ICU Acinetobacter in Romania: High Carbapenem Non-Susceptibility, Delayed Active Therapy

A six-year, single-center intensive care unit cohort from Romania describes an endemic Acinetobacter ecology marked by predominance of Acinetobacter baumannii, extensive carbapenem non-susceptibility, and frequent delays before patients received an in vitro active agent despite early empiric antibiotic starts.

In the authors’ reporting, these treatment-timing observations coincided with high short-term all-cause mortality in the ICU population studied. The overall picture presented in the ICU Acinetobacter cohort is of a high-resistance setting in which prompt initiation of antibiotics did not consistently translate into early microbiologically active coverage.

The investigators included 234 adult ICU episodes with a first clinical isolate of Acinetobacter spp. collected from January 2019 through December 2024. Across the cohort, A. baumannii represented 87.6% of isolates (205/234), and carbapenem non-susceptibility among Acinetobacter spp. was reported in 89.3% (209/234). Clinical samples were predominantly endotracheal aspirates (68.8%), followed by wound or other purulent secretions (17.5%), blood cultures (10.3%), and urine cultures (3.4%). The authors frame these organism and specimen patterns as the backdrop for very limited carbapenem activity.

Empiric systemic antibiotics were initiated within 24 hours of culture collection in 95.7% of episodes (224/234), yet 49.6% (116/234) received at least one empiric agent later confirmed to be in vitro active, according to the authors. In analyses centered on time to active therapy (TTAT), the median TTAT from index culture collection to first active agent was 6 days (IQR 4–7), and active therapy within 72 hours occurred in 8.5% (20/234). When stratified by EUCAST carbapenem category, TTAT was reported as shorter in EUCAST R episodes than in S/I episodes (median 5 vs 6 days), while early active coverage remained uncommon across categories. The authors describe an ongoing gap between early initiation and subsequent confirmation of activity against the recovered isolate.

Thirty-day all-cause mortality was 73.1% (171/234) in the overall cohort. In the authors’ comparison restricted to A. baumannii, mortality did not differ between carbapenem non-susceptible (EUCAST I+R) and carbapenem-susceptible (EUCAST S) episodes (73.2% vs 72.0%; p = 1.00). They also report that survivors had longer hospital stays and accumulated more ventilator days than non-survivors, and that age was higher among those who died, while selected severity scores did not differ significantly in the unadjusted comparisons presented. As reported, mortality remained high across resistance strata in this ICU cohort.

In multivariable analysis, increasing age was independently associated with 30-day mortality (OR 1.36 per 10-year increase, 95% CI 1.04–1.90), and the authors report acceptable model discrimination (AUC 0.74). For early inflammatory biomarkers, neutrophil-to-lymphocyte ratio and C-reactive protein during the first 72 hours did not differ between CRAB and non-CRAB episodes, which the authors interpret as limited discriminatory value for early phenotype differentiation in this setting.

In discussion and conclusions, the authors state that the findings align with empiric strategies tailored to unit epidemiology and emphasize improved diagnostics and stewardship interventions in high-burden ICUs. They conclude that high resistance and delayed receipt of in vitro active therapy were common in the cohort they analyzed.

Key Takeaways:

• The cohort was dominated by A. baumannii, and carbapenem non-susceptibility was reported as a pervasive phenotypic pattern across Acinetobacter isolates.
• Empiric antibiotics were started early in most episodes, but investigators report that confirmed in vitro activity was less frequent and TTAT was often prolonged.
• Thirty-day mortality was high overall; age was independently associated with mortality in the adjusted model, and early NLR/CRP did not distinguish resistant from susceptible episodes as reported.