Our results document the spread of the IMP13-producing clone in a small private surgical clinic, rather than in an intensive care unit of a university hospital as previously described for this clone.
For two of the four isolates, the MIC of carbapenem drugs was in the susceptibility range. Consequently, the detection of the cluster of UTIs was delayed for several weeks. Our findings confirm the previously described heterogeneous carbapenem-resistance phenotype of the IMP13-producing clone
 and suggest that the spread of the clone may be, at least partially, facilitated by difficulties in its detection.
Despite extensive investigations prompted by this cluster of UTIs associated with the IMP-13 P. aeruginosa, no evidence was found for the involvement of the water supplies or contaminated medical or surgical devices. It is possible that the delay between the time of contamination of the patients and our investigations hindered the identification of a contaminated source. However, in all but one case, the time separating surgery and invasive acts from the onset of clinical signs of UTI was long. These observations do not argue for contamination of the patients during surgery or invasive acts.
By contrast, the temporal superposition of hospitalisation for the infected patients, and the observation of insufficiently strict application of hygiene practices in the urology unit and before cystography suggest that the UTI cases may have resulted from cross-transmission during post-surgery hospitalisation. The rate of faecal carriage of P. aeruginosa among elderly urology patients is frequently high, confirming the ability of this bacterium to colonise the urinary tract of such patients
. In view of the frequent and invasive nature of urinary care acts following surgery, and the well-described epidemic potential of the IMP-13 clone
[1, 3, 11], we suggest that elderly urology patients should be considered to be at high risk of patient-to-patient cross-transmission of IMP-13-producing P. aeruginosa.
Concordant with MLST results, the recent sequence analysis of the blaIMP13 gene in numerous IMP-13-P. aeruginosa isolates from distant French regions showed that our epidemic isolates and the Italian clone have very similar characteristics
. Nevertheless, no link with countries experiencing IMP-13-P. aeruginosa outbreaks was identified in the index case history, and the origin of the IMP-13 clone involved in this outbreak remains unclear. Further studies are needed to improve our knowledge of the epidemiology of this highly resistant clone.
Because of the association between the use of broad-spectrum antibiotics and multidrug resistance, a campaign has been run to promote systematic microbiological documentation of UTI: this should favour the use of specific rather than broad-spectrum antibiotics. The early implementation of barrier precautions around infected patients successfully prevented the further spread of the carbapenem-resistant clone in the clinic. We suggest that, when facing a limited cluster of infections, such precautions may be sufficient for infection control, without the need for more extreme measures, such as cohorting, for example.