Individualized prediction of mortality using multiple inflammatory markers in patients on dialysis

Associations between liver function parameters and poor clinical outcomes in peritoneal dialysis patients

Patients with end-stage renal disease (ESRD) have significantly lower survival rates compared with the general population of the same age. Peritoneal dialysis (PD) is an effective treatment for patients with ESRD, but the clinical outcome of PD patients is still not promising. The survival of PD patients is associated with various clinical factors, and exploring some valid risk predictors may be beneficial for this population. In this review, by integrating the latest research, we summarized the association of some common and novel liver function parameters (ALT, AST, ALP, GGT, serum bilirubin, pre-albumin, albumin, albumin-globulin ratio [AGR], serum ferritin, and hyaluronic acid) with clinical outcomes in PD patients. It may contribute to a better understanding of potential risk factors and help to develop strategies to prevent the disease progression.

Predicting mortality risk in dialysis: Assessment of risk factors using traditional and advanced modeling techniques within the Monitoring Dialysis Outcomes initiative

INTRODUCTION

Several factors affect the survival of End Stage Kidney Disease (ESKD) patients on dialysis. Machine learning (ML) models may help tackle multivariable and complex, often non-linear predictors of adverse clinical events in ESKD patients. In this study, we used advanced ML method as well as a traditional statistical method to develop and compare the risk factors for mortality prediction model in hemodialysis (HD) patients.

MATERIALS AND METHODS

We included data HD patients who had data across a baseline period of at least 1 year and 1 day in the internationally representative Monitoring Dialysis Outcomes (MONDO) Initiative dataset. Twenty-three input parameters considered in the model were chosen in an a priori manner. The prediction model used 1 year baseline data to predict death in the following 3 years. The dataset was randomly split into 80% training data and 20% testing data for model development. Two different modeling techniques were used to build the mortality prediction model.

FINDINGS

A total of 95,142 patients were included in the analysis sample. The area under the receiver operating curve (AUROC) of the model on the test data with XGBoost ML model was 0.84 on the training data and 0.80 on the test data. AUROC of the logistic regression model was 0.73 on training data and 0.75 on test data. Four out of the top five predictors were common to both modeling strategies.

DISCUSSION

In the internationally representative MONDO data for HD patients, we describe the development of a ML model and a traditional statistical model that was suitable for classification of a prevalent HD patient's 3-year risk of death. While both models had a reasonably high AUROC, the ML model was able to identify levels of hematocrit (HCT) as an important risk factor in mortality. If implemented in clinical practice, such proof-of-concept models could be used to provide pre-emptive care for HD patients.

Development and Validation of a Nomogram Based on Nutritional Indicators and Tumor Markers for Prognosis Prediction of Pancreatic Ductal Adenocarcinoma

Purpose

This study aimed to develop and validate a nomogram with preoperative nutritional indicators and tumor markers for predicting prognosis of patients with pancreatic ductal adenocarcinoma (PDAC).

Methods

We performed a bicentric, retrospective study including 155 eligible patients with PDAC. Patients were divided into a training group (n = 95), an internal validation group (n = 34), an external validation group (n = 26), and an entire validation group (n = 60). Cox regression analysis was conducted in the training group to identify independent prognostic factors to construct a nomogram for overall survival (OS) prediction. The performance of the nomogram was assessed in validation groups and through comparison with controlling nutritional status (CONUT) and prognostic nutrition index (PNI).

Results

The least absolute shrinkage and selection operator (LASSO) regression, univariate and multivariate Cox regression analysis revealed that serum albumin and lymphocyte count were independent protective factors while CA19-9 and diabetes were independent risk factors. The concordance index (C-index) of the nomogram in the training, internal validation, external validation and entire validation groups were 0.777, 0.769, 0.759 and 0.774 respectively. The areas under curve (AUC) of the nomogram in each group were 0.861, 0.845, 0.773, and 0.814. C-index and AUC of the nomogram were better than those of CONUT and PNI in the training and validation groups. The net reclassification index (NRI), integrated discrimination improvement (IDI) and decision curve analysis showed improvement of accuracy of the nomogram in predicting OS and better net benefit in guiding clinical decisions in comparison with CONUT and PNI.

Conclusions

The nomogram incorporating four preoperative nutritional and tumor markers including serum albumin concentration, lymphocyte count, CA19-9 and diabetes mellitus could predict the prognosis more accurately than CONUT and PNI and may serve as a clinical decision support tool to determine what treatment options to choose.

Risk factors for mortality in patients undergoing peritoneal dialysis: a systematic review and meta-analysis

Aim

Inconsistent investigations of the risk factors for all-cause mortality in patients undergoing peritoneal dialysis (PD) were reported. The present meta-analysis aimed to assess the impact of some clinical characteristics on the risk of mortality in PD patients.

Methods

PubMed and Embase were systematically searched for studies evaluating the risk factors for all-cause mortality in PD patients. Hazard ratio (HR) and 95% confidence interval (CI) were derived using a random-effect or fixed-effect model considering the heterogeneity across studies.

Result

A total of 26 studies were included in this meta-analysis in accordance with the inclusion and exclusion criteria. Age, primary cardiovascular diseases, diabetes mellitus, and high level of alkaline phosphatase showed significant positive associations with elevated risk of all-cause and cardiovascular mortality in PD patients, while hemoglobin acted as a benefit factor. Furthermore, early onset of peritonitis, high peritoneal transport status, elevated body mass index and high-sensitivity C-reactive protein could also considerably increase the risk of all-cause mortality. The absolute serum level of magnesium, potassium, and uric acid required to improve survival in PD patients should be verified further.

Conclusions

Multiple factors could affect the risk of mortality in PD patients.

Abnormal iron status is associated with an increased risk of mortality in patients on peritoneal dialysis

BACKGROUND AND AIMS

Iron deficiency is prevalent, but there is limited data about the relationship between iron status and poor outcomes in chronic kidney disease patients undergoing peritoneal dialysis (PD). We aimed to investigate the association between iron status and mortality in PD patients.

METHODS AND RESULTS

This retrospective study was conducted on incident PD patients from January 2006 to December 2016 and followed up until December 2018. Patients were categorized into four groups according to baseline serum transferrin saturation (percent) and ferritin levels (ng/ml): reference (20-30%, 100-500 ng/ml), absolute iron deficiency (<20%, <100 ng/ml), function iron deficiency (FID) (<20%, >100 ng/ml), and high iron (>30%, >500 ng/ml). Among the 1173 patients, 77.5% had iron deficiency. During a median follow-up period of 43.7 months, compared with the reference group, the FID group was associated with increased risk for all-cause [adjusted hazard ratio (aHR) 1.87, 95% confidence interval (95% CI) 1.05-3.31, P = 0.032], but not cardiovascular (CV) mortality. Additionally, the high iron group had a more than four-fold increased risk of both all-cause and CV mortality [aHR 4.32 (95% CI 1.90-9.81), P < 0.001; aHR 4.41 (95% CI 1.47-13.27), P = 0.008; respectively].

CONCLUSION

FID and high iron predict worse prognosis of patients on PD.

Comparison of risk models for mortality and cardiovascular events between machine learning and conventional logistic regression analysis

AIMS

Non-linear models by machine learning may identify different risk factors with different weighting in comparison to conventional linear models.

METHODS AND RESULTS

The analyses were performed in 15,933 patients included in the Shinken Database (SD) 2004-2014 (n = 22,022) for whom baseline data of blood sampling and ultrasound cardiogram and follow-up data at 2 years were available. Using non-linear models with machine learning software, 118 risk factors and their weighting of risk for all-cause mortality, heart failure (HF), acute coronary syndrome (ACS), ischemic stroke (IS), and intracranial hemorrhage (ICH) were identified, where the top two risk factors were albumin/hemoglobin, left ventricular ejection fraction/history of HF, history of ACS/anti-platelet use, history of IS/deceleration time, and history of ICH/warfarin use. The areas under the curve of the developed models for each event were 0.900, 0.912, 0.879, 0.758, and 0.753, respectively.

CONCLUSION

Here, we described our experience with the development of models for predicting cardiovascular prognosis by machine learning. Machine learning could identify risk predicting models with good predictive capability and good discrimination of the risk impact.

Heart Valve Disease Predict Mortality in Hemodialysis Patients: A Single Center Experience

Our aim is to investigate the clinical and laboratory findings affecting the mortality of the patients in 3 years follow-up who underwent hemodialysis at our center. In this retrospective, observational cohort study, 432 patients who underwent hemodialysis at our center for at least 5 months were included. The first recorded data and subsequent clinical findings of patients who died and survived were compared. Two hundred and ninety patients survived, 142 patients died. The mean age of the patients who died was higher (63.4 ± 12.3 years, vs. 52 ± 16.1 years, P = 0.0001), 60.5% of them had coronary artery disease (P = 0.0001), 93.7% of them had a heart valve disease. Duration of hemodialysis (survived 57 [21-260] months; died 44 [5-183] months, P = 0.000) was lower in patients who died. Serum potassium level before dialysis (5.1 ± 0.6; 4.9 ± 0.7 mEq/L, P = 0.030), parathyroid hormone (435 [4-3054]; 304 [1-3145] pg/mL, P = 0.0001), albumin (3.9 ± 0.4; 3.8 ± 0.4 mg/dL, P = 0.0001) and Kt/V (1.48 ± 0.3; 1.40 ± 0.3, P = 0.019) levels were lower, C-reactive protein (5[1-208]; 8.7[2-256] mg/L, P = 0.000) levels were higher in patients who died. Logistic regression analysis showed age (OR = 1.1), coronary artery disease (OR = 1.7) and more than one heart valve disease (OR = 2.4) are independent risk factors for mortality. Potassium level before dialysis (OR = 0.60), parathyroid hormone (OR = 0.99), and higher Kt/V (OR = 0.28) were found to be an advantage for survival. Age, coronary artery disease and especially pathology in more than one heart valve are risk factors for mortality. Heart valve problems might develop because of malnutrition and inflammation caused by the chronic renal failure.