Insulin resistance (IR) is associated with diabetes mellitus, cardiovascular disease (CV), and mortality. Few studies have used machine learning to predict IR in the non-diabetic population. In this prospective cohort study, we trained a predictive model for IR in the non-diabetic populations using the US National Health and Nutrition Examination Survey (NHANES, from JAN 01, 1999 to DEC 31, 2012) database and the Taiwan MAJOR (from JAN 01, 2008 to DEC 31, 2017) database. We analysed participants in the NHANES and MAJOR and participants were excluded if they were aged 0.8 (highest being XGboost, 0.87). In the test group, all AUC were also >0.80 (highest being XGboost, 0.88). Among all 9 features (age, gender, race, body mass index, fasting plasma glucose (FPG), glycohemoglobin, triglyceride, total cholesterol and high-density cholesterol), BMI had the highest value of feature importance on IR (0.43 for XGboost and 0.47 for RF algorithms). All participants from the TWB database were separated into the IR group and the non-IR group according to the XGboost algorithm. The Kaplan–Meier survival curve showed a significant difference between the IR and non-IR groups (p < 0.0001 for CV mortality, and p = 0.0006 for all-cause mortality). Therefore, the XGboost model has clear clinical implications for predicting IR, aside from CV and all-cause mortality. To predict IR in non-diabetic patients with high accuracy, only 9 easily obtained features are needed for prediction accuracy using our machine learning model. Similarly, the model predicts IR patients with significantly higher CV and all-cause mortality. The model can be applied to both Asian and Caucasian populations in clinical practice. 10.13039/501100010101 Taichung Veterans General Hospital , Taiwan and 10.13039/501100001691 Japan Society for the Promotion of Science KAKENHI Grant Number JP21KK0293.

Idiopathic membranous nephropathy (iMN) is the major cause of end-stage renal disease (ESRD). Recent guidelines suggest limiting immunosuppressants only to high risk patients for ESRD. The present study is aimed at identifying new predictors for the renal outcome of iMN patients. We conducted a retrospective cohort study covering a period from January 2003 to December 2013. We enrolled participants who had received their first renal biopsy at our medical center in Taiwan with the diagnosis of iMN. Clinical, pathological and laboratory data were collected from medical records. Analyses with Mann–Whitney U test was used for continuous variables and Chi-square test for categorical variables. The Kaplan-Meier curve was used for the analyses of patient survival and renal survival. Youden index was used for evaluating the performance of a dichotomous diagnostic test for renal and patient outcomes. Cox proportional hazard regression was used to determine factors affecting renal survival.A total of 99 patients with renal biopsy-confirmed idiopathic iMNs were enrolled. C3 level ≤114 mg/dl predicted patient outcome (p < 0.001) with good predictive power (AUC = 0.736). The univariate analysis showed that risk factors for poor renal outcome were older age (HR = 1.04, p = 0.002), high BUN (HR = 1.03, p < 0.001), poor baseline renal function (HR = 1.30 and p < 0.001 for higher serum creatinine; HR = 0.97 and p < 0.001 for higher eGFR; HR = 1.06 and p < 0.001 for urine PCR), C3 ≤ 93.4 mg/dl (HR = 2.15, p = 0.017), NLR > 3.34 (HR = 3.30, p < 0.001) and PLR > 14.48 (HR = 2.54, p = 0.003). Stage of iMN did not fully account for the risk of ESRD. This is the first evidence that serum levels of C3 ≤ 93.4 mg/dl predicted poor renal outcomes with good predictive power. Easily obtained markers, NLR > 3.34 also predicted poor renal outcomes.

Background: The effect of glucose control, especially variability of glycated hemoglobin (HbA1c), on estimated glomerular filtration rate (eGFR) decline in type 2 diabetes is still debatable. Methods: We used tertiles of coefficient of variation (CV) to determine the variability of HbA1c (HbA1c_CV). Mixed model repeated measures (MMRM) were used to evaluate the annual eGFR decline rate. Results: In 1383 type 2 diabetic patients, we found the greater the HbA1c_CV, the greater the eGFR decline ( p = 0.01, −0.99 in low, −1.73 in mid, and −2.53 ml/min/1.73 m²/year in high HbA1c_CV). Regardless of eGFR (⩾60 or <60 ml/min/1.73 m²), the same result holds ( p = 0.019 and p = 0.007, respectively). In subgroup analysis of baseline HbA1c (%) (HbA1c < 7, 7 ⩽ HbA1c < 9, and HbA1c ⩾ 9), tertiles of HbA1c_CV showed similar effects on annual decline of eGFR ( p = 0.193, 0.300, 0.182, respectively), although a trend for a steeper decline in renal function in the highest HbA1c_CV tertile was observed for all HbA1c strata, and even for HbA1c < 7%. A similar behavior was observed in patients with macroalbuminuria or normoalbuminuria ( p = 0.219, and 0.109, respectively), with a significant trend in those with microalbuminuria ( p = 0.019). Even in patients with HbA1c < 7, high HbA1c_CV also predicts rapid eGFR decline. Before macroalbuminuria, minimizing HbA1c_CV also has renal benefit. Conclusions: HbA1c variability is an independent risk factor for deterioration of renal function. Even with well-controlled HbA1c levels (<7%), patients with high HbA1c_CV still experienced faster eGFR decline. Early minimization of glycemic variability (before macroalbuminuira) can curb deterioration of renal function. Monitoring and lowering of HbA1c_CV is highly recommended for diabetic care.

Background: Chronic kidney disease (CKD) is a complex syndrome without a definitive treatment. For these patients, insulin resistance (IR) is associated with worse renal and patient outcomes. Until now, no predictive model using machine learning (ML) has been reported on IR in CKD patients. Methods: The CKD population studied was based on results from the National Health and Nutrition Examination Survey (NHANES) of the USA from 1999 to 2012. The homeostasis model assessment of IR (HOMA-IR) was used to assess insulin resistance. We began the model building process via the ML algorithm (random forest (RF), eXtreme Gradient Boosting (XGboost), logistic regression algorithms, and deep neural learning (DNN)). We compared different receiver operating characteristic (ROC) curves from different algorithms. Finally, we used SHAP values (SHapley Additive exPlanations) to explain how the different ML models worked. Results: In this study population, 71,916 participants were enrolled. Finally, we analyzed 1,229 of these participants. Their data were segregated into the IR group (HOMA IR > 3, n = 572) or non-IR group (HOMR IR ≤ 3, n = 657). In the validation group, RF had a higher accuracy (0.77), specificity (0.81), PPV (0.77), and NPV (0.77). In the test group, XGboost had a higher AUC of ROC (0.78). In addition, XGBoost also had a higher accuracy (0.7) and NPV (0.71). RF had a higher accuracy (0.7), specificity (0.78), and PPV (0.7). In the RF algorithm, the body mass index had a much larger impact on IR (0.1654), followed by triglyceride (0.0117), the daily calorie intake (0.0602), blood HDL value (0.0587), and age (0.0446). As for the SHAP value, in the RF algorithm, almost all features were well separated to show a positive or negative association with IR. Conclusion: This was the first study using ML to predict IR in patients with CKD. Our results showed that the RF algorithm had the best AUC of ROC and the best SHAP value differentiation. This was also the first study that included both macronutrients and micronutrients. We concluded that ML algorithms, particularly RF, can help determine risk factors and predict IR in patients with CKD.

Introduction: Hyperuricemia and diabetes mellitus (DM) are associated with increased mortality risk in patients with chronic kidney disease (CKD). Here we aimed to evaluate the independent and joint risks of these two conditions on mortality and end stage kidney disease (ESKD) in CKD-patients.Methods: This retrospective cohort study enrolled 4380 outpatients (with CKD stage 3–5) with mortality and ESKD linkage during a 7-year period (from 2007 to 2013). All-causes mortality and ESKD risks were analyzed by multivariable-adjusted Cox proportional hazards models (adjusted for age, sex, smoke, previous coronary arterial disease, blood pressure, and medications for hyperlipidemia, hyperuricemia and renin–angiotensin system inhibitors).Results: Overall, 40.5% of participants had DM and 66.4% had hyperuricemia. In total, 356 deaths and 932 ESKD events occurred during the 7 years follow-up. With the multivariate analysis, increased risks for all-cause mortality were: hyperuricemia alone, HR = 1.48 (1–2.19); DM alone, and HR = 1.52 (1.02–2.46); DM and hyperuricemia together, HR = 2.12 (1.41–3.19). Similar risks for ESKD were: hyperuricemia alone, HR = 1.34 (1.03–1.73); DM alone, HR = 1.59 (1.15–2.2); DM and hyperuricemia together, HR = 2.46 (1.87–3.22).Conclusions: DM and hyperuricemia are strongly associated with higher all-cause mortality and ESKD risk in patients with CKD stage 3–5. Hyperuricemia is similar to DM in terms of risk for all-cause mortality and ESKD. DM and hyperuricemia when occurred together further increase both risks of all-cause mortality and ESKD.

Contrast-induced acute kidney injury (CI-AKI) is associated with high risks of morbidity and mortality. Hyperbilirubinemia might have some renal protection but with no clear cutoff value for protection. Related studies are typically on limited numbers of patients and only in conditions of vascular intervention. We performed this study to elucidate CI-AKI in patients after contrast-enhanced computed tomography (CCT). The outcomes were CI-AKI, dialysis and mortality. Patients were divided to three groups based on their serum levels of total bilirubin: ≤1.2 mg/dl, 1.3–2.0 mg/dl, and >2.0 mg/dl. We enrolled a total of 9,496 patients who had received CCT. Patients with serum total bilirubin >2.0 mg/dl were associated with CI-AKI. Those undergoing dialysis had the highest incidence of PC-AKI (p2mg/dl were associated with CI-AKI (OR = 1.89, 1.53–2.33 of 95% CI), dialysis (OR = 1.40, 1.01–1.95 of 95% CI) and mortality (OR = 1.63, 1.38–1.93 of 95% CI) after adjusting for laboratory data and all comorbidities (i.e., cerebrovascular disease, coronary artery disease, peripheral arterial disease, and acute myocardial infarction, diabetes mellitus, hypertension, gastrointestinal bleeding, cirrhosis, peritonitis, ascites, hepatoma, shock lung and colon cancer). We concluded that total bilirubin level >2 mg/dl is an independent risk factor for CI-AKI, dialysis and mortality after CCT. These patients also had high risks for cirrhosis or hepatoma. This is the first study providing evidence that hyperbilirubinemia (total bilirubin >2.0 mg/dl) being an independent risk factor for CI-AKI, dialysis and mortality after receiving CCT. Most patients with total bilirubin >2.0mg/dl had cirrhosis or hepatoma.