This article is more than 5 years old. Information may no longer be current.
Novel algorithm predicted risk for CIN in patients undergoing PCI
Click Here to Manage Email Alerts
A novel, easy-to-use algorithm can reliably predict the risk for contrast-induced nephropathy and new requirement for dialysis among patients undergoing PCI, study results found.
Random forest models were developed using 46 baseline clinical and laboratory variables to estimate the risk for contrast-induced nephropathy (CIN) in patients undergoing PCI in a cohort of 68,573 procedures performed at 46 hospitals from January 2010 to June 2012 in Michigan. Of the procedures, 48,001 (70%) were randomly selected for developing the model and 20,572 (30%) for validating it.
Hitinder S. Gurm, MD, with the University of Michigan, Ann Harbor, and fellow colleagues chose the 15 most influential variables for a model to estimate risk for CIN and new requirement for dialysis in an independent validation data set using area under the receiver-operating characteristic curve (AUC). These variables included patient presentation (PCI indication, PCI status, CAD presentation, cardiogenic shock, HF within 2 weeks and pre-PCI left ventricular ejection fraction), clinical history (diabetes history and treatment), patient characteristics (age, weight and height) and pre-procedural lab values (creatine kinase, serum creatinine, hemoglobin, troponin I and troponin T).
Hitinder S. Gurm
The models had excellent calibration and discrimination for CIN (AUC for full model, 0.85; for reduced model, 0.84; P <.01) and for new requirement for dialysis (AUC for both models, 0.88; P=0.82; net reclassification improvement for CIN, 2.92%; P=.06).
“Standard clinical and laboratory variables that are routinely collected in patients undergoing PCI can be easily and reliably used to predict the risk of CIN and [new requirement for dialysis] using a novel computational tool,” Gurm and colleagues wrote. “The robust discrimination and calibration of this method, combined with the ease of use for simplified bedside prediction, makes this model an easy tool to apply clinically.”
The CIN calculator can be accessed at https://bmc2.org/calculators/cin.
Disclosure: Gurm has received research funding from Blue Cross Blue Shield of Michigan and the NIH and Agency for Healthcare Research & Quality.
Collapse
Click Here to Manage Email Alerts