After uneventful use of cefotaxime and ceftazidime as first line therapy for three years in our neonatal intensive care unit we isolated cephalosporin-resistant Enterobacter cloacae (CREC) strains which caused clusters of cases or colonization and/or serious neonatal infection. By using two or more typing methods, at least five different strains with similar patterns of antimicrobial sensitivities were identified. The results of a case-control study did not support the notion that the use of third generation cephalosporins was associated with colonization and infection by CREC. The outbreak was brought under control by interrupting the transmission of the epidemic strain D, by measures such as cohort nursing, diligent handwashing before and after procedures, and through environmental cleaning as well as by decontamination with glutaraldehyde after dismantling of the blood gas analyser believed to have acted as a persistent reservoir. Our experience highlights the danger of inadequate supervision and maintenance of equipment used for near-patient testing and the need to monitor such equipment not only in terms of its calibration and analytical performance but also microbiologically.