Molecular Epidemiology of Candidemia: Evidence of Clusters of Smoldering Nosocomial Infections

• Published in: Clinical Infectious Diseases
• Main Authors: Rós Ásmundsdóttir, Lena, Erlendsdóttir, Helga, Haraldsson, Gunnsteinn, Guo, Hong, Xu, Jianping, Gottfredsson, Magnús
• Format: Text
• Language: English
• Published: Oxford University Press 2008
• Subjects: Major Articles; Iceland
• Online Access: http://cid.oxfordjournals.org/cgi/content/short/47/2/e17; https://doi.org/10.1086/589298

Background. Invasive fungal infections pose a serious threat to hospitalized patients worldwide. In particular, the prevalence of clusters of nosocomial infection among patients with candidemia remains unknown. The aim of this study was to investigate the molecular epidemiology of candidemia in a nationwide setting in Iceland during a 16-year period. Methods. The genotypes of all available fungal bloodstream isolates during 1991–2006 ( n = 219) were determined by polymerase chain reaction fingerprinting with use of 4 separate primers. Clusters were defined as isolation of ⩾2 strains with genotypes that had ⩾90% relatedness in the same hospital within a period of 90 days. Results. Candida albicans represented 61.6% of isolates, followed by Candida glabrata (13.7%), Candida tropicalis (9.1%), and Candida parapsilosis (8.7%). Polymerase chain reaction fingerprinting revealed 35 clones of C. albicans , 10 clones of C. glabrata , 7 clones of C. tropicalis , 4 clones of C. parapsilosis , and 5 clones of Candida dubliniensis . Overall, 18.7%–39.9% of all infections were part of nosocomial clusters, most commonly caused by C. albicans , C. parapsilosis , and C. tropicalis . Most clusters involved 2 cases and disproportionately affected patients in adult and neonatal intensive care units ( P = .045). The 7-day (16%) and 30-day (32%) case-fatality rates among cluster-associated cases did not differ statistically significantly from those for sporadic nosocomial infections. None of the clusters were identified by the hospital surveillance team. Conclusions. In an unselected patient population, as many as one-third of all cases of candidemia may be attributable to nosocomial clusters. The risk is dependent on hospital wards and patient populations; it is highest in intensive care units. Small clusters are not identified by routine hospital surveillance.