Background Cell adhesion molecules (CAMs) have been implicated in cardiac allograft rejection. However, previous studies have used qualitative analysis of immunohistochemical data and did not exclude patients with infection or malignancy.Methods and Results We analyzed 40 endomyocardial biopsy specimens from 25 cardiac transplant patients and 8 specimens from patients undergoing cardiac surgery. Patients with evidence of infection or malignancy were excluded. Specimens were stained with monoclonal antibodies against ICAM-1, E-selectin, VCAM-1, and PECAM-1 (which labels all vessels). ICAM-1 expression was assessed by counting ICAM-1-positive vessels and dividing by the total number of vessels (measured by PECAM staining). Specimens were scored as positive or negative for VCAM-1 and E-selectin. We also determined whether serum-soluble ICAM-1 levels (sICAM) correlated with rejection by evaluating 145 serum specimens from 48 cardiac transplant patients and 8 specimens from patients undergoing diagnostic cardiac catheterization. ICAM-1 was present on 50% to 60% of vessels in normal and nonrejecting specimens. Specimens with histologically significant rejection (focal moderate, moderate, or severe) had an increased percentage of ICAM-1-positive vessels: focal moderate, 77%; moderate/severe, 92% (P<.01). E-selectin expression did not differ between groups. VCAM-1 frequently was not present on rejecting specimens. No correlation was noted between sICAM levels and the presence or absence of rejection.Conclusions (1) ICAM-1 expression is strongly correlated with histologically significant cardiac allograft rejection. (2) The use of PECAM-1 staining as a vascular marker permits quantitative analysis of ICAM-1 expression. (3) VCAM-1 and E-selectin are not consistently increased during cardiac allograft rejection. (4) sICAM levels do not accurately reflect endomyocardial biopsy results. (Circulation. 1994;89:1760-1768.)