Journal of Clinical Monitoring and Computing最新文献

Purpose: Impaired cerebral pressure autoregulation is common and detrimental after acute brain injuries. Based on the prevalence of delayed cerebral ischemia in aneurysmal subarachnoid hemorrhage (aSAH) patients compared to traumatic brain injury (TBI), we hypothesized that the type of autoregulatory disturbance and the optimal PRx range may differ between these two conditions. The aim of this study was to determine the optimal PRx ranges in relation to functional outcome following aSAH and TBI, respectively.

Methods: In this observational study, 487 aSAH patients and 413 TBI patients, treated in the neurointensive care, Uppsala, Sweden, between 2008 and 2018, were included. The percentage of good monitoring time (%GMT) of PRx was calculated within 8 intervals covering the range from -1.0 to + 1.0, and analyzed in relation to favorable outcome (GOS-E 5 to 8).

Results: In multiple logistic regressions, a higher %GMTs of PRx in the intervals -1.0 to -0.5 and + 0.75 to + 1.0 were independently associated with a lower rate of favorable outcome in the aSAH cohort. In a similar analysis in the TBI cohort, only positive PRx in the interval + 0.75 to + 1.0 was independently associated with a lower rate of favorable outcome.

Conclusion: Extreme PRx values in both directions were unfavorable in aSAH, possibly as high PRx could indicate proximal vasospasm with exhausted distal vasodilatory reserve, while very negative PRx could reflect myogenic hyperreactivity with suppressed cerebral blood flow. Only elevated PRx was unfavorable in TBI, possibly as pressure passive vessels may be a more predominant pathomechanism in this disease.

Purpose: To this day there is no consensus regarding evidence of usefulness of Intraoperative Neurophysiological Monitoring (IONM). Randomized controlled trials have not been performed in the past mainly because of difficulties in recruitment control subjects. In this study, we propose the use of Bayesian Networks to assess evidence in IONM.

Methods: Single center retrospective study from January 2020 to January 2022. Patients admitted for cranial neurosurgery with intraoperative neuromonitoring were enrolled. We built a Bayesian Network with utility calculation using expert domain knowledge based on logistic regression as potential causal inference between events in surgery that could lead to central nervous system injury and postoperative neurological function.

Results: A total of 267 patients were included in the study: 198 (73.9%) underwent neuro-oncology surgery and 69 (26.1%) neurovascular surgery. 50.7% of patients were female while 49.3% were male. Using the Bayesian Network´s original state probabilities, we found that among patients who presented with a reversible signal change that was acted upon, 59% of patients would wake up with no new neurological deficits, 33% with a transitory deficit and 8% with a permanent deficit. If the signal change was permanent, in 16% of the patients the deficit would be transitory and in 51% it would be permanent. 33% of patients would wake up with no new postoperative deficit. Our network also shows that utility increases when corrective actions are taken to revert a signal change.

Conclusions: Bayesian Networks are an effective way to audit clinical practice within IONM. We have found that IONM warnings can serve to prevent neurological deficits in patients, especially when corrective surgical action is taken to attempt to revert signals changes back to baseline properties. We show that Bayesian Networks could be used as a mathematical tool to calculate the utility of conducting IONM, which could save costs in healthcare when performed.

Background: Robotic-assisted laparoscopic radical prostatectomy (RALP) requires pneumoperitoneum and steep Trendelenburg position. Our aim was to investigate the influence of the combination of pneumoperitoneum and Trendelenburg position on mechanical power and its components during RALP.

Methods: Sixty-one prospectively enrolled patients scheduled for RALP were studied in supine position before surgery, during pneumoperitoneum and Trendelenburg position and in supine position after surgery at constant ventilatory setting. In a subgroup of 17 patients the response to increasing positive end-expiratory pressure (PEEP) from 5 to 10 cmH₂O was studied.

Results: The application of pneumoperitoneum and Trendelenburg position increased the total mechanical power (13.8 [11.6 - 15.5] vs 9.2 [7.5 - 11.7] J/min, p < 0.001) and its elastic and resistive components compared to supine position before surgery. In supine position after surgery the total mechanical power and its elastic component decreased but remained higher compared to supine position before surgery. Increasing PEEP from 5 to 10 cmH₂O within each timepoint significantly increased the total mechanical power (supine position before surgery: 9.8 [8.4 - 10.4] vs 12.1 [11.4 - 14.2] J/min, p < 0.001; pneumoperitoneum and Trendelenburg position: 13.8 [12.2 - 14.3] vs 15.5 [15.0 - 16.7] J/min, p < 0.001; supine position after surgery: 10.2 [9.4 - 10.7] vs 12.7 [12.0 - 13.6] J/min, p < 0.001), without affecting respiratory system elastance.

Conclusion: Mechanical power in healthy patients undergoing RALP significantly increased both during the pneumoperitoneum and Trendelenburg position and in supine position after surgery. PEEP always increased mechanical power without ameliorating the respiratory system elastance.

Purpose: To determine the precise induction dose, an objective assessment of individual propofol sensitivity is necessary. This study aimed to investigate whether preinduction electroencephalogram (EEG) data are useful in determining the optimal propofol dose for the induction of general anesthesia in healthy adult patients.

Methods: Seventy healthy adult patients underwent total intravenous anesthesia (TIVA), and the effect-site target concentration of propofol was observed to measure each individual's propofol requirements for loss of responsiveness. We analyzed preinduction EEG data to assess its relationship with propofol requirements and conducted multiple regression analyses considering various patient-related factors.

Results: Patients with higher relative delta power (ρ = 0.47, p < 0.01) and higher absolute delta power (ρ = 0.34, p = 0.01) required a greater amount of propofol for anesthesia induction. In contrast, patients with higher relative beta power (ρ = -0.33, p < 0.01) required less propofol to achieve unresponsiveness. Multiple regression analysis revealed an independent association between relative delta power and propofol requirements.

Conclusion: Preinduction EEG, particularly relative delta power, is associated with propofol requirements during the induction of general anesthesia. The utilization of preinduction EEG data may improve the precision of induction dose selection for individuals.

Objective: This study aims to analyze the risk factors for early postoperative brain injury in patients undergoing cardiovascular surgery and explore the predictive value of transcranial color Doppler (TCCD) and regional cerebral oxygen saturation (rSO₂) for detecting early postoperative brain injury in cardiovascular surgery patients.

Methods: A total of 55 patients undergoing cardiovascular surgery with cardiopulmonary bypass in Changzhou No.2 The People's Hospital of Nanjing Medical University were included in this study. Neuron-specific enolase (NSE) concentration was measured 24 h after operation. Patients were divided into brain injury (NSE ≥ 16.3 ng/mL) and normal (0 < NSE < 16.3 ng/mL) groups according to the measured NSE concentration. The clinical outcomes between the two groups were compared, including decreased rSO₂ and cerebral blood flow (as measured by TCCD) levels. The risk factors of early postoperative brain injury were analyzed by multivariate logistic regression analysis, and the significant variables were analyzed by receiver operating characteristic (ROC) analysis.

Results: A total of 50 patients were included in this study, with 20 patients in the brain injury group and 30 patients in the normal group. Cardiopulmonary bypass time (min) (107 ± 29 vs. 90 ± 28, P = 0.047) and aortic occlusion time (min) (111 (IQR 81-127) vs. 87 (IQR 72-116), P = 0.010) were significantly longer in the brain injury group than in the normal group. Patients in the brain injury group had greater decreased rSO₂ (%) (27.0 ± 7.3 vs. 17.5 ± 6.1, P < 0.001) and cerebral blood flow (%) (44.9 (IQR 37.8-69.2) vs. 29.1 (IQR 12.0-48.2), P = 0.004) levels. Multivariate logistic regression analysis suggested that decreased rSO₂ and cerebral blood flow levels, aortic occlusion time, and history of atrial fibrillation were independent risk factors for early postoperative brain injury (P < 0.05). ROC analysis reported that the best cutoff values for predicting early postoperative brain injury were 21.4% and 37.4% for decreased rSO₂ and cerebral blood flow levels, respectively (P < 0.05).

Conclusion: The decreased rSO₂ and cerebral blood flow levels, aorta occlusion time, and history of atrial fibrillation were independent risk factors for early postoperative brain injury. TCCD and rSO₂ could effectively monitor brain metabolism and cerebral blood flow and predict early postoperative brain injury.

The Inferior Vena Cava (IVC) is commonly utilized to evaluate fluid status in the Intensive Care Unit (ICU),with more recent emphasis on the study of venous congestion. It is predominantly measured via subcostal approach (SC) or trans-hepatic (TH) views, and automated border tracking (ABT) software has been introduced to facilitate its assessment. Prospective observational study on patients ventilated in pressure support ventilation (PSV) with 2 × 2 factorial design. Primary outcome was to evaluate interchangeability of measurements of the IVC and the distensibility index (DI) obtained using both M-mode and ABT, across both SC and TH. Statistical analyses comprised Bland-Altman assessments for mean bias, limits of agreement (LoA), and the Spearman correlation coefficients. IVC visualization was 100% successful via SC, while TH view was unattainable in 17.4% of cases. As compared to the M-mode, the IVC-DI obtained through ABT approach showed divergences in both SC (mean bias 5.9%, LoA -18.4% to 30.2%, ICC = 0.52) and TH window (mean bias 6.2%, LoA -8.0% to 20.4%, ICC = 0.67). When comparing the IVC-DI measures obtained in the two anatomical sites, accuracy improved with a mean bias of 1.9% (M-mode) and 1.1% (ABT), but LoA remained wide (M-mode: -13.7% to 17.5%; AI: -19.6% to 21.9%). Correlation was generally suboptimal (r = 0.43 to 0.60). In PSV ventilated patients, we found that IVC-DI calculated with M-mode is not interchangeable with ABT measurements. Moreover, the IVC-DI gathered from SC or TH view produces not comparable results, mainly in terms of precision.

Hand hygiene among anesthesia personnel is important to prevent hospital-acquired infections in operating rooms; however, an efficient monitoring system remains elusive. In this study, we leverage a deep learning approach based on operating room videos to detect alcohol-based hand hygiene actions of anesthesia providers. Videos were collected over a period of four months from November, 2018 to February, 2019, at a single operating room. Additional data was simulated and added to it. The proposed algorithm utilized a two-dimensional (2D) and three-dimensional (3D) convolutional neural networks (CNNs), sequentially. First, multi-person of the anesthesia personnel appearing in the target OR video were detected per image frame using the pre-trained 2D CNNs. Following this, each image frame detection of multi-person was linked and transmitted to a 3D CNNs to classify hand hygiene action. Optical flow was calculated and utilized as an additional input modality. Accuracy, sensitivity and specificity were evaluated hand hygiene detection. Evaluations of the binary classification of hand-hygiene actions revealed an accuracy of 0.88, a sensitivity of 0.78, a specificity of 0.93, and an area under the operating curve (AUC) of 0.91. A 3D CNN-based algorithm was developed for the detection of hand hygiene action. The deep learning approach has the potential to be applied in practical clinical scenarios providing continuous surveillance in a cost-effective way.

Purpose: Growing evidence shows the complex interaction between lung and kidney in critically ill patients. The renal resistive index (RRI) is a bedside measurement of the resistance of the renal blood flow and it is correlated with kidney injury. The positive end-expiratory pressure (PEEP) level could affect the resistance of renal blood flow, so we assumed that RRI could help to monitoring the changes in renal hemodynamics at different PEEP levels. Our hypothesis was that the RRI at ICU admission could predict the risk of acute kidney injury in mechanical ventilated critically ill patients.

Methods: We performed a prospective study including 92 patients requiring mechanical ventilation for ≥ 48 h. A RRI ≥ 0.70, was deemed as pathological. RRI was measured within 24 h from ICU admission while applying 5,10 and 15 cmH₂O of PEEP in random order (PEEP trial).

Results: Overall, RRI increased from 0.62 ± 0.09 at PEEP 5 to 0.66 ± 0.09 at PEEP 15 (p < 0.001). The mean RRI value during the PEEP trial was able to predict the occurrence of AKI with AUROC = 0.834 [95%CI 0.742-0.927]. Patients exhibiting a RRI ≥ 0.70 were 17/92(18%) at PEEP 5, 28/92(30%) at PEEP 10, 38/92(41%) at PEEP 15, respectively. Thirty-eight patients (41%) exhibited RRI ≥ 0.70 at least once during the PEEP trial. In these patients, AKI occurred in 55% of the cases, versus 13% remaining patients, p < 0.001.

Conclusions: RRI seems able to predict the risk of AKI in mechanical ventilated patients; further, RRI values are influenced by the PEEP level applied.

Trial registration: Clinical gov NCT03969914 Registered 31 May 2019.

Prediction and avoidance of intraoperative hypotension (IOH) can lead to less postoperative morbidity. Machine learning (ML) is increasingly being applied to predict IOH. We hypothesize that incorporating demographic and physiological features in an ML model will improve the performance of IOH prediction. In addition, we added a "dial" feature to alter prediction performance. An ML prediction model was built based on a multivariate random forest (RF) trained algorithm using 13 physiologic time series and patient demographic data (age, sex, and BMI) for adult patients undergoing hepatobiliary surgery. A novel implementation was developed with an adjustable, multi-model voting (MMV) approach to improve performance in the challenging context of a dynamic, sliding window for which the propensity of data is normal (negative for IOH). The study cohort included 85% of subjects exhibiting at least one IOH event. Males constituted 70% of the cohort, median age was 55.8 years, and median BMI was 27.7. The multivariate model yielded average AUC = 0.97 in the static context of a single prediction made up to 8 min before a possible IOH event, and it outperformed a univariate model based on MAP-only (average AUC = 0.83). The MMV model demonstrated AUC = 0.96, PPV = 0.89, and NPV = 0.98 within the challenging context of a dynamic sliding window across 40 min prior to a possible IOH event. We present a novel ML model to predict IOH with a distinctive "dial" on sensitivity and specificity to predict first IOH episode during liver resection surgeries.

Purpose: Postoperative Delirium (POD) has an incidence of up to 65% in older patients undergoing cardiac surgery. We aimed to develop two dynamic nomograms to predict the risk of POD in older patients undergoing cardiac surgery.

Methods: This was a single-center retrospective cohort study, which included 531 older patients who underwent cardiac surgery from July 2021 to June 2022 at Nanjing First Hospital, China. Univariable and multivariable logistic regression were used to identify the significant predictors used when constructing the models. We evaluated the performances and accuracy, validated, and estimated the clinical utility and net benefit of the models using the receiver operating characteristic (ROC), the 10-fold cross-validation, and decision curve analysis (DCA).

Results: A total of 30% of the patients developed POD, the significant predictors in the preoperative model were ASA ( p < 0.001 OR = 3.220), cerebrovascular disease (p < 0.001 OR = 2.326), Alb (p < 0.037 OR = 0.946), and URE (p < 0.001 OR = 1.137), while for the postoperative model they were ASA (p = 0.044, OR = 1.737), preoperative MMSE score (p = 0.005, OR = 0.782), URE (p = 0.017 OR = 1.092), CPB duration (p < 0.001 OR = 1.010) and APACHE II (p < 0.001, OR = 1.353). The preoperative and postoperative models achieved satisfactory predictive performances, with AUC values of 0.731 and 0.799, respectively. The web calculators can be accessed at https://xxh152.shinyapps.io/Pre-POD/ and https://xxh152.shinyapps.io/Post-POD/ .

Conclusion: We established two nomogram models based on the preoperative and postoperative time points to predict POD risk and guide the flexible implementation of possible interventions at different time points.