Abstract: PO1974
Diagnosis-Specific Combination of Cystatin C- and Creatinine-Based eGFR
Session Information
- Pediatric Nephrology: AKI, Dialysis, Transplant, CKD, and Nephrotic Syndrome
November 04, 2021 | Location: On-Demand, Virtual Only
Abstract Time: 10:00 AM - 12:00 PM
Category: Pediatric Nephrology
- 1700 Pediatric Nephrology
Authors
- Bokenkamp, Arend, Amsterdam UMC Locatie VUmc, Amsterdam, Noord-Holland, Netherlands
- Oostdam, Tobias, Amsterdam UMC Locatie VUmc, Amsterdam, Noord-Holland, Netherlands
- Björk, Jonas, Lunds Universitet, Lund, Sweden
- Nyman, Ulf, Skanes universitetssjukhus Malmo, Malmo, Skåne, Sweden
- Goffin, Karolien, Katholieke Universiteit Leuven Universitaire Ziekenhuizen Leuven, Leuven, Flanders, Belgium
- Åsling Monemi, Kajsa, Karolinska Universitetssjukhuset, Stockholm, Sweden
- Hansson, Magnus Daniel, Karolinska Universitetssjukhuset, Stockholm, Sweden
- Berg, Ulla B., Karolinska Universitetssjukhuset, Stockholm, Sweden
- Littmann, Karin, Karolinska Universitetssjukhuset, Stockholm, Sweden
- Grubb, Anders O., Lunds Universitet, Lund, Sweden
- Pottel, Hans, Katholieke Universiteit Leuven Universitaire Ziekenhuizen Leuven, Leuven, Flanders, Belgium
- Den bakker, Emil, Amsterdam UMC Locatie VUmc, Amsterdam, Noord-Holland, Netherlands
Background
The arithmetic mean of a creatinine- and a cystatin C-based GFR estimation (eGFR) has higher accuracy than either of the two. Smaller studies indicate that the relative contribution of the creatinine- and the cystatin C-based equation should be adapted based on underlying diagnosis.
Methods
Retrospective analysis of 1712 plasma clearance GFR measurements from four pediatric nephrology centers. eGFR was calculated using the height-based Full Age Spectrum equation using creatinine (FAS-creat) and the FAS-cys equation for cystatin C. α describes the contribution of FAS-creat, (1-α) the contribution of FAS-cys. 2/3 of the cohort (mean age 11.8 years, mGFR 93.8 ml/min/1.73m2) was used to determine the α-values yielding the highest P30 accuracy globally (FASα) and in diagnosis subgroups. These α-values were validated in the remaining 1/3 of the cohort assessing accuracy, bias and precision.
Results
Globally, the optimal α-value was 0.3 [95% CI 0.2 – 0.4, Figure]. Lower α-values were determined for spina bifida (0), glomerulonephritis (0.2), and liver disease (0.25), while CAKUT, kidney transplantation and tubulointerstitial disease had α-values between 0.35 and 0.55.
Accuracy of FAS0.3 in the validation cohort was 90.2%, which was significantly higher than accuracy of the arithmetic mean (87.4%, p<0.05). Using the diagnosis-specific α-values rather than FAS0.3 improved bias (+2.5 vs +1.3 ml/min/1.73m2, p < 0.01) and precision (14.1 vs 13.8 ml/min/1.73m2, p < 0.05), while accuracy was unchanged. Only in children with spina bifida, a clinically relevant improvement was observed when using the diagnosis-specific α-value.
Conclusion
For calculation of the weighted mean, a fixed mix of 30% FAS-creat and 70% FAS-cys is optimal and yields very high accuracy overall. A disease-specific adaption (i.e. 100% cystatin C eGFR) is clinically relevant only for patients with spina bifida.