BACKGROUND: Identification of the infecting organism is critical to the successful management of deep prosthetic joint infections about the hip and the knee. However, the number of culture specimens and which culture specimens are best to identify these organisms is unknown. METHODS: We evaluated 113 consecutive patients with infected total hip and total knee arthroplasties and correlated the type of culture specimen and number of specimens taken during surgery to the likelihood of a positive culture result. From these data, we subsequently developed a model to maximize culture yield at the time of surgical intervention. After exclusions, 74 patients meeting the Musculoskeletal Infection Society criteria were left for final analysis. RESULTS: From this cohort, 63 of 74 patients had a positive culture result (85%). The odds of a fluid culture result being positive was 35 of 47 (0.75), whereas the likelihood of tissue cultures yielding a positive result was 164 of 245 (0.67; P = .313). The sample designated "best culture" specimen was the only culture with a positive result in 1 of 48 cases in which a best culture was identified. The optimal number of cultures needed to yield a positive test result was 4 (specificity = 0.61 and sensitivity = 0.63). Increasing the number of samples increases sensitivity but reduces specificity. CONCLUSION: A minimum of 4 tissue cultures from representative areas is necessary to maximize the chance of identifying the infecting organism during management of the infected total hip and total knee arthroplasties. The designation of the best culture specimen for additional testing is arbitrary and may not be clinically efficacious.
BACKGROUND: Identification of the infecting organism is critical to the successful management of deep prosthetic joint infections about the hip and the knee. However, the number of culture specimens and which culture specimens are best to identify these organisms is unknown. METHODS: We evaluated 113 consecutive patients with infected total hip and total knee arthroplasties and correlated the type of culture specimen and number of specimens taken during surgery to the likelihood of a positive culture result. From these data, we subsequently developed a model to maximize culture yield at the time of surgical intervention. After exclusions, 74 patients meeting the Musculoskeletal Infection Society criteria were left for final analysis. RESULTS: From this cohort, 63 of 74 patients had a positive culture result (85%). The odds of a fluid culture result being positive was 35 of 47 (0.75), whereas the likelihood of tissue cultures yielding a positive result was 164 of 245 (0.67; P = .313). The sample designated "best culture" specimen was the only culture with a positive result in 1 of 48 cases in which a best culture was identified. The optimal number of cultures needed to yield a positive test result was 4 (specificity = 0.61 and sensitivity = 0.63). Increasing the number of samples increases sensitivity but reduces specificity. CONCLUSION: A minimum of 4 tissue cultures from representative areas is necessary to maximize the chance of identifying the infecting organism during management of the infected total hip and total knee arthroplasties. The designation of the best culture specimen for additional testing is arbitrary and may not be clinically efficacious.
Keywords:
best culture; culture negative; micro-organism identification; number of culture specimens; periprosthetic joint infection; total joint arthroplasty