Getting it right first time: The importance of a structured tissue sampling protocol for diagnosing fracture-related infections

Abstract

Introduction: Fracture-related infection (FRI) is an important complication following surgical fracture management. Key to successful treatment is an accurate diagnosis. To this end, microbiological identification remains the gold standard. Although a structured approach towards sampling specimens for microbiology seems logical, there is no consensus on a culture protocol for FRI. The aim of this study is to evaluate the effect of a structured microbiology sampling protocol for fracture-related infections compared to ad-hoc culture sampling. Methods: We conducted a pre-/post-implementation cohort study that compared the effects of implementation of a structured FRI sampling protocol. The protocol included strict criteria for sampling and interpretation of tissue cultures for microbiology. All intraoperative samples from suspected or confirmed FRI were compared for culture results. Adherence to the protocol was described for the post-implementation cohort. Results: In total 101 patients were included, 49 pre-implementation and 52 post-implementation. From these patients 175 intraoperative culture sets were obtained, 96 and 79 pre- and post-implementation respectively. Cultures from the pre-implementation cohort showed significantly more antibiotic use during culture sampling (P = 0.002). The post-implementation cohort showed a tendency more positive culture sets (69% vs. 63%), with a significant difference in open wounds (86% vs. 67%, P = 0.034). In all post-implementation culture sets causative pathogens were cultured more than once per set, in contrast to pre-implementation. Despite stricter tissue sampling and culture interpretation criteria, the number of polymicrobial infections was similar in both cohorts, approximately 29% of all culture sets and 44% of all positive culture sets. Significantly more polymicrobial cultures were found in early infections in the post-implementation cohort (P = 0.048). This indicates a better yield in the new protocol. Conclusion: A standardised protocol for intraoperative sampling for bacterial identification in FRI is superior than an ad-hoc approach. It has a positive effect on both surgeon and microbiologist by increasing awareness about the problem at hand. This resulted in more microbiologically confirmed infections and more certainty when identifying causative pathogens.