Background: Malnutrition and inflammation are highly prevalent and tightly regulated with each other in chronic kidney disease (CKD) patients. Inflammation can lead to malnutrition in patients with sufficient nourishment, while malnutrition may also induce an inflammatory response. This study investigated whether the albumin-globulin ratio (AGR) can predict the mortality risk in CKD patients. Methods: We enrolled 956 stage 3–5 CKD patients retrospectively at a medical center. Patients’ baseline characteristics including demographics, laboratory data, pharmacotherapy, and comorbidities were collected for statistical adjustments. The study patients were stratified into three AGR groups according to similar magnitudes of hazards for mortality as follows: low AGR group, AGR ≤ 1.0; moderate AGR group, 1.1 ≤ AGR < 1.3; high AGR group, AGR ≥1.3. Multivariate Cox proportional hazard analysis was performed to evaluate the association of the AGR with the study outcomes, including overall and cardiovascular disease (CVD) mortality. Results: During a median follow-up duration of 2.44 years, 108 (11.3%) deaths were recorded and 50 patients died from CVD. In adjusted model 1, the moderate AGR group was associated with hazard ratios (HR) of 0.57 (95% CI = 0.36–0.90, p = 0.016) and 0.52 (95% CI = 0.28–0.98, p = 0.043) for all-cause and CVD mortality compared with the low AGR group, respectively. The high AGR group was associated with HRs of 0.49 (95% CI = 0.27–0.90, p = 0.021) and 0.27 (95% CI = 0.1–0.74, p = 0.01) for all-cause and CVD mortality compared with the low AGR group, respectively. Similar results were obtained in the adjusted model 2 (inverse probability of the group weighted Cox model). In addition, the association between the AGR and mortality risk remained significant when the AGR was treated as a continuous variable. Conclusion: AGR is a significant biomarker predicting overall and cardiovascular mortality risk independent of various important factors amongst stage 3–5 CKD patients. We suggest that the AGR may be a simple and inexpensive measurement for detecting CKD patients at risk of mortality.

Background: Nutrition and inflammation have been implicated in predicting mortality in patients on peritoneal dialysis (PD). Serum albumin and globulin can be regarded for the nutritional and inflammatory status. However, there is lack of data to evaluate the synergistic effect of albumin and globulin on mortality prediction. Methods: In 554 patients initiating PD from January 2001 to July 2016, we divided them into four groups by the combination of two categories of low vs. high albumin and low vs. high globulin. The median values for albumin and globulin were chosen to classify them into low or high groups. Their associations with all-cause and cardiovascular (CV) mortality were examined in Cox regression models adjusted for confounding clinical and laboratory data. Results: Patients, 52.91 ± 15.2 years old and 47.8% men, had a median (interquartile range) value of 3.3 (2.9–3.8) g/dL for albumin and 2.8 (2.5–3.2) g/dL for globulin, respectively. Patients with low albumin and high globulin had the highest all-cause mortality and CV mortality, with adjusted hazard ratios of 3.87 (95% CI 1.83–8.20, p < 0.001) and 5.65 (95% CI 2.23–14.34, p < 0.001), respectively, compared with those with a high albumin and low globulin having the lowest mortality rate. Sensitivity analyses further confirmed this relationship. Conclusions: A patient profile of either low albumin or high globulin is linked to a higher risk for mortality, particularly for a profile of both low albumin and high globulin compared with one without either of them. Further studies are needed to explore the mechanisms underlying this phenomenon and how to improve clinical outcomes in those high-risk patients.

There is little information about the association between stroke and kidney diseases. We aimed to investigate the impact of stroke on long-term renal outcomes. In this large population-based retrospective cohort study, we identified 100,353 subjects registered in the National Health Insurance Research Database of Taiwan from January 1, 2000, through December 31, 2012, including 33,451 stroke patients and 66,902 age-, sex- and Charlson’s comorbidity index score-matched controls. The incidence rate of chronic kidney disease (CKD) was higher in the stroke than in the control cohort (17.5 vs. 9.06 per 1000 person-years). After multivariate adjustment, the risk of developing CKD was significantly higher in patients with stroke (adjusted hazard ratio [aHR] 1.43, 95% confidence interval [CI] 1.36–1.50, P<0.001). Subgroup analysis showed that stroke patients <50 years (aHR 1.61, P<0.001) and those with concomitant diabetes mellitus (aHR 2.12, P<0.001), hyperlipidemia (aHR 1.53, P<0.001) or gout (aHR 1.84, P<0.001) were at higher risk of incident CKD. Additionally, the risks of progression to advanced CKD and end-stage renal disease (ESRD) were significantly higher for stroke patients (aHRs, 1.22 and 1.30; P = 0.04 and P = 0.008, respectively), independent of age, sex, comorbidities and long-term medications. Stroke is associated with higher risks for incident CKD, decline in renal function and ESRD. Younger stroke patients, as well as those with concomitant diabetes mellitus, hyperlipidemia or gout are at greater risk for kidney diseases.

Background: General severity of illness scores are not well calibrated to predict mortality among patients receiving renal replacement therapy (RRT) for acute kidney injury (AKI). We developed machine learning models to make mortality prediction and compared their performance to that of the Sequential Organ Failure Assessment (SOFA) and HEpatic failure, LactatE, NorepInephrine, medical Condition, and Creatinine (HELENICC) scores. Methods: We extracted routinely collected clinical data for AKI patients requiring RRT in the MIMIC and eICU databases. The development models were trained in 80% of the pooled dataset and tested in the rest of the pooled dataset. We compared the area under the receiver operating characteristic curves (AUCs) of four machine learning models (multilayer perceptron [MLP], logistic regression, XGBoost, and random forest [RF]) to that of the SOFA, nonrenal SOFA, and HELENICC scores and assessed calibration, sensitivity, specificity, positive (PPV) and negative (NPV) predicted values, and accuracy. Results: The mortality AUC of machine learning models was highest for XGBoost (0.823; 95% confidence interval [CI], 0.791–0.854) in the testing dataset, and it had the highest accuracy (0.758). The XGBoost model showed no evidence of lack of fit with the Hosmer–Lemeshow test (p > 0.05). Conclusion: XGBoost provided the highest performance of mortality prediction for patients with AKI requiring RRT compared with previous scoring systems.

Background: Hyperkalemia is a common complication of chronic kidney disease (CKD). Hyperkalemia is associated with mortality, CKD progression, hospitalization, and high healthcare costs in patients with CKD. We developed a machine learning model to predict hyperkalemia in patients with advanced CKD at an outpatient clinic. Methods: This retrospective study included 1,965 advanced CKD patients between January 1, 2010, and December 31, 2020 in Taiwan. We randomly divided all patients into the training (75%) and testing (25%) datasets. The primary outcome was to predict hyperkalemia (K⁺ > 5.5 mEq/L) in the next clinic vist. Two nephrologists were enrolled in a human-machine competition. The area under the receiver operating characteristic curves (AUCs), sensitivity, specificity, and accuracy were used to evaluate the performance of XGBoost and conventional logistic regression models with that of these physicians. Results: In a human-machine competition of hyperkalemia prediction, the AUC, PPV, and accuracy of the XGBoost model were 0.867 (95% confidence interval: 0.840–0.894), 0.700, and 0.933, which was significantly better than that of our clinicians. There were four variables that were chosen as high-ranking variables in XGBoost and logistic regression models, including hemoglobin, the serum potassium level in the previous visit, angiotensin receptor blocker use, and calcium polystyrene sulfonate use. Conclusions: The XGBoost model provided better predictive performance for hyperkalemia than physicians at the outpatient clinic.