SP432THE ERROR OF ESTIMATED GFR Y TYPE 2 DIABETES

Abstract

INTRODUCTION AND AIMS: Type 2 diabetes (T2DM) is one of the major causes of renal disease worldwide. Renal function is the main outcome in studies that evaluate new strategies to prevent the progression of renal disease, and is usually assessed by estimation formulas (eGFR). Nevertheless, formulas have an error of630% when compared with measured GFR. This has been observed in diverse populations (CKD, kidney transplantation, ADPKD, etc) including diabetics. However, no study evaluated the performance of all the available formulas to estimate GFR, including those creatinine and cystatin-C based. We aimed to assess the performance of 68 formulas with mGFR determined by the plasma clearance of iohexol in patients with T2DMincluding the whole range of renal function. METHODS: We analyzed 346 patients with T2DM. All patients underwent the plasma clearance of iohexol. Simulstaneously, creatinine and cystatin-C were determined to calculate eGFR by formulas. The agreement between mGFR and eGFR was assessed by total deviation index (TDI) and concordance correlation coefficient (CCC). TDI indicates how wide the limits of the differences between mGFR and eGFR need to be to include the majority of eGFR estimations. Empirical TDI was calculated for a theoretical TDI of 10% and a coverage probability of 90%. The CCC is a statistic that combines meaningful components of accuracy and precision. Patients were also analyzed according with CKD stages. RESULTS: Patients characteristics included: mean age 62.4±11.0, 251 (69%) were male. mGFR ranged from 9.4 to 173.7 ml/min. Table 1 shows the agreement of 10 representative formulas based on creatinine and cystatin-C for the overall cohort. CCC ranged from 0.87 (0.85) to 0.94 (0.92), for the Rule≥cr and CKD-EPI≥cr equations, respectively (Table 1), showing moderate accuracy and precision. TDI ranged from 78.25 (84.97) to 40.68 (44.69) for the Rule-cr and CKD-EPI-cr and Stevens equations, respectively (Table 1). TDI averaged about 55%, indicating that 90% of the estimations of GFR showed an error ranging from-55 to+55% when compared with mGFR. Similar results were observed when the patients were analyzed according with CKD stages and also for the 58 remaining formulas (data not shown). Clinical examples are shown in table 2, including results of the 10 selected formulas. Similar results were observed with the remaining formulas. CONCLUSIONS: In conclusion, we observed that creatinine and cystatin-C formulas have a wide error in reflecting mGFR and do not allow the correct assessment of renal function in diabetics. Caution is needed in the use and interpretation of formulas in this population at risk of CKD (Table presented).