The association between fluid balance and mortality in patients with ARDS was modified by serum potassium levels: a...

Background and Objective. Acute respiratory distress syndrome (ARDS) is characterized by pulmonary edema and may benefit from conservative fluid management. However, conflicting results exist in the literature. The study aimed to investigate the association between mean fluid balance and mortality outcome in ARDS patients who required invasive mechanical ventilation. Methods. The study was a secondary analysis of a prospectively collected dataset obtained from the NHLBI Biologic Specimen and Data Repository Information Coordinating Center. ARDS patients with invasive mechanical ventilation were eligible. Demographic and laboratory data were extracted from the dataset. Multivariable regression model was built by stepwise selection of covariates. A fractional polynomial approach was used to test the linearity of mean fluid balance in the model. The potential interactions of mean fluid balance with other variables were tested. Main Results. A total of 282 patients were eligible for the analysis, including 61 non-survivors with a mortality rate of 21.6%. After stepwise regression analysis, mean fluid balance remained to be an independent predictor of death (OR: 1.00057; 95% CI [1.00034-1.00080]). The two-term model obtained using fractional polynomial analysis was not superior to the linear model. There was significant interaction between mean fluid balance and serum potassium levels (p = 0.011). While the risk of death increased with increasing mean fluid balance at potassium levels of 1.9, 2.9 , 3.9 and 4.9 mmol/l, the risk decreased at potassium level of 5.9 mmol/l. Conclusion. The present study demonstrates that more positive fluid balance in the first 8 days is significantly associated with increased risk of death. However, the relationship between mean fluid balance and mortality can be modified by serum potassium levels. With hyperkalemia, more positive fluid balance is associated with reduced risk of death.