A modular tibial insert exchange is a seemingly attractive benign and simple surgical alternative when compared to full knee revision. Unfortunately, the results have been less than satisfactory with modular insert exchange for polyethylene wear and knee instability. Babis et al reported the results of 56 isolated insert exchanges performed for wear or instability. The re-revision rate was 25% at a mean follow-up of 3-years and the cumulative survival rate was only 63.5% at 5.5 years. In another study, 27% (6/22) insert exchanges for wear required re-revision within 5 years. Bert et al reported scoring and damage in 89% of 55 retained components considered candidates for isolated insert exchange. Such damage could account for accelerated wear of a new insert. These studies are misleading. The new insert must be polyethylene not prone to oxidation and accelerated wear. In a recent study of 177 revisions for wear and osteolysis, the survivorship of insert exchange using non-irradiated poly was 100%. Insert exchange does not correct the problem of a poor tibial locking mechanism. Whiteside and Katerberg reported 3 failures in 49 insert exchanges, fabrication of the tibial locking mechanism was used to address this problem. With revision for instability, insert exchange must provide full stability in both flexion and extension. Tibial insert exchange must correct the underlying cause of failure that led to the revision surgery. Full knee revision is a complex procedure that brings with it increased risks of perioperative complications such as infection and should be reserved only for cases that will not do well with simple insert exchange.
Implant wear and osteolysis have been a major cause of failure of total knee implants. In the mid-1990s, manufacturers recognised the impact of oxidation on wear with implants sterilised by gamma irradiation in air and changed their methods of sterilisation. This has resulted in a dramatic reduction in wear. In retrieval studies, non-irradiated polyethylene has not shown the fatigue type of failure associated with oxidation. The percentage of revisions done at the Anderson Clinic for polyethylene wear for osteolysis has dropped from 44% in the late 1990s to 4% in the past decade. With the continued use of polyethylene free of oxidation, we anticipate a further reduction in the need for revision surgery secondary to wear and osteolysis. Highly cross-linked polyethylene was introduced to further reduce wear with total knee implants. Higher levels of irradiation used increases crosslinking in the polyethylene but the material strength is reduced. Although volumetric wear is reduced, the wear particles are smaller in size and potentially more bio-reactive. The Manufacturer And User Device Experience Database (MAUDE) reports describe early implant breakage and osteolysis of highly cross-linked polyethylene inserts. Implants that were highly crosslinked with quenched free radicals demonstrate increased levels of oxidation after retrieval unlike, never implanted components “off-the-shelf”. Backside wear remains a concern as non-modular implants have better long-term survivorship compared to their modular counterparts. These reports should temper enthusiasm for using highly cross-linked material in knees or modular tibial components until longer term clinical and retrieval studies have been completed.
A previous incision, usually near the midline, is always present and should be used whenever possible. It is preferable to use the most recent and/or most lateral longitudinal old incision. In addition, the surgeon should follow general principles of surgical exposure; i.e., avoiding isolated strips of tissue and crossing transverse incisions at right angles. Whenever adherent skin or marginal skin is present, a sham incision, tissue expansion, and tissue transfer should be considered. Protecting the patellar tendon from avulsion is essential during revision surgery. In routine revision cases, a 15.75 mm pin should be placed in the center of the tibial tubercle to avoid distal avulsion. To avoid interfering with the stem location of a revision component, the pin should be directed toward the lateral plateau. Whenever the patient has less than 90° of knee flexion, the extensor mechanism has lost its elasticity and must be relaxed either proximally, by dividing the quadriceps tendon, or distally, by performing a tibial tubercle osteotomy. The classic rectus snip is performed at the musculotendinous junction of the rectus femoris. The rectus snip may improve knee flexion at the expense of a minor extension lag. If the extensor mechanism is relaxed distally, the tubercle osteotomy should be 4 to 6 cm in length, taking care to preserve the fascial attachments on the lateral side of the osteotomised segment. After surgery, the fragment is reattached with heavy gauge wire. A disadvantage is that the osteotomy is subcutaneous and postoperative wound drainage in this location can be problematic. In some instances capsular scar is extensive. The scar must be excised and the medial and lateral gutters cleared to achieve knee flexion. In extreme cases, the knee may need to be destabilised in flexion to gain exposure. This is accomplished either with a peel of the capsule and ligaments from the medial side of the femur, or with an epicondylar osteotomy that detaches the collateral ligaments from the medial condyle. The osteotomised bone wafer is reattached with heavy sutures to the medial condyle prior to closure.
In 1972, unicondylar arthroplasty (UKA) was introduced, along with total knee arthroplasty (TKA), as an option for managing gonarthrosis. Although the early clinical results with the first generation of implants were equivalent to those of total knee arthroplasty, little interest in UKA was sustained. If unicondylar arthroplasty is to realise a role in the management of degenerative arthritis, even as a temporising procedure, the results must be predictable and reproducible. Patient satisfaction must be equivalent to or better than that of TKA. Finally, the conversion of UKA to TKA must be uncomplicated, avoiding complex reconstructive procedures and the use of revision implants. UKA achieves these goals. As documented by such things as reduced blood loss and risk of infection, morbidity has always been less with unicondylar arthroplasty. Patients with both a UKA and a TKA on the contralateral side generally prefer the unicondylar knee. This is partly because a UKA provides a superior range of motion and better function with such activities as stair climbing. Adapting the surgery to an outpatient operative procedure using a minimally invasive incision has enhanced patient satisfaction. In most studies, the revision of a failed unicondylar arthroplasty using primary TKA components has been predictable and durable. Osteolysis has not been reported with failed UKA; therefore bone defects usually are minimal. If major tibial bone defects are present, a revision tibial component and proper bone defect management will achieve excellent results. In conclusion, we cannot only justify UKA as a temporising procedure, but also as a definitive procedure with long-term results that are comparable to TKA for gonarthrosis.
We examined 86 polyethylene inserts, retrieved from total and unicompartmental knee prostheses after an average of 39.5 months in situ, grading them from 0 to 3 for seven modes of polyethylene degradation. Severe wear, with delamination or deformation, was observed in 51% of the implants, and was associated with time in situ, lack of congruency, thin polyethylene, third-body wear debris, and heat-pressed polyethylene. Significant under-surface cold flow was identified in some areas of unsupported polyethylene, and was associated with delamination in the load-bearing areas of thin inserts above screw holes in the underlying metal tray. We recommend the use of thicker polyethylene inserts, particularly in young, active patients and in designs with screw holes in the tibial baseplate. Thin polyethylene inserts which are at risk for accelerated wear and premature failure should be monitored radiographically at annual intervals.
