It has been widely reported that the serum anion gap is significantly associated with mortality in intensive care unit (ICU); however, it remains unknown whether the association is present in aortic aneurysm (AA) patients. The present study aimed to investigate the association between the admission serum anion gap and ICU mortality in AA patients. Data extracted from a publicly accessible clinical database using a modifiable data mining technique were analyzed retrospectively, mainly by employing multivariable logistic regression analysis. The primary study outcome was ICU mortality. A total of 273 patient records were analyzed. The ICU mortality was 8.79% (24/273). The median serum anion gap was significantly higher in non‑survivors [17.50 mEq/l, interquartile range (IQR) 15.75‑22.50 mEq/l] compared with survivors [13.00 mEq/l, IQR 11.00‑15.00 mEq/l, P<0.001]. Multivariate analysis resulted in identification of a clear association between admission serum anion gap and ICU mortality in AA patients [odds ratio (OR) 1.38 per 1 mEq/l increase, 95% confidence interval (CI) 1.08‑1.76]. The area under the receiver operating characteristic curve showed an outstanding discrimination ability in predicting ICU mortality (area under curve 0.8513, 95% CI 0.7698‑0.9328). In conclusion, admission serum anion gap may serve as a strong predictor of ICU mortality for AA patients.

Background: There has not been a well-accepted prognostic model to predict the mortality of aortic aneurysm patients in intensive care unit after open surgery repair. Otherwise, our previous study found that anion gap was a prognosis factor for aortic aneurysm patients. Therefore, we wanted to investigate the relationship between anion gap and mortality of aortic aneurysm patients in intensive care unit after open surgery repair. Methods: From Medical Information Mart for Intensive Care III, data of aortic aneurysm patients in intensive care unit after open surgery were enrolled. The primary clinical outcome was defined as death in intensive care unit. Univariate analysis was conducted to compare the baseline data in different groups stratified by clinical outcome or by anion gap level. Restricted cubic spline was drawn to find out the association between anion gap level and mortality. Subgroup analysis was then conducted to show the association in different level and was presented as frost plot. Multivariate regression models were built based on anion gap and were adjusted by admission information, severity score, complication, operation and laboratory indicators. Receiver operating characteristic curves were drawn to compare the prognosis ability of anion gap and simplified acute physiology score II. Decision curve analysis was finally conducted to indicate the net benefit of the models. Results: A total of 405 aortic aneurysm patients were enrolled in this study and the in-intensive-care-unit (in-ICU) mortality was 6.9%. Univariate analysis showed that elevated anion gap was associated with high mortality (P value < 0.001), and restricted cubic spline analysis showed the positive correlation between anion gap and mortality. Receiver operating characteristic curve showed that the mortality predictive ability of anion gap approached that of simplified acute physiology score II and even performed better in predicting in-hospital mortality (P value < 0.05). Moreover, models based on anion gap showed that 1 mEq/L increase of anion gap improved up to 42.3% (95% confidence interval 28.5–59.8%) risk of death. Conclusions: The level of serum anion gap was an important prognosis factor for aortic aneurysm mortality in intensive care unit after open surgery.

Background: Prevention is highly involved in reducing the incidence of post-thrombotic syndrome (PTS). We aimed to develop accurate models with machine learning (ML) algorithms to predict whether PTS would occur within 24 months.Materials and methods: The clinical data used for model building were obtained from the Acute Venous Thrombosis: Thrombus Removal with Adjunctive Catheter-Directed Thrombolysis study and the external validation cohort was acquired from the Sun Yat-sen Memorial Hospital in China. The main outcome was defined as the occurrence of PTS events (Villalta score ≥5). Twenty-three clinical variables were included, and four ML algorithms were applied to build the models. For discrimination and calibration, F scores were used to evaluate the prediction ability of the models. The external validation cohort was divided into ten groups based on the risk estimate deciles to identify the hazard threshold.Results: In total, 555 patients with deep vein thrombosis (DVT) were included to build models using ML algorithms, and the models were further validated in a Chinese cohort comprising 117 patients. When predicting PTS within 2 years after acute DVT, logistic regression based on gradient descent and L1 regularization got the highest area under the curve (AUC) of 0.83 (95% CI:0.76–0.89) in external validation. When considering model performance in both the derivation and external validation cohorts, the eXtreme gradient boosting and gradient boosting decision tree models had similar results and presented better stability and generalization. The external validation cohort was divided into low, intermediate, and high-risk groups with the prediction probability of 0.3 and 0.4 as critical points.Conclusion: Machine learning models built for PTS had accurate prediction ability and stable generalization, which can further facilitate clinical decision-making, with potentially important implications for selecting patients who will benefit from endovascular surgery.

Objectives: We aimed to use machine learning (ML) algorithms to risk stratify the prognosis of critical pulmonary embolism (PE). Material and methods: In total, 1229 patients were obtained from MIMIC-IV database. Main outcomes were set as all-cause mortality within 30 days. Logistic regression (LR) and simplified eXtreme gradient boosting (XGBoost) were applied for model constructions. We chose the final models based on their matching degree with data. To simplify the model and increase its usefulness, finally simplified models were built based on the most important 8 variables. Discrimination and calibration were exploited to evaluate the prediction ability. We stratified the risk groups based on risk estimate deciles. Results: The simplified XGB model performed better in model discrimination, which AUC were 0.82 (95% CI: 0.78–0.87) in the validation cohort, compared with the AUC of simplified LR model (0.75 [95% CI: 0.69—0.80]). And XGB performed better than sPESI in the validation cohort. A new risk-classification based on XGB could accurately predict low-risk of mortality, and had high consistency with acknowledged risk scores. Conclusions: ML models can accurately predict the 30-day mortality of critical PE patients, which could further be used to reduce the burden of ICU stay, decrease the mortality and improve the quality of life for critical PE patients.

Peripheral artery disease (PAD) is often asymptomatic but increases the risk of developing cardiovascular events. Due to the uncertainties regarding the quality of related guidelines and a lack of clear-cut evidence, we performed a systematic review and critical appraisal of these guidelines to evaluate their consistency of the recommendations in asymptomatic PAD population. Guidelines in English between January 1st, 2000 to December 31th, 2017 were screened in databases including Medline via PubMed, EMBASE, the G-I-N International Guideline Library, the National Guidelines Clearinghouse, the Canadian Medication Association Infobase and the National Library for Health. Those guidelines containing recommendations on screening and treatment for asymptomatic PAD were included, and three reviewers evaluated the quality of the guidelines using Appraisal of Guidelines Research and Evaluation (AGREE) II instrument. Related recommendations were then fully extracted and compared by two reviewers. Fourteen guidelines were included finally and the AGREE scores ranged from 39 to 73%. Most of included guidelines scored low in Rigor of development and Editorial independence, and only two guidelines (ACCF/AHA, AHA/ACC) reached the standard on Conflict of Interest from Institute of Medicine (IOM). Eight guidelines recommended screening at different strength while the others found insufficient evidence or were against screening. Conflicting recommendations on treatment were found in the target value of the lipid lowering and antiplatelet therapy. The treatment policies in three guidelines (BWG, CEVF, ESC) appeared more aggressive, but they had low transparency between guideline developer and industry or did not reach the standard of IOM. Current guidelines on asymptomatic PAD varied in the methodological quality and fell short of the standard in the rigor of development and editorial independence. Conflicting recommendations were found both on the screening and treatment. More effort is needed to provide clear-cut evidences with high quality and transparency among guideline developer and industry.

Current metabolomic aging clocks inadequately capture individual heterogeneity in biological aging trajectories, constraining their clinical utility. Here, we developed a metabolomic age clock in the UK Biobank (n = 196,790) using a comprehensive panel of 249 plasma metabolites. This framework was trained to predict phenotypic age (PhenoAge), a validated composite biomarker that integrates clinical chemistry across multiple systems, and was evaluated for its utility to predict incident cardiovascular diseases (CVDs) and dementia. We found that this new measure accurately predicted actual PhenoAge (Pearson's r = 0.90) and was significantly associated with the incidence of seven CVDs, including major adverse cardiovascular events, atherosclerotic cardiovascular disease, myocardial infarction, stroke, aortic stenosis, heart failure, and abdominal aortic aneurysm, but not dementia. Furthermore, metabolomic aging was associated with biological, physical, and cognitive age-related phenotypes, comprising telomere length, frailty index, and reaction time. Incorporating the metabolomic age clock with PREVENT (Predicting Risk of CVD Events) risk score modestly improved the performance, as measured by C-statistic and net reclassification index. Genetic analyses revealed 91 genomic loci and 168 genes (e.g., SERPINA1, FADS cluster), with tissue-enrichment analysis highlighting the liver's significant role in metabolic aging. By bridging metabolomic profiles with multisystem aging information, this framework provides a measure of biological aging that is associated with age-related functional status and cardiovascular risk.