30 September 2021 in JMIR Medical Informatics
Background: Intraoperative hypotension has an adverse impact on postoperative outcomes. However, it is difficult to predict and treat intraoperative hypotension in advance according to individual clinical parameters. Objective: The aim of this study was to develop a prediction model to forecast 5-minute intraoperative hypotension based on the weighted average ensemble of individual neural networks, utilizing the biosignals recorded during noncardiac surgery. Methods: In this retrospective observational study, arterial waveforms were recorded during noncardiac operations performed between August 2016 and December 2019, at Seoul National University Hospital, Seoul, South Korea. We analyzed the arterial waveforms from the big data in the VitalDB repository of electronic health records. We defined 2s hypotension as the moving average of arterial pressure under 65 mmHg for 2 seconds, and intraoperative hypotensive events were defined when the 2s hypotension lasted for at least 60 seconds. We developed an artificial intelligence–enabled process, named short-term event prediction in the operating room (STEP-OP), for predicting short-term intraoperative hypotension. Results: The study was performed on 18,813 subjects undergoing noncardiac surgeries. Deep-learning algorithms (convolutional neural network [CNN] and recurrent neural network [RNN]) using raw waveforms as input showed greater area under the precision-recall curve (AUPRC) scores (0.698, 95% CI 0.690-0.705 and 0.706, 95% CI 0.698-0.715, respectively) than that of the logistic regression algorithm (0.673, 95% CI 0.665-0.682). STEP-OP performed better and had greater AUPRC values than those of the RNN and CNN algorithms (0.716, 95% CI 0.708-0.723). Conclusions: We developed STEP-OP as a weighted average of deep-learning models. STEP-OP predicts intraoperative hypotension more accurately than the CNN, RNN, and logistic regression models. Trial Registration: ClinicalTrials.gov NCT02914444; https://clinicaltrials.gov/ct2/show/NCT02914444.
31 December 2021 in Kidney Research and Clinical Practice
Background: Hepatocyte growth factor (HGF)/cMet pathway is necessary for repair and regeneration following acute kidney injury (AKI). We evaluated the clinical potential of plasma HGF and soluble cMet as prognostic biomarkers for severe AKI requiring continu-ous renal replacement therapy (CRRT). Methods: One hundred thirty-six patients with severe AKI who participated in the VENUS (volume management under body composi-tion monitoring in critically ill patients on CRRT) trial between 2017 and 2019 were enrolled in this study. We investigated associa-tions between plasma HGF and cMet concentrations and all-cause mortality. Results: Plasma HGF and soluble cMet levels were positively correlated. Patients were divided into three groups based on their HGF and soluble cMet concentrations. The day D 0, D2, and D7 highest concentration HGF groups had significantly higher in-hospital mor-tality after adjusting for sex, body mass index, Acute Physiology and Chronic Health Evaluation II, and age-adjusted Charlson comor-bidity index score, especially on D7 (hazard ratio, 4.26; 95% confidence interval, 1.71-10.62; p = 0.002). D7 soluble cMet level was also associated with mortality. Receiver operating characteristic curve analysis indicated that D7 HGF and soluble cMet levels were best at predicting mortality. Addition of plasma HGF and soluble cMet to conventional prognostic indices significantly improved the predictive value for mortality on D7. However, plasma HGF and soluble cMet were not associated with fluid status. Conclusion: Plasma HGF and soluble cMet levels were significant predictors of the outcomes of severe AKI patients undergoing CRRT. There was no correlation between plasma HGF and soluble cMet levels and fluid balance.
20 October 2022 in Blood Purification
VIVENUS study investigators
VENUS study investigators
+7
Introduction: This prospective cohort study investigated the clinical role of circulating tumor necrosis factor receptor (cTNFR) levels as prognostic biomarkers in severe acute kidney injury (AKI) patients requiring continuous renal replacement therapy (CRRT). Methods: We enrolled 136 patients from 7 hospitals participating in the VENUS (VolumE maNagement Under body composition monitoring in critically ill patientS on CRRT) trial from July 2017 to October 2019. The levels of cTNFR1 and cTNFR2 were measured using plasma samples collected on days 0 (D0), 2 (D2), and 7 (D7). Patients were divided into high- and low-cTNFR groups based on their receptor concentrations. Results: D0 concentrations of cTNFR1 and cTNFR2 were positively correlated with one another (R2 = 0.37, p < 0.001). The high-cTNFR1 group displayed a higher in-hospital mortality rate than the low-TNFR1 group (p = 0.002). Moreover, the mortality rate was significantly higher in the high-TNFR1 group than in the low-TNFR1 group after adjusting for age, sex, and acute physiology, and chronic health evaluation II scores (hazard ratio 1.82, 95% confidence interval 1.09–3.03, p = 0.025). D2 and D7 cTNFR1 levels were also associated with in-hospital mortality; contrastingly, cTNFR2 levels were not associated with this outcome. Additionally, patients were divided into three groups according to the change in cTNFR levels from D0 to D2 (ΔcTNFR). Those in the highest ΔcTNFR tertile had a higher mortality rate than the remaining patients (p = 0.033 for ΔcTNFR1; p = 0.025 for ΔcTNFR2). Patients who underwent AKI-to-chronic kidney disease transition had higher concentrations of cTNFR1 (p = 0.014). Discussion/Conclusion: Plasma cTNFR1 concentrations at CRRT initiation and changes in cTNFR1 and 2 levels immediately following CRRT initiation are significant biomarkers for predicting the outcomes of patients with severe AKI.
10 September 2021 in PLOS ONE
[This corrects the article DOI: 10.1371/journal.pone.0239126.].
24 June 2021 in JMIR Preprints
BACKGROUND Intraoperative hypotension has an adverse impact on postoperative outcomes. However, it is difficult to predict and treat intraoperative hypotension in advance according to individual clinical parameters. OBJECTIVE The aim of this study was to develop a prediction model to forecast 5-minute intraoperative hypotension based on the weighted average ensemble of individual neural networks, utilizing the biosignals recorded during noncardiac surgery. METHODS In this retrospective observational study, arterial waveforms were recorded during noncardiac operations performed between August 2016 and December 2019, at Seoul National University Hospital, Seoul, South Korea. We analyzed the arterial waveforms from the big data in the VitalDB repository of electronic health records. We defined 2s hypotension as the moving average of arterial pressure under 65 mmHg for 2 seconds, and intraoperative hypotensive events were defined when the 2s hypotension lasted for at least 60 seconds. We developed an artificial intelligence–enabled process, named short-term event prediction in the operating room (STEP-OP), for predicting short-term intraoperative hypotension. RESULTS The study was performed on 18,813 subjects undergoing noncardiac surgeries. Deep-learning algorithms (convolutional neural network [CNN] and recurrent neural network [RNN]) using raw waveforms as input showed greater area under the precision-recall curve (AUPRC) scores (0.698, 95% CI 0.690-0.705 and 0.706, 95% CI 0.698-0.715, respectively) than that of the logistic regression algorithm (0.673, 95% CI 0.665-0.682). STEP-OP performed better and had greater AUPRC values than those of the RNN and CNN algorithms (0.716, 95% CI 0.708-0.723). CONCLUSIONS We developed STEP-OP as a weighted average of deep-learning models. STEP-OP predicts intraoperative hypotension more accurately than the CNN, RNN, and logistic regression models. CLINICALTRIAL ClinicalTrials.gov NCT02914444; https://clinicaltrials.gov/ct2/show/NCT02914444.