Development and validation of an interpretable machine learning model-Predicting mild cognitive impairment in a high-risk stroke population.

Background: Mild cognitive impairment (MCI) is considered a preclinical stage of Alzheimer's disease (AD). People with MCI have a higher risk of developing dementia than healthy people. As one of the risk factors for MCI, stroke has been actively treated and intervened. Therefore, selecting the high-risk population of stroke as the research object and discovering the risk factors of MCI as early as possible can prevent the occurrence of MCI more effectively. Methods: The Boruta algorithm was used to screen variables, and eight machine learning models were established and evaluated. The best performing models were used to assess variable importance and build an online risk calculator. Shapley additive explanation is used to explain the model. Results: A total of 199 patients were included in the study, 99 of whom were male. Transient ischemic attack (TIA), homocysteine, education, hematocrit (HCT), diabetes, hemoglobin, red blood cells (RBC), hypertension, prothrombin time (PT) were selected by Boruta algorithm. Logistic regression (AUC = 0.8595) was the best model for predicting MCI in high-risk groups of stroke, followed by elastic network (ENET) (AUC = 0.8312), multilayer perceptron (MLP) (AUC = 0.7908), extreme gradient boosting (XGBoost) (AUC = 0.7691), and support vector machine (SVM) (AUC = 0.7527), random forest (RF) (AUC = 0.7451), K-nearest neighbors (KNN) (AUC = 0.7380), decision tree (DT) (AUC = 0.6972). The importance of variables suggests that TIA, diabetes, education, and hypertension are the top four variables of importance. Conclusion: Transient ischemic attack (TIA), diabetes, education, and hypertension are the most important risk factors for MCI in high-risk groups of stroke, and early intervention should be performed to reduce the occurrence of MCI.

Improvement of Definite Diagnosis of Familial Hypercholesterolemia Using an Expanding Genetic Analysis.

Background: The deeper understanding of the complex hereditary basis of familial hypercholesterolemia (FH) has raised the rationale of genetic testing, which has been underutilized in clinical practice. Objectives: The present study aimed to explore the variant spectrum of FH in an expanding manner and compare its diagnostic performance. Methods: A total of 169 Chinese individuals (124 index cases and 45 relatives) with clinical definite/probable FH were consecutively enrolled. Next-generation sequencing was performed for genetic analysis of 9 genes associated with hypercholesterolemia (major genes: LDLR, APOB, and PCSK9; minor genes: LDLRAP1, LIPA, STAP1, APOE, ABCG5, and ABCG8) including the evaluations of small-scale variants and large-scale copy number variants (CNVs). Results: Among the 169 clinical FH patients included, 98 (58.0%) were men. A total of 85 (68.5%) index cases carried FH-associated variants. The proportion of FH caused by small-scale variants in LDLR, APOB, and PCSK9 genes was 62.1% and then increased by 6.5% when other genes and CNVs were further included. Furthermore, the variants in LDLR, APOB, and PCSK9 genes occupied 75% of all FH-associated variants. Of note, there were 8 non-LDLR CNVs detected in the present study. Conclusions: LDLR, APOB, and PCSK9 genes should be tested in the initial genetic screening, although variants in minor genes also could explain phenotypic FH, suggesting that an expanding genetic testing may be considered to further explain phenotypic FH.

Activation of the CaR-CSE/H2S pathway confers cardioprotection against ischemia-reperfusion injury.

Ischemia-reperfusion (I/R) injury is a multifactorial process triggered when an organ is subjected to transiently reduced blood supply. The result is a cascade of pathological complications and organ damage due to the production of reactive oxygen species following reperfusion. The present study aims to evaluate the role of activated calcium-sensing receptor (CaR)-cystathionine γ-lyase (CSE)/hydrogen sulfide (H2S) pathway in I/R injury. Firstly, an I/R rat model with CSE knockout was constructed. Transthoracic echocardiography, TTC and HE staining were performed to determine the cardiac function of rats following I/R Injury, followed by TUNEL staining observation on apoptosis. Besides, with the attempt to better elucidate how CaR-CSE/H2S affects I/R, in-vitro culture of human coronary artery endothelial cells (HCAECs) was conducted with gadolinium chloride (GdCl3, a CaR agonist), H2O2, siRNA against CSE (siCSE), or W7 (a CaM inhibitor). The interaction between CSE and CaM was subsequently detected. Plasma oxidative stress indexes, H2S and CSE, and apoptosis-related proteins were all analyzed following cell apoptosis. We found that H2S elevation led to the improvement whereas CSE knockdown decreased cardiac function in rats with I/R injury. Moreover, oxidative stress injury in I/R rats with CSE knockout was aggravated, while the increased expression of H2S and CSE in the aortic tissues resulted in alleviated the oxidative stress injury. Moreover, increased H2S and CSE levels were found to inhibit cell apoptotic ability in the aortic tissues after I/R injury, thus attenuating oxidative stress injury, accompanied by inhibited expression of apoptosis-related proteins. In HCAECs following oxidative stress treatment, siCSE and CaM inhibitor were observed to reverse the protection of CaR agonist. Coimmunoprecipitation assay revealed the interaction between CSE and CaM. Taken together, all above-mentioned data provides evidence that activation of the CaR-CSE/H2S pathway may confer a potent protective effect in cardiac I/R injury.

Accurate Prediction of Coronary Heart Disease for Patients With Hypertension From Electronic Health Records With Big Data and Machine-Learning Methods: Model Development and Performance Evaluation.

BACKGROUND: Predictions of cardiovascular disease risks based on health records have long attracted broad research interests. Despite extensive efforts, the prediction accuracy has remained unsatisfactory. This raises the question as to whether the data insufficiency, statistical and machine-learning methods, or intrinsic noise have hindered the performance of previous approaches, and how these issues can be alleviated. OBJECTIVE: Based on a large population of patients with hypertension in Shenzhen, China, we aimed to establish a high-precision coronary heart disease (CHD) prediction model through big data and machine-learning. METHODS: Data from a large cohort of 42,676 patients with hypertension, including 20,156 patients with CHD onset, were investigated from electronic health records (EHRs) 1-3 years prior to CHD onset (for CHD-positive cases) or during a disease-free follow-up period of more than 3 years (for CHD-negative cases). The population was divided evenly into independent training and test datasets. Various machine-learning methods were adopted on the training set to achieve high-accuracy prediction models and the results were compared with traditional statistical methods and well-known risk scales. Comparison analyses were performed to investigate the effects of training sample size, factor sets, and modeling approaches on the prediction performance. RESULTS: An ensemble method, XGBoost, achieved high accuracy in predicting 3-year CHD onset for the independent test dataset with an area under the receiver operating characteristic curve (AUC) value of 0.943. Comparison analysis showed that nonlinear models (K-nearest neighbor AUC 0.908, random forest AUC 0.938) outperform linear models (logistic regression AUC 0.865) on the same datasets, and machine-learning methods significantly surpassed traditional risk scales or fixed models (eg, Framingham cardiovascular disease risk models). Further analyses revealed that using time-dependent features obtained from multiple records, including both statistical variables and changing-trend variables, helped to improve the performance compared to using only static features. Subpopulation analysis showed that the impact of feature design had a more significant effect on model accuracy than the population size. Marginal effect analysis showed that both traditional and EHR factors exhibited highly nonlinear characteristics with respect to the risk scores. CONCLUSIONS: We demonstrated that accurate risk prediction of CHD from EHRs is possible given a sufficiently large population of training data. Sophisticated machine-learning methods played an important role in tackling the heterogeneity and nonlinear nature of disease prediction. Moreover, accumulated EHR data over multiple time points provided additional features that were valuable for risk prediction. Our study highlights the importance of accumulating big data from EHRs for accurate disease predictions.

Predicting Cardiovascular Outcomes by Baseline Lipoprotein(a) Concentrations: A Large Cohort and Long-Term Follow-up Study on Real-World Patients Receiving Percutaneous Coronary Intervention.

Background Although several studies have indicated that lipoprotein(a) is a useful prognostic predictor for patients following percutaneous coronary intervention (PCI), previous observations have somewhat been limited by either small sample size or short-term follow-up. Hence, this study aimed to evaluate the impact of lipoprotein(a) on long-term outcomes in a large cohort of stable coronary artery disease patients after PCI. Methods and Results In this multicenter and prospective study, we consecutively enrolled 4078 stable coronary artery disease patients undergoing PCI from March 2011 to March 2016. They were categorized according to both the median of lipoprotein(a) levels and lipoprotein(a) values of <15 (low), 15 to 30 (medium), and ≥30 mg/dL (high). All patients were followed up for occurrence of cardiovascular events, including cardiovascular death, nonfatal myocardial infarction, and stroke. During an average of 4.9 years of follow-up, 315 (7.7%) cardiovascular events occurred. The events group had significantly higher lipoprotein(a) levels than the nonevents group. Compared with the low lipoprotein(a) group, Kaplan-Meier analysis showed that the high lipoprotein(a) group had a significantly lower cumulative event-free survival rate, and multivariate Cox regression analysis further revealed that the high lipoprotein(a) group had significantly increased cardiovascular events risk. Moreover, adding continuous or categorical lipoprotein(a) to the Cox model led to a significant improvement in C-statistic, net reclassification, and integrated discrimination. Conclusions With a large sample size and long-term follow-up, our data confirmed that high lipoprotein(a) levels could be associated with a poor prognosis after PCI in stable coronary artery disease patients, suggesting that lipoprotein(a) measurements may be useful for patient risk stratification before selective PCI.

Differences In Gastrointestinal Safety Profiles Among Novel Oral Anticoagulants: Evidence From A Network Meta-Analysis.

BACKGROUND: There is no consensus at present regarding the differences in the risk of GI bleeding across various NOAC regimens. Therefore, we performed a network meta-analysis to compare the risk of gastrointestinal bleeding after different NOAC regimens. METHODS: PubMed, Cochrane, Web of Science, Clinicaltrial.gov and Clinicaltrialresults.org were searched for randomized controlled trials (RCTs) assessing gastrointestinal bleeding of all NOAC regimens from inception to January 2018. The primary endpoint was major gastrointestinal (MGI) bleeding. The meta-regression was performed to access the association between the MGI bleeding events and mortality. The network meta-analysis was carried out with the Bayesian random-effect model. RESULTS: A total of 25 RCTs, including 139,392 patients, were identified. Meta-regression analysis showed that MGI bleeding was correlated with fatal bleeding events (odds ratios [OR], 1.76; 95% confidence interval [CI], 1.13-2.77], P=0.015). The network meta-analysis results showed that compared to the conventional regimens, rivaroxaban was associated with increased risk of MGI bleeding (OR, 1.37; 95% credible interval [CrI], 1.00-1.85), but not the apixaban (OR, 0.77; 95% CrI, 0.53-1.07]), edoxaban (OR, 0.86; 95%CrI, 0.52-1.18) and dabigatran etexilate (OR, 1.22; 95% CrI, 0.82-1.69). Compared to rivaroxaban, apixaban (OR, 0.56; 95% CrI, 0.35-0.88) and edoxaban (OR, 0.62; 95% CrI, 0.35-0.96) showed a significantly lower risk of MGI bleeding. Apixaban had the highest probability of being the safest option with regard to the risk of MGI bleeding (89.1%), followed by edoxaban (77.4%), conventional therapy (51.4%), dabigatran etexilate (23.8%) and rivaroxaban (8.3%). CONCLUSION: The risk of GI bleeding significantly varies among different NOAC regimens, and evidence shows that apixaban and edoxaban had the most favorable MGI bleeding safety profile, while rivaroxaban and dabigatran etexilate were the least safe.

Challenges in the use of zinc and its alloys as biodegradable metals: Perspective from biomechanical compatibility.

To date, more than fifty articles have been published on the feasibility studies of zinc and its alloys as biodegradable metals. These preliminary in vitro and in vivo studies showed acceptable biodegradability and reasonable biocompatibility in bone and blood microenvironments for the experimental Zn-based biodegradable metals and, for some alloy systems, superior mechanical performance over Mg-based biodegradable metals. For instance, the Zn-Li alloys exhibited higher UTS (UTS), and the Zn-Mn alloys exhibited higher elongation (more than 100%). On the one hand, similar to Mg-based biodegradable metals, insufficient strength and ductility, as well as relatively low fatigue strength, may lead to premature failure of medical devices. On the other hand, owing to the low melting point of the element Zn, several new uncertainties with regard to the mechanical properties of biomedical zinc alloys, including low creep resistance, high susceptibility to natural aging, and static recrystallization (SRX), may lead to device failure during storage at room temperature and usage at body temperature. This paper comprehensively reviews studies on these mechanical aspects of industrial Zn and Zn alloys in the last century and biomedical Zn and Zn alloys in this century. The challenges for the future design of biomedical zinc alloys as biodegradable metals to guarantee 100% mechanical compatibility are pointed out, and this will guide the mechanical property design of Zn-based biodegradable metals. STATEMENT OF SIGNIFICANCE: Previous studies on mechanical properties of industrial Zn and Zn alloys in the last century and biomedical Zn and Zn alloys in this century are comprehensively reviewed herein. The challenges for the future design of zinc-based biodegradable materials considering mechanical compatibility are pointed out. Common considerations such as strength, ductility, and fatigue behaviors are covered together with special attention on several new uncertainties including low creep resistance, high susceptibility to natural aging, and static recrystallization (SRX). These new uncertainties, which are not significantly observed in Mg-based and Fe-based materials, are largely due to the low melting point of the element Zn and may lead to device failure during storage at room temperature and clinical usage at body temperature. Future studies are urgently needed on these topics.

KRT1 gene silencing ameliorates myocardial ischemia-reperfusion injury via the activation of the Notch signaling pathway in mouse models.

Myocardial ischemia and reperfusion injury (MIRI) includes major drawbacks, such as excessive formation of free radicals and also overload of calcium, which lead to cell death, tissue scarring, and remodeling. The current study aims to explore whether KRT1 silencing may ameliorate MIRI via the Notch signaling pathway in mouse models. Myocardial tissues were used for the determination of the positive rate of KRT1 protein expression, apoptosis of myocardial cells, creatine kinase (CK) and lactate dehydrogenase (LDH) expression, expression of related biomarkers as well as myocardial infarction area. The transfected myocardial cells were treated with KRT1-siRNA, Jagged1, and DAPT (inhibitor of Notch-1 signaling pathway). The expression of KRT1, NICD, Hes1, Bcl-2, and Bax protein was detected. The MTT assay was applied for cell proliferation and flow cytometry was used for cell apoptosis. Mice with MIRI had a higher positive rate of KRT1 protein expression, apoptosis of myocardial cells, CK and LDH expression, myocardial infarction area, increased expression of MDA, NO, SDH, IL-1, IL-6, TNF-α, CRP, KRT1, Bax protein, CK, and LDH, and decreased expression of SOD, NICD, Hes1, and Bcl-2. The downregulation of KRT1 led to decreased expression of KRT1 and Bax protein, increased expression of NICD, Hes1, and Bcl-2, decreased cell apoptosis, and improved cell proliferation. The inhibition of the Notch signaling pathway leads to reduced expression of Bax, increased expression of NICD, Hes1, and Bcl 2, and also decreased cell apoptosis and increased cell proliferation. Our data conclude that KRT1 silencing is able to make MIRI better by activating the Notch signaling pathway in mice.

Exogenous hydrogen sulfide ameliorates high glucose-induced myocardial injury & inflammation via the CIRP-MAPK signaling pathway in H9c2 cardiac cells.

AIMS: Hydrogen sulfide (H2S) is a novel signaling molecule with potent cytoprotective actions. In this study, we hypothesize that exogenous H2S may protect cardiac cells against high glucose (HG)-induced myocardial injury and inflammation with the involvement of the CIRP-MAPK signaling pathway. MAIN METHODS: H9c2 cardiac cells cultured under HG conditions were transfected with siRNA and different inhibitor for detecting the effects of sodium hydrogen sulfide (NaHS) (a H2S donor) on cell biological processes. The cardiac cell viability and LDH activity were determined by CCK-8 and LDH kit. ELISA was employed to measure the levels of inflammatory factors, while 2',7'-dichlorofluorescein diacetate (DCFH-DA) to evaluate reactive oxygen species (ROS). Mitochondrial membrane potential (MMP) was identified by rhodamine 123 staining. TUNEL staining and Hoechst 33258 staining were employed to observe cardiac cell apoptosis. Besides, we determined the expression of CIRP-MAPK signaling pathway- and apoptosis-related factors by protein immunoblot analysis. KEY FINDINGS: HG culturing induced toxicity, LDH, higher level of inflammatory factors, ROS, MMP, and apoptosis in cardiac cells, attenuated the viability of cardiac cells, and activated the CIRP-MAPK signaling pathway. Notably, CIRP silencing aggravated the above condition. H2S or blockade of the MAPK signaling pathway reversed the above conditions induced by HG. SIGNIFICANCE: The present study provides evidence for the protective effect of exogenous H2S on HG-induced myocardial injury and inflammation in H9c2 cardiac cells and suggests that the activation of CIRP-MAPK signaling pathway might be one of the mechanisms underlying the protective effect of H2S.

The role of microRNA-1 targeting of MAPK3 in myocardial ischemia-reperfusion injury in rats undergoing sevoflurane preconditioning via the PI3K/Akt pathway.

Recent studies have uncovered the vital roles played by microRNAs in regulating cardiac injury. Among them, the cardiac enriched microRNA-1 (miR-1) has been extensively studied and proven to be detrimental to cardiac myocytes. Hence, the current study aimed to explore whether miR-1 affects myocardial ischemia-reperfusion injury (MIRI) in rats undergoing sevoflurane preconditioning and the underlying mechanism. After successful model establishment, rats with MIRI were transfected with mimics or inhibitors of miR-1, or siRNA against MAPK3, and then were injected with sevoflurane. A luciferase reporter gene assay was conducted to evaluate the targeting relationship between miR-1 and MAPK3. Reverse transcription quantitative polymerase chain reaction and Western blot analysis were employed to evaluate the expressions of miR-1, MAPK3, phosphatidylinositol 3-kinase (PI3K), and Akt. Additionally, the concentration of lactate dehydrogenase (LDH) was determined. Cell apoptosis and viability were assessed using TUNEL and cell counting kit-8 assays, and the ischemic area at risk and infarct size were detected using Evans blue and triphenyltetrazolium chloride staining. MAPK3 was found to be the target gene of miR-1. miR-1 expressed at a high level whereas MAPK3 expressed at a low level in MIRI rats. Overexpressing miR-1 or silencing MAPK3 blocked the PI3K/Akt pathway to increase cell apoptosis, ischemic area at risk, and infarct area but decreased cell viability and increased LDH concentration. In contrast, miR-1 downregulation abrogated the effects induced by miR-1 mimics or siRNA against MAPK3. These findings indicate that inhibition of miR-1 promotes MAPK3 to protect against MIRI in rats undergoing sevoflurane preconditioning through activation of the PI3K/Akt pathway.