Extended-spectrum beta-lactamase-producing enterobacteriaceae : epidemiology and dynamics of fecal carriage

Abstract

ESBL-producing Enterobacteriaceae (EPE) has become a major cause of community acquired urinary tract infection (UTI), and fecal carriage of EPE is emerging worldwide. The aims of this thesis were to study the molecular epidemiology of ESBL-enzymes in Stockholm (I) to evaluate treatment alternatives to the carbapenems for EPE (I-III), and to study the duration of fecal carriage and identify factors associated with prolonged carriage (IV).

Paper I describes a consecutive collection of EPE (n=169). The distribution of ESBL-enzymes and clonal relatedness of the isolates was determined with PCR, DNA sequencing and pulsed-field gel electrophoresis (PFGE). Antimicrobial activity was evaluated using gradient test, broth microdilution and disk diffusion, and the susceptibility test methods were compared for parenteral !-lactams. We found that CTX-M-15 (75%) and CTX-M-14 (23%) were the dominating genotypes, that the collection was largely polyclonal and that 41% of E. coli belonged to the international clone sequence type (ST) 131. We concluded that there are several oral (mecillinam, nitrofurantoin, fosfomycin) and parenteral (piperacillin-tazobactam, tigecycline, temocillin) treatment alternatives for E. coli but few for K. pneumoniae. We also showed that susceptibility rates obtained with Etest and disk diffusion (DD) were not in agreement with the reference method broth microdilution for piperacillin- tazobactam (TZP). Etest and DD are therefore not reliable to detect resistance to TZP, with the breakpoints used at the time of the study.

In paper II the novel cephalosporin CXA-101 (later known as ceftolozane) in combination with tazobactam (CXA-201) was evaluated against the same collection of isolates as in paper I, and compared to other !-lactam/!-lactamase inhibitors. MICs were determined with broth microdilution and susceptibility to CXA-201 was 88-98%, depending on the concentration of tazobactam and the tentative breakpoint used. All ACL-resistant and 94% of the TZP-resistant isolates were CXA-201 susceptible. We concluded that ceftolozane-tazobactam (CXA-201) is a potential future therapeutic option against EPE, especially for TZP-resistant isolates.

Paper III evaluates the clinical and bacteriological activity of pivmecillinam for patients treated for lower UTI caused by an EPE (n=8). The clinical cure (resolved UTI symptoms after completed treatment) was high (8/8) but bacteriological cure (< 103 CFU/ml at follow-up after 30 days) was low (2/8), although none of the patients with persisting bacteriuria relapsed within 6 months.

In paper IV we studied the duration and dynamics of ESBL-carriage. A cohort of patients (n=61) were followed with fecal samples and questionnaires about antimicrobial treatment and risk factors for EPE, 1, 3, 6 and 12 months after EPE infection. EPE strains were subjected to PFGE, PCR for phylogrouping, detection of CTX-M phylogroup, pabB (ST131) and virulence genes and PCR based replicon typing. Patient and strain related variables were compared for carriers and non-carriers at 12 months. We concluded that EPE carriage is common 12 months after infection (43%) and that persisting carriage may be associated with E. coli phylogroup B2 and CTX-M-9. The strain background frequently changes throughout the carriage and negative samples do not imply eliminated carriage.

This knowledge will hopefully contribute to providing better medical care of patients with infection caused by EPE. It may also prove important for defining patients in require of prolonged isolation in single rooms or cohorts. Thereby the spread of EPE in hospitals and long term care facilities can be limited.