background
Acute kidney injury (AKI) is a common comorbidity of critically ill patients, which refers to a clinical syndrome characterized by a rapid decrease in renal excretory function, with the accumulation of products of nitrogen metabolism such as urea and creatinine and other clinical unmeasured waste products. According to research statistics, AKI causes 2 million deaths per year, and 50% of critically ill patients develop AKI.When kidney injury occurred, serum creatinine due to the kidney's own compensation mechanism, could not reflect sensitively the change of kidney injury and would lag behind the progress of AKI. The urine volume would be affected by clinical drugs, diuretics, and so on, which also could not truly reflect the progress of the kidney injury.Therefore, it was urgent and significant to find early sensitive and specific diagnostic biomarkers or models to improve the long-term survival rate and postoperative recovery of patients with AKI.
parameter description
α1MG（mg/g.Cr）: The ratio of urine α1MG concentration（mg/ml）to urine creatinine (g/ml)
L-FABP（μg/g.Cr）: The ratio of urine L-FABP concentration（μg/ml）to urine creatinine (g/ml)
IGFBP7（μg/g.Cr）: The ratio of urine IGFBP7 concentration（μg/ml）to urine creatinine (g/ml)
Logistics Regression Model Introduction:
Logistic Regression is a statistical learning method used to address classification problems, particularly widely applied in binary classification. Despite the term 'regression' in its name, it is, in fact, a classification algorithm used to predict the probability of an observation belonging to a certain category.
The expression equation of our novel model, U-AKIpredTM as follows:
The equation of P(AKI) = ez/(1+ez) was the probability of AKI in a patient.