Abstract 16276: A Web-based Tool for Predicting Acute Kidney Injury in ST-elevation Myocardial Infarction
ST-elevation myocardial infarction (STEMI) requires emergent diagnostic catheterization and therapeutic percutaneous coronary intervention. These therapies may cause acute kidney injury (AKI), a serious procedural complication that contributes to morbidity and mortality in STEMI patients. Identifying patients at risk for renal complications may save lives and reduce healthcare costs, but there is no validated tool for predicting AKI in STEMI patients.
Methods: The UT-Methodist STEMI program is a racially diverse series of 1144 consecutive patients. We derived and validated the UT-AKI score to predict the risk of AKI in patients referred for STEMI (and not currently on dialysis). AKI was categorized by the modified AKI Network and RIFLE indices. The UT-AKI score was derived from logistic regression analysis performed on a randomly selected derivation dataset using variables with significant univariate relationships to AKI and markers of renal function. Receiver operator characteristic (ROC) curves for the UT-AKI were assessed in the derivation and validation datasets. Statistical significance was measured by appropriate parametric and non-parametric tests; a p<0.05 was considered significant.
Results: The incidence of AKI was similar in patients undergoing diagnostic catheterization with or without PCI (12% vs. 14%, p=0.34). ROC curve analysis of the validation data set showed that the UT-AKI score (AUC 0.76 (95% CI 0.70-0.82)) was significantly better at classifying patients at risk for AKI than the ACEF (AUC 0.65 (95% CI: 0.58-0.73)), AGEF (AUC 0.65 (95% CI: 0.58-0.72)), or Mehran (AUC 0.67 (95% CI: 0.59-0.75)). The classification accuracy for the UT-AKI index was comparable for milder and severe forms of AKI.
Conclusions: In the largest analysis of STEMI patients to date, we found no significant difference in the risk of AKI from PCI vs. diagnostic catheterization. The validation study indicates that the UT-AKI index was significantly better than other scores at classifying which patients will develop AKI. Accurate prediction of AKI may help physicians to personalize therapies and develop novel strategies to reduce kidney injury in patients with STEMI.
Author Disclosures: B.R. Zambetti: None. R.T. Ellis: None. I. Hwang: None. M. Chumpia: None. M.R. Heckle: None. B. Smith: None. M.T. Crouse: None. N. Efeovbokhan: None. F. Thomas: None. R.N. Khouzam: None. U.N. Ibebuogu: None. G.L. Reed: None.
- © 2015 by American Heart Association, Inc.