35 patients with an infected total knee arthroplasty were operated with a two-stage revision protocol including the use of custom hand-made antibiotic loaded articulating spacers.

Spacers were built intraoperatively, without specific tools, regarless the defect being considered cavitary or segmentary.

Patients were allowed to walk with an orthosis. Range of motion (ROM) with the articulating spacer averaged 80° and after reimplantation 106.5°. All but two patients in our series were treated with a combination of antibiotics including rifampicin and the antibiotics used in the spacers constituted from 7.5% of its final weight.

Reimplantation was successfully performed in 33 out of 35 cases at an average time of 10.2 weeks, excluding a patient were we had to wait 2.5 years. An extended exposure at reimplantation was necessary in 21% of the patients (five “Q-snip” and two anterior tibial tuberosity osteotomies).

According to the Knee Society Score (KSS) the results were considered excellent or good in 27 patients (84.8%), and fair or poor in 6 patients (14.2%) out of the 33 reimplantated. No significant differences related to the micro-organism or the time elapsed until reimplantation, as compared for ROM and functional and clinical KSS were found but early infections had significant worst ROM than late or sub-acute infections. Intercondylar constrained designs had better functional, clinic KSS and ROM that posterostabilised designs or hinge designs without significant differences.

Our modified technique for custom made spacers can be applied in any surgical theatre with a minimum cost. Our results are comparable to those reported in the literature, demonstrating the consistency of the two-stage reimplantation protocol despite multiple modifications and different dosages of antibiotic used in the cement spacers.

Introduction: Defects in rotational alignment of the femoral component in total knee replacements (TKR) may cause poor alignment of the extensor apparatus. There are numerous references concerning the correct alignment of the femoral component of a prosthesis: transepicondylar axis, anteroposterior axis, and posterior condylar axis.

Materials and methods: Computer-assisted measurement of the relative differences between the transepicondylar axis, anteroposterior axis and posterior condylar axis in 38 TKR patients, excluding those with varus or valgus deformity greater than 15 degrees.

Results: The difference between the anteroposterior axis and the transepicondylar axis was 3.13 degrees of external rotation in the former.

Between the posterior condylar axis and the transepicondylar axis it was 1.18 degrees of internal rotation in the former.

Between the anteroposterior axis and the posterior condylar axis it was 5.51 degrees of external rotation of the former.

Conclusions: Probably the transepicondylar axis is the best landmark to enable reproducing the biomechanics of the knee in a patient bearing a prosthesis, although it is often difficult to reproduce it precisely. Several studies have noted errors among observers that are too great to make us feel certain that we are doing the best thing.

Although it is accepted that the perpendicular to the anteroposterior axis is reliable and corresponds to 4° of external rotation in relation to the posterior condylar axis, we have observed significant differences from one patient to another.

It would seem preferable to use a combination of the different axes, which we can do with a surgical browser.