Recent studies show that high agreement between estimated GFR based on serum creatinine (eGFR_crea) and cystatin C (eGFR_cys) indicates high accuracy of the mean of the two estimates.  When the difference between the two (∆eGFR) is <40%, a P30 accuracy rate of more than 90% has been documented in a research population (den Bakker et al, ClinChimActa 474(2017):38-43). This was the case in ca. 80% of the measurements.

Aim

 To assess the relationship between eGFR_crea, eGFR_cys  and ∆eGFR in a broad pediatric nephrological population and identify factors influencing the difference between eGFR_crea and eGFR_cys.

Material and methods:

Retrospective analysis on 464 patients from a pediatric nephrology outpatient clinic in whom creatinine and cystatin C were measured simultaneously. The FASage equation was used for creatinine, FAScys for cystatin C. ∆eGFR was calculated as (|eGFRcrea-eGFRcys|)/(0.5x(eGFRcrea+eGFRcys). Concordance between eGFR_crea and eGFR_cys was measured using Pearson’s correlation. Diagnosis and steroid use were assessed as potential confounding factors on ∆eGFR using non-parametric tests

Results:

 There was a highly significant correlation of 0.411 between eGFR_crea and eGFR_cys (p=0.000). The distribution of ∆eGFR was 0-9%: 27.6%, 10-19%: 28.2%, 20-39%: 33%, >40%: 11.2%. None of the factors studied had a significant influence on ∆eGFR although spina bifida patients trended to have higher ∆eGFR, which was due to high eGFR_creat (median 123.4ml/min/1.73m² [IQR 42.6] vs 111.4 [30.8], p=0.000).

Conclusions:

 In a broad pediatric nephrological population, ∆eGFR is below 40% in almost 90% of measurements indicating that the average of eGFR_crea and eGFR_cys can be used as a reasonable estimate of GFR in the vast majority of patients. In patients with neural tube defects, the cystatin C-based eGFR should be used as shown previously.