To determine the cause and mode of transmission of a cluster of infections due to Enterobacter cloacae.

Retrospective cohort study in a neonatal intensive-care unit (NICU) from December 1996 to January 1997; environmental and laboratory investigations.

60 infants hospitalized in the NICU during the outbreak period.

Odds ratios (OR) linking E cloacae colonization or infection and various exposures. All available E cloacae isolates were typed and characterized by contour-clamped homogenous electric-field electrophoresis to confirm possible cross-transmission.

Of eight case-patients, two had bacteremia; one, pneumonia; one, soft-tissue infection; and four, respiratory colonization. Infants weighing <2,000 g and born before week 33 of gestation were more likely to become cases (P<.001). Multivariate analysis indicated that the use of multidose vials was independently associated with E cloacae carriage (OR, 16.3; 95% confidence interval [CI95], 1.8-∞ P=011). Molecular studies demonstrated three epidemic clones. Cross-transmission was facilitated by understaffing and overcrowding (up to 25 neonates in a unit designed for 15), with an increased risk of E cloacae carriage during the outbreak compared to periods without understaffing and overcrowding (relative risk, 5.97; CI95, 2.2-16.4). Concurrent observation of healthcare worker (HCW) handwashing practices indicated poor compliance. The outbreak was terminated after decrease of work load, increase of hand antisepsis, and reinforcement of single-dose medication.

Several factors caused and aggravated this outbreak: (1) introduction of E cloacae into the NICU, likely by two previously colonized infants; (2) further transmission by HCWs' hands, facilitated by substantial overcrowding and understaffing in the unit; (3) possible contamination of multidose vials with E cloacae. Overcrowding and understaffing in periods of increased work load may result in outbreaks of nosocomial infections and should be avoided.