Niklas Jonsson1, Patrik Gille-Johnson2, Claes-Roland Martling3, Shengyuan Xu4, Per Venge4, Johan Mårtensson3. 1. Department of Physiology and Pharmacology, Karolinska Institutet, 171 77, Stockholm, Sweden. Electronic address: niklas.jonsson@sll.se. 2. Department of Medicine, Karolinska Institutet, 171 77, Stockholm, Sweden. 3. Department of Physiology and Pharmacology, Karolinska Institutet, 171 77, Stockholm, Sweden. 4. Departments of Medical Sciences and Clinical Chemistry, Uppsala University, 751 05 Uppsala, Sweden.
Abstract
PURPOSE: To assess the value of dimeric neutrophil-gelatinase associated lipocalin (NGAL) as an early marker of bacterial infection and its response to antibiotic therapy in intensive care unit (ICU) patients. MATERIALS & METHODS: We measured daily plasma dNGAL in 198 patients admitted to a mixed ICU. Likelihood of infection was determined with International Sepsis Forum criteria. We measured dNGAL in 145 healthy controls to establish normal values. RESULTS: ICU patients had higher dNGAL than healthy controls. A suspected or confirmed infection was independently associated with 90% (95% CI 15-215%) higher dNGAL than absence of infection. We observed no association between acute kidney injury and dNGAL. Diagnostic accuracy at antibiotic treatment initiation, assessed with area under the receiver-operating characteristics curve (AUC-ROC), for dNGAL was 0.70 (95% CI 0.60-0.79). AUC-ROC for dNGAL 24 h before antibiotic treatment initiation was 0.54 (95% CI 0.41-0.66). The mean (95% CI) change of dNGAL in the first 2 days after appropriate antibiotic therapy initiation was -31 (-49,-13)%. CONCLUSIONS: In our cohort of ICU patients, plasma dNGAL was associated with presence of bacterial infections independent of AKI but it performed poor as a predictor of infections. Following antibiotic therapy, dNGAL markedly decreased-supporting further exploration of dNGAL-guided antibiotic de-escalation.
PURPOSE: To assess the value of dimeric neutrophil-gelatinase associated lipocalin (NGAL) as an early marker of bacterial infection and its response to antibiotic therapy in intensive care unit (ICU) patients. MATERIALS & METHODS: We measured daily plasma dNGAL in 198 patients admitted to a mixed ICU. Likelihood of infection was determined with International Sepsis Forum criteria. We measured dNGAL in 145 healthy controls to establish normal values. RESULTS: ICU patients had higher dNGAL than healthy controls. A suspected or confirmed infection was independently associated with 90% (95% CI 15-215%) higher dNGAL than absence of infection. We observed no association between acute kidney injury and dNGAL. Diagnostic accuracy at antibiotic treatment initiation, assessed with area under the receiver-operating characteristics curve (AUC-ROC), for dNGAL was 0.70 (95% CI 0.60-0.79). AUC-ROC for dNGAL 24 h before antibiotic treatment initiation was 0.54 (95% CI 0.41-0.66). The mean (95% CI) change of dNGAL in the first 2 days after appropriate antibiotic therapy initiation was -31 (-49,-13)%. CONCLUSIONS: In our cohort of ICU patients, plasma dNGAL was associated with presence of bacterial infections independent of AKI but it performed poor as a predictor of infections. Following antibiotic therapy, dNGAL markedly decreased-supporting further exploration of dNGAL-guided antibiotic de-escalation.