Carbapenem use in French hospitals: A nationwide survey at the patient level.

The objective of this study was to evaluate the characteristics of carbapenem use in French healthcare settings in order to guide future actions. Healthcare facilities voluntarily participated in a nationwide cross-sectional survey in 2011. Medical data and reasons for carbapenem treatment (CPR) and discontinuation were recorded for all patients treated with carbapenems. A total of 2338 patients were recorded by 207 facilities. The median duration of CPR was 8 days, and 31.4% of patients received CPR for >10 days. An antibiotic consultant was involved in the initial choice of CPR in 36.8% of cases. CPR was chosen on an empirical (EP) basis for 1229 patients (52.6%), mainly because of severe sepsis (48.6%) or a perceived risk of bacterial resistance (33.7%). Among EP patients, de-escalation was more frequent in the case of intervention of an antibiotic consultant (35.1%) than without intervention (22.9%) (P<0.01). Among the 1109 patients receiving CPR initially based on bacteriological results, 607 (54.7%) had ESBL-producing Enterobacteriaceae and 397 (35.8%) had Gram-negative bacilli susceptible to at least one β-lactam other than carbapenems or to fluoroquinolones. Among the latter, de-escalation was performed in 59 cases (14.9%). The intervention of an antibiotic consultant did not favour de-escalation in this group. In conclusion, carbapenems are frequently used for treating suspected or confirmed multidrug-resistant bacteria, and overall CPR duration is long. De-escalation is frequently not implemented despite isolates being susceptible to other drugs. More frequent antibiotic consultant intervention may help to decrease carbapenem use in the case of EP treatment.