predictor factors for sepsis diagnosis, length of icu stay and mortality in icu

introduction: in a prospective study, we aimed to assess the value of enhanced rdw to predict sepsis and evaluate factors affecting length of icu stay and in-hospital mortality among sepsis patients. material and method: after exclusion of 27 patients, we included 160 adult patients with suspicious sepsis admitted at university affiliated hospital with 33 icu-beds from 2010 to 2012. nighty patients were diagnosed with sepsis and the source of infection was defdined. receiver–operating characteristic (roc) curves were used to examine the sepsis predictions from rdw, apache ii scores, and combination of them. the primary endpoint of this analysis was icu mortality. the secondary endpoints were length of stay at icu and hospital.a linear regression analysis was used to study risk factors for longer icu stay and we used logistic regression analysis to predict factors affecting in-hospital mortality. results: the addition of elevated rdw value to apacheii score in critically illness states enhanced the auc for predicting sepsis and its differentiation from sirs.female patients and those with numerous co-morbidities or aki and those on mechanical ventilation significantly stayed longer at icu. moreover, the patients with higher apache ii score died significantly more than others. conclusion: the addition of elevated rdw value to apacheii score in critically illness helps to differentiate sepsis from sirs.