Early operative infection should be a rare occurrence today with the use of improved sterile technique , more efficient surgery with shorter operating times and perioperative antibiotics. Most surgeons should experience an early operative infection rate between 0 and 0.3%. Infection is most likely to occur in the multiply operated, diabetic, or morbidly obese patient or when there has been a wound healing issue. Many factors likely accounted for the low (0%) primary infection rate in this series [12].

Late metastatic infection from a remote focus can occur throughout the life of the patient. Common sites that can cause a bacteremia that seeds the joint are the oral cavity, gastrointestinal and urinary tract, or skin site (usually cellulitis). Some of these cases are inevitable despite education of patients and treating physicians about their possibility. They are currently responsible for about 12% of reoperations (32 among the 264 knees reoperated in this series) and may eventually become the leading cause as other issues are minimized. The incidence of late metastatic infection was 0.6% at mean follow-up of 15 years with an occurrence rate of 0.04% per year of follow-up.

Patients must be educated as to the prevention of late infections Dental prophylaxis has become optional. Some surgeons recommend dental prophylaxis for 2 years following the arthroplasty while others recommend lifetime coverage.

Isolated femoral component loosening has been very rare in this series whether the component was cemented or cementless. Failure of either femoral fixation method accounted for approximately 3% of reoperations. The incidence of femoral loosening was 0.2% at mean 15-year follow-up occurring at a rate of 0.01% per year.

The success of a cementless femur could be dependent on the quality of primary fixation at the time of the arthroplasty. In this series, suitability for cementless fixation was subjectively determined by testing the difficulty of removing the femoral trial using an extractor with a slap hammer.

Both cemented and cementless femurs are vulnerable to late loosening in the presence of severe osteolysis, and this data is included in the section on wear and lysis.

Isolated cemented tibial component loosening is also infrequent and also accounts for approximately 2% of reoperations. It also has an extremely low incidence of 0.01% per year failure over the first 15 years. As seen on the femoral side, tibial loosening is more prevalent in the presence of wear debris-induced osteolysis.

Cementless tibial components are more likely to loosen than cemented components with the incidence dependent on their design and the accuracy of implantation. Components fixed with screws or those using metals with high in growth potential are more likely to enjoy long-term success. In this series, there were only 40 cementless porous titanium tibias implanted, none with screw fixation. At mean 25-year follow-up, four had loosened for an incidence of 10%. Conclusions about the success of cementless tibial fixation using current designs cannot be judged by the results in this series of this outdated design.

As mentioned earlier, patellar problems were the leading cause of reoperation some 25 years ago. These complications included worn metal-backed patellar components, patellar fracture when large central fixation lugs were in vogue, and patellar loosening when fixation was via a small central fixation lug. Since the mid-1980s three-pegged all-polyethylene patellar components (as were used in this series) became the state of the art. Metal-backed patellar components were abandoned in this series in 1986. Of 87 implanted, nine have worn through and required revision [13].

With this fixation method, both loosening and fractures were rare. Of 4699 all-poly 3 lugged patellas implanted, there have been three reoperations for loosening, one for a fracture, and one for a patellar clunk. Shearing off of the three lugs occurred in four cases with an early design. Its etiology required the presence of an abnormal shearing force caused by imbalance in the quadriceps mechanism. The conformity of the prosthetic articulation tended to keep the patella located in the trochlear groove, while the soft tissue imbalance pulled the patella toward the lateral side, resulting in the shear forces. Once the junction between the lug and the patellar component was reinforced by manufacturers, this complication has not been seen.