We investigated whether determination in fine-needle aspiration biopsy (FNAB) specimens of cells expressing granzymes (Grs) and Fas ligand would provide a reliable, easy, and quantitative measure of rejection activity in the transplanted liver. Retrospectively, 13 FNAB specimens obtained during clinical acute rejection, 10 FNAB specimens obtained during subclinical rejection, 12 FNAB specimens obtained during cytomegalovirus (CMV) infection, and 26 FNAB specimens obtained in the absence of rejection or infection were included on the study. Cytospin preparations of FNAB and peripheral-blood specimens were immunocytochemically stained for Fas-ligand and Gr, and increments in the liver were calculated by subtracting frequencies of positive cells in blood from those in FNAB specimens. Only sporadically Fas ligand-expressing, but many Gr-expressing, cells were detected in FNAB specimens. Increments in Gr-positive (Gr(+)) cells were significantly greater in FNAB specimens obtained during clinical rejection (median, 70 Gr(+) cells; range, 0 to 312 Gr(+) cells; P = .006) and tended to be greater in FNAB specimens obtained during subclinical rejection (median, 62 Gr(+) cells; range, 5 to 113 Gr(+) cells; P = .09) compared with those obtained in the absence of rejection (median, 16 Gr(+) cells; range, 0 to 103 Gr(+) cells). Increments obtained during clinical or subclinical rejection did not differ from those obtained during CMV infection (median, 27 Gr(+) cells; range, 6 to 212 Gr(+) cells). With the exclusion of specimens obtained during CMV infection, the sensitivity of Gr determination in FNAB specimens for the diagnosis of acute rejection (either clinical or subclinical) was 70%, and specificity, 69%. In FNAB specimens obtained during clinical and subclinical acute rejection episodes after liver transplantation, increased numbers of Gr-expressing cells were present; in the absence of CMV infection, their quantification provides a measure for rejection activity with moderate accuracy.