We evaluated the role of circulating intercellular adhesion molecule-1 in the pathogenesis of hepatolithiasis. From December 1994 to May 1995, 40 patients with hepatolithiasis were included. All the patients met the following criteria: (1) presence of hepatolithiasis, (2) no obvious clinical evidence of an associated intrahepatic cholangiocarcinoma, (3) no clinical manifestation of cholangitis for at least 72 hr, (4) no immunomodulatory agents in the last three weeks, and (5) no blood transfusion in the last three weeks. Venous blood samples were collected both before surgery and at least three months after complete clearance of the stones, and the serum concentrations of circulating intercellular adhesion molecule-1 were measured with a sandwich enzyme immunoassay method. Fifteen healthy subjects were used as a control group. Bile specimens routinely obtained during surgery were cultured for aerobes and anaerobes. The x-ray films of cholangiography were all reviewed in detail. The mean value (834 +/- 128 ng/ml) of circulating intercellular adhesion molecule-1 (ICAM-1) in the patient group before surgery was significantly higher than that (346 +/- 68 ng/ml) of the control group (P < 0.01). The mean value (677 +/- 139 ng/ml) of circulating ICAM-1 in the patient group at least three months after complete clearance of the stones was significantly lower than that (834 +/- 128 ng/ml) of the patients before surgery (P < 0.01), but this mean value (677 +/- 139 ng/ml) was still significantly higher than that (346 +/- 68 ng/ml) of the control group (P < 0.01). Bacteria was present in the bile of all patients. The total number of bacterial species was 135, and there were an average of 3.4 bacterial species cultured per patient. Intrahepatic stricture was demonstrated in cholangiography in 33 patients (82.5%). In addition to the high incidence of intrahepatic bile duct strictures and bile infection, a significant elevation in circulating intercellular adhesion molecule-1 (sICAM-1) was shown in patients with hepatolithiasis. Our preliminary results seem to be promising and the real role of sICAM-1 deserves further investigation and elucidation.