The Importance Of Multi-site Intra-operative Tissue Sampling In The Diagnosis Of Hip And Knee Periprosthetic Joint Infection - Results From A Single Centre Study

Introduction: The primary aim of this study was to determine whether the tissue type and anatomical location of intra-operative samples influences the accuracy of culture in the diagnosis of periprosthetic joint infection (PJI). The secondary aim was to create a predictive model of PJI using other known patient variables. Methods: A retrospective cohort of 3460 intra-operative samples from 887 patients was identified. The data was then analysed to compare intra-operative culture results (positive or negative) to the chosen gold standard of clinical diagnosis made by the treating team (infected or non-infected prosthetic joint). The intra-operative samples were grouped according to their labelling at the time of collection. Results: No single tissue type or anatomical location had both high sensitivity and specificity. The highest specificity for an anatomical location was hip bursa with 100%, for tissue type it was synovium with 93%. Sensitivity was highest in the anatomical locations for hip capsule (68%) and in the tissue types for pus (83%). Data analysis was performed to create a model for PJI and identified pre-operative predictors of PJI (increased white cell count, knee joint and non-revision surgery) which when used in combination with intra-operative culture results increased the sensitivity. Conclusion: Sample type and anatomical location influenced the reliability of the diagnosis of PJI however, no single sample type had higher diagnostic accuracy than samples combined thereby highlighting the necessity of obtaining multiple intra-operative samples in the diagnosis of PJI. The variation in predictive values of tissue types as well as improvement in sensitivity when combined with patient factors indicates that types of intra-operative sampling and the overall diagnostic pathway should vary depending on the individual case.