The volume of joint replacement surgery has risen steadily in recent years, because the population is aging and increasingly wishes to reserve a high functional status onward into old age. Infection is among the more common complications of joint replacement surgery, arising in 0.2% to 2% of patients, or as many as 9% in special situations such as the implantation of megaprostheses. The associated morbidity and mortality are high. It is thus very important to minimize risk factors for infection and to optimize the relevant diagnostic and therapeutic strategies.This review is based on pertinent publications retrieved by a selective search in PubMed, including current guidelines and expert recommendations.The crucial diagnostic step is joint biopsy for the identification of the pathogenic organism, which succeeds with over 90% sensitivity and specificity. If the prosthesis is firmly anchored in bone, the pathogen is of a type that responds well to treatment, and symptomatic infection has been present only for a short time, then rapidly initiated treatment can save the prosthesis in 35-90% of cases. The pillars of treatment are thorough surgical care (radical débridement) and targeted antibiotic therapy. On the other hand, if the prosthesis is loose or the pathogen is of a poorly treatable type, the infection can generally only be cured by a change of the prosthesis. This can be performed in either one or two procedures, always in conjunction with systemic antibiotic therapy tailored to the specific sensitivity and resistance pattern of the pathogen.The risk of infection of an artificial joint is low, but the overall prevalence of such infections is significant, as the number of implanted joints is steadily rising. Artificial joint infections should be treated by a standardized algorithm oriented toward the recommendations of current guidelines. Many of these recommendations, however, are based only on expert opinion, as informative studies providing high-grade evidence are lacking. Thus, for any particular clinical situation, there may now be multiple therapeutic approaches with apparently comparable efficacy. Randomized trials are urgently needed.