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.

IntroductionFever is frequently observed in critically ill patients. An independent association of fever with increased mortality has been observed in non-neurological critically ill patients with mixed febrile etiology. The association of fever and antipyretics with mortality, however, may be different between infective and non-infective illness.MethodsWe designed a prospective observational study to investigate the independent association of fever and the use of antipyretic treatments with mortality in critically ill patients with and without sepsis. We included 1,425 consecutive adult critically ill patients (without neurological injury) requiring > 48 hours intensive care admitted in 25 ICUs. We recorded four-hourly body temperature and all antipyretic treatments until ICU discharge or 28 days after ICU admission, whichever occurred first. For septic and non-septic patients, we separately assessed the association of maximum body temperature during ICU stay (MAXICU) and the use of antipyretic treatments with 28-day mortality.ResultsWe recorded body temperature 63,441 times. Antipyretic treatment was given 4,863 times to 737 patients (51.7%). We found that treatment with non-steroidal anti-inflammatory drugs (NSAIDs) or acetaminophen independently increased 28-day mortality for septic patients (adjusted odds ratio: NSAIDs: 2.61, P = 0.028, acetaminophen: 2.05, P = 0.01), but not for non-septic patients (adjusted odds ratio: NSAIDs: 0.22, P = 0.15, acetaminophen: 0.58, P = 0.63). Application of physical cooling did not associate with mortality in either group. Relative to the reference range (MAXICU 36.5°C to 37.4°C), MAXICU ≥ 39.5°C increased risk of 28-day mortality in septic patients (adjusted odds ratio 8.14, P = 0.01), but not in non-septic patients (adjusted odds ratio 0.47, P = 0.11).ConclusionsIn non-septic patients, high fever (≥ 39.5°C) independently associated with mortality, without association of administration of NSAIDs or acetaminophen with mortality. In contrast, in septic patients, administration of NSAIDs or acetaminophen independently associated with 28-day mortality, without association of fever with mortality. These findings suggest that fever and antipyretics may have different biological or clinical or both implications for patients with and without sepsis.Trial registrationClinicalTrials.gov: NCT00940654

Use of the LMA in smaller children results in more airway obstruction, higher ventilatory pressures, larger inspiratory leak, and more complications than in older children.

SummaryWe performed a randomised comparison of the i-gel TM and the Laryngeal Mask Airway (LMA) Classic TM in children aged less than a year who were undergoing general anaesthesia for elective surgery. Fifty-four infants were randomly allocated to either the i-gel or the LMA Classic. We measured performance characteristics, fibreoptic views through the device and complications. Success rate at first insertion attempt was 100% (27/27) in the i-gel group compared with 88% (23/26) in the LMA Classic group. Insertion of the device was considered easy in 26/27 (96%) patients in the i-gel group compared with 18/26 (69%) patients in the LMA Classic group (p = 0.009). There were no differences between the groups in insertion times, fibreoptic views through the device, airway leak pressures or complications. We conclude that the i-gel was considered easier to insert than the LMA Classic in infants.