Performance of glomerular filtration rate estimation equations in Congolese healthy adults: The inopportunity of the ethnic correction

Abstract

Context and objective: In the estimation of glomerular filtration rate (GFR), ethnicity is an important determinant. However, all existing equations have been built solely from Caucasian and Afro-American populations and they are potentially inaccurate for estimating GFR in African populations. We therefore evaluated the performance of different estimated GFR (eGFR) equations in predicting measured GFR (mGFR).

Methods: In a cross-sectional study, 93 healthy adults were randomly selected in the general population of Kinshasa, Democratic Republic of the Congo, between June 2015 and April 2016. We compared mGFR by plasma clearance of iohexol with eGFR obtained with the Modified Diet in Renal Disease (MDRD) equation with and without ethnic factor, the Chronic Kidney Disease Epidemiology (CKD-EPI) serum creatinine (SCr)-based equation, with and without ethnic factor, the cystatin C-based CKD-EPI equation (CKD-EPI SCys) and with the combined equation (CKD-EPI SCrCys) with and without ethnic factor. The performance of the equations was studied by calculating bias, precision and accuracy within 30% (P30) of mGFR.

Results: There were 48 women and 45 men. Their mean age was 45.0±15.7 years and the average body surface area was 1.68±0.16m2. Mean mGFR was 92.0±17.2 mL/min/1.73m2 (range of 57 to 141 mL/min/1.73m2). Mean eGFRs with the different equations were 105.5±30.1 and 87.2±24.8 mL/min/1.73m2 for MDRD with and without ethnic factor, respectively; 108.8±24.1 and 94.3x20.9 mL/min/1.73m2 for CKD-EPI SCr with and without ethnic factor, respectively, 93.5±18.6 mL/min/1.73m2 for CKD-EPI SCys; 93.5±18.0 and 101±19.6 mL/min/ 1.73m2 for CKD-EPI SCrCys with and without ethnic factor, respectively. All equations slightly overestimated mGFR except MDRD without ethnic factor which underestimated by -3.8±23.0 mL/min /1.73m2. Both CKD-EPI SCr and MDRD with ethnic factors highly overestimated mGFR with a bias of 17.9±19.2 and 14.5±27.1 mL/min/1.73m2, respectively. There was a trend for better P30 for MDRD and CKD-EPI SCr without than with the ethnic factor [86.0% versus 79.6% for MDRD (p = 0.21) and 81.7% versus 73.1% for the CKD-EPI SCr equations (p = 0.057)]. CKD-EPI SCrCys and CKD-EPI SCys were more effective than creatinine-based equations.

Conclusion: In the Congolese healthy population, MDRD and CKD-EPI equations without ethnic factors had better performance than the same equations with ethnic factor. The equations using Cys C (alone or combined with SCr) performed better than the creatinine-based equations.

Fig 1. Bland and Altman analysis between mGFR and creatinine-based equations (with and without the ethnic factor).

All values are expressed in mL/min/1.73m². CKD-EPI SCr: Chronic Kidney Disease-Epidemiology Collaboration equation based on serum creatinine only, with ethnic factor; CKD-EPI SCr nef: CKD-EPI without ethnic factor. MDRD: Modification of Diet in Renal Disease study equation with ethnic factor; MDRD nef: MDRD without ethnic factor. The central line represents the mean difference between measured and estimated GFR, whereas the upper and lower lines represent the limits of agreement (mean difference ± 2SD).

Fig 2. Bland and Altman analysis between mGFR and cystatin C-based and combined equations (with and without the ethnic factor).

All values are expressed in mL/min/1.73m². CKD-EPI SCys CKD-EPI equation based on cystatin C only; CKD-EPI SCrCys: CKD-EPI combining creatinine and cystatin C with ethnic factor. CKD-EPI SCrCys nef: CKD-EPI combining serum creatinine and cystatin C without ethnic factor. The central line represents the mean difference between measured and estimated GFR, whereas the upper and lower lines represent the limits of agreement (mean difference ± 2SD).