Acute kidney injury (AKI) after liver transplantation has been reported to be associated with increased mortality. Recently, machine learning approaches were reported to have better predictive ability than the classic statistical analysis. We compared the performance of machine learning approaches with that of logistic regression analysis to predict AKI after liver transplantation. We reviewed 1211 patients and preoperative and intraoperative anesthesia and surgery-related variables were obtained. The primary outcome was postoperative AKI defined by acute kidney injury network criteria. The following machine learning techniques were used: decision tree, random forest, gradient boosting machine, support vector machine, naïve Bayes, multilayer perceptron, and deep belief networks. These techniques were compared with logistic regression analysis regarding the area under the receiver-operating characteristic curve (AUROC). AKI developed in 365 patients (30.1%). The performance in terms of AUROC was best in gradient boosting machine among all analyses to predict AKI of all stages (0.90, 95% confidence interval [CI] 0.86–0.93) or stage 2 or 3 AKI. The AUROC of logistic regression analysis was 0.61 (95% CI 0.56–0.66). Decision tree and random forest techniques showed moderate performance (AUROC 0.86 and 0.85, respectively). The AUROC of support the vector machine, naïve Bayes, neural network, and deep belief network was smaller than that of the other models. In our comparison of seven machine learning approaches with logistic regression analysis, the gradient boosting machine showed the best performance with the highest AUROC. An internet-based risk estimator was developed based on our model of gradient boosting. However, prospective studies are required to validate our results.

Son YG, et al. J Dent Anesth Pain Med. 2017 Mar;17(1):1-12. https://doi.org/10.17245/jdapm.2017.17.1.1

Jeon K, et al. J Korean Med Sci. 2019 Mar;34(14):e110. https://doi.org/10.3346/jkms.2019.34.e110

Recent reports suggest that not all critically ill patients with delirium share the same consequences. The outcomes of surgical intensive care unit patients with postoperative delirium were evaluated depending on the onset and duration of delirium. A total of 527 patients who were admitted from the operating theater and cared for in the surgical intensive care unit for >24 hours were evaluated for delirium using the Confusion Assessment Method for intensive care unit, 3 times a day. Patients were analyzed according to the onset time and duration of delirium. Patients were classified into 4 groups according to the onset and duration of delirium: no delirium, early brief delirium (delirium for <1 day on postoperative day 0), late brief delirium (delirium for 24 hours (persistent). Persistent delirium was associated with longer intensive care unit (4.6 [1.1-53.3] vs 1.6 [1.1-37.5] days) and hospital duration of stay (24 [3-112] vs 16 [2-225] days) and higher hospital mortality (14.5% vs 2.2%) compared to no delirium (P < .01). For postoperative intensive care unit patients, intensive care unit and hospital duration of stay did not seem to differ between patients with early brief delirium or no delirium, whereas patients with late brief or persistent delirium seemed to show longer intensive care unit and hospital duration of stay and higher mortality.