Rescue ultrafiltration indication in acute heart failure hospitalizations: time trends in temporary use, predictors and clinical implications

Background and Aims Diuretic resistance is a severe event in patients (pts) hospitalized for acute decompensated heart failure (AHF). Among these pts, rescue ultrafiltration (rescue UF) is a therapeutic option to minimize risk and shorten length of stay (LOS). Our aim was to study rescue UF usage time trends, predictors and impact on outcomes. Methods Pts over 18 years-old, admitted to Intensive Care Telemetry Unit in our Hospital for AHF between July, 2011 and Aug, 2024 were included in an observational and cross-sectional study. Those who received a heart transplant or on chronic dialysis were excluded from analysis. Period identification was based on modality in diuretic use. Inverse probability weighting using most relevant predictors was used to generate two groups with similar baseline characteristics. Logistic regression and time to event analysis were adjusted to determine outcomes. Analysis was conducted using R Core Team, 2024. Results From a total of 1880 pts, 200 developed diuretic resistance, and 47 from them received rescue UF. While we found no differences in gender (p= 0.3), rescue UF were younger (68 vs 75 years old; p< 0.001), without differences neither in left ventricular ejection fraction (41 vs 40%; p= 0.9) nor in etiology of heart disease (coronary heart disease 26 vs 26%; p= 0.9; idiopathic 23 vs 15%; p = 0.06). Rescue UF pts had higher prior admission rate (51 vs 39%; p= 0.046) and longer LOS (15 vs 7 days; p< 0.001). Their admission creatinine was higher (1.90 vs 1.10 mg/dl; p<0.001), as was proportion of pts with glomerular filtration rate (GFR) < 30 ml/min/1.73 m2 (42 vs 7.4%; p< 0.001). Concerning to HF phenotypes, predominant hypoperfusion (21 vs 10%) and right-sided HF (35 vs 23%) were more prevalent in rescue UF, while pulmonary congestion was less frequent (44 vs 66; global p< 0.001). Neither In-hospital (HR 0.89; CI95% 0.48-1.65; p=0.71), nor post-discharge mortality (HR 0.71; CI95% 0.25-1.96; p=0.5) or readmission rates (HR 1.25; CI95% 0.70-2.22; p=0.5) were higher after receiving rescue UF. Age under 65 (OR 0.33; CI95% 0.21-0.54; p< 0.001); 2017-2022 period (OR 0.69; CI95% 0.49-0.96; p= 0.03) and 2023-2024 period (OR 0.30; CI95% 0.09-0.78; p< 0.001) were independent predictors for a lesser rescue UF indication, while predominant pulmonary congestion phenotype (OR 1.6; CI95% 1-2.41; p= 0.05) and GFR <30 ml/min/1.73 m2 (OR 4.4; CI95% 2.7-7.5; p< 0.001) were independent predictors for higher rescue UF indication. Conclusions Diuretic resistance portends poor outcomes in AHF admissions. Under more profound tubular blockade, there is a lower likelihood of requiring UF. Admission GFR <30 ml/min/1.73 m2 coupled with predominant pulmonary congestion phenotype remains the two most important features to consider for an early invasive strategy to lessen LOS and improve short-term outcomes. It is also important to note that neither ventricular function, nor etiology of heart disease had impact in UF decision making.