Tailoring diuretic therapy in acute heart failure: insight into early diuretic response predictors

Few data exist to help physicians in the use of diuretics to provide the greatest symptomatic benefit with the least adverse effect to patients and to select the subset of patients who require a more aggressive diuretic strategy and monitoring. The aim of this study is to identify early predictors of diuretic response in a selected group of patients with acutely decompensated chronic heart failure (ADCHF). This was an observational, retrospective secondary analysis of a study including 100 patients with ADCHF. The mean +/- A standard deviation (SD) of age was 76.0 +/- A 10.9 years. Sixty-one patients were female. After three days of inpatient treatment, 16 (16 %) patients maintained or increased i.v. furosemide dose (slow diuretic response, SDR). This group of patients had more indirect signs of fluid overload, including greater body mass index increment. The other 84 patients had greater congestion relief and had decreased i.v. furosemide dose or were switched furosemide to oral route (fast diuretic response, FDR). Admission day factors predicting SDR were: higher levels of pUr (mean +/- A SD, 69.6 +/- A 20.9 vs. 52.5 +/- A 19.8, p = 0.002), higher levels of pUr/pCr ratio (mean +/- A SD, 58.3 +/- A 15.2 vs. 49.6 +/- A 15.1, p = 0.036), higher levels of albuminuria [median (IQR), 131.5 (396.9) vs. 47.1 (143.6), p = 0,011], higher levels of red cell distribution width (RDW) [median (IQR), 16.0 (1.9) vs. 15.1 (1.5), p = 0.039], lower levels of HgB (mean +/- A SD, 11.5 +/- A 1.8 vs. 12.6 +/- A 2.1, p = 0.04) and higher levels of hsTnT [median (IQR), 0.05 (0.05) vs. 0.03 (0.03), p = 0,026]. By multivariate analysis, the strongest independent early predictors of SDR were: pUr [OR (95 % CI), 1.04 [1.01-1.07], p = 0.006] and RDW [OR (95 % CI), 1.47 (1.07-2.02), p = 0.018]. During the first 3 days of hospitalization, the strongest independent factor associated with SDR was NTproBNP increase or decrease by less than 30 % from day 1 to day 3 [OR (95 % CI), 4.84 (1.14-20.55), p = 0.032]. Use of high-dose spironolactone was associated with FDR [OR (95 % CI), 0.17 (0.03-0.85), p = 0.031]. High RDW and high levels of pUr at admission are strong predictors of slower diuretic response. No change or increase in NTproBNP in the first 3 days of treatment is associated with slower diuretic response. On the other hand, the use of high-dose spironolactone is associated with faster diuretic response.