Osteoarthritis Index [4, 5] (WOMAC) as well as the Knee Society score [12] was obtained in all patients. We performed standardised fluoroscopic assisted stress radiography to test varus-valgus stability in extension with equal medial and lateral force of 15 N (Telos®, Marburg, Germany, SE 2000; Fig. 1) and 80° flexion according to the established technique described by Stähelin [26].

Fig. 1
figure 1

Standardised fluoroscopic assisted stress radiography to test varus-valgus stability in 80° flexion

Full size image

After similar warm-up programme each patient was investigated according to a standardised measurement using an isokinetic muscle force protocol with 60° and 180° per second (Biodex™—3 dynamometer; Biodex Medical Systems, Inc., New York, USA).

Statistical analysis

Assessment for all patients in both groups was performed by one orthopaedic surgeon. Statistical analysis was performed using the paired Student t test. Two-tailed values of P < 0.05 were considered statistically significant. For comparing WOMAC score, Knee Society score, medio-lateral stability and muscle force between both groups, box plots were used. The box height is the interquartile range which represents half of all values. Twenty-five percent of values are higher and 25% of values are lower than the box. The median is displayed as a horizontal line across each box. The minimum and maximum values are represented with the vertical lines. Analysis of the data was performed using the SPSS statistical package (version 11.5, SPSS Inc., Chicago, USA).

Results

WOMAC score

No statistical differences were found between both groups for WOMAC A, B, C and total score. For details see Fig. 2.

Fig. 2
figure 2

Boxplots for WOMAC score representing the lower quartile (25th percentile) and upper quartile (75th percentile), the interquartile range and the median

Full size image

Knee society score

No statistical differences were found for the Knee score, the functions score and the Total Knee Society score. For details see Fig. 3.

Fig. 3
figure 3

Boxplots for Knee Society score representing the lower quartile (25th percentile) and upper quartile (75th percentile), the interquartile range and the median

Full size image

Postoperative leg alignment

For postoperative leg alignment which was measured on standardised weight bearing long leg X-rays, we found statistically significant better values for the CAS group. Deviation from the neutral leg axis: CAS 1.32° (SD 1.09); conv 2.24° (SD 1.32; P = 0.04).

Varus-valgus stability

The standardised fluoroscopically investigated varus-valgus stability in extension and flexion gave no statistically better values for one or the other group. Stability was measured in degrees. In extension we found for varus stress CASvarus 3.88°, SD ±1.63°; convvarus 3.08°, SD ±1.29°; P = 0.05. Values for valgus stress in extension revealed: CASvalgus 3.28°, SD ±1.51°; convvalgus 2.8°, SD ±0.89°; P = 0.12.

Lateral opening with 80° Flexion and varus stress showed following values: CASvarus 6.4°, SD ±3.5°; convvarus 6.76°, SD ±3.4°; P = 0.37, concerning medial opening with valgus stress no statistically values could be detected: CASvalgus 4.04°, SD ±1.34°; convvalgus 3.4°; SD ±1.98°; p = 0.09.

Biodex™ measurement

Biodex™ measurement showed no statistically significant values for isokinetic muscle force. For details see Fig. 4.

Fig. 4
figure 4

Boxplots for Biodex™ Measurement representing the lower quartile (25th percentile) and upper quartile (75th percentile), the interquartile range and the median

Full size image

Range of motion

Range of motion was slightly different but no statistically values were found. Maximal flexion in the CAS group was 113° (SD 12.91) and in the freehand group 108.8° (SD 13.73; P = 0.06).

Discussion

In recent years computer assisted surgery has gained very much interest in total knee arthroplasty with superior results concerning axial leg alignment compared to the conventional technique [14–20]. In our study we found a postoperative leg alignment within the range of 3°varus–3°valgus in all patients. The accuracy of component alignment was better in the CAS group as well. To our knowledge there is no previous study that describes clinical parameters and fluoroscopically measured ligamentous stability in a follow-up collective of 2 years postoperatively after computer assisted versus conventional freehand total knee replacement. Most of the currently available studies are leg alignment studies without any clinical follow-up according to patient’s satisfaction and knee function. As navigation in total knee replacement is performed only since a few years, we aimed to elucidate the effect of navigation to the functional outcome and patient’s satisfaction 2 years after primary implantation. We focussed on WOMAC score, Knee Society score, range of motion and ligamentous stability. One of the authors performed a 2 year outcome study comparing CAS TKR fixed versus mobile bearing [17]. Another study by Martin et al. [20] investigated CAS TKR image less versus image based technique.

According to WOMAC score we found good results for both groups, but without statistically significant differences between both groups (P = 0.45). The total score is comparable to other studies in the current literature [1]. When focussing on the Knee Society score described by Insall et al. [12] no statistically significant differences for either the knee sub-score or the function sub-score could be seen. However, there was a tendency of better values in the CAS group in both the sub scores as the total score.

In terms of range of motion as outcome criteria we could see better values in the CAS group (mean 113° versus 108°) compared to the freehand technique but without statistically significant difference (P = 0.06). Other authors describe a similar range of motion in their freehand outcome studies [9].

One of the cornerstones in TKR is the postoperative joint stability. It is well known that a stable joint over the full range of motion is superior to instable joints according to patient’s satisfaction, polyethylene wear and longevity. In our patients we did not see any statistically significant difference in terms of ligamentous stability as well for full extension as for 80° flexion. Lateral and medial opening was within a range of 2.5°–3.5° in extension in both groups and slightly higher in flexion as recommended by most knee surgeons.

To evaluate isokinetic muscle force objectively, Biodex™ measurement for each patient at angular velocities of 60 and 180° per second was performed. In general the quadriceps femoris muscle is stronger than the flexor muscles, which we also found in our patients. No differences were seen for the flexion peak torque at both angular velocities in the CAS group compared to the freehand group. Up to now there is no comparable work, which investigates muscle force with biodex® measurement, although it is a useful and standardised tool.

Although we did not see statistically significant differences in our series except for the postoperative leg alignment the authors recommend the use of a navigation system because of the described advantages such as better leg alignment and better component orientation [3, 27]. Our study has some limitations. We present only a small number of patients on a short term investigation of 2 years. This might result in some bias. Further, inaccuracies in determining the leg axis on postoperative weight bearing long leg radiographs might be another reason for deviations of the postoperative mechanical leg axis. Haaker et al. [11] reported an inaccuracy of 2° when analysing the mechanical leg axis on long leg radiographs, while it might be even 5° when using short leg radiographs. Nevertheless, this possible systematically inaccuracy is similar for all three groups. Flexion stability measurement has been performed in a standard technique. Although the thigh was fixed to the measurement tool (Fig. 1) inaccuracies in terms of rotation of the hip are possible. It is hard to quantify these. Nevertheless, these inaccuracies are the same for both groups.

Conclusion

As we could show in the current study, there are no statistically significant differences for 2 year follow-up results between CAS TKR and freehand TKR concerning WOMAC and Knee Society score and Biodex™ measurement. The Knee Society score and range of motion were better in patients operated on using CAS TKR. However, the differences were not statistically significant. Postoperative leg alignment is statistically better in the CAS group. Therefore we conclude that the hypothetical advantages of computer assistance in total knee replacement for functional parameters and patient’s satisfaction is still not yet proven.