Advertisement
American Journal of Kidney Diseases

The Modified CKiD Study Estimated GFR Equations for Children and Young Adults Under 25 Years of Age: Performance in a European Multicenter Cohort

Published:March 25, 2022DOI:https://doi.org/10.1053/j.ajkd.2022.02.018
      To the Editor:
      The CKiD creatinine- and cystatin C–based glomerular filtration rate (GFR) estimating equations have recently been modified by incorporation of continuous age- and sex-dependent k values to yield less biased internal validation results in chronic kidney disease (CKD) patients aged under 25 years.
      • Pierce C.B.
      • Muñoz A.
      • Ng D.K.
      • Warady B.A.
      • Furth S.L.
      • Schwartz G.J.
      Age- and sex-dependent clinical equations to estimate glomerular filtration rates in children and young adults with chronic kidney disease.
      Here we report an external validation of these CKiDU25 equations in the European Kidney Function Consortium (EKFC) multicenter cohort (Table 1, Item S1).
      • Björk J.
      • Nyman U.
      • Berg U.
      • et al.
      Validation of standardized creatinine and cystatin C GFR estimating equations in a large multicentre European cohort of children.
      ,
      • Björk J.
      • Nyman U.
      • Larsson A.
      • Delanaye P.
      • Pottel H.
      Estimation of the glomerular filtration rate in children and young adults using the CKD-EPI equation with age-adjusted creatinine values.
      In this cohort of children and young adults, the vast majority have a measured GFR (mGFR) ≥75 mL/min/1.73 m2, thus resembling a setting where CKD is screened for. Comparisons were made with 3 creatinine and 3 cystatin C GFR equations (Items S2 and S3) applicable for the entire lifespan starting from 2 years of age,
      • Björk J.
      • Nyman U.
      • Larsson A.
      • Delanaye P.
      • Pottel H.
      Estimation of the glomerular filtration rate in children and young adults using the CKD-EPI equation with age-adjusted creatinine values.
      • Grubb A.
      • Horio M.
      • Hansson L.O.
      • et al.
      Generation of a new cystatin C-based estimating equation for glomerular filtration rate by use of 7 assays standardized to the international calibrator.
      • Pottel H.
      • Delanaye P.
      • Schaeffner E.
      • et al.
      Estimating glomerular filtration rate for the full age spectrum from serum creatinine and cystatin C.
      • Björk J.
      • Nyman U.
      • Delanaye P.
      • et al.
      A novel method for creatinine adjustment makes the revised Lund-Malmö GFR equation applicable in children.
      • Pottel H.
      • Björk J.
      • Courbebaisse M.
      • et al.
      Development and validation of a modified full age spectrum creatinine-based equation to estimate glomerular filtration rate: a cross-sectional analysis of pooled data.
      including the adult cystatin C–based CKD-EPI equation,
      • Inker L.A.
      • Schmid C.H.
      • Tighiouart H.
      • et al.
      Estimating glomerular filtration rate from serum creatinine and cystatin C.
      shown to have an acceptable performance in children.
      • Björk J.
      • Nyman U.
      • Berg U.
      • et al.
      Validation of standardized creatinine and cystatin C GFR estimating equations in a large multicentre European cohort of children.
      Plasma and renal clearance methods for mGFR were used as the reference test and plasma/serum creatinine and cystatin C assays were traceable to international standards (Item S1). Comparisons focused on bias, precision, and accuracy, with P30 accuracy as the main performance metric.
      Table 1Patient Characteristics of the Cohorts
      Children With Creatinine and Cystatin C Measured (N = 2,293)Young Adults (N = 1,816)
      With Creatinine Measured (n = 1,816)With Cystatin C Measured (n = 348)
      Age, y11.9 (2.3-17.8)20.0 (18.0-24.6)18.9 (18.0-24.1)
      Female sex949 (41%)846 (47%)144 (41%)
      Body mass index, kg/m218 (14-29)21 (16-31)22 (15-35)
      Body surface area, m21.29 (0.54-2.05)1.68 (1.29-2.14)1.73 (1.33-2.36)
      Plasma/serum creatinine, μmol/L52 (19-155)75 (41-191)78 (38-196)
      Plasma/serum cystatin, mg/L0.96 (0.61-2.72)0.96 (0.62-2.39)
      mGFR, mL/min/1.73 m297 (28-169)92 (31-141)91 (30-134)
      mGFR <75 mL/min/1.73 m2503 (22%)543 (30%)95 (27%)
      Children defined as aged 2.0-17.9 years; young adults as 18.0-24.9 years. Continuous variables given as median (2.5 and 97.5 percentiles).
      For creatinine-based equations, the overall distribution of CKiDU25 eGFR followed mGFR more closely at higher levels of kidney function than did eGFRs calculated with CKD-EPI40, EKFC, or LMR18 (Fig S1). On the other hand, CKiDU25cr yielded overestimations exceeding 50 mL/min/1.73 m2 more often than the other creatinine-based equations (Fig S2). Overall, P30 for CKiDU25cr exceeded 80% in both children and young adults. In children, it was not different from CKD-EPI40, but was lower than for EKFC and LMR18 (Table 2). In young adults, it was not different from EKFC but was lower than for CKD-EPI40 and LMR18. In both age groups, the accuracy for CKiDU25cr was lower in male than female patients (Table S1).
      Table 2Bias, Precision, and Accuracy of eGFR Equations in Children and Young Adults
      Creatinine EquationsCystatin C Equations
      CKiDU25CKD-EPI40EKFCLMR18CKiDU25CAPACKD-EPIcysFAS
      Children (n = 2,293 for creatinine equations and for cystatin C equations)
      Bias, median1.3 (0.6; 2.2)−5.3 (−6.1; −4.5)−1.6 (−2.4; −0.4)−4.5 (−5.3; −3.7)−12.8 (−13.8; −11.8)−0.6 (−1.7; 0.3)−2.5 (−3.3; −1.3)−4.2 (−4.8; −3.4)
      Precision, IQR23.623.923.423.024.226.924.523.9
      Accuracy, P3083.8 (82.3; 85.3)83.9 (82.4; 85.4)85.2 (83.8; 86.7)86.3 (84.9; 87.8)82.6 (81.0; 84.1)82.0 (80.4; 83.6)84.3 (82.9; 85.8)85.9 (84.5; 87.3)
      P30 differenceReference0.1 (−1.4; 1.6)1.4 (0.1; 2.7)2.5 (1.1; 3.8)Reference−0.6 (−2.4; 1.2)1.8 (0.1; 3.4)3.4 (1.9; 4.8)
      Young adults (n = 1,816 for creatinine equations and n = 348 for cystatin C equations)
      Bias, median2.1 (1.3; 2.9)−1.7 (−2.8; −1.1)2.5 (1.8; 3.1)−2.6 (−3.5; −1.5)−11.5 (−13.6; −8.6)−0.2 (−1.9; 1.4)2.0 (−0.6; 4.3)3.7 (1.7; 5.3)
      Precision, IQR22.321.821.321.522.323.522.223.4
      Accuracy, P3082.8 (81.0; 84.5)85.4 (83.7; 87.0)84.0 (82.3; 85.7)86.1 (84.5; 87.7)83.9 (80.0; 87.8)84.8 (81.0; 88.5)83.3 (79.4;87.2)84.2 (80.4; 88.0)
      P30 differenceReference2.6 (1.2; 4.0)1.3 (−0.2; 2.7)3.4 (1.9; 4.8)Reference0.9 (−3.6; 5.3)−0.6 (−5.3; 4.1)0.3 (−4.5; 5.0)
      Differences in P30 were evaluated statistically using 95% CIs for paired proportions. Abbreviations and definitions: bias, median error eGFR – mGFR (given with 95% CI); precision, IQR of eGFR – mGFR, expressed in mL/min/1.73 m2; P30, accuracy expressed in percentage of GFR estimates within ±30% of mGFR (given with 95% CI).; IQR, interquartile range; CKiDU25, Chronic Kidney Disease in Children equation for individuals under 25 years; CKD-EPI40, Chronic Kidney Disease Epidemiology equation based on creatinine values adjusted for individuals under 40 years
      • Björk J.
      • Nyman U.
      • Larsson A.
      • Delanaye P.
      • Pottel H.
      Estimation of the glomerular filtration rate in children and young adults using the CKD-EPI equation with age-adjusted creatinine values.
      ; EKFC, European Kidney Function Consortium equation
      • Pottel H.
      • Björk J.
      • Courbebaisse M.
      • et al.
      Development and validation of a modified full age spectrum creatinine-based equation to estimate glomerular filtration rate: a cross-sectional analysis of pooled data.
      ; LMR18, Lund-Malmö revised equation based on creatinine values adjusted for individuals under 18 years
      • Björk J.
      • Nyman U.
      • Delanaye P.
      • et al.
      A novel method for creatinine adjustment makes the revised Lund-Malmö GFR equation applicable in children.
      ; CAPA, Caucasian, Asian, Paediatric and Adult equation
      • Grubb A.
      • Horio M.
      • Hansson L.O.
      • et al.
      Generation of a new cystatin C-based estimating equation for glomerular filtration rate by use of 7 assays standardized to the international calibrator.
      ; CKD-EPIcys, CKD-EPI cystatin C equation; FAS, Full Age Spectrum equation.
      • Pottel H.
      • Delanaye P.
      • Schaeffner E.
      • et al.
      Estimating glomerular filtration rate for the full age spectrum from serum creatinine and cystatin C.
      For cystatin C–based equations, CKiDU25 exhibited marked underestimation in both children and young adults, and in both male and female patients; the overall distribution of CKiDU25cys eGFR thus deviated markedly from the mGFR distribution (Table S1, Figs S3 and S4). Overall, P30 for CKiDU25cys exceeded 80% in children and young adults. In children, it was not different from CAPA, but was lower than for CKD-EPI and FAS (Table 2). In young adults, P30 was not different from the other cystatin C equations but results were hampered by statistical imprecision, as reflected by the wide confidence intervals.
      For simplicity and to limit the number of comparisons, stratification by the eGFR threshold
      • Delanaye P.
      • Jager K.J.
      • Bokenkamp A.
      • et al.
      CKD: a call for an age-adapted definition.
      75 mL/min/1.73 m2 was limited to the best-performing equations (Table S2). CKiDU25cr was more accurate than EKFC in patients with lower eGFR, whereas EKFC was more accurate in those with higher eGFR. These findings were consistent irrespective of whether eGFR based on CKiDU25cr or EKFC was used for stratifying. CKiDU25cys was less accurate than FAS, and was negatively biased both below and above the eGFR threshold of 75 mL/min/1.73 m2.
      A study limitation was that 44% and 68% of the children and young adults, respectively, were included in the development cohort for the EKFC equation, and 20% and 4.5% of the children and young adults, respectively, were included in the development cohort for the CAPA equation (see Item S1 for more details). This means that the performance estimates for these equations may be somewhat upwardly biased. Also, it was not possible to stratify results by ethnic origin, as such data were not available; thus, the generalizability of the results to specific population groups is uncertain.
      CKiDU25cr is less suitable for automatic laboratory reporting than the height-independent lifespan equations, since height is often missing at the time of analysis. Another drawback is the potential for implausible changes in eGFR at the transition from CKiDU25cr to an adult equation at age 25, which is avoided by using lifespan equations. For CKiDU25cys a particular concern is the observed underestimation in both children and young adults, which may lead to inflation of false-positives when screening for CKD.
      In conclusion, if the focus is on overall estimation accuracy in children and young adults in settings where the vast majority can be expected to have near-normal or normal GFR, then our results do not provide strong arguments for using any of the two CKiDU25 equations instead of lifespan equations. However, if the aim is to screen for and detect CKD, then including patient’s height and using CKiDU25cr may be advantageous.

      Article Information

      Additional Information

      UN, JB, UB, AB, LB, AG, HP, and PD are European Kidney Function Consortium (EKFC) Members.

      Authors’ Contributions

      Research idea and study design: UN and JB (equal contribution); data acquisition: UB, AB, LD, KG, AG, MH, AL, KL, KÅ; data analysis/interpretation: UN, JB, HP, PD; statistical analysis: JB, HP; supervision or mentorship: HP and PD (equal contribution). Each author contributed important intellectual content during manuscript drafting or revision and agrees to be personally accountable for the individual’s own contributions and to ensure that questions pertaining to the accuracy or integrity of any portion of the work, even one in which the author was not directly involved, are appropriately investigated and resolved, including with documentation in the literature if appropriate.

      Support

      Swedish Research Council (Vetenskapsrådet; grant no. 2019 – 00198). The EKFC is supported by the European Renal Association. The funders had no role in study design, data collection, analysis, reporting, or the decision to submit for publication.

      Financial Disclosure

      The authors declare that they have no relevant financial interests.

      Peer Review

      Received June 1, 2021. Evaluated by 2 external peer reviewers, with direct editorial input from a Statistics/Methods Editor, an Associate Editor, and the Editor-in-Chief. Accepted in revised form February 7, 2022.

      Supplementary Material

      References

        • Pierce C.B.
        • Muñoz A.
        • Ng D.K.
        • Warady B.A.
        • Furth S.L.
        • Schwartz G.J.
        Age- and sex-dependent clinical equations to estimate glomerular filtration rates in children and young adults with chronic kidney disease.
        Kidney Int. 2021; 99: 948-956
        • Björk J.
        • Nyman U.
        • Berg U.
        • et al.
        Validation of standardized creatinine and cystatin C GFR estimating equations in a large multicentre European cohort of children.
        Pediatr Nephrol. 2019; 34: 1087-1098
        • Björk J.
        • Nyman U.
        • Larsson A.
        • Delanaye P.
        • Pottel H.
        Estimation of the glomerular filtration rate in children and young adults using the CKD-EPI equation with age-adjusted creatinine values.
        Kidney Int. 2021; 99: 940-947
        • Grubb A.
        • Horio M.
        • Hansson L.O.
        • et al.
        Generation of a new cystatin C-based estimating equation for glomerular filtration rate by use of 7 assays standardized to the international calibrator.
        Clin Chem. 2014; 60: 974-986
        • Pottel H.
        • Delanaye P.
        • Schaeffner E.
        • et al.
        Estimating glomerular filtration rate for the full age spectrum from serum creatinine and cystatin C.
        Nephrol Dial Transplant. 2017; 32: 497-507
        • Björk J.
        • Nyman U.
        • Delanaye P.
        • et al.
        A novel method for creatinine adjustment makes the revised Lund-Malmö GFR equation applicable in children.
        Scand J Clin Lab Invest. 2020; 80: 456-463
        • Pottel H.
        • Björk J.
        • Courbebaisse M.
        • et al.
        Development and validation of a modified full age spectrum creatinine-based equation to estimate glomerular filtration rate: a cross-sectional analysis of pooled data.
        Ann Intern Med. 2021; 174: 183-191
        • Inker L.A.
        • Schmid C.H.
        • Tighiouart H.
        • et al.
        Estimating glomerular filtration rate from serum creatinine and cystatin C.
        N Engl J Med. 2012; 367: 20-29
        • Delanaye P.
        • Jager K.J.
        • Bokenkamp A.
        • et al.
        CKD: a call for an age-adapted definition.
        J Am Soc Nephrol. 2019; 30: 1785-1805