To investigate the role of aldosterone in sodium retention and ascites in cirrhosis, the urinary sodium excretion, sodium balance and urinary excretion of aldosterone-18-glucuronide (UAidV) were serially measured in 11 rats undergoing cirrhosis induction with carbon tetrachloride (CT) and phenobarbital (CT rats) and in 10 control rats which received phenobarbital. All CT rats developed ascites, seven within the ninth week after starting the program and four within the 10th week. One week before the onset of ascites, CT rats and control rats were different with respect to sodium excretion (1.41 ± 0.15 vs. 1.82 ± 0.1 mEq per day), sodium balance (0.57 ± 0.12 vs. 0.20 ± 0.09 mEq per day) and UAidV (67.8 ± 9.5 vs. 25.7 ± 1.7 ng per day). These differences were more pronounced within the week in which ascites was detected in CT rats. Before these 2 weeks, both groups did not differ with respect to these parameters. In the 132 urine samples obtained in CT rats, there was a correlation between sodium excretion and UAidV (r = -0.53; p < 0.001). Twenty-one additional CT rats were divided into two groups. Eleven animals were given spironolactone (20 mg per day s.c. in olive oil) from the 6th week, and 10 only received olive oil. Thirteen weeks after starting the program, all rats not treated with spironolactone had sodium retention and ascites (in five rats, ascites appeared within the ninth week and in five within the tenth week); this occurred in only one animal treated with spironolactone. This difference could not be due to a protective effect of spironolactone against carbon tetrachloride hepatotoxicity since both groups of rats were similar with respect to portal pressure and severity of hepatic lesions. These results are consistent with the concept that aldosterone plays an important but not the sole role in the pathogenesis of ascites in cirrhosis.