Abstract: To analyze the influencing factors for the duration of continuous renal replacement therapy (CRRT) in the treatment of acute kidney injury (AKI) patients, and to establish a model to predict the duration of CRRT in AKI patients.  Methods  A total of 102 AKI patients treated in the Intensive Care Unit (ICU) of Jiangsu Province Hospital from April 2018 to January 2020 were retrospectively analyzed. They were divided into <6 days group and ≥6 days group based on the median duration of CRRT [6.0 (2.3, 10.4) d]. Univariate analysis and binary logistic regression were used to find out the independent influencing factors for CRRT duration. The R software was then used to construct a prediction nomogram model, and the prediction value of this model was evaluated.  Results  Among the 102 AKI patients, 53 cases were in the <6 days group and 49 cases were in the ≥6 days group. Univariate analysis indicated that α1-MG (t=2.009,          P=0.047), β2-MG (t=2.402, P=0.018), Scr (t=2.395, P=0.019) and neutrophil gelatinase-associated lipocalin (NGAL; t=2.030, P=0.045) were lower in the <6 days group than in the ≥6 days, while mean arterial pressure (MAP; t=2.425, P=0.017) and urine volume (t=2.058, P=0.042) were higher in the <6 days group than in the ≥6 days. Binary logistic regression showed that urine volume (OR=1.131, 95% CI: 1.017~1.024), NGAL (OR=0.987, 95% CI: 0.982~0.993), α1-MG (OR=0.860, 95% CI: 0.767~0.976) and β2-MG (OR=0.755, 95% CI: 0.572~0.996) were the independent predictors for CRRT duration in AKI patients (P<0.05). The C-index of the nomogram model was 0.894 (95% CI: 0.723~0.896). The calibration curve showed that the average absolute error of prediction possibility was 0.029, indicating a better discrimination and accuracy of the prediction model.  Conclusion  The nomogram model constructed on the differences of urine volume, NGAL, α1-MG and β2-MG can better predict the duration of CRRT in AKI patients, and provide useful references for CRRT treatment in AKI patients.

Key words: Acute kidney injury, Continuous renal replacement therapy, Treatment duration, Prediction model, Nomogram