Multicentre validation of a machine learning model for predicting respiratory failure after noncardiac surgery.

BACKGROUND: Postoperative respiratory failure is a serious complication that could benefit from early accurate identification of high-risk patients. We developed and validated a machine learning model to predict postoperative respiratory failure, defined as prolonged (>48 h) mechanical ventilation or reintubation after surgery. METHODS: Easily extractable electronic health record (EHR) variables that do not require subjective assessment by clinicians were used. From EHR data of 307,333 noncardiac surgical cases, the model, trained with a gradient boosting algorithm, utilised a derivation cohort of 99,025 cases from Seoul National University Hospital (2013-9). External validation was performed using three separate cohorts A-C from different hospitals comprising 208,308 cases. Model performance was assessed by area under the receiver operating characteristic (AUROC) curve and area under the precision-recall curve (AUPRC), a measure of sensitivity and precision at different thresholds. RESULTS: The model included eight variables: serum albumin, age, duration of anaesthesia, serum glucose, prothrombin time, serum creatinine, white blood cell count, and body mass index. Internally, the model achieved an AUROC of 0.912 (95% confidence interval [CI], 0.908-0.915) and AUPRC of 0.113. In external validation cohorts A, B, and C, the model achieved AUROCs of 0.879 (95% CI, 0.876-0.882), 0.872 (95% CI, 0.870-0.874), and 0.931 (95% CI, 0.925-0.936), and AUPRCs of 0.029, 0.083, and 0.124, respectively. CONCLUSIONS: Utilising just eight easily extractable variables, this machine learning model demonstrated excellent discrimination in both internal and external validation for predicting postoperative respiratory failure. The model enables personalised risk stratification and facilitates data-driven clinical decision-making.

Intraoperative Normal Saline Administration and Acute Kidney Injury in Patients Undergoing Liver Transplantation.

BACKGROUND: Normal saline is still used in patients undergoing living donor liver transplantation (LDLT) with normonatremia. We investigated whether the normal saline administered during LDLT is associated with the increased risk of acute kidney injury (AKI) compared with the balanced crystalloids. METHODS: We reviewed 1011 cases undergoing LDLT. The primary exposure variable was normal saline administered intraoperatively compared with the balanced crystalloid. To compare the risk of AKI after adjusting for potential confounders of baseline characteristics and surgical parameters, a propensity score matching analysis was performed. As a sensitivity analysis, ordinal logistic regression analysis was performed for AKI using inverse probability of treatment weighting (IPTW). RESULTS: The incidence of AKI was significantly higher in the saline group (n = 88/174, 50.6%) than in the balanced group (n = 67/174, 38.5%) after matching (P = .010). The incidence of stage 2 or 3 AKI was also significantly higher in the saline group (n = 26/174, 14.9%) than in the balanced group (n = 43/174, 24.7%) after matching (P = .022). The length of hospital stay was significantly longer in the saline group than in the balanced group after matching. Ordinal logistic regression analysis using IPTW showed that the saline group showed a significant association of saline administration with the risk of AKI (odds ratio 1.23, 95% CI 1.05-1.28, P = .013). CONCLUSION: Our propensity score analysis using propensity score matching and IPTW showed that normal saline administration during LDLT is associated with a high risk of postoperative AKI and longer hospital stays. However, our results should be interpreted carefully due to the relatively long period of data collection.