181 Clinical sub-phenotypes of Staphylococcus aureus bacteraemia 8 The five classes identified by LCA in the Edinburgh and ARREST cohorts represented distinct clinical sub-phenotypes when considering their association with class-defining clinical variables, and were replicated in the two analyses (Figure 1). Sub-phenotype A was associated with older age, co-morbidity, and SAB from unknown or SSTI source. Sub-phenotype B was associated with nosocomial SAB, bacteraemia originating from an intravenous catheter, younger age, less co-morbidity, and lack of any metastatic foci. Sub-phenotype C was associated with community-acquired SAB from unknown source, with higher CRP, and with the presence of metastatic foci of infection. Subphenotype D was associated with chronic kidney disease, intravenous catheter source, and nosocomial or healthcare associated acquisition. In the Edinburgh cohort, 17/39 predicted members of this sub-phenotype received haemodialysis. Sub-phenotype E was associated with community-acquired SAB, younger age, IDU, liver disease, and with endocarditis. In the Edinburgh cohort, the source of SAB in 32/37 predicted members of this sub-phenotype was IDU (categorised as ‘other’ source in the LCA since this category did not exist in the classification used in ARREST). In the ARREST cohort, SSTI is the source enriched in sub-phenotype E consistent with acquisition through IDU. In the Edinburgh cohort, 14/37 predicted members of this sub-phenotype had infected deep vein thrombophlebitis and 3/37 had an infected pseudoaneurysm. A B C D E Age Male Dementia Vascular disease Cardiac prosthetic material Chronic kidney disease Liver disease Person who injects drugs Heart rate Temperature Haemoglobin Creatinine C-reactive protein MRSA Endocarditis Other metastatic foci -2 0 2 z-score: A B C D E Age Dementia Chronic kidney disease Liver disease Person who injects drugs Heart rate Temperature Haemoglobin Creatinine C-reactive protein MRSA Endocarditis Other metastatic foci 0 150 N patients 0 300 N patients Community Healthcare Nosocomial Unknown IV catheter SSTI Other Respiratory Urine Community Healthcare Nosocomial Unknown IV catheter SSTI Other Respiratory Urine Edinburgh cohort ARREST cohort A B C D Age Male Dementia Vascular disease Cardiac prosthetic material Chronic kidney disease Liver disease Temperature Creatinine Endocarditis Other metastatic foci Community Healthcare Nosocomial Unknown IV catheter SSTI Other Respiratory Urine Acquisition* Source of SAB* 0 150 N patients SAFO cohort Acquisition* Source of SAB* Acquisition* Source of SAB* Figure 1: Comparison of class-defining variables between SAB sub-phenotypesVertical bars show the number of patients assigned to each sub-phenotype. Cells are shaded according to row z-score (i.e. comparing subphenotypes) except ‘Acquisition’ and ‘Source of SAB’, where shading is by column z-score (i.e. comparing within each sub-phenotype). Intensity of red shading reflects a more positive z-score (i.e. above the mean) and intensity of blue shading reflects a more negative z-score (i.e. below the mean). In the SAFO trial, people with Child Pugh C liver cirrhosis, MRSA infection, and people who inject drugs were not recruited. IV: intravenous; SSTI: skin or soft tissue infection; MRSA: methicillin-resistant S. aureus.
RkJQdWJsaXNoZXIy MTk4NDMw