A/Prof. Ling Li1, Dr Jannah Baker1, Dr Amith Shetty1,2
1Macquarie University
2NSW Ministry of Health
Introduction: Sepsis is a medical emergency associated with extensive mortality worldwide. Frequently, no causative pathogen is detected in blood cultures(BC) or urine cultures(UC) taken from sepsis patients. This paper aimed to compare clinical outcomes of suspected sepsis patients by BC and UC positivity.
Methods: This retrospective study included patients(≥16 years) with suspected sepsis presenting to the emergency department (ED) at four hospitals between 2017-2019. Primary outcomes were in-hospital mortality and intensive care unit(ICU) admission. Four culture groups were defined based on the positivity of BC and UC ordered within 24 hours of triage. Multilevel modelling was applied, adjusting for patient characteristics and clinical factors.
Results: Of 8108 suspected sepsis ED presentations without antibiotics given prior to culture orders, 6255(77%) were culture negative (BC-UC-); 1239(15%) BC- UC+; 431(5%) BC+UC-; and 183(2%) BC+UC+. Patients with bacteraemia not associated with urinary tract infection (BC+ UC-) had a higher risk of death (adjusted odds ratio [AOR] 1.76) and 7-day re-presentation (AOR 1.64) relative to BC- and UC+ patients. Bacteraemic (BC+) patients were nearly three times more likely to be admitted to ICU regardless of UC results (AOR 2.90 for UC-; 2.93 for UC+) and had longer hospital stays (adjusted relative risk 1.36 for UC-;1.48 for UC+).
Conclusion: While bacteraemia was associated with worse patient outcomes, those with urinary tract infection had less severe outcomes among bacteraemic patients with suspected sepsis. Clinicians should ensure patients presenting with bacteraemic sepsis and negative urine cultures are closely monitored to tailor treatment accordingly for improved patient outcomes.
Biography: A/Prof Ling Li leads the health analytics research at Centre for Health Systems and Safety Research, Australian Institute of Health Innovation, Macquarie University. She has extensive experience in evaluating the impacts of health interventions with a strong record in study design, complex health data linkage and statistical modelling. She is nationally recognised for her research in sepsis, including evaluation and optimisation of sepsis early detection tools, and publication of the first nationwide sepsis epidemiology report. Her research has contributed to a number of significant changes into policy and practice.