Performance of estimated glomerular filtration rate equations across ethnic groups in the AIM CKD UK cohort: a multicentre cross-sectional study

Abstract

Summary Background Accurate assessment of kidney function using estimated glomerular filtration rate (eGFR) is fundamental for chronic kidney disease diagnosis and management, although performance is variable in African and Asian cohorts. We evaluated creatinine-based eGFR performance in an ethnically diverse United Kingdom cohort. Methods We conducted a multi-centre, cross-sectional study. We included adults who had paired measured GFR (mGFR) and serum creatinine tests between 2009 and 2022. eGFR was calculated using five validated creatinine-based equations. The primary outcome was eGFR agreement with mGFR determined by assessing correlation, difference, precision and 30% accuracy (P30) with stratification by ethnicity. Findings The study included 15,879 participants (mean age 54 ± 16 years; 8893/15,879 (56%) male). On average, all equations overestimated mGFR. There were significant variations in difference (p < 0.001) and accuracy (p < 0.001) between all equations. Lund-Malmo Revised (LMR) had the least difference (1.4 mL/min/1.73 m2; 95% CI: 1.1–1.6 mL/min/1.73 m2) and highest accuracy (P30: 86.0%; 95% CI: 85.5–86.6%), followed by European Kidney Function Consortium (EKFC). eGFR had the greatest difference and lowest accuracy in South Asians, compared with other ethnicity groups. mGFR was substantially overestimated in young adults (<25 years), low BMI (<20 kg/m2) and hypoalbuminaemia (<20 g/L). Hypothetical prescribing analyses for dapagliflozin (upper GFR threshold: 75 mL/min/1.73 m2) showed potential prescribing ineligibility based on eGFR alone in 29% (1933/6586) to 52% (3433/6586) of eligible participants. Interpretation LMR and EKFC showed improved agreement with mGFR, particularly in South Asians and young adults. Further prospective validation is recommended to support changes in clinical practice. Funding Kidney Research UK.