An acute kidney injury prediction nomogram based on neurosurgical intensive care unit profiles

Acute kidney injury (AKI) is a common and serious complication with high mortality within the neural-critical care unit, and can limit the treatment of osmotic diuresis and body fluid equilibrium. Given its seriousness, it is necessary to find a tool to predict the likelihood of AKI and to prevent its occurrence.

In this retrospective study, patients' clinical profiles, laboratory test results, and doctors' prescriptions were collected. Least absolute shrinkage and selection operator (LASSO) regression was used to select variables, and a logistic regression model was then applied to find independent risk factors for AKI. Based on the results of multivariate analysis, we established a nomogram to evaluate the probability of AKI, which was verified through the use of a receiver operating characteristic (ROC) curve and its calibration curves.

Risk factors given by logistic regression were Glasgow Coma Scale (GCS) classification (1.593; 95% CI: 0.995-2.549; P=0.0523), coefficient of variation (CV) of GCS (1.017; 95% CI: 0.995-1.04; P=0.1367), hypertension (2.238; 95% CI: 1.124-4.456; P=0.0219), coronary heart disease (2.924; 95% CI: 1.2-7.126; P=0.0182), pneumonia within 7 days (3.032; 95% CI: 1.511-6.085; P=0.0018), heart failure within 7 days (6.589; 95% CI: 2.235-19.42; P=0.0006), furosemide (1.011; 95% CI: 1.005-1.016; P<0.0001), torasemide (1.028; 95% CI: 0.976-1.082; P=0.297), dopamine (1; 95% CI: 1-1.001, P=0.3297), and norepinephrine (1.007; 95% CI: 1-1.015; P=0.0474). The area under the curve (AUC) of the prediction model was 0.8786, and the calibration curves showed that the model had a good ability to predict AKI occurrence.

This study presents an AKI prediction nomogram based on LASSO, logistic regression, and clinical risk factors. The clinical use of the nomogram may allow for the timely detection of AKI occurrence and thus improve the prognosis of patients.