Plain radiographs are a poor indication of the overall coronal, sagittal and axial alignment of a total knee arthroplasty (TKA). We describe a new CT method that allows the mechanical axis in both planes to be defined and seven alignment characteristics to be defined.

A GE Light Speed multislice CT scanner performed a high-speed helical scan from the acetabular roof to the talus in 100 patients following TKA. The knees were scanned in a supine position with the legs in a neutral position. The images were reformatted in coronal, sagittal and axial planes and the mechanical and anatomical axes identified. The femoral component (varus/ valgus, flexion/extension, rotation) as well as the tibial – (varus/valgus, posterior slope and rotation) are measured. Coupled femoro-tibial rotational alignment was assessed by superimposition of the femoral and tibial axial images. The accuracy of this technique has been checked by using a mechanical FARO-arm.

The technique has a low intraobserver error rate of 9% (in each case less than 1 degree) and an accuracy of 3mm in a three-dimensional plane, as determined against an independent FARO arm technique. The CT analysis of 100 patients shows normal tibial baseplate rotation to be 8–12 degrees from the tibial tuberosity.

Conclusion: The CT protocol is the first single radiographic investigation that characterizes all the alignment parameters of a TKA. It sets an excellent standard in planning revision knee surgery and provides a valuable tool in assessing alignment of painful knee replacements as well as in outcome measures of TKA.

Introduction and Aims: A total knee replacement (TKR) needs to be well aligned for it to function well and avoid premature failure. Computer Assistance system has been designed to improve that alignment. The rationale for these systems rests on the assumptions that we currently have an alignment problem. To date, there have been no comprehensive survey of alignment outcomes and therefore no evidence of the size of the problem.

Method: The Perth CT protocol was used to measure the alignment of 164 primary TKRs, male/female = 76/88, mean age 69.4 years, performed by seven surgeons. One hundred and twenty-four patients were approached in the immediate post-operative phase and 40 within three years of surgery. Six alignment parameters were measured. The Perth Alignment Index was used as a measure of the quality of outcome. A parameter was judged perfect if it was within two degrees of the target alignment.

Results: In the coronal plane femoral alignment was ‘perfect’ in 74% (value range −3 to +6 degrees) and tibial alignment in 84% (value range −5 to +7 degrees). In the sagittal plane femoral flexion/extension was perfect in 75% (range −3 to +9 degrees), while tibial slope was perfect in 44% (value range −6 to +16 degrees). Femoral component rotation was perfect in 66% (range −8 to +9 degrees) and femorotibial matching was 45% (−14 to +13degrees). The mean PAI was 5.11:1.98, with only 14 patients (9%) having a perfect score.

Conclusion: The alignment of conventional jig-assisted TKR is far from perfect and efforts to improve the results seem to be justified.

Introduction and Aims: The Western Australian experience with the introduction and use of two different computer assistance systems for total knee replacement is presented.

Method: A multi-parameter CT assessment of TKR alignment has been devised. This has been used in a controlled cadaver study, a randomised controlled patient trial, two prospective clinical and radiological audits using the Stryker and the BrainLAB systems. A prospective audit of non-computer assisted TKRs is available for comparison. A total of 240 patients have had either primary or revision computer-assisted TKRs.

Results: Both computer assistance systems provide some advantages in alignment over the non-CAS TKRs. They provide excellent control in revision operations where bone loss needs to be augmented with bone graft. There is however significant morbidity associated with anchoring pin sites. There are pitfalls in accepting software upgrades. The cost-benefit analysis still needs to done.

Conclusion: Computer-assisted TKR has made a promising start but it is still in its infancy and the current versions may not be ready for widespread introduction

Introduction Patients are not universally pleased with their total knee replacements and satisfaction rates of 77% are quoted. This is in spite of quoted 10 year survivals for implants being in excess of 95%. We have looked at mal-alignment as a potential cause of the painful TKR.

Methods Twenty-five patients (21 male), mean age 78 (range 68 to 88 years), are presented. Each consulted the author because of pain in their total knee replacement. Twenty had had a primary replacement and five had at least one further procedure after the replacement. There was no evidence of infection in any of these. The Perth CT Protocol was used to investigate the alignment of the knees.

Results This group of patients had, on average, 8.6° of cumulative mal-alignment with 2.6 (of six) parameters being more than two degrees mal-aligned. The greatest deviations were in femorotibial matching (error rate 80%) with femoral rotation (error rate 68%) mal-alignment being the next most common. In a consecutive series of 42 primary, computer assisted knee replacements the mean mal-alignment index was 2.3:1.3. In patients with a cumulative error of six or less other causes of pain were found.

Conclusions It is suggested that mal-alignment of a knee replacement is a potential cause of pain and that a cumulative score of more than six probably represents the symptom threshold.

In relation to the conduct of this study, one or more of the authors is in receipt of a research grant from a non-commercial source.

Introduction: We describe a CT method that allows the seven alignment characteristics of a knee arthroplasty to be defined in a single investigation.

Method: A multislice CT scanner, scans in 2.5mm slices from the acetabular roof to the dome of the talus with the legs in a standard position.

The mechanical and anatomical axes are identified, from 3 dimensional landmarks, in both AP and lateral planes. The coronal and sagittal alignment of the pros-theses is then measured against the axes.

The rotation of the femoral component is measured relative to the transepicondylar axis. Tibial rotation was measured with reference to the posterior tibial condyles and the tibial tuberosity. Coupled femorotibial rotational alignment was assessed by superimposition of the femoral and tibial axial images.

The results of 100 scans show a low inter and intra observer error rate whilst independent assessment shows a mean measurement error of 3mm in a three dimensional plane. The radiation dose is 2.7mSV.

Conclusions: The technique provides the only currently available measure of all the alignment characteristics required to assess the quality of a knee arthroplasty. It will become a gold standard in planning revision surgery and provide a valuable tool in assessing alignment of painful knee replacements.

Aim To compare the new technique of computer assisted knee arthroplasty (CAK) against the current gold standard conventional jig based technique (JBK).

Methods Seventy-Five consecutive patients underwent knee replacement and were randomly allocated to either the CAK or JBK group. Pre and postoperative Knee society scores were collected. Post-operative CT scans were performed according to the Perth CT Knee Arthroplasty protocol and pre and post operative Maquet views of the limb performed. Intra operative soft tissue release together with post operative pain scores and blood loss where also assessed.

Results CT scans performed show a statistically significant improvement in component alignment when using computer assisted surgery for femoral varus/valgus (p=0.032), femoral rotation (p=0.001), tibial varus/valgus (p=0.047) tibial posterior slope (p=0.0001), tibial rotation (p=0.011) and femoraltibial mismatch (p=0.037). Standing Maquet limb alignment was also improved (p=0.004) as was blood loss (p=0.0001). CAK surgery took longer- a mean increase of 13minutes(p=0.0001).

Conclusions This is the first controlled study to assess all seven alignment characteristics of knee arthroplasty in these two groups of patients.

The improvement in alignment resulted in this trial being stopped prematurely as 6 out of 7 of the initial variables had reached significance. It shows a clear improvement in component alignment with computer navigation.

Introduction The aim of this study was to compare the new technique of computer assisted knee arthroplasty (CAK) against the conventional jig based technique (JBK).

Methods Seventy-five consecutive patients underwent knee replacement and were randomly allocated to either the CAK or JBK group. Post-operative CT scans were performed according to the Perth CT Knee Arthroplasty protocol to assess the accuracy of alignment. This measures seven parameters of alignment to an accuracy of one degree. Pre and post-operative Maquet views of the limb were also performed. Intra-operative soft tissue release together with post-operative pain scores and blood loss where also assessed.

Results CT scans performed show a statistically significant improvement in component alignment when using computer assisted surgery for femoral varus/valgus (p=0.032),femoral rotation (p=0.001),tibial varus/valgus (p=0.047) tibial posterior slope (p=0.0001), tibial rotation (p=0.011) and femoraltibial mismatch (p=0.037). Standing Maquet limb alignment was also improved (p=0.004) as was blood loss (p=0.0001). CAK surgery took longer, a mean increase of 13 minutes (p=0.0001).

Conclusions This is the first controlled study to assess all seven alignment characteristics of knee arthroplasty and use them to compare outcomes in conventional and computer assisted operations. It shows a clear improvement in component alignment with computer navigation.

In relation to the conduct of this study, one or more of the authors is in receipt of a research grant from a non-commercial source

Introduction Computer assistance can be valuable in positioning of knee prostheses when the bone interface is in the form of bone graft. The efficacy of this technique can be checked using the Perth CT Protocol for knee prosthesis alignment.

Methods Fourteen patients are presented who had an allograft revision total knee replacement. The entire prosthesis had to be removed and this resulted in bone deficits sufficently severe to require bone grafting. The Stryker computer navigation system was used. The final outcome was subsequently checked using a multi-slice CT which provided a six paramenter evaluation of the alignment of the knee prosthesis.

Results The technique produces excellent alignment of both components in the coronal plane, less good results in sagittal plane and the greatest problems are in the axial plane with femorotibial mismatch occuring in 50%. The mean mal-alignment index is 4.0:1.4. This compares with an index of 2.6:1.3 in navigated primary TKRs.

Conclusions Computer assistance provides significant help in the revision total knee replacement but does not produce perfect alignment in every case. Further refinement of the techniques are still needed.

Introduction: Plain radiographs are a poor indication of the overall axial and rotational alignment of a total knee replacement. A CT method is described to overcome these problems. Method: A multislice CT scanner took 1mm contingous slices from the acetabular roof to the dome of the talus with the legs immobile in a standard position. The scans are stored and reformatted to allow axial, coronal and sagittal images. The centre of femoral head, distal femur, tibial plateau and ankle joints are identiþed in three dimensions. The mechanical and anatomical axes are identiþed. The alignment of the prostheses is then measured against the mechanical axis in both AP and lateral planes. The rotation of the femoral component is measured relative to the transepicondylar axis. Tibial rotation was measured with reference to the posterior tibial condyles and the tibial tuberosity. Coupled femorotibial rotational alignment was assessed by superimposition of the femoral and tibial axial images. The accuracy of this technique has been checked by using a Ferro Arm which mechanically validates the mechanical axis. Results: The technique provides the only currently available measure of all the alignment characteristics required to assess the quality of a knee replacement. Conclusions: The Perth protocol provides an accurate assessment of alignment of the femoral and tibial components to a measured mechanical axis and rotational measurements of both components both individually and