The whole blood pharmacokinetics of intravenously administered 99mTc-disofenin (DISIDA) has been studied in normal subjects and patients with documented liver disease. The apparent overall whole blood disposition rates of radioactivity were calculated from serial blood data, in order to evaluate liver clearance of DISIDA. The measurements obtained clearly discriminated 9 normal subjects from 7 patients with severe liver disease causing jaundice--1233 mls/min vs 384 mls/min (P less than 0.002). Nine subjects with liver disease of insufficient severity to cause jaundice also had clearly abnormal DISIDA disposition--642 ml/min (P less than 0.05 for difference to controls). The time activity curves from all subjects showed biexponential elimination of blood activity, with a rapid (T1/2 = 3.8 min) and a slow disposition phase (T1/2 = 75 min) in normals. These curves were fitted by computer to the timed rate of hepatic uptake, simultaneously obtained by gamma imaging over the liver. It was not possible to satisfactorily fit these using a model which assumed distribution of a single compound within two body compartments. However, another which assumed the administration of two radioactive agents satisfactorily fitted the two types of data. This conclusion is consistent with our animal experiments which indicate the existence of two compounds in injected DISIDA with contrasting high and low hepatic extraction efficiency (Fraser et al. 1988). A pharmacokinetic approach to DISIDA disposition can yield quantitative information which discriminates different degrees of liver dysfunction, but the mechanisms involved are more complicated than previously thought, so that further study should permit very precise quantification.