What Is Known
- Lipid emulsions are an important component of parenteral nutrition but they have been implicated in liver injury.
- Mixed lipid emulsions are thought to be less hepatotoxic than traditional soybean-based emulsions.
- Studies comparing the 2 formulations in the pediatric population under longer-term use are limited.
What Is New
- Our prospective cohort study suggests that prolonged (≥4 weeks) use of mixed lipid emulsions in hospitalized children is more hepatoprotective than soybean-based emulsions.
Lipid emulsions are an essential part of parenteral nutrition (PN). Until recently, the predominant source of fat in these emulsions was soybean oil. Soybean oil-based lipid formulations (SL) have been implicated in the development of intestinal failure associated liver disease (IFALD) (1–3). IFALD is the most common long-term complication in patients on PN (4,5) and its severity ranges from asymptomatic elevations in liver enzymes, conjugated bilirubin, or serum bile acids, to end-stage liver disease requiring liver transplantation.
Although the pathogenesis of IFALD is multifactorial, the choice of lipid emulsion appears to play a role (2). Third-generation mixed lipid emulsions (ML) may be more hepatoprotective (6). One such emulsion is SMOFlipid (Fresenius Kabi, Homburg, Germany), a 20% lipid emulsion comprised of soybean oil (30%), medium-chain triglycerides from coconut oil (30%), olive oil (25%), and fish oil (15%). Studies in adults suggest that ML is associated with lower levels of liver enzymes and serum markers of cholestasis compared with SL, such as Intralipid (Baxter, Deerfield, IL) (7). Few pediatric studies have indicated that ML is well-tolerated, and its use is associated with lower serum bilirubin levels compared with SL (8–11). These studies are, however, limited to neonates, infants with short bowel syndrome or stable children receiving PN at home. It remains unclear whether the hepatoprotective benefit is also seen with longer use in acutely unwell hospitalized children, including those beyond the neonatal period and with varying medical conditions. The objective of this study was to compare markers of hepatobiliary injury associated with prolonged (≥4 weeks) use of ML against SL in a cohort of hospitalized children.
This was a cohort study of patients exposed to PN between 2000 and 2016 at the Hospital for Sick Children (Toronto, Canada). ML became available for use in our institution in 2013, before which SL was the only available lipid emulsion (along with fish oil-based emulsions, which were only used for those with documented IFALD).
Patients from the pediatric, general surgery, intestinal failure, gastroenterology, and cardiology wards who were <18 years old with exposure to ML for ≥4 weeks at the time of study entry were recruited. Patients with interruptions in PN for >2 days, concurrent exposure to fish-oil based emulsions (Omegaven, Fresenius Kabi), preexisting liver disease, and oncological diagnoses (as they often have hepatitis and/or cholestasis secondary to other etiologies) were excluded. Different teams may have cared for patients with surgical diagnoses (eg, gastroschisis) over the course of the study.
Patients receiving ML provided informed consent to enrollment. Data for patients on ML were collected prospectively except for information before recruitment (ie, first 4 weeks of PN/ML exposure), which were collected retrospectively. Decisions around PN prescriptions, enteral intake, medications, and blood work were at the discretion of the primary medical team of the patients, which included specialized registered dietitians. These patients were matched for age (±6 months for those <24 months of age, and ±2 years for those >24 months of age, unless at or beyond puberty) and primary diagnosis to a historical cohort (2000–2013) of patients who were exposed to SL for a similar duration of time.
Data collected included the patients’ age, sex, gestational age, birth weight, reports of central line-associated blood stream infections (which was verified using blood culture results), diagnosis, and indication for PN. Weekly data collection from the onset to the termination of PN included PN prescriptions and enteral intake; weight, length (age <24 months), or height (≥24 months); and z scores (calculated using the WHO Anthro for personal computers SPSS Macro, Version 3.2.2 (World Health Organization, Geneva, Switzerland; 2011: http://www.who.int/childgrowth/software/en/)); and biochemical markers of hepatobiliary injury, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-glutamyl transferase (GGT), alkaline phosphatase (ALP), unconjugated bilirubin, and conjugated bilirubin.
The study was approved by the research ethics board and was in compliance with the ethical principles outlined in the declaration of Helsinki.
This was a retrospective analysis of prospectively collected data (ML patients) and retrospective chart review data (SL patients). Comparisons between the paired subjects in each group were performed using the McNemar test for categorical variables and the related samples Wilcoxon signed-rank test for continuous variables. Significance threshold was set at 0.05. Sample size was calculated based on the data by Goulet et al (9) where mean conjugated bilirubin at 4 weeks of PN exposure was 11 μmol/L for the SL group and 8 μmol/L for the ML groups, with a standard error of the mean of 3 μmol/L. Based on that, data from 16 patients provides sufficient power to detect differences in conjugated bilirubin between the groups.
The univariable effect of type of lipid administered (ML vs SL) over time on liver biochemistry was explored by comparing median values at baseline and weeks 4, 8, and 12 using the related samples Wilcoxon signed-rank test. We also used a generalized estimating equations (GEE) approach (12). Briefly, this method compares the overall trajectory of a variable of interest (eg, liver biochemistry values) over time. It also enables multivariable modeling, to account for the effect of covariates on the main outcome of interest. It is robust in the case of missing data; high quality estimates are still generated with this approach even if patients have different duration of follow-up or number of samples taken (13).
SPSS Statistics for Windows, Version 23.0 (Armonk, NY: IBM Corp) was used for all analyses.
A total of 40 patients (20 exposed to ML, and 20 to SL) were included in the study. All patients were hospitalized during exposure to PN. Baseline characteristics were similar between the groups (Table 1). The median year in which PN was started was 2014 (2013–2016) for the ML group and 2012 (2003–2013) for the SL group. Diagnoses leading to PN requirement were gastroschisis (N = 12, 30%), inflammatory bowel disease (N = 8, 20%), intestinal obstruction (N = 4, 5%), intestinal atresia/web (N = 6, 15%), dysmotility (N = 2, 5%), short bowel syndrome (N = 2, 5%), intussusception (N = 2, 5%), cystic fibrosis/meconium ileus (N = 2, 5%), and malrotation/volvulus (N = 2, 5%). Four and 5 patients in the ML and SL groups, respectively, were older than age 3. Mean age of subjects with inflammatory bowel disease was 13 years in the ML group and 16 years in the SL group.
Baseline weight-for-age and length/height-for-age z scores were comparable between the 2 groups (P = 0.781 and 0.151, respectively). Median weight-for-age z score were similar between the ML and SL groups over time (at 4 weeks: −1.4 vs −1.7, P = 0.528; at 8 weeks: −1.9 vs −2.0, P = 0.374; at 12 weeks: −2.1 vs −2.2, unable to compute P-value due to missing paired data). A GEE approach also showed no difference in the change in weight-for-age z scores over time between the SL and ML groups, controlling for the effects of lipid type, gestational age, and birth weight (P = 0.061 for the interaction between lipid type and time).
Baseline biochemical markers of liver injury (AST, ALT, GGT, ALP, and conjugated bilirubin) were similar between the groups (P = 0.655, 0.655, 0.317, 0.317, 0.173, respectively). Changes in the markers over time were determined using the GEE approach. Simple univariable analysis demonstrated a significant difference in the trajectory of conjugated bilirubin for those on ML compared with SL (P < 0.001; Fig. 1). There was no difference in AST, ALT, or ALP trajectories. There were, however, significant outliers for both AST and ALT; 4 patients had high AST and ALT at baseline which fell within the first week to near-normal levels. A sensitivity analysis with these patients removed demonstrated significantly lower AST and ALT trajectories for the ML group (P < 0.001 for both analyses). Multivariable analysis, controlling for the effects of gestational age and the percentage of total calories delivered from parenteral nutrition, confirmed the persistent benefit of ML on the trajectory of conjugated bilirubin (P < 0.001), but not AST, ALT, and ALP (P = 0.336, P = 0.735, P = 0.575 respectively). When the AST and ALT outliers were removed, a benefit to ML administration on both AST and ALT over time was, however, shown (P < 0.001 for both analyses).
This study shows that prolonged use of ML in hospitalized children is associated with significantly lower conjugated bilirubin compared with SL, suggesting that ML could be used preferentially over SL in hospitalized children with varied medical conditions (such as those with inflammatory bowel disease, intussusception, and bowel obstruction), including those outside the neonatal/infantile period.
Compared with SL, ML's favorable composition—its higher omega-3 to omega-6 ratio, supplemental EPA and DHA (14), higher vitamin E content (15), and lower phytosterol content (14)—renders it less likely to contribute to IFALD. Clinically, ML has been shown to be more hepatoprotective compared with SL in adults. A randomized controlled trial by Klek et al showed that adults with intestinal failure exposed to ML had lower serum transaminase and bilirubin levels, and higher α-tocopherol following a 4-week period compared with those exposed to SL (7). Similar conclusions of favorable hepatobiliary biochemistries were made in a meta-analysis of 6 randomized controlled trials comparing liver enzymes in the postoperative period in adults exposed to ML and SL (16) and another study assessing adults in the critical care unit (17).
Pediatric studies to date have echoed the adult literature. Some of these data have been summarized in a recent meta-analysis (18). The most robust pediatric literature is in the neonatal and infantile population (8,10,11,19). Studies outside of this age group have been of shorter duration (≤4 weeks) of lipid exposure (9,20) or in clinically stable patients on home PN (21). Nonetheless, these studies all reach the same conclusion that ML is associated with lower transaminases and/or conjugated bilirubin levels, suggesting decreased hepatotoxicity.
One study, similar to ours, addressed the use of ML in hospitalized children, including older children (20). Pichler et al retrospectively reviewed the change in liver biochemistries in those who switched from SL to ML (or Lipofundin, LCT/MCT lipid emulsion; B. Braun Medical Supplies, Berlin, Germany) following a rise in total bilirubin >50 μmol/L or ALT, GGT, and/or ALP above 1.5 times normal with prior use of SL. The study cohort also included neonates who received ML (or LCT/MCT lipid emulsion) with no prior exposure to SL. Of the 127 children, 71 received ML (56 received LCT/MCT lipid emulsion) with only 35 patients beyond the neonatal period at ML introduction. The duration of ML in the entire cohort of 71 patients ranged from 3 to 311 days, and their liver biochemistries were comparable to our results.
Considering the results of our study and existing literature, there are reassuring data to support that ML is less hepatotoxic than SL. It should, however, be noted that IFALD has been reported in the context of ML use in children with intestinal failure (22). In these patients, ML discontinuation and fish oil emulsion introduction was required for the resolution of cholestasis.
Our study is limited in that it was observational as our institution mandated the use of ML for all patients who receive PN for more the 2 weeks (unless specifically contraindicated) since 2013 due to concerns regarding IFALD with long-term SL use. In addition, some patients from the SL cohort were cared for a decade before their ML matches. This difference may result in different clinical practice (such as practices around cycling of PN, composition of PN, enteral feeding, and prevention and management of central line-associated blood stream infections) that may confound results. Furthermore, our and previous studies assess biochemical evidence, but do not compare using histological data. Studies are needed to evaluate histological evidence of hepatocellular injury. Lastly, laboratory data were not always available, limiting the power to detect differences between the groups at single time points. The use of the GEE model, however, allowed us to overcome this limitation.
Children on long-term PN and exposed to ML demonstrated decreased evidence of biochemical liver injury, while maintaining similar growth patterns, compared with those exposed to SL. This study corroborates earlier data that ML use is favorable compared with SL, particularly in those requiring prolonged PN.
1. Lauriti G, Zani A, Aufieri R, et al. Incidence, prevention, and treatment of parenteral nutrition
-associated cholestasis and intestinal failure-associated liver disease in infants and children: a systematic review. JPEN J Parenter Enteral Nutr
2. Lacaille F, Gupte G, Colomb V, et al. ESPGHAN Working Group of Intestinal Failure and Intestinal Transplantation. Intestinal failure-associated liver disease: a position paper of the ESPGHAN Working Group of Intestinal Failure and Intestinal Transplantation. J Pediatr Gastroenterol Nutr
3. Lee WS, Sokol RJ. Intestinal microbiota, lipids, and the pathogenesis of intestinal failure-associated liver disease. J Pediatr
4. Colomb V, Dabbas-Tyan M, Taupin P, et al. Long-term outcome of children receiving home parenteral nutrition
: a 20-year single-center experience in 302 patients. J Pediatr Gastroenterol Nutr
5. Pichler J, Horn V, Macdonald S, et al. Intestinal failure-associated liver disease in hospitalised children. Arch Dis Child
6. Hojsak I, Colomb V, Braegger C, et al. ESPGHAN Committee on Nutrition. ESPGHAN Committee on Nutrition Position Paper. Intravenous lipid emulsions and risk of hepatotoxicity
in infants and children: a systematic review and meta-analysis. J Pediatr Gastroenterol Nutr
7. Klek S, Chambrier C, Singer P, et al. Four-week parenteral nutrition
using a third generation lipid emulsion (SMOFlipid
)—a double-blind, randomised, multicentre study in adults. Clin Nutr
8. Diamond IR, Grant RC, Pencharz PB, et al. Preventing the progression of intestinal failure-associated liver disease in infants using a composite lipid emulsion: a pilot randomized controlled trial of SMOFlipid
. JPEN J Parenter Enteral Nutr
9. Goulet O, Antebi H, Wolf C, et al. A new intravenous fat emulsion containing soybean oil, medium-chain triglycerides, olive oil, and fish oil: a single-center, double-blind randomized study on efficacy and safety in pediatric patients receiving home parenteral nutrition
. JPEN J Parenter Enteral Nutr
10. Rayyan M, Devlieger H, Jochum F, et al. Short-term use of parenteral nutrition
with a lipid emulsion containing a mixture of soybean oil, olive oil, medium-chain triglycerides, and fish oil: a randomized double-blind study in preterm infants. JPEN J Parenter Enteral Nutr
2012; 36 (1 suppl):81S–94S.
11. Tomsits E, Pataki M, Tolgyesi A, et al. Safety and efficacy of a lipid emulsion containing a mixture of soybean oil, medium-chain triglycerides, olive oil, and fish oil: a randomised, double-blind clinical trial in premature infants requiring parenteral nutrition
. J Pediatr Gastroenterol Nutr
12. Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics
13. Ballinger GA. Using generalized estimating equations for longitudinal data analysis. Organ Res Methods
14. Meisel JA, Le HD, de Meijer VE, et al. Comparison of 5 intravenous lipid emulsions and their effects on hepatic steatosis in a murine model. J Pediatr Surg
15. Adolph M. Lipid emulsions in parenteral nutrition
. Ann Nutr Metab
16. Tian H, Yao X, Zeng R, et al. Safety and efficacy of a new parenteral lipid emulsion (SMOF) for surgical patients: a systematic review and meta-analysis of randomized controlled trials. Nutrition reviews
17. Antebi H, Mansoor O, Ferrier C, et al. Liver function and plasma antioxidant status in intensive care unit patients requiring total parenteral nutrition
: comparison of 2 fat emulsions. JPEN J Parenter Enteral Nutr
18. Finn KL, Chung M, Rothpletz-Puglia P, et al. Impact of providing a combination lipid emulsion compared with a standard soybean oil lipid emulsion in children receiving parenteral nutrition
: a systematic review and meta-analysis. JPEN J Parenter Enteral Nutr
19. Skouroliakou M, Konstantinou D, Koutri K, et al. A double-blind, randomized clinical trial of the effect of omega-3 fatty acids on the oxidative stress of preterm neonates fed through parenteral nutrition
. Eur J Clin Nutr
20. Pichler J, Simchowitz V, Macdonald S, et al. Comparison of liver function with two new/mixed intravenous lipid emulsions in children with intestinal failure. Eur J Clin Nutr
21. Muhammed R, Bremner R, Protheroe S, et al. Resolution of parenteral nutrition
-associated jaundice on changing from a soybean oil emulsion to a complex mixed-lipid emulsion. J Pediatr Gastroenterol Nutr
22. Lee S, Park HJ, Yoon J, et al. Reversal of Intestinal failure-associated liver disease by switching from a combination lipid emulsion containing fish oil to fish oil monotherapy. JPEN J Parenter Enteral Nutr