Although circulating levels of soluble intercellular adhesion molecule-1 (sICAM-1) are frequently used as an indicator of the severity of different immune, inflammatory, or neoplastic diseases, little is known about the factors that govern plasma sICAM-1 concentration and its relationship to the membranous form of ICAM-1 (mICAM-1) expressed on vascular endothelial cells. Plasma sICAM-1 concentration (measured by enzyme-linked immunosorbent assay) and mICAM-1 expression (measured using the dual radiolabeled monoclonal antibody technique) in different vascular beds (eg, lung, small intestine, and spleen) were monitored in wild-type (C57BL) and ICAM-1-deficient mice, before and after administration of tumor necrosis factor (TNF)-alpha. In wild-type mice, TNF-alpha elicited time-dependent increases in lung and intestine mICAM-1 (plateau achieved at 12 hours), with a corresponding increase in plasma sICAM-1 (peaked at 5 hours and then declined). The initial increases in mICAM-1 and pulmonary leukocyte sequestration (measured as lung myeloperoxidase activity) induced by TNF-alpha preceded any detectable elevation in sICAM-1. In ICAM-1-deficient mice, plasma sICAM-1 was reduced by approximately 70%, with > 95% reductions of mICAM-1 in lung and intestine, and > 75% reduction in splenic accumulation of anti-ICAM-1 antibody. Although TNF-alpha doubled plasma sICAM-1 in ICAM-1-deficient mice, mICAM-1 was unaffected in all tissues. Either splenectomy or pretreatment with cycloheximide resulted in an attenuated TNF-induced increase in sICAM-1, without affecting mICAM-1 expression. These findings indicate that plasma sICAM-1 concentration does not accurately reflect the level of ICAM-1 expression on endothelial cells in different vascular beds.