Abstract Background Chronic Kidney Disease (CKD) Stage based on estimated GFR governs the decision between treatment choices eg dialysis, transplant etc. There is no customized eGFR formula for Indians, and hardly any data is available regarding concordance between CKD staging by various formulae on Indian population. The CKD-EPI-Pakistan formula, developed through a collaboration involving original authors of MDRD/CKD-EPI, is the Asian formula ethno-geographically closest to India. However it has not been tested on Indians. Methods We compared 4 eGFR formulae: MDRD(4-factor), CKD-EPI-2009, and CKD-EPI-2021-Creatinine and CKD-EPI-Pakistan for 126 patients undergoing kidney function testing in a tertiary care center in Eastern India. Input for race was non-black in all. Results Thirty six of 126 (29%) were discordant by 1 stage between 1 or more formula-pairs, and Kendall's W was 0.034366339 suggesting weak concordance. Among the MDRD CKD3b and CKD4 cases, CKD-EPI Pakistan re-classified 12% by one stage higher. In the same cases, CKD-EPI-2009 and 2021 both assigned 1 stage lower than MDRD in about 18% and 35% respectively. For MDRD-CKD3a cases too CKD-Epi-Pakistan predicted one stage higher in 50%, whereas both CKD-EPI-2009 and 2021 gave one stage lower in 25%. CKD5 staging were mutually concordant, except Pak formula giving 66% more CKD5. Conclusions There is a significant under-staging of CKD stages of Indian patients using western CKD-EPI equations compared to MDRD, whereas CKD-EPI-Pakistan over-stages them especially in the severe stages. If Asian CKD-EPI-Pakistan is considered biologically appropriate for Indians, about 20-30% of treatment-needing patients might currently be under-classified and denied life-saving treatment modalities e.g. Dialysis/CRRT/Transplant, and it might become worse if CKD-EPI 2009 or 2021 is adapted instead of MDRD, especially for higher CKD Grades. The mutual discordance highlights urgent need of having formulae customized for Indian population validated over a large number of patients locally using measured GFR.
Read full abstract