The value of monitoring the serum activity of SGOT, as well as markers of hepatitis B virus and hepatitis A virus infections, in the patients and staff of two dialysis units has been assessed retrospectively. Sera were checked each month for SGOT and HBsAg on 406 patients and 170 staff members over a 4-year period. Anti-HBc, anti-HBs, and anti-hepatitis A antibodies were assayed on the stored sera. Only 30% of the patients had normal SGOT values (less than 65 units per ml) on all occasions. Most of the abnormal values were less than 100 units per ml and could not be explained. Viral hepatitis was a reasonable explanation for only half of those instances where the SGOT value was greater than 100 units per ml. Hepatitis A virus contributed nothing to the problem of dialysis-associated liver disease. Testing for HBsAg alone missed approximately 40% of the hepatitis B events acquired in the unit. Only two of these episodes were of epidemiologic importance, however, because the rest were recognized only after there was serologic resolution of the infection. There was a high frequency of potentially "false positive" reactions with all the antibodies tested. It is not cost-effective to monitor dialysis patients and staff regularly with anti-HBc, anti-HBs, or antibodies against hepatitis A. Initial screening with anti-HBc and anti-HBs on entry to the unit is of value but weak positive results must be interpreted with caution since approximately half of such results will prove to be nonspecific.