AKI prediction model in acute aortic dissection surgery: nomogram development and validation.

OBJECTIVES

This multicenter study developed and internally validated a biomarker-enhanced risk prediction nomogram integrating hemodynamic parameters and novel urinary biomarkers to stratify postoperative acute kidney injury (AKI) risks in patients undergoing emergency surgical repair for acute Stanford Type A aortic dissection (ATAAD).

METHODS

A cohort of 1,277 patients from the China Aortic Dissection Alliance (CADA) registry was chronologically split into derivation (70%,  = 894) and validation (30%,  = 383) sets. LASSO regression with 10-fold cross-validation (λ1SE criterion) was applied to identify non-redundant predictors from 34 candidate variables (e.g., cardiac dysfunction [LVEF <50% or INTERMACS 1-3]) and elevated urinary biomarkers. Multivariable logistic regression refined these predictors to establish independent risk factors for the final nomogram. Model performance was evaluated using the concordance index (C-index), area under the receiver operating characteristic curve (AUC-ROC), calibration plots (Brier score and Hosmer-Lemeshow test), and decision curve analysis (DCA) to quantify clinical utility.

RESULTS

Multivariable analysis identified seven independent predictors of postoperative AKI: preexisting cardiac dysfunction (adjusted odds ratio [aOR] = 2.17; 95% CI: 1.68-3.56), microvascular complications of diabetes (aOR = 3.26; 2.71-4.34), baseline renal impairment (aOR = 1.72; 1.36-3.29), blood urea nitrogen (BUN) ≥ 20 mg/dL (aOR = 2.19; 1.57-3.64), glomerular filtration rate (GFR) < 90 mL/min/1.73 m (aOR = 1.47; 1.02-2.13), serum creatinine >1.3 mg/dL (aOR = 3.28; 2.58-3.75), and peripheral vasculopathy (aOR = 1.78; 1.12-2.32). The model demonstrated strong discrimination (training AUC-ROC: 0.830 [0.802-0.858]; internal validation AUC-ROC: 0.786 [0.737-0.834]), calibration (Brier scores: 0.138 training, 0.141 validation), and clinical utility (net reclassification improvement [NRI] = 0.21,  = 0.001), with optimal decision thresholds at 40-60% probability.

CONCLUSION

The nomogram demonstrates superior preoperative discriminative accuracy in AKI following ATAAD repair surgery. External validation via the VASCUNET registry is planned to confirm generalizability.