Risk Prediction Models for Atherosclerotic Cardiovascular Disease in Patients with Chronic Kidney Disease: The CRIC Study

Association of retinal age gap with chronic kidney disease and subsequent cardiovascular disease sequelae: a cross-sectional and longitudinal study from the UK Biobank

Background

Chronic kidney disease (CKD) increases the risk of cardiovascular disease (CVD) and is more prevalent in older adults. Retinal age gap, a biomarker of aging based on fundus images, has been previously developed and validated. This study aimed to investigate the association of retinal age gap with CKD and subsequent CVD complications.

Methods

A deep learning model was trained to predict the retinal age using 19 200 fundus images of 11 052 participants without any medical history at baseline. Retinal age gap, calculated as retinal age predicted minus chronological age, was calculated for the remaining 35 906 participants. Logistic regression models and Cox proportional hazards regression models were used for the association analysis.

Results

A total of 35 906 participants (56.75 ± 8.04 years, 55.68% female) were included in this study. In the cross-sectional analysis, each 1-year increase in retinal age gap was associated with a 2% increase in the risk of CKD prevalence [odds ratio 1.02, 95% confidence interval (CI) 1.01-1.04, P = .012]. A longitudinal analysis of 35 039 participants demonstrated that 2.87% of them developed CKD in follow-up, and each 1-year increase in retinal age gap was associated with a 3% increase in the risk of CKD incidence (hazard ratio 1.03, 95% CI 1.01-1.05, P = .004). In addition, a total of 111 CKD patients (15.81%) developed CVD in follow-up, and each 1-year increase in retinal age gap was associated with a 10% increase in the risk of incident CVD (hazard ratio 1.10, 95% CI 1.03-1.17, P = .005).

Conclusions

We found that retinal age gap was independently associated with the prevalence and incidence of CKD, and also associated with CVD complications in CKD patients. This supports the use of this novel biomarker in identifying individuals at high risk of CKD and CKD patients with increased risk of CVD.

Machine learning model for cardiovascular disease prediction in patients with chronic kidney disease

Introduction

Cardiovascular disease (CVD) is the leading cause of death in patients with chronic kidney disease (CKD). This study aimed to develop CVD risk prediction models using machine learning to support clinical decision making and improve patient prognosis.

Methods

Electronic medical records from patients with CKD at a single center from 2015 to 2020 were used to develop machine learning models for the prediction of CVD. Least absolute shrinkage and selection operator (LASSO) regression was used to select important features predicting the risk of developing CVD. Seven machine learning classification algorithms were used to build models, which were evaluated by receiver operating characteristic curves, accuracy, sensitivity, specificity, and F1-score, and Shapley Additive explanations was used to interpret the model results. CVD was defined as composite cardiovascular events including coronary heart disease (coronary artery disease, myocardial infarction, angina pectoris, and coronary artery revascularization), cerebrovascular disease (hemorrhagic stroke and ischemic stroke), deaths from all causes (cardiovascular deaths, non-cardiovascular deaths, unknown cause of death), congestive heart failure, and peripheral artery disease (aortic aneurysm, aortic or other peripheral arterial revascularization). A cardiovascular event was a composite outcome of multiple cardiovascular events, as determined by reviewing medical records.

Results

This study included 8,894 patients with CKD, with a composite CVD event incidence of 25.9%; a total of 2,304 patients reached this outcome. LASSO regression identified eight important features for predicting the risk of CKD developing into CVD: age, history of hypertension, sex, antiplatelet drugs, high-density lipoprotein, sodium ions, 24-h urinary protein, and estimated glomerular filtration rate. The model developed using Extreme Gradient Boosting in the test set had an area under the curve of 0.89, outperforming the other models, indicating that it had the best CVD predictive performance.

Conclusion

This study established a CVD risk prediction model for patients with CKD, based on routine clinical diagnostic and treatment data, with good predictive accuracy. This model is expected to provide a scientific basis for the management and treatment of patients with CKD.

The Prognostic Role of Serum β-Trace Protein Levels among Patients on Maintenance Hemodialysis

Cardiovascular (CV) diseases are the most commonly encountered etiology of mortality in patients having kidney failure. β-Trace protein (BTP) is a biomarker of glomerular filtration function as well as a potential predictor of adverse CV outcomes. This study aimed to determine the prognostic value of BTP in patients on chronic hemodialysis (HD). A total of 96 patients undergoing HD were enrolled. Baseline variables were collected, and the patients were tracked for 3 years. Twenty-five patients died at 3 years. Those who experienced mortality were noted to have higher serum concentrations of BTP and a higher incidence of diabetes mellitus (DM). The area under the receiver operating characteristic curve for serum BTP distinguishing mortality from survival was 0.659 (95% confidence interval [CI], 0.555-0.752; p = 0.027). After the adjustment of variables potentially affecting survival rates, BTP levels above the median (adjusted hazard ratio [aHR]: 2.913, 95% CI, 1.256-6.754; p = 0.013), the presence of DM (aHR: 2.474, 95% CI, 1.041-5.875; p = 0.040), and low serum albumin (aHR: 0.298, 95% CI, 0.110-0.806; p = 0.017) independently correlated with survival in HD patients. Serum BTP is a novel biomarker for predicting overall outcomes in HD patients.

Trajectories of atherosclerotic cardiovascular disease risk scores as a predictor for incident chronic kidney disease

BACKGROUND

The relationship between atherosclerosis and renal function is well established. Atherosclerotic cardiovascular disease (ASCVD) risk scores reflect atherosclerotic burden, which changes over time. We investigated the association between ASCVD risk trajectories and incident chronic kidney disease (CKD) using data from a large community-based Korean cohort with up to 16 years of follow-up.

METHODS

We analyzed data from 5032 participants without CKD from the baseline survey of the Korean Genome and Epidemiology Study Ansan-Ansung cohort. Participants were categorized into stable or increasing ASCVD risk groups based on the revised ASCVD risk pooled cohort equation over a median period of exposure of 5.8 years. Incident CKD was defined as two consecutive events of an estimated glomerular filtration rate < 60 mL/min/1.73 m2.

RESULTS

During a median 9.9 years of event accrual period, 449 (8.92%) new-onset CKD cases were identified. Multiple Cox proportional regression analyses showed that the hazard ratio (95% confidence interval) for incident CKD in the increasing group, compared to the stable group, was 2.13 (1.74-2.62) in the unadjusted model and 1.35 (1.02-1.78) in the fully-adjusted model. Significant relationships were maintained in subgroups of individuals in their 50s, without diabetes mellitus or hypertension. The prevalence of proteinuria was consistently higher in the increasing group than that in the stable group.

CONCLUSIONS

An increasing trend in ASCVD risk scores independently predicted adverse renal outcomes in patients without diabetes mellitus or hypertension. Continuous monitoring of ASCVD risk is not only important for predicting cardiovascular disease but also for predicting CKD.

Associations between inflammatory and angiogenic proteomic biomarkers, and cardiovascular events and mortality in relation to kidney function

Background

The links between chronic kidney disease (CKD) and the high burden of cardiovascular disease remain unclear. We aimed to explore the association between selected inflammatory and angiogenic biomarkers, kidney function and long-term outcome in patients with an acute coronary syndrome (ACS) and to test the hypothesis that CKD status modifies this association.

Methods

A total of 1293 ACS patients hospitalized between 2008 and 2015 were followed until 31 December 2017. Plasma was collected on days 1-3 after admission. A total of 13 biomarkers were a priori identified and analysed with two proteomic methods, proximity extension assay or multiple reaction monitoring mass spectrometry. Boxplots and multiple linear regression models were used to study associations between biomarkers and kidney function and adjusted standardized Cox regression with an interaction term for CKD was used to assess whether CKD modified the association between biomarkers and major adverse cardiovascular events and death (MACE+).

Results

The concentrations of nine biomarkers-endothelial cell-specific molecule-1 (ESM-1), fibroblast growth factor 23 (FGF-23), fractalkine (CX3CL1), interleukin-1 receptor antagonist (IL-1RA), interleukin-18 (IL-18), monocyte chemotactic protein-1 (MCP-1), placenta growth factor (PlGF), transmembrane immunoglobulin 1 (TIM-1) and vascular endothelial growth factor A (VEGFA)-were inversely associated with kidney function. ESM-1, FGF-23 and TIM-1 showed associations with MACE+. Only FGF23 remained independently associated after adjustment for the other biomarkers (hazard ratio per standard deviation increase 1.34; 95% Bonferroni corrected confidence interval 1.19-1.50). None of the biomarkers showed an interaction with CKD.

Conclusions

The concentrations of 9 of the 13 prespecified inflammatory and angiogenic proteomic biomarkers increased when kidney function declined. Only FGF-23 demonstrated an independent association with MACE+, and this association was not modified by CKD status. These findings further support FGF-23 as an independent prognostic marker in ACS patients with and without CKD.

Development of machine learning-based models to predict 10-year risk of cardiovascular disease: a prospective cohort study

BACKGROUND

Previous prediction algorithms for cardiovascular diseases (CVD) were established using risk factors retrieved largely based on empirical clinical knowledge. This study sought to identify predictors among a comprehensive variable space, and then employ machine learning (ML) algorithms to develop a novel CVD risk prediction model.

METHODS

From a longitudinal population-based cohort of UK Biobank, this study included 473 611 CVD-free participants aged between 37 and 73 years old. We implemented an ML-based data-driven pipeline to identify predictors from 645 candidate variables covering a comprehensive range of health-related factors and assessed multiple ML classifiers to establish a risk prediction model on 10-year incident CVD. The model was validated through a leave-one-center-out cross-validation.

RESULTS

During a median follow-up of 12.2 years, 31 466 participants developed CVD within 10 years after baseline visits. A novel UK Biobank CVD risk prediction (UKCRP) model was established that comprised 10 predictors including age, sex, medication of cholesterol and blood pressure, cholesterol ratio (total/high-density lipoprotein), systolic blood pressure, previous angina or heart disease, number of medications taken, cystatin C, chest pain and pack-years of smoking. Our model obtained satisfied discriminative performance with an area under the receiver operating characteristic curve (AUC) of 0.762±0.010 that outperformed multiple existing clinical models, and it was well-calibrated with a Brier Score of 0.057±0.006. Further, the UKCRP can obtain comparable performance for myocardial infarction (AUC 0.774±0.011) and ischaemic stroke (AUC 0.730±0.020), but inferior performance for haemorrhagic stroke (AUC 0.644±0.026).

CONCLUSION

ML-based classification models can learn expressive representations from potential high-risked CVD participants who may benefit from earlier clinical decisions.

A deep learning system for retinal vessel calibre improves cardiovascular risk prediction in Asians with chronic kidney disease

Backgraund

Cardiovascular disease (CVD) and mortality is elevated in chronic kidney disease (CKD). Retinal vessel calibre in retinal photographs is associated with cardiovascular risk and automated measurements may aid CVD risk prediction.

Methods

Retrospective cohort study of 860 Chinese, Malay and Indian participants aged 40-80 years with CKD [estimated glomerular filtration rate (eGFR) <60 ml/min/1.73 m2] who attended the baseline visit (2004-2011) of the Singapore Epidemiology of Eye Diseases Study. Retinal vessel calibre measurements were obtained by a deep learning system (DLS). Incident CVD [non-fatal acute myocardial infarction (MI) and stroke, and death due to MI, stroke and other CVD] in those who were free of CVD at baseline was ascertained until 31 December 2019. Risk factors (established, kidney, and retinal features) were examined using Cox proportional hazards regression models. Model performance was assessed for discrimination, fit, and net reclassification improvement (NRI).

Results

Incident CVD occurred in 289 (33.6%) over mean follow-up of 9.3 (4.3) years. After adjusting for established cardiovascular risk factors, eGFR [adjusted HR 0.98 (95% CI: 0.97-0.99)] and retinal arteriolar narrowing [adjusted HR 1.40 (95% CI: 1.17-1.68)], but not venular dilation, were independent predictors for CVD in CKD. The addition of eGFR and retinal features to established cardiovascular risk factors improved model discrimination with significantly better fit and better risk prediction according to the low (<15%), intermediate (15-29.9%), and high (30% or more) risk categories (NRI 5.8%), and with higher risk thresholds (NRI 12.7%).

Conclusions

Retinal vessel calibre measurements by DLS were significantly associated with incident CVD independent of established CVD risk factors. Addition of kidney function and retinal vessel calibre parameters may improve CVD risk prediction among Asians with CKD.

A 4-Variable Model to Predict Cardio-Kidney Events and Mortality in Chronic Kidney Disease: The Chronic Renal Insufficiency Cohort (CRIC) Study

INTRODUCTION

Current prognostic models for chronic kidney disease (CKD) are complex and were designed to predict a single outcome. We aimed to develop and validate a simple and parsimonious prognostic model to predict cardio-kidney events and mortality.

METHODS

Patients from the CRIC Study (n = 3,718) were randomly divided into derivation (n = 2,478) and validation (n = 1,240) cohorts. Twenty-nine candidate variables were preselected. Multivariable Cox regression models were developed using stepwise selection for various cardio-kidney endpoints, namely, (i) the primary composite outcome of 50% decline in estimated glomerular filtration rate (eGFR) from baseline, end-stage renal disease, or cardiovascular (CV) mortality; (ii) hospitalization for heart failure (HHF) or CV mortality; (iii) 3-point major CV endpoints (3P-MACE); (iv) all-cause death.

RESULTS

During a median follow-up of 9 years, the primary outcome occurred in 977 patients of the derivation cohort and 501 patients of the validation cohort. Log-transformed N-terminal pro-B-type natriuretic peptide (NT-proBNP), log-transformed high-sensitive cardiac troponin T (hs-cTnT), log-transformed albuminuria, and eGFR were the dominant predictors. The primary outcome risk score discriminated well (c-statistic = 0.83) with a proportion of events of 11.4% in the lowest tertile of risk and 91.5% in the highest tertile at 10 years. The risk model presented good discrimination for HHF or CV mortality, 3P-MACE, and all-cause death (c-statistics = 0.80, 0.75, and 0.75, respectively). The 4-variable risk model achieved similar c-statistics for all tested outcomes in the validation cohort. The discrimination of the 4-variable risk model was mostly superior to that of published models.

CONCLUSION

The combination of NT-proBNP, hs-cTnT, albuminuria, and eGFR in a single 4-variable model provides a unique individual prognostic assessment of multiple cardio-kidney outcomes in CKD.

Heart Failure-Type Symptom Score Trajectories in CKD: Findings From the Chronic Renal Insufficiency Cohort (CRIC) Study

RATIONALE & OBJECTIVE

Quality of life in chronic kidney disease (CKD) is impaired by a large burden of symptoms including some that overlap with the symptoms of heart failure (HF). We studied a group of individuals with CKD to understand the patterns and trajectories of HF-type symptoms in this setting.

STUDY DESIGN

Prospective cohort study.

SETTING & PARTICIPANTS

3,044 participants in the Chronic Renal Insufficiency Cohort (CRIC) without prior diagnosis of HF.

PREDICTORS

Sociodemographics, medical history, medications, vital signs, laboratory values, echocardiographic and electrocardiographic parameters.

OUTCOME

Trajectory over 5.5 years of a HF-type symptom score (modified Kansas City Cardiomyopathy Questionnaire [KCCQ] Overall Summary Score with a range of 0-100 where<75 reflects clinically significant symptoms).

ANALYTICAL APPROACH

Latent class mixed models were used to model trajectories. Multinomial logistic regression was used to model relationships of predictors with trajectory group membership.

RESULTS

Five trajectories of KCCQ score were identified in the cohort of 3,044 adults, 45% of whom were female, and whose median age was 61 years. Group 1 (41.7%) had a stable high score (minimal symptoms, average score of 96); groups 2 (35.6%) and 3 (15.6%) had stable but lower scores (mild symptoms [average of 81] and clinically significant symptoms [average of 52], respectively). Group 4 (4.9%) had a substantial worsening in symptoms over time (mean 31-point decline), and group 5 (2.2%) had a substantial improvement (mean 33-point increase) in KCCQ score. A majority of group 1 was male, without diabetes or obesity, and this group had higher baseline kidney function. A majority of groups 2 and 3 had diabetes and obesity. A majority of group 4 was male and had substantial proteinuria. Group 5 had the highest proportion of baseline cardiovascular disease (CVD).

LIMITATIONS

No validation cohort available, CKD management changes in recent years may alter trajectories, and latent class models depend on the missing at random assumption.

CONCLUSIONS

Distinct HF-type symptom burden trajectories were identified in the setting of CKD, corresponding to different baseline characteristics. These results highlight the diversity of HF-type symptom experiences in individuals with CKD.

The Cardiovascular Literature-Based Risk Algorithm (CALIBRA): Predicting Cardiovascular Events in Patients With Non-Dialysis Dependent Chronic Kidney Disease

Background and Objectives

Cardiovascular (CV) disease is the main cause of morbidity and mortality in patients suffering from chronic kidney disease (CKD). Although it is widely recognized that CV risk assessment represents an essential prerequisite for clinical management, existing prognostic models appear not to be entirely adequate for CKD patients. We derived a literature-based, naïve-bayes model predicting the yearly risk of CV hospitalizations among patients suffering from CKD, referred as the CArdiovascular, LIterature-Based, Risk Algorithm (CALIBRA).

Methods

CALIBRA incorporates 31 variables including traditional and CKD-specific risk factors. It was validated in two independent CKD populations: the FMC NephroCare cohort (European Clinical Database, EuCliD®) and the German Chronic Kidney Disease (GCKD) study prospective cohort. CALIBRA performance was evaluated by c-statistics and calibration charts. In addition, CALIBRA discrimination was compared with that of three validated tools currently used for CV prediction in CKD, namely the Framingham Heart Study (FHS) risk score, the atherosclerotic cardiovascular disease risk score (ASCVD), and the Individual Data Analysis of Antihypertensive Intervention Trials (INDANA) calculator. Superiority was defined as a ΔAUC>0.05.

Results

CALIBRA showed good discrimination in both the EuCliD® medical registry (AUC 0.79, 95%CI 0.76-0.81) and the GCKD cohort (AUC 0.73, 95%CI 0.70-0.76). CALIBRA demonstrated improved accuracy compared to the benchmark models in EuCliD® (FHS: ΔAUC=-0.22, p<0.001; ASCVD: ΔAUC=-0.17, p<0.001; INDANA: ΔAUC=-0.14, p<0.001) and GCKD (FHS: ΔAUC=-0.16, p<0.001; ASCVD: ΔAUC=-0.12, p<0.001; INDANA: ΔAUC=-0.04, p<0.001) populations. Accuracy of the CALIBRA score was stable also for patients showing missing variables.

Conclusion

CALIBRA provides accurate and robust stratification of CKD patients according to CV risk and allows score calculations with improved accuracy compared to established CV risk scores also in real-world clinical cohorts with considerable missingness rates. Our results support the generalizability of CALIBRA across different CKD populations and clinical settings.