Abstract
Background: Bloodstream infections (BSIs) are a well-recognized complication of parenteral nutrition (PN). However, their epidemiology and clinical consequences are incompletely described. Methods: A retrospective cohort study was performed, from 2002 to 2009, of all hospital inpatients who were administered PN, outside the intensive care setting, at a major tertiary hospital in Queensland, Australia. Results: In 780 episodes of PN administration, 120 BSIs occurred, giving an incidence of 10.0/1000 PN-days. The majority of PN-associated BSIs were classified as central line-associated (n = 98, 81.7%). Candida spp. were the most frequent pathogens. Observed BSI management revealed that over 8% of intravascular devices were inappropriately retained, over 30% of empirical antibiotic therapy was inappropriate, and 62% of antifungal therapy was delayed ≥ 48 h. All-cause hospital mortality was over 2-fold greater in patients with a PN-associated BSI compared to those without (17.9% vs 8.3%, crude odds ratio (OR) 2.4, 95% confidence interval (CI) 1.29–4.35, p = 0.002). BSI was identified as an independent risk factor for mortality (adjusted OR 3.54, 95% CI 1.76–7.12, p < 0.001). Low baseline albumin levels and a requirement for intravenous insulin infusion (a marker of sustained hyperglycaemia) were independent risk factors for the development of PN-associated BSIs. Conclusions: PN-associated BSI in hospital inpatients is common and is associated with mortality. The implementation of standardized evidence-based infection prevention strategies, particularly targeting IVD maintenance, is a priority. PN-associated BSI management pathways require optimization, with timely IVD removal and appropriate antimicrobial therapy. Depending on local epidemiology patterns, empirical antifungal therapy should be considered.
Acknowledgements
The authors would like to acknowledge Kath Murrell, Parenteral Nutrition Clinical Nurse Consultant at the Princess Alexandra Hospital, for collection and provision of data.
Declaration of interest: The authors declare no conflict of interests.