Comparisons in the same patients of platelet-bound anti-platelet antibody (APA) levels (direct test) with serum APA levels (indirect test) frequently do not give the same results. Indirect test results frequently are negative or marginally elevated even though platelet-bound antibody is greatly increased on direct testing. The most likely cause for these differences between the two tests is insufficient binding of serum antibody to test platelets in vitro. In an effort to enhance in vitro antibody binding, we examined four test modifications affecting platelet attachment. Rabbit anti-human platelet antisera (absorbed with their specific human donor leukocytes to remove HLA- and granulocyte-specific antibodies) were combined under various conditions with human platelets and tested in a modified antiglobulin consumption test utilizing rabbit IgG. The test conditions varied: incubation time of the serum APA-platelet mixture at 37 degrees C; age of the test platelet pool; ionic strength of the mixture (varied through incorporation of a low-ionic-strength solution (LISS); and vortex agitation for 30 seconds prior to the addition of sera. Optimal attachment in standard phosphate-buffered saline took place in 60 minutes using test platelets stored for either 3 or 30 days, and yielded mean values for rabbit immune sera of 9.9 fg of IgG per platelet. Preincubation controls yielded a mean of 0.7 to 1.5 fg per platelet. Incorporation of LISS or vortexing of test platelets did not affect the results of control, but both modifications substantially increased test values of immune sera to 15.8 and 17.3 fg of IgG per platelet, respectively. Combining LISS with vortexing did not further increase values.