Comparative Determination of Glomerular Filtration Rate Estimation Formulae in Type 2 Diabetic Patients: An Observational Study.

Abstract

Assessing glomerular filtration rate (GFR) involves collecting timed urine samples for 24 hours, requiring significant time and resources in the clinical setting. Using predictive GFR formulae to assess renal function may be a better alternative. Our goal was to determine which predictive GFR formula had the highest level of concordance with the GFR that has been measured in a resource-poor setting. This is an observational study. We selected fifty (50) individuals diagnosed with type 2 diabetes (T2DM) in Kumasi, Ghana. The sociodemographic and clinical characteristics were obtained using a structured questionnaire. Urine was obtained from each subject over 24 hours. The levels of glucose (FBG) and creatinine in patients' blood, as well as the levels of creatinine in their urine, were measured after the patients had fasted overnight. Participants had a mean age of 57.4 ± 10.7 (years), BMI of 27.8 ± 4.1 (kg/m2), FBG of 9.0 ± 3.1 (mmol/L), and creatinine concentrations of 95.6 ± 29.1 (μmol/L). A Krouwer plot was used to compare the measured GFR with three formulae: Chronic Kidney Disease Epidemiology (CKD-EPI), Modification of Diet in Renal Disease (MDRD), and Cockroft-Gault (CG) for GFR prediction. Among the 3 estimates, CG showed nonsignificance (p > 0.05) with the measured GFR. The primary finding was that the GFR calculated using the CG formula was not different from the GFR measured, suggesting that CG is the most appropriate alternative GFR estimate among a cross-section of T2DM patients in Ghana.