Development and validation of a machine learning model to predict postoperative delirium using a nationwide database: A retrospective, observational study.

STUDY OBJECTIVE: Postoperative delirium is a neuropsychological syndrome that typically occurs in surgical patients. Its onset can lead to prolonged hospitalization as well as increased morbidity and mortality. Therefore, it is important to promptly identify its signs. This study aimed to develop and validate a machine learning predictive model for postoperative delirium using extensive population data. DESIGN: Retrospective observational study. SETTING: Japanese Diagnosis Procedure Combination inpatient data. Data were used for internal (2016.4-2018.12) and temporal validation (2019.01-2019.10). PATIENTS: Patients aged ≥65 years who underwent general anesthesia for surgical procedure. MEASUREMENTS: The primary outcome was postoperative delirium, which was defined as a condition requiring newly prescribed antipsychotic drugs or assignment of the corresponding insurance claim code after the date of surgery. We trained and tuned the optimal machine-learning model through 10-fold cross-validation using the selected optimal area under the receiver operating characteristic curve (AUC) value. In the temporal validation, we measured the performance of our model. MAIN RESULTS: The analysis included 557,990 patients. The light-gradient boosting machine models showed a higher AUC value (0.826 [95% confidence interval (CI): 0.822-0.829]) than the other models. Regarding performance, the model had a recall value of 0.124 (95% CI: 0.119-0.129) and precision value of 0.659 (95% CI: 0.641-0.677]). This performance was sustained in the temporal validation (AUC, 0.815 [95% CI: 0.811-0.818]). At a sensitivity of 0.80, the model achieved a specificity of 0.672 (95% CI: 0.670-0.674]), a negative predictive value of 0.975 (95% CI: 0.974-0.975), and a positive predictive value of 0.176 (95% CI: 0.176-0.179). CONCLUSIONS: Using extensive Diagnostic Procedure Combination data, we successfully created and validated a machine learning model for predicting postoperative delirium. This model may facilitate prediction of postoperative delirium.

Preoperative echocardiography predictive analytics for postinduction hypotension prediction.

PURPOSE: Hypotension is a risk factor for adverse perioperative outcomes. Preoperative transthoracic echocardiography has been extended for preoperative risk assessment before noncardiac surgery. This study aimed to develop a machine learning model to predict postinduction hypotension risk using preoperative echocardiographic data and compared it with conventional statistic models. We also aimed to identify preoperative echocardiographic factors that cause postinduction hypotension. METHODS: In this retrospective observational study, we extracted data from electronic health records of patients aged >18 years who underwent general anesthesia at a single tertiary care center between April 2014 and September 2019. Multiple supervised machine learning classification techniques were used, with postinduction hypotension (mean arterial pressure <55 mmHg from intubation to the start of the procedure) as the primary outcome and 95 transthoracic echocardiography measurements as factors influencing the primary outcome. Based on the mean cross-validation performance, we used 10-fold cross-validation with the training set (70%) to select the optimal hyperparameters and architecture, assessed ten times using a separate test set (30%). RESULTS: Of 1,956 patients, 670 (34%) had postinduction hypotension. The area under the receiver operating characteristic curve using the deep neural network was 0.72 (95% confidence interval (CI) = 0.67-0.76), gradient boosting machine was 0.54 (95% CI = 0.51-0.59), linear discriminant analysis was 0.56 (95% CI = 0.51-0.61), and logistic regression was 0.56 (95% CI = 0.51-0.61). Variables of high importance included the ascending aorta diameter, transmitral flow A wave, heart rate, pulmonary venous flow S wave, tricuspid regurgitation pressure gradient, inferior vena cava expiratory diameter, fractional shortening, left ventricular mass index, and end-systolic volume. CONCLUSION: We have created developing models that can predict postinduction hypotension using preoperative echocardiographic data, thereby demonstrating the feasibility of using machine learning models of preoperative echocardiographic data for produce higher accuracy than the conventional model.

Comparison of total intravenous with inhalational anesthesia in terms of postoperative delirium and complications in older patients: a nationwide population-based study.

PURPOSE: Postoperative delirium incidences are increasing in older adults. A Cochrane Review found no significant difference in the incidence of postoperative delirium between total intravenous anesthesia (TIVA) and inhalational anesthesia (IA). This study evaluated the differences in postoperative delirium and morbidity between patients who underwent either TIVA or IA. METHODS: A nationwide Japanese inpatient database was used to retrospectively compare differences in postoperative delirium and composite morbidity between patients older than 65 years, who underwent general anesthesia (TIVA or IA). The primary outcome was postoperative delirium. The secondary outcomes were: morbidity incidence, length of hospital stay, and mortality. A 1:3 propensity score analysis of patients who underwent all surgical procedures was conducted according to covariates, to calculate odds ratios and their 95% confidence intervals (CIs). Sensitivity analyses were conducted using an instrumental variable analysis of the proportion of TIVA by hospital scale, stabilized inverse probability of treatment weighting analyses, limiting the definitions of postoperative delirium, and subgroup analysis. RESULTS: Of 738,600 patients, 149,540 received TIVA and 589,060 received IA. After 1:3 propensity score matching, the adjusted odds ratios for postoperative delirium and composite morbidity were 0.93 (95% CI 0.91-0.95) and 0.94 (95% CI 0.90-0.97), respectively, for TIVA concerning IA. There were no differences in the length of the intensive care unit and hospital stay, or hospital mortality. The findings were consistent with the sensitivity analyses. CONCLUSIONS: This study demonstrated that TIVA was related to a slightly decreased postoperative delirium and incidence of morbidity compared to IA.

Comparing the effects of peripheral nerve block and general anesthesia with general anesthesia alone on postoperative delirium and complications in elderly patients: a retrospective cohort study using a nationwide database.

INTRODUCTION: The difference between the effects of peripheral nerve block (PNB) with general anesthesia (GA) and GA alone on the patients' postoperative clinical outcomes remains unknown. We assessed whether there is a difference in postoperative delirium and composite morbidity between patients receiving GA with PNB and GA alone using a national clinical database in Japan. METHODS: We compared the outcomes of patients receiving GA with PNB and GA alone from April 2016 to October 2019. The primary outcome was postoperative delirium, defined as a status requiring newly prescribed antipsychotic drugs or that given the code of a reimbursable disease after the surgery date. The secondary outcome was morbidity incidence as the occurrence of at least one of any of the following life-threatening complications. We conducted propensity score-matched analyses using covariates for patients who underwent any surgical procedure. We used instrumental variables and restricted the definition of postoperative delirium and subgroup for sensitivity analyses. RESULTS: Of 653,759 patients, 90,358 received GA-PNB and 563,401 received only GA. After 1:4 propensity score matching, 89,754 patients were included in the GA-PNB and 359,015 in the GA. The adjusted ORs for postoperative delirium and composite morbidity were 0.96 (95% CIs 0.94 to 0.99; p<0.01), 0.80 (95% CIs 0.76 to 0.83; p<0.001), respectively, for the GA-PNB concerning the GA. For sensitivity analyses, findings were also consistent with instrumental variables and subgroup analyses. DISCUSSION: This retrospective, nationwide cohort study demonstrated that GA-PNB was associated with a small reduction in the likelihood of postoperative delirium and a moderate reduction in the likelihood of composite morbidity.

Skin temperature changes after ultrasound-guided supra-inguinal fascia iliaca block: a prospective observational study.

PURPOSE: Ultrasound-guided supra-inguinal fascia iliaca block (SFIB) is widely used as regional anesthesia of the hip and thigh. It is difficult to judge the blocking effect and the spreading local anesthesia. We hypothesize that the effect and spread of the block could be proven objectively by a rise in the temperature. In this prospective observational study, the broad regional rise in skin temperature of twenty patients who were scheduled for hip surgery was measured using an infrared thermographic camera at multiple intervals following ultrasound-guided SFIB. METHODS: Infrared thermographic imaging of skin temperature at the femoral, obturator, and lateral femoral cutaneous nerve sites was performed before and at 5-min intervals after ultrasound-guided SFIB for up to 15-min post-injection. The primary outcomes are skin surface temperature. Sensory block was assessed immediately after the final infrared thermographic image acquisition using the cold test. RESULTS: Compared to pre-injection baseline, temperature increased by 1.2 °C [95% confidence interval (CI) 0.4-2.0 °C] after 5 min, 1.2 °C (95% CI 0.4-2.0 °C) after 10 min, and 0.9 °C (95% CI 0.4-2.1°C) after 15 min. The cold test response was reduced in all cases at the femoral and lateral femoral cutaneous nerve sites and in 13 cases at the obturator nerve site. The sensitivity and specificity of the temperature increase to cold loss were 96% and 63%, respectively when we defined >0°C as the clinical threshold. CONCLUSIONS: Successful SFIB significantly enhanced skin temperature at the hip and thigh in all cases, suggesting that infrared surface thermography can be used as an objective assessment tool for adequate analgesia. TRIAL REGISTRATION: University Hospital Medical Information Network Clinical Trials Registry ( UMIN 000037866 ). Registered 31 August 2019.

[A case of severe embolic complications due to warfarin withdrawal].

We report a case of three severe embolic complications due to warfarin withdrawal. An 83-year-old man with hypertension, angina pectoris and atrial fibrillation underwent bladder biopsy under spinal anesthesia after 13 days of warfarin withdrawal. On the second postoperative day, the patient complained of chest pain and was diagnosed as acute myocardial infarction. Embolus was successfully removed by suctioning. Warfarin and heparin therapy was started after that. On the 6th postoperative day, the patient complained of abdominal pain and was diagnosed as superior mesenteric artery embolism. After suctioning of the thrombus and monteplase injection, symptoms disappeared. On the 9th postoperative day, paralysis on the right side of his body and aphasia appeared. Stroke was suspected. Coma advanced day by day and he died due to brain herniation on the 16th postoperative day. In this patient we should have assessed the risk of the thromboembolic complication and planned the appropriate anticoagulation with closer cooperation with his attending physicians.