Background Antimicrobial stewardship includes a de-escalation strategy for antimicrobial treatment. Treatment of urinary tract infections (UTI) offers a good opportunity to de-escalate antimicrobial treatment.

Purpose To evaluate the prescription of carbapenems for complicated UTIs in patients requiring hospitalisation and to measure the frequency with which de-escalation was performed.

Materials and methods A retrospective three-month study (March to June 2012) in a tertiary university hospital, of inpatients empirically treated with carbapenems for UTIs. The variables analysed were: number of patients prescribed carbapenems for UTI, requests for urinary cultures, analytical data such as leukocytes and C-reactive protein (CRP) at hospital admission, microorganisms isolated, duration of treatment, length of hospital stay and de-escalation performed. The Oncology-Haematology, Orthopaedics &Trauma, Neurosurgery and Neurology departments were excluded. De-escalation was defined as the replacement of the empirical carbapenem treatment by an antibiotic with a narrower spectrum.

Results A total of 433 prescriptions with carbapenems were recorded over these months, 52 (12%) for UTIs. The mean age was 73.5 ± 12.79 years and sex ratio: 60% Female. The most prescribed carbapenems were ertapenem (36.5%) and meropenem (30.8%). Out of a total of 27 requests for a urinary culture and sensitivity testing for carbapenem prescriptions for UTI, 17 (62.9%) were positive. The average length of treatment with carbapenems was 5 ± 2.28 days, 42% of carbapenem prescriptions were suspended due to decisions to change the antibiotic treatment (including de-escalation), 46% were discharged and 12% died. The most common pathogen isolated was Escherichia coli (70.5%), 75% of which were sensitive to amoxicillin-clavulanic acid and 96% to fosfomycin. Less common were Pseudomonas aeruginosa (8.3%) and Enterobacter cloacae (8.3%). Total treatments subject to de-escalation represented 66.6% (18/27).

Conclusions The rate of carbapenem prescriptions for UTI was low. Meropenem and ertapenem were the most prescribed. De-escalations was undertaken in a high proportion. Based on these data, more requests need to be made for microbiological tests.

No conflict of interest.