Postoperative pain and neurocognitive outcomes after noncardiac surgery: a systematic review and dose-response meta-analysis.

BACKGROUND: Postoperative delirium (POD) and postoperative cognitive dysfunction (POCD) are common after noncardiac surgery. Postsurgical pain is frequent and can persist as chronic postsurgical pain (CPSP). The association between postsurgical pain and POD or POCD is biologically plausible. We conducted this systematic review to evaluate the association between acute postsurgical pain or CPSP and POD or POCD in adults undergoing noncardiac surgery. METHODS: We followed Preferred Reporting Items for Systematic Review and Meta-Analyses. We searched MEDLINE, EMBASE, Cochrane, CINAHL and PSYCHINFO up to May 2023. We included cohort, case-control, and cross-sectional studies of any language. Pairs of reviewers independently screened studies, extracted data and assessed the risk of bias using the CLARITY tool and the Joanna Briggs Institute checklist. We assessed the certainty of evidence using the Grading of Recommendations Assessment, Development, and Evaluation approach. Where possible, we conducted random-effects meta-analyses to summarise our findings. RESULTS: We analysed 30 studies (>9000 participants) that assessed the association between acute postoperative pain and POD/POCD. Dose-response meta-analyses found that postoperative pain intensity was associated with occurrence of POD (adjusted relative risk [aRR]/unit of pain intensity: 1.26; 95% confidence interval [CI]: 1.17-1.35; low certainty of evidence) and risk of developing POD (aRR/unit of pain intensity: 1.18; 95% CI: 1.08-1.30; low certainty of evidence). There was very low certainty of evidence regarding the association between postoperative pain and POCD. No studies assessed the association between CPSP and POCD. Residual confounding and substantial methodological variability between studies prevented pooling data from many of the included studies and lowered certainty of evidence. CONCLUSIONS: Dose-response meta-analyses found that postoperative pain intensity was associated with occurrence of and risk of developing POD. SYSTEMATIC REVIEW PROTOCOL: PROSPERO-CRD42021192105.

Effect of dexmedetomidine on postoperative high-sensitivity cardiac troponin T in patients undergoing video-assisted thoracoscopic surgery: a prospective, randomised controlled trial.

BACKGROUND: One-lung ventilation and intrathoracic operations during thoracoscopic surgery often result in intraoperative hypoxaemia and haemodynamic fluctuations, resulting in perioperative myocardial injury. Dexmedetomidine, an alpha-2 (α-2) agonist, has demonstrated myocardial protection. We hypothesize that the routine intravenous administration of dexmedetomidine could reduce the extent of myocardial injury during video-assisted thoracoscopic surgery (VATS). METHODS: The study included patients aged ≥ 45 years, classified as American Society of Anesthesiologists physical status I-III, who underwent general anesthesia for video-assisted thoracoscopic surgery. The patients were randomly assigned to either the intervention group, receiving general anesthesia with dexmedetomidine, or the control group, receiving general anesthesia without dexmedetomidine. Patients in the intervention group received a loading dose of dexmedetomidine (0.5 µg·kg-1) before anesthesia induction, followed by a continuous infusion (0.5 µg·kg-1·h-1) until the completion of the surgery. Placebos (saline) were administered for the control group to match the treatment. The primary outcome assessed was the high-sensitivity cardiac troponin T on postoperative day 1. Additionally, the incidence of myocardial injury after noncardiac surgery (MINS) was noted. RESULTS: A total of 110 participants completed this study. The median [interquartile range (IQR)] concentration of hs-cTnT on postoperative day 1 was lower in the intervention group compared with the control group (7 [6-9] vs. 8 [7-11] pg·ml-1; difference in medians,1 pg·ml-1; 95% confidence interval [CI], 0 to 2; P = 0.005). Similarly, on postoperative day 3, the median [IQR] concentration of hs-cTnT in the intervention group was also lower than that in the control group (6 [5-7] vs. 7 [6-9]; difference in medians,1 pg·ml-1; 95%CI, 0 to 2; P = 0.011). Although the incidence of MINS was not statistically significant (the intervention group vs. the control group, 3.8% vs. 9.1%, P = 0.465), there was a decreasing trend in the incidence of MINS in the intervention group. CONCLUSION: The administration of perioperative dexmedetomidine in patients ≥ 45 years undergoing video-assisted thoracoscopic surgery could lower the release of postoperative hs-cTnT without reducing incidence of myocardial injury. TRIAL REGISTRATION: chictr.org.cn (ChiCTR2200063193); prospectively registered 1 September 2022.

Desflurane Versus Sevoflurane and Postoperative Cardiac Biomarkers in Older Adults Undergoing Low- to Moderate-Risk Noncardiac Surgery-Secondary Analysis of a Prospective, Observer-Blinded, Randomized Clinical Trial.

Background/Objectives: Previous preclinical studies have shown that desflurane might have the most significant cardioprotective effect of all volatile anesthetics. However, data regarding the cardioprotective effects of desflurane versus sevoflurane are lacking. Therefore, we evaluated the effect of the maintenance of anesthesia using desflurane versus sevoflurane on the postoperative maximum concentrations of cardiac biomarkers in older adults undergoing low- to moderate-risk noncardiac surgery. Methods: In this secondary analysis of a prospective randomized trial, we included all 190 older adults undergoing low- to moderate-risk noncardiac surgery. Patients were randomized to receive desflurane or sevoflurane for the maintenance of anesthesia. We administered desflurane or sevoflurane, aiming at a BIS value of 50 ± 5. The cardiac-specific biomarkers included troponin T, NT-proBNP, and copeptin, which were measured preoperatively, within one hour after surgery, and on the second postoperative day. Results: There were no significant differences between the desflurane and sevoflurane groups in the postoperative maximum concentrations of troponin T (11 ng.L-1 [8; 16] versus 13 ng.L-1 [9; 18]; p = 0.595), NT-proBNP (196 pg.mL-1 [90; 686] versus 253 pg.mL-1 [134; 499]; p = 0.288), or copeptin (19 pmol.L-1 [7; 58] versus 12 pmol.L-1 [6; 41]; p = 0.096). We also observed no significant differences in the troponin T, NT-proBNP, or copeptin concentrations between the desflurane and sevoflurane groups at any measured timepoint (all p > 0.05). Conclusions: In contrast to preclinical studies, we did not observe a significant difference in the postoperative maximum concentrations of cardiac biomarkers. It seems likely that desflurane does not exert significant clinical meaningful cardioprotective effects in older adults. Thus, our results do not support the use of desflurane in patients undergoing low- to moderate-risk noncardiac surgery.

Outcomes after noncardiac surgery in patients with left ventricular assist devices: a systematic review.

Background: The number of patients living with left ventricular assist devices (LVADs) has gradually increased in the past decade. Non-cardiac surgery (NCS) in patients with LVAD poses a unique situation with its inherent challenges. Aim: We conducted a comprehensive review to investigate the perioperative complications and mortality associated with emergent or elective NCS in patients with LVAD. Method: A comprehensive literature search for any papers referring to continuous LVAD patients with NCS. All publications with at least five durable LVAD patients who had NCS were eligible for inclusion. Result: Twenty articles matching our criteria were found and included in our study. This systematic review included 6,476 LVAD patients who underwent 6,824 NCS. There were 5-3,216 LVAD patients with NCS in each study. The median age was between 39 and 65 years, and most of the patients (78.8%) were male. Thirty-day postoperative mortality ranged from 0% to 60%. Eight studies reported no death within the 30 days of the operation. Common complications include gastrointestinal (GI) bleeding, intracranial bleeding, infection, acute kidney injury (AKI), urinary tract infection (UTI), stroke, sepsis, pneumonia, and VAD exchange. Emergent abdominal surgery had the highest (up to 60%) mortality rate, and vascular and neurological operations had the highest complication rates. Due to the diverse range of patients in each publication and the combination of outcomes presented in various publications, a meta-analysis was not conducted. Conclusion: In LVAD patients, noncardiac surgery may be performed effectively and safely. LVAD patients who undergo non-cardiac surgery may require more transfusions due to their complex coagulopathies. However, perioperative management of LVAD patients undergoing emergent NCS should be optimized to reduce mortality. Systematic Review Registration: https://osf.io/fetsb/.

Heart failure diagnostic accuracy, intraoperative fluid management, and postoperative acute kidney injury: a single-centre prospective observational study.

BACKGROUND: The accurate diagnosis of heart failure (HF) before major noncardiac surgery is frequently challenging. The impact of diagnostic accuracy for HF on intraoperative practice patterns and clinical outcomes remains unknown. METHODS: We performed an observational study of adult patients undergoing major noncardiac surgery at an academic hospital from 2015 to 2019. A preoperative clinical diagnosis of HF was defined by keywords in the preoperative assessment or a diagnosis code. Medical records of patients with and without HF clinical diagnoses were reviewed by a multispecialty panel of physician experts to develop an adjudicated HF reference standard. The exposure of interest was an adjudicated diagnosis of heart failure. The primary outcome was volume of intraoperative fluid administered. The secondary outcome was postoperative acute kidney injury (AKI). RESULTS: From 40 659 surgeries, a stratified subsample of 1018 patients were reviewed by a physician panel. Among patients with adjudicated diagnoses of HF, those without a clinical diagnosis (false negatives) more commonly had preserved left ventricular ejection fractions and fewer comorbidities. Compared with false negatives, an accurate diagnosis of HF (true positives) was associated with 470 ml (95% confidence interval: 120-830; P=0.009) lower intraoperative fluid administration and lower risk of AKI (adjusted odds ratio:0.39, 95% confidence interval 0.18-0.89). For patients without adjudicated diagnoses of HF, non-HF was not associated with differences in either fluids administered or AKI. CONCLUSIONS: An accurate preoperative diagnosis of heart failure before noncardiac surgery is associated with reduced intraoperative fluid administration and less acute kidney injury. Targeted efforts to improve preoperative diagnostic accuracy for heart failure may improve perioperative outcomes.

2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM Guideline for Perioperative Cardiovascular Management for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.

AIM: The "2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM Guideline for Perioperative Cardiovascular Management for Noncardiac Surgery" provides recommendations to guide clinicians in the perioperative cardiovascular evaluation and management of adult patients undergoing noncardiac surgery. METHODS: A comprehensive literature search was conducted from August 2022 to March 2023 to identify clinical studies, reviews, and other evidence conducted on human subjects that were published in English from MEDLINE (through PubMed), EMBASE, the Cochrane Library, the Agency for Healthcare Research and Quality, and other selected databases relevant to this guideline. STRUCTURE: Recommendations from the "2014 ACC/AHA Guideline on Perioperative Cardiovascular Evaluation and Management of Patients Undergoing Noncardiac Surgery" have been updated with new evidence consolidated to guide clinicians; clinicians should be advised this guideline supersedes the previously published 2014 guideline. In addition, evidence-based management strategies, including pharmacological therapies, perioperative monitoring, and devices, for cardiovascular disease and associated medical conditions, have been developed.

Neuraxial anesthesia for patients with severe pulmonary arterial hypertension undergoing urgent open abdominal surgeries: two case reports.

BACKGROUND: There is no consensus regarding the choice of anesthetic method for patients with pulmonary hypertension (PH). We report two cases in which neuraxial anesthesia was safely performed without general anesthesia during open abdominal surgery in patients with severe PH. CASE PRESENTATION: Case 1: A 59-year-old woman had an atrial septal defect and a huge abdominal tumor with a mean pulmonary arterial pressure (PAP) of 39 mmHg and pulmonary vascular resistance (PVR) of 3.5 Wood units. Case 2: A 23-year-old woman who had hereditary pulmonary artery hypertension (mean PAP, 65 mmHg; PVR, 16.45 Wood units). Both patients underwent open abdominal surgery under neuraxial anesthesia without circulatory collapse with intraoperative administration of vasoconstrictors. CONCLUSION: Although anesthetic care must be personalized depending on the pathology and severity of PH, neuraxial anesthesia may be an option for patients with severe PH undergoing abdominal surgery.

Aspirin Monotherapy vs No Antiplatelet Therapy in Stable Patients With Coronary Stents Undergoing Low-to-Intermediate Risk Noncardiac Surgery.

BACKGROUND: Current guidelines recommend the perioperative continuation of aspirin in patients with coronary drug-eluting stents (DES) undergoing noncardiac surgery. However, supporting evidence is limited. OBJECTIVES: This study aimed to compare continuing aspirin monotherapy vs temporarily holding all antiplatelet therapy before noncardiac surgery in patients with previous DES implantation. METHODS: We randomly assigned patients who had received a DES >1 year previously and were undergoing elective noncardiac surgery either to continue aspirin or to discontinue all antiplatelet agents 5 days before noncardiac surgery. Antiplatelet therapy was recommended to be resumed no later than 48 hours after surgery, unless contraindicated. The primary outcome was a composite of death from any cause, myocardial infarction, stent thrombosis, or stroke between 5 days before and 30 days after noncardiac surgery. RESULTS: A total of 1,010 patients underwent randomization. Among 926 patients in the modified intention-to-treat population (462 patients in aspirin monotherapy group and 464 patients in the no-antiplatelet therapy group), the primary composite outcome occurred in 3 patients (0.6%) in the aspirin monotherapy group and 4 patients (0.9%) in the no antiplatelet group (difference, -0.2 percentage points; 95% CI: -1.3 to 0.9; P > 0.99). There was no stent thrombosis in either group. The incidence of major bleeding did not differ significantly between groups (6.5% vs 5.2%; P = 0.39), whereas minor bleeding was significantly more frequent in the aspirin group (14.9% vs 10.1%; P = 0.027). CONCLUSIONS: Among patients undergoing low-to-intermediate risk noncardiac surgery >1 year after stent implantation primarily with a DES, in the setting of lower-than-expected event rates, we failed to identify a significant difference between perioperative aspirin monotherapy and no antiplatelet therapy with respect to ischemic outcomes or major bleeding. (Perioperative Antiplatelet Therapy in Patients With Drug-eluting Stent Undergoing Noncardiac Surgery [ASSURE-DES]; NCT02797548).

2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM Guideline for Perioperative Cardiovascular Management for Noncardiac Surgery: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.

AIM: The "2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM Guideline for Perioperative Cardiovascular Management for Noncardiac Surgery" provides recommendations to guide clinicians in the perioperative cardiovascular evaluation and management of adult patients undergoing noncardiac surgery. METHODS: A comprehensive literature search was conducted from August 2022 to March 2023 to identify clinical studies, reviews, and other evidence conducted on human subjects that were published in English from MEDLINE (through PubMed), EMBASE, the Cochrane Library, the Agency for Healthcare Research and Quality, and other selected databases relevant to this guideline. STRUCTURE: Recommendations from the "2014 ACC/AHA Guideline on Perioperative Cardiovascular Evaluation and Management of Patients Undergoing Noncardiac Surgery" have been updated with new evidence consolidated to guide clinicians; clinicians should be advised this guideline supersedes the previously published 2014 guideline. In addition, evidence-based management strategies, including pharmacological therapies, perioperative monitoring, and devices, for cardiovascular disease and associated medical conditions, have been developed.

Desflurane versus sevoflurane anesthesia and postoperative recovery in older adults undergoing minor- to moderate-risk noncardiac surgery - A prospective, randomized, observer-blinded, clinical trial.

STUDY OBJECTIVE: The effect of volatile anesthetics on postoperative recovery in older adults is still not entirely clear. Thus, we evaluated the effect of desflurane versus sevoflurane anesthesia on speed of postoperative recovery in older adults eligible for same-day discharge. We further evaluated the incidence of postoperative nausea and vomiting (PONV), bispectral index (BIS) values, and S100B concentrations. DESIGN: Single-center, prospective, observer-blinded, randomized clinical trial. SETTING: Operating room. PATIENTS: 190 patients ≥65 years of age and scheduled for minor- to moderate-risk noncardiac surgeries. INTERVENTIONS: Goal-directed administration of desflurane versus sevoflurane for maintenance of anesthesia with an intraoperative goal of BIS 50 ± 5. MEASUREMENTS: The primary outcome was the time to anesthesia recovery, which was defined as the time between arrival at the post-anesthesia care unit (PACU) and reaching criteria for discharge from PACU, based on modified Aldrete score ≥ 12 points. Modified Aldrete scores were assessed at PACU arrival and thereafter in five-minute intervals. PONV was evaluated during PACU stay and the first three postoperative days, BIS values were recorded during PACU stay, and S100B values were measured before and after surgery, and on the second postoperative day. MAIN RESULTS: 95 patients were randomized to receive desflurane, and 95 patients to receive sevoflurane. We did not observe a significant difference in median duration of postoperative recovery between the groups (desflurane: 0 min [0;0]; sevoflurane: 0 min [0;0]; p = 0.245). 77 patients (81.1%) in the desflurane group and 84 patients (88.4%) in the sevoflurane group already had Aldrete scores ≥12 points upon arrival at PACU (p = 0.277). There was also no significant difference in the incidences of PONV (p = 0.606), postoperative BIS values (p = 0.197), and postoperative maximum S100B concentrations (p = 0.821) between the groups. CONCLUSIONS: Despite previous reports, we did not observe significant faster recovery times after desflurane anesthesia. Both volatile anesthetics may be appropriate for same-day discharge in older adults.

Visualising myocardial injury after noncardiac surgery: a case series using postoperative cardiovascular MRI.

Myocardial injury after noncardiac surgery (MINS) and perioperative myocardial injury are associated with increased morbidity and mortality. Both are diagnosed by a perioperative increase in troponin, yet there is controversy if MINS is a genuine myocardial insult. We applied postoperative cardiovascular magnetic resonance T2 mapping techniques to visualise acute myocardial injury (i.e. oedema) in six patients with multiple cardiovascular risk factors who underwent aortic surgery. The burden of myocardial oedema was substantially higher in four patients with elevated troponin qualifying for MINS, compared with patients without MINS. The data and images suggest that MINS represents genuine myocardial injury.

Assessment of perioperative cardiac risk using preoperative quantitative flow ratio in patients with coronary artery disease undergoing noncardiac surgery: a retrospective cohort study.

Background: Quantitative flow ratio (QFR) is a novel diagnostic modality for the functional testing of coronary artery stenosis, but evidence concerning the postoperative prognostic implication of QFR in noncardiac surgery (NCS) of patients with coronary artery disease (CAD) is limited. The purpose of this study was to examine the role of QFR in perioperative risk prediction in patients with coronary heart disease. Methods: This retrospective cohort study was conducted in The First Affiliated Hospital of Wenzhou Medical University between 2013 and 2022, and consecutively included patients with CAD who had undergone NCS <1 year after coronary angiography. The primary endpoint was major adverse cardiovascular events (MACEs), which were defined as a composite of cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, cardiopulmonary arrest, malignant ventricular arrhythmia (MVA), congestive heart failure, and revascularization. Univariate and multifactorial Cox regression was used to identify the independent risk factors for perioperative cardiovascular events and to construct new models. The area under the curve (AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were used to compare the newly constructed model with existing traditional models. Results: Among the 929 participants enrolled (median age 68 years; 72.0% male), the primary endpoint was met in 67 (7.2%) patients within 30 days of follow-up. There was no significant difference in the incidence of the primary endpoint between patients with QFR <0.75 and those with "gray zone" lesions (0.75≤ QFR ≤0.8) (log-rank P=0.325). Patients with QFR <0.75 and those with "gray zone" lesions (0.75≤ QFR ≤0.8) had a higher incidence of primary endpoint events compared to patients with QFR >0.8. [QFR <0.75 vs. QFR >0.8: adjusted hazard ratio (HR) =20.70, P<0.001; 0.75≤ QFR ≤0.8 vs. QFR >0.8: HR =15.99, P<0.001]. The independent predictors of MACEs events within 30 days after NCS were albumin level [HR =0.92, 95% confidence interval (CI): 0.87-0.98; P=0.008], emergency surgery (HR =4.12, 95% CI: 1.66-10.23; P=0.002), and QFR ≤0.8 (HR =15.92, 95% CI: 5.96-42.51; P<0.001). In addition, adjusting the original Revised Cardiac Risk Index (RCRI) with QFR ≤0.8 as a risk factor significantly improved the risk stratification of postoperative adverse events, with the adjusted AUC rising from 0.574 to 0.740 (P<0.001). Conclusions: QFR ≤0.8 could independently predict perioperative cardiovascular adverse events in patients with CAD undergoing NCS and improve the predictive value of original predictive index. Gray-zone lesions (0.75≤ QFR ≤0.8) should be actively treated.

Risk factors for acute kidney injury in preterm neonates after noncardiac surgery: a single-center retrospective cohort study.

Postoperative acute kidney injury (AKI) is a common complication that is associated with chronic kidney disease, early postsurgical mortality, and prolonged hospital stays. Preterm neonates who undergo surgery are at risk factors for AKI due to underdeveloped kidneys. To date, little is known about the incidence and perioperative risk factors for AKI in preterm neonates undergoing noncardiac surgery. Preterm neonates who underwent noncardiac surgery between January May 1, 2020, and February 28, 2023, were enrolled in the trial according to the inclusion criteria. Both multivariable and logistic regression analyses were used to analyze the associations between characteristic data and AKI. In total, 106 preterm neonates met the inclusion criteria, and 25 preterm neonates (23.6%) developed postoperative AKI. Multivariate analysis revealed that the factors associated with AKI were gestational age < 32 weeks [OR: 4.88; 95% CI (1.23-19.42)], preoperative sepsis [OR: 3.98; 95% CI (1.29-12.28)], and intraoperative hypotension [OR: 3.75; 95% CI (1.26-11.15)]. Preterm neonates who developed AKI were more likely to have longer hospital length of stays (38 [18,69] days vs. 21[12,46]) and higher medical costs (93,181.6 [620450.0,173,219.0] ï¿¥ vs. 58,134.6 [31015.1,97,224,1) ï¿¥ than neonates who did not develop AKI. Preterm neonates who underwent noncardiac surgery had a high incidence of AKI. Independent risk factors for AKI in preterm neonates who underwent noncardiac surgery were low gestational age, preoperative sepsis, and intraoperative hypotension. Preterm neonates who developed AKI were more likely to have longer hospital stays and higher medical costs.

Coronary heart disease increases the risk of perioperative ischemic stroke after noncardiac surgery: A retrospective cohort study.

OBJECTIVE: To investigate the association between coronary heart disease (CHD) and the risk of perioperative ischemic stroke in patients undergoing noncardiac surgery. METHODS: This retrospective study evaluated the incidence of ischemic stroke within 30 days after a noncardiac surgery. A cohort of 221,541 patients who underwent noncardiac surgery between January 2008 and August 2019 was segregated according to whether they were diagnosed with CHD. Primary, sensitivity, and subgroup logistic regression analyses were conducted to confirm that CHD is an independent risk factor for perioperative ischemic stroke. Propensity score matching analysis was used to account for the potential residual confounding effect of covariates. RESULTS: Among the 221,541 included patients undergoing noncardiac surgery, 484 patients (0.22%) experienced perioperative ischemic stroke. The risk of perioperative ischemic stroke was higher in patients with CHD (0.7%) compared to patients without CHD (0.2%), and multivariate logistic regression analysis showed that CHD was associated with a significantly increased risk of perioperative ischemic stroke (odds ratio (OR), 3.7943; 95% confidence interval (CI) 2.865-4.934; p < 0.001). In a subset of patients selected by propensity score matching (PSM) in which all covariates between the two groups were well balanced, the association between CHD and increased risk of perioperative ischemic stroke remained significantly significant (OR 1.8150; 95% CI, 1.254-2.619; p = 0.001). In the subgroup analysis stratified by age, preoperative ß-blockers, and fibrinogen-to-albumin ratio (FAR), the association between CHD and perioperative ischemic stroke was stable (p for interaction >0.05). Subgroup analyses also showed that CHD was significantly increased the risk of perioperative ischemic stroke in the preoperative mean arterial pressure (MAP) ≥94.2 mmHg subgroups (p for interaction <0.001). CONCLUSION: CHD is significantly associated with an increased risk of perioperative ischemic stroke and is an independent risk factor for perioperative ischemic stroke after noncardiac surgery. Strict control of preoperative blood pressure may reduce the risk of perioperative ischemic stroke for patients with CHD undergoing noncardiac surgery.

Development and Validation of an Explainable Machine Learning Model for Predicting Myocardial Injury After Noncardiac Surgery in Two Centers in China: Retrospective Study.

Background: Myocardial injury after noncardiac surgery (MINS) is an easily overlooked complication but closely related to postoperative cardiovascular adverse outcomes; therefore, the early diagnosis and prediction are particularly important. Objective: We aimed to develop and validate an explainable machine learning (ML) model for predicting MINS among older patients undergoing noncardiac surgery. Methods: The retrospective cohort study included older patients who had noncardiac surgery from 1 northern center and 1 southern center in China. The data sets from center 1 were divided into a training set and an internal validation set. The data set from center 2 was used as an external validation set. Before modeling, the least absolute shrinkage and selection operator and recursive feature elimination methods were used to reduce dimensions of data and select key features from all variables. Prediction models were developed based on the extracted features using several ML algorithms, including category boosting, random forest, logistic regression, naïve Bayes, light gradient boosting machine, extreme gradient boosting, support vector machine, and decision tree. Prediction performance was assessed by the area under the receiver operating characteristic (AUROC) curve as the main evaluation metric to select the best algorithms. The model performance was verified by internal and external validation data sets with the best algorithm and compared to the Revised Cardiac Risk Index. The Shapley Additive Explanations (SHAP) method was applied to calculate values for each feature, representing the contribution to the predicted risk of complication, and generate personalized explanations. Results: A total of 19,463 eligible patients were included; among those, 12,464 patients in center 1 were included as the training set; 4754 patients in center 1 were included as the internal validation set; and 2245 in center 2 were included as the external validation set. The best-performing model for prediction was the CatBoost algorithm, achieving the highest AUROC of 0.805 (95% CI 0.778-0.831) in the training set, validating with an AUROC of 0.780 in the internal validation set and 0.70 in external validation set. Additionally, CatBoost demonstrated superior performance compared to the Revised Cardiac Risk Index (AUROC 0.636; P<.001). The SHAP values indicated the ranking of the level of importance of each variable, with preoperative serum creatinine concentration, red blood cell distribution width, and age accounting for the top three. The results from the SHAP method can predict events with positive values or nonevents with negative values, providing an explicit explanation of individualized risk predictions. Conclusions: The ML models can provide a personalized and fairly accurate risk prediction of MINS, and the explainable perspective can help identify potentially modifiable sources of risk at the patient level.

Implementation of a software-based decision support tool for guideline-appropriate preoperative evaluation: a prospective agreement study.

BACKGROUND: Guideline adherence in the medical field leaves room for improvement. Digitalised decision support helps improve compliance. However, the complex nature of the guidelines makes implementation in clinical practice difficult. METHODS: This single-centre prospective study included 204 adult ASA physical status 3-4 patients undergoing elective noncardiac surgery at a German university hospital. Agreement of clearance for surgery between a guideline expert and a digital guideline support tool was investigated. The decision made by the on-duty anaesthetists (standard approach) was assessed for agreement with the expert in a cross-over design. The main outcome was the level of agreement between digital guideline support and the expert. RESULTS: The digital guideline support approach cleared 18.1% of the patients for surgery, the standard approach cleared 74.0%, and the expert approach cleared 47.5%. Agreement of the expert decision with digital guideline support (66.7%) and the standard approach (67.6%) was fair (Cohen's kappa 0.37 [interquartile range 0.26-0.48] vs 0.31 [0.21-0.42], P=0.6). Taking the expert decision as a benchmark, correct clearance using digital guideline support was 50.5%, and correct clearance using the standard approach was 44.6%. Digital guideline support incorrectly asked for additional examinations in 31.4% of the patients, whereas the standard approach did not consider conditions that would have justified additional examinations before surgery in 29.4%. CONCLUSIONS: Strict guideline adherence for clearance for surgery through digitalised decision support inadequately considered patients, clinical context. Vague formulations, weak recommendations, and low-quality evidence complicate guideline translation into explicit rules. CLINICAL TRIAL REGISTRATION: NCT04058769.

Associations of inflammatory biomarkers with morbidity and mortality after noncardiac surgery: A systematic review and meta-analysis.

BACKGROUND: Noncardiac surgery is associated with an inflammatory response. Whether increased inflammation in the perioperative period is associated with subsequent morbidity and mortality is unknown. METHODS: MEDLINE, EMBASE, and CENTRAL were systematically searched from date of inception until May 2023. Longitudinal studies were included if they reported multivariable adjusted associations of biomarkers measured preoperatively and/or within 10 days after surgery with at least one prespecified adverse outcome in noncardiac surgery patients. Data were extracted independently and in duplicate. Risk estimates were pooled using DerSimonian-Laird random-effects models and reported as summary odds ratios (ORs) with 95% CIs. The outcomes were all-cause mortality and major adverse cardiovascular events. RESULTS: Fifty-two studies with a total of 121,849 patients were included. The median follow-up was 56 [IQR, 28-63] months and the average age was 57 (±3) years. Elevated preoperative C-reactive protein (CRP) levels were associated with a higher risk of mortality (OR 1.57, 95% CI 1.29-1.90, I2 = 93%, 28 studies). This association was stronger in non-cancer surgery populations (OR 2.10, 95% CI 1.92-2.31, I2 = 0%, 4 studies) when compared to cancer surgery populations (OR 1.51, 95% CI 1.26-1.81, I2 = 83%, 24 studies) (p for subgroup difference = 0.001). Similarly, higher postoperative CRP levels were associated with all-cause mortality (OR 1.61, 95% CI 1.17-2.20, I2 = 90%, 7 studies). Higher preoperative CRP levels were associated with major cardiovascular events (OR 2.11, 95% CI 1.51-2.94, I2 = 0%, 2 studies). Other preoperatively measured biomarkers associated with all-cause mortality were fibrinogen (OR 1.48, 95% CI 1.05-2.09, I2 = 52%, 5 studies), interleukin-6 (OR 1.17, 95% CI 1.07-1.28, I2 = 27%, 3 studies), and tumour necrosis factor-alpha (OR 1.37, 95% CI 1.16-1.61, I2 = 0%, 2 studies). CONCLUSION AND RELEVANCE: Inflammatory biomarker levels in the perioperative period were associated with all-cause mortality and adverse cardiovascular events in patients undergoing noncardiac surgery.