1. The mechanism of renal protection by angiotensin-converting enzyme inhibition is still the subject of debate. Inhibition of proteinuria might play a role. If so, a good antiproteinuric response to angiotensin-converting enzyme inhibition should predict subsequent protection against renal structural damage. This hypothesis has not been tested in models where treatment is started after the renal disease is well established, i.e. models that mimic the clinical situation. 2. We therefore investigated this hypothesis in 96 male Wistar rats with established adriamycin nephrosis. Reduction of proteinuria was achieved by lisinopril (0, 2, 5 and 10 mg day-1 kg-1) on two different sodium diets (0.3% and 0.05% NaCl). Therapy started 6 weeks after adriamycin (at stable proteinuria) and was continued for 6 weeks. 3. Lisinopril reduced blood pressure by 32 +/- 4% and proteinuria by an average of 72 +/- 7%, with stabilization after 2 weeks. Considerable interindividual differences in antiproteinuric response was found. Glomerulosclerosis score was reduced by 15 +/- 5%. All the effects of angiotensin-converting enzyme inhibitors were enhanced by sodium depletion, but sodium depletion in itself did not affect blood pressure (124 +/- 4 mmHg), proteinuria (664 +/- 68 mg/day) or glomerulosclerosis score (30 +/- 5%). Interestingly, the more proteinuria was reduced initially in an individual rat, the less sclerosis was found in the long term in that rat. 4. In conclusion, angiotensin-converting enzyme inhibition lowers proteinuria and prevents glomerulosclerosis in established adriamycin nephrosis. These effects are enhanced by sodium depletion. The individual short-term antiproteinuric effect predicts the protection against ultimate glomerular damage. This is consistent with the hypothesis that reduction of proteinuria is a mechanism by which angiotensin-converting enzyme inhibitors exert renoprotection.