Two approaches have been used to produce universal group O donor red blood cells (RBCs) from groups A, B, and AB RBCs. The first involves cleavage of the terminal immunodominant sugars from carbohydrate chains on the RBC membrane, using specific enzymes, to produce so-called enzyme-converted group O (ECO) RBCs. ECO RBCs have been produced from whole units of B RBCs and transfused successfully to humans. Group A RBCs (especially A(1) RBCs) have been more difficult. New sources of enzymes have produced ECO RBCs from A(1) and A(2) RBCs that do not react with powerful monoclonal anti-A. Unfortunately, there are still problems encountered with polyclonal human antibodies (i.e. cross-matching). The second approach interferes with an antibody reaching its specific antigen on the RBC membrane by bonding polyethylene glycol (PEG) to the RBC. PEG will attract water molecules, yielding a combination that may block most RBC antigens, including A and B antigens. Initial excitement generated by preliminary reports of the possibility of producing 'stealth' PEG-RBCs were tempered by the findings of in vitro serological problems and possible reduced in vivo RBC survival. Many of these problems were solved, but recent findings that PEG is immunogenic in animals and humans, and that PEG antibodies can shorten the survival of PEG-RBCs (in rabbits) and pegylated proteins (e.g. PEG-asparaginase) in humans, are disturbing, suggesting that 'stealth' RBCs may never become a reality.