Indexing glomerular filtration rate (GFR) for body surface area (BSA) is often realized without arising any questions. However, physiological basis for such an indexation are very poor. Indeed, indexing GFR for BSA to avoid variation due to differences in body size necessarily implies that GFR is a linear function of BSA and that the intercept of this linear function is zero. Moreover, when GFR is indexed for BSA, the relation indexed GFR-BSA must completely disappear. These physiological prerequisites are not found for BSA indexation. We will review the history of this indexation and will underline errors and defaults. Different equations to estimate BSA exist and will be discussed. The choice of "1.73m(2)" will be also criticized. Moreover, indexing GFR for BSA has little impact on GFR results in "normal" body size population. Nevertheless, this indexation will have strong consequences in very lean (such as anorexia) and in obese patients. We will discuss possible alternatives proposed to substitute for BSA indexation.