Daily ReportOct 29, 2026
Anesthesiology, October 29 edition
We read 39 papers and selected 3.
Summary
The most impactful studies examined postoperative delirium prevention, comparative anesthetic safety, and large-scale perioperative risk prediction. A randomized noninferiority trial found remimazolam-based anesthesia was noninferior to propofol for delirium after neurovascular surgery while reducing hypotensive burden, whereas a meta-analysis supported dexmedetomidine over propofol for delirium prevention in older adults receiving regional anesthesia. A multicenter study showed that adding intraoperative and early postoperative variables modestly improved prediction of major complications, although calibration remained institution dependent.
Research Themes
- Postoperative delirium prevention and anesthetic selection
- Noninferiority evaluation of newer intravenous anesthetics
- Machine learning for perioperative risk prediction
Selected Articles
1. Effect of remimazolam versus propofol anaesthesia on postoperative delirium in neurovascular surgery: a randomised noninferiority trial.
In 686 patients undergoing elective neurovascular surgery, postoperative delirium occurred in 12.8% of patients receiving remimazolam and 10.8% receiving propofol. Remimazolam met the prespecified noninferiority criterion for delirium and produced a lower hypotensive burden and higher operating-room extubation rate.
Impact: This rigorous randomized trial directly addresses whether a reversible benzodiazepine can replace propofol in a high-risk neurosurgical setting without worsening delirium. It also links anesthetic choice to clinically relevant hemodynamic and recovery outcomes.
Clinical Implications: Remimazolam-based total intravenous anesthesia may be considered when rapid emergence, reduced hypotension, or operating-room extubation is desirable, provided that delirium risk is monitored and local experience with flumazenil reversal is adequate.
Key Findings
- Postoperative delirium occurred in 12.8% of the remimazolam group and 10.8% of the propofol group.
- Remimazolam was noninferior to propofol for postoperative delirium using a 10-percentage-point noninferiority margin.
- Hypotensive burden below a mean arterial pressure of 65 mm Hg was lower with remimazolam, and operating-room extubation was more frequent.
Methodological Strengths
- Single-centre, double-blind randomized noninferiority design with prespecified margin.
- Large sample for a focused neurovascular surgical population and validated delirium assessments performed twice daily.
Limitations
- The study was conducted at a single centre, which may limit generalizability to other neuroanesthetic practices.
- The observed delirium incidence was lower than anticipated when the noninferiority margin was defined.
Future Directions: Multicenter pragmatic trials should assess remimazolam across broader neurosurgical populations, evaluate longer-term cognitive outcomes, and examine cost-effectiveness and optimal flumazenil strategies.
BACKGROUND: Propofol-based total intravenous anaesthesia (TIVA) is commonly used in neurovascular surgery, in which intraoperative neurophysiological monitoring limits inhalation anaesthetic use. However, propofol causes hypotension, and prolonged infusions can delay emergence. Remimazolam, a benzodiazepine whose effects are reversible with flumazenil, has been associated with less intraoperative hypotension than propofol, but its use raises concerns about postoperative delirium (POD). We tested whether remimazolam-based TIVA is noninferior to propofol-based TIVA for POD. METHODS: In this single-centre, double-blind, randomised noninferiority trial, adults undergoing elective intracranial aneurysm clipping, arteriovenous malformation resection, microvascular decompression, or superficial temporal artery-middle cerebral artery bypass received remimazolam-based TIVA with flumazenil reversal at emergence, or propofol-based TIVA, both with remifentanil.
2. Dexmedetomidine versus propofol for the prevention of postoperative delirium in older adults undergoing regional anaesthesia: a systematic review and meta-analysis.
This PRISMA-compliant meta-analysis included five randomized trials and 1,613 older adults receiving regional anesthesia. Dexmedetomidine was associated with approximately half the risk of postoperative delirium compared with propofol, with low heterogeneity and moderate GRADE certainty.
Impact: Postoperative delirium is a major modifiable complication in older surgical patients. The consistent effect across randomized trials provides clinically useful comparative evidence for selecting intraoperative sedation during regional anesthesia.
Clinical Implications: Dexmedetomidine may be preferred over propofol for intraoperative sedation in selected older adults undergoing regional anesthesia, while monitoring for bradycardia, hypotension, and delayed recovery.
Key Findings
- Five randomized controlled trials involving 1,613 patients were included.
- Postoperative delirium occurred in 7.7% of patients receiving dexmedetomidine versus 16.7% receiving propofol.
- Dexmedetomidine reduced delirium risk with RR 0.48, 95% CI 0.36-0.64, and low heterogeneity of I² = 13.4%.
Methodological Strengths
- Systematic review and meta-analysis restricted to randomized controlled trials.
- Protocol registration, validated delirium instruments, GRADE assessment, and low statistical heterogeneity strengthen confidence in the estimate.
Limitations
- Only five trials were available, and variations in surgical procedures, sedation protocols, and delirium assessment timing may affect applicability.
- The evidence was rated moderate rather than high certainty, so residual bias and imprecision cannot be excluded.
Future Directions: Future multicenter trials should standardize dexmedetomidine dosing, compare clinically relevant recovery outcomes, and identify older patients most likely to benefit or experience adverse hemodynamic effects.
Postoperative delirium (POD) is a common complication in older adults undergoing surgery and is associated with adverse clinical outcomes. The choice of intraoperative sedative may influence its occurrence, particularly in the setting of regional anesthesia. A systematic review and meta-analysis of randomized controlled trials comparing dexmedetomidine versus propofol in patients aged ≥60 years was conducted in accordance with PRISMA guidelines, with protocol registration in PROSPERO. The primary outcome was the incidence of POD assessed using validated instruments (CAM or CAM-ICU). Five randomized controlled trials involving a total of 1,613 patients were included. The incidence of POD was 7.7% in the dexmedetomidine groups and 16.7% in the propofol groups. The meta-analysis showed a significant reduction in POD with dexmedetomidine (RR 0.48, 95% CI 0.36-0.64; p < 0.001), with low heterogeneity (I² = 13.4%).
3. Machine Learning-Based Prediction of Major Postoperative Complications After Noncardiac Surgery Under General Anesthesia Using Routinely Available Electronic Health Record Data: Retrospective Multicenter Development and Validation Study.
In 35,702 adults from three tertiary hospitals, models incorporating intraoperative and early postoperative variables consistently outperformed preoperative-only models for predicting a 30-day composite of major complications. However, calibration varied across facilities and risk estimates were generally overextreme, indicating that local recalibration and prospective workflow evaluation are necessary.
Impact: This is one of the largest anesthesiology-focused evaluations of routinely available perioperative machine learning data. It demonstrates both the incremental value of real-time perioperative information and the practical limitation that discrimination alone does not ensure transportable clinical risk estimates.
Clinical Implications: Perioperative machine learning models may help prioritize postoperative reassessment and surveillance, but they should not be deployed as universal decision tools without institution-specific recalibration, prospective validation, and predefined action thresholds.
Key Findings
- The 30-day composite major complication rate was 10.9% among 35,702 patients.
- Adding intraoperative and early postoperative variables increased AUROC from approximately 0.784-0.791 to 0.810-0.817.
- Calibration varied by facility, with perioperative models showing overly extreme predictions and overall risk overestimation.
Methodological Strengths
- Large multicenter cohort using routinely collected electronic health record, claims, and anesthesia information system data.
- Leave-one-facility-out internal-external cross-validation evaluated transportability across hospitals and included discrimination, calibration, probabilistic, and decision-curve metrics.
Limitations
- The retrospective design may permit residual confounding, missing-data bias, and changes in clinical practice over time.
- Facility-specific calibration differences limit direct portability, and decision-curve analysis did not establish a validated intervention or triage threshold.
Future Directions: Prospective implementation studies should test whether model-guided postoperative reassessment improves outcomes, establish clinically actionable thresholds, address fairness across patient groups, and continuously recalibrate models across institutions.
BACKGROUND: Postoperative complications after noncardiac surgery under general anesthesia remain a major cause of morbidity, mortality, and resource-intensive care. Conventional risk assessment tools rely mainly on preoperative information and may not capture perioperative physiological and procedural factors associated with postoperative risk. OBJECTIVE: This retrospective multicenter observational study aimed to develop and validate electronic health record (EHR)-based machine learning (ML) models for predicting major postoperative complications and to evaluate their discrimination, overall probabilistic performance, calibration, and potential clinical utility. METHODS: We analyzed 35,702 adults who underwent noncardiac surgery under general anesthesia at 3 tertiary hospitals between January 2021 and March 2025. Preoperative, intraoperative, and early postoperative variables were extracted from diagnosis procedure combination (DPC) claims, EHRs, and anesthesia information systems. The primary outcome was a 30-day composite of in-hospital death, unplanned reoperation, prolonged mechanical ventilation, new initiation of renal replacement therapy, and postoperative infection.