Vetrix Anesthesiology

An interpretable machine learning model for predicting emergence agitation in children: a multicenter development and validation study

Citation: Zhao Q, Zhang Y, An R, Yi B, Huang G. An interpretable machine learning model for predicting emergence agitation in children: a multicenter development and validation study. BMC Anesthesiol. 2026;[epub ahead of print]. This multicenter retrospective study from two Chinese hospitals developed and externally validated several machine learning models to predict emergence agitation in children after general anesthesia. Using five routine perioperative predictors, a compact multilayer perceptron model showed excellent discrimination internally but only moderate performance and suboptimal calibration in an external cohort. Because variable selection, model choice, and key predictors are highly data- and center-dependent, the tool is best seen as a proof-of-concept rather than a ready-made decision aid for routine paediatric anesthesia practice. Study at a glance - Design and setting: Retrospective multicenter prediction-model study using electronic records from two tertiary hospitals in China: Center I (Third Affiliated Hospital of Zunyi Medical University) for model development and internal validation, and Center II (First Affiliated Hospital of Army Medical University) as an independent external validation cohort. Children aged 3–12 years with American Society of Anesthesiologists physical status I–II undergoing elective surgery under general anesthesia were included. - Participants and primary outcome: A total of 445 pediatric patients were analyzed (321 in the development center, 124 in the external validation center). In Center I, emergence agitation occurred in 95 children, an incidence of 29.6%; in Center II, the reported incidence was about 25.8% (32–33 events, with slight inconsistencies across sections). The primary outcome was emergence agitation within 30 minutes after post-anesthesia care unit admission, defined as Pediatric Anesthesia Emergence Delirium (PAED) score >10, after pain was assessed and treated using the FLACC scale to limit misclassification from pain-related distress. - Predictors and main model: From 63 perioperative variables, the authors used univariable screening followed by least absolute shrinkage and selection operator (LASSO) regression to select predictors, then trained six algorithms (logistic regression, support vector machine, multilayer perceptron, random forest, extreme gradient boosting, and Light Gradient Boosting Machine). The final five-variable clinical model used parental educational level, preoperative alanine aminotransferase (ALT), postoperative patient-controlled analgesia (PCA) pump use, postoperative antagonist (reversal agent) use, and extubation suctioning frequency. A multilayer perceptron (MLP) was chosen as the primary clinical model because it performed best in external validation; a support vector machine was used for detailed interpretability analyses with SHAP values. - Key performance results: In internal holdout validation at Center I, discrimination was high across models, with area under the receiver operating curve (AUC) around 0.87–0.92; the support vector machine achieved AUC 0.918 (95% confidence interval 0.844–0.973, Brier score 0.098), and logistic regression AUC 0.915. In external validation at Center II, performance dropped noticeably. The primary clinical MLP model achieved AUC 0.705 (95% confidence interval 0.59–0.804) with a Brier score of 0.190, reflecting only moderate discrimination and imperfect calibration; other models performed worse (e.g., logistic regression AUC 0.587, LightGBM AUC 0.494). No decision-curve or net benefit analyses were reported. - Risk of bias and applicability: Using a structured prediction-model appraisal, overall risk of bias was judged high, mainly due to the analysis domain. Concerns include data-driven predictor selection from many candidates, testing and informally selecting among six algorithms, limited optimism correction (formally reported only for the support vector machine), and a relatively small, case-mix–different external validation cohort. Applicability is further constrained because three of the five final predictors (PCA pump use, antagonist use, suctioning frequency) are early postoperative management decisions that vary by center and over time, rather than stable baseline risk factors. - Practice implications: For practising clinicians, this study underscores that emergence agitation after pediatric anesthesia is common and potentially predictable, and it highlights perioperative features—such as parental education, preoperative ALT, and postoperative analgesia and reversal strategies—that may correlate with risk. However, the current models should not be used as stand-alone decision aids: external performance is only moderate, calibration is imperfect, and the models depend strongly on center-specific management choices. At present, these tools are best viewed as research prototypes and a stimulus for locally developed and rigorously validated prediction models, rather than as ready-to-implement clinical calculators.

Clinical Performance in Critical Care Simulation Under Sleep Deprivation: Effects of Power Napping in the R-NAP Randomized Controlled Trial

Citation: Schmidt L, Genty F, Delaire T, Valero B, Rey I, Galan L, Mairet-Mabboux S, Douplat M, Schlatter S, Rimmele T, Mazza S, Lilot M. Clinical Performance in Critical Care Simulation Under Sleep Deprivation: Effects of Power Napping in the R-NAP Randomized Controlled Trial. Anesthesiology. 2026; doi:10.1097/ALN.0000000000006135. In this single-centre randomized controlled trial, anesthesia and intensive care residents finishing an overnight on-site shift were assigned either to a brief supervised nap opportunity or to quiet wakefulness before a high-fidelity critical care simulation. The nap group achieved higher overall and non-technical performance scores without reported harms, but the trial is small, uses simulation rather than real patient outcomes, and has some missing data, so confidence in the size of benefit is moderate and the findings mainly support, rather than replace, broader fatigue management strategies. Study at a glance - Design and setting: Prospective individually randomized parallel-group behavioural trial conducted in a single French university simulation centre, comparing a supervised thirty-minute nap opportunity versus quiet wakefulness after an overnight on-site shift in anesthesia and intensive care residents. - Participants: Thirty-five second to fifth year anesthesia and intensive care residents were randomized (nineteen to nap, sixteen to no nap); twenty-seven with complete actigraphy sleep data were included in the primary adjusted analysis, with similar baseline characteristics between groups. - Primary outcome: Overall simulated clinical performance (sum of technical and non-technical scores, range zero to two hundred) after the overnight shift was higher with a nap; the adjusted mean difference was 14.84 points in favour of the nap group, with a ninety-five percent confidence interval from 2.8 to 26.88 and a P value of 0.018, yielding moderate certainty that a brief nap improves overall simulated performance. - Key secondary outcomes: Total non-technical skills score (zero to one hundred) was higher with a nap, with an adjusted mean difference of 11.03 points (ninety-five percent confidence interval 2.22 to 19.84; P value 0.016). Among Ottawa Global Rating Scale subscales, Overall performance (mean difference 0.77; ninety-five percent confidence interval 0.05 to 1.48; P value 0.036), Leadership (0.73; ninety-five percent confidence interval 0.05 to 1.41; P value 0.037), and Resource utilization (1.02; ninety-five percent confidence interval 0.03 to 2.00; P value 0.043) favoured the nap group, whereas purely technical checklist scores showed smaller, imprecise differences. - Harms and safety: No intervention-related adverse events or harms were reported in either the nap group (zero of nineteen) or the control group (zero of sixteen) during or after the brief nap or quiet wakefulness periods. - Risk of bias and certainty: Overall risk of bias was judged as having some concerns, mainly due to missing outcome data (eight of thirty-five randomized residents excluded from the primary model because of missing actigraphy) and lack of participant blinding, although assessors were blinded and outcomes were structured and video-based. Using this and the reasonably precise effect estimate, certainty in the primary outcome was rated as Moderate. - Applicability and limitations: Findings apply most directly to anesthesia and intensive care residents in similar academic settings and to high-fidelity simulation; they are indirect for attending physicians, other specialties, or real-world patient outcomes. The trial is small and single-centre, with multiple secondary and exploratory analyses without adjustment for multiplicity, so apparent benefits on individual subscales should be interpreted cautiously and used to support, not replace, broader fatigue risk management policies.

Outcome Differences Between General and Neuraxial Anesthesia for Hip Fracture by Frailty and Age in the Elderly: A Retrospective Cohort Study

Citation: Giannakis P, Restrepo M, Stone AB, Zhuang ST, Wang J, Cozowicz C, et al. Outcome Differences Between General and Neuraxial Anesthesia for Hip Fracture by Frailty and Age in the Elderly: A Retrospective Cohort Study. Anesth Analg. 2026;XXX(00):00–300. doi:10.1213/ANE.0000000000008062 Using a large United States hospital claims database, Giannakis and colleagues compared neuraxial versus general anesthesia for more than six hundred thousand hip fracture surgeries across age and frailty strata. Neuraxial anesthesia was associated with very small differences in in-hospital mortality and a composite of major complications, a clearer reduction in high opioid use, and slightly more discharges home, but also small increases in some complications and intensive care admissions. Because anesthesia type was not randomized and key clinical confounders and outcomes were captured only through billing codes, overall certainty is very low and the results should inform, not dictate, anesthetic choice. Study at a glance - Design and setting: Retrospective cohort study using the Premier Healthcare Database, including 623,122 adults undergoing surgical treatment of hip fracture in United States hospitals between 2016 and 2023. Exposure was anesthesia type (general vs neuraxial) coded from billing data; outcomes (in-hospital mortality, major complications, intensive care unit admission, length of stay, opioid use, discharge disposition) were defined from ICD-10-CM diagnosis codes and billing records. Associations were estimated with mixed-effects multivariable logistic regression adjusted for demographics, comorbidities, hospital characteristics, procedure type, peripheral nerve block use, fracture type, and time to surgery. - Primary outcome – composite of death and major complications: The prespecified primary endpoint was a composite of in-hospital mortality, respiratory complications, cardiac complications, acute renal failure, and delirium. Overall, neuraxial anesthesia versus general anesthesia was associated with an adjusted odds ratio (OR) of 0.97 (95% confidence interval [CI] 0.94–0.997; p=0.053), a very small relative difference compatible with little to no effect. Given the nonrandomized, claims-based design and serious residual confounding, GRADE certainty for this outcome is Very Low; the apparent benefit could easily be due to unmeasured differences between patients selected for each technique. - In-hospital mortality: In-hospital death was lower in the neuraxial group overall, with an adjusted OR of 0.83 (95% CI 0.74–0.93; p=0.003), and a more pronounced association in older, more frail subgroups (for example, OR 0.77, 95% CI 0.65–0.91 in patients aged ≥87 years with intermediate/high frailty). However, choice of anesthesia is strongly confounded by clinical status, cognitive function, and hemodynamic reserve, which are incompletely measured in claims. With Serious overall risk of bias and no advanced causal methods, GRADE certainty for any mortality benefit is Very Low. - Key secondary outcomes – opioid use, discharge home, length of stay: Neuraxial anesthesia was associated with a moderate reduction in high postoperative opioid use (overall adjusted OR 0.69, 95% CI 0.66–0.72; p<0.001), consistent across age and frailty strata, and with slightly higher odds of discharge to home among survivors (overall OR 1.08, 95% CI 1.04–1.12; p<0.001). Prolonged length of stay (≥75th percentile) showed a very small reduction with neuraxial anesthesia (overall OR 0.97, 95% CI 0.94–0.998; p=0.046). High opioid use is a process measure rather than a direct patient-important endpoint, and discharge disposition and length of stay are influenced by social and system factors; all three outcomes are rated Very Low certainty due to serious confounding and, for opioid use, additional indirectness. - Potential harms – respiratory, cardiac, and ICU outcomes: Across the overall cohort, neuraxial anesthesia was associated with slightly higher rates of several coded complications and intensive care unit use: respiratory complications (OR 1.06, 95% CI 1.01–1.10; p=0.03), cardiac complications (OR 1.07, 95% CI 1.02–1.12; p=0.008), and intensive care unit admission (OR 1.07, 95% CI 1.03–1.12; p=0.002). Subgroup and sensitivity analyses showed some heterogeneity by age, frailty, and hospital neuraxial use, but effects remained small. Because these outcomes rely on diagnosis codes without validation and are highly susceptible to confounding by severity and practice patterns, GRADE certainty is Very Low, and the direction of true effect is uncertain. - Risk of bias, certainty, and practice implications: Overall risk of bias is judged Serious due to residual confounding by indication, selection related to coding completeness, and outcome misclassification from claims data. All appraised outcomes, including the primary composite, mortality, complications, length of stay, opioid use, intensive care unit admission, and discharge home, are rated Very Low certainty with GRADE. Clinically, the study suggests that neuraxial and general anesthesia for hip fracture have broadly similar in-hospital risks, with neuraxial associated with less high opioid use and more home discharges but also small increases in some complications, all very uncertain. These findings should not, on their own, drive a major practice shift; instead, anesthetic choice should remain individualized, and system-level improvements in timely surgery, hemodynamic management, multimodal analgesia, delirium prevention, and early mobilization are likely to have larger and more reliable impact than anesthesia type alone.

Effect of low-dose propofol infusion with sevoflurane versus propofol-only total intravenous anesthesia on postoperative nausea and vomiting in high-risk patients: a single-blind randomized controlled clinical trial

Citation: Keck WL, Deng E, Liu WM, Kanekar R, Lomivorotov V, Sharma S. Effect of low-dose propofol infusion with sevoflurane versus propofol-only total intravenous anesthesia on postoperative nausea and vomiting in high-risk patients: a single-blind randomized controlled clinical trial. BMC Anesthesiol. 2026;26:136. doi:10.1186/s12871-026-03649-7 Single-center randomized trial in high-risk adults undergoing elective laparoscopic surgery compared a hybrid anesthetic (low-dose propofol infusion plus sevoflurane) with propofol-only total intravenous anesthesia, with all patients receiving dexamethasone and ondansetron prophylaxis. Rates of postoperative nausea and vomiting over 24 hours and rescue antiemetic use were similar, and adjusted odds ratios had very wide confidence intervals spanning benefit and harm. Overall certainty for differences between techniques in postoperative nausea and vomiting or antiemetic use is very low, so these findings are best viewed as exploratory rather than practice-changing. Study at a glance - Design and setting: Prospective, single-blind, individually randomized controlled trial in a single United States academic medical center, enrolling high-risk adults undergoing elective laparoscopic general, gynecologic, or urologic surgery under general anesthesia. - Participants: Sixty-five adults aged 18 years or older with a history of postoperative nausea and vomiting and/or motion sickness, predominantly American Society of Anesthesiologists physical status II–III and mostly female, were randomized to hybrid anesthesia (n=32) or propofol-only total intravenous anesthesia (n=33). - Intervention and comparator: Hybrid group: induction with propofol plus opioids and neuromuscular blockade, then low-dose propofol infusion combined with sevoflurane for maintenance under bispectral index guidance. Comparator group: propofol-based total intravenous anesthesia for maintenance under bispectral index guidance without volatile agents. Both groups received standardized prophylaxis with dexamethasone 4 milligrams after induction and ondansetron 4 milligrams at closure, similar fluid and vasoactive management, and opioid-based analgesia. - Primary outcome: 24-hour postoperative nausea and vomiting: Cumulative postoperative nausea and vomiting (any nausea, retching, or vomiting within 24 hours after surgery) occurred in 19 of 32 patients (59%) in the hybrid group versus 14 of 33 (42%) in the propofol-only group. Adjusted odds ratio 1.59; 95% confidence interval 0.51 to 4.93; p=0.80. GRADE certainty for this outcome is rated Very Low due to some concerns about selective reporting and very serious imprecision. - Secondary outcomes: early postoperative nausea and vomiting and rescue antiemetics: Post-anesthesia care unit postoperative nausea and vomiting occurred in 9 of 32 (28%) hybrid versus 7 of 33 (21%) propofol-only patients; adjusted odds ratio 1.83; 95% confidence interval 0.52 to 6.48; p=0.34 (Very Low certainty). Rescue antiemetic use in the post-anesthesia care unit occurred in 9 of 32 (28%) versus 7 of 33 (21%); adjusted odds ratio 1.59; 95% confidence interval 0.51 to 4.92; p=0.48 (Very Low certainty). - Secondary outcomes: 24-hour rescue antiemetic use and overall antiemetic consumption: Within 24 hours, rescue antiemetics were used in 12 of 32 (38%) hybrid versus 8 of 33 (24%) propofol-only patients; adjusted odds ratio 1.84; 95% confidence interval 0.56 to 6.04; p=0.31 (Very Low certainty). Overall, any postoperative rescue antiemetic (post-anesthesia care unit or 24 hours) was given to 16 of 32 hybrid versus 13 of 33 propofol-only patients (no adjusted estimate reported; Very Low certainty). - Pain and opioid use: Visual analog scale pain scores in the post-anesthesia care unit and total perioperative morphine milligram equivalents did not differ meaningfully between groups (for example, median total morphine equivalents 20 vs 27.5, p=0.19), and no clear analgesic or opioid-sparing advantage of the hybrid technique was demonstrated; these outcomes were not formally graded for certainty in this appraisal. - Risk of bias and certainty: Randomization, protocol adherence, and outcome measurement were generally robust, with low risk of bias for most domains. However, incomplete visibility of the prespecified analysis plan and selective emphasis on adjusted models led to an overall risk-of-bias judgement of "some concerns." All key postoperative nausea and vomiting and antiemetic outcomes (24-hour and post-anesthesia care unit) were rated Very Low certainty by GRADE because of this risk-of-bias concern and very serious imprecision from the small sample and wide confidence intervals. - Bottom line for practice: In high-risk adults undergoing elective laparoscopic surgery with standardized dexamethasone and ondansetron prophylaxis, hybrid low-dose propofol plus sevoflurane did not clearly reduce 24-hour postoperative nausea and vomiting, early postoperative nausea and vomiting, or rescue antiemetic use compared with propofol-only total intravenous anesthesia. Given the Very Low certainty of evidence, these findings should not on their own drive a change in anesthetic technique for the purpose of postoperative nausea and vomiting prevention; choice of hybrid versus propofol-only anesthesia should instead be guided by other clinical considerations until larger, higher-certainty trials are available.

Effect of intermittent recruitment manoeuvre combined with positive end expiratory pressure on postoperative atelectasis in obese patients: a randomized controlled trial

Citation: Qiao Y, Zhang H, Yu J, You J, Hou K, Zhao T, Feng H. Effect of intermittent recruitment manoeuvre combined with positive end expiratory pressure on postoperative atelectasis in obese patients: a randomized controlled trial. BMC Anesthesiology. 2026;[Epub ahead of print]. In a single-centre randomized controlled trial of obese adults undergoing elective laparoscopic sleeve gastrectomy, Qiao and colleagues compared three intraoperative ventilation strategies: zero end-expiratory pressure with a single recruitment manoeuvre, low positive end expiratory pressure alone, and the combination of positive end expiratory pressure plus intermittent recruitment manoeuvres. Moderate-certainty evidence suggests that the combined strategy substantially reduces computed tomography-defined postoperative atelectasis and improves intraoperative lung mechanics and oxygenation, but very-low-certainty data mean that any effects on short-term clinical complications such as hypoxemia or pneumonia remain highly uncertain. Study at a glance - Design and setting: Single-centre, parallel-group randomized controlled trial in obese adults (body mass index 30–60 kilograms per square metre), American Society of Anesthesiologists class I–III, aged 18–60 years, undergoing elective laparoscopic sleeve gastrectomy under general anaesthesia at a tertiary hospital in China. Ninety-five patients were randomised and ninety (30 per arm) were included in the analysed population; follow-up for outcomes extended to 48 hours postoperatively. Overall risk of bias was judged as having some concerns. - Interventions: Three intraoperative ventilation strategies were compared: (1) recruitment manoeuvre group: volume-controlled ventilation with zero end-expiratory pressure throughout and a single standardized recruitment manoeuvre at the end of surgery (positive end expiratory pressure ramped up to 30 centimetres of water for 30 seconds, then stepped down); (2) positive end expiratory pressure group: volume-controlled ventilation with 5 centimetres of water positive end expiratory pressure maintained until extubation, without recruitment manoeuvres; (3) combined group: 5 centimetres of water positive end expiratory pressure plus two recruitment manoeuvres (30 minutes after pneumoperitoneum creation and again at the end of surgery), using the same 30 centimetres of water pattern. Anaesthetic drugs, inspired oxygen fraction, and positioning were standardized across groups. - Primary outcome: postoperative atelectasis: The primary endpoint was incidence of computed tomography-defined postoperative atelectasis within 48 hours, with atelectasis defined as non-ventilated lung areas between −100 and +100 Hounsfield units on chest computed tomography. Atelectasis occurred in 14 of 30 patients (46.7%) in the recruitment manoeuvre group, 13 of 30 (43.3%) in the positive end expiratory pressure group, and 4 of 30 (13.3%) in the combined group, indicating a large relative reduction with the combined strategy. Among patients with atelectasis, the proportion of atelectatic lung volume was also lowest in the combined group (2.56%±0.51%) versus the recruitment manoeuvre group (5.94%±1.01%) and the positive end expiratory pressure group (4.13%±0.83%). Certainty of evidence for both incidence and extent of atelectasis was rated as moderate (downgraded one level for trial-level risk of bias). - Secondary outcomes: physiology and mechanics: Dynamic lung compliance at the end of surgery (T3) was higher with positive end expiratory pressure and with the combined strategy (22.42±2.88 and 23.37±3.06 millilitres per centimetre of water, respectively) than with recruitment manoeuvre alone (18.16±1.98). Arterial partial pressure of oxygen at 10 minutes after extubation (T4) was highest in the combined group (83.17±14.91 millimetres of mercury) compared with the positive end expiratory pressure group (72.47±12.52) and the recruitment manoeuvre group (68.13±10.00), and respiratory index at T4 was lowest (0.74±0.34 vs 1.22±0.44 and 1.33±0.49). These findings, supported by consistent trends in oxygenation index and pulse oximetry, are judged to be of moderate certainty (one-level downgrade for risk of bias). - Clinical pulmonary complications and safety: Postoperative hypoxemia, pneumonia, and pleural effusion within 48 hours were recorded as pulmonary complications, but exact per-arm event counts were only available in a supplementary table not fully accessible in the extracted data. Events appeared infrequent across all groups, and no clear between-group differences could be reliably quantified. Because of the small sample, sparse events, and incomplete numerical reporting, certainty for effects on these clinical complications is very low (downgraded one level for risk of bias and two levels for very serious imprecision). Hemodynamic parameters such as mean arterial pressure and heart rate, as well as peak airway pressures, remained broadly similar between groups, and no major safety signal related to recruitment manoeuvres or positive end expiratory pressure was reported.