Acute kidney injury (AKI) represents a life-threatening condition among hospitalized patients, where early prediction enables prevention. Despite advances in existing models, clinical implementation remains hindered by excessive false positive rates (70%-94%) and lack of actionable clinical insights. We conduct a multi-center retrospective cohort study and develop a two-model large language model framework: AKI-PM (Prediction Model) for predicting AKI occurrence within 24 hours and AKI-RAM (Risk Attribution Model) for providing explainable risk attribution. Using a cohort of 140,637 hospital admissions across four geographically diverse Chinese hospitals, we demonstrate that AKI-PM achieves high predictive performance in internal validation (area under curve 0.95, positive predictive value 0.68) and maintains robust generalizability across external sites after few-shot (area under curve 0.92-0.96, positive predictive value 0.69-0.74). Crucially, AKI-RAM provides structured, clinically actionable risk explanations by distinguishing modifiable from non-modifiable factors and offering tailored recommendations. In a clinical evaluation of 200 cases from four independent hospitals by six nephrologists, AKI-RAM receives high scores across eight dimensions (Likert scale: 4.18-4.88) with moderate to good inter-rater reliability (intraclass correlation coefficients: 0.680-0.803). This integrated framework addresses critical limitations in AI-driven clinical prediction by combining accuracy with interpretability, offering a scalable solution for early AKI prevention in diverse healthcare settings.

Background:Monogenic causes are increasingly recognized in end-stage kidney disease (ESKD), but the real-world diagnostic efficacy of exome sequencing in unselected dialysis cohorts is still being defined. Methods:We conducted a prospective study enrolling 317 adult ESKD patients from a single center in Taiyuan, China, regardless of presumed etiology. Whole-exome sequencing (WES) was performed on peripheral blood DNA. Variants were curated and classified per the American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) 2015 and Association for Clinical Genomic Science (ACGS) guidelines, with only "pathogenic" or "likely pathogenic" findings considered diagnostic. Results:The cohort was 59% male, mean ESKD onset 53.2 ± 14.3 years. A definitive monogenic diagnosis emerged in 7.3% (23/317) of patients, in line with multicenter and international studies. Genes most frequently implicated were PKD1 (3.5% of cohort; 47.8% of genetically diagnosed) and COL4A3/4/5 (1.9%; 26.1% of diagnosed), reflecting global trends of autosomal dominant polycystic kidney disease and Alport syndrome as major genetic contributors in adult ESKD. Notably, mutations in ACTN4, PAX2, COQ8B or INF2, causing hereditary steroid-resistant nephrotic syndrome, led to significantly earlier ESKD onset (mean 31.3 years) compared with PKD1 or COL4-related cases. Inconclusive genetic findings were present in 7.9% (25/317). Most patients reported no family history of kidney disease, indicating the limitations of clinical suspicion alone. Conclusions:In a real-world Chinese dialysis cohort, WES provided a molecular diagnosis in 7.3% of cases, demonstrating clinical utility for risk stratification, family counseling, donor selection and actionable therapy. These findings underscore the need for routine integration of genetic testing in ESKD care irrespective of family history, especially to clarify ambiguous cases and optimize management.

Natural medicines, particularly Traditional Chinese Medicine (TCM), are gaining global recognition for their therapeutic potential in addressing human symptoms and diseases. TCM, with its systematic theories and extensive practical experience, provides abundant resources for healthcare. However, the effective application of TCM requires precise syndrome diagnosis, determination of treatment principles, and prescription formulation, which demand decades of clinical expertise. Despite advancements in TCM-based decision systems, machine learning, and deep learning research, limitations in data and single-objective constraints hinder their practical application. In recent years, large language models (LLMs) have demonstrated potential in complex tasks, but lack specialization in TCM and face significant challenges, such as too big model scale to deploy and issues with hallucination. To address these challenges, we introduce Tianyi with 7.6-billion-parameter LLM, a model scale proper and specifically designed for TCM, pre-trained and fine-tuned on diverse TCM corpora, including classical texts, expert treatises, clinical records, and knowledge graphs. Tianyi is designed to assimilate interconnected and systematic TCM knowledge through a progressive learning manner. Additionally, we establish TCMEval, a comprehensive evaluation benchmark, to assess LLMs in TCM examinations, clinical tasks, domain-specific question-answering, and real-world trials. The extensive evaluations demonstrate the significant potential of Tianyi as an AI assistant in TCM clinical practice and research, bridging the gap between TCM knowledge and practical application.

AIM:Peritoneal fibrosis (PF) is a progressive complication of long-term peritoneal dialysis, driven primarily by high-glucose-induced mesothelial injury, epithelial-mesenchymal transition, and chronic inflammation. Emerging evidence implicates autophagy dysfunction as a pivotal contributor to PF progression. Finerenone, a novel non-steroidal mineralocorticoid receptor antagonist, has shown potent anti-fibrotic and anti-inflammatory effects in cardiorenal diseases, yet its role in peritoneal fibrosis remains incompletely understood. This study aimed to investigate whether Finerenone alleviates PF by regulating autophagy activity in peritoneal mesothelial cells. METHODS:A mouse model of peritoneal fibrosis was established by daily intraperitoneal injection of 4.25% glucose-based peritoneal dialysis solution for 28 days. Finerenone (10 mg/kg) was administered orally concomitantly with dialysis fluid for 28 days. In vitro, human peritoneal mesothelial cells (HPMCs) were stimulated with TGF-β1 (2 ng/mL, 36 h) to induce fibrotic transformation. Cellular morphology, collagen deposition (Masson staining), and fibrosis markers (α-SMA, Fibronectin, Collagen I, and E-cadherin) were assessed by immunohistochemistry, immunofluorescence, and Western blot. Autophagy activity was evaluated by LC3-II/LC3-I ratio, p62 degradation, Beclin-1 and Atg5 expression, and chloroquine (CQ) co-treatment. Finally, Atg5 was silenced using siRNA to confirm its role in Finerenone-induced autophagy. RESULTS:Finerenone markedly attenuated peritoneal thickening and collagen accumulation in vivo and suppressed TGF-β1-induced fibrotic phenotype in HPMCs. Finerenone decreased α-SMA, Fibronectin, and Collagen I expression while restoring E-cadherin levels. Mechanistically, Finerenone reversed TGF-β1-induced autophagy blockade, as evidenced by increased LC3-II/LC3-I and Beclin-1, upregulated Atg5, and decreased p62 accumulation, indicating enhanced autophagic flux. The autophagy inhibitor CQ abolished these protective effects, leading to LC3-II and p62 accumulation. Moreover, Atg5 knockdown counteracted Finerenone-mediated suppression of fibrosis-related proteins, confirming that Finerenone's anti-fibrotic effects depend on Atg5-mediated autophagy activation. Finerenone also reduced pro-inflammatory cytokines (TGF-β1, IL-6, and IL-1β), suggesting that autophagy restoration alleviates inflammation in parallel. CONCLUSION:Finerenone alleviates peritoneal fibrosis, at least in part, by restoring Atg5-mediated autophagic flux and suppressing inflammation in peritoneal mesothelial cells. These findings provide mechanistic insight into its antifibrotic effects and support autophagy modulation as a potential therapeutic strategy for peritoneal dialysis-associated complications.

OBJECTIVE:This study aimed to present our experience in applying robotic-assisted surgery for managing pancreatic disorders in children. METHODS:A retrospective analysis was conducted on pediatric patients who underwent robot-assisted pancreatic procedures at two centers from April 2020 to August 2025. Information regarding preoperative status, intraoperative details, and postoperative outcomes was collected. RESULTS:A total of 49 pediatric patients received pancreatic surgery during the study period, including 19 boys and 30 girls, with a mean age of 10.3 ± 2.6 years. Surgical indications were pancreatic masses (n = 46), pancreatic trauma (n = 2), and congenital hyperinsulinism (n = 1). The mean operation time was 233.4 ± 117.3 min, and average blood loss was 59.4 ± 81.9 ml. Six patients (12.2 %) converted to open surgery. The total complication rate was 38.8 % (19/49), with major complications (Clavien-Dindo ≥ III) in 6.1 % (3/49). Postoperative pancreatic fistula (POPF) occurred in 7 patients (14.3 %), all of which were Grade B; two of these patients required ultrasound-guided external drainage. Two patients were readmitted within 30 days for intra-abdominal infection. The median follow-up was 23.4 months (range 3-62 months), and all remained alive without complications. CONCLUSION:Robot-assisted surgery for pediatric pancreatic disorders appears to be safe and feasible. The robotic platform can be employed for various complex pancreatic operations with satisfactory results.

ObjectiveTo explore the application effect of follow⁃up management based on the integrated theory of health behavior change in outpatients undergoing hemodialysis.MethodsBy using the convenience sampling method,a total of 66 outpatients undergoing hemodialysis in a tertiary grade A hospital in Shanxi province from April to August 2024 were selected as the research subjects,and randomly divided into control group and intervention group,with 33 cases in each.The patients in control group received routine care,and the patients in intervention group implemented follow⁃up management intervention plan based on the integrated theory of health behavior change on the basis of routine care.The interdialytic weight gain rate,objective biochemical indicators(serum potassium,serum calcium,serum phosphorus,serum albumin,hemoglobin),health literacy level,self⁃management level and treatment compliance effect of the two groups before the intervention and at 3 months and 6 months after the intervention were compared.ResultsThe time effect,intergroup effect and interaction effect of interdialytic weight gain rate,objective biochemical indicators,health literacy level,self⁃management level and treatment compliance of the two groups were statistically significant(P<0.05).At 3 months and 6 months after the intervention,all indicators of the intervention group were better than those in control group(P<0.05).ConclusionsFollow⁃up management based on the integrated theory of health behavior change could enhance the health literacy level,self⁃management ability and treatment compliance of outpatients undergoing hemodialysis,and reduce interdialytic weight gain rate and blood potassium levels during dialysis,improve calcium and phosphorus metabolism and nutritional status.

To compare the performance of predictive models for cardiovascular event (CVE) in patients undergoing peritoneal dialysis (PD) based on machine learning algorithm and Cox proportional hazard regression. This study included patients underwent PD catheterization in our center from January 1, 2010, to July 31, 2022. The patients were randomly divided into training and validation sets in a 7:3 ratio. Cox regression, extreme gradient boosting (XGBoost), and random survival forest (RSF) models were developed using the training set and validated using the validation set. The time-dependent area under the curve (AUC) and concordance index (C-index) were used to evaluate the discriminative ability of predictive models. A total of 318 patients were enrolled in this study. 110 (34.6

Early prediction of acute kidney injury (AKI) may provide a crucial opportunity for AKI prevention. To date, no prediction model targeting AKI among general hospitalized patients in developing countries has been published. Here we show a simple, real-time, interpretable AKI prediction model for general hospitalized patients developed from a large tertiary hospital in China, which has been validated across five independent, geographically distinct, different tiered hospitals. The model containing 20 readily available variables demonstrates consistent, high levels of predictive discrimination in validation cohort, with AUCs for serum creatinine-based AKI and severe AKI within 48 h ranging from 0.74–0.85 and 0.83–0.90 for transported models and from 0.81–0.90 and 0.88–0.95 for refitted models, respectively. With optimal probability cutoffs, the refitted model could predict AKI at a median of 72 (24–198) hours in advance in internal validation, and 54–90 h in advance in external validation. Broad application of the model in the future may provide an effective, convenient and cost-effective approach for AKI prevention. Early prediction of Acute kidney injury (AKI) may be crucial for AKI prevention. Here the authors present a simple, real-time, interpretable, AKI prediction model for hospitalized patients, based on routinely collected electronic health records data.

Introduction Although the impact of air pollutants on infectious diseases is well-known, there is limited evidence regarding its effects on peritoneal dialysis (PD) patients. This study aimed to investigate the association between air pollutants and PD-related peritonitis. Methods This is an observational study affiliated to the PD Telemedicine-assisted Platform Cohort Study (PDTAP study), which is a national-level cohort study in China. The primary outcome was PD-related peritonitis, and the secondary outcomes were peritonitis-related death and transfer to hemodialysis. The pollution data were obtained from China High Air Pollutants according to the patients' place of residence. The association between pollutants and outcomes was evaluated by cause-specific Cox proportional hazard regression model. The patients were divided into the high-pollution group and low-pollution group according to the median value of PM2.5 (53.90 mu g/m(3)) and the WHO standard of PM2.5 (35.00 mu g/m(3)). Results A total of 7439 PD patients from all 7 geographical regions across China were enrolled between June 2016 and April 2019. There were 1585 patients who developed peritonitis during follow-up. The pollution was most severe in the north and central regions of China. Patients in the high-pollution group were characterized by older age, higher BMI, lower income, from rural and non-university affiliated hospitals, and had more comorbidities and better residual renal function. In multivariate analysis, PM2.5 and its components (SO4, NO3, NH4, OM, and BC), PM10, NO2, and CO were associated with increased peritonitis risk (P < 0.001-0.027). Additionally, following the propensity score matching to control for key individual-level covariates, the association between PM2.5 and its components, NO2, and CO with elevated peritonitis risk remained significant (P < 0.001). Conclusion In this national large-scale Chinese PD cohort study, air pollutants were found to be associated with increased risk for peritonitis.

This study aimed to investigate the potential relationship between baseline glucose to lymphocyte ratio (GLR) levels and the first episode of peritonitis in patients treated with peritoneal dialysis (PD). A total of 314 patients treated with PD were included and divided into three groups based on GLR tertiles: tertile 1 (GLR ≤ 4.23); tertile 2 (4.23 < GLR ≤ 5.96), and tertile 3 (GLR > 5.96). The relationships between GLR and the first peritonitis episode were analyzed using Kaplan–Meier curves and multivariable Cox regression models. Competitive risk analysis, subgroup and sensitivity analyses were performed to validate the robustness of the findings. During a median follow-up of 27 months, 123 (39.17%) patients developed the first episode of peritonitis. The incidence of the first peritonitis episode increased with the higher GLR tertiles (tertile 1: 32.08%, tertile 2: 37.50%, tertile 3: 48.08%). Kaplan–Meier curves revealed significant differences in the cumulative incidence of the first peritonitis episode among the GLR tertiles (Log-Rank test, P = 0.018). After full adjustment for confounding factors, patients in tertile 3 remained at significantly higher risk for the first episode of peritonitis compared to those in tertile 1 (HR 2.633, 95% CI 1.223–5.668, P = 0.013). Competitive risk models and sensitivity analysis further confirmed this association. Our study suggests that elevated GLR is associated with an increased risk of the first peritonitis episode in patients with PD.

BackgroundThere are inequalities in resource allocation and services across peritoneal dialysis (PD) centers in China. This study aimed to explore the association between hospital type (university-affiliated vs. non-university-affiliated hospitals) and clinical outcomes in PD patients.MethodsData from the Peritoneal Dialysis Telemedicine-assisted Platform cohort was analyzed. The primary outcome was all-cause mortality, while secondary outcomes included hemodialysis transfer and first-episode PD-related peritonitis. Univariable and multivariable Fine-Gray models were used to calculate subdistribution hazard ratios (SHRs). Propensity-score matched analyses and sensitivity analyses restricted to incident patients were also performed.ResultsA total of 7416 PD patients were enrolled (June 2016 to April 2019), with a median follow-up of 29.0 months. University-affiliated hospitals' patients (n = 4806) were younger, had better nutritional status, and higher socio-economic status than those in non-university-affiliated hospitals (n = 2610). University-affiliated hospitals exhibited a lower risk for all-cause mortality (SHR: 0.72, 95% confidence interval (CI): 0.61-0.85, p < 0.001), higher hemodialysis transfer (SHR: 1.31, 95% CI: 1.08-1.60, p < 0.01), but no association with first-episode peritonitis in multivariable analyses. After propensity-score matching, university-affiliated hospitals were still associated with lower all-cause mortality (SHR: 0.74, 95% CI: 0.61-0.91, p < 0.01) and a higher risk of hemodialysis transfer (SHR: 1.52, 95% CI: 1.19-1.94, p < 0.01). Comparable results for all-cause mortality and first-episode peritonitis also found in incident patients.ConclusionIn China, PD patients in university-affiliated hospitals had lower mortality but a higher risk of hemodialysis transfer. Further studies are needed to understand these findings and inform future practices and resource allocations.

Peritoneal dialysis (PD)-related peritonitis is a common complication with high morbidity and mortality, and empirical antibiotic regimens vary across countries. Despite some research, inconsistent results and design limitations highlight the need to reassess the association between these regimens and outcomes. This study was affiliated with the PD Telemedicine-assisted Platform (PDTAP) study. The primary outcome was peritonitis-associated death, and the secondary outcomes were peritonitis-associated hemodialysis transfer and subsequent peritonitis within 6 months. Propensity score matching and logistic regression were used to access the relationship between empirical antibiotic administration and outcomes. Altogether, 1431 patients experienced a first episode of peritonitis from June 1, 2016, to April 30, 2019. Among them, 1203 patients were assigned to the cefazolin-based group (n = 637) or to the vancomycin-based group (n = 566) based on administration of empirical antibiotics against Gram-positive bacteria. Compared to the cefazolin-based group, patients in the vancomycin-based group were older, had a longer PD duration, and reported higher income, along with a greater prevalence of diabetes, cardiovascular disease, and peritonitis history (P < 0.05 for all). Both groups exhibited similar rates of peritonitis-associated death and subsequent peritonitis within 6 months (P > 0.05 for all), however, the vancomycin-based group was more prone to to hemodialysis transfer (11.00

Background and hypothesis:Iron metabolism markers, such as transferrin saturation (TSAT) and ferritin, are crucial in anemia management in patients with CKD and those undergoing dialysis, yet optimal levels remain unelucidated. Methods:We conducted a prospective multicenter cohort study using data from the nationwide Peritoneal Dialysis Telemedicine-based Management Platform (PDTAP) to analyze TSAT, ferritin, and hemoglobin (Hb) levels, and their associations with mortality in the peritoneal dialysis (PD) population. Results:Our study included 4429 PD patients, analyzing data through restricted cubic splines and Cox regression models, adjusted for multiple confounders. Non-linear associations between Hb levels and TSAT/ferritin were observed. Hb levels increased with TSAT up to 40%, then plateaued, whereas ferritin levels increased with the decline of Hb. Further, ferritin levels above 200 ng/mL were independently linked to increased mortality risk [hazard ratio (HR) 1.207, 95% confidence interval (CI) 1.134-1.286], with this effect decreasing as high-sensitivity C-reactive protein levels rose. This risk was notably significant in patients with a history of cardiovascular disease. A ferritin/Hb ratio >2 was associated with increased risk of mortality after adjusting for demographic, nutritional factors and erythropoiesis agents (HR 1.219, 95% CI 1.144-1.299). The ferritin/Hb ratio demonstrated superior predictive ability for iron responsiveness compared with ferritin alone. Conclusion:Serum ferritin level exceeding 200 ng/mL was indepently associated with a higher risk of martality in Chinese PD population. Monitoring the ferritin/Hb ratio may help assess the relative iron content in the body and provide reference for iron supplementation among patients undergoing PD.

Although more and more cardiovascular risk factors have been verified in peritoneal dialysis (PD) populations in different countries and regions, it is still difficult for clinicians to accurately and individually predict death in the near future. We aimed to develop and validate machine learning-based models to predict near-term all-cause and cardiovascular death. Machine learning models were developed among 7539 PD patients, which were randomly divided into a training set and an internal test set by five random shuffles of 5-fold cross-validation, to predict the cardiovascular death and all-cause death in 3 months. We chose objectively collected markers such as patient demographics, clinical characteristics, laboratory data, and dialysis-related variables to inform the models and assessed the predictive performance using a range of common performance metrics, such as sensitivity, positive predictive values, the area under the receiver operating curve (AUROC), and the area under the precision recall curve. In the test set, the CVDformer models had a AUROC of 0.8767 (0.8129, 0.9045) and 0.9026 (0.8404, 0.9352) and area under the precision recall curve of 0.9338 (0.8134,0.9453) and 0.9073 (0.8412, 0.9164) in predicting near-term all-cause death and cardiovascular death, respectively. The CVDformer models had high sensitivity and positive predictive values for predicting all-cause and cardiovascular deaths in 3 months in our PD population. Further calibration is warranted in the future.

Hypokalemia has been associated with an increased risk of peritoneal dialysis (PD)-associated peritonitis. However, hypokalemia is commonly associated with malnutrition, inflammation, and severe coexisting comorbidities, which thus are suspected of being potential confounders. This study was aimed at testing whether hypokalemia was independently associated with the occurrence and prognosis of PD-associated peritonitis. A national-level dataset from the Peritoneal Dialysis Telemedicine-assisted Platform Cohort (PDTAP) Study was used to explore the independent association of serum potassium with PD-associated peritonitis. Unmatched and propensity score-adjusted multivariate competing risk models, as well as univariate competing risk models following 1:1 propensity score matching, were conducted to balance potential biases between patients with and without hypokalemia. The association between potassium levels prior to peritonitis and treatment failure due to peritonitis was also investigated. During a median follow-up of 25.7 months in 7220 PD patients, there was a higher incidence of peritonitis in patients with serum potassium below 4.0 mmol/L compared to those with higher serum levels (677 [0.114/patient-year] vs. 914 [0.096/patient-year], P = 0.001). After adjusting for demographics, laboratory tests, residual renal function, and medication use, baseline potassium levels below 4.0 mmol/L were not linked to an increased risk of peritonitis, with a hazard ratio of 0.983 (95

ABSTRACT Background To explore the cut-off values of haemoglobin (Hb) on adverse clinical outcomes in incident peritoneal dialysis (PD) patients based on a national-level database. Methods The observational cohort study was from the Peritoneal Dialysis Telemedicine-assisted Platform (PDTAP) dataset. The primary outcomes were all-cause mortality, major adverse cardiovascular events (MACE) and modified MACE (MACE+). The secondary outcomes were the occurrences of hospitalization, first-episode peritonitis and permanent transfer to haemodialysis (HD). Results A total of 2591 PD patients were enrolled between June 2016 and April 2019 and followed up until December 2020. Baseline and time-averaged Hb <100 g/l were associated with all-cause mortality, MACE, MACE+ and hospitalizations. After multivariable adjustments, only time-averaged Hb <100 g/l significantly predicted a higher risk for all-cause mortality {hazard ratio [HR] 1.83 [95% confidence interval (CI) 1.19–281], P = .006}, MACE [HR 1.99 (95% CI 1.16–3.40), P = .012] and MACE+ [HR 1.77 (95% CI 1.15–2.73), P = .010] in the total cohort. No associations between Hb and hospitalizations, transfer to HD and first-episode peritonitis were observed. Among patients with Hb ≥100 g/l at baseline, younger age, female, use of iron supplementation, lower values of serum albumin and renal Kt/V independently predicted the incidence of Hb <100 g/l during the follow-up. Conclusion This study provided real-world evidence on the cut-off value of Hb for predicting poorer outcomes through a nation-level prospective PD cohort.

AbstractBackground & aimsDiabetes is known to increase the risk of gallstone disease. This study assesses the impact of sodium-glucose cotransporter-2 inhibitors (SGLT2i) on the incidence of biliary diseases, relative to sulfonylureas, in patients with type 2 diabetes mellitus (T2DM) and lithogenic diet (LD)-fed mice.MethodsA retrospective cohort analysis was performed on T2DM patient data who commenced SGLT2i or sulfonylurea therapy from January 1, 2017, to September 1, 2022-sourced from Nanjing Medical University’s database. They were matched using propensity scores (PS) and inverse probability of treatment weighting (IPTW). Follow-up for developing biliary diseases was conducted up to the earliest relevant end-point. Cox models, PS matching, and sensitivity analyses, including standard mortality ratio weighting (SMRW), were applied to determine hazard ratios (HRs) and confidence intervals (CIs). Parallelly, LD-fed C57BL/6J mice were administered SGLT2i or sulfonylureas to corroborate findings in animal models.ResultsFrom the 1,901 patients analyzed over an average of 2.83 years, SGLT2i therapy correlated with a substantially lower risk of developing biliary diseases (HR 0.595, 95% CI 0.410-0.863), particularly among defined subgroups. A downward trend in risk was observed with extended use beyond two years. Concordant data from the mouse model pointed towards SGLT2i mitigating gallstone formation, with putative mechanisms including reduced liver injury and dyslipidemia, as well as improved gallbladder motility and bile acid production.ConclusionSGLT2i potentially reduces the risk of biliary diseases compared to sulfonylureas, meriting further clinical investigation.