The study presented here aimed to establish a predictive model for heart failure (HF) and all-cause mortality in peritoneal dialysis (PD) patients with machine learning (ML) algorithm. We retrospectively included 1006 patients who initiated PD from 2010 to 2016. XGBoost, random forest (RF), and AdaBoost were used to train models for assessing risk for 1-year and 5-year HF hospitalization and mortality. The performance was validated using fivefold cross-validation. The optimal ML algorithm was used to construct the models to predictive the risk of the HF and all-cause mortality. The prediction performance of ML methods and Cox regression was compared. Over a median follow-up of 49 months. Two hundred and ninety-eight patients developed HF required hospitalization; 199 patients died during the follow-up. The RF model (AUC = 0.853) was the best performing model for predicting HF, and the XGBoost model (AUC = 0.871) was the best model for predicting mortality. Baseline moderate or severe renal disease, systolic blood pressure (SBP), body mass index (BMI), age, Charlson Comorbidity Index (CCI) score were strongly associated with HF hospitalization, whereas age, CCI score, creatinine, age, high-density lipoprotein cholesterol (HDL-C), total cholesterol, baseline estimated glomerular filtration rate (eGFR) were the most significant predictors of mortality. For all the above endpoints, the ML models demonstrated better discrimination than Cox regression. We developed and validated a novel method to predict the risk factors of HF and all-cause mortality that integrates readily available clinical, laboratory, and electrocardiographic variables to predict the risk of HF among PD patients.
1 November 2014 in Peritoneal Dialysis International: Journal of the International Society for Peritoneal Dialysis
Introduction: The morbidity and mortality of acute kidney injury (AKI) are increasing. Epigenetic regulation and immune cell infiltration are thought to be involved in AKI. However, the relationship between epigenetic regulation and immune cell infiltration in AKI has not been elucidated. This study was conducted to identify the differentially expressed genes (DEGs), differentially expressed RNA methylation genes (DEMGs), and infiltrated immune cells in the kidneys of ischemia-reperfusion induced-acute kidney injury (IRI-AKI) models and further explore their relationships in IRI-AKI. Methods: This is a bioinformatic analysis using R programming language in 3 selected IRI-AKI datasets from the Gene Expression Omnibus (GEO) database, including 16 IRI-AKI kidney tissues and 10 normal kidney tissues. The DEGs were screened, and enrichment pathways were analyzed using gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The DEMGs and core DEMGs were identified using the R package. The ROC curve was plotted to predict disease occurrence of 7 core DEMGs. The correlation of 7 core DEMGs and other genes was analyzed using Pearson’s correlation test. The gene set enrichment analysis (GSEA) of each DEMG was conducted using the R package. The upstream miRNAs and transcript factors of 7 core DEMGs were predicted based on the RegNetwork database and Cytoscape software. The STITCH database was used to predict the possible binding compounds of the 7 core DEMGs. Immune cell infiltration in kidney tissues between the IRI-AKI group and control group was evaluated using the R package. Results: A total of 2,367 DEGs were obtained, including 1,180 upregulated and 1,187 downregulated genes in IRI-AKI kidney associated with the cell structure, proliferation, molecule binding/interaction, and signaling pathways such as the leukocyte migration and chemokine signaling pathways. Ten DEMGs were identified, with Ythdf1, Rbm15, Trmt6, Hnrnpc, and Dnmt1 being significantly upregulated, while Lrpprc, Cyfip2, Mettl3, Ncbp2, and Nudt7 were significantly downregulated in IRI-AKI tissues. The molecules interacting with 7 core DEMGs were identified. Significant changes in the infiltration of 8 types of immune cells were observed in IRI-AKI kidneys compared to normal controls. The significant correlation between 6 core DEMGs and the infiltration of immune cells was observed. Conclusion: IRI may induce AKI through RNA methylation to regulate the expression of genes involved in immune cell infiltration.
The study aimed to predict the risks of Major adverse cardiac events (MACE) in patients undergoing peritoneal dialysis (PD) with machine learning (ML) algorithm. In addition, we added the time factor and predicted the risk factors for MACE during the 1-year and 5-year follow-up. This retrospective study included 1006 PD patients from January 2010 to December 2016. XGBoost, Random Forest (RF) and Adaboost were used to train models for assessing risk of 1-year and 5-year MACE. The optimal ML algorithm was used to construct the models to predict the risk of the MACE end point. 409 patients developed MACE during the follow-up. The RF model (AUC = 0.80) was optimal for overall MACE prediction. The three most influential variables, ranked in descending order of importance were Parathyroid hormone, Congestive heart failure and Age.114 patients developed MACE during the first-year follow-up. The XGBoost model (AUC = 0.86) performed best for 1-year MACE. The three most influential variables, ranked in descending order of importance were High-Density Lipoprotein Cholesterol (HDL-C), Age and Calcium. 331 patients developed MACE during the 5-year follow-up. The RF model (AUC = 0.75) was the best predicting model for 5-year MACE. The three most influential variables, ranked in descending order of importance were Age, Creatinine and estimated Glomerular Filtration Rate. We developed and validated a novel algorithm to predict the risk factors of MACE in PD patients.