Cargando…

2408. Genotypic Corroboration of Epidemiologically Linked Clusters to Detect Outbreaks of C. difficile at a Tertiary Care Hospital

BACKGROUND: The Society for Healthcare Epidemiology of America (SHEA) recommends that surveillance for healthcare facility-onset C. difficile infections (HO CDI) be conducted to detect elevated rates or outbreaks of CDI and stratify data by hospital unit when possible to facilitate detection of clus...

Descripción completa

Detalles Bibliográficos
Autores principales: Bayoumi, Nagla, McMillen, Tracy, Aslam, Anoshé, Kamboj, Mini
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Oxford University Press 2019
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6810324/
http://dx.doi.org/10.1093/ofid/ofz360.2086
Descripción
Sumario:BACKGROUND: The Society for Healthcare Epidemiology of America (SHEA) recommends that surveillance for healthcare facility-onset C. difficile infections (HO CDI) be conducted to detect elevated rates or outbreaks of CDI and stratify data by hospital unit when possible to facilitate detection of clusters. At Memorical Sloan Kettering Cancer Center, strain typing of isolates using multi-locus sequence typing (MLST) is performed routinely and in real time to inform control efforts. Genotyping can conclusively establish or debunk transmission events based on routine surveillance. Management of C. difficile outbreaks is time and resource intensive. METHODS: A retrospective analysis was conducted to identify all nosocomial C. difficile cases between July 2013 and July 2018. Based on Memorial Sloan Kettering’s baseline surveillance data, a cluster of C. difficile was defined as three or more hospital-acquired cases (as defined by NHSN) on the same inpatient unit within a 7-day period. Data were analyzed to quantify the number of clusters observed and determine genetic relatedness among cases to detect an outbreak. RESULTS: A total of 1,116 HO CDI cases occurred during the 5-year time period. Annual nosocomial rates of CDI remained stable (P = 0.052). Eighty clusters were identified; 63 clusters had 3 cases within each cluster, 16 were each made up of 4 cases, and 1 cluster consisted of 5 cases. Two clusters had strain typing concordance amongst all 3 cases; strain type 42 and strain type 1. Among all the epidemiologically linked clusters over the 5-year period, only 2.5% were genetically linked suggestive of true outbreaks. CONCLUSION: The majority of HO-CDI clusters detected on clinical surveillance are non-clonal. Genotyping should be routinely used to corroborate clusters identified on microbiological surveillance before costly outbreak control interventions are deployed. [Image: see text] DISCLOSURES: All authors: No reported disclosures.