Investigation of an outbreak of Serratia marcescens in a neonatal unit via a case-control study and molecular typing.

Background: In March 2004, infection or colonization with Serratia marcescens affected one third of all neonates in a newborn services unit (NBS).
Methods: We performed a case-control study and automated ribotyping.
Results: Forty-nine cases were compared with 64 controls. The overall mean length of stay (LOS) in the NBS was 67 days for cases and 36 days for controls, P = .005. Cases were of lower mean birth weight than controls (1566 g vs 1968 g, respectively, P = .02). Risk factors that trended toward significance for S marcescens acquisition included the following: premature rupture of membranes (odds ratio [OR], 2.7; 95% confidence interval [95% CI]: 1.0-7.1; P = .05), vaginal delivery at our hospital (OR, 2.1; 95% CI: 0.9-4.6; P = .06), intubation at delivery (OR, 2.3; 95% CI: 0.9-5.2; P = .05), mechanical ventilation (OR, 2.1; 95% CI: 0.9-4.4; P = .06), and theophylline treatment (OR, 2.5; 95% CI: 1.1-5.4; P = .02). Multiple logistic regression analysis revealed vaginal delivery at our hospital (OR, 3.4; 95% CI: 1.4-8.2; P = .007) and LOS >30 days (OR, 4.4; 95% CI: 1.8-10.6; P = .001) as independent risk factors for S marcescens acquisition. Ribotyping of specimens revealed 5 restriction patterns.
Conclusion: Cases were of lower birth weight than controls, were born by vaginal delivery at our hospital, had longer LOS in NBS, and had greater requirements for respiratory support. Ribotyping of specimens revealed that this outbreak was not clonal.