Introduction Lipid infusions as part of parenteral nutrition (PN) can adversely affect the liver function as measured by the observed rise in serum LFT's during PN administration. Using full dose of lipid infusions in patients with deranged LFT's is not advisable by current consensus. In the event of sepsis, or significantly raised LFT's, the lipid infusions are either reduced or completely removed from the PN regime. There is an absence of information in the literature as to the impact of various lipid formulations on the LFT's within the critically ill paediatric population.

Hypothesis To investigate the impact of lipid infusions on LFT's and outcomes of sepsis and death.

Methods A retrospective cohort study including patients who received PN between 2007 and 2011 on a tertiary paediatric intensive care unit. Patients who received PN were identified from the pharmacy records. Out of 116 records identified, 50 records were suitable or available for final analysis. Records of theses patients were scanned and observations relating to their period of PN administration were recorded.

Results A total of 50 patients (48% females; n=24) were included during the study period. The median age was 2 months (IQR 0,12). The majority of diagnoses were cardiac (41/49, 84%); 41% of whom were HLHS (17/41). The median number of days on PN therapy was 13 (IQR 8, 28). Patients were grouped by the type of lipid they received: SMOF, Lipofundin, and Intralipid. The majority of patients received Lipofundin (58%; n=29) followed by SMOF (24%; n=12). One patient received no lipids.

The median duration of PN in the intralipid, SMOF, and lipofundin groups were 8.5, 30.5, and 12 days respectively (n=49, p=0.0056), and they received a median dose in g/kg/day of 2.5, 2.75, and 3 respectively (n=49, p=0.724). Death occurred in 2/8 (25%), 3/12 (25%), and 4/27 (15%) in those who received intralipid, SMOF, and lipofundin respectively (n=48, p=0.67). Sepsis was present in 5/8 (63%), 11/12 (92%), and 15/27 (56%) in those who received intralipid, SMOF, and lipofundin respectively (n=31, p=0.08). Of the septic patients, sepsis occurred post PN initiation in 1/5 (20%), 3/10 (30%), and 4/15 (27%) in those who received intralipid, SMOF, and lipofundin respectively (n=30, p=1.00)

Logistic regression for the effect of duration of PN on sepsis shows a 3.4% rise in the odd ratio (OR) of developing sepsis post PN for each day on PN, (OR-1.0339, 95% CI 1.0108–1.0577, p=0.004). Likewise, patients on longer term PN were more likely to have sepsis (OR-1.094, CI-1.0140–1.2138, p=0.024).

Regression analyses for changes in LFT's for the different lipids showed no relationship between the dose or type of lipids and LFT or bilirubin changes from baseline.

Conclusions Patients who received SMOF had a similar mortality rate compared to those on other type of lipids despite their higher proportion with sepsis and their significantly prolonged time on PN. There was no relationship between the lipid dose and LFT's suggesting that adequate doses of lipids in the presence of sepsis and raised LFT's are not contraindicated. Although firm conclusions cannot be drawn, data from this study should prove useful in the assessment of SMOF use in septic and post surgical children and in the planning of a future large scale study.