摘要Background The equations for estimating glomerular filtration rate (GFR) based on creatinine have been found to have limitations and have not been generalizable across all populations. Equations based on cystatin C provide an alternative method to estimate GFR. Whether the equation based on cystatin C alone or combined creatinine would improve GFR estimates has not been validated among Chinese patients with chronic kidney disease (CKD) and diabetes. The aim of this study was to compare the performance of the modification of diet in renal disease (MDRD) equation based on creatinine with the five cystatin C-based formulae for estimation of GFR in patients with CKD and diabetes. Methods A total of 166 patients with CKD and 91 patients with type 2 diabetes were enrolled in this study. Cystatin C was measured by using the particle-enhanced immunonephelometric method and estimated formulae proposed by five different investigator teams (Stevens, Ma, Rule, Macisaac and Perkins). The plasma clearance of ~(99m)Tc-DTPA was determined as measured GFR (mGFR).Results For CKD patients, the bias and accuracy for the Ma and Macisaac equations were superior compared with the MDRD, and the mean results for the Ma formula were closer to mGFR than the other equations in CKD stages 2-5. The differences between Macisaac and mGFR in CKD stages 2-4 were significantly less than those in CKD stage 1 or 5.Stevens and Rule's formulae revealed a similar bias and accuracy compared with the MDRD equation. The MDRD formula had a higher accuracy in CKD stages 3-5 as compared with the results in other stages. For diabetic patients, the mean results between Macisaac and mGFR were closer than those of other equations in mGFR≥90 ml.min~(-1).1.73 m~(-2) stage, in GFR 60-89 ml·min~(-1)·1.73 m~(-2) stage, the MDRD formula showed the smallest difference compared with other equations. All equations overestimated GFR in the cases with GFR <60 mi.min~(-1)·1.73 m~(-2) stages. The MDRD formula had a greater accuracy within 50% of mGFR than the equations based on cystatin C in diabetic patients. Perkins formula showed a large positive bias and low accuracy, therefore it may not be suitable for assessing GFR in patients with CKD and diabetes. Conclusions The formulae for estimating GFR based on cystatin C or creatinine have different trends and accuracies in patients with CKD and diabetes, especially in patients with various GFR levels. The equations based on cystatin C provide less accurate results than MDRD formulae, at least in the diabetic patients. Therefore, whether the formulae based on cystatin C are superior to MDRD formula requires further investigation in large diverse populations.