Abstract
BACKGROUND: Despite reduced infectious disease mortality and improved survival, infectious diseases continue to pose health threats due to their contagiousness, societal harm, and morbidity. Empiric antibiotic therapy, often prescribed without knowledge of the causative pathogen, faces challenges from rising antibiotic resistance. This study explores the potential of prior positive culture results to guide empiric antibiotic therapy. METHODS: Data from King Abdullah University Hospital (Jan 2014-Dec 2019) included adult patients with recurrent bacterial infections (pneumonia, sepsis, UTIs, wounds). Excluded cases included: mixed infections, transfers, <14 days or >12 months between episodes. The study compared bacterial growth and sensitivity patterns between previous and recent cultures. RESULTS: The study included 970 episodes from 650 patients, mainly UTIs (60.3%) and gram-negative bacteria (77.9%). The study found that (65.1%) of culture pairs matched. Empirical therapy was accurate in (71.8%) of cases. Further, accuracy of selected empiric antibiotic therapy was significantly predicted (p < 0.001) by: type of infection, type of antibiotics, and concordance with prior microbiologic data. Multivariate logistic analysis showed blood culture as less predictive of pending identity (OR: 0.234, P < 0.001) compared to urine culture; and prior affirmed gram negative bacterial culture was less predictive (OR: 0.606, P = 0.021) compared to gram positive bacterial culture. CONCLUSION: This study underscores the potential of prior positive culture results in guiding empiric antibiotic therapy, enhancing accuracy and identity agreement. Future research should explore this approach in different infection contexts and across multiple centers. Reducing the indiscriminate use of broad-spectrum antibiotics is essential to combat antibiotic resistance.