Reproducible clinically abnormal reactions to cow's milk protein (CMP) may be due to the interaction between one or more milk proteins and one or more immune mechanisms, possibly any of the four basic types of hypersensitivity reactions. At present, evidence for type I, III and IV reactions against CMP has been demonstrated. Immunologically mediated reactions, mainly immediate IgE-mediated reactions are defined as cow's milk protein allergy (CMPA). Non immunologically reactions against CMP are defined as cow's milk protein intolerance (CMPI). Many studies on "cow's milk allergy'" have not investigated the immunological basis of the clinical reactions. It is not possible to differentiate between CMPA and CMPI solely on clinical symptoms. No single laboratory test is diagnostic of CMPA/CMPI. Therefore, the diagnosis still has to be based on strict well-defined elimination and milk challenge procedures. Before 1950 CMPA/CMPI was rarely diagnosed. Since 1970 widely varying estimates of the incidence from 1.8% to 7.5% have been reported, mainly reflecting differences in diagnostic criteria and study design. Based on strict diagnostic criteria the incidence of confirmed CMPA/CMPI in infancy seems to be about 2-5% in developed countries. Symptoms suggestive of CMPA/CMPI may be encountered in about 5-15% of infants emphasizing the importance of controlled elimination/milk challenge. In breastfed infants reproducible clinical reactions to CMP in human milk have been reported in about 0.5%. Most infants with CMPA/CMPI develop symptoms before one month of age, often within one week after introduction of cow's milk based formula. The majority have > or = 2 symptoms and symptoms from > or = 2 organ systems. About 50%-70% have cutaneous symptoms, 50-60% gastrointestinal symptoms, and about 20-30% respiratory symptoms. In exclusively breast-fed infants with CMPA/CMPI severe atopic eczema is a predominant symptom. Debut of CMPA/CMPI after 12 months is extremely rare. The basic treatment is complete avoidance of CMP. In infancy a proven hypoallergenic CM substitute is needed. Due to clinically important residual allergenicity in some hypoallergenic formulae controlled clinical testing is necessary in each case before use. Goat's milk proteins share identity with CMP Raw untreated cow's milk and unhomogenized cow's milk is as allergenic as normal pasteurized and homogenized milk products. The prognosis of CMPA/CMPI is good with a remission rate about 45-50% at one year, 60-75% at two years, and 85-90% at three years. Associated adverse reactions to other foods develop in about 50%, and allergy against inhalants in 50-80% before puberty.(ABSTRACT TRUNCATED AT 400 WORDS)