Researchers compared August 1987 and September 1991 data on 281 HIV-positive blood donors, 16-59 years old, with those of 1076 randomly selected controls to determine factors associated with HIV infections they could develop simple and inexpensive deferral criteria to prevent donations from high-risk persons who may transmit HIV during the seronegative window period. The blood bank records were at Siriraj Hospital in Thailand. HIV-positive donors were significantly more likely to be men than were controls (98.9% vs. 76.3%; odds ratio = 2.64). They were also more likely to test positive for syphilis (VDRL test) (8.2% vs. 2.3%; OR = 3). A higher proportion of HIV-positive blood donors were 21-30 years old than were controls (66.2% vs. 40.5%; OR = 2.2). A greater percentage of blood donors who were asked by friends or relatives of transfused patients to donate blood (i.e., replacement donors) were HIV positive than controls (55.5% vs. 40.4%; (OR = 1.5). All these independent risk factors were significant at the 5% level. Paid donors were less likely to be HIV positive than controls (2.8% vs. 9.9%; odds ratio [OR] = 0.6). Since 40.4% of all HIV-negative donors were replacement donors, the researchers decided not to include replacement donorship in the analysis for potential deferral criteria. They used the records of 214 HIV-positive donors with all 3 other risk factors in 1990 and 1991 to determine the potential deferral criteria. The probabilities of HIV positivity of blood donors with 2 and 3 risk factors were 1 and 2.2% (respectively), which meant that they were 4.9 and 4.1 times (respectively) more likely to be HIV positive than were other donors. Using these criteria would have eliminated 1.5% and 31.2% (respectively) of all HIV-negative donors. Elimination of 31.2% is too high to be acceptable. The researchers suggest an additional p24 antigen testing for donors at high risk for HIV infection, particularly 21-30 year old men.