After over 4 decades of experience with total knee arthroplasty, many lessons have been learned regarding surgical technique.

These include exposure issues, alignment methods, bone preparation, correction of deformity, implantation techniques and wound closure.

Where is the proper placement of the skin incision relative to the tibial tubercle? How does one safely evert the patella in the obese or ankylosed knee? Can a tibial tubercle osteotomy be avoided in the ankylosed knee? How does one protect the patellar tendon insertion from avulsing? How do you protect the soft tissues from debris and contamination and minimise the potential for infection? Can exposure be maintained if there are few surgical assistants? How do you find the lateral inferior genicular vessels and minimise postoperative bleeding? How do you know where to enter the intramedullary femoral canal for placement of the distal femoral alignment device? How can you avoid notching the anterior femoral cortex when in-between sizes or there is a pre-existing dysplastic trochlea? How can you correct a varus deformity without performing a formal MCL release? An inverted cruciform lateral retinacular release effectively corrects a severe valgus deformity and avoids the need for an LCL release. Trimming the posterior femoral condyles and removing posterior osteophytes is best accomplished using a trial femoral component as a template. Zone 4 femoral bone-cement radiolucencies can be minimised using the “smear” technique. The best indicator of potential postoperative flexion is not preoperative flexion but is intraoperative flexion against gravity measured after capsular closure.

Maltracking of the patella associated with TKA is usually the result of several factors coming together in the same patient. Causes of maltracking include residual valgus limb alignment, valgus placement of the femoral component, patella alta, poor prosthetic geometry, internal rotation of the femoral or tibial component, excessive patellar thickness, asymmetric patellar preparation, failure to perform a lateral release when indicated, capsular dehiscence, and dynamic instability.

Prior to wound closure after implantation of total knee arthroplasty, patellar tracking should be evaluated to assess the potential need for lateral release. The incidence of lateral release in the past was quite high in some series.

Most experienced surgeons will report a lateral release rate less than 5% for varus knees. It is usually higher for valgus knees because they are often associated with patella alta and preoperative subluxation. The classic intraoperative test for patellar tracking has been referred to as the “rule of no thumb” In this test, first suggested by Fred Ewald, the patella is returned to the trochlear groove in extension with the capsule unclosed. The knee is then passively flexed and one assesses whether or not the patella tracks congruently without capsular closure. If it does and the medial facet of the patellar component contacts the medial aspect of the trochlea no lateral release need be considered. If the patella dislocates or tilts, lateral release may be necessary. The test should be repeated with 1 suture closing the capsule at the level of the superior pole. If tracking then becomes congruent without excessive tension on the suture, no release is necessary. If tilting still persists, some surgeons like to assess tracking with the tourniquet deflated so that any binding effect on the quadriceps can be eliminated from the test. A tight PCL can also impart apparent patellar tilt as the femoral component is drawn posteriorly while the tibia (with its tubercle) moves anteriorly.

Lateral unicompartmental replacement is performed less frequently than medial replacement and is technically more difficult. The ratio of medial to lateral arthroplasties is approximately 10:1.

Differences in technique include the following:

The patella is more vulnerable to impingement on the leading edge of the femoral component and must be carefully recessed. Because the wear pattern in lateral disease is more posterior than in medial disease, there is often residual cartilage on the distal femoral condyle. This is also the case when UKA is performed for the sequella of a lateral plateau fracture. To avoid this impingement, residual cartilage should probably be removed from the distal condyle before its resection and the femoral component should be under-sized anteriorly.

Initial tibial resection should be very conservative to avoid the need for very thick tibial components to restore alignment and stability.

Err toward shifting the femoral component laterally and the tibial component medially to maximise M-L congruency.

Consider a medial parapatellar approach (avoiding the anterior horn of the medial meniscus) to facilitate visibility and component alignment.

To consider bilateral simultaneous knee replacement, both knees must have significant structural damage. It is best if the patient can't decide which knee is more bothersome. In borderline cases, ask the patient to pretend that the worse knee is normal and if so, would they be seeing you for consideration of knee replacement on the less involved side. If the answer to this question is “yes,” consider the patient a potential candidate for bilateral knee replacement. If the answer is “no,” recommend operating only on the worse knee, and expect that the operation on the second knee can probably be delayed for a considerable period of time.

Strong indications for bilateral simultaneous TKA are bilateral severe angular deformity, bilateral severe flexion contracture, and anesthesia difficulties, i.e., patients who are anatomically or medically difficult to anesthetise, such as some adult or juvenile rheumatoid arthritis patients or patients with severe ankylosing spondylitis.

Relative indications for bilateral simultaneous TKA include the need for multiple additional surgical procedures to achieve satisfactory function and financial or social considerations for the patient. Contraindications to bilateral TKA include medical infirmity (especially cardiac), a reluctant patient, and a patient with a very low pain threshold.

When performing bilateral simultaneous TKA, both limbs are prepped and draped at the same time. An initial dose of an intravenous antibiotic is given (usually 1g of a cephalosporin) before inflation of the tourniquet. Surgery begins on the more symptomatic side or on either side if neither knee is significantly worse than the other. The reason for starting on the more symptomatic side is in case surgery has to be discontinued after only one procedure owing to anesthetic considerations.

After the components have been implanted on the first side, the tourniquet is deflated and a second dose of intravenous antibiotic is administered (usually 500mg of a cephalosporin). After the joint capsule is closed and flexion against gravity is measured, one team completes the subcutaneous and skin closure on the first side while the other team inflates the second tourniquet and begins the exposure of the second side. When the second tourniquet is deflated, a third dose of antibiotic is given (usually 500mg of a cephalosporin for a total dose of 2g for both knees).

Because of concern about the potential for cross-contamination of the knee wounds when instruments used during the final stages of skin closure on the first knee are maintained on the field and used on the second knee, they should probably be handed off the field and outer surgical gloves changed.

Most patients will report after their complete recovery that they are glad they did both knees at the same time. A patient who has any uncertainty about proceeding with bilateral surgery should have only one knee done at a time. In many cases, the second side receives a “reprieve,” becoming more tolerable after the first side has been operated on.

To consider bilateral simultaneous knee replacement, both knees must have significant structural damage. It is best if the patient can't decide which knee is more bothersome. In borderline cases, ask the patient to pretend that the worse knee is normal and if so, would they be seeing you for consideration of knee replacement on the less involved side. If the answer to this question is “yes,” consider the patient a potential candidate for bilateral knee replacement. If the answer is “no,” recommend operating only on the worse knee, and expect that the operation on the second knee can probably be delayed for a considerable period of time.

Strong indications for bilateral simultaneous TKA are bilateral severe angular deformity, bilateral severe flexion contracture, and anesthesia difficulties, i.e., patients who are anatomically or medically difficult to anesthetise, such as some adult or juvenile rheumatoid arthritis patients or patients with severe ankylosing spondylitis.

Relative indications for bilateral simultaneous TKA include the need for multiple additional surgical procedures to achieve satisfactory function and financial or social considerations for the patient. Contraindications to bilateral TKA include medical infirmity (especially cardiac), a reluctant patient, and a patient with a very low pain threshold.

When performing bilateral simultaneous TKA, both limbs are prepped and draped at the same time. An initial dose of an intravenous antibiotic is given (usually 1 gram of a cephalosporin) before inflation of the tourniquet. Surgery begins on the more symptomatic side or on either side if neither knee is significantly worse than the other. The reason for starting on the more symptomatic side is in case surgery has to be discontinued after only one procedure owing to anesthetic considerations.

After the components have been implanted on the first side, the tourniquet is deflated and a second dose of intravenous antibiotic is administered (usually 500 mg of a cephalosporin). After the joint capsule is closed and flexion against gravity is measured, one team completes the subcutaneous and skin closure on the first side while the other team inflates the second tourniquet and begins the exposure of the second side. When the second tourniquet is deflated, a third dose of antibiotic is given (usually 500 mg of a cephalosporin for a total dose of 2 g for both knees).

Because of concern about the potential for cross-contamination of the knee wounds when instruments used during the final stages of skin closure on the first knee are maintained on the field and used on the second knee, they should probably be handed off the field and outer surgical gloves changed.

Most patients will report after their complete recovery that they are glad they did both knees at the same time. A patient who has any uncertainty about proceeding with bilateral surgery should have only one knee done at a time. In many cases, the second side receives a “reprieve,” becoming more tolerable after the first side has been operated on.

CURRENT INDICATIONS

The ideal patient for unicompartmental arthroplasty has been described as an elderly sedentary individual with significant joint space loss isolated to either the medial or lateral compartment. Angular deformity should be no more than 5 or 10 degrees off a neutral mechanical axis. Ideal weight is below 180 pounds. Pre-operative flexion contracture should be less than 15 degrees. At surgery, the anterior cruciate ligament is ideally intact and there is no evidence of inflammatory synovitis. (Kozinn, Scott, 1989) Indications for the procedure have broadened today because of the availability of less invasive operative techniques and more rapid recovery with UKA. Because of its conservative nature, the procedure is being thought of as a conservative first arthroplasty in the middle-aged patient. Because of its less invasive nature with more rapid recovery and potentially less medical morbidity, it is being considered as the “last arthroplasty” in the octogenarian or older.

There are basically 4 ways advocated to determine the proper femoral component rotation during TKA: (1) The Trans-epicondylar Axis, (2) Perpendicular to the “Whiteside Line,” (3) Three to five degrees of external rotation off the posterior condyles, and (4) Rotation of the component to a point where there is a balanced symmetric flexion gap. This last method is the most logical and functionally, the most appropriate. Of interest is the fact that the other 3 methods often yield flexion gap symmetry, but the surgeon should not be wed to any one of these individual methods at the expense of an unbalanced knee in flexion.

In correcting a varus knee, the knee is balanced first in extension by the appropriate medial release and then balanced in flexion by the appropriate rotation of the femoral component. In correcting a valgus knee, the knee can be balanced first in flexion by the femoral component rotation since balancing in extension almost never involves release of the lateral collateral ligament (LCL) but rather release of the lateral retinaculum. If a rare LCL release is anticipated for extension balancing, then it would be performed prior to determining the femoral rotation since the release may open up the lateral flexion gap to a point where even more femoral component rotation is needed to close down that lateral gap.

It is important to know and accept the fact that some knees will require internal rotation of the femoral component to yield flexion gap symmetry. The classic example of this is a knee that has previously undergone a valgus tibial osteotomy that has led to a valgus tibial joint line. In such a case, if any of the first 3 methods described above is utilised for femoral component rotation, it will lead to a knee that is very unbalanced in flexion being much tighter laterally than medially. A LCL release to open the lateral gap will be needed, increasing the complexity of the case. My experience has shown that intentional internal rotation of the femoral component when required is well-tolerated and rarely causes problems with patellar tracking. It is also of interest to note that mathematical calculations reveal that internally rotating a femoral component as much as 4 degrees will displace the trochlear groove no more that 2–3 mm (depending on the FC size), an amount easily compensated for by undersizing the patellar component and shifting it medially those few mm.

There are basically 3 ways to determine the proper tibial component rotation during TKA: (1) Anatomically cap the tibial cut surface with an asymmetric tibial component, (2) Align the tibial rotation relative to a fixed anatomic tibial landmark (most surgeons use this method and align relative to the medial aspect of the tibial tubercle), (3) Rotate the tibial component to a point where there is rotational congruency in extension between the femoral and tibial articulating surfaces. This third method must be used with fixed bearing arthroplasties (especially with conforming articulations) to avoid rotational incongruency between the components during weight-bearing that can create abnormal and deleterious torsional forces on posterior stabilised posts, insert tray interfaces and bone-cement interfaces. Rotating platform articulations can tolerate rotational mismatch unless it is to a point where the polyethylene insert rotates excessively and causes symptomatic soft tissue impingement.

Minimally invasive total knee arthroplasty is growing in popularity. It appears to reduce blood loss, reduce hospital stay, improve post-operative quadriceps function and shorten post-operative recovery. We show our experience of minimally invasive TKA with a computer navigation system.

The first series compared forty MICA TKA and forty conventional computer assisted total knee arthroplasties (CATKA). Component positioning was assessed radiographically with long leg Maquet views. Knee Society Scores (KSS) were recorded pre-operatively and at 6, 12, 18 months. Length of stay and recovery of straight leg raise was also recorded. A second series of fifty MICATKA patients were assessed post-operatively for component alignment using long leg Maquet views. Twenty-two of these patients had assessment of femoral rotation using CT.

In the first series pre-operative KSS showed no significant difference between the two groups. Post-operatively the mean femoral component alignment was 89.7 degrees for MICATKA and 90.2 for CATKA. The mean tibial component alignment was 89.7 degrees for both. KSS at 6, 12, 18 months were statistically better in the MICATKA (p<000.1). Straight leg raise was achieved by day one in 93% of the MICATKA compared to 30% of the CATKA. Length of stay for MICATKA was a mean of 3.25 days with CATKA a mean of 6 days. In the second series the mean femoral component varus/valgus angle was 89.98 degrees, the mean tibial component varus/valgus angle was 89.91 degrees and the mean femoral component rotation was 0.6 degrees of external rotation.

MICATKA is a safe procedure with reproducible results. Alignment is equivalent to CATKA. It gives statistically significant improvement in KSS compared to the open procedure. The length of stay and time to straight leg raise are also reduced. At 2 years follow-up we have seen no revisions and no evidence of loosening radiographically.

Introduction: The data reporting clinical and radiological outcomes after modern unicompartmental knee arthroplasty (UKA) for spontaneous avascular osteonecrosis are limited. We hypothesized that UKA for spontaneous osteonecrosis may be as reliable and durable as it is for osteoarthritis.

Materials and Methods: We retrospectively reviewed 40 cemented UKA operated for spontaneous osteonecrosis of the knee in two different centers between 1989 and 2004. Twenty-six patients were women and 14 men, mean patient age was 67 years (range, 45 to 84) and mean body mass index was 27.4 Kg/m2 (range, 18 to 44). Clinical and radiological evaluations were performed by an independent observer in each center according to the same protocol at a minimum follow-up of 3 years (mean 8 years; range, 3–17.5 years).

Results: The mean Knee Society Knee score improved from 60 preoperatively to 95 post-operatively. The mean Knee Society Function score improved from 50 preoperatively to 89 post-operatively. Restoration of an appropriate lower-limb mechanical axis was achieved for 36 knees (90%). Two knees were revised for aseptic loosening. The Kaplan–Meier survivorship was 95% at 12 years.

Discusssion and conclusion: Our data suggest UKA is a reasonable solution for restoring clinical function and radiological lower-limb alignment for spontaneous osteonecrosis of the knee, with a durable survivorship.

This study was to assess the accuracy of fixed posterior condylar referencing cutting blocks to the accuracy of combined epicondylar/AP axis referencing in femoral component rotation using a computer navigation system.

Seventy-five consecutive patients undergoing TKRs were randomized into two groups. The first received femoral component rotation by a computerized method that combined the epicondylar axis and Whitesides AP axis measurements to determine rotation. The second group had a zero or three-degree posterior referencing external rotation block, depending on which was closest to the epicondylar axis. All patients underwent axial CT scans of the distal femur to determine component rotation around the surgical epicondylar axis.

Femoral component alignment with the combined method as compared to fixed posterior alignment guides is statistically improved (p=0.001). In the posterior referencing group 43% were correctly rotated to the epicondylar axis but another 43% were malrotated by 3 degrees or more. The mean malrotation was 1.72 degrees (range 0–5) In the combined group 82% were correctly rotated and 11% were malrotated by 3 degrees or more. The mean malrotation was 0.51 degrees (range 0–4).

Conclusion: A combined computerized method of using the surgical epicondylar axis and Whitesides AP axis produces superior results when aiming for neutral femoral component rotation. Fixed posterior referencing blocks will produce errors in malrotation in over 50% of cases.

There are only a limited number of long term studies of total knee arthroplasty and few with a minimum fifteen year survivorship of a modular fixed bearing posterior cruciate-retaining prosthesis. This consecutive series of 139 total knee arthroplasties (109 patients), using the non-conforming posterior cruciate-retaining Press Fit Condylar (PFC®) system was followed for a minimum of 15 years (range 15.0 to 16.9 years). The patella were resurfaced with an all-polyethylene component in 83% of knees. The tibial component was always cemented, while a porous-coated femoral component was used in 84% of knees. Fifty-nine knees (45 patients) were followed up for a minimum of 15 years. Fifty-seven patients (70 knees) had died and five (8 knees) were too ill to assess. Survivorship of the prosthesis was confirmed for 98.6% of the prosthesis, as only two patients (2 knees) were lost to follow-up.

The mean Knee Society Score and Function Score were 96 and 78 respectively. The total incidence of radiolucent lines was 13%, with 2% around the femur, 11% around the tibia, and 0% around the patella. None of these lines were of any clinical relevance. There was no evidence of progressive radiolucent lines or component loosening, and one case of zone 4 femoral osteolysis.

There were five re-operations for any indication, of which four were for polyethylene insert wear. There was also one loose cemented femoral component after more than 15 years. The survival without need for revision for any reason was 99% at 10 years and 95.6% (worst-case scenario of 94.2%) at 15 years.

This single-surgeon series with a minimum 15 year follow-up shows that the modular fixed bearing posterior cruciate retaining total knee arthroplasty of the PFC system can provide excellent and predictable long term results in tri-compartmental arthritis of the knee.

To compare the new technique of computer assisted knee arthroplasty (CAK) against the current gold standard conventional jig based technique (JBK), 75 consecutive patients underwent knee replacement and were randomly allocated to either the CAK or JBK group. The CAK surgery was performed using a freehand technique that avoids violation of the medullary canal. Pre-and post-operative 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 postoperative pain scores and blood loss were also assessed.

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 femoral-tibial 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).

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. The reduction in blood loss in this surgery through not violating the medullary canal will also be beneficial.

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.

A consecutive series of 139 total knee arthroplasties (109 patients, average age 67 years), using a non-conforming posterior cruciate-retaining prosthesis was followed for 15 years (range, 15.0 to 16.9 years).

Forty-five patients (59 knees) were clinically and radiographically evaluated, 57 (70 knees) had died, five patients (8 knees) were too ill to assess, two patients (2 knees) were considered lost to follow-up. Survivorship analysis was performed using worst case scenario analysis and failure defined as re-operation for any reason.

Results: In this series there were five re-operations for any indication, four of which were for polyethylene insert wear. One loose cemented femoral component was revised at 15 years. The survival without revision or need for any re-operation was 92.6% at 15 years. The mean Knee Society Score and Function Score at 15 year follow-up are 96 and 78 respectively. The prevalence of radiolucent lines was 13%. None of these lines were clinically relevant. There was no evidence of progressive radiolucent lines, and one case of asymptomatic femoral osteolysis.

Conclusions: This single-surgeon series with a minimum 15 year follow-up shows good clinical and radiological results with excellent survivorship of a modular fixed bearing posterior cruciate retaining total knee arthroplasty system.

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: 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.

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 In an attempt to decrease the incidence of posterior hip dislocation following a posterior approach, a simple capsulorrhaphy was utilized in 255 consecutive primary total hip arthroplasties performed by one surgeon.

Methods All patients were reviewed at a minimum of two years post-operatively and no patient was lost to follow-up. One patient sustained a posterior hip dislocation, while there were no anterior hip dislocations. The dislocation rate of 0.39 is equal to or less than the rates of dislocation reported in the literature using a direct lateral approach.

Conclusions We postulate that this capsular repair creates not only a static restraint but also a capsule and gluteus medius mediated proprioceptive feedback to guard against extremes of internal rotation of the hip.

Introduction There are only a limited number of long term studies of total knee arthroplasty but none with a minimum 15 year survivorship of a modular fixed bearing posterior cruciate-retaining prosthesis.

Methods We present a consecutive series of 139 total knee arthroplasties (109 patients, average age 67 years), using a non-conforming posterior cruciate-retaining prosthesis, followed for a minimum of 15 years (range, 15.0 to 16.9 years). The patella was resurfaced with an all-polyethylene component in 83% of knees. The tibial component was always cemented, while a porous coated femoral component was used in 84% of knees. Fortyfive patients (59 knees) were followed-up for a minimum of 15 years, 57 (70 knees) had died, five patients (8 knees) were too ill to assess, two patients (two knees) were considered lost to follow-up.

Results In this series there were five re-operations, four of which were for polyethylene insert wear. At two of these, the patella was exchanged for early surface wear and one patella was resurfaced for the first time. There was one loose cemented femoral component after more than 15 years. The survival without revision or need for revision for any reason was 99% at 10 years and 95.6% (worst case scenario of 94.2%) at 15 years. The mean Knee Society Score and Function Score at 15 year follow-up was 96 and 78 respectively. The total incidence of radiolucent lines was 13%, with two percent around the femur, 11% around the tibia and zero percent around the patella. None of these lines were of any clinical relevance. There was no evidence of progressive radiolucent lines or component loosening, and one case of zone four femoral osteolysis.

Conclusions This single-surgeon series with a minimum 15 year follow-up, and excellent clinical, radiological and survivorship results provides a benchmark upon which other long term studies of modular fixed bearing posterior cruciate retaining total knee arthroplasty can be compared.

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: 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