Introduction and objectives: Creatinine increases during acute heart failure treatment may occur in the setting of hemoconcentration and decongestion, but the ranges of creatinine increase associated with lower risk, safety, or harm remain poorly defined.

Methods: We studied 2043 consecutive hospitalizations for acute heart failure between 2014 and 2021. The main exposures were the maximum in-hospital increase in creatinine (ΔCr) and hemoglobin change (ΔHb). Models used restricted cubic splines for ΔCr, continuous ΔHb, and their interaction. Mortality and composite endpoints were analyzed with Cox models; first all-cause and heart failure rehospitalization were analyzed using Fine-Gray models. The primary ΔHb contrast was the observed interquartile range reporting contrast (+1.3 g/dL), rather than a subgroup cutoff.

Results: The median age was 77 [69-83] years and 57.6% of hospitalizations involved male patients. In the context of hemoconcentration, ΔCr ranges associated with lower risk were identified for mortality (0.10-0.18 mg/dL), first all-cause rehospitalization (0.06-0.25 mg/dL), first heart failure rehospitalization (0.00-0.32 mg/dL), and the composite of death or first all-cause rehospitalization (0.10-0.23 mg/dL). Harm thresholds were observed above 0.61, 0.98, and 0.91 mg/dL, respectively; no harm threshold was identified for heart failure rehospitalization.

Conclusions: Among patients hospitalized with acute heart failure who achieved hemoconcentration, small creatinine increases were associated with lower risk, whereas larger increases were associated with a loss of benefit and, for some endpoints, harm. These findings support a continuous, context-dependent interpretation of worsening renal function during decongestive therapy.