'%3e%3cline%20id='Line_151'%20data-name='Line%20151'%20x2='29.522'%20transform='translate(22.5%2022.368)'%20fill='none'%20stroke='%23fff'%20stroke-linecap='round'%20stroke-width='4'/%3e%3cline%20id='Line_152'%20data-name='Line%20152'%20x2='29.522'%20transform='translate(22.5%2032.443)'%20fill='none'%20stroke='%23fff'%20stroke-linecap='round'%20stroke-width='4'/%3e%3cline%20id='Line_153'%20data-name='Line%20153'%20x2='29.522'%20transform='translate(22.5%2042.519)'%20fill='none'%20stroke='%23fff'%20stroke-linecap='round'%20stroke-width='4'/%3e%3c/g%3e%3c/svg%3e)
'%20fill='%232c3e50'%20fill-rule='evenodd'/%3e%3c/svg%3e)
'%20d='M14,8a6,6,0,1,0,6,6A6,6,0,0,0,14,8Zm0,10.839A4.839,4.839,0,1,1,18.839,14,4.837,4.837,0,0,1,14,18.839Zm1.5-2.526L13.441,14.82a.292.292,0,0,1-.119-.235V10.613a.291.291,0,0,1,.29-.29h.774a.291.291,0,0,1,.29.29v3.428l1.616,1.176a.29.29,0,0,1,.063.406l-.455.627A.292.292,0,0,1,15.5,16.313Z'%20transform='translate(-8%20-8)'%20fill='%232c3e50'/%3e%3c/svg%3e)
'%20d='M4.1,11.944C.642,6.929,0,6.414,0,4.571a4.571,4.571,0,0,1,9.142,0c0,1.843-.642,2.358-4.1,7.373a.572.572,0,0,1-.94,0Zm.47-5.468a1.9,1.9,0,1,0-1.9-1.9A1.9,1.9,0,0,0,4.571,6.476Z'%20transform='translate(0)'%20fill='%232c3e50'/%3e%3c/svg%3e)
Linezolid versus vancomycin as empirical therapy in external ventricular drain related infections in icu patients
Position du problème et objectif(s) de l’étude
Background: External ventricular drains (EVD) are invasive devices commonly used to reduce and monitor intracranial pressure. Primarily caused by Gram-positive bacteria, EVD-related infections are frequent in post-neurosurgical settings, with a significant impact on patient outcomes. Vancomycin (VAN) is the first-line empirical treatment; however, with more favorable pharmacokinetics, linezolid (LZD) has emerged as an alternative to glycopeptides. Objective: To compare LZD and VAN as empirical treatments for EVD-related infections
Matériel et méthodes
Method: This retrospective study was conducted in a French university hospital. ICU patients over 18 years old with EVD-related infections between 2017 and 2025 were considered for inclusion. Exclusion criteria were the isolation of Gram negative bacteria in cerebrospinal fluid (CSF) samples. All patients were treated in accordance with the 2017 IDSA guidelines. The primary outcome was the reintroduction of new antibiotics for EVD infection recurrence at day 15. Secondary outcomes included 1-month mortality rates, duration of EVD, and evolution of CSF biological parameters. A combined statistical approach was used. First, we used univariate analysis of baseline characteristics and patients outcomes. Then, to mitigate differences between groups and limit confounding bias, we used a propensity score (PS) approach with inverse probability of treatment weighting (IPTW). Variables included in the PS calculation were the patient age, SAPS II score, and time to surgical revision for EVD removal or replacement (i.e., the time interval between CSF samples).
Résultats & Discussion
Results: Fifty-nine patients with Gram-positive EVD-related infections were included, 14 (23.7%) received LZD and 45 (76.3%) received VAN as empirical treatments. Univariate analysis showed no significant differences in infection recurrence rates (LZD group : 0 (0%) and VAN group : 5 (11%), p=0.32), median EVD duration (LZD group : 32 (26-42) days and VAN group : 27 (19-36) days, p=0.36) and 1-month mortality rates (LZD group : 1 (7.1%) and VAN group : 3 (6.7%), p>0.99). The IPTW procedure improved covariate balance between groups. In the IPTW cohort, LZD empirical treatment resulted in lower infection recurrence rates (LZD group : 0 (0%) and VAN group : 6 (10%), p=0.035). Median EVD duration (LZD group : 29 (21-37) days and VAN group : 27 (18-35) days, p=0.37) and 1-month mortality rates (LZD group : 2 (3.4%) and VAN group : 4 (7.3%), p=0.51) remained comparable. In all analyses, changes in bacterial inoculum and CSF biological parameters between first and second samples did not differ between treatment groups. The incidence of drug-related adverse events (acute renal failure, cytopenia) was also similar.
Conclusion
Conclusion: Empirical linezolid treatment for EVD-related infections may be associated with reduced recurrence rates compared to glycopeptides. However, this finding is limited by statistical power, as significance was only observed in the IPTW pseudo-population with a larger sample size. Further multicenter studies with larger cohorts are needed to confirm these results and optimize empirical antibiotic strategies for EVD-related infections in the ICU.
Auteurs
Eddy BOUCHOU (1) , Pauline GAREL (1), Anne-Claire TOBENAS-DUJARDIN (2), Kevin ALEXANDRE (3), Thomas CLAVIER (4), Charles BAULIER (1) - (1)Rouen University Hospital, Department Of Anesthesiology, Critical Care And Perioperative Medicine, Rouen, France, (2)Rouen University Hospital, Department Of Neurosurgery, Rouen, France, (3)Univ Rouen Normandie, Université De Caen Normandie, Inserm, Normandie Univ, Rouen, France, (4)Rouen University Hospital, Department Of Anesthesiology, Critical Care And Perioperative Medicine; Umr 1096 Inserm– Envi, Ufr Santé, University Of Rouen, Rouen, France