The effect of leukocyte antibodies detected under different conditions on the fate in vivo of granulocytes was studied using 111-indium-labeled granulocytes. Sera from patients were tested by granulocyte agglutination (GA), granulocytotoxicity (GC), granulocyte immunofluorescence (GIF), lymphocytotoxicity (LC), and antibody-dependent lymphocyte-mediated granulocytotoxicity. Granulocytes from donors to be studied were labeled with 111-indium and injected. Then the intravascular recovery and survival or tissue localization was determined in 93 studies. Antibodies detected by granulocyte agglutination were associated with a significant reduction in recovery (6.7% v 30.8% in controls; P < .001) and T 1/2 (0.3 hours v 5.6 hours in control; P = .002). When all possible combinations of serum reactivity were considered, reactivity in the GA plus GIF assays had the best correlation with decreased recovery (R2 = .49; P < .001) and T 1/2 (R2 = .73; P < .001). When the relationship between the strength of antibody reactivity and the recovery and t 1/2 were analyzed, the best relationship was between the combination of LC and GIF with recovery (R2 = .62; P = .001). Because of the general availability of the HLA (LC) testing, the role of LC reactivity was investigated in other ways. There was a strong relationship between sera highly reactive by LC and those reactive by GIF. These highly reactive sera were also associated with reduced recovery and t 1/2 . The influence of specific HLA antigen mismatches was also studied. When donor and recipient were mismatched for the HLA-A2, B8, or BW44 antigens, there was a significant reduction in either recovery, t 1/2 , or both. Tissue localization was studied by body scans in patients with and without known sites of inflammation. Antibodies detected by a combination of GA and GIF caused abnormal pulmonary sequestration of granulocytes (three cases) and failure of granulocytes to localize at known sites of inflammation (three cases). HLA (LC) antibodies did not alter tissue localization despite the presence of the corresponding HLA antigens on granulocytes. It appears that GA, GIF, or a combination of these tests is the most effective predictor of altered in vivo fate of granulocytes. However, sera highly reactive by LC and GIF probably define a group of highly immunized patients in whom granulocyte recovery and t 1/2 are also reduced. Mismatching for certain HLA antigens is also associated with reduced granulocyte recovery and survival. At present, GA with or without the immunofluorescence assay, is the most effective predictor of altered in vivo granulocyte activity. LC, if interpreted in a specific manner, may also define a group of highly immunized patients in whom in vivo granulocyte activity is altered.