Total amount: € 0,00
Indexed/Abstracted in: e-psyche, EMBASE, PubMed/MEDLINE, Neuroscience Citation Index, Science Citation Index Expanded (SciSearch), Scopus
Impact Factor 1,651
Online ISSN 1827-1855
Prisco L. 1, Iscra F. 1, Ganau M. 2, Berlot G. 1
1 Department of Perioperative Medicine, Intensive Care and Emergency, Azienda Ospedaliera Universitaria Mista, Ospedali Riuniti di Trieste, Cattinara Hospital, Trieste, Italy;
2 Graduate School of Nanotechnology, University of Trieste, Trieste, Italy
AIM: Early hyperglycemia is a feature of traumatic brain injured (TBI) patients. The aim of our study was to analyze the impact of early hyperglycemia on in-ICU mortality in isolated TBI and its correlations with other factors responsible for secondary injury.
METHODS: We studied admission values (AV) and worse values in the first 48 hours (WV 48 h) of 112 ICU TBI patients (mortality 29.6%) of blood glucose (BG), base excess (BE), mean arterial pressure (MAP), PaO2/FiO2 ratio and serum hemoglobin (Hb). Predictive strength as the area under the receiver operating curves (AUROC) and correlation between all variables were calculated.
RESULTS: Data are expressed as median, 1st-3rd quartile. Both BG AV (147.5, 126-182 mg/dL; AUROC 0.716, P=0.0002) and WV 48 h (156.5, 132-192 mg/dL; AUROC 0.721, P=0.0001) are predictive of mortality. AV and WV 48 h are respectively: PaO2/FiO2 (366.8, 237.2-477.6 vs. 320, 214.4-426; P=0.05), MAP (90, 80-100.5 vs. 75, 66-83 mmHg; P<0.0001) and Hb (11.4, 9.7-13.1 vs. 10.6, 9-12.2 g/dL; P<0.02). BG AV and WV 48 h correlates with: age (r=0.419, P<0.0001 and r=0.489, P<0.0001), PaO2/FiO2 AV (r -0.223, P<0.03 and r -0.236, P<0.02), PaO2/FiO2 WV 48 h (r -0.215, P<0.03 and r -0.279, P<0.005) and MAP WV 48 h (r -0.216, P<0.03 and r -0.261, P<0.007).
CONCLUSION: Early hyperglycemia is a major predictor of mortality and correlates with other factors responsible for secondary injury. Early hyperglycemia seems to be a marker of inflammatory reaction responsible for early cardiovascular and respiratory impairment.