Diuretic Response: Clinical and Hemodynamic Predictors and Relation to Clinical Outcome

Objective The aims of this work were to investigate the clinical and hemodynamic profile underlying the response to loop diuretics in acute decompensated heart failure (ADHF), and to compare the relative usefulness of measures of diuretic resistance for predicting mortality. Methods and Results We studied 475 patients with ADHF, of whom 241 underwent right heart catheterization. Linear regression models were used to identify factors that affected urine output. Loop diuretics response was estimated as 1) net fluid loss per 40 mg furosemide equivalents and 2) urine output produced per 40 mg furosemide equivalents. In a multivariable regression model, key independent predictors of urine output included diuretic dose (partial r = 0.44), baseline renal function (partial r = 0.38), systolic blood pressure (partial r = 0.26), and fluid intake (partial r = 0.31; all P < .0001). Among hemodynamic variables, elevated right atrial pressure was associated with greater urine output (partial r = 0.19; P = .002). The partial correlation attributable to diuretic dose (partial R2 = 0.19) accounted for approximately one-half of the variance in urine output explained by the model (model R2 = 0.40).Cox regression models demonstrated inverse relationships between quartiles of net fluid loss (P = .004) and quartiles of urine output (P = .04) per 40 mg furosemide and 6-month mortality. When comparing nested models, the model based on net fluid loss was better than the model based on urine output for the prediction of mortality (χ2 = 8.1; 3 df; P = .04). Conclusions In patients with ADHF, beyond diuretic dose, other parameters including renal function, hemodynamic status, the degree of volume overload, and fluids intake also affect urine output. Measures of loop diuretic response are associated with short-term mortality.