In a prospective, randomised study the effects of orally administered bifidobacteria on the intestinal microflora were investigated in 100 preterm and term neonates under intensive care conditions during the first 21 days of life. The 50 infants (group with bifidobacteria) received lyophilized bifidobacteria (Töpfer Bifidus) via nasogastral tube with an initial dosage of 3 times daily 1.25 x 10(8) bifidobacteria on day 2 of life and a daily dosage of 6 times 1.25 x 10(8) bifidobacteria on day 3 until day 21 of life. The other 50 infants (control group) did not receive bifidobacteria. The preterm and term neonates were fed either with pasteurized mother's milk or milk from healthy female donors (n = 79) or with an infant formula (Alfaré, n = 13) or initially with Alfaré and thereafter with mother's milk (n = 8). The intestinal microflora of preterm and term neonates under intensive care conditions could be influenced by the oral administration of bifidobacteria. The administration of bifidobacteria resulted in the group of inoculated infants in a significantly earlier colonization of bifidobacteria (8.1 +/- 3.9 days of life) than in the control group (11.3 +/- 4.7 days of life). On day 7 a bifidobacterial dominance (> 90% of the intestinal microflora) could be found in 26% of infants with inoculation of bifidobacteria and only in 2% of the control group (p < 0.001). These significant differences could be shown until day 21 of life. A difference in septicemia frequency between the two groups could not be demonstrated. At the beginning of the infection a bifidobacterial dominance was found in only one of 23 cases of septicemia.