A103-14 Antifungal Use With and Without Fungal Diagnoses Among Patients With Septic Shock

Abstract Rationale Septic shock from fungal infections has higher mortality and diagnostic complexity compared to bacterial pathogens. The role of empiric antifungal therapy in the undifferentiated patient with new septic shock is uncertain. Methods We developed an initial cohort across 5 Johns Hopkins Health System hospitals and validated findings in the Premier Health Database, which represents approximately one-quarter of all U.S. hospitalizations across 892 facilities. The cohort included all admitted adults with septic shock, defined by infection-related diagnostic codes with administration of a vasopressor in the medication administration record (MAR). Outcomes of interest were presence of a high-risk fungal infection—defined by presence of ICD-10 codes for a high-risk fungal infection—and receipt of intravenous antifungal therapy as recorded in the MAR. Results Among 23,232 Hopkins patients with septic shock, 1,082 (4.7%) had a high risk fungal diagnosis and 1,828 (7.9%) received an intravenous antifungal within one day of shock onset. Most antifungal recipients did not have a high-risk fungal diagnosis (86.6%, 1,583/1,828). Conversely, few (22.6%, 245/1,082) patients with a high-risk fungal diagnosis received antifungal therapy within one day of shock onset. Mean time to antifungal administration was 4.2 hours after initiation of vasopressors within the subgroup of patients with a high-risk fungal diagnosis who received antifungals within one day of shock onset. Findings were similar in the Premier validation cohort which included 544,948 patients with septic shock, of whom 11,405 (2.1%) had a high-risk fungal diagnosis and 18,652 (3.4%) received an intravenous antifungal within one day of shock onset. Again, most (87.6%, 16,333/18,652) antifungal recipients did not carry a high-risk fungal diagnosis and few (20.3%, 2,319/11,405) patients with a high-risk fungal diagnosis received an intravenous antifungal medication within one day of shock onset. In multivariable analysis for receipt of antifungals, patient-level factors such as additional vasopressor use (odds ratio OR=1.47 for each additional vasopressor, 95% confidence interval CI 1.45-1.50) were of the same magnitude as the variation in practice between hospitals (median odds ratio 1.58, 95%CI: 1.55-1.61; Figure). Conclusion In septic shock, most intravenous antifungals are administered to patients without a diagnosed high-risk fungal infection, while many patients with diagnosed high-risk fungal infections do not receive timely intravenous antifungals. Together, this suggests that tailored antifungal prescribing supported by precise risk prediction could improve antifungal use, safety, stewardship, and outcomes. This abstract is funded by: NHLBI, Hopkins Business of Health Initiative