Comparation of two cystatin C-based eGFR equations in assessing risk of all-cause mortality and incident cardiovascular disease Authors: Guangda Xin, Qianyu Li, Chen Sheng, Yining Zha, Kailiang Cheng Published: 2024 ABSTRACT BackgroundChronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 and European Kidney Function Consortium (EKFC) 2023 both recently updated the equations to estimate the glomerular filtration rate (eGFR) using cystatin C; however, little is known about the benefits of using the equations for the risk stratification of health outcomes. We conducted this longitudinal study to compare the cystatin C CKD-EPI and EKFC equations to track the risks of cardiovascular disease and all-cause mortality among Chinese adults.MethodsWe used data from China Health and Retirement Longitudinal Study (CHARLS) between 2011 and 2018. Adjusted logistic regression models and restricted cubic spline functions were used to evaluate the relationships of cystatin C-based eGFR values with incidence of cardiovascular disease and mortality.ResultsA total of 6 496 participants were finally included in this study. The mean age of the participants was 59.6 (± 9.5) years, including 2996 (46.1%) males. There were 473 deaths and 1996 cases of cardiovascular disease observed during a maximum follow-up of 7.0 years. Using cystatin C-based CKD-EPI equation, people of eGFR < 60 mL/min/1.73 m2 had an increased risk of mortality (risk ratio [RR], 1.527; 95% CI, 1.068–2.178) and incident cardiovascular disease (RR, 1.363; 95% CI, 1.006–1.844), compared to those of eGFR ≥ 90 mL/min/1.73 m2. On the contrary, we did not observe significant associations of eGFR levels by EKFC equation with mortality nor cardiovascular disease.ConclusionsThe findings indicated that cystatin C-based eGFR using CKD-EPI equation is more closely associated with all-cause mortality and cardiovascular disease compared to EKFC equation among Chinese adults. The cystatin C-based eGFR by CKD-EPI equation should be monitored in health practice, which needs further validation in other populations.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12986-024-00870-4. Introduction Estimated glomerular filtration rate (eGFR) is crucial for diagnosing and staging chronic kidney disease, and a lower eGFR is also an independent predictor of mortality [1]. Routinely, creatinine is widely used as the surrogate marker of renal dysfunction, while cystatin C has emerged as a potential biomarker and is increasingly adopted to evaluate eGFR levels [2–4]. Some studies showed that cystatin C-based eGFR has a closer association with kidney function, vascular disease and mortality compared to creatinine [5–11]. Moreover, two studies conducted in different cohorts reported that cystatin C-based eGFR is more accurate than that estimated using cystatin C and creatinine combined Eqs. [12, 13], highlighting the independent usefulness of cystatin C to estimate eGFR on the risk stratification of mortality in routine practice. In 2021, Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) updated the cystatin C-based eGFR equation from the previous version without race [14, 15]. Previous equations of eGFR use serum creatinine or cystatin C and incorporate demographic characteristics (age, sex, and race) to estimate renal function. However, race in eGFR equations is not a biologic component, which is neglected in the newly developed equation presenting a higher accuracy. In early 2023, the European Kidney Function Consortium (EKFC) updated a cystatin C-based eGFR equation with a scaling factor constant to race and sex [16]. This study declared that this equation improved the accuracy of renal function evaluation over other commonly used equations developed in Europe and the United States. However, there is limited evidence supporting which equation which equation is most practical for predicting the adverse health outcomes among the Asian population. To clarify this issue, this study aimed to investigate the associations of cystatin C-based eGFR derived by two most updated equations, i.e., 2021 CKD-EPI and 2023 EKFC equations, with the risk of all-cause mortality and cardiovascular disease using a large nation-representative cohort. Materials and methods Study design and population China Health and Retirement Longitudinal Study (CHARLS) cohort was designed to collect high-quality data among individuals aged 45 and above in China, aiming to promote interdisciplinary research on population aging. This national population-based cohort study (http://charls.pku.edu.cn/) have conducted four regular biannual surveys during 2011–2012 (as baseline), 2013–2014, 2015–2016, and 2017–2018. The participants were recruited using a multistage stratified probability proportional-to-size sampling strategy, covering about 10 000 households, 450 villages, 150 counties or districts of 28 provinces both in rural and urban areas. Separate weights for households, individuals, and biomarkers were developed to correct for nonresponse and sampling-frame errors. Details of the study design and cohort profile have been described in a previous study [17]. CHARLS projected used computer assisted personal interviewing (CAPI) technology to conduct the face-to-face interview and data collection including demographic information and health related issues. Physical measurements and blood sample collection were performed by interviewers at interviewees’ household in the CHARLS cohort. This current study was a secondary analysis of CHARLS project. We excluded individuals due to (1) no available data on cystatin C or out of the detectable range, (2) age < 45 years, and (3) having been diagnosed with coronary heart disease, stroke, or cancer at baseline survey. The detailed flowchart of the inclusion and exclusion process is shown in Fig. 1. Fig. 1Flowchart of this current study. Abbreviation CHD, coronary heart disease. Serum cystatin C detectable range: ~2.5 mg/L Flowchart of this current study. Abbreviation CHD, coronary heart disease. Serum cystatin C detectable range: ~2.5 mg/L The CHARLS national cohort was ethnically approved by the Biomedical Ethics Committee of Peking University. This study was conducted and reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline. Measurements of covariates Demographic characteristics (age, sex, residence, education level and marriage status), lifestyle behaviors and history of diseases were collected via standard and uniformed questionnaires. Smoking status was divided into never, former, and current. Residence type was grouped into rural and urban. Education degree was divided into primary, secondary, and tertiary. Marriage status was dichotomously grouped into married, and others (including never married, divorced, and widowed). Anthropometric measurements were conducted for body weight and height. Body mass index (BMI) was calculated, and overweight was defined using the cutoff of BMI ≥ 24.0 kg/m2 and obesity was defined using the cutoff of BMI ≥ 28.0 kg/m2 for the Chinese population. Systolic blood pressure and diastolic blood pressure were measured three times and then calculated as the average of measurements. Hypertension was defined as systolic pressure ≥ 140 mmHg, diastolic pressure ≥ 90 mmHg or using any antihypertensive medication or a self-reported history of hypertension diagnosis by physicians [18]. Diabetes was defined as fasting glucose ≥ 7.0 mmol/L, self-reported diagnosis history of diabetes or using any glucose-lowering medication according to the American Diabetes Association [19]. Venous blood samples were collected and subsequently tested. Serum cystatin C (mg/L) was measured using a particle-enhanced turbimetric assay and the coefficient of variation (CV) was less than 5% for the measure platform. Serum creatinine (µmol/L) was measured using rate-blanked and compensated Jaffe method. High-sensitivity C-reactive protein (hs-CRP) was measured using an immunoturbidimetric assay. Assays of cystatin C and hs-CRP were conducted on a Hitachi chemistry analyzer (Hitachi, Tokyo, Japan). An enzymatic colorimetric test was used for lipids and glucose measurements. Non high-density lipoprotein (non-HDL) cholesterol was calculated using the standard lipid profile. Exposure and outcome Our independent variable of interest was eGFR (mL/min/1.73 m2) corresponding to 2021 CKD-EPI [15] and 2023 EKFC [16] equations. We classified participants into three groups using eGFR levels ≥ 90, 60–89.9 and < 60 mL/min/1.73 m2. The death information was collected from death registration centres, which was then certificated by asking their family members or community managers in three follow-up surveys (2013, 2015, and 2018). Diagnosis history of cardiovascular disease including heart disease and stroke was self-reported and collected at each follow-up survey. In accordance with previous studies [20, 21], cardiovascular disease events were confirmed by the following questions “Have you been told by a doctor that you have been diagnosed with a heart attack, angina, coronary heart disease, heart failure, or other heart problems?” or “Have you been told by a doctor that you have been diagnosed with a stroke?”. Participants who responded affirmatively to either of the two questions were defined as having cardiovascular disease. The outcomes were assessed by trained interviewers that are harmonized to international leading aging surveys in the Health and Retirement Study (HRS) [22]. Statistical analysis The baseline characteristics of continuous variables were described using the mean (standard deviation). Frequencies (proportions) were used for categorical variables. The associations of cystatin C-based eGFR groups with mortality and cardiovascular disease were analyzed using binomial logistic regression models. To control the potential confounding factors, two models were step-by-step constructed: model 1 was adjusted for age (continuous) and sex; model 2 was further adjusted for residence (rural, urban), education degree (primary, secondary, third), marriage status (married, others), smoking (current, former, never), current alcohol use (yes, no), BMI group (normal, overweight, obesity), hypertension (yes, no), diabetes(yes, no), systolic blood pressure (continuous), triglyceride (continuous), nonHDL cholesterol (continuous), glucose (continuous). For sensitivity analyses. we adjusted for hs-CRP level (continuous), or serum creatinine level (continuous) in addition to covariates included in model 2 to test the robust effects. A two-sided P value of 0.05 was statistically significant in the analysis. All statistical analyses were performed based on R packages and software (version 4.1.0). Results Baseline characteristics The baseline characteristics of the enrolled participants are shown in Table 1. Out of the 6496 participating, the mean age was 59.6 (SD 9.5) years, and 3500 were females (53.9%). During a follow-up up to 7.0 years, 473 cases of death were observed. Table S1 and Table S2 show the characteristics according to eGFR categories using CKD-EPI and EKFC equations. People with eGFR < 60 mL/min/1.73 m2 were more likely to be older, males, achieving lower education level. The proportions of participants of poor eGFR (defined as < 60 mL/min/1.73 m2) were 16.6% and 23.7% when applying CKD-EPI and EKFC equations. When compared to CKD-EPI 2021 equation, EKFC equation tends to underestimate the eGFR level, yielding a worse renal function status, especially for those with normal eGFR under CKD-EPI equation (Fig. 2). Table 1Characteristics of 6 496 participantsCharacteristicsOverallParticipants, No.6 496Death, n (%)473(7.3)Age, years, mean (SD)59.57(9.53)Sex, n (%) Male2996(46.1) Female3500(53.9)Residence, n (%) Rural5225(80.5) Urban1267(19.5)Marriage, married, n (%)5727(88.2)Educational level, n (%) Primary4455(68.6) Secondary1302(20.1) Third734(11.3)Smoking status, n (%) Never3892(60.0) Former456(7.0) Current2139(33.0)Current alcohol use, n (%)2251(34.7)BMI a, kg/m223.76(6.04)Hypertension, n (%)2770(42.6)Diabetes, n (%)1197(18.4)Data are presented as mean (SD) or number (%), as appropriateAbbreviation SD, standard deviation; BMI, body mass indexa Calculated as weight in kilograms divided by height in meters squared Table 1Characteristics of 6 496 participantsCharacteristicsOverallParticipants, No.6 496Death, n (%)473(7.3)Age, years, mean (SD)59.57(9.53)Sex, n (%) Male2996(46.1) Female3500(53.9)Residence, n (%) Rural5225(80.5) Urban1267(19.5)Marriage, married, n (%)5727(88.2)Educational level, n (%) Primary4455(68.6) Secondary1302(20.1) Third734(11.3)Smoking status, n (%) Never3892(60.0) Former456(7.0) Current2139(33.0)Current alcohol use, n (%)2251(34.7)BMI a, kg/m223.76(6.04)Hypertension, n (%)2770(42.6)Diabetes, n (%)1197(18.4)Data are presented as mean (SD) or number (%), as appropriateAbbreviation SD, standard deviation; BMI, body mass indexa Calculated as weight in kilograms divided by height in meters squared Characteristics of 6 496 participants Data are presented as mean (SD) or number (%), as appropriate Abbreviation SD, standard deviation; BMI, body mass index a Calculated as weight in kilograms divided by height in meters squared Fig. 2Distribution and classification matrix of cystatin C–based eGFR using CKD-EPI and EKFC equations. Abbreviation eGFR: estimated glomerular filtration rate; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium Distribution and classification matrix of cystatin C–based eGFR using CKD-EPI and EKFC equations. Abbreviation eGFR: estimated glomerular filtration rate; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium eGFR levels and mortality Mortality rates of all-cause death in subgroups of eGFR categories (CKD-EPI equation) ≥ 90, 60–89.9 and < 60 mL/min/1.73 m2 were 6.2%, 7.5% and 8.7%, respectively. Mortality rates for deaths in subgroups according to the eGFR categories (EKFC equation) were 6.1%, 7.6% and 7.7%, correspondingly. Figure 3 (A, B) shows the dose-responsive relationship between eGFR by CKD-EPI and EKFC equations with mortality. Using cystatin C-based CKD-EPI equation, people of eGFR < 60 mL/min/1.73 m2 had a significantly increased risk of mortality (adjusted risk ratio [RR], 1.527; 95% CI, 1.068–2.178), compared to those with normal eGFR levels (≥ 90 mL/min/1.73 m2) as shown in Table 2. However, people of decreased eGFR groups were not significantly associated with a higher mortality risk when applying EKFC equation. Fig. 3Dose-responsive relationships between eGFR by CKD-EPI and EKFC equations with mortality and cardiovascular disease. Abbreviation eGFR: estimated glomerular filtration rate; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium. The relationship was estimated using restricted cubic splines with knots at the 5th, 50th, and 95th percentiles Dose-responsive relationships between eGFR by CKD-EPI and EKFC equations with mortality and cardiovascular disease. Abbreviation eGFR: estimated glomerular filtration rate; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium. The relationship was estimated using restricted cubic splines with knots at the 5th, 50th, and 95th percentiles Table 2Associations of cystatin C–based estimated GFR with all-cause mortality using CKD-EPI and EKFC equationsNo. Death / TotalModel 1Model 2RR (95% CI)P valueRR (95% CI)P valueCKD-EPI eGFRcys>=90127/2042Ref60 ~ 89.9252/33761.215(0.968–1.532)0.0961.271(0.984–1.652)0.069< 6094/10781.427(1.032–1.968)0.0311.527(1.068–2.178)0.021EKFC eGFRcys>=9088/1449Ref60 ~ 89.9267/35101.254(0.958–1.652)0.1031.283(0.948–1.754)0.112< 60118/15371.229(0.84–1.801)0.2881.320(0.865–2.017)0.198Abbreviation: RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function ConsortiumeGFR unit: ml/minute/1.73 m2Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol Table 2Associations of cystatin C–based estimated GFR with all-cause mortality using CKD-EPI and EKFC equationsNo. Death / TotalModel 1Model 2RR (95% CI)P valueRR (95% CI)P valueCKD-EPI eGFRcys>=90127/2042Ref60 ~ 89.9252/33761.215(0.968–1.532)0.0961.271(0.984–1.652)0.069< 6094/10781.427(1.032–1.968)0.0311.527(1.068–2.178)0.021EKFC eGFRcys>=9088/1449Ref60 ~ 89.9267/35101.254(0.958–1.652)0.1031.283(0.948–1.754)0.112< 60118/15371.229(0.84–1.801)0.2881.320(0.865–2.017)0.198Abbreviation: RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function ConsortiumeGFR unit: ml/minute/1.73 m2Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol Associations of cystatin C–based estimated GFR with all-cause mortality using CKD-EPI and EKFC equations Abbreviation: RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium eGFR unit: ml/minute/1.73 m2 Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol eGFR levels and cardiovascular disease Figure 3 (C, D) shows the dose-responsive relationship between eGFR by CKD-EPI and EKFC equations with stroke. Incidences of cardiovascular disease in subgroups of eGFR categories (CKD-EPI equation) ≥ 90, 60–89.9 and < 60 mL/min/1.73 m2 were 26.5%, 30.8% and 38.5%, respectively. The rates in subgroups according to the eGFR categories using EKFC equation were 25.9%, 29.7% and 37.5%, correspondingly. Using cystatin C-based CKD-EPI equation, people of eGFR < 60 mL/min/1.73 m2 had a significantly increased risk of cardiovascular disease (RR, 1.363; 95% CI, 1.006–1.844), compared to those with normal eGFR levels (≥ 90 mL/min/1.73 m2). However, people of decreased eGFR groups were not significantly associated with a higher risk of incident cardiovascular disease when applying EKFC equation (Table 3). Table 3Associations of cystatin C–based estimated GFR with cardiovascular disease incidence using CKD-EPI and EKFC equationsNo. CVD / TotalModel 1Model 2RR (95% CI)P valueRR (95% CI)P valueCKD-EPI eGFRcys≥ 90542/2042Ref60 ~ 89.91039/33761.125(0.92–1.38)0.2551.133(0.907–1.419)0.275< 60415/10781.427(1.084–1.876)0.0111.363(1.006–1.844)0.045EKFC eGFRcys≥ 90375/1449Ref60 ~ 89.91044/35100.943(0.744–1.201)0.6320.922(0.709–1.204)0.545< 60577/15371.098(0.794–1.518)0.5741.097(0.768–1.568)0.613Abbreviation RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function ConsortiumeGFR unit: ml/minute/1.73 m2Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol Table 3Associations of cystatin C–based estimated GFR with cardiovascular disease incidence using CKD-EPI and EKFC equationsNo. CVD / TotalModel 1Model 2RR (95% CI)P valueRR (95% CI)P valueCKD-EPI eGFRcys≥ 90542/2042Ref60 ~ 89.91039/33761.125(0.92–1.38)0.2551.133(0.907–1.419)0.275< 60415/10781.427(1.084–1.876)0.0111.363(1.006–1.844)0.045EKFC eGFRcys≥ 90375/1449Ref60 ~ 89.91044/35100.943(0.744–1.201)0.6320.922(0.709–1.204)0.545< 60577/15371.098(0.794–1.518)0.5741.097(0.768–1.568)0.613Abbreviation RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function ConsortiumeGFR unit: ml/minute/1.73 m2Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol Associations of cystatin C–based estimated GFR with cardiovascular disease incidence using CKD-EPI and EKFC equations Abbreviation RR, risk ratio; CI, confidence interval; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium eGFR unit: ml/minute/1.73 m2 Model 1 was adjusted for age and sex; model 2 was further adjusted for residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol Sensitivity and subgroup analysis The results remained consistent after additionally adjusting for hs-CRP or serum creatinine level (Table 4). After additionally adjusting for hs-CRP (sensitivity analysis 1), people of eGFR < 60 mL/min/1.73 m2 by CKD-EPI equation were associated a higher risk of mortality (RR, 1.518; 95% CI, 1.060–2.167) and cardiovascular disease (RR, 1.364; 95% CI, 1.006–1.848) compared with the reference group. When further adjusting for creatinine concentrations (sensitivity analysis 2), cystatin C based eGFR levels using CKD-EPI equation were still significantly associated with all-cause death and cardiovascular disease. In subgroups analyses, we found that the associations of eGFR levels using CKD-EPI equation with mortality and cardiovascular disease were largely consistent between male and female, but more significant among those < 65 years. The EKFC equation based eGFR level was not significantly associated with mortality nor cardiovascular disease, irrespective of the whole population and subgroups stratified by age and sex. Table 4Sensitivity analyses of cystatin C–based estimated GFR with all-cause mortality and cardiovascular disease using CKD-EPI and EKFC equationsCKD-EPI EquationEKFC EquationRR (95% CI)P valueRR (95% CI)P value Outcome: mortality Sensitivity 1 ≥ 90Ref60 ~ 89.91.270(0.983–1.650)0.0701.281(0.946–1.751)0.114< 601.518(1.060–2.167)0.0221.310(0.858–2.003)0.212 Sensitivity 2 ≥ 90Ref60 ~ 89.91.325(1.02–1.730)0.0371.312(0.965–1.801)0.087< 601.723(1.166–2.541)0.0061.406(0.897–2.207)0.137 Age < 65 ≥ 90Ref60 ~ 89.91.295(0.975–1.729)0.0771.201(0.87–1.672)0.270< 601.721(1.058–2.724)0.0241.493(0.887–2.469)0.124 Age ≥ 65 ≥ 90Ref60 ~ 89.90.939(0.515–1.837)0.8451.628(0.794–6.27)0.186< 601.146(0.613–2.288)0.6821.559(0.766–5.909)0.201 Male ≥90Ref60 ~ 89.91.35(0.899–2.067)0.1571.332(0.796–2.334)0.293<601.873(1.098–3.208)0.0221.58(0.825–3.099)0.174 Female ≥ 90Ref60 ~ 89.91.229(0.882–1.723)0.2271.284(0.878–1.895)0.202< 601.58(1.083–2.074)0.0301.098(0.617–1.946)0.750 Outcome: CVD Sensitivity 1 ≥ 90Ref60 ~ 89.91.133(0.907–1.419)0.2740.922(0.709–1.204)0.546< 601.364(1.006–1.848)0.0451.097(0.767–1.569)0.613 Sensitivity 2 ≥ 90Ref60 ~ 89.91.213(0.868–1.691)0.2570.881(0.676–1.154)0.353< 601.190(1.009–1.272)0.0480.954(0.652–1.395)0.807 Age < 65 ≥ 90Ref60 ~ 89.91.162(0.928–1.459)0.1930.918(0.702–1.205)0.536< 601.471(1.087–2.148)0.0490.887(0.477–1.553)0.690 Age ≥ 65 ≥ 90Ref60 ~ 89.90.932(0.587–1.538)0.7441.039(0.499–2.532)0.925< 601.222(0.762–2.033)0.4201.269(0.597–3.135)0.568 Male ≥ 90Ref60 ~ 89.90.923(0.695–1.234)0.5850.923(0.695–1.234)0.585< 601.321(1.010–1.922)0.0421.321(0.907–1.922)0.146 Female ≥ 90Ref60 ~ 89.91.368(1.031–1.828)0.0321.054(0.766–1.46)0.748< 601.51(1.007–2.255)0.0451.186(0.743–1.889)0.473Abbreviation RR, risk ratio; CI, confidence interval; hs-CRP, high-sensitivity C-reactive protein; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium; CVD, cardiovascular diseaseeGFR unit: ml/minute/1.73 m2Analyses were adjusted for age, sex, residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol, if applicableSensitivity analysis 1: hs-CRP level was additionally adjusted; sensitivity analysis 2: serum creatinine level was additionally adjusted Table 4Sensitivity analyses of cystatin C–based estimated GFR with all-cause mortality and cardiovascular disease using CKD-EPI and EKFC equationsCKD-EPI EquationEKFC EquationRR (95% CI)P valueRR (95% CI)P value Outcome: mortality Sensitivity 1 ≥ 90Ref60 ~ 89.91.270(0.983–1.650)0.0701.281(0.946–1.751)0.114< 601.518(1.060–2.167)0.0221.310(0.858–2.003)0.212 Sensitivity 2 ≥ 90Ref60 ~ 89.91.325(1.02–1.730)0.0371.312(0.965–1.801)0.087< 601.723(1.166–2.541)0.0061.406(0.897–2.207)0.137 Age < 65 ≥ 90Ref60 ~ 89.91.295(0.975–1.729)0.0771.201(0.87–1.672)0.270< 601.721(1.058–2.724)0.0241.493(0.887–2.469)0.124 Age ≥ 65 ≥ 90Ref60 ~ 89.90.939(0.515–1.837)0.8451.628(0.794–6.27)0.186< 601.146(0.613–2.288)0.6821.559(0.766–5.909)0.201 Male ≥90Ref60 ~ 89.91.35(0.899–2.067)0.1571.332(0.796–2.334)0.293<601.873(1.098–3.208)0.0221.58(0.825–3.099)0.174 Female ≥ 90Ref60 ~ 89.91.229(0.882–1.723)0.2271.284(0.878–1.895)0.202< 601.58(1.083–2.074)0.0301.098(0.617–1.946)0.750 Outcome: CVD Sensitivity 1 ≥ 90Ref60 ~ 89.91.133(0.907–1.419)0.2740.922(0.709–1.204)0.546< 601.364(1.006–1.848)0.0451.097(0.767–1.569)0.613 Sensitivity 2 ≥ 90Ref60 ~ 89.91.213(0.868–1.691)0.2570.881(0.676–1.154)0.353< 601.190(1.009–1.272)0.0480.954(0.652–1.395)0.807 Age < 65 ≥ 90Ref60 ~ 89.91.162(0.928–1.459)0.1930.918(0.702–1.205)0.536< 601.471(1.087–2.148)0.0490.887(0.477–1.553)0.690 Age ≥ 65 ≥ 90Ref60 ~ 89.90.932(0.587–1.538)0.7441.039(0.499–2.532)0.925< 601.222(0.762–2.033)0.4201.269(0.597–3.135)0.568 Male ≥ 90Ref60 ~ 89.90.923(0.695–1.234)0.5850.923(0.695–1.234)0.585< 601.321(1.010–1.922)0.0421.321(0.907–1.922)0.146 Female ≥ 90Ref60 ~ 89.91.368(1.031–1.828)0.0321.054(0.766–1.46)0.748< 601.51(1.007–2.255)0.0451.186(0.743–1.889)0.473Abbreviation RR, risk ratio; CI, confidence interval; hs-CRP, high-sensitivity C-reactive protein; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium; CVD, cardiovascular diseaseeGFR unit: ml/minute/1.73 m2Analyses were adjusted for age, sex, residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol, if applicableSensitivity analysis 1: hs-CRP level was additionally adjusted; sensitivity analysis 2: serum creatinine level was additionally adjusted Sensitivity analyses of cystatin C–based estimated GFR with all-cause mortality and cardiovascular disease using CKD-EPI and EKFC equations Abbreviation RR, risk ratio; CI, confidence interval; hs-CRP, high-sensitivity C-reactive protein; eGFR: estimated glomerular filtration rate; cys: cystatin C; CKD-EPI: Chronic Kidney Disease Epidemiology Collaboration; EKFC: European Kidney Function Consortium; CVD, cardiovascular disease eGFR unit: ml/minute/1.73 m2 Analyses were adjusted for age, sex, residence, marriage, education level, BMI group, smoking status, current drinking, hypertension, systolic blood pressure, diabetes, glucose, triglyceride, nonHDL cholesterol, if applicable Sensitivity analysis 1: hs-CRP level was additionally adjusted; sensitivity analysis 2: serum creatinine level was additionally adjusted Discussion In this study, we observed that decreased eGFR estimated by cystatin C using 2021 CKD-EPI equation is significantly associated with all-cause mortality and cardiovascular disease using a large cohort of unselected middle-aged and older adults, even independent of creatinine level. But we did not observe any significant associations between eGFR by EKFC equation and adverse outcomes among the Chinese population. Our data showed that there was a consistent difference in eGFR by CKD-EPI 2021 and EKFC equations, which is in line with previous studies [23, 24]. Giavarina D et al. reported similar results that the EKFC equation underestimated GFR compared with the CKD-EPI equation in an Italian cohort [25]. One possible explanation is that EKFC equation would be applicable across the full age spectrum with a constant scaling factor to sex using Q value [26]. Applying CKD-EPI and EKFC equations, the prevalence of chronic kidney diseases defined by eGFR < 60 mL/min/1.73 m2 were 16.6% and 23.7% among the middle-aged and older adults, respectively. Using cystatin C-based CKD-EPI equation, people of decreased eGFR were associated a 52.7% excess risk of all-cause mortality and 36.3% excess risk of cardiovascular disease compared to those with normal eGFR status. Of note, the significant association was still observed after creatinine additionally adjusted in the model. However, we did not observe a significant association between cystatin C-based eGFR derived from EKFC equation and mortality nor cardiovascular disease. The prevalence rate of chronic kidney disease is relatively high in China. Previous cross-sectional surveys from a national sample of Chinese adults showed that the total prevalence of chronic kidney disease reached 8.2-10.8%, being a severe public health issue in China [27, 28]. Due to the complicated process and expenditures to directly measure GFR by inulin or radioisotope renography methods, eGFR is a practical surrogate instead, which has been used to reflect the renal function in clinical settings. Given this issue, several equations were developed among different populations using demographic and blood biomarkers to calculate eGFR. The equation of CKD-EPI 2009 was the most widely used one [29]. In 2021, the CKD-EPI cystatin-C based eGFR equation was updated with the exclusion of race information, given that race is only a social factor. Another publication from the EKFC in early 2023 claimed that the EKFC equation based on cystatin-C without including race and sex outperformed the current equations among the US [16]. Europe and Africa. However, the performance and predictive capacity of EKFC equation has limitedly been validated in Chinese population. And the clinical relevance remailed totally unknown. Data from CHARLS cohort showed that the proportion of poor eGFR when applying the new EKFC equation is much higher than that estimated using CKD-EPI equation. There were about 31.5% participants of normal eGFR indicated by CKD-EPI equation reclassified into moderately decreased eGFR group when using EKFC equation. Due to lacking directly measured GFR data in CHARLS cohort, we were unable to compare the accuracy of these two equations among Chinese population. However, this study supplemented the longitudinal association between eGFR based on two equations and risk of mortality and cardiovascular disease, suggesting the applicability from a public health perspective. According to a previous survey study, the EKFC equation seems to underestimate the GFR level and yield poor renal function data [24]. This current study for the first investigated the clinical relevance of cystatin-C based eGFR using CKD-EPI equation and the updated EKFC equation in terms of all-cause mortality risk and cardiovascular disease. We found that cystatin-C based eGFR derived from CKD-EPI equation was an independent risk factor for death and cardiovascular disease onset among Chinese population, but there was no significant effect of eGFR by EKFC equation. The association between eGFR and education level is significant, in regardless of equations. Generally, higher education levels are associated with better health outcomes, including kidney function. Individuals with higher education levels often have greater access to healthcare, healthier lifestyles, and better awareness of chronic disease management, which can contribute to optimal eGFR values [30]. The current findings did not deny the accuracy of EKFC equation but suggested that the current EKFC equation is not a useful indication for mortality risk or cardiovascular disease onset in its current form among middle aged and older Chinese adults. Most of the current eGFR equations have been modified for Asian populations [31–33]. Further research is needed to determine the Chinese-modified EKFC equation for utilization and risk stratification in clinical practice. There are also other studies reporting the associations between cystatin C or cystatin C based eGFR and mortality risk. Cystatin C has been showed as a more sensitive marker than creatinine in reflecting the renal function, specifically for the older population. Cystatin C based eGFR is less biased by potential factors, such as muscle mass, dehydration status, dietary intake, and other comorbidities [34, 35]. A study [36] showed that eGFR using cystatin C is mostly associated with the risk of cardiovascular events and mortality, rather than creatinine-based measures. A study [37] among older men showed that higher cystatin C-based eGFR was associated with an increased risk of frailty or death, not creatinine or creatinine-based GFR level. A cohort study [38] reported that the eGFR equation based on a combination of cystatin C and creatinine may be preferable than that estimated using creatinine equation to predict the risk of mortality. The Chronic Renal Insufficiency Cohort (CRIC) study [39] also suggested that large differences exist between cystatin C-based eGFR and creatinine-based eGFR, which may be informative of mortality risk. A cohort study among Chinese population also showed that the CKD-EPI cystatin C-based equation can best predict the risk of all-cause death [8]. Our study also supplemented the evidence that the cystatin C-based eGFR by CKD-EPI equation outperformed the newly updated EKFC equation for the risk stratification of mortality and cardiovascular disease. Our study still recommended the routine monitoring of cystatin C and the CKD-EPI equation to calculate the eGFR level in the current stage. Limitations There are several limitations to be acknowledged. First, the specific cause of death was lacking in the CHARLS database, which stops us from further analyzing the effects of eGFR functions on cause-specific mortality, such as cardiovascular mortality, cancer, or metabolic related mortality. Second, the CHARLS collected data among participants aged 45 years and older, which means the association found in this current analysis may not be fully generalized to the whole population. Further comparation studies are needed, especially for the clinical relevance of EKFC equation in the Asian population. Third, this is an observational study design, which is potential biased by confounding factors. Conclusion The findings suggest that cystatin C-based eGFR using CKD-EPI equation is more closely associated with all-cause mortality and cardiovascular disease compared to EKFC equation among Chinese adults. This study emphasized that CKD-EPI cystatin C-based eGFR should be monitored in routine practice, which needs further validation in other populations. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1: Additional file: Table S1 [Characteristics of 6 496 participants according eGFR categories using CKD-EPI equation]. Table S2 [Characteristics of 6 496 participants according eGFR categories using EKFC equation]. Supplementary Material 1: Additional file: Table S1 [Characteristics of 6 496 participants according eGFR categories using CKD-EPI equation]. Table S2 [Characteristics of 6 496 participants according eGFR categories using EKFC equation].