Daily Anesthesiology Research Analysis
Three large-scale perioperative studies stand out today: a statewide analysis shows opioid-free discharge after bariatric surgery rose fourfold without worsening outcomes; a single-center cohort of 1,555 liver transplants supports the safety of early extubation with very low reintubation rates; and a multicenter machine-learning model accurately predicts postoperative pulmonary complications after neurosurgery, outperforming existing risk scores.
Summary
Three large-scale perioperative studies stand out today: a statewide analysis shows opioid-free discharge after bariatric surgery rose fourfold without worsening outcomes; a single-center cohort of 1,555 liver transplants supports the safety of early extubation with very low reintubation rates; and a multicenter machine-learning model accurately predicts postoperative pulmonary complications after neurosurgery, outperforming existing risk scores.
Research Themes
- Opioid stewardship and enhanced recovery after surgery
- Early extubation strategies in transplant anesthesia
- Machine-learning risk prediction for postoperative pulmonary complications
Selected Articles
1. Variation in opioid-free discharge after metabolic surgery from 2018 to 2023: a state-wide analysis from the Michigan Bariatric Surgery Collaborative.
In a statewide registry of 54,276 metabolic-bariatric operations, opioid-free discharge rose from 7.7% to 32.1% with only 0.4% filling opioids within 30 days. This strategy did not increase complications and was associated with fewer ED visits, suggesting safety and effectiveness of opioid stewardship at discharge.
Impact: This large-scale, real-world analysis demonstrates that opioid-free discharge after bariatric surgery is scalable and safe, with measurable reductions in ED utilization. It provides compelling evidence to shift postoperative prescribing norms.
Clinical Implications: Programs can implement opioid-free discharge pathways with monitoring, patient education, and non-opioid analgesia, expecting comparable complication rates and potentially fewer ED visits.
Key Findings
- Opioid-free discharge increased from 7.7% (2018) to 32.1% (2023).
- Only 0.4% of opioid-free patients filled an opioid prescription within 30 days.
- Opioid-free discharge was associated with fewer ED visits (7.7% vs 8.2%) with similar 30-day complication rates (7.6% vs 7.3%).
Methodological Strengths
- Very large, statewide, procedure-specific registry with risk-adjusted 30-day outcomes
- Contemporary multi-year trend analysis with surgeon-level comparisons
Limitations
- Observational design may suffer residual confounding (e.g., pain severity, patient preference not fully captured)
- Generalizability outside the collaborative and to other procedures uncertain
Future Directions: Prospective implementation trials integrating patient-reported pain outcomes and tailored non-opioid protocols; exploration of scalability to other surgeries and diverse health systems.
BACKGROUND: Efforts have been made to reduce opioid prescribing after metabolic-bariatric surgery (MBS) given the increased risk for misuse. Variation in prevalence of opioid-free discharge following MBS and its impact on outcomes remains unclear. OBJECTIVES: To evaluate variation in opioid prescribing practices after MBS and the impact of opioid-free discharge on outcomes. SETTING: MBS programs participating in a state-wide quality improvement collaborative. METHODS: Using a state-wide bariatric-specific data registry, all patients who underwent MBS between 2018 and 2023 and had opioid prescribing data were identified (n = 54,276). Patient characteristics and 30-day risk-adjusted outcomes were compared between patients who were and were not prescribed opioids at discharge. Surgeon and practice characteristics were also compared between the top and bottom quartiles of opioid-free discharge. RESULTS: The prevalence of opioid-free discharge increased from 7.7% to 32.1% over the study period. Only .4% of patients, who were opioid-free at discharge, obtained an opioid prescription within 30 days of discharge. Opioid-free discharge was associated with lower rates of emergency department (ED) visits (7.7% vs 8.2%, P = .0008), despite similar complication rates (7.6% vs 7.3%, P = .7261). There were no significant differences in age, case volume, or practice types between surgeons in the top quartile and bottom quartile for opioid-free discharge. CONCLUSIONS: Opioid-free discharge after MBS has increased in prevalence with extremely low failure rates without negatively impacting ED visit rates. Variation in opioid prescribing persists and may be due to patient-specific factors as well as surgeon-specific preference.
2. Development and multicenter validation of machine learning models for predicting postoperative pulmonary complications after neurosurgery.
Using 7,533 development cases with temporal (n=2,824) and external (n=2,300) validation, a DNN model and an 11-feature LR nomogram predicted PPCs within 7 days after neurosurgery (AUC ~0.83). The LR nomogram outperformed ARISCAT and LAS VEGAS scores, indicating potential for clinical decision support.
Impact: Provides validated, parsimonious and full-feature ML tools tailored to neurosurgery, outperforming widely used generic scores and enabling targeted perioperative prevention.
Clinical Implications: Integrate the nomogram into preoperative assessment to identify high-risk patients for lung-protective ventilation, early mobilization, and respiratory therapy; monitor model drift with ongoing calibration.
Key Findings
- PPC incidence ~9–9.5% across development, temporal, and external datasets.
- DNN achieved AUC 0.835 (Brier 0.069) in temporal validation; LR/XGBoost performed closely.
- An 11-feature LR nomogram outperformed ARISCAT (AUC 0.672) and LAS (0.663) with AUC 0.824.
Methodological Strengths
- Temporal and external validation across multiple centers with standardized PPC definitions (EPCO)
- Comparison of six ML algorithms with both full and LASSO-selected feature sets
Limitations
- Retrospective data extraction; unmeasured confounders and practice variability may persist
- Feature importance inconsistency between LR-SHAP and multivariable analyses
Future Directions: Prospective implementation with clinician-in-the-loop decision support; assess impact on PPC reduction and cost; periodic recalibration across institutions.
BACKGROUND: Postoperative pulmonary complications (PPCs) are major adverse events in neurosurgical patients. This study aimed to develop and validate machine learning models predicting PPCs after neurosurgery. METHODS: PPCs were defined according to the European Perioperative Clinical Outcome standards as occurring within 7 postoperative days. Data of cases meeting inclusion/exclusion criteria were extracted from the anesthesia information management system to create three datasets: The development (data of Huashan Hospital, Fudan University from 2018 to 2020), temporal validation (data of Huashan Hospital, Fudan University in 2021) and external validation (data of other three hospitals in 2023) datasets. Machine learning models of six algorithms were trained using either 35 retrievable and plausible features or the 11 features selected by Lasso regression. Temporal validation was conducted for all models and the 11-feature models were also externally validated. Independent risk factors were identified and feature importance in top models was analyzed. RESULTS: PPCs occurred in 712 of 7533 (9.5%), 258 of 2824 (9.1%), and 207 of 2300 (9.0%) patients in the development, temporal validation and external validation datasets, respectively. During cross-validation training, all models except Bayes demonstrated good discrimination with an area under the receiver operating characteristic curve (AUC) of 0.840. In temporal validation of full-feature models, deep neural network (DNN) performed the best with an AUC of 0.835 (95% confidence interval [CI]: 0.805-0.858) and a Brier score of 0.069, followed by Logistic regression (LR), random forest and XGBoost. The 11-feature models performed comparable to full-feature models with very close but statistically significantly lower AUCs, with the top models of DNN and LR in temporal and external validations. An 11-feature nomogram was drawn based on the LR algorithm and it outperformed the minimally modified Assess respiratory RIsk in Surgical patients in CATalonia (ARISCAT) and Laparoscopic Surgery Video Educational Guidelines (LAS VEGAS) scores with a higher AUC (LR: 0.824, ARISCAT: 0.672, LAS: 0.663). Independent risk factors based on multivariate LR mostly overlapped with Lasso-selected features, but lacked consistency with the important features using the Shapley additive explanation (SHAP) method of the LR model. CONCLUSIONS: The developed models, especially the DNN model and the nomogram, had good discrimination and calibration, and could be used for predicting PPCs in neurosurgical patients. The establishment of machine learning models and the ascertainment of risk factors might assist clinical decision support for improving surgical outcomes. TRIAL REGISTRATION: ChiCTR 2100047474; https://www.chictr.org.cn/showproj.html?proj=128279 .
3. Safety and Feasibility of Early Extubation in Liver Transplantation: Experience in 1555 Patients.
In 1,555 adult liver transplants, 62% underwent early extubation with only 3.2% requiring ventilation within 48 hours and no increase in postoperative pneumonia. Notably, even in the highest quartiles of MELD-Na and blood loss, one-third of patients were safely extubated early.
Impact: Provides strong real-world evidence that early extubation is feasible and safe in liver transplantation, including select high-risk subgroups, supporting broader protocol adoption.
Clinical Implications: Transplant anesthesia teams can adopt early extubation pathways with defined selection criteria and rescue plans, potentially reducing ICU utilization without increasing pulmonary complications.
Key Findings
- Early extubation was achieved in 62% (969/1555) of liver transplants.
- Only 3.2% required ventilation within 48 h postoperatively (1.1% reintubation for respiratory failure; 2.1% continued intubation after reoperation).
- No significant difference in postoperative pneumonia; 34% of highest-risk quartile (MELD-Na >34 and EBL >5 L) were extubated early.
Methodological Strengths
- Large single-center cohort over a decade with pre-specified primary outcome
- Stratified analyses by MELD-Na and blood loss quartiles including high-risk subgroups
Limitations
- Retrospective, single-center design with potential selection bias
- Limited detail on standardized extubation criteria across providers
Future Directions: Prospective multicenter protocols to test standardized early extubation criteria and evaluate ICU length of stay, costs, and long-term outcomes.
BACKGROUND: Early extubation after liver transplantation can decrease cost and intensive care unit lengths of stay, but its adoption remains limited because of safety concerns. We assessed the feasibility and safety of early extubation at a liver transplant center with a high early extubation rate. We analyzed subgroups of high-risk patients, including high model for end-stage liver disease-sodium (MELD-Na) score, high intraoperative blood loss, and patients undergoing simultaneous liver-kidney transplantation. METHODS: We included all adult liver transplantations performed at a single center between June 2012 and July 2022. Patients were divided into 2 groups: (1) those extubated early (ie, in the operating room or within the first hour of intensive care unit admission) and (2) those who underwent delayed extubation. The primary outcome was reintubation within 48 h after early extubation. Rates of early extubation were analyzed separately for quartiles of MELD-Na score and intraoperative blood loss. RESULTS: Of 1555 patients, 969 (62%) were extubated early. Of these, 31 patients (3.2%) required mechanical ventilation within 48 h postoperatively: 11 patients (1.1%) were reintubated for respiratory failure and 20 (2.1%) remained intubated after reoperation. There was no difference in postoperative pneumonia between the groups ( P = 0.059). Early extubation rates inversely correlated with the quartiles of MELD-Na score and estimated blood loss. In the highest quartile for MELD-Na (>34) and estimated blood loss (>5 L), 34% of patients were extubated early. CONCLUSIONS: Early extubation of properly selected patients after liver transplantation is safe and associated with a low rate of reintubation, even among select groups of high-risk patients.