Gastrointestinal carriage is a major reservoir of K. pneumoniae infection in intensive care patients

Abstract BackgroundKlebsiella pneumoniae is an opportunistic pathogen and a leading cause of hospital-associated (HA) infections. Patients in intensive care units (ICUs) are particularly at risk, and outbreaks are frequently reported in ICUs. K. pneumoniae is also part of the healthy human microbiome, providing a potential reservoir for HA infection. However, the frequency of K. pneumoniae gut colonization and its contribution to HA infections are not well characterized. MethodsWe conducted one-year prospective cohort study of ICU patients. Participants (n=498) were screened for rectal and throat carriage of K. pneumoniae shortly after admission, and clinical information was extracted from hospital records.K. pneumoniae isolated from screening swabs and clinical diagnostic samples were characterized using whole genome sequencing. Genomic and epidemiological data were combined to identify likely transmission events. Results and ConclusionsK. pneumoniae carriage frequencies were estimated at 6% (95% CI, 3%-8%) amongst ICU patients admitted direct from the community, and 19% (95% CI, 14% – 51%) amongst those who had recent contact with healthcare. Gut colonisation on admission was significantly associated with subsequent K. pneumoniae infection (infection risk 16% vs 3%, OR=6.9, p<0.001), and genome data indicated a match between carriage and infection isolates in most patients. Five likely transmission chains were identified, resulting in six infections (12% of K. pneumoniae infections in ICU). In contrast, 49% of K. pneumoniae infections were caused by a strain that was unique to the patient, and 48% of patients with K. pneumoniae infections who participated in screening were positive for prior colonisation. These data confirm K. pneumoniae colonisation is a significant risk factor for subsequent infection in ICU, and indicate that half of all K. pneumoniae infections result from patients’ own microbiota. Screening for colonisation on admission could limit risk of infection in the colonised patient and others.