Associations of Postoperative Hemoglobin Level and Clinical Outcomes in Patients Undergoing Mitral Valve Surgery.

BACKGROUND: Postoperative hemoglobin could indicate useful information for transfusion practices. The aim of this study was to investigate the association of optimal hemoglobin level and clinical outcomes after mitral valve surgery (MVS). METHODS: This investigation was a multicenter observational cohort study including 1,518 patients undergoing mitral valve surgery from 2016 through 2018. Patients were separated into six predefined groups based on initial postoperative hemoglobin (< 7.5 g/dL, 7.5 - 8.4 g/dL, 8.5 - 9.4 g/dL, 9.5 - 10.4 g/dL, 10.5 - 11.4 g/dL, ≥ 11.5 g/dL). Multivariable regression analysis was used to adjust laboratory results and surgical features of patients to evaluate the relationships between initial hemoglobin after MVS and clinical outcomes. RESULTS: Patients with initial postoperative hemoglobin below 7.5 g/dL had longer length of stays [mean (95% confidence interval [CI]), 1.9 (1.093 - 1.367)] in comparison with the reference group of 9.5 - 10.4 g/dL. Similarly, for those with hemoglobin below 7.5 g/dL, the odds (95% CI) for secondary outcomes included myocardial infraction 11.801 (1.353 - 22.966) and thrombosis 5.113 (1.340 - 9.508). However, for clinical outcomes, there was no significant difference between the five groups with hemoglobin greater than 7.5 g/dL. CONCLUSIONS: In patients after MVS, initial postoperative hemoglobin values below 7.5 g/dL was associated with worse outcomes compared to other values. Given similar outcomes between hemoglobin more than 7.5 g/dL groups, targeting treatment to an initial postoperative hemoglobin value at the lower value may be more desirable.

Biomimetic nanotherapy: core-shell structured nanocomplexes based on the neutrophil membrane for targeted therapy of lymphoma.

BACKGROUND: Non-Hodgkin's lymphoma (NHL) is a malignant disease of lymphoid tissue. At present, chemotherapy is still the main method for the treatment of NHL. R-CHOP can significantly improve the survival rate of patients. Unfortunately, DOX is the main cytotoxic drug in R-CHOP and it can lead to adverse reactions. Therefore, it is particularly important to uncover new treatment options for NHL. RESULTS: In this study, a novel anti-tumor nanoparticle complex Nm@MSNs-DOX/SM was designed and constructed in this study. Mesoporous silica nanoparticles (MSNs) loaded with Doxorubicin (DOX) and anti-inflammatory drugs Shanzhiside methylester (SM) were used as the core of nanoparticles. Neutrophil membrane (Nm) can be coated with multiple nanonuclei as a shell. DOX combined with SM can enhance the anti-tumor effect, and induce apoptosis of lymphoma cells and inhibit the expression of inflammatory factors related to tumorigenesis depending on the regulation of Bcl-2 family-mediated mitochondrial pathways, such as TNF-α and IL-1ß. Consequently, the tumor microenvironment (TME) was reshaped, and the anti-tumor effect of DOX was amplified. Besides, Nm has good biocompatibility and can enhance the EPR effect of Nm@MSNs-DOX/SM and increase the effect of active targeting tumors. CONCLUSIONS: This suggests that the Nm-modified drug delivery system Nm@MSNs-DOX/SM is a promising targeted chemotherapy and anti-inflammatory therapy nanocomplex, and may be employed as a specific and efficient anti-Lymphoma therapy.

The Effect of Autophagic Activity on the Function of Apheresis Platelets and on the Efficacy of Clinical Platelet Transfusion.

Platelet activation and survival jointly determine the efficacy of clinical platelet transfusion. This study aimed to discuss the effect of autophagic activity on activation and aggregation of apheresis platelets and on the efficacy of clinical platelet transfusion. In this study, we investigated the effects of autophagic activity of apheresis platelets for different blood types and after different storage durations on platelet activation and aggregation functions. By Western blot, immunofluorescence, and RT-qPCR detection, we found that with the prolongation of the storage duration, the expressions of both autophagy-related proteins and genes were upregulated in apheresis platelets and their expressions were insignificantly higher in the apheresis platelets of type A and O blood than in those of type B and type AB blood. After RAPA/IGF-1 pretreatment, there was a significant increase/reduction in autophagic activity. After RAPA and IGF-1 pretreatment, an opposite variation trend was observed with platelet activation and aggregation. Autophagic activity of platelets correlated negatively with the efficacy of clinical platelet transfusion. These research findings provide a theoretical basis for effective clinical platelet transfusion.

Neonatal alloimmune thrombocytopenia caused by anti-HPA antibodies in pregnant Chinese women: a study protocol for a multicentre, prospective cohort trial.

BACKGROUND: Neonatal alloimmune thrombocytopenia (NAIT), caused by maternal antibodies raised against alloantigens carried on foetal platelets, is a very common haematological abnormality in newborns worldwide. However, baseline data on NAIT in China are lacking. Therefore, this study seeks to explore the incidence of alloantibody against the human platelet antigen (HPA) in pregnant women and its associations with NAIT in China. METHODS: A multicentre, prospective cohort study design will be used, and 55,497 pregnant women will be recruited for the first screening of the anti-HPA antibody at 12 to 28 weeks of gestational age. Subjects who are positive in the first screening for the anti-HPA antibody will be included in the exposure group. Re-tests of the antibody titre, antigen-specificity and genotyping of HPA and HLA will be conducted during admission. A ratio of 1:1 paired individuals with the same ethnicity and parity but testing negative for the anti-HPA antibody will be randomly selected to be included in the non-exposure group. NAIT will be diagnosed in the newborns on day one of the birth. The HPA of the neonates in the exposure group will also be genotyped by sequencing. Associations of maternal HLA with the occurrence of the anti-HPA antibody and correlation of the severity of NAIT with the titre of the anti-HPA antibody will be further analysed. DISCUSSION: The study is expected to provide baseline data on NAIT in China. Besides, we hope to find out a population who expresses particular HLA molecules has significant higher risk of HPA alloimmunization in Chinese individuals. We also hope to find a Chinese-specific cut-off antibody titre for the prediction of the severity of NAIT and to provide a means to evaluate the necessity of antenatal treatment. TRIAL REGISTRATION: ClinicalTrials.gov: NCT02934906 (date registered: 13.10.2016).

AP-57/C10orf99 is a new type of multifunctional antimicrobial peptide.

Antimicrobial peptides (AMPs) are an evolutionarily conserved component of the innate immune response that provides host defence at skin and mucosal surfaces. Here, we report the identification and characterization of a new type human AMPs, termed AP-57 (Antimicrobial Peptide with 57 amino acid residues), which is also known as C10orf99 (chromosome 10 open reading frame 99). AP-57 is a short basic amphiphilic peptide with four cysteines and a net charge +14 (MW = 6.52, PI = 11.28). The highest expression of AP-57 were detected in the mucosa of stomach and colon through immunohistochemical assay. Epithelium of skin and esophagus show obvious positive staining and strong positive staining were also observed in some tumor and/or their adjacent tissues, such as esophagus cancer, hepatocellular carcinoma, squamous cell carcinoma and invasive ductal carcinoma. AP-57 exhibited broad-spectrum antimicrobial activities against Gram-positive Staphylococcus aureus, Actinomyce, and Fungi Aspergillus niger as well as mycoplasma and lentivirus. AP-57 also exhibited DNA binding capacity and specific cytotoxic effects against human B-cell lymphoma Raji. Compared with other human AMPs, AP-57 has its distinct characteristics, including longer sequence length, four cysteines, highly cationic character, cell-specific toxicity, DNA binding and tissue-specific expressing patterns. Together, AP-57 is a new type of multifunctional AMPs worthy further investigation.

Opposite effects of estradiol and progesterone on woman's disgust processing.

Background: Ovarian hormones play a critical role in emotion processing, which may be a major reason for the high rates of major depressive disorders in women. However, the exact roles of estradiol and progesterone in emotional processing remain unclear. To this end, we performed behavioral and rs-fMRI studies on the effects ovarian hormones on disgust emotion. Methods: In Experiment 1, 95 Chinese female undergraduates completed the single category implicit association test (SC-IAT) and explicit measures of disgust intensity task, 32 in the menstrual phase, 30 in the follicular phase, and 33 in the luteal phase. In Experiment 2, A total of 25 healthy female undergraduates completed three sessions of the rs-fMRI. The menstrual group was scanned during cycle days 2-5, the follicular group during cycle during days 10-13, and the luteal group was scanned 3-7 days before the next menstruation. Results: The behavioral results showed that women during the luteal phase had higher D scores and shorter response times (RTs) to disgust stimuli compared to the menses and follicular phases. In contrast, women during the follicular phase had fewer feelings of disgust and longer RTs to pathogen stimuli compared with that during the menses and luteal phases, but this effect was moderated by the intensity of the stimuli. rs-fMRI studies showed that women during the luteal phase have higher functional connectivity in the salience network than those in the follicular phase. Compared with the menstrual phase, women have lower functional connectivity in the amygdala during the follicular phase. Conclusion: In summary, a more negative attitude to disgust stimuli and the enhanced functional connectivity of the salience network during the luteal phase may be associated with high progesterone levels, whereas lower disgust feelings and reduced functional connectivity of the amygdala during the follicular phase may be associated with high estradiol levels.

Identification of risk factors for infection after mitral valve surgery through machine learning approaches.

Background: Selecting features related to postoperative infection following cardiac surgery was highly valuable for effective intervention. We used machine learning methods to identify critical perioperative infection-related variables after mitral valve surgery and construct a prediction model. Methods: Participants comprised 1223 patients who underwent cardiac valvular surgery at eight large centers in China. The ninety-one demographic and perioperative parameters were collected. Random forest (RF) and least absolute shrinkage and selection operator (LASSO) techniques were used to identify postoperative infection-related variables; the Venn diagram determined overlapping variables. The following ML methods: random forest (RF), extreme gradient boosting (XGBoost), Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), AdaBoost, Naive Bayesian (NB), Logistic Regression (LogicR), Neural Networks (nnet) and artificial neural network (ANN) were developed to construct the models. We constructed receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC) was calculated to evaluate model performance. Results: We identified 47 and 35 variables with RF and LASSO, respectively. Twenty-one overlapping variables were finally selected for model construction: age, weight, hospital stay, total red blood cell (RBC) and total fresh frozen plasma (FFP) transfusions, New York Heart Association (NYHA) class, preoperative creatinine, left ventricular ejection fraction (LVEF), RBC count, platelet (PLT) count, prothrombin time, intraoperative autologous blood, total output, total input, aortic cross-clamp (ACC) time, postoperative white blood cell (WBC) count, aspartate aminotransferase (AST), alanine aminotransferase (ALT), PLT count, hemoglobin (Hb), and LVEF. The prediction models for infection after mitral valve surgery were established based on these variables, and they all showed excellent discrimination performance in the test set (AUC > 0.79). Conclusions: Key features selected by machine learning methods can accurately predict infection after mitral valve surgery, guiding physicians in taking appropriate preventive measures and diminishing the infection risk.

Altered Functional Connectivity of Ventral Striatum Subregions in De-novo Parkinson's Disease with Depression.

Although various studies have reported a high prevalence of depression among Parkinson's disease (PD) patients, the pathophysiological mechanism of depression in PD (DPD) is still unclear. The core region of the reward network, the ventral striatum (VS), is critical in the occurrence and development of DPD. This study aimed to explore the altered functional connectivity (FC) of VS subregions in DPD. We recruited 20 DPD patients, 37 non-depressed PD (NDPD) patients, and 41 healthy controls (HC) matched in age, gender, and years of education. The patients' diagnosis with PD was de-novo. We then used resting-state functional magnetic resonance imaging to detect the FC differences of VS subregions among these groups. The FC between the left ventral caudate (vCa_L) and the left middle occipital gyrus (MOG.L) was significantly increased in DPD than in NDPD patients or HC. Compared with HC, NDPD patients exhibited significantly increased FCs between bilateral ventromedial putamen and the left paracentral lobule, the right ventromedial putamen (vmPu_R), and the right precentral gyrus, the vmPu_R, and the left precuneus. Besides, a significant negative correlation was found between the FC values of the vCa_L with the MOG.L and the HAMD-17 scores in the DPD group. The hyperconnectivity between vCa_L and the MOG.L might be viewed as a compensatory mechanism for depression in the early stage of PD. This study provides new insight into the neural mechanism of depression in the early stage of PD and contributes to explore the potential neuroimaging markers for DPD.

The Relationship Between 5-Hydroxytryptamine and Its Metabolite Changes With Post-stroke Depression.

Post-stroke depression (PSD) is the most common and serious sequelae of stroke. Approximately 33% of stroke survivors were affected by PSD. However, many issues (e.g., incidence, diagnostic marker, and risk factor) related to PSD remained unclear. The "monoamine hypothesis" is a significant hypothesis for depression, which suggests that three monoamines play a key role in depression. Therefore, most current antidepressants are developed to modulate the monoamines on PSD treatment, and these antidepressants have good effects on patients with PSD. However, the potential mechanisms of three monoamines in PSD are still unclear. Previously, we proposed "three primary emotions," which suggested a new model of basic emotions based on the three monoamines. It may provide a new way for PSD treatment. In addition, recent studies have found that monoamine-related emotional intervention also showed potential effects in the treatment and prevention of PSD. This study discusses these issues and attempts to provide a prospect for future research on PSD.

Corrigendum: Machine Learning for the Prediction of Complications in Patients After Mitral Valve Surgery.

[This corrects the article DOI: 10.3389/fcvm.2021.771246.].

Metabolomics changes in brain-gut axis after unpredictable chronic mild stress.

BACKGROUND: Major depressive disorder is a leading cause of disability worldwide, affecting up to 17 % of the general population. The neural mechanisms of depression, however, are yet to be uncovered. Recently, attention has been drawn to the effects of dysfunctional brain-gut axis on depression, and many substances have been suggested to be involved in the communication between the gut and brain, such as ghrelin. METHODS: We herein systematically examined the changes of metabolomics after unpredictable chronic mild stress (UCMS)-induced depression-like behaviors in rats and compared the altered metabolites in the hippocampus and jejunum samples. RESULTS: Our results show that many metabolites significantly changed with UCMS both in the hippocampus and jejunum, such as L-glutamine, L-tyrosine, hydroxylamine, and 3-phosphoglyceric acid. Further studies suggested that these changes are the reasons for anxiety-like behaviors and depression-like behaviors in UCMS rats and also are the reasons for hippocampal neural plasticity. CONCLUSIONS: Coexistence of brain and gut metabolic changes in UCMS-induced depressive behavior in rats suggests a possible role of brain-gut axis in depression. This study provides insights into the neurobiology of depression.

Corrigendum: Machine Learning for the Prediction of Complications in Patients After Mitral Valve Surgery.

[This corrects the article DOI: 10.3389/fcvm.2021.771246.].

Dissemination of bla NDM-5 and mcr-8.1 in carbapenem-resistant Klebsiella pneumoniae and Klebsiella quasipneumoniae in an animal breeding area in Eastern China.

Animal farms have become one of the most important reservoirs of carbapenem-resistant Klebsiella spp. (CRK) owing to the wide usage of veterinary antibiotics. "One Health"-studies observing animals, the environment, and humans are necessary to understand the dissemination of CRK in animal breeding areas. Based on the concept of "One-Health," 263 samples of animal feces, wastewater, well water, and human feces from 60 livestock and poultry farms in Shandong province, China were screened for CRK. Five carbapenem-resistant Klebsiella pneumoniae (CRKP) and three carbapenem-resistant Klebsiella quasipneumoniae (CRKQ) strains were isolated from animal feces, human feces, and well water. The eight strains were characterized by antimicrobial susceptibility testing, plasmid conjugation assays, whole-genome sequencing, and bioinformatics analysis. All strains carried the carbapenemase-encoding gene bla NDM-5, which was flanked by the same core genetic structure (IS5-bla NDM-5-ble MBL-trpF-dsbD-IS26-ISKox3) and was located on highly related conjugative IncX3 plasmids. The colistin resistance gene mcr-8.1 was carried by three CRKP and located on self-transmissible IncFII(K)/IncFIA(HI1) and IncFII(pKP91)/IncFIA(HI1) plasmids. The genetic context of mcr-8.1 consisted of IS903-orf-mcr-8.1-copR-baeS-dgkA-orf-IS903 in three strains. Single nucleotide polymorphism (SNP) analysis confirmed the clonal spread of CRKP carrying-bla NDM-5 and mcr-8.1 between two human workers in the same chicken farm. Additionally, the SNP analysis showed clonal expansion of CRKP and CRKQ strains from well water in different farms, and the clonal CRKP was clonally related to isolates from animal farms and a wastewater treatment plant collected in other studies in the same province. These findings suggest that CRKP and CRKQ are capable of disseminating via horizontal gene transfer and clonal expansion and may pose a significant threat to public health unless preventative measures are taken.

Monoamine Neurotransmitters Control Basic Emotions and Affect Major Depressive Disorders.

Major depressive disorder (MDD) is a common and complex mental disorder, that adversely impacts an individual's quality of life, but its diagnosis and treatment are not accurately executed and a symptom-based approach is utilized in most cases, due to the lack of precise knowledge regarding the pathophysiology. So far, the first-line treatments are still based on monoamine neurotransmitters. Even though there is a lot of progress in this field, the mechanisms seem to get more and more confusing, and the treatment is also getting more and more controversial. In this study, we try to review the broad advances of monoamine neurotransmitters in the field of MDD, and update its effects in many advanced neuroscience studies. We still propose the monoamine hypothesis but paid special attention to their effects on the new pathways for MDD, such as inflammation, oxidative stress, neurotrophins, and neurogenesis, especially in the glial cells, which have recently been found to play an important role in many neurodegenerative disorders, including MDD. In addition, we will extend the monoamine hypothesis to basic emotions; as suggested in our previous reports, the three monoamine neurotransmitters play different roles in emotions: dopamine-joy, norepinephrine-fear (anger), serotonins-disgust (sadness). Above all, this paper tries to give a full picture of the relationship between the MDD and the monoamine neurotransmitters such as DA, NE, and 5-HT, as well as their contributions to the Three Primary Color Model of Basic Emotions (joy, fear, and disgust). This is done by explaining the contribution of the monoamine from many sides for MDD, such the digestive tract, astrocytes, microglial, and others, and very briefly addressing the potential of monoamine neurotransmitters as a therapeutic approach for MDD patients and also the reasons for its limited clinical efficacy, side effects, and delayed onset of action. We hope this review might offer new pharmacological management of MDD.

High-dose methotrexate combined with rituximab improves the survival rate of patients with primary central nervous system lymphoma.

PURPOSE: To explore the efficacy of high-dose methotrexate (HD-MTX) combined with rituximab (R) in the treatment of primary central nervous system lymphoma (PCNSL). METHODS: 108 PCNSL patients were randomly divided into Rituximab group (n=54) or control group (n=54). The patients in Rituximab group were treated with HD-MTX + R chemotherapy, while those in control group were given HD-MTX combined with whole brain radiotherapy (WBRT). The therapeutic effect, incidence rate of adverse reactions and the SF-36 score were compared between the two groups. RESULTS: The overall response rate was overtly higher in Rituximab group than that in control group (81.5% vs. 57.4%). After treatment, the scores of physical function, physical competence, health condition, social function and emotional function in the SF-36 scale were notably higher in Rituximab group than those in control group. The 1-year overall survival (OS) rate was 83.3% (45/54) and 63.0% (34/54), 1-year progression-free survival (PFS) rate was 70.4% (38/54) and 46.3% (25/68), 3-year OS rate was 57.4% (31/54) and 31.5% (17/54), and 3-year PFS rate was 27.8% (15/54) and 14.8% (8/54) in Rituximab group and control group, respectively. The results of log-rank test showed that the OS and PFS rates in Rituximab group were obviously better than those in control group. CONCLUSION: Compared with HD-MTX combined with WBRT, HD-MTX combined with R can remarkably improved the quality of life and survival rate of patients with PCNSL, with tolerable adverse reactions and is worthy of popularization and application in clinical practice.

Effects of ginkgo diterpene lactone on brain inflammation and oxidative stress in rats with cognitive impairment of cerebral small vessel disease.

PURPOSE: This study aims to investigate the effects of ginkgo diterpene lactone on brain inflammation and oxidative stress in rats with cognitive impairment of cerebral small vessel disease (CSVD). METHODS: We equally assigned 40 Sprague Dawley (SD) rats to the observation group (OG) and the control group (CG) and modeled them for cognitive impairment of CSVD. Rats in OG were given ginkgo diterpene lactone for treatment, while those in CG were injected with the same amount of normal saline. The learning and memorizing ability of rats was tested by the water maze. The oxidative stress and inflammatory response in rats were evaluated. The levels of vascular endothelial growth factor (VEGF) and endostatin (ES) mRNA in the hippocampus of rats were measured. Vascular smooth muscle cells of rats were sampled for cell viability and apoptosis assays. RESULTS: Rats from OG were superior to those from CG in the learning and memorizing ability. After treatment, rats from OG had markedly lower malondialdehyde (MDA) levels and higher levels of superoxide dismutase (SOD), reduced glutathione (GSH), and glutathione peroxidase (GSH-Px) than those from CG (all P < 0.05). Concentrations of interleukin-18 (IL-18), transforming growth factor-ß1 (TGF-ß1), tumor necrosis factor-α (TNF-α), and amyloid ß-protein 1-40 (Aß1-40) were markedly lower in OG than in CG (all P < 0.05). Rats from OG had markedly higher VEGF levels and lower ES mRNA levels than those from CG (all P < 0.05). Cell viability gradually decreased in rats from OG after treatment and was markedly lower than that in rats from CG (P < 0.05). Cell apoptosis rate was markedly higher in OG than in CG (P < 0.05). CONCLUSION: Ginkgo diterpene lactone can inhibit oxidative stress and inflammatory response in rats with cognitive impairment of CSVD to a certain degree.

The Aversion Function of the Limbic Dopaminergic Neurons and Their Roles in Functional Neurological Disorders.

The Freudian theory of conversion suggested that the major symptoms of functional neurological disorders (FNDs) are due to internal conflicts at motivation, especially at the sex drive or libido. FND patients might behave properly at rewarding situations, but they do not know how to behave at aversive situations. Sex drive is the major source of dopamine (DA) release in the limbic area; however, the neural mechanism involved in FND is not clear. Dopaminergic (DAergic) neurons have been shown to play a key role in processing motivation-related information. Recently, DAergic neurons are found to be involved in reward-related prediction error, as well as the prediction of aversive information. Therefore, it is suggested that DA might change the rewarding reactions to aversive reactions at internal conflicts of FND. So DAergic neurons in the limbic areas might induce two major motivational functions: reward and aversion at internal conflicts. This article reviewed the recent advances on studies about DAergic neurons involved in aversive stimulus processing at internal conflicts and summarizes several neural pathways, including four limbic system brain regions, which are involved in the processing of aversion. Then the article discussed the vital function of these neural circuits in addictive behavior, depression treatment, and FNDs. In all, this review provided a prospect for future research on the aversion function of limbic system DA neurons and the therapy of FNDs.

Effects of Coronavirus-19 Induced Loneliness on Mental Health: Sleep Quality and Intolerance for Uncertainty as Mediators.

Objective: The aim of the study is to investigate effects of loneliness on individual's mental health and the mediating effects of intolerance of uncertainty and sleep quality in the post Coronavirus-19 period, especially for the young people. Methods: The questionnaires used in this study include UCLA loneliness scale (UCLA-3), the Pittsburgh Sleep Quality Index (PSQI), intolerance for uncertainty (IU) and the Chinese version of DASS-21. A total number of 289 subjects were recruited in the study, which includes 209 females (72.3%), 80 males (27.7%); and 212 students (73.4%), 77 working staffs (26.6%). Results: The results showed that: (1) people have high levels of loneliness, anxiety, depression and stress, and poor sleep quality; (2) the mediating effect of intolerance for uncertainty in the relationship of loneliness and mental health is significant (effect size = 0.178, 95% CI confidence interval: [0.115, 0.241]), and the mediating effects of sleep quality in the relationship between loneliness and mental health is significant (effect size = 0.127, 95% CI confidence interval: [0.017, 0.239]). Conclusion: Loneliness invokes a stronger self-concerned inadaptability to threat response and may lead to more mental diseases through more serious intolerance for uncertainty and insomnia.

Ability of a Machine Learning Algorithm to Predict the Need for Perioperative Red Blood Cells Transfusion in Pelvic Fracture Patients: A Multicenter Cohort Study in China.

Background: Predicting the perioperative requirement for red blood cells (RBCs) transfusion in patients with the pelvic fracture may be challenging. In this study, we constructed a perioperative RBCs transfusion predictive model (ternary classifications) based on a machine learning algorithm. Materials and Methods: This study included perioperative adult patients with pelvic trauma hospitalized across six Chinese centers between September 2012 and June 2019. An extreme gradient boosting (XGBoost) algorithm was used to predict the need for perioperative RBCs transfusion, with data being split into training test (80%), which was subjected to 5-fold cross-validation, and test set (20%). The ability of the predictive transfusion model was compared with blood preparation based on surgeons' experience and other predictive models, including random forest, gradient boosting decision tree, K-nearest neighbor, logistic regression, and Gaussian naïve Bayes classifier models. Data of 33 patients from one of the hospitals were prospectively collected for model validation. Results: Among 510 patients, 192 (37.65%) have not received any perioperative RBCs transfusion, 127 (24.90%) received less-transfusion (RBCs < 4U), and 191 (37.45%) received more-transfusion (RBCs ≥ 4U). Machine learning-based transfusion predictive model produced the best performance with the accuracy of 83.34%, and Kappa coefficient of 0.7967 compared with other methods (blood preparation based on surgeons' experience with the accuracy of 65.94%, and Kappa coefficient of 0.5704; the random forest method with an accuracy of 82.35%, and Kappa coefficient of 0.7858; the gradient boosting decision tree with an accuracy of 79.41%, and Kappa coefficient of 0.7742; the K-nearest neighbor with an accuracy of 53.92%, and Kappa coefficient of 0.3341). In the prospective dataset, it also had a food performance with accuracy 81.82%. Conclusion: This multicenter retrospective cohort study described the construction of an accurate model that could predict perioperative RBCs transfusion in patients with pelvic fractures.

Machine Learning for the Prediction of Complications in Patients After Mitral Valve Surgery.

Background: This study intended to use a machine learning model to identify critical preoperative and intraoperative variables and predict the risk of several severe complications (myocardial infarction, stroke, renal failure, and hospital mortality) after cardiac valvular surgery. Study Design and Methods: A total of 1,488 patients undergoing cardiac valvular surgery in eight large tertiary hospitals in China were examined. Fifty-four perioperative variables, such as essential demographic characteristics, concomitant disease, preoperative laboratory indicators, operation type, and intraoperative information, were collected. Machine learning models were developed and validated by 10-fold cross-validation. In each fold, Recursive Feature Elimination was used to select key variables. Ten machine learning models and logistic regression were developed. The area under the receiver operating characteristic (AUROC), accuracy (ACC), Youden index, sensitivity, specificity, F1-score, positive predictive value (PPV), and negative predictive value (NPV) were used to compare the prediction performance of different models. The SHapley Additive ex Planations package was applied to interpret the best machine learning model. Finally, a model was trained on the whole dataset with the merged key variables, and a web tool was created for clinicians to use. Results: In this study, 14 vital variables, namely, intraoperative total input, intraoperative blood loss, intraoperative colloid bolus, Classification of New York Heart Association (NYHA) heart function, preoperative hemoglobin (Hb), preoperative platelet (PLT), age, preoperative fibrinogen (FIB), intraoperative minimum red blood cell volume (Hct), body mass index (BMI), creatinine, preoperative Hct, intraoperative minimum Hb, and intraoperative autologous blood, were finally selected. The eXtreme Gradient Boosting algorithms (XGBOOST) algorithm model presented a significantly better predictive performance (AUROC: 0.90) than the other models (ACC: 81%, Youden index: 70%, sensitivity: 89%, specificity: 81%, F1-score:0.26, PPV: 15%, and NPV: 99%). Conclusion: A model for predicting several severe complications after cardiac valvular surgery was successfully developed using a machine learning algorithm based on 14 perioperative variables, which could guide clinical physicians to take appropriate preventive measures and diminish the complications for patients at high risk.