Aim:
To identify modifiable risk factors for developing bacteraemia caused by E. coli.
Methods:
This study encompassed three workpackages;
1. A matched case-control study using linked health data from the Secure Anonymised Information Linkage (SAIL) databank in Swansea University (2005 - 2011, n=227,871).
2. Whole genome sequencing of 720 E. coli bacteraemia (ECB) isolates from 2013-2014, in collaboration with the School of Biosciences, Cardiff University.
3. Enhanced surveillance of 496 consecutive ECB cases in Welsh residents in 2014, in collaboration with local infection prevention and control teams and regional health protection teams.
Results:
1. Linked health data analysis is feasible for public health investigation of ECB. Four variables were strongly associated with ECB: positive urine culture (previous 3 months), diagnosis likely to have led to hospital antibiotic prescription (previous 3 months), high comorbidity score, and age 85+ years.
2. Four lineages dominated (62% of isolates): MLST Sequence Types ST131, ST69, ST95, ST73. Variability in resistance gene carriage was observed; ST131 and ST69 isolates had increased carriage compared to ST73 and ST95. There was no evidence of phylo-geography, indicating the ECB increase is due to a global, population-level increase.
3. The majority of cases were community onset (75%), 50% likely of urinary source, 68% resistant to at least one antimicrobial class, 18% of cases died within 30 days.
Conclusions:
Interventions to reduce the ECB burden were subsequently developed, targeting UTI management and antimicrobial resistance: a UTI working group has developed key standards for UTI prevention, treatment and management (the ‘UTI 9’).