Loop and distal diuretics are the basic drugs for the treatment of ascites. Although pharmacologic studies indicate that the natriuretic potency of loop diuretics is much greater than that of distal diuretics, there are no studies comparing the efficacy of these drugs in cirrhosis. Forty nonazotemic cirrhotic patients with ascites and avid sodium retention were randomly allocated into two groups. Group 1 contained 21 patients treated with furosemide; group 2 contained 19 patients treated with spironolactone. The initial doses were 80 and 150 mg/day, respectively. These doses were increased to 160 and 300 mg/day, respectively, if there was no response. Cases not responding to furosemide and spironolactone were later treated with spironolactone and furosemide, respectively. In group 1, 11 of the 21 patients responded to furosemide, while in group 2, 18 of the 19 patients responded to spironolactone (p less than 0.01). Of the 10 patients in group 1 not responding to furosemide, 9 responded later to spironolactone. The diuretic response to furosemide and spironolactone was related to the activity of the renin-aldosterone system. Patients with higher renin and aldosterone did not respond to furosemide and required 300 mg/day of spironolactone to achieve a diuretic response. These results indicate that (a) at the dosages used in the study, spironolactone is more effective than furosemide in nonazotemic cirrhosis with ascites, and (b) the activity of the renin-aldosterone system influences the diuretic response to furosemide and spironolactone in these patients.