Systematic review and meta-analyses of the effect of lipid emulsion on microbial growth in parenteral nutrition

Abstract

As lipid in parenteral nutrition (PN) purportedly enhances microbial growth, recommendations limit infusion of lipid PN (or lipid emulsion) from a single container to 24h (48h for lipid-free PN). However, the associated evidence base is ambiguous. To examine factors affecting microbial growth in PN. A systematic review with meta-analyses examined effects of nutrients on microbial growth in PN infusates over a 48-h period using the growth ratio {GR=log10[colony-forming units (cfu)/mL at 48h/cfu/mL at time zero]}. Factors influencing GR in PN included glucose, microbial species, temperature, osmolarity, presence of vitamins, trace elements and lipid, and amino acid profile. Using unmatched datasets (N=306), a general linear model found that lipid inclusion in PN represented 3.3% of the variability, which was less than that due to glucose concentration (5.8%), microbial species (35.3%) and microbe-infusate interaction (4.4%). Using matched datasets (N=38 pairs), lipid inclusion in PN represented 5.4% of the variability (P=0.076), which was less than that due to glucose concentration (8.5%; P=0.025), microbial species (75.5%; P<0.001) and microbe-infusate interaction (13.3%; P=0.382). Using meta-analyses of matched datasets, the presence of lipid in PN at fixed glucose concentrations did not significantly increase GR of Candida albicans, Escherichia coli or Staphylococcus epidermidis (P=0.352, P=0.025 and P=0.494, respectively; overall P=0.175). Lipid inclusion in PN is only one of several factors that may influence microbial growth in PN. Any recommendations about the duration of PN infusion from a single container should account for all these factors, and should be weighted according to microbial species likely to contaminate PN.