OBJECTIVEWe analyse the effectiveness of short courses of adequate treatment in patients with episodes of carbapenemase-producing Enterobacterales bloodstream-infections (CPE-BSI).METHODSPatients with first monomicrobial CPE-BSI episodes who received ≥72 h of appropriate treatment from 2014-2022 were selected. Detection of CPE was established on the basis of phenotypic antibiogram and confirmation by PCR and/or immunochromatographic methods. Patients were classified in short treatment group (STG) those who received 3-10 days of appropriate treatment, and long treatment (LTG) those receiving >10 days. Unfavourable outcome consisted in a composite of global 30-day mortality and/or persistent bacteremia and/or recurrent bacteremia. Inverse probability of treatment weighting (IPTW) analysis was performed to compare the outcome between the two study groups.RESULTSWe included 105 CPE-BSI episodes: 99 were caused by OXA-48-like, 4 VIM and 2 KPC carbapenemases. Thirty-nine patients (37.1%) were included in the STG and 66 (62.9%) in LTG. The STG group presented frequent treatment with ceftazidime-avibactam (43.6% vs. 24.2%, P = 0.03) and lower in-hospital stay (21 days vs. 32 days, P = 0.02). Overall, 28 patients (26.7%) presented unfavourable outcome: IPTW analysis showed no differences in the outcome between STG to LTG groups (24.2% vs. 30.8%, weighted-risk difference 6.6%, P = 0.44). Patients with unfavourable outcome presented more frequently source other than urinary-biliary (46.4% vs. 23.4%, P = 0.02), received less frequently ceftazidime-avibactam (14.3% vs. 37.7%, P = 0.02) and presented frequently with absence of source control when indicated (28.6% vs. 13.0%, P = 0.06).CONCLUSIONSShort treatment durations for CPE-BSI episodes may be effective, as long as they are appropriate and source control is performed.