A comparison of definitions (RIFLE, AKIN, AND KDIGO) of acute kidney injury for prediction of outcomes in adults after isolated coronary artery bypass graft (CABG) surgery

Document Type



Nephrology; Medicine; Surgery; Cardiothoracic Surgery


Introduction: Acute kidney injury (AKI) occurs in as many as 40% of patients after cardiac surgery and is associated with an increased risk of mortality and morbidity, predisposes patients to a longer hospitalization, requires additional treatments, and increases the hospital costs. At present, there are three widely accepted consensus definitions providing uniform criteria for the diagnosis of AKI; RIFLE, AKIN and KDIGO. Each of these definition systems have their own benefits and limitations for predicting the degree of AKI as well as adverse outcomes (need for RRT, morbidity and mortality) in patients undergoing cardiac surgery. Having a standard definition for diagnosing and classifying AKI would enhance our ability to improve the management of patients and their clinical outcomes.The aim of this study is to compare the three AKI criteria (RIFLE (eGFR), AKIN and KDIGO) for their ability to predict all-cause mortality and morbidity after isolated CABG surgery in adult patients.
Methods: A single center retrospective review was conducted on adults who had undergone isolated CABG surgery during January 2013- January 2017 at Aga Khan University Hospital Karachi, Pakistan. Patients with known chronic kidney disease or a baseline Serum Creatinine of >1.1 and 1.3 mg/dL respectively for female and males were excluded. AKI was assessed on three definitions and estimated glomerular filtration rate (eGFR) was computed using standardized CKD-EPI-PK equations. Comparative ROC curves were built and Area under the Curve with sensitivity and specificity of each definitions were computed on percent change and the outcomes.
Results: A total of 1508 patients were analyzed. Mean age of participants was 59.43 (±1.12) years and 82.6% were males. Patient with AKI were older and more likely to be diabetic and hypertensive. Their perfusion and cross clamp time and morbidities were higher than their counterpart. Incidence of AKI was 33.7%, 34.4% and 57.5% on AKIN, KDIGO and RIFLE (based on change in eGFR) respectively. Area under the curve for 30 day mortality was AKIN: [0.786 (0.764 to 0.806)], KDIGO: [0.796 (0.775 to 0.816)], and RIFLE [0.844 (0.825 to 0.862)]. However discrimination power for morbidity was <0.7 and was undesirable.
Conclusions: AKIN and KDIGO are comparable to estimate AKI, while RIFLE (eGFR based) definition though overestimates the incidence of AKI, however has excellent discriminatory power to predict mortality compared to other definitions. Since eGFR provides age and gender adjusted estimates of AKI rather absolute change in renal function over the course of recovery phase, it should be integrated for a subset of population undergoing surgical intervention.


Pagination are not provided by the author/publisher

Publication ( Name of Journal)

Nephrology Dialysis Transplantation