Predictive Value Of Extravascular Lung Water In Critically Ill Patients

Objectives:The study is to analyze whether EVLW (extravascular lung water) measured by single indicator transpulmonary thermodilution technique can predict ICU (intensive care unit) mortality of critically ill patients, and to compare the predictive value of EVLW in different periods or indexed to different body weights.Methods:We reviewed the clinical data of patients who were monitored with pulse induced contour cardiac output (PiCCO) technique from June 2007 to December 2009 in medical intensive care unit(MICU), Peking Union Medical College Hospital. We recorded the initial EVLW (EVLW1), maximum EVLW (EVLWmax), and EVLWmax within 24 hours of ICU admission (EVLWmax24) if available. We calculated daily mean EVLW (EVLWmean), and mean EVLW during ICU stay. The values of EVLW indexed to actual body weight, predicted body weight and adjusted body weight were calculated respectively by the formula (EVLW absolute value/body weight). We compared the difference of variables between survivor group and nonsurvivor group and evaluated whether these variables can predict ICU mortality of critically ill patients.Results:The values of EVLW indexed to actual body weight, predicted body weight and adjusted body weight were significantly increased in nonsurvivors compared with survivors. Multiple Logistic regression analysis showed that APACHE(acute physiology and chronic health evaluation)â…¡(OR 1.138,95%CI 1.031-1.256, P=0.010), age (OR 1.058,95%CI 1.010-1.109, P=0.016) and EVLWmax indexed to predicted body weight (pbwEVLWmax) (OR 1.151,95%CI 1.042-1.272. P=0.006) were independent predictors of ICU mortality in critically ill patients. The areas under ROC (receiver operating characteristic) curves of APACHEâ…¡, age. and pbwEVLWmax were 0.704, 0.627, and 0.675 respectively, without statistical significance between each other. Spearman correlation analysis showed that APACHEâ…¡, age and pbwEVLWmax were positively correlated with ICU mortality (r= 0.341,0.288,0.210; P=0.001,0.002, 0.026). ICU mortality significantly increased as pbwEVLWmax increased (P<0.001). The difference of daily EVLWmean indexed to predicted body weight between survivors and nonsurvivors became statistically significant.Conclusions:Among all EVLW-derived variables, pbwEVLWmax was the only independently prognostic indicator of ICU mortality. The predictive value of pbwEVLWmax was comparable to APACHEâ…¡. However, dynamic changes of EVLW might help to predict patient outcome.