Development and validation of a new equation to estimate glomerular filtration rate in Argentinian adults

Latin Americans are underrepresented in the current equations used to estimate glomerular filtration rate (GFR), and the applicability of their predictions to this population may therefore be questionable. The objective of this study was to develop a new equation to estimate GFR based on data from Argentina. A cross-sectional study was conducted. We included 583 Argentinian adults in development sample (DS) and 78 in temporary validation sample (TVS). Urinary iothalamate clearance (reference method), and different predictive variables were used to develop candidate equations to estimate GFR. The performance was assessed through 10-fold cross-validation in DS, and by direct validation in TVS, using root mean squared error (RMSE), correlation (r), bias, P15, P30 and correct classification percentage (CC%) in CKD stages. The Argentinian equation (AE) chosen is based on a quasi-likelihood model which predicts GFR from creatinine, age, sex, single kidney, albumin and urea. Within the previous creatinine based equations (MDRD, MCQ, CKD-EPI and EKFC), the ones with the best performance were CKD-EPI 2009 and 2021. In the DS, AE showed lower RMSE, similar r, higher P15, median bias closer to zero and higher CC% compared to CKD-EPI 2009, and very slight differences in comparison to CKD-EPI 2021. In the TVS, AE presented lower RMSE, higher (or equal) r, P30 and CC%, and median bias closer to 0 compared to CKD-EPI in its two versions. In addition, it presented higher P15 than CKD-EPI 2009. In conclusion, AE presented a better performance to estimate GFR in Argentinian people and its use could have a positive impact on the clinical management of these patients.