Résumé
PURPOSE: To determine whether initial corrected anion gap (C(o)AG), base excess caused by unmeasured anions (BEua), and strong ion gap (SIG) can predict the morbidity of critically ill patients admitted to emergency department (ED). METHODS: 138 patients who visited the critical section of the ED and were admitted to intensive care unit (ICU) were enrolled. We calculated the C(o)AG, BEua, and SIG from the initial blood samples of the patients and initial logistic organ dysfunction score (LODS) also. Then we measured the LODS at the last day of ICU stay again. Comparing with the initial LODS, we divided the patients into two groups based on the changes of the values: favorable group and poor group. RESULTS: There was a significant difference in the mean AGcorr (p=0.007), BEua (p=0.008), SIG (p=0.037) between favorable and poor group. The area under the receiver operating characteristic (AUROC) curves for morbidity prediction were relatively small: 0.66 (95% CI, 0.56-0.77) for C(o)AG, 0.65 (95% CI, 0.54-0.76) for BEua, and 0.59 (95% CI, 0.49-0.70) for SIG. CONCLUSION: We found the initial unmeasured anions at the ED of the patients who eventually showed improved LODS during ICU period are significantly different to those of the other patients. But they failed to show enough capability of discriminating the morbidities between two groups.