Hypoallergenic formulas are processed by enzymatic hydrolysis of different protein sources such as bovine casein/whey and soy followed by further processing such as heat treatment and/or ultrafiltration, or they are based on amino acid mixtures. The products have been classified according to the degree of protein hydrolysis as 'extensively' or 'partially' hydrolysed protein products. Product properties may be characterized by biochemical techniques, and reduction of allergenicity may be assessed in vitro with various immunological methods, and in vivo with skin prick tests, patch tests and challenge tests. In vitro tests do not predict the allergenic effects in humans, and at present there is no evidence of a specific threshold of immunogenic protein. Only pure amino acid mixtures are considered to be nonallergenic. Other 'hypoallergenic' products contain residual allergenicity. At present, the potential of a product for treatment and prevention of food allergy can only be determined by clinical trials using scientifically appropriate standards. It has been recommended that dietary products for treatment of cow's milk protein allergy in infants should be tolerated by at least 90% (with 95% confidence) of infants with documented cow's milk protein allergy. Some extensively hydrolysed products and amino-acid-based products have met these criteria. Formulas intended for prevention should have a very low, if any, allergenic activity until otherwise proven. So far there are no firm criteria available for the design of hypoallergenic foods for prevention. Newborns included in prevention studies should be from high-risk families; they should be randomized at birth and fed the formula when supplements are needed for at least the first 4-6 months of life. Follow-up should be at least 18 months, and children should be investigated when symptoms appear. Validated clinical criteria, including controlled food challenges, should be used for diagnosis. Infants fed formulas that claim to prevent or delay allergy should have a statistically significant lower prevalence of allergy when compared with infants fed a standard cow's milk formula. Feeding high-risk infants a documented hypoallergenic formula combined with avoidance of solid foods during the first 4-6 months reduces the cumulative incidence of cow's milk protein allergy and atopic dermatitis as compared with a standard cow's-milk-based formula. Partially hydrolysed formulas may have an effect, though it seems to be less than that of extensively hydrolysed formulas at present. Thus, if exclusive breast-feeding for 4-6 months is not possible in high-risk infants, a documented hypoallergenic formula and avoidance of solid foods are recommended for the first 4 months of life.