Accurate renal function measurements are important for the diagnosis and treatment of kidney disease, proper medication dosing, interpretation of possible uremic symptoms, and decision-making regarding when to initiate renal replacement therapy. Because the use of highly accurate filtration markers to measure renal function has traditionally been limited by cumbersome and costly techniques and the involvement of radioactivity (among other factors), renal function is typically estimated by using specially derived prediction equations. These formulae usually use serum creatinine levels, i.e., a marker of filtration that is insensitive to mild/moderate decreases in GFR. Although attempts have been made to validate certain renal function prediction equations among patients with chronic kidney disease (CKD) with abnormal serum creatinine levels, this is the first study to specifically evaluate the predictive performance of these equations for patients with CKD and serum creatinine levels in the normal range. The results of eight prediction equations for 109 patients with CKD and serum creatinine levels of < or =1.5 mg/dl were compared with standard iohexol GFR values. The most accurate results were obtained with the Cockroft-Gault and Bjornsson equations. The most precise formulae were the Modification of Diet in Renal Disease Study equations, although they were highly biased. Even the most accurate results exhibited levels of error that made them suboptimal for clinical treatment of these patients. These results suggest that measurement of GFR with endogenous or exogenous filtration markers might be the most prudent strategy for the assessment of renal function in the CKD population with normal serum creatinine levels. Further studies are needed to confirm the generalizability of these findings for this patient subgroup.