Abstract
BACKGROUND: Loop diuretics are common therapy for emergency department (ED) patients with acute heart failure (AHF). Diuretic resistance (DR) is a term used to describe blunted natriuretic response to loop diuretics. It would be important to detect DR prior to it becoming clinically apparent, so early interventions can be initiated. However, several definitions have been proposed, and it is not clear if they identify similar patients. We compared these definitions and described the clinical characteristics of patients who fulfilled them. METHODS: To qualify for this secondary analysis of 1033 ED patients with AHF, all patients needed to receive intravenous diuretics in the ED and have urine available within 24 h of their ED evaluation. A poor diuretic response, suggesting DR, was characterized by (1) a fractional sodium excretion (FeNa) of less than 0.2%; (2) spot urinary sodium of less than 50 meq/L; and (3) a urinary Na/K ratio <1.0. McNemar's test was used to compare the different cohorts identified by the three definitions. Secondary analyses evaluated associations between each DR definition and hospital length of stay (LOS), ED revisits and rehospitalizations for AHF, and mortality using the Wilcoxon rank-sum tests and linear regression or Pearson chi-square test and logistic regression, as appropriate. RESULTS: The median age of the 187 patients was 64, and 50% were African-American. There were 5.9% of patients with a FeNa less than 0.2%, 17.1% had urinary sodium less than 50 meq/L, and 10.7% had a urinary Na/K ratio <1.0. The three definitions identified significantly different patients with very little overlap (p < 0.02 for all comparisons). There were 37 (19.8%) patients who were readmitted to the ED or hospital or died within 30 days of ED evaluation. Patients with spot urinary sodium less than 50 meq/L were more likely to be readmitted (p = 0.03). CONCLUSIONS: The patient proportion with poor natriuresis and DR varies depending on the definition used. Early ED therapy would be impacted at different rates if clinical decisions are made based on these definitions. These findings need to be further explored in a prospective ED-based study. TRIAL REGISTRATION: ClinicalTrials.gov, NCT00508638.