Abstract
Background/Objectives: Acute kidney injury (AKI) is a common complication in hospitalized patients and carries a substantial risk of chronic kidney disease (CKD), dialysis dependence, and mortality. Although novel biomarkers such as NGAL, KIM-1, and cystatin C have shown promise, their high cost and limited availability restrict their use in routine practice, particularly in developing countries where CKD incidence is rising. The trajectory of serum creatinine decline after its peak may provide a simple, low-cost, and universally available prognostic marker for renal recovery. Methods: This retrospective cohort study included 817 adult patients diagnosed with AKI between January 2015 and December 2024. The creatinine decline rate was calculated as the difference between peak and discharge creatinine divided by hospital stay (mg/dL/day). Patients were stratified into rapid or slow decline groups according to the median value (0.19 mg/dL/day). Post-discharge outcomes, including CKD development, readmission, dialysis requirement, and mortality, were evaluated at 3, 6, and 12 months. Receiver operating characteristic (ROC) analysis was performed to determine the optimal cutoff for predicting renal recovery. Results: Patients in the rapid decline group (n = 409) were younger and had fewer comorbidities and shorter hospital stays than those in the slow decline group (n = 408). The ROC analysis yielded an AUC of 0.78 (95% CI 0.73-0.82, p < 0.001) with an optimal cutoff of 0.18 mg/dL/day (sensitivity 76%, specificity 71%). At 12 months, CKD (18.6% vs. 34.3%), dialysis requirement (3.4% vs. 8.8%), readmission (29.8% vs. 41.2%), and mortality (9.3% vs. 14.2%) were all significantly higher in the slow decline group (all p < 0.05). In multivariable analysis, faster creatinine decline independently predicted renal recovery (OR = 1.36 per 0.1 mg/dL/day, 95% CI 1.22-1.53, p < 0.001), along with younger age, higher serum albumin, and shorter hospital stay. In the longitudinal GEE model, both time (p = 0.004) and group effects (p < 0.001) remained significant, with an interaction effect (p = 0.018) indicating greater eGFR improvement over time among patients with rapid creatinine decline. Conclusions: The rate of creatinine decline is an independent predictor of long-term renal recovery following AKI. This simple and inexpensive parameter may complement novel biomarkers and serve as a practical risk-stratification tool in diverse clinical settings, especially where resources are limited. Prospective multicenter studies integrating albuminuria and emerging biomarkers are warranted to validate and expand these findings.