Spectrum, dose, and duration of antibiotic exposure and risk of intensive care unit–acquired carbapenem-resistant gram-negative bacteria: a prospective cohort study

Background While antibiotic exposure is a known key risk factor for acquiring Carbapenem-resistant Gram-negative bacteria (CR-GNB) in the ICU, the independent contributions and relative importance of its core dimensions—spectrum, dose, and duration—remain poorly understood. This study aimed to clarify these specific relationships to inform the optimization of antibiotic stewardship strategies. Methods We prospectively enrolled consecutive adult patients admitted to 4 ICUs at a university hospital between March 2024 and January 2025. Patients were screened for CR-GNB upon admission and weekly. Antibiotic exposure was quantified by spectrum (Antibiotic Spectrum Index per antibiotic day [ASI]), dose (Defined Daily Doses [DDDs]), and duration (Length of Therapy [LOT]). The primary outcome was ICU-acquired CR-GNB. We used interval-censored Cox regression to assess associations. Restricted cubic splines were used to explore potential non-linear relationships, and relative importance analysis was performed to compare the impact of the exposure metrics. Results Overall, 151 of 422 patients (35.8%) acquired CR-GNB during their ICU stay, with a median follow-up of 12.0 days (interquartile range, 8.0–17.0). ASI per antibiotic day was independently associated with an increased risk of ICU-acquired CR-GNB (adjusted Hazard Ratio [aHR] per 1-unit increase, 1.14; 95% Confidence Interval [CI] 1.09–1.19; P < 0.001), exhibiting a non-linear J-shaped relationship ( P for nonlinearity = 0.027). In contrast, after full adjustment, DDDs were not significantly associated with CR-GNB acquisition (aHR per 1-unit increase, 0.89; 95% CI 0.69–1.15; P = 0.374), despite displaying a non-linear inverted U-shaped relationship ( P for nonlinearity < 0.001). Similarly, LOT showed no significant independent association in the fully adjusted model (aHR per 1-day increase, 1.03; 95% CI 0.97–1.11; P = 0.214), although a non-linear trend suggested increasing risk with longer durations ( P for nonlinearity < 0.001). Relative importance analysis identified ASI per antibiotic day as the most critical factor ( P < 0.001), significantly outweighing both DDDs and LOT ( P > 0.05). Conclusions This study identifies ASI per antibiotic day as the principal independent risk factor for ICU-acquired CR-GNB, significantly outweighing the adjusted impact of DDDs or LOT. Therefore, prioritizing antibiotic spectrum optimization is crucial for stewardship strategies targeting CR-GNB prevention in the ICU. Trial registration Chinese Clinical Trial Registry Identifier ChiCTR2400081352. Registered 28 February 2024.