Search

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. 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. 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. 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. [ABSTRACT FROM AUTHOR]

Background: The endotracheal cuff pressure depends on the airway pressure during positive-pressure ventilation. A high endotracheal cuff pressure may be related to intraoperative coughing, which can be detrimental during neurosurgery. We investigated the incidence of intraoperative coughing and its association with peak inspiratory pressure (PIP) during neurosurgery under general anesthesia without neuromuscular blockade. Methods: This retrospective study divided 1656 neurosurgical patients who underwent total intravenous anesthesia without additional neuromuscular blockade after tracheal intubation into high (PIP > 21.6 cmH2O, n = 318) and low (PIP ≤ 21.6 cmH2O, n = 1338) PIP groups. After propensity score matching, 206 patients were selected in each group. Demographic, preoperative, surgical, and anesthetic data were collected retrospectively from electronic medical records and continuous ventilator, infusion pump, and bispectral index data from a data registry. Results: Intraoperative coughing occurred in 30 (1.8%) patients, including 9 (0.5%) during the main surgical procedure. Intraoperative coughing was more frequent in the high PIP group than in the low PIP group before (14/318 [4.4%] vs. 16/1338 [1.2%], P < 0.001) and after (13/206 [6.3%] vs. 1/206 [0.5%], P = 0.003) propensity score matching. In multivariable logistic regression analysis after propensity score matching, a high PIP (odds ratio [95% confidence interval] 14.22 [1.81-111.73], P = 0.012), tidal volume divided by predicted body weight (mL/kg, 1.36 [1.09–1.69], P = 0.006), and surgical duration (min, 1.01 [1.00–1.01], P = 0.025) predicted intraoperative coughing. Conclusion: The incidence of intraoperative coughing was 1.8% in neurosurgical patients undergoing general anesthesia without neuromuscular blockade and might be associated with a high PIP. [ABSTRACT FROM AUTHOR]

Background: The endotracheal cuff pressure depends on the airway pressure during positive-pressure ventilation. A high endotracheal cuff pressure may be related to intraoperative coughing, which can be detrimental during neurosurgery. We investigated the incidence of intraoperative coughing and its association with peak inspiratory pressure (PIP) during neurosurgery under general anesthesia without neuromuscular blockade. Methods: This retrospective study divided 1656 neurosurgical patients who underwent total intravenous anesthesia without additional neuromuscular blockade after tracheal intubation into high (PIP > 21.6 cmH2O, n = 318) and low (PIP ≤ 21.6 cmH2O, n = 1338) PIP groups. After propensity score matching, 206 patients were selected in each group. Demographic, preoperative, surgical, and anesthetic data were collected retrospectively from electronic medical records and continuous ventilator, infusion pump, and bispectral index data from a data registry. Results: Intraoperative coughing occurred in 30 (1.8%) patients, including 9 (0.5%) during the main surgical procedure. Intraoperative coughing was more frequent in the high PIP group than in the low PIP group before (14/318 [4.4%] vs. 16/1338 [1.2%], P < 0.001) and after (13/206 [6.3%] vs. 1/206 [0.5%], P = 0.003) propensity score matching. In multivariable logistic regression analysis after propensity score matching, a high PIP (odds ratio [95% confidence interval] 14.22 [1.81-111.73], P = 0.012), tidal volume divided by predicted body weight (mL/kg, 1.36 [1.09–1.69], P = 0.006), and surgical duration (min, 1.01 [1.00–1.01], P = 0.025) predicted intraoperative coughing. Conclusion: The incidence of intraoperative coughing was 1.8% in neurosurgical patients undergoing general anesthesia without neuromuscular blockade and might be associated with a high PIP. [ABSTRACT FROM AUTHOR]

Background: Clinical outcome after unplanned extubation (UE) in patients admitted to the surgical intensive care unit (SICU) has not been fully investigated. In this study we assessed in-hospital mortality of patients with UE and determined whether UE is a predictor of in-hospital mortality. Finally, we sought to identify predictors of reintubation after UE in mechanically ventilated patients in the SICU. Methods: Medical charts of patients ( n = 4,407) admitted to the SICU between October 2007 and December 2011 were reviewed retrospectively. Results: Eighty-five episodes of UE occurred in 81 patients. Patients with UE required emergency surgery more frequently and had higher ICU and hospital mortality rates, reintubation rate, and APACHE II scores and longer mechanical ventilation (MV) and ICU stay than patients without UE ( P < 0.05 for all associations). Multivariate analysis revealed that reintubation (odds ratio [95 % confidence interval]: 4.14 [2.58-6.67]; P < 0.001), APACHE II scores (1.14 [1.12-1.17]; P < 0.001), emergency surgery (1.73 [1.18-2.53]; P = 0.005), and chronic neurologic disease (2.11 [1.30-3.41]; P = 0.002) were associated with hospital mortality. Reintubation was necessary in 17 patients. On multivariate analysis, a score on the Richmond Agitation-Sedation Scale (RASS, 0.48 [0.31-0.76]; P = 0.001), PaO/FiO ratio (0.99 [0.99-1.00]; P = 0.048), and MV duration before UE (1.46 [1.08-1.98]; P = 0.014) were independently associated with reintubation after UE. Conclusions: Our results indicated that although patients with UE had high in-hospital mortality, UE was not directly associated with in-hospital mortality. Reintubation, chronic neurologic disease, emergency operation, and higher APACHE II score were related to increased in-hospital mortality. A low RASS score, a low PaO/FiO ratio, and long MV duration before UE were related to reintubation after UE. [ABSTRACT FROM AUTHOR]