Red Cell Distribution Width and Acute Kidney Injury After Aortic Valve Replacement: A Retrospective Cohort Study From the MIMIC‐IV Database

Background Acute kidney injury (AKI) is a common and serious complication in ICU patients after aortic valve replacement (AVR), which can lead to increased morbidity and mortality. Red cell distribution width (RDW) has emerged as a prognostic marker for adverse outcomes in cardiovascular diseases. However, its association with AKI following AVR remains underexplored, particularly among those in intensive care units (ICUs). We examined the association between ICU‐admission RDW and postoperative AKI and 365‐day mortality in AVR patients. Methods This retrospective cohort study analyzed 2267 ICU‐admitted AVR patients via the MIMIC‐IV database (2008–2022). AKI was defined by KDIGO criteria. Adult AVR patients with ICU admission were included. The 365‐day all‐cause mortality was the secondary outcome. The 14.5% RDW cutoff was selected according to the upper limit of the clinical reference range; the observed dose–response relationship was linear, so no inflection point was identified. Multivariable logistic regression and Cox proportional hazards models with progressive adjustments were employed. All models were fitted by standard maximum likelihood estimation; Firth penalized‐likelihood regression was used only as a sensitivity analysis. The incremental predictive value was evaluated using the area under the curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI) metrics. Results In a cohort of 2267 patients with a mean age of 71.1 ± 13.0 years, of whom 61.9% were male, 1782 individuals (78.6%) experienced AKI, an incidence notably higher than in general cardiac surgery cohorts, likely reflecting the ICU‐admitted study population. The cutoff for RDW was 14.5%. In partially adjusted models (Model 4), each 1% increase in RDW at ICU admission was associated with 8.1% higher AKI odds (OR 1.081, 95% CI 1.003–1.165, p = 0.040) and 15.2% higher 365‐day mortality hazard (HR 1.152, 95% CI 1.079–1.230, p < 0.001). High RDW (≥ 14.5%) was independently associated with AKI (OR 1.315, 95% CI 1.016–1.702, p = 0.037). After additional adjustment for baseline creatinine and hemoglobin, the AKI association was attenuated and no longer statistically significant (Model 5, OR 1.076, 95% CI 0.994–1.166, p = 0.071), whereas the mortality association remained significant (HR 1.102, 95% CI 1.024–1.185, p = 0.009). Adding RDW did not significantly improve discrimination for AKI (ΔAUC 0.004, p = 0.62; NRI 0.099, IDI 0.002) or 365‐day mortality (ΔAUC 0.004, p = 0.48; NRI 0.229, IDI 0.006). Conclusions In patients admitted to the ICU for AVR, the relationship between RDW and AKI was primarily accounted for by anemia and baseline renal function. After adjusting for baseline creatinine and hemoglobin levels, RDW was no longer significantly correlated with AKI and did not enhance the predictive accuracy for AKI or 365‐day mortality. However, RDW remained independently associated with 365‐day mortality and with moderate‐to‐severe AKI (stage ≥ 2), a secondary endpoint without formal multiple‐testing correction, indicating that it serves as a marker of overall physiological reserve rather than a specific or incremental predictor of AKI. Consequently, its clinical utility for AKI risk stratification in this patient population is limited.