1. Albumin is normally excreted as a mixture of intact protein and fragments that are produced during renal passage. The purpose of this study was to investigate the ratio of intact versus degraded forms of excreted albumin to ascertain whether changes in this ratio could account for the apparent increase in albumin excretion seen in diabetes, as measured by standard radioimmunoassay techniques. 2. Four-week male Sprague-Dawley rats with streptozotocin-induced diabetes and age-matched control rats were intravenously injected with [3H]albumin. Urine collected over 2 h was analysed by size exclusion chromatography and radioimmunoassay. A standard radioimmunoassay found a 7-fold increase in albumin excretion rate in diabetic rats, whereas there was only a 2-fold increase in albumin excretion (intact plus fragments). Urine analysed by size exclusion chromatography showed severe degradation for control rats (% monomer=4+/-2%); in diabetic rats there was a significant amount of monomer albumin excreted, along with moderately degraded and heavily degraded albumin (% monomer=17+/-5%). 3. This study has shown that the radioimmunoassay, which specifically detects intact albumin, considerably underestimates the amount of total urinary albumin which consists of intact and degraded material. The increase in albumin excretion rate observed in diabetes as measured by radioimmunoassay is mainly due to a change in the amount of intact albumin excreted and this is specifically due to the inhibition of albumin degradation at a post-glomerular site and not due to the onset of any type of glomerular 'shunt' pathway.