RECALIBRATION OF EKFC eGFR EQUATION FOR INDIAN POPULATION: AN INTERNAL VALIDATION STUDY

Recently, the European Kidney Function Consortium (EKFC) developed a new creatinine-based eGFR equation intended to be applicable across the full age spectrum. The EKFC approach “rescales” serum creatinine to an age- and sex-specific normal median value, denoted Q. The original EKFC equation was developed and validated in European cohorts, and it demonstrated improved accuracy and lower bias compared to prior equations in those populations. No study to date has evaluated the EKFC equation in an Indian population, which is ethnically and biologically distinct from the European populations from which EKFC was derived. The present study evaluates the performance of the EKFC equation in a large Indian cohort against measured GFR. We then derive Indian-specific Q from a healthy sub-cohort and apply these to recalibrate the EKFC creatinine equation. In addition to the Indian Q-value we also tested the interest of using a specific Indian mGFR median value, i.e. 90.25 (referred here as K) instead of 107.3.

Indian adults (>18 years) were recruited comprising healthy individuals and chronic kidney disease (CKD) patients. Healthy participants had blood pressure <140/90 mmHg, fasting glucose <126 mg/dL, eGFR (CKD-EPI 2009) >60 mL/min/1.73 m², and normal urinary protein indices. mGFR was measured by plasma iohexol clearance (120, 180, 240 min). Indian-specific Q values were derived from healthy individuals, and a population-specific constant (K) from those under age 40 years. We evaluated EKFC (original creatinine), and recalibrated EKFC equations incorporating Indian Q (ISQ), Indian constant (ISK), and both (ISQ+ISK). Performance was assessed by bias, precision (RMSE), and accuracy (P30: % of eGFR within 30% of mGFR).

A total of 1,174 participants (615 healthy, 559 CKD) were included. The mean age of study population was 48 years with equal gender distribution. The healthy group had median serum creatinine (Q values) of 0.81 mg/dL (males) and 0.63 mg/dL (females). Median mGFR among healthy participants <40 years was 90.25 (IQR 80–103) mL/min/1.73 m². The overall mean mGFR was 59.8±31.9 mL/min/1.73 m². Mean eGFRs were (mL/min/1.73 m²): 69.8±36.1 (EKFCcr), 64.7±35.5 (EKFC_ISQ), 58.7±30.4 (EKFC_ISK), and 55.4±29.5 (EKFC_ISQ+ISK). The EKFCcr equation showed a mean bias of –9.9 (95% CI: –11.1 to –8.7) mL/min/1.73 m² with P30=58.1%. EKFC_ISQ reduced bias to –4.9 (–6.0 to –3.7) and improved P30 to 64%. EKFC_ISK achieved a bias of 1.2 (0.2–2.2) and P30=69%. EKFC_ISQ+ISK has a bias of 5.4 (–4.4 to 6.4) and P30=64.2%. In individuals with mGFR ≥60 mL/min/1.73 m², EKFCcr bias was –11.1 mL/min/1.73 m², with P30=66%. EKFC_ISQ and EKFC_ISK improved bias to –5.1 and 4.6 mL/min/1.73 m², with P30 of 74.7% and 84.6%, respectively. EKFC_ISQ+ISK showed bias 9.6 mL/min/1.73 m², and P30=81.5%. Among those with mGFR <60, EKFCcr bias was –8.8 mL/min/1.73 m², with P30~50% while recalibration with ISQ improved bias (–4.7 mL/min/1.73 m²,) and P30 (~54%).

Incorporating Indian-specific Q and K values markedly enhanced the performance of EKFC equations in Indian adults, particularly at higher GFR levels. While bias and accuracy improved, performance remained suboptimal in advanced CKD. Population-specific recalibration thus provides more reliable GFR estimates for Indian individuals.