A validation study of the kidney failure risk equation in advanced chronic kidney disease according to disease aetiology with evaluation of discrimination, calibration and clinical utility

Predictive Value of Kidney Failure Risk Equation and Neutrophil Gelatinase-Associated Lipocalin for Chronic Kidney Disease Progression in Chinese Population - A Retrospective Study

Objective

To analyze the independent associations of the Kidney Failure Risk Equation (KFRE) and neutrophil gelatinase-associated lipocalin (NGAL) with end-stage renal disease (ESRD) among patients with chronic kidney disease (CKD) stages 3-5 in China and evaluate their predictive values for ESRD.

Patients and Methods

A total of 716 patients with CKD stages 3-5 at the time of the initial renal medicine referral were retrospectively enrolled, and the study outcome was the observed incidence of ESRD at 2 years after the initial referral. Baseline characteristics were collected, and relevant laboratory indexes, including neutrophil gelatinase-associated lipocalin (NGAL), were detected. The binary logistic regression model was used to analyze the independent associations, and the receiver operating characteristic (ROC) curve was used to assess the predictive values.

Results

The 2-year incidence of ESRD was 20.5% (147/716). The 4-variable KFRE, 8-variable KFRE and NGAL were independently associated with ESRD after adjusting for potential confounding factors. The AUCs of the 4-variable KFRE, 8-variable KFRE and NGAL for predicting ESRD among patients with CKD stages 3-5 were 0.711 [standard error (SE): 0.026, 95% confidence interval (CI): 0.662-0.761], 0.725 (SE: 0.025, 95% CI: 0.677-0.774) and 0.736 (SE: 0.024, 95% CI: 0.686-0.785), respectively. The AUC of the 4-variable KFRE plus NGAL was significantly higher than those of the 4-variable KFRE and NGAL alone (0.900 vs 0.711, Z = 6.297, P < 0.001; 0.900 vs 0.736, Z = 5.795, P < 0.001), and the AUC of the 8-variable KFRE plus NGAL was also significantly higher than those of the 8-variable KFRE and NGAL alone (0.911 vs 0.725, Z = 6.491, P < 0.001; 0.911 vs 0.736, Z = 6.298, P < 0.001).

Conclusion

The KFRE was able to independently predict progression of CKD stage 3-5 to ESRD in Chinese population. The addition of NGAL to the KFRE was able to elevate the predictive value when applied in predicting 2-year ESRD.

Adding biomarker change information to the kidney failure risk equation improves predictive ability for dialysis dependency in eGFR <30 ml/min/1.73 m2

Background

Although the kidney failure risk equation (KFRE), a well-known predictive model for predicting dialysis dependency, is useful, it remains unclear whether the addition of biomarker changes to the KFRE model in patients with an estimated glomerular filtration rate (eGFR) <30 ml/min/1.73 m2 will improve its predictive value.

Methods

We retrospectively identified adults with eGFR <30 ml/min/1.73 m2 without dialysis dependency, and available health checkup data for two successive years using a large Japanese claims database (DeSC, Tokyo, Japan). We dichotomized the entire population into a training set (50%) and a validation set (the other half). To assess the incremental value in the predictive ability for dialysis dependency by the addition of changes in eGFR and proteinuria, we calculated the difference in the C-statistics and net reclassification index (NRI).

Results

We identified 4499 individuals and observed 422 individuals (incidence of 45.2 per 1000 person-years) who developed dialysis dependency during the observation period (9343 person-years). Adding biomarker changes to the KFRE model improved C-statistics from 0.862 to 0.921, with an improvement of 0.060 (95% confidence intervals (CI) of 0.043-0.076, P < .001). The corresponding NRI was 0.773 (95% CI: 0.637-0.908), with an NRI for events of 0.544 (95% CI of 0.415-0.672) and NRI for non-events of 0.229 (95% CI of 0.186-0.272).

Conclusions

The KFRE model was improved by incorporating yearly changes in its components. The added information may help clinicians identify high-risk individuals and improve their care.

Representation of multimorbidity and frailty in the development and validation of kidney failure prognostic prediction models: a systematic review

BACKGROUND

Prognostic models that identify individuals with chronic kidney disease (CKD) at greatest risk of developing kidney failure help clinicians to make decisions and deliver precision medicine. It is recognised that people with CKD usually have multiple long-term health conditions (multimorbidity) and often experience frailty. We undertook a systematic review to evaluate the representation and consideration of multimorbidity and frailty within CKD cohorts used to develop and/or validate prognostic models assessing the risk of kidney failure.

METHODS

We identified studies that described derivation, validation or update of kidney failure prognostic models in MEDLINE, CINAHL Plus and the Cochrane Library-CENTRAL. The primary outcome was representation of multimorbidity or frailty. The secondary outcome was predictive accuracy of identified models in relation to presence of multimorbidity or frailty.

RESULTS

Ninety-seven studies reporting 121 different kidney failure prognostic models were identified. Two studies reported prevalence of multimorbidity and a single study reported prevalence of frailty. The rates of specific comorbidities were reported in a greater proportion of studies: 67.0% reported baseline data on diabetes, 54.6% reported hypertension and 39.2% reported cardiovascular disease. No studies included frailty in model development, and only one study considered multimorbidity as a predictor variable. No studies assessed model performance in populations in relation to multimorbidity. A single study assessed associations between frailty and the risks of kidney failure and death.

CONCLUSIONS

There is a paucity of kidney failure risk prediction models that consider the impact of multimorbidity and/or frailty, resulting in a lack of clear evidence-based practice for multimorbid or frail individuals. These knowledge gaps should be explored to help clinicians know whether these models can be used for CKD patients who experience multimorbidity and/or frailty.

SYSTEMATIC REVIEW REGISTRATION

This review has been registered on PROSPERO (CRD42022347295).

The Impact of Primary Renal Diagnosis on Prognosis and the Varying Predictive Power of Albuminuria in the NURTuRE-CKD Study

INTRODUCTION

The definition of CKD is broad, which neglects the heterogeneity of risk across primary renal diseases.

METHODS

The National Unified Renal Translational Research Enterprise (NURTuRE)-CKD is an ongoing UK, prospective multicenter cohort study of 2,996 adults with an eGFR of 15-59 mL/min/1.73 m2 or eGFR ≥60 mL/min/1.73 m2 with a urine albumin-to-creatinine ratio (uACR) >30 mg/mmol. Outcomes and predictive performance of eGFR and uACR were subcategorized by ERA-EDTA primary renal diagnosis (PRD) codes.

RESULTS

2,638 participants were included, with baseline median eGFR of 33.5 mL/min/1.73 m2 and uACR 29.8 mg/mmol. Over a median 49.2 months follow-up, 630 (23.9%) experienced kidney failure (KF), and 352 (13.3%) died before KF, the median eGFR slope was -1.97 mL/min/1.73 m2/year. There were significant differences in risk across the PRD, persisting after adjustment for age, sex, baseline eGFR, and modifiable risk factors (blood pressure, HbA1c, and renin-angiotensin-aldosterone system inhibitors). Diabetic kidney disease (DKD), glomerulonephritis, and familial/hereditary nephropathy were associated with the greatest risk, while tubulointerstitial disease and vasculitis carried a low risk of KF. eGFR had good predictive accuracy across all PRD. However, the addition of uACR showed variable benefit, depending on the PRD. The largest benefit was seen in vasculitis, renal vascular, and DKD groups, but uACR added no predictive value to the familial/hereditary group.

CONCLUSION

Significant differences in the risk of kidney-related outcomes occurred across the various primary renal diagnoses persisting after adjustment for age, sex, baseline eGFR, and modifiable risk factors. Albuminuria's discriminatory ability as a biomarker of progression varies by diagnosis. CKD care should, therefore, take a personalized approach that always considers the primary renal diagnosis.

Investigation of end-stage kidney disease risk prediction in an ethnically diverse cohort of people with type 2 diabetes: use of kidney failure risk equation

INTRODUCTION

The four variable kidney failure (KF) risk equation (KFRE) is recommended to estimate KF risk (ie, need for dialysis or kidney transplantation). Earlier referral to clinical kidney services for people with high-risk of kidney failure can ensure appropriate care, education and support are in place pre-emptively. There are limited data on investigating the performance of KFRE in estimating risk of end-stage kidney disease (ESKD) in people with type 2 diabetes mellitus (T2DM) and chronic kidney disease (CKD). The primary ESKD endpoint event was defined as estimated glomerular filtration rate (eGFR) <10 mL/min/1.73 m2 and secondary endpoint eGFR <15 mL/min/1.73 m2.

RESEARCH DESIGN AND METHODS

We studied 7296 people (30% women, 41% African-Caribbean, 45% Caucasian) with T2DM and CKD (eGFR median (range) 48 (15-59) mL/min/1.73 m2) were included at two hospitals in London (median follow-up 10.2 years). Time to ESKD event was the endpoint and Concordance index (C-index) was used to assess KFRE's discrimination of those experiencing ESKD from those who did not. Mean (integrated calibration index (ICI)) and 90th percentile (E90) of the difference between observed and predicted risks were used as calibration metrics.

RESULTS

Of the cohort 746 (10.2%) reached ESKD primary event (135 (1.9%) and 339 (4.5%) over 2 and 5 years, respectively). Similarly, 1130 (15.5%) reached the secondary endpoint (270 (3.7%) and 547 (7.5%) over 2 and 5 years, respectively). The C-index for the primary endpoint was 0.842 (95% CI 0.836 to 0.848) and 0.816 (95% CI 0.812 to 0.820) for 2 and 5 years, respectively. KFRE 'under-predicted' ESKD risk overall and by ethnic group. Likewise, the C-index for secondary endpoint was 0.843 (0.839-0.847) and 0.801 (0.798-0.804) for 2 and 5 years, respectively. KFRE performance analysis performed more optimally with the primary endpoint with 50% enhancement of the calibration metrics than with the secondary endpoint. KFRE recalibration improved ICI by 50% and E90 by more than 78%.

CONCLUSIONS

Although derived for predicting KF, KFRE also demonstrated good discrimination for ESKD outcome. Further studies are needed to identify variables/biomarkers that may improve KFRE's performance/calibration and to aid the development of other predictive models to enable early identification of people at risk of advanced stages of CKD prior to onset of KF.

Estimates of eskd risk and timely kidney replacement therapy education

BACKGROUND

Kidney replacement therapy (KRT) needs preparation and its timing is difficult to predict. Nephrologists' predictions of kidney failure risk tend to be more pessimistic than the Kidney Failure Risk Equation (KFRE) predictions. We aimed to explore how physicians' risk estimate related to referral to KRT education, vs. the objective calculated KFRE.

METHODS

Prospective observational study of data collected in chronic kidney disease (CKD) clinics of the Veterans Affairs Medical Center San Diego and the University of California, San Diego. The study included 257 participants who were aged 18 years or older, English speaking, prevalent CKD clinic patients, with estimated glomerular filtration rate (eGFR) < 60 mL/min per 1.73 m2 (MDRD equation). The exposure consisted of end stage kidney disease (ESKD) risk predictions. Nephrologists' kidney failure risk estimations were assessed: "On a scale of 0-100%, without using any estimating equations, give your best estimate of the risk that this patient will need dialysis or a kidney transplant in 2 years." KFRE was calculated using age, sex, eGFR, serum bicarbonate, albumin, calcium, phosphorus, urine albumin/creatinine ratio. The outcomes were the pattern of referral to KRT education (within 90 days of initial visit) and kidney failure evaluated by chart review. The population was divided into groups either by nephrologists' predictions or by KFRE. Referral to KRT education was examined by group and sensitivity and specificity were calculated based on whether participants reached kidney failure at 2 years.

RESULTS

A fifth were referred for education by 90 days of enrollment. Low risk patients by both estimates had low referral rates. In those with nephrologists' predictions ≥ 15% (n = 137), sensitivity was 71% and specificity 76%. In those with KFRE ≥ 15% (n = 55), sensitivity was 85% and specificity 41%.

CONCLUSIONS

Although nephrologists tend to overestimate patients' kidney failure risk, they do not appear to act on this overestimation, as the rates of KRT education referrals are lower than expected when a nephrologist identifies a patient as high risk.

CLINICAL TRIAL NUMBER

Not applicable.

Use of kidney failure risk equation as a tool to evaluate referrals from primary care to specialist nephrology care

BACKGROUND

With rising costs and burden of chronic kidney disease (CKD), timely referral of patients to a kidney specialist is crucial. Currently, Kidney Health Australia (KHA) uses a 'heat map' based on severity and not future risk of kidney failure, whereas the kidney failure risk equation (KFRE) score predicts future risk of progression.

AIMS

Evaluate whether a KFRE score assists with timing of CKD referrals.

METHODS

Retrospective cohort of 2137 adult patients, referred to tertiary hospital outpatient nephrologist between 2012 and 2020, were analysed. Referrals were analysed for concordance with the KHA referral guidelines and, with the KFRE score, a recommended practice.

RESULTS

Of 2137 patients, 626 (29%) did not have urine albumin-to-creatinine ratio (UACR) measurement at referral. For those who had a UACR, the number who met KFRE preferred referral criteria was 36% less than KHA criteria. If the recommended KFRE score was used, then fewer older patients (≥40 years) needed referral. Positively, many diabetes patients were referred, even if their risk of kidney failure was low, and 29% had a KFRE over 3%. For patients evaluated meeting KFRE criteria, a larger proportion (76%) remained in follow-up, with only 8% being discharged.

CONCLUSIONS

KFRE could reduce referrals and be a useful tool to assist timely referrals. Using KFRE for triage may allow those patients with very low risk of future kidney failure not be referred, remaining longer in primary care, saving health resources and reducing patients' stress and wait times. Using KFRE encourages albuminuria measurement.

Evaluation of a proteomic signature coupled with the kidney failure risk equation in predicting end stage kidney disease in a chronic kidney disease cohort

BACKGROUND

The early identification of patients at high-risk for end-stage renal disease (ESRD) is essential for providing optimal care and implementing targeted prevention strategies. While the Kidney Failure Risk Equation (KFRE) offers a more accurate prediction of ESRD risk compared to static eGFR-based thresholds, it does not provide insights into the patient-specific biological mechanisms that drive ESRD. This study focused on evaluating the effectiveness of KFRE in a UK-based advanced chronic kidney disease (CKD) cohort and investigating whether the integration of a proteomic signature could enhance 5-year ESRD prediction.

METHODS

Using the Salford Kidney Study biobank, a UK-based prospective cohort of over 3000 non-dialysis CKD patients, 433 patients met our inclusion criteria: a minimum of four eGFR measurements over a two-year period and a linear eGFR trajectory. Plasma samples were obtained and analysed for novel proteomic signals using SWATH-Mass-Spectrometry. The 4-variable UK-calibrated KFRE was calculated for each patient based on their baseline clinical characteristics. Boruta machine learning algorithm was used for the selection of proteins most contributing to differentiation between patient groups. Logistic regression was employed for estimation of ESRD prediction by (1) proteomic features; (2) KFRE; and (3) proteomic features alongside KFRE.

RESULTS

SWATH maps with 943 quantified proteins were generated and investigated in tandem with available clinical data to identify potential progression biomarkers. We identified a set of proteins (SPTA1, MYL6 and C6) that, when used alongside the 4-variable UK-KFRE, improved the prediction of 5-year risk of ESRD (AUC = 0.75 vs AUC = 0.70). Functional enrichment analysis revealed Rho GTPases and regulation of the actin cytoskeleton pathways to be statistically significant, inferring their role in kidney function and the pathogenesis of renal disease.

CONCLUSIONS

Proteins SPTA1, MYL6 and C6, when used alongside the 4-variable UK-KFRE achieve an improved performance when predicting a 5-year risk of ESRD. Specific pathways implicated in the pathogenesis of podocyte dysfunction were also identified, which could serve as potential therapeutic targets. The findings of our study carry implications for comprehending the involvement of the Rho family GTPases in the pathophysiology of kidney disease, advancing our understanding of the proteomic factors influencing susceptibility to renal damage.

Multivariate canonical correlation analysis identifies additional genetic variants for chronic kidney disease

Chronic kidney diseases (CKD) have genetic associations with kidney function. Univariate genome-wide association studies (GWAS) have identified single nucleotide polymorphisms (SNPs) associated with estimated glomerular filtration rate (eGFR) and blood urea nitrogen (BUN), two complementary kidney function markers. However, it is unknown whether additional SNPs for kidney function can be identified by multivariate statistical analysis. To address this, we applied canonical correlation analysis (CCA), a multivariate method, to two individual-level CKD genotype datasets, and metaCCA to two published GWAS summary statistics datasets. We identified SNPs previously associated with kidney function by published univariate GWASs with high replication rates, validating the metaCCA method. We then extended discovery and identified previously unreported lead SNPs for both kidney function markers, jointly. These showed expression quantitative trait loci (eQTL) colocalisation with genes having significant differential expression between CKD and healthy individuals. Several of these identified lead missense SNPs were predicted to have a functional impact, including in SLC14A2. We also identified previously unreported lead SNPs that showed significant correlation with both kidney function markers, jointly, in the European ancestry CKDGen, National Unified Renal Translational Research Enterprise (NURTuRE)-CKD and Salford Kidney Study (SKS) datasets. Of these, rs3094060 colocalised with FLOT1 gene expression and was significantly more common in CKD cases in both NURTURE-CKD and SKS, than in the general population. Overall, by using multivariate analysis by CCA, we identified additional SNPs and genes for both kidney function and CKD, that can be prioritised for further CKD analyses.

Nephrology Nurses: Essential Professionals in Sustainable Kidney Care

Purpose

The increasing frequency of extreme climate events underscores the need for urgent action on climate change. The health care system contributes 4.6% of greenhouse gas emissions (GHGs) in Canada; thus, it is a major contributor to the country's carbon footprint. Kidney care in particular can involve high amounts of waste (eg, plastic and consumable waste associated with dialysis, transportation, emissions, energy, and water consumption). Therefore, sustainability initiatives within the health care system, and especially in the context of kidney care, have great potential to make a positive impact on planetary health. Here, we outline ways in which nephrology nurses can expand our duty of care to the environment and incorporate sustainability into our work.

Sources of information

A small advisory group of nephrology nurses in partnership with the Canadian Association of Nurses for the Environment (CANE) assessed ways that sustainable practices can be incorporated into nephrology nursing. Drawing on the Planetary Health Care model used by the Canadian Society of Nephrology: Sustainable Nephrology Action Planning (SNAP) committee, we assessed how the model could be adapted in the context of kidney care using 3 main actionable themes in their work: reducing the demand for health services, matching the supply of health services with demand, and reducing emissions from the supply of health services. We also reviewed and selected real-world examples of initiatives pursued by colleagues.

Key findings

Through this established framework, we provide recommendations and case examples for nephrology nurses to expand our duty of care to the environment. We describe nursing-led strategies used in Canada to improve environmental sustainability in kidney programs and consider their applicability to other renal programs. In 1 case example, we show how a simple nurse-led initiative at a single dialysis clinic can lower plastic waste and associated costs by $2042.59 per year. More broadly, we provide recommendations and actions for nephrology nurses to improve environmental sustainability in kidney care.

Limitations

Nurses in Canada have many responsibilities within limited timeframes, making it essential to choose sustainable practices that do not exacerbate burnout and high workloads. For sustainable practices to be successful, nurses must integrate them into their existing workflows.

Risk Prediction and Management of Chronic Kidney Disease in People Living with Type 2 Diabetes Mellitus

People with type 2 diabetes mellitus have increased risk of chronic kidney disease and atherosclerotic cardiovascular disease. Improved care delivery and implementation of guideline-directed medical therapy have contributed to the declining incidence of atherosclerotic cardiovascular disease in high-income countries. By contrast, the global incidence of chronic kidney disease and associated mortality is either plateaued or increased, leading to escalating direct and indirect medical costs. Given limited resources, better risk stratification approaches to identify people at risk of rapid progression to end-stage kidney disease can reduce therapeutic inertia, facilitate timely interventions and identify the need for early nephrologist referral. Among people with chronic kidney disease G3a and beyond, the kidney failure risk equations (KFRE) have been externally validated and outperformed other risk prediction models. The KFRE can also guide the timing of preparation for kidney replacement therapy with improved healthcare resources planning and may prevent multiple complications and premature mortality among people with chronic kidney disease with and without type 2 diabetes mellitus. The present review summarizes the evidence of KFRE to date and call for future research to validate and evaluate its impact on cardiovascular and mortality outcomes, as well as healthcare resource utilization in multiethnic populations and different healthcare settings.

Strategies to Guide Preemptive Waitlisting and Equity in Waittime Accrual by Race/Ethnicity

BACKGROUND

Use of eGFR to determine preemptive waitlisting eligibility may contribute to racial/ethnic disparities in access to waitlisting, which can only occur when the eGFR falls to ≤20 ml/min per 1.73 m 2 . Use of an alternative risk-based strategy for waitlisting may reduce these inequities ( e.g. , a kidney failure risk equation [KFRE] estimated 2-year risk of kidney failure) rather than the standard eGFR threshold for determining waitlist eligibility. Our objective was to model the amount of preemptive waittime that could be accrued by race and ethnicity, applying two different strategies to determine waitlist eligibility.

METHODS

Using electronic health record data, linear mixed models were used to compare racial/ethnic differences in preemptive waittime that could be accrued using two strategies: estimating the time between an eGFR ≤20 and 5 ml/min per 1.73 m 2 versus time between a 25% 2-year predicted risk of kidney failure (using the KFRE, which incorporates age, sex, albuminuria, and eGFR to provide kidney failure risk estimation) and eGFR of 5 ml/min per 1.73 m 2 .

RESULTS

Among 1290 adults with CKD stages 4-5, using the Chronic Kidney Disease Epidemiology Collaboration equation yielded shorter preemptive waittime between an eGFR of 20 and 5 ml/min per 1.73 m 2 in Black (-6.8 months; 95% confidence interval [CI], -11.7 to -1.9), Hispanic (-10.2 months; -15.3 to -5.1), and Asian/Pacific Islander (-10.3 months; 95% CI, -15.3 to -5.4) patients compared with non-Hispanic White patients. Use of a KFRE threshold to determine waittime yielded smaller differences by race and ethnicity than observed when using a single eGFR threshold, with shorter time still noted for Black (-2.5 months; 95% CI, -7.8 to 2.7), Hispanic (-4.8 months; 95% CI, -10.3 to 0.6), and Asian/Pacific Islander (-5.4 months; -10.7 to -0.1) individuals compared with non-Hispanic White individuals, but findings only met statistical significance criteria in Asian/Pacific Islander individuals. When we compared potential waittime availability using a KFRE versus eGFR threshold, use of the KFRE yielded more equity in waittime for Black ( P = 0.02), Hispanic ( P = 0.002), and Asian/Pacific Islander ( P = 0.002) patients.

CONCLUSIONS

Use of a risk-based strategy was associated with greater racial equity in waittime accrual compared with use of a standard single eGFR threshold to determine eligibility for preemptive waitlisting.

Clinical Decision Support Tools in the Electronic Medical Record

The integration of clinical decision support (CDS) tools into electronic medical record (EMR) systems has become common. Although there are many benefits for both patients and providers from successful integration, barriers exist that prevent consistent and effective use of these tools. Such barriers include tool alert fatigue, lack of interoperability between tools and medical record systems, and poor acceptance of tools by care providers. However, successful integration of CDS tools into EMR systems have been reported; examples of these include the Statin Choice Decision Aid, and the Kidney Failure Risk Equation (KFRE). This article reviews the history of EMR systems and its integration with CDS tools, the barriers preventing successful integration, and the benefits reported from successful integration. This article also provides suggestions and strategies for improving successful integration, making these tools easier to use and more effective for care providers.

Implementation of the Kidney Failure Risk Equation in a United States Nephrology Clinic

Introduction

The kidney failure risk equation (KFRE) estimates a person's risk of kidney failure and has great potential utility in clinical care.

Methods

We used mixed methods to explore implementation of the KFRE in nephrology clinics.

Results

KFRE scores were integrated into the electronic health record at Johns Hopkins Medicine and were displayed to nephrology providers. Documentation of KFRE scores increased over time, reaching 25% of eligible outpatient nephrology clinic notes at month 11. Three providers documented KFRE scores in >75% of notes, whereas 25 documented scores in <10% of notes. Surveys and focus groups of nephrology providers were conducted to probe provider views on the KFRE. Survey respondents (n = 25) reported variability in use of KFRE for decisions such as maintaining nephrology care, referring for transplant evaluation, or providing dialysis modality education. Provider perspectives on the use of KFRE, assessed in 2 focus groups of 4 providers each, included 3 common themes as follows: (i) KFRE scores may be most impactful in the care of specific subsets of people with chronic kidney disease (CKD); (ii) there is uncertainty about KFRE risk-based thresholds to guide clinical care; and (iii) education of patients, nephrology providers, and non-nephrology providers on appropriate interpretations of KFRE scores may help maximize their utility.

Conclusion

Implementation of the KFRE was limited by non-uniform provider adoption of its use, and limited knowledge about utilization of the KFRE in clinical decisions.

CKD Progression Prediction in a Diverse US Population: A Machine-Learning Model

Rationale & Objective

Chronic kidney disease (CKD) is a major cause of morbidity and mortality. To date, there are no widely used machine-learning models that can predict progressive CKD across the entire disease spectrum, including the earliest stages. The objective of this study was to use readily available demographic and laboratory data from Sonic Healthcare USA laboratories to train and test the performance of machine learning-based predictive risk models for CKD progression.

Study Design

Retrospective observational study.

Setting & Participants

The study population was composed of deidentified laboratory information services data procured from a large US outpatient laboratory network. The retrospective data set included 110,264 adult patients over a 5-year period with initial estimated glomerular filtration rate (eGFR) values between 15-89 mL/min/1.73 m2.

Predictors

Patient demographic and laboratory characteristics.

Outcomes

Accelerated (ie, >30%) eGFR decline associated with CKD progression within 5 years.

Analytical Approach

Machine-learning models were developed using random forest survival methods, with laboratory-based risk factors analyzed as potential predictors of significant eGFR decline.

Results

The 7-variable risk classifier model accurately predicted an eGFR decline of >30% within 5 years and achieved an area under the curve receiver-operator characteristic of 0.85. The most important predictor of progressive decline in kidney function was the eGFR slope. Other key contributors to the model included initial eGFR, urine albumin-creatinine ratio, serum albumin (initial and slope), age, and sex.

Limitations

The cohort study did not evaluate the role of clinical variables (eg, blood pressure) on the performance of the model.

Conclusions

Our progressive CKD classifier accurately predicts significant eGFR decline in patients with early, mid, and advanced disease using readily obtainable laboratory data. Although prospective studies are warranted, our results support the clinical utility of the model to improve timely recognition and optimal management for patients at risk for CKD progression.

Plain-Language Summary

Defined by a significant decrease in estimated glomerular filtration rate (eGFR), chronic kidney disease (CKD) progression is strongly associated with kidney failure. However, to date, there are no broadly used resources that can predict this clinically significant event. Using machine-learning techniques on a diverse US population, this cohort study aimed to address this deficiency and found that a 5-year risk prediction model for CKD progression was accurate. The most important predictor of progressive decline in kidney function was the eGFR slope, followed by the urine albumin-creatinine ratio and serum albumin slope. Although further study is warranted, the results showed that a machine-learning model using readily obtainable laboratory information accurately predicts CKD progression, which may inform clinical diagnosis and management for this at-risk population.

Application of a Machine Learning Algorithm to Develop and Validate a Prediction Model for Ambulatory Non-Arrivals

BACKGROUND

Non-arrivals to scheduled ambulatory visits are common and lead to a discontinuity of care, poor health outcomes, and increased subsequent healthcare utilization. Reducing non-arrivals is important given their association with poorer health outcomes and cost to health systems.

OBJECTIVE

To develop and validate a prediction model for ambulatory non-arrivals.

DESIGN

Retrospective cohort study.

PATIENTS OR SUBJECTS

Patients at an integrated health system who had an outpatient visit scheduled from January 1, 2020, to February 28, 2022.

MAIN MEASURES

Non-arrivals to scheduled appointments.

KEY RESULTS

There were over 4.3 million ambulatory appointments from 1.2 million adult patients. Patients with appointment non-arrivals were more likely to be single, racial/ethnic minorities, and not having an established primary care provider compared to those who arrived at their appointments. A prediction model using the XGBoost machine learning algorithm had the highest AUC value (0.768 [0.767-0.770]). Using SHAP values, the most impactful features in the model include rescheduled appointments, lead time (number of days from scheduled to appointment date), appointment provider, number of days since last appointment with the same department, and a patient's prior appointment status within the same department. Scheduling visits close to an appointment date is predicted to be less likely to result in a non-arrival. Overall, the prediction model calibrated well for each department, especially over the operationally relevant probability range of 0 to 40%. Departments with fewer observations and lower non-arrival rates generally had a worse calibration.

CONCLUSIONS

Using a machine learning algorithm, we developed a prediction model for non-arrivals to scheduled ambulatory appointments usable for all medical specialties. The proposed prediction model can be deployed within an electronic health system or integrated into other dashboards to reduce non-arrivals. Future work will focus on the implementation and application of the model to reduce non-arrivals.

CHA2DS2VASc score predicts risk of end stage renal disease in patients with atrial fibrillation: Long-term follow-up study

Background

End stage renal disease (ESRD) is an increasing worldwide epidemic disease. CHA2DS2-VASc score is a well-established predictor of cardiovascular outcome among atrial fibrillation (AF) patients.

Objective

The aim of this study was to test whether CHA2DS2-VASc score is a good predictor for incident ESRD events.

Methods

This is a retrospective cohort study (from January 2010 to December 2020) with median follow-up of 61.7 months. Clinical parameters and baseline characteristics were recorded. The endpoint was defined as ESRD with dialysis dependent.

Results

The study cohort comprised 29,341 participants. Their median age was 71.0 years, 43.2% were male, 21.5% had diabetes mellitus, 46.1% had hypertension, and mean CHA2DS2-VASc score was 2.89. CHA2DS2-VASc score was incrementally associated with the risk of ESRD status during follow-up. Using the univariate Cox model, we found a 26% increase in ESRD risk with an increase of one point in the CHA2DS2-VASc score (HR 1.26 [1.23-1.29], P < 0.001). And using the multi-variate Cox model adjusted by initial CKD stage, we still observed a 5.9% increase in risk of ESRD with a one-point increase in the CHA2DS2-VASc score (HR 1.059 [1.037-1.082], P < 0.001). The CHA2DS2-VASC score and the initial stage of CKD were associated with the risk of ESRD development in patients with AF.

Conclusions

Our results first confirmed the utility of CHA2DS2-VASC score in predicting progression to ESRD in AF patients. The efficiency is best in CKD stage 1.

Contemporary risk prediction models in chronic kidney disease: when less is more

PURPOSE OF REVIEW

Clinicians have an ever-increasing number of prediction tools at their disposal for estimating the risk of kidney failure in their patients. This review aims to summarize contemporary evidence for chronic kidney disease (CKD) risk prediction models across the spectrum of kidney function, and explore nuances in the interpretation of risk estimates.

RECENT FINDINGS

A European study using predominantly laboratory data has extended kidney failure prediction to patients with more preserved estimated glomerular filtration rate. For older patients with advanced CKD, prediction tools that censor for death (such as the Kidney Failure Risk Equation) overestimate the risk of kidney failure, especially over time horizons longer than 2 years. This problem can be addressed by accounting for the competing risk of death, as shown in well designed validation studies. The clinical utility of kidney failure risk prediction tools is being increasingly tested at a population level to inform policy and referral guidelines.

SUMMARY

There is welcome trend to validate existing prediction tools in diverse clinical settings and identify their role in clinical practice. Clinicians should be cognizant of overestimating kidney failure risk in older patients with advanced CKD due to the competing risk of death. For moderate CKD and for short-term predictions, the Kidney Failure Risk Equation remains the most widely validated prediction tool.

Machine Learning Improves Upon Clinicians' Prediction of End Stage Kidney Disease

Background and Objectives

Chronic kidney disease progression to ESKD is associated with a marked increase in mortality and morbidity. Its progression is highly variable and difficult to predict.

Methods

This is an observational, retrospective, single-centre study. The cohort was patients attending hospital and nephrology clinic at The Canberra Hospital from September 1996 to March 2018. Demographic data, vital signs, kidney function test, proteinuria, and serum glucose were extracted. The model was trained on the featurised time series data with XGBoost. Its performance was compared against six nephrologists and the Kidney Failure Risk Equation (KFRE).

Results

A total of 12,371 patients were included, with 2,388 were found to have an adequate density (three eGFR data points in the first 2 years) for subsequent analysis. Patients were divided into 80%/20% ratio for training and testing datasets.

ML model had superior performance than nephrologist in predicting ESKD within 2 years with 93.9% accuracy, 60% sensitivity, 97.7% specificity, 75% positive predictive value. The ML model was superior in all performance metrics to the KFRE 4- and 8-variable models.

eGFR and glucose were found to be highly contributing to the ESKD prediction performance.

Conclusions

The computational predictions had higher accuracy, specificity and positive predictive value, which indicates the potential integration into clinical workflows for decision support.