The social network of cystic fibrosis centre care and shared Pseudomonas aeruginosa strain infection: a cross-sectional analysis

Background Person-to-person transmission is a potential pathway of Pseudomonas aeruginosa acquisition in cystic fibrosis. Reports of cross-infection of shared cystic-fibrosis-specific P aeruginosa strains across large geographical distances are concerning. Therefore, we aimed to assess the extent to which patient movement between cystic fibrosis centres contributes to dissemination. Methods We did a cross-sectional study to assess movement of patients with cystic fibrosis who were infected with P aeruginosa between Sept 3, 2007, and June 16, 2010, at 18 Australian cystic fibrosis centres. We applied social network analysis to patient movement data from P aeruginosa-infected patients to assess the role of patient mobility in P aeruginosa genotype prevalence. We generated networks linking treatment centres based on the movement of patients attending adult and paediatric cystic fibrosis centres, and compared these with the movement of patients infected with all P aeruginosa strains, unique strains, and predominant Australian shared strains (AUST-01 and AUST-02). We summarised connectivity using degree centrality, in-degree centrality, out-degree centrality, and k-core estimates. Infection control and surveillance practices were also assessed by use of a questionnaire. Findings 983 patients (mean age 25 years [SD 10]; 551 [56%] male) provided 2887 P aeruginosa isolates for ERIC-PCR genotyping, which yielded 531 distinct genotypes: 493 unique strains in 373 patients and 38 shared strains in 610 patients. AUST-01 infections were associated with higher in-degree centrality (p=0 . 004) and k-core (p=0 . 005) estimates and AUST-02 infections with higher degree centrality (p=0 . 002), out-degree centrality (p=0 . 002), and k-core (p=0 . 007) estimates for the previous health-care facilities; associations for the present cystic fibrosis centre were not significant. These findings were significant for adult patients (AUST-01 in-degree centrality p=0 . 004 and k-core p=0 . 005; AUST-02 degree centrality p=0 . 004, out-degree centrality p=0 . 003, and k-core p=0 . 007), but not for paediatric patients. By contrast, infections with unique strains were associated with a lower k-core estimate for the present cystic fibrosis centre overall (p<0 . 0001); this finding was significant in adults (p<0 . 0001), but not in paediatric patients. Interpretation Our results show that the connectivity of cystic fibrosis centres, as measured by the movement of patients, seems to be an important risk factor for the acquisition of shared P aeruginosa strain infections. These results show the importance of prioritising infection control interventions (eg, prospective molecular surveillance for shared P aeruginosa strains, strict universal infection control precautions, and hospital design and ventilation) to limit P aeruginosa cross-infection between patients with cystic fibrosis.