Factors Predicting Survival in ICU Patients Requiring Fluid Resuscitation

Rationale Fluid resuscitation is common in ICU and has been increasingly investigated in the last two decades. However factors predicting survival of patients undergoing fluid resuscitation is not known. We examined the data from the SAFE study to examine the factors predicting survival in ICU patients undergoing fluid resuscitation. MethodsData from SAFE study (A Comparison of Albumin and Saline for Fluid Resuscitation in the Intensive Care Unit) was accessed from The George Institute for Global Health Australia after appropriate approvals. Demographic data, baseline hemodynamic data and fluid balance were extracted and examined. Univariate and logistic regression analysis was done to examine the survivors and non survivors. Data is presented as mean and standard deviation or median and interquartile range based on the distribution of the data. Results There were 6841 patient in the dataset in whom the survival data at day 28 was available. There were 5396 patients (78.8%) who survived till day 28. Survivors were younger [56.6 (19.1) vs. 65.8 (16.2) years], lower APACHE II score at admission [17.5 (7.4)vs. 23.8 (7.9)] and more likely to have an elective surgery [1254 (23.2%) vs 86 (6.0%)], all p<0.001. Survivors also had a lower baseline heart rate (per minute) [90.5 (23.0) vs. 96.9 (24.6)],higher mean arterial blood pressure (mm Hg) [(78.6 (16.3) vs. 75.8 (16.6)], lower central venous pressure (mm Hg) [8.4 (4.5) vs. 9.6 (5.3)] and a lower day one fluid balance (litres) [2.26 (1.31-3.59) vs. 2.72 (1.48- 4.63)], all p <0.001. Using logistic regression (AUC ROC 0.78), factors [OR(SE), p value] which predicted non survival were age [1.027 (0.002) p<0.001], organ failure score[1.073(0.015) p=0.000], APACHE II score [1.071 (0.006) p=0.000], admission after elective surgery [0.368 (0.059) p=0.000], higher heart rate [1.006 (0.002) p=0.001] and higher day 1 fluid balance [1.103 (0.016) p=0.000]. Conclusion - Patients requiring fluid resuscitation in the intensive care unit, a higher fluid balance on day 1 was the strongest and only modifiable predictor of mortality at day 28.