Incidence and clinical manifestation of cow's milk protein intolerance (CMPI) were studied in 1158 unselected newborn infants followed prospectively from birth to 1 year of age. No food changes were required in 914 infants who were used as healthy controls. When CMPI was suspected (211 infants), diagnostic dietary interventions according to a standard protocol were performed. After exclusion of lactose intolerance, two positive cow's milk elimination/challenge tests were considered diagnostic of CMPI. Two hundred and eleven symptomatic infants were examined for possible CMPI. A large group of 80 infants improved on a lactose reduced formula. In 87/211 infants CMPI was excluded (sick controls). Finally CMPI was proven in 26 infants. The calculated incidence rate for CMPI was 2.8%. The principal symptoms in infants with CMPI were gastrointestinal, dermatological and respiratory in 50%, 31% and 19% respectively. A positive family history for atopy (first or second degree relatives) was more frequent in either CMPI infants (65%), or sick controls (63%) when compared to either healthy controls (35%) or infants improving on a low lactose formula (51%). Differences between patients with CMPI and sick controls were only found for the presence of atopy in at least 2 first degree relatives [(5/26 in CMPI infants and 4/87 in sick controls (P < 0.05)] and for multiorgan involvement [10/26 infants with CMPI as opposed to 12/87 in the sick control group (P < 0.02)]. These statistical differences are too weak to be of clinical value.