At the age of 5 years, the prevalence of atopic manifestations was analysed in 58 formula-fed "at risk" infants because of a history of atopic disease in at least two first degree relatives. Infants were randomly assigned to receive either a partial whey-hydrolysate formula (n: 28) or a regular cow's milk formula (n: 30) during the first 6 months of life; thereafter, feeding was unrestricted. Only non-breastfed infants were included. The groups did not differ in risk factors or in known confounding factors possibly influencing the incidence of manifestations suggestive of atopic disease. At 6 months, the prevalence of cow's milk protein (CMP) sensitivity was significantly decreased in the hydrolysate group (7% versus 43%; P: 0.002). At the age of 12 (21% versus 53%; P: 0.029), 36 (25% versus 57%; P: 0.018) and 60 months (29% versus 60%; P: 0.016) there was still a significant difference in the number of atopic manifestations, if calculated cumulatively. There was no difference between the groups if only the new cases after the age of 6 months were considered. Eczema was less frequent in the whey-hydrolysate group, but only during the 1st year of life, suggesting a decreased prevalence of CMP sensitivity. During the first 6 months, diarrhoea of non-infectious origin occurred in 8/30 infants (27%) of the adapted formula group, and in no infant in the hydrolysate group. "Colic as single manifestation" was considered of "allergic" origin in 1/28 infants in the hydrolysate group, and in 4/30 infants in the adapted formula group.(ABSTRACT TRUNCATED AT 250 WORDS)