A multimodal stacked ensemble model for cardiac output prediction utilizing cardiorespiratory interactions during general anesthesia
Ontology highlight
ABSTRACT:
SUBMITTER: Dervishi A
PROVIDER: S-EPMC10980739 | biostudies-literature | 2024 Jan
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Multimodal general anesthesia (MMGA) is a strategy that utilizes the well-known neuroanatomy and neurophysiology of nociception and arousal control in designing a rational and clinical practical paradigm to regulate the levels of unconsciousness and antinociception during general anesthesia while mitigating side effects of any individual anesthetic. We sought to test the feasibility of implementing MMGA for seniors undergoing cardiac surgery, a high-risk cohort for hemodynamic instability, delirium, and post-operative cognitive dysfunction. Twenty patients aged 60 or older undergoing on-pump coronary artery bypass graft (CABG) surgery or combined CABG/valve surgeries were enrolled in this non-randomized prospective observational feasibility trial, wherein we developed MMGA specifically for cardiac surgeries. Antinociception was achieved by a combination of intravenous remifentanil, ketamine, dexmedetomidine, and magnesium together with bupivacaine administered as a pecto-intercostal fascial block. Unconsciousness was achieved by using electroencephalogram (EEG)-guided administration of propofol along with the sedative effects of the antinociceptive agents. EEG-guided MMGA anesthesia was safe and feasible for cardiac surgeries, and exploratory analyses found hemodynamic stability and vasopressor usage comparable to a previously collected cohort. Intraoperative EEG suppression events and postoperative delirium were found to be rare. We report successful use of a total intravenous anesthesia (TIVA)-based MMGA strategy for cardiac surgery and establish safety and feasibility for studying MMGA in a full clinical trial. Clinical Trial Number: www.clinicaltrials.gov; identifier NCT04016740 (https://clinicaltrials.gov/ct2/show/NCT04016740).
| S-EPMC8551571 | biostudies-literature
Project description:Artificial intelligence (AI) technology has huge scope in developing models to predict the survival rate of critically ill patients in the intensive care unit (ICU). The availability of electronic clinical data has led to the widespread use of various machine learning approaches in this field. Innovative algorithms play a crucial role in boosting the performance of models. This study uses a stacked ensemble model to predict mortality in ICU by incorporating the clinical severity scoring results, in which several machine learning algorithms are employed to compare the performance. The experimental results show that the stacked ensemble model achieves good performance compared with the model without integrating the severity scoring results, which has the area under curve (AUC) of 0.879 and 0.862, respectively. To improve the performance of prediction, two feature subsets are obtained based on different feature selection techniques, labeled as SetS and SetT. Evaluation performances show that the SEM based on the SetS achieves a higher AUC value (0.879 and 0.860). Finally, the SHapley Additive exPlanations (SHAP) analysis is employed to interpret the correlation between the risk features and the outcome.
| S-EPMC9722283 | biostudies-literature
Project description:BackgroundCircadian rhythm involved with physiology has been reported to affect pharmacokinetics or pharmacodynamics. We hypothesized that circadian variations in physiology disturb anesthesia and eventually affect recovery after anesthesia.MethodsA retrospective cohort study initially included 107,406 patients (1 June 2016-6 June 2021). Patients were classified into morning or afternoon surgery groups. The primary outcome was daytime variation in PACU (post-anesthesia care unit) recovery time and Steward score. Inverse probability weighting (IPW) approach based on propensity score and univariable/multivariable linear regression were used to estimate this outcome.ResultsOf 28,074 patients, 13,418 (48%) patients underwent morning surgeries, and 14,656 (52%) patients underwent afternoon surgeries. LOWESS curves and IPW illustrated daytime variation in PACU recovery time and Steward score. Before adjustment, compared to morning surgery group, afternoon surgery group had less PACU recovery time (median [interquartile range], 57 [46, 70] vs. 54 [43, 66], p < 0.001) and a higher Steward score (5.62 [5.61, 5.63] vs. 5.66 [5.65, 5.67], p < 0.001). After adjustment, compared to morning surgery group, afternoon surgery group had less PACU recovery time (58 [46, 70] vs. 54 [43, 66], p < 0.001). In multivariable linear regression, morning surgery is statistically associated with an increased PACU recovery time (coefficient, -3.20; 95% confidence interval, -3.55 to -2.86).Among non-cardiac surgeries, daytime variation might affect recovery after general anesthesia. These findings indicate that the timing of surgery improves recovery after general anesthesia, with afternoon surgery providing protection.KEY MESSAGESIn this retrospective cohort study of 28,074 participants, the afternoon surgery group has a higher Steward score than the morning surgery group.In multivariable linear regression, morning surgery is statistically associated with an increased PACU recovery time.Among non-cardiac surgeries, daytime variation affects the recovery after general anesthesia, with afternoon surgery providing protection.
| S-EPMC10035960 | biostudies-literature
Project description:ObjectivesCardiac output (CO) measurements in the ICU are usually based on invasive techniques, which are technically complex and associated with clinical complications. This study aimed to compare CO measurements obtained from a noninvasive photoplethysmography-based device to a pulse contour cardiac output device in ICU patients.DesignObservational, prospective, comparative clinical trial.SettingSingle-center general ICU.PatientsPatients admitted to the general ICU monitored using a pulse contour cardiac output device as per the decision of the attending physician.InterventionsParallel monitoring of CO using a photoplethysmography-based chest patch device and pulse contour cardiac output while the medical team was blinded to the values obtained by the noninvasive device.Measurements and main resultsSeven patients (69 measurements) were included in the final analysis. Mean CO were 7.3 ± 2.0 L/m and 7.0 ± 1.5 L/m for thermodilution and photoplethysmography, respectively. Bland-Altman showed that the photoplethysmography has a bias of 0.3 L/m with -1.6 and 2.2 L/m 95% limit of agreement (LOA) and a bias of 2.4% with 95% LOA between -25.7% and 30.5% when calculating the percentage of difference from thermodilution. The values obtained by thermodilution and photoplethysmography were highly correlated (r = 0.906).ConclusionsThe tested chest patch device offers a high accuracy for CO compared to data obtained by the pulse contour cardiac output and the thermodilution method in ICU patients. Such devices could offer advanced monitoring capabilities in a variety of clinical settings, without the complications of invasive devices.
| S-EPMC8812679 | biostudies-literature
Project description:BackgroundThe optimal type of anesthesia for acute vertebrobasilar artery occlusion (VBAO) remains controversial. We aimed to assess the influence of anesthetic management on the outcomes in VBAO patients received endovascular treatment (EVT).MethodsPatients underwent EVT for acute VBAO at 21 stroke centers in China were retrospectively enrolled and compared between the general anesthesia (GA) group and non-GA group. The primary outcome was the favorable outcome, defined as a modified Rankin Scale (mRS) score 0-3 at 90 days. Secondary outcomes included functional independence (90-day mRS score 0-2), and the rate of successful reperfusion. The safety outcomes included all-cause mortality at 90 days, the occurrence of any procedural complication, and the rate of symptomatic intracranial hemorrhage (sICH). In addition, we performed analyses of the outcomes in subgroups that were defined by Glasgow Coma Scale (GCS) score (≤8 or >8).ResultsIn the propensity score matched cohort, there were no difference in the primary outcome, secondary outcomes and safety outcomes between the two groups. Among patients with a GCS score of 8 or less, the proportion of successful reperfusion was significantly higher in the GA group than the non-GA group (aOR, 3.57, 95% CI 1.06-12.50, p = 0.04). In the inverse probability of treatment weighting-propensity score-adjusted cohort, similar results were found.ConclusionsPatients placed under GA during EVT for VBAO appear to be as effective and safe as non-GA. Furthermore, GA might yield better successful reperfusion for worse presenting GCS score (≤8).RegistrationURL: http://www.chictr.org.cn/; Unique identifier: ChiCTR2000033211.
| S-EPMC9932259 | biostudies-literature
Project description:Our study examined the disaggregation of inflation components in Nigeria using the stacked ensemble approach, a machine learning algorithm capable of compensating the weakness of an ensemble and a base learner with the strength of another. This approach gives flexibility of a synergistic performance of stacking each base learner and produces a formidable model that yields a high level of accuracy and predictive ability. We analyzed the test data, out-of-sample, and our analyses reveals a robust inflation prediction results. In particular, we show that food CPI is the most important driver for headline urban, and rural inflation while bread and cereals is the most important driver for food inflation in Nigeria. Also, biscuits, agric rice, garri white were found to be among the top main drivers of bread and cereal inflation. Our study further shows that some components of the CPI baskets that majorly drive inflation were assigned lower weights. Hence, attention to CPI weights only, without recourse to understanding the tipping source, may undermined a successful control of inflation in Nigeria. Tracing and tracking the source of inflation to the least sub-component will help resolve inflation problem. Supplementary Information The online version contains supplementary material available at 10.1007/s43546-022-00384-2.
| S-EPMC9734342 | biostudies-literature
Project description:PurposeTo assess the feasibility of ferumoxytol-enhanced anesthesia-free cardiac MRI in neonates and young infants for complex congenital heart disease (CHD).Materials and methodsWith Institutional Review Board approval, 21 consecutive neonates and young infants (1 day to 11 weeks old; median age of 3 days) who underwent a rapid two-sequence (MR angiography [MRA] and four-dimensional [4D] flow) MRI protocol with intravenous ferumoxytol without sedation (n = 17) or light sedation (n = 4) at 3 Tesla (T) (except one case at 1.5T) between June 2014 and February 2016 were retrospectively identified. Medical records were reviewed for indication, any complications, if further diagnostic imaging was performed after MRI, and surgical findings. Two radiologists scored the images in two sessions on a 5-point scale for overall image quality and delineation of various anatomical structures. Confidence interval of proportions for likelihood of requiring additional diagnostic imaging after MRI was determined. For the possibility of reducing the protocol to a single rapid sequence, Wilcoxon-rank sum test was used to assess whether 4D flow and MRA significantly differed in anatomical delineation.ResultsOne of 21 patients (4.8%, 80% confidence interval 0-11%) required additional imaging, a computed tomography angiography to assess lung parenchyma and peripheral pulmonary arteries. Only 1 of 13 patients (7.7%) with operative confirmation had a minor discrepancy between radiology and operative reports (80% confidence interval 0-17%). 4D flow was significantly superior to MRA (P < 0.05) for the evaluation of systemic arteries, valves, ventricular trabeculae, and overall quality. Using Cohen's kappa coefficient, there was good interobserver agreement for the evaluation of systemic arteries by 4D flow (κ = 0.782), and systemic veins and pulmonary arteries by MRA (κ > 0.6). Overall 4D flow measurements (mean κ = 0.64-0.74) had better internal agreement compared with MRA (mean κ = 0.30-0.64).ConclusionFerumoxytol-enhanced cardiac MRI, without anesthesia, is feasible for the evaluation of complex CHD in neonates and young infants, with a low likelihood of need for additional diagnostic studies. The decreased risk by avoiding anesthesia must be balanced against the potential for adverse reactions with ferumoxytol.Level of evidence2 J. MAGN. RESON. IMAGING 2017;45:1407-1418.
| S-EPMC5373922 | biostudies-literature
Project description:PurposeCough during emergence from anesthesia is a common problem and may cause adverse events. Monotherapy faces uncertainty in preventing emergence cough due to individual differences. We aimed to evaluate the efficacy and safety of multimodal intervention for preventing emergence cough in patients following nasal endoscopic surgery.MethodsIn this double-blind randomized trial, 150 adult patients undergoing nasal endoscopic surgery were randomly allocated into three groups. For the control group (n = 50), anesthesia was performed according to clinical routine, no intervention was provided. For the double intervention group (n = 50), normal saline 3 mL was sprayed endotracheally before intubation, 0.4 μg/kg dexmedetomidine was infused over 10 min after intubation, and target-controlled remifentanil infusion was maintained at an effect-site concentration of 1.5 ng/mL before extubation after surgery. For the multimodal intervention group (n = 50), 0.5% ropivacaine 3 mL was sprayed endotracheally before intubation, dexmedetomidine and remifentanil were administered as those in the double intervention group. The primary endpoint was the incidence of emergence cough, defined as single cough or more from end of surgery to 5 min after extubation.ResultsThe incidences of emergence cough were 98% (49/50) in the control group, 90% (45/50) in the double group, and 70% (35/50) in the multimodal group, respectively. The incidence was significantly lower in the multimodal group than those in the control (relative risk 0.71; 95% CI 0.59 to 0.86; p < 0.001) and double (relative risk 0.78; 95% CI 0.63 to 0.95; p = 0.012) groups; the difference between the double and control groups was not statistically significant (relative risk 0.92; 95% CI 0.83 to 1.02; p = 0.20). The severity of sore throat was significantly lower in the multimodal group than that in the control group (median difference-1; 95% CI -2 to 0; p = 0.016). Adverse events did not differ among the three groups.ConclusionFor adult patients undergoing endonasal surgery, multimodal intervention including ropivacaine topical anesthesia before intubation, dexmedetomidine administration after intubation, and remifentanil infusion before extubation after surgery significantly reduced emergence cough and was safe.
| S-EPMC10847355 | biostudies-literature
Project description:Promoters are short consensus sequences of DNA, which are responsible for transcription activation or the repression of all genes. There are many types of promoters in bacteria with important roles in initiating gene transcription. Therefore, solving promoter-identification problems has important implications for improving the understanding of their functions. To this end, computational methods targeting promoter classification have been established; however, their performance remains unsatisfactory. In this study, we present a novel stacked-ensemble approach (termed SELECTOR) for identifying both promoters and their respective classification. SELECTOR combined the composition of k-spaced nucleic acid pairs, parallel correlation pseudo-dinucleotide composition, position-specific trinucleotide propensity based on single-strand, and DNA strand features and using five popular tree-based ensemble learning algorithms to build a stacked model. Both 5-fold cross-validation tests using benchmark datasets and independent tests using the newly collected independent test dataset showed that SELECTOR outperformed state-of-the-art methods in both general and specific types of promoter prediction in Escherichia coli. Furthermore, this novel framework provides essential interpretations that aid understanding of model success by leveraging the powerful Shapley Additive exPlanation algorithm, thereby highlighting the most important features relevant for predicting both general and specific types of promoters and overcoming the limitations of existing 'Black-box' approaches that are unable to reveal causal relationships from large amounts of initially encoded features.
| S-EPMC7986616 | biostudies-literature
Project description:BackgroundEarly detection of postoperative increase in C-reactive protein (CRP) correlates with postoperative complications. The present study examined the association between preoperative / intraoperative factors and postoperative CRP levels, with development and validation of a prediction model of early postoperative CRP level, for prophylactic management of postoperative complications in patients undergoing surgery under general anesthesia.Material and methodsMultivariate regression analysis was retrospectively performed to determine the independent factor of CRP levels on postoperative day (POD) 1 and to develop a prediction model. Validation of the prediction model was prospectively performed. Data from 316 adult patients on perioperative variables were retrospectively obtained in a training cohort in patients undergoing elective non-cardiac surgery. In a validation cohort, 88 patients undergoing mastectomy and 68 patients undergoing laparoscopic colon surgery were prospectively utilized to evaluate the value of the prediction model. Major complications after surgery were defined as the Clavien-Dindo grade IIIa or greater.ResultsDuration of surgery, mean nociceptive response (NR) during surgery as intraoperative nociception level, and preoperative CRP level were selected to set up the prediction model of CRP level on POD1 (P < 0.0001). In the validation cohort, the predicted CRP levels on POD1 significantly correlated with the measured CRP after mastectomy (P < 0.0001) and laparoscopic colon surgery (P = 0.0001). Receiver-operating characteristic curve analysis showed that the predicted CRP levels on POD1 was significantly associated with major complications after mastectomy (P = 0.0259) and laparoscopic colon surgery (P = 0.0049). The measured and predicted CRP levels significantly increased in the order of severity of postoperative complications (P < 0.01).ConclusionIncreases in duration of surgery, intraoperative nociceptive level and preoperative CRP level were selected to predict early increases in CRP level after non-cardiac surgery under general anesthesia. Predicted CRP levels on POD1 were likely associated with severity of postoperative complications.
| S-EPMC6886833 | biostudies-literature