Chronic kidney disease (CKD) is generally regarded as a final common pathway of several renal diseases, often leading to end-stage kidney disease (ESKD) and a need for renal replacement therapy. Estimated GFR (eGFR) has been used to predict this outcome recognizing its robust association with renal disease progression and the eventual need for dialysis in large, mainly cross-sectional epidemiological studies. However, GFR is implicitly limited as follows: (1) GFR reflects only one of the many physiological functions of the kidney; (2) it is dependent on several non-renal factors; (3) it has intrinsic variability that is a function of dietary intake, fluid and cardiovascular status, and blood pressure especially with impaired autoregulation or medication use; (4) it has been shown to change with age with a unique non-linear pattern; and (5) eGFR may not correlate with GFR in certain conditions and disease states. Yet, many clinicians, especially our non-nephrologist colleagues, tend to regard eGFR obtained from a simple laboratory test as both a valid reflection of renal function and a reliable diagnostic tool in establishing the diagnosis of CKD. What is the validity of these beliefs? This review will critically reassess the limitations of such single-focused attention, with a particular focus on inter-individual variability. What does science actually tell us about the usefulness of eGFR in diagnosing CKD?