Journal of Clinical Monitoring and Computing最新文献

Emergency surgeries are resource-intensive procedures with high variability in operating room occupation time (OT) and hospital length of stay (LOS), complicating scheduling and capacity planning. Manual estimates by surgeons are frequently inaccurate, especially in emergency settings. Machine learning models (MLMs) have shown good predictive performance in elective surgery, but their applicability to emergency contexts remains underexplored. We conducted a retrospective, single-center study on 3,117 emergency procedures performed at the Pitié-Salpêtrière hospital, a major trauma center, between 2015 and 2018. Preoperative data available at the time of surgical scheduling were used to train four regression models for OT and LOS prediction: Ridge Regression, Random Forest, XGBoost, and a Multi-Layer Perceptron. Model performance was evaluated using Mean Absolute Error, Root Mean Square Error, Mean Absolute Percentage Error, and operational metrics: proportion of OT predictions within 20% of actual value (Within20) and LOS within fixed-day thresholds. RF and XGB outperformed manual estimates for OT, with RF achieving a MAE of 32 min and Within20 of 60%, improving surgeon estimates by 13%. For LOS, XGB was the best performing model with a MAE of 5 days and RMSE of 12 days. As measured through MAPE, prediction performance varied across specialties, with better accuracy in digestive and maxillofacial procedures. As for elective cases, MLMs can improve OT and LOS predictions in emergency surgery, though predictive performance remains moderate. Future work should refine models through enriched data, clinically relevant thresholds, and integration into decision-support tools to enhance emergency surgical care coordination.

To evaluate whether a web-based Intraoperative Glycemic Protocol Calculator (IGPC) improves provider compliance with intraoperative glycemic management protocols during cardiac surgery. Single-center retrospective cohort study conducted between August - October 2022 (pre-intervention) and April - June 2023 (post-intervention). Tertiary care academic hospital. Adult patients undergoing coronary artery bypass grafting and/or valve surgery requiring cardiopulmonary bypass. Implementation of the IGPC, a web-based clinical decision support tool designed to automate insulin dosing recommendations intraoperatively. Protocol adherence, defined as appropriate insulin administration within five minutes of glucose measurement, was compared before and after IGPC implementation. Among 143 patients, IGPC use significantly increased adherence across all intraoperative phases: Pre-CPB (65.5% to 80.2%, p = 0.017), On-CPB (53.0% to 75.1%, p < 0.001), and Post-CPB (34.8% to 58.8%, p < 0.001). Rates of severe hypoglycemia remained low and unchanged (0.1% in both groups; p = 0.772), and intraoperative hyperglycemia rates were similar (4.2% vs. 4.1%; p = 0.995). Implementation of the IGPC significantly improved real-time adherence to intraoperative glycemic control protocols without increasing adverse glycemic events. However, rates of intraoperative hyperglycemia and hypoglycemia remained unchanged between the pre- and post-intervention phases. These findings highlight the utility of clinical decision support tools in enhancing protocol compliance during high-acuity cardiac surgeries.

Purpose: Postoperative pain remains a significant adverse event after thoracic surgery, to which intraoperative nociception contributes. To measure intraoperative nociception, the Nociception Level (NOL) monitor is one option. This study aims to assess the NOL's utility for measuring intraoperative nociception and predicting acute postoperative pain and opioid consumption.

Methods: This observational study included 114 thoracic surgery patients (37 thoracotomy, 77 minimally invasive surgery) utilizing two analgesic approaches (40 peridural anaesthesia, 74 intercostal nerve block). NOL's utility to monitor responses to incision and nerve block was assessed at population (Wilcoxon-signed-rank tests) and individual level (ROC-analyses). NOL's predictive utility for postoperative pain and opioid consumption was analysed at population (Mann-Whitney-U tests) and individual level (multivariable linear regression).

Results: Population NOL significantly increased after incision and decreased after nerve block (p < 0.01/ p < 0.01) and individual detection of nociception was significantly better than chance (AUCs: 0.68 [95%CI 0.61-0.75] / 0.62 [95%CI 0.53-0.72]). However, NOL did not differ significantly between thoracotomy and minimally invasive surgery (p = 0.12) or peridural anaesthesia and nerve block (p = 0.16), despite significantly different postoperative pain and opioid consumption (p < 0.01). Multivariable analyses showed no significant effect of NOL on postoperative pain or opioid consumption.

Conclusion: NOL captures intraoperative stress with an accuracy allowing to differentiate large nociception changes at both the population and individual level after skin incision and nerve block. However, NOL was unable to differentiate between patients with plausibly different nociception levels, like patients undergoing different surgical or analgesic techniques. Consequently, while NOL can detect large nociceptive changes, its current accuracy may be insufficient to reliably guide individual analgesia in clinical practice.

To determine the effect of monitoring the Analgesia Nociception Index (ANI) on intraoperative opioid use, postoperative recovery, and analgesia in patients receiving preoperative bilateral erector spinae plane block (ESPB) for gynecological surgery under general anesthesia. Eighty patients classified in the American Society of Anesthesiologists physical status I-III scheduled for hysterectomy under general anesthesia were included in the study. After ultrasound-guided ESPB, patients were divided into 2 groups: control and ANI. In the control group, the intraoperative remifentanil infusion dose was adjusted using conventional methods; in the ANI group, the dose was adjusted according to ANI values of 50-70. Intraoperative remifentanil consumption, postoperative pain scores, additional analgesic requirements, and complications were recorded. Intraoperative remifentanil consumption was lower in the ANI group than in the control group (p < 0.001). Numerical rating scale (NRS) scores and requirements for additional analgesics in the postoperative recovery unit were both lower in the ANI group (p < 0.05). There were no significant differences between the groups in terms of nausea or vomiting in the recovery unit. ANI monitoring in patients undergoing gynecological surgery under general anesthesia with ESPB reduced opioid consumption during the intraoperative period. Intraoperative ANI monitoring enabled individualized opioid administration and guided determination of the required dose of analgesic agent.

Capillary refill time (CRT) is a vaso-occlusive test that allows the non-invasive assessment of skin perfusion. A vascular occlusive test (VOT) induces transient ischemia similar to that used in preconditioning ischemia. We hypothesized that CRT could be influenced by local tissue compression mimicking ischemic preconditioning when repeated measurements are performed. In healthy volunteers (n = 30), CRTs were performed twice on the index and middle fingers of the dominant hand and the index finger of the non-dominant hand at 15-minute intervals on the first day. On the second day, two CRT measurements were taken at 30-minute intervals. No significant differences were observed in CRT measurements repeated at 15- and 30-minute intervals. Additionally, baseline CRT values did not significantly differ between the fingers of the dominant and non-dominant hands on either study day. Repeated CRT measurements are not influenced by local ischemic preconditioning in the finger over short intervals.

Purpose: To evaluate the inter- and intraobserver reproducibility of sublingual microcirculatory indices measured using the GlycoCheck system, including the perfused boundary region (PBR), vascular density (VD), and red blood cell filling (RBCF), in patients under general anesthesia without any motion artifacts.

Methods: Fifty patients who received general anesthesia for laparoscopic gastrointestinal surgery were included in this study. After the induction of general anesthesia, the leading observer and one of the five subobservers took two and one measurements of sublingual microcirculation with the GlycoCheck system, respectively. Inter- and intraobserver reproducibility was assessed using intraclass correlation coefficients (ICC). Interobserver reproducibility was calculated using the first measurements of the leading observer and subobservers, and intraobserver reproducibility was calculated using two consecutive measurements of the leading observer.

Results: The interobserver reproducibility of a single measurement was poor for all three parameters. The interobserver ICCs for PBR were 0.13 [95% CI: -0.15, 0.39], for VD was - 0.01 [95%CI: -0.29, 0.27], and for RBCF were 0.31 [95%CI: -0.45, 0.78]. The intraobserver ICCs for PBR was 0.32 [95% CI: 0.05, 0.55] for all 50 cases, 0.17 [95% CI: -0.25, 0.53] for the first 25 cases, and 0.46 [95% CI: 0.09, 0.72] for the second 25 cases. The Bland-Altman plots indicated that the measurement errors were random.

Conclusion: In patients under general anesthesia, single PBR, VD, and RBCF measurements using the GlycoCheck system showed poor interobserver reproducibility. Although the intraobserver reproducibility of PBR measurements was poor, improving measurement proficiency might improve reproducibility. Further research is required to establish measurement methods that achieve better reproducibility and adequate observer training.

The Nociception Level (NOL) index of the PMD-200™ monitor measures intraoperative nociception-antinociception balance. Because it relies on photoplethysmography, it may be affected by pacemaker interference. We evaluated its response to pacemaker stimulation in the absence of nociceptive input. Mechanically ventilated adults after elective cardiac surgery were studied. NOL index, bispectral index, mean arterial pressure, and heart rate were recorded every minute for 35 min across seven five-minute periods: baseline (pacemaker off), pacing at 90 beats.min^- 1, pacing at 110 beats min^- 1, pacemaker off (washout), pacing at 110 beats min^- 1 (rechallenge), after PMD-200™ recalibration at 110 beats min^- 1, and continued monitoring at 110 beats min^- 1. Data were analysed with mixed-model repeated measures (random intercept for patient, time fixed; bispectral index covariate for NOL). Results are least-square adjusted means ± (standard error), comparing the last minute of each period. Twenty patients were analysed. Pacemaker-induced heart rate changes significantly affected NOL over time (F = 28.420, p < 0.001). Compared with baseline 2.1 ± (1.74), pacing at 90 beats min^- 1 increased NOL to 8.4 ± (1.73) (p = 1.000) and at 110 beats min^- 1 to 18.4 ± (1.73) (p < 0.001). Stopping pacing returned NOL to 1.1 ± (1.73) (p = 1.000), which rose again at 110 beats.min^- 1 rechallenge to 18.0 ± (1.73) (p < 0.001). Recalibration restored baseline values 1.1 ± (1.73) (p = 1.000), with stability maintained during continued monitoring 1.5 ± (1.73) (p = 1.000). The NOL index captured the studied nociception-antinociception balance during pacemaker stimulation when recalibrated to the paced rate. ClinicalTrials.gov: NCT06696781 on 17.11.2024.

Introduction: Haemoglobin measurement is an essential parameter for quantifying anaemia and often used for guiding transfusion decisions. Conventional methods require blood sampling and are invasive. Results are intermittent, discontinuous and obtained after a reasonable acquisition time. Hemoglobinemia by pulsed co-oximetry is non-invasive, immediate and offers the advantage of continuous monitoring. The aim of this systematic review is to assess the diagnostic accuracy of pulsed co-oximetry compared with reference biological determinations in perioperative management.

Methods: The review was registered in PROSPERO and performed according to the PRISMA statement. Searches in Pubmed, Cochrane Library and Scopus databases were performed from January 2000 to February 2024 for studies comparing non-invasive haemoglobin measurement with invasive methods. The QUADAS-2 scale was used to assess the risk of bias. For data analysis, Review Manager 5.4.1 software was employed, using the inverse variance method and a random-effects model to calculate the mean difference (MD) and 95% confidence intervals. Sensitivity analysis were performed in order to assess the influence of site of blood sampling (arterial or venous), revision model reference of the Masimo finger sensor, the geographical location of the study centre, the risk of bias classification, the population type and the type of study.

Results: The meta-analysis included 36 studies involving 1888 patients. Meta-analysis revealed a mean difference between the non-invasive and invasive methods of 0.13 g.dL-1 (95% confidence interval [CI]: 0.10- 0.36) (P-value > 0.05). Sensitivity analyses showed no statistically significant difference between the two methods. There was a very good homogeneity among the studies (I² = 0%). Trending analysis was considered acceptable in a majority of the studies.

Conclusion: The results obtained support the reliability of pulsed co-oximetry. Considering the potential benefits of this parameter, it seems rational to integrate this technology perioperatively to guide standard clinical practices for optimizing the management of surgical patients.