Background: Approximately one-third of patients with acute decompensated heart failure (ADHF) treated with an intravenous (iv) loop diuretic at a relatively high dose (>80 mg/day of furosemide, or an equivalent dose of another loop diuretic), exhibit worsening renal function (WRF) after a single course of iv infusions or iv bolus injections maintained for several days. WRF is currently defined as an increase in serum creatinine >0.3 mg/dL (WRF-Cr) or a decrease in the estimated glomerular filtration rate of ≥20% (WRF-GFR) compared to baseline measurements. Furthermore, small increases in serum creatinine in the high-normal range of its values are indicative of significant reductions in estimated glomerular filtration rate (eGFR) due to the exponential relationship between serum creatinine and eGFR. Therefore, underestimating this relationship could lead to an erroneous quantitative estimate of new-onset renal dysfunction, diuretic-related.
Methods: The relationship between baseline serum creatinine (exposure variable) and the risk of diuretic-related WRF (dichotomous outcome variable), expressed either as WRF-Cr or as WRF-GFR, was assessed by logistic regression analysis. For this purpose, medical records with a diagnosis of previous ADHF were collated, and retrospectively analyzed. The eGFR was calculated using the equation "Modification of Diet in Renal Disease" (MDRD). The WRF was inferred from measurements of serum creatinine that had been made daily during the scheduled courses of intravenous diuretic therapy.
Results: Thirty-eight patients with chronic heart failure (CHF) and history of a previous episode of ADHF were enrolled in the study. An increase higher than 0.3 mg/dL of serum creatinine (WRF-Cr) was detected in 14 of 38 patients (36.8%). In addition, a decrease of ≥20% in GFR (WRF-GFR) was detected in 14 of 38 patients (36.8%). However, a poor concordance between the two criteria was found (Cohen's Kappa =0.208, 95% CI: -0.110 to 0.526). WRF-Cr and WRF-GFR showed opposing relations with baseline serum creatinine. In fact, the risk of WRF-Cr appeared positively associated with baseline serum creatinine (odds ratio =33.56; 95% CI:2.93- 384.18 P=0.0047), while the risk of WRF-GFR was inversely associated with the same analyte (odds ratio =0.0393; 95% CI: 0.0039 to 0.3966 P=0.0061).
Conclusions: The criterion to discontinue the iv diuretic or to reduce its dosage in the presence of WRF-Cr for patients with ADHF or resistance to oral diuretic should be joined with the useful notion that this finding indicates a significant reduction of eGFR only for values of serum creatinine in the normal or near-normal ranges.