Improved Mortality Prediction in Dialysis Patients Using Specific Clinical and Laboratory Data

Development and validation of a prediction model for all-cause mortality in maintenance dialysis patients: a multicenter retrospective cohort study

BACKGROUND

The mortality risk varies considerably among individual dialysis patients. This study aimed to develop a user-friendly predictive model for predicting all-cause mortality among dialysis patients.

METHODS

Retrospective data regarding dialysis patients were obtained from two hospitals. Patients in training cohort (N = 1421) were recruited from the Fifth Affiliated Hospital of Sun Yat-sen University, and patients in external validation cohort (N = 429) were recruited from the First Affiliated Hospital of Guangzhou University of Traditional Chinese Medicine. The follow-up endpoint event was all-cause death. Variables were selected by LASSO-Cox regression, and the model was constructed by Cox regression, which was presented in the form of nomogram and web-based tool. The discrimination and accuracy of the prediction model were assessed using C-indexes and calibration curves, while the clinical value was assessed by decision curve analysis (DCA).

RESULTS

The best predictors of 1-, 3-, and 5-year all-cause mortality contained nine independent factors, including age, body mass index (BMI), diabetes mellitus (DM), cardiovascular disease (CVD), cancer, urine volume, hemoglobin (HGB), albumin (ALB), and pleural effusion (PE). The 1-, 3-, and 5-year C-indexes in the training set (0.840, 0.866, and 0.846, respectively) and validation set (0.746, 0.783, and 0.741, respectively) were consistent with comparable performance. According to the calibration curve, the nomogram predicted survival accurately matched the actual survival rate. The DCA showed the nomogram got more clinical net benefit in both the training and validation sets.

CONCLUSIONS

The effective and convenient nomogram may help clinicians quantify the risk of mortality in maintenance dialysis patients.

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 of a risk prediction model for infection-related mortality in patients undergoing peritoneal dialysis

Background

Assessment of infection-related mortality remains inadequate in patients undergoing peritoneal dialysis. This study was performed to develop a risk model for predicting the 2-year infection-related mortality risk in patients undergoing peritoneal dialysis.

Methods

The study cohort comprised 606 patients who started and continued peritoneal dialysis for 90 at least days and was drawn from the Fukuoka Peritoneal Dialysis Database Registry Study in Japan. The patients were registered from 1 January 2006 to 31 December 2016 and followed up until 31 December 2017. To generate a prediction rule, the score for each variable was weighted by the regression coefficients calculated using a Cox proportional hazard model adjusted by risk factors for infection-related mortality, including patient characteristics, comorbidities, and laboratory data.

Results

During the follow-up period (median, 2.2 years), 138 patients died; 58 of them of infectious disease. The final model for infection-related mortality comprises six factors: age, sex, serum albumin, serum creatinine, total cholesterol, and weekly renal Kt/V. The incidence of infection-related mortality increased linearly with increasing total risk score (P for trend <0.001). Furthermore, the prediction model showed adequate discrimination (c-statistic = 0.79 [0.72–0.86]) and calibration (Hosmer–Lemeshow test, P = 0.47).

Conclusion

In this study, we developed a new model using clinical measures for predicting infection-related mortality in patients undergoing peritoneal dialysis.

THE ASSOCIATION OF LOW TRIIODOTHYRONINE SYNDROME, PROTEIN-ENERGY WASTING AND CHRONIC INFLAMMATION WITH MORTALITY IN PREVALENT DIALYSIS PATIENTS

SUMMARY – This prospective study in prevalent dialysis patients investigated prognostic properties of low triiodothyronine syndrome, protein-energy wasting and chronic inflammation. Ninety-four prevalent dialysis patients were followed-up for a median of 39 months. Demographic, anthropometric and biochemical parameters were collected at baseline. Univariate and multivariate analysis was done using Cox regression analysis. ROC curve analysis using survival status as a classification variable was performed with the goal of determining optimal cut-off values for numerical variables. In our population, low total triiodothyronine (hazard ratio (HR) 2.19, p=0.038), catheter as vascular access (HR 2.76, p=0.023), higher vintage (HR 1.01, p=0.014) and higher Charlson comorbidity index (HR 1.28, p=0.017) were statistically significantly associated with inferior survival. In our group of steady-state dialysis patients, total triiodothyronine seemed to be the strongest predictor of inferior survival among thyroid hormones. Taking this parameter into account, it was possible to identify patients at an increased risk of death even after adjustment for other prognostically relevant variables. However, after further adjustment for significant risk factors, the impact of C-reactive protein and albumin on survival disappeared due to the overlapping prognostic properties. We concluded that triiodothyronine was an independent prognostic factor in our study group.

Development and External Validation of Risk Scores for Cardiovascular Hospitalization and Rehospitalization in Patients With Diabetes

CONTEXT

Cardiovascular disease (CVD) is a common and costly reason for hospitalization and rehospitalization among patients with type 2 diabetes.

OBJECTIVE

This study aimed to develop and externally validate two risk-prediction models for cardiovascular hospitalization and cardiovascular rehospitalization.

DESIGN

Two independent prospective cohorts.

SETTING

The derivation cohort includes 4704 patients with type 2 diabetes from 18 general practices in Cambridgeshire. The validation cohort comprises 1121 patients with type 2 diabetes from post-trial follow-up data.

MAIN OUTCOME MEASURE

Cardiovascular hospitalization over 2 years and cardiovascular rehospitalization after 90 days of the prior CVD hospitalization.

RESULTS

The absolute rate of cardiovascular hospitalization and rehospitalization was 12.5% and 6.7% in the derivation cohort and 16.3% and 7.0% in the validation cohort. Discrimination of the models was similar in both cohorts, with C statistics above 0.70 and excellent calibration of observed and predicted risks.

CONCLUSION

Two prediction models that quantify risks of cardiovascular hospitalization and rehospitalization have been developed and externally validated. They are based on a small number of clinical measurements that are available for patients with type 2 diabetes in many developed countries in primary care settings and could serve as the tools to screen the population at high risk of cardiovascular hospitalization and rehospitalization.

Development and validation of risk prediction models for cardiovascular mortality in Chinese people initialising peritoneal dialysis: a cohort study

Cardiovascular disease is the leading cause of death among patients receiving peritoneal dialysis. We aimed to develop and validate a risk prediction model for cardiovascular death within 2 years after the initiation of peritoneal dialysis (PD). A cohort including all patients registered with the Henan Peritoneal Dialysis Registry (HPDR) between 2007 and 2014. Multivariate logistic regression analysis was used to develop the risk prediction model. The HPDR data was randomly divided into two cohorts with 60% (1,835 patients) for model derivation, and 40% (1,219 patients) for model validation. The absolute rate of cardiovascular mortality was 14.2% and 14.4 in the derivation and validation cohort, respectively. Age, body mass index, blood pressure, serum lipids, fasting glucose, sodium, albumin, total protein, and phosphorus were the strongest predictors of cardiovascular mortality in the final model. Discrimination of the model was similar in both cohorts, with a C statistic above 0.70, with good calibration of observed and predicted risks. The new prediction model that has been developed and validated with clinical measurements that are available at the point of initiation of PD and could serve as a tool to screen for patients at high risk of cardiovascular death and tailor more intensive cardio-protective care.

Prediction models for the mortality risk in chronic dialysis patients: a systematic review and independent external validation study

OBJECTIVE

In medicine, many more prediction models have been developed than are implemented or used in clinical practice. These models cannot be recommended for clinical use before external validity is established. Though various models to predict mortality in dialysis patients have been published, very few have been validated and none are used in routine clinical practice. The aim of the current study was to identify existing models for predicting mortality in dialysis patients through a review and subsequently to externally validate these models in the same large independent patient cohort, in order to assess and compare their predictive capacities.

METHODS

A systematic review was performed following the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. To account for missing data, multiple imputation was performed. The original prediction formulae were extracted from selected studies. The probability of death per model was calculated for each individual within the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD). The predictive performance of the models was assessed based on their discrimination and calibration.

RESULTS

In total, 16 articles were included in the systematic review. External validation was performed in 1,943 dialysis patients from NECOSAD for a total of seven models. The models performed moderately to well in terms of discrimination, with C-statistics ranging from 0.710 (interquartile range 0.708-0.711) to 0.752 (interquartile range 0.750-0.753) for a time frame of 1 year. According to the calibration, most models overestimated the probability of death.

CONCLUSION

Overall, the performance of the models was poorer in the external validation than in the original population, affirming the importance of external validation. Floege et al's models showed the highest predictive performance. The present study is a step forward in the use of a prediction model as a useful tool for nephrologists, using evidence-based medicine that combines individual clinical expertise, patients' choices, and the best available external evidence.