This study forms a part of an investigation into the extent to which the type of renal damage influences the renal clearance of drugs. We have already demonstrated an effect of different types of experimental renal failure (ERF) on the renal clearance of two cations: cimetidine, a drug that is filtered and secreted by the nephron, and lithium, which is filtered and reabsorbed by more than one segment of the nephron. In this report the renal clearance of cephalexin (CLCEX) is investigated, a drug that has a different mode of renal elimination, since it is filtered, secreted, and reabsorbed by the proximal tubules. The aim was to extend our earlier studies to an organic anion, and to provide an opportunity to evaluate the feasibility of using the renal clearance of N-1-methylnicotinamide (NMN) to predict the renal clearance of anionic drugs in renal failure. Different models of site-specific ERF have been developed in the rat; proximal tubular necrosis (induced by cisplatin), papillary necrosis (induced by 2-bromoethylamine), and glomerulonephritis (induced by sodium aurothiomalate or by antiglomerular basement membrane antibodies). Glomerular function (GFR) was assessed by the clearance of inulin (CLNULIN), and tubular function was assessed by the clearance of endogenous NMN (CLNMN). OUr results show that even if the models of ERF used were not absolutely site-specific, glomerular function and tubular function did not decrease to the same extent in the different ERF. Therefore, glomerulo-tubular imbalance existed, which is incompatible with the "intact nephron hypothesis," i.e., the site of the damage along the nephron and not only the degree of renal dysfunction, is a potential source of variability in the clearance of certain drugs. The renal clearance of cephalexin was estimated more accurately by CLNMN than GFR (r = 0.90). We conclude that the clearance of the endogenous cation NMN can be used to predict the renal clearance of drugs that are not only filtered by the glomeruli but also secreted and/or reabsorbed by the proximal tubules, and in the limited examples investigated appears to apply to both anionic and cationic compounds. In this respect the GFR alone was not an adequate parameter for the prediction of the renal clearance of such drugs.