OBJECTIVES: The purpose of this study was to present the evidence for the use of prophylactic antibiotics in total knee arthroplasty.

METHODS: To analyze the use of prophylactic antibiotics and occurrence of infections, the retrospective cohort was constructed using health insurance claims data. Following a review of international guidelines, and Health Insurance Review & Assessment Service survey items that evaluated the appropriateness of prophylactic antibiotic treatment, prophylactic antibiotics were defined as 1st through 4th generation cephalosporin, streptomycin, other aminoglycosides, glycopeptide antibacterials, and intravenous antibiotics of the quinolone. To analyze the incidence rate of infections according to the appropriateness use of prophylactic antibiotics, patients were classified as follows: administration of prophylactic antibiotics for two days or less from the start of convalescence during hospitalization for total knee arthroplasty (adherent prophylactic antibiotic group), and administration of prophylactic antibiotics lasting longer than two days (non-adherent prophylactic antibiotic group).

RESULTS: From 2008 to 2016, the total number of insurance claims for the first total knee arthroplasty was 473,034. The average days of antibiotic use per one total knee arthroplasty gradually reduced from 8.87 days in 2008 to 7.79 days in 2016. The analysis results using Cox's proportional hazard model for surgical site infection in all patients showed a nonsignificant difference (HR: 0.87, 95% CI: 0.68, 1.12, p=0.27) between the adherent and nonadherent groups. The analysis of subgroups according to age, type of antibiotic, type of medical care institution, and the number of antibiotic administration days per medical care institution also showed a nonsignificant difference in the total number of surgical site infections between the groups.

CONCLUSIONS: Current real-world practice shows that the use of prophylactic antibiotics has gradually reduced and there was no significant difference in surgical site infection between the adherent and nonadherent groups.