Enterobacter cloacae and Escherichia hermannii Infection and Sepsis in Pediatric Patients due to Contaminated Parenteral Infusion Solutions

Background:

On August 20th, 2010, Enterobacter cloacae and Escherichia hermannii infection and sepsis occured in  11 patients from the pediatric clinic of our university medical center. 4 patients were from the perinatal intensive care unit (PNI), 4 from the intermediate care unit (IMC) and 3 from the intensive care unit (ICU).  In 2 patients E. cloacae and E. hermannii were isolated from blood culture. On the following day, contaminated parenteral mixed infusion was regarded as possible source.  Infusions were stopped and all affected patients treated with adequate antibiotics. 3 patients (premature newborns) died, the other recovered after days to weeks.  Microbiological cultures of residual infusates revealed presence of E. cloacae and E. hermannii indicating that contamination had occurred.

Objective:

To determine the probable source of the E. cloacae and E. hermannii contamination of the infusates.

Methods:

Microbiological cultures of all available infusion bags applied to the patients and of all available bottles used for the preparation of the mixed infusions were performed. Further, all  steps of the production process in the hospital pharmacy were reviewed. Environmental cultures were obtained from various surfaces in the laminar air flow room and adjacent rooms. For molecular typing PFGE was used. For the determination of growth characteristics of E. cloacae and E. hermannii, the microorganisms were inoculated in mixed infusion and amino acid solution and monitored by colony count, turbidity and endotoxin content.

Results:

All parenteral infusion solutions used for the 11 patients were prepared under aseptic conditions in the hospital pharmacy using an automated mixing device. A total of 12 single solutions were used for the preparation. Of the available residual bottles, the amino acid solution was also contaminated by E. cloacae and E. hermannii with a concentration of  ~ 1,200 cfu/ml. PFGE revealed clonal identity for the isolates from blood cultures, parenteral mixed solution and amino acid solution. Environmental monitoring yielded E. cloacae isolates from the sink and from other surfaces, but all displayed PFGE patterns different from the patients isolates.

Conclusions:

At the moment, the current hypothesis is that contamination of the amino acid bottle was responsible for the infections and that it occurred long before entry of the bottle in the hospital pharmacy, most probably caused by a thin invisible crack. Further epidemiological and microbiological investigations are still in progress.