Outbreak of Serratia marcescens bloodstream infections in patients receiving parenteral nutrition prepared by a compounding pharmacy

Abstract

Background: Compounding pharmacies often prepare parenteral nutrition (PN) and must adhere to rigorous standards to avoid contamination of the sterile preparation. In March 2011, Serratia marcescens bloodstream infections (BSIs) were identified in 5 patients receiving PN from a single compounding pharmacy. An investigation was conducted to identify potential sources of contamination and prevent further infections.

Methods: Cases were defined as S. marcescens BSIs in patients receiving PN from the pharmacy between January and March 2011. We reviewed case patients' clinical records, evaluated pharmacy compounding practices, and obtained epidemiologically directed environmental cultures. Molecular relatedness of available Serratia isolates was determined by pulsed-field gel electrophoresis (PFGE).

Results: Nineteen case patients were identified; 9 died. The attack rate for patients receiving PN in March was 35%. No case patients were younger than 18 years. In October 2010, the pharmacy began compounding and filter-sterilizing amino acid solution for adult PN using nonsterile amino acids due to a national manufacturer shortage. Review of this process identified breaches in mixing, filtration, and sterility testing practices. S. marcescens was identified from a pharmacy water faucet, mixing container, and opened amino acid powder. These isolates were indistinguishable from the outbreak strain by PFGE.

Conclusions: Compounding of nonsterile amino acid components of PN was initiated due to a manufacturer shortage. Failure to follow recommended compounding standards contributed to an outbreak of S. marcescens BSIs. Improved adherence to sterile compounding standards, critical examination of standards for sterile compounding from nonsterile ingredients, and more rigorous oversight of compounding pharmacies is needed to prevent future outbreaks.

Keywords: Serratia; compounding; contamination; nutrition; outbreak.

Figure 1.

Epidemic curve of Serratia marcescens bloodstream infections among patients in hospitals receiving parenteral nutrition (PN) from the pharmacy, by PN exposure, Alabama, April 2010–March 2011. Abbreviation: PN, parenteral nutrition.

Figure 2.

Summary of the parenteral nutrition (PN) compounding process at the pharmacy and breaches in practice that potentially contributed to contamination of amino acids with Serratia marcescens. 1, The pharmacy tap water faucet (the postulated source of introduction of S. marcescens) was used to clean a large mixing container. 2, Amino acid powders were compounded in sterile water in the mixing container (on occasion up to 48 hours prior to filter-sterilization); failure to filter solution immediately following this process could have allowed for increased bacterial growth and endotoxin production. 3, The solution was passed through a 0.2-µm sterilizing filter into individual sterile bags. A larger upstream filter (prefilter) was not used to reduce the bioburden of bacteria or remove excessive particulate matter visible in solution; this particulate matter could have interfered with efficiency of the sterilizing filter. Also, replacement of the clogged filter during filtration caused a break in the sterile system and could serve as a source for downstream contamination. 4, The filtered amino acid solution was stored in individual sterile bags pending use in PN; suboptimal sampling for sterility testing failed to detect contamination with S. marcescens. 5, Contaminated amino acids were eventually incorporated into adult PN preparations and administered to patients at 6 Alabama hospitals. Abbreviation: PFGE, pulsed-field gel electrophoresis.

Figure 3.

Results of pulsed-field gel electrophoresis (PFGE) of environmental and blood isolates of Serratia marcescens. All environmental isolates from the pharmacy were indistinguishable by PFGE. Isolates from 14 of the 19 case patients were available for testing; 13 were indistinguishable to environmental isolates and 1 isolate (patient 10) differed by one band. The “non-PN related” blood specimen was obtained in March 2011 from a patient at hospital B who did not receive parenteral nutrition. Abbreviation: PN, parenteral nutrition.