The present prospective study was undertaken to evaluate the usefulness of urinary chloride concentration in determining the cause of an abrupt decline in renal function. 99 patients from diverse clinical settings with multiple causes of acute renal failure were evaluated. Urinary chloride concentrations of less than 20 mEq/l were observed in most cases of reversible prerenal azotemia (20 of 21 cases) and were observed in more frequently than urinary sodium concentration of less than 20 mEq/l (13 of 21 cases, p less than 0.01). Only prerenal azotemia accompanying diuretic use was associated with high urinary chloride concentrations (57 +/- 7 mEq/l). When prerenal azotemia occurred in the setting of metabolic alkalosis with bicarbonaturia, urinary chloride was low (4.0 +/- 1.0 mEq/l) while urinary sodium was high (65.0 +/- 19.0 mEq/l). In patients with oliguric and nonoliguric acute tubular necrosis, and in patients with acute exacerbations of chronic renal failure, mean urinary chloride concentration ranged from 40 to 67 mEq/l and mean fractional excretions of chloride ranged from 7.2 to 8.4%. Only 11% of patients with oliguric and nonoliguric acute tubular necrosis had urinary chloride concentrations of less than 20 mEq/l. Urinary chloride concentrations exhibited greater sensitivity and equivalent specificity as urinary sodium concentrations in differentiating patients with reversible prerenal azotemia from those with oliguric and nonoliguric acute tubular necrosis.