All major studies have incorporated the use of prolonged courses of parenteral or oral antibiotic therapy in the management of two-stage revision of an infected total knee arthroplasty. We present a series of 59 consecutive patients, all with microbiologically-proven deep infection of a total knee arthroplasty, in whom a prolonged course of antibiotic therapy was not routinely used. The mean follow-up was 56.4 months (24 to 114). Of the 38 patients who underwent a staged exchange, infection was successfully eradicated in 34 (89%) but recurrent or persistent infection was present in four (11%). Our rate of cure for infection is similar to that reported elsewhere. We conclude that a prolonged course of antibiotic therapy seems not to alter the incidence of recurrent or persistent infection. The costs of the administration of antibiotics are high and such a regime may be unnecessary.