Risk factors for acquisition of Serratia marcescens in a surgical intensive care unit

Between January 1996 and May 1997, a four-fold increased rate of isolation of Serratia marcescens was observed amongst patients admitted to the surgical Intensive Care Unit (SICU) of the Leiden University Medical Center compared to the preceding years. Random amplification of polymorphic DNA showed the involvement of genotypically distinct strains, implicating multiple different sources. After improvement of hygienic measures the frequency of isolation of S. marcescens returned to baseline. A case-control study was performed to assess patient-related risk factors for acquisition of S. marcescens. Nineteen cases and 38 controls were included. Hospital- and SICU-stay were significantly longer in case patients than in controls. By univariate analysis, statistically significant differences were found in body weight, the duration of mechanical ventilatory support, the cumulative use of antimicrobial agents, the use of aminoglycosides, parenteral nutrition and tube feeding. The sum of the number of days per invasive device (deep intravenous lines, arterial lines, wound drains and urinary catheters) was higher in cases than in controls (P = 0.08). Categorically, a cumulative number of device-days > 25 was a statistically significant risk factor for acquisition of S. marcescens. Multivariable logistic regression analysis showed that body weight, parenteral feeding and mechanical ventilation were independent predictors of acquisition of S. marcescens. As transmission of S. marcescens appears to be by the hands of personnel, the identified risk factors may act by necessitating an increased frequency and intensity of direct contacts.