A cohort of 78 infants of gestational age less than 34 weeks was examined for Ureaplasma urealyticum colonization and neonatal morbidity. Ureaplasma urealyticum was cultured from nasopharyngeal, endotracheal and blood-culture samples. A child was considered as being colonized if any sample was positive. The children with perinatal U. urealyticum colonization (n = 11; 14%) differed from those with no colonization (n = 67) in two important aspects: (i) they had higher leucocyte counts on the first (18.6 vs 12.4 10(9)) and the second (29.0 vs 15.4 10(9)) days of life (p = 0.01, both days); and (ii) they more often needed high-frequency oscillatory ventilation (45% vs 13%, p = 0.02). This study showed that U. urealyticum colonization is associated with signs of the host defence response together with symptoms of respiratory tract involvement suggesting the pathogenicity of U. urealyticum in premature infants.