Abstract
Aim
An instrumented blood culture system automatically flags when growth within the culture medium has been detected (‘work in progress’), and subsequently when the organism has been identified.
We explore using this data to switch patients to oral therapy within 72 hours post-surgery, reducing costs and improving antimicrobial stewardship.
Method
This retrospective review focused on clinically significant culture-positive bone and joint infections over a 5-month period in 2022. Two cohorts were defined as either having positive intraoperative microbiology at <72 hours or at ≥72 hours.
Results
150 patients were included. 133/150(88%) exhibited microbial growth <72hours. Of these, 98/133(74%) had all organisms identified <72-hours, and 34/133(26%) had additional organisms ≥72 hours. 19/151(12%) patients had their first positive cultures ≥72hrs from sampling.
The most common isolates identified within 72 hours were S. aureus(30%), Enterobacteriaceae (26%), and Coagulase-negative Staphylococcus (CoNS)(19%). If no growth was observed by 48 hours, there was a 69.6% probability that subsequent growth wouldn't occur; this probability increases to 81.9% by 72 hours, 88.7% by 96 hours, 91.0% by 120 hours, and 95.0% by 144 hours (see figure 1). The most common isolates identified ≥72 hours were CoNS(28%), Cutibacterium acnes(16%) and S. aureus(12%). Assessing oral antibiotic regimes for isolates identified after 72 hours demonstrated that linezolid would cover isolates from 96% of patients, tetracyclines 92% of patients, clindamycin 85% of patients, and ciprofloxacin and rifampicin would cover 80% of patients. Vancomycin and meropenem, our standard empirical therapy, gave the best cover at 96% of patients.
Conclusions
This study suggests there is sufficient microbiological information at 72 hours for most patients to allow transition to a targeted regimen. If there has been no detection of growth when using an instrumented blood culture system by 72 hours, it is likely that there will be no growth.
For any tables or figures, please contact the authors directly.
