Sofía de la Villa, Nuria Fernández‐Hidalgo, Francesc Escrihuela-Vidal, Rosa Escudero-Sánchez, Itxasne Cabezón, Lucía Boix-Palop, Beatriz Díaz‐Pollán, Ane Josune Goikoetxea, María José García-Pais, Lucía Ramos‐Merino, María Teresa Pérez-Rodríguez, Ángela Crespo, Laura Rio, José M. Bellón, Patricia Muñóz, MRSA-GEIRAS-SEIMC study group, Dàmaris Berbel, Luís Buzón, David Campany, Álex García Tellado, Inmaculada Grau, José Manuel Guerra-Laso, Joan Roig-Sanchis, Celia Sánchez-Martínez, O. Sanz, Fiorana Silvante, Belén Viñado, Luciana Urbina, Ana Verónica Halperín, Mariona Xercavins
Abstract Background We aimed to identify and evaluate clinical subphenotypes in a cohort of patients with methicillin-resistant Staphylococcus aureus bacteremia (MRSAB) and to assess their association with all-cause 90-day mortality. Methods This post hoc analysis of the MRSA-GEIRAS-SEIMC study was conducted across 15 Spanish hospitals. MRSAB in adult patients from 2019 to 2022 were included. Clinical subphenotypes were identified using a combination of principal component analysis and latent class analysis based on age, sex, comorbidities, SOFA score, creatinine levels, metastatic foci, source, and acquisition. The 90-day mortality associated with each subphenotype was estimated using the Kaplan–Meier method. Cox regression was performed to assess the risk of death. Results A total of 419 MRSAB were included. Four distinct subphenotypes were identified: S1 was associated with younger age, community acquisition, and unknown or skin and soft-tissue infection source; S2 was associated with older age, female sex, high burden of comorbidities, and healthcare-related acquisition; S3 was linked to a catheter source and nosocomial acquisition; and S4 was predominantly associated with the presence of heart valve prostheses, and metastatic foci. Significant differences in all-cause 90-day mortality were observed across subphenotypes: 20.0% in S1, 47.4% in S2, 26.2% in S3, and 35.1% in S4 (P < .01). Cox regression indicated an increased 90-day mortality risk in S2 (HR, 2.98; 95% CI, 1.59–5.56) and S4 (HR, 1.99; 95% CI, 1.16–3.42) compared with S1. Conclusions We identified 4 distinct clinical subphenotypes of MRSAB associated with prognostic outcomes. Further investigation is needed to implement them into clinical practice.