In human medicine drug allergy is a well-established side-effect of the therapeutic use of antibiotics, especially the beta-lactams. Side-effects caused by macrolides are uncommon and only a very few of these seem to be caused by allergic mechanisms. Clinically, drug allergy is characterized by a spectrum of reactions ranging from mild skin rashes to angio-oedema or life-threatening anaphylaxis. Concern has been expressed that antibiotic residues in meat and other foods might be responsible for similar hypersensitivity reactions in a small number of individuals. This review assesses the potential risk of such reactions in general, but focuses on allergy to penicillin and macrolide residues in particular. In relation to the risk of primary sensitization, it is unlikely that residues could contribute to the overall immune response in view of the very low levels that are likely to be encountered in comparison with the high levels received during therapeutic use. No evidence has been found that any individual has become sensitized by residues of either penicillins or macrolides. Furthermore, the oral route is much less sensitizing than parenteral administration and immunochemical studies with penicillin indicate that hapten-protein complexes formed in vivo are unlikely to be immunogenic because of their low dose, low epitope density and binding to autologous carrier proteins. For performed allergens, the epitope density was also too low to be immunogenic. Because of the ubiquitous nature of penicillin-producing moulds in nature and the extensive use of beta-lactam antibiotics in human medicine, it is unlikely that epidemiological studies could be undertaken that could allow quantification of the minimal risk. The risk of allergic reactions in pre-sensitized individuals can be assessed similarly and again it is concluded that factors such as dose, oral administration and low epitope density make it unlikely that a significantly antigenic derivative could be formed. However, a review of the literature on penicillin hypersensitivity revealed a very small number of previously sensitized individuals from whom there is reasonable clinical and documentary evidence that penicillin residues in milk triggered an allergic reaction, usually a rash. Although these cases are very rare (less than 10 cases reported in the last 25 years), they illustrate the continuing need to control antibiotic residues vigilantly. Animal models have not proved useful for predicting the risk of hypersensitivity reactions to drugs, since allergy in man is determined by genetic and other factors and no validated methods exist to determine a no-effect level.(ABSTRACT TRUNCATED AT 400 WORDS)