Evaluating Equations for Estimated Glomerular Filtration Rate (eGFR) in Patients with Sickle Cell Disease (SCD)

Background: Assessment of kidney function in sickle cell disease poses a unique dilemma, both due to the presence of hyperfiltration and the increased tubular secretion of creatinine, the metabolite most often used in eGFR assessment. In the general population, formulae utilizing serum creatinine have been developed to estimate GFR. The Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) have developed the most recent formulae. Until recently, the CKD-EPI-2009 creatinine-based formula has been the standard for estimating eGFR and included black race as a variable in the formula. This formula has been noted to overestimate eGFR in SCD patients, but its bias is improved when the black race variable is excluded. Among available formulae, the CKD-EPI-2012 cystatin C-based formula was shown to have the least bias in SCD patients compared to radionucleotide measured GFR but loses its accuracy when combined with creatinine. The recently developed CKD-EPI-2021 formula which excludes race as a variable is becoming the standard formula in most clinical laboratories. In this multicenter study, we sought to evaluate the agreement of various CKD-EPI formulae in SCD patents.

Methods: Patients with HbSS or HbSb0 thalassemia between ages 18 and 65 years were enrolled in a prospective cohort study to assess the natural history of kidney disease in SCD. We excluded patients with bone marrow transplantation, diabetic nephropathy, cancer, connective tissue disease, other glomerular diseases, known infection with hepatitis (B or C) or HIV, and end-stage kidney disease on dialysis. Participants were enrolled during routine clinical visits at four centers. At baseline, blood samples and urine were obtained, and kidney function was assessed using each sites clinical laboratory to measure serum creatinine. Serum cystatin C was also measured at baseline using a commercially available ELISA kit. We calculated eGFR using each of the CKD-EPI formula, excluding the race variable in the CKD-EPI-2009 formula. Spearman’s rank correlation coefficients were calculated for each formula compared to the others. Prevalence of CKD was assessed using the CKD-EPI-2009, CKD-EPI-2012 (cystatin C) and CKD-EPI-2021 formulae and defined as the presence of albuminuria or an eGFR ≤ 90 ml/min/1.73m2. Prevalence of CKD derived from the CKD-EPI-2012 was compared to the other two formulae using McNemar’s Chi-squared test with 95% confidence intervals determined using bootstrap replications.

Results: Our cohort enrolled 284 subjects with HbSS or HbSb0 thalassemia. For these analyses, 277 patients with baseline serum creatinine were included. The mean age of included subjects was 32.9 (SD ±11.5) years and 57% were female (N=158). The median eGFR of these subjects using the creatinine-based CKD-EPI-2009 formula was 123.9 (IQR 105.9, 134.0) ml/min/1.73m2. Cystatin C was available in 270 patients and had median value of 0.72 (IQR 0.59, 0.92) mg/L. Correlation between the several eGFR formulae are presented in Table 1. CKD-EPI-2009 and CKD-EPI-2021 creatinine-based formulae were very highly correlated with each other; both had similar but less correlation with the CKD-EPI-2012 cystatin C-based formula, which has been shown previously to have the least bias when compared to measured GFR. The prevalence of CKD using the creatinine-based CKD-EPI-2009 formula was 51.3% and 50.5% using the creatinine-based CKD-EPI-2021 formula. When assessed with the cystatin C-based CKD-EPI-2012 formula, CKD prevalence was higher at 55.7% (p<0.0001) than that seen with both of creatinine-based CKD-EPI formulae.

Conclusions: When comparing the new CKD-EPI-2021 creatinine-based formula to the CKD-EPI-2009 formula (excluding the black race variable), eGFR calculation remains very similar in SCD patients. The CKD-EPI-2009 formula is well recognized to overestimate eGFR in SCD patients. Our data would suggest the newer formula would have similar bias. Using the cystatin C-based CKD-EPI-2012 formula, the prevalence of CKD is higher in this SCD cohort than using creatinine-based CKD-EPI formulae, although our data are limited by lack of directly measured GFR. We anticipate the newer CKD-EPI-2021 creatinine-based formula would similarly underestimate prevalence of CKD in SCD. Use of cystatin C for assessing kidney function in SCD patients should be considered, and resources should be utilized to make this measure more widely and readily available.