LAUSR

Background Empirical antimicrobial therapy (EAT) is a challenge for community-acquired, hospital-acquired and ventilator-associated pneumonia, particularly in the context of the increasing occurrence of third-generation cephalosporin-resistant Enterobacterales (3GCR-E), including extended-spectrum beta-lactamase Enterobacterales (ESBL-E) and high-level expressed AmpC cephalosporinase-producing Enterobacterales (HLAC-E). To prevent the overuse of broad-spectrum antimicrobial therapies, such as carbapenems, we assessed the performance of screening for intestinal carriage of HLAC-E in addition to ESBL-E to predict 3GCR-E (ESBL-E and/or HLAC-E) presence or absence in respiratory samples in ICU, and to evaluate its potential impact on carbapenem prescription. Materials and methods This monocentric retrospective observational study was performed in a surgical ICU during a 4-year period (January 2013–December 2016). Patients were included if they had a positive culture on a respiratory sample and a previous intestinal carriage screening performed by rectal swabbing within 21 days. Sensitivity, specificity, positive (PPV) and negative (NPV) predictive values and likelihood ratios were calculated for the screening for intestinal carriage of ESBL-E, HLAC-E and 3GCR-E (ESBL-E and/or HLAC-E) as predictor of their absence/presence in respiratory samples. Impact of HLAC-E and ESBL-E reporting on EAT was also studied. Results 765 respiratory samples, retrieved from 468 patients, were analyzed. ESBL-E prevalence was 23.8% in rectal swab and 4.4% in respiratory samples. HLAC-E prevalence was 9.0% in rectal swabs and 3.7% in respiratory samples. Overall, the 3GCR-E prevalence was 31.8% in rectal swabs and 7.7% in respiratory samples. NPVs were 98.8%, 98.0% and 96.6% for ESBL-E, HLAC-E and 3GCR-E, respectively. Over the study period, empirical antimicrobial therapy was initiated for 315 episodes of respiratory infections: 228/315 (72.4%) were associated with negative intestinal carriage screening for both HLAC-E and ESBL-E, of whom 28/228 (12.3%) were treated with carbapenems. Of 23/315 (7.3%) cases with screening for positive intestinal carriage with HLAC-E alone, 10/23 (43.5%) were treated with carbapenems. Conclusion Systematic screening and reporting of HLAC-E in addition to ESBL-E in intestinal carriage screening could help to predict the absence of 3GCR-E in respiratory samples of severe surgical ICU patients. This could improve the appropriateness of EAT in ICU patients with HAP and may prevent the overuse of carbapenems.