Early loop-diuretic efficiency and subsequent worsening renal function in critically Ill adults with a coded acute heart-failure diagnosis: a 12-hour landmark analysis of MIMIC-IV

Background Early urine output after intravenous loop-diuretic therapy is often interpreted relative to dose, but whether this dose-indexed measure provides prognostic information beyond urine output and loop-diuretic dose modeled separately remains uncertain. Methods We conducted a retrospective 12-hour landmark study using MIMIC-IV version 3.1. Adults with an acute or acute-on-chronic heart-failure discharge code in any diagnosis position who received an intravenous furosemide or bumetanide bolus within 24 h of ICU admission were included. Diuretic efficiency was cumulative urine output from t0 through < t0 + 12 h divided by cumulative loop-diuretic dose, expressed as mL per 40 mg furosemide equivalent. Worsening renal function (WRF) was a creatinine increase of at least 0.3 mg/dL strictly after the landmark through 48 h. Logistic models contained strictly landmark-available clinical predictors alone (M0), plus urine output and dose separately (M3), or plus diuretic efficiency (M4). Missing predictors were multiply imputed without using WRF as an imputation predictor; outcome-free imputation and model fitting were repeated within 500 bootstrap resamples. Results The primary cohort included 2,206 patients; 2,167 had evaluable creatinine, and 462 (21.3%) developed WRF. Optimism-corrected AUCs (empirical intervals) were 0.667 (0.639 to 0.692) for M0, 0.705 (0.678 to 0.731) for M3, and 0.703 (0.674 to 0.727) for M4. Compared with M0, M4 changed corrected AUC by 0.036 (0.019 to 0.049) and Brier score by -0.007 (-0.010 to -0.004). M4 did not outperform M3 (paired corrected AUC difference, -0.002 [-0.008 to 0.006]). In the exploratory principal-diagnosis subgroup (n = 286; 51 events), the conditional paired-bootstrap AUC increment was 0.002 (-0.020 to 0.022). Conclusions In this retrospectively defined, heterogeneous ICU population with an acute or acute-on-chronic heart-failure discharge code, lower record-derived early diuretic efficiency was associated with subsequent creatinine-defined worsening renal function. Adding diuretic efficiency to a strictly landmark-available clinical model increased discrimination only slightly; however, the ratio did not outperform urine output and loop-diuretic dose modeled separately. This direct comparison clarifies the prognostic information captured by the ratio and provides a benchmark for future studies, but it does not establish clinical utility. Prospective evaluation using standardized urine-output and creatinine assessment and clinically defined heart-failure populations is needed.