The introduction of parenteral nutrition (PN) to paediatric practice several decades ago made possible the survival of infants and children with chronic intestinal failure. Unfortunately, many of these children experience potentially life-threatening complications including PN-associated liver disease (PNALD) (1,2). Up to 60% of children receiving long-term PN experience some degree of hepatic dysfunction (3). This may be manifested as minor liver test derangements, overt cholestasis, or end-stage liver disease and liver failure. There are recognised risk factors for PNALD in children, including prematurity, age, massive loss of small bowel, repeated laparotomies, absence of enteral feeding, necrotising enterocolitis, and sepsis (1,4). The development of cholestasis in children receiving PN is an ominous event. In a study of prognostic factors in children with short bowel syndrome (SBS), the presence of cholestasis was associated with a 50% mortality compared with 95% survival in those without cholestasis (1).
The aetiology of PNALD disease is unknown (5). Speculation has focused on possible nutrient excesses or deficiencies, and potentially toxic constituents. (6–8) In particular, the possible role of lipid emulsions has been a focus of concern (9).
Until recently, most lipid emulsions were based on soybean or safflower oil. Specific fatty acids have potent biologic effects, for example, influencing inflammatory processes and modulating immune function (10). There are concerns that PN lipids derived solely from vegetable oils may have adverse effects particularly in relation to the response to stress or sepsis. Various alternative lipid emulsions have therefore been introduced, containing, for example, medium-chain triglycerides (MCTs), monounsaturated fatty acids, or fish oils (10).
If a PN regimen could be devised that was effective in treating or preventing PNALD, then this would be a major medical breakthrough, potentially avoiding the need for liver transplantation and improving the survival rate in these children. Reports from a group in Boston suggested that in infants with PNALD, discontinuation of a standard soybean oil–based lipid and administering a much reduced amount of lipid in the form of a fish oil–based emulsion led to the resolution of cholestasis in some children (11-14).
We describe the resolution of jaundice in children receiving PN when changed from a soybean lipid emulsion (Intralipid) to a complex mixed-type lipid emulsion derived from soybean, coconut, olive, and fish oils. It is important to note that the total dose of lipid was not reduced in our patients. We compared these children with a historic cohort of children with PNALD who remained on Intralipid, following their progress during a 6-month period.
In January 2007, we began a practice of changing from Intralipid 20% (Fresenius Kabi) to SMOFlipid (Fresenius Kabi) as the PN lipid emulsion in any child who developed cholestatic jaundice due to PNALD. Intralipid 20% contains soybean oil 200 g/L and has an energy content of 8.4 MJ (2000 kcal/L). SMOFlipid contains soybean oil 60.0 g/L, MCT, 60.0 g/L, olive oil 50.0 g/L, and fish oil, rich in ω-3 acids 30.0 g/L, and has an energy content of 8.4 MJ/L (equal to 2000 kcal/L). It also contains α-tocopherol.
During a 1-year period, the change to SMOFlipid was made in 8 children, all of whom had a total bilirubin >70 μmol/L. No children with a total bilirubin >70 μmol/L remained on Intralipid during this period. No other alterations to management were implemented in these cases. Routine measures aimed at reducing the risk of PNALD continued. These measures included management by a multidisciplinary team, treatment with oral ursodeoxycholic acid, intravenous antibiotics for suspected and proven central venous catheter infections, maximisation of enteral nutrition (EN), and cycling on and off PN each day to reduce the total duration of administration. The 8 children remained dependent on PN during a follow-up period of at least 6 months. We examined their outcome during this time in terms of PN and EN dependence and bilirubin concentration. All of the children receiving PN for ≥6 months during this period who fulfilled the criterion for inclusion had a total bilirubin >70 μmol/L and all received SMOFlipid.
We compared these children with a historic cohort of children with PNALD and hyperbilirubinemia who were managed during a 3-year period immediately before the introduction of SMOFlipid usage. There were 2 criteria for inclusion in this comparison group: first, the serum bilirubin should have been persistently >70 μmol/L on 4 successive weekly estimates; second, they should have subsequently remained dependent on PN for at least 6 months to facilitate comparison with the SMOFlipid group. Nine children received PN for ≥6 months during this period, and each of the children fulfilled the second inclusion criterion regarding bilirubin concentrations.
The children in both cohorts were investigated to exclude causes of liver disease other than PNALD. Routine investigations included abdominal ultrasound scanning, screening for α1-antitrypsin deficiency, hypothyroidism, congenital infections (TORCH screen), galactosemia, and metabolic disorders (plasma amino acids and urinary organic acids).
For both groups, demographic, clinical, and laboratory data were obtained from the medical, pharmacy, dietetic, and nutritional care team records with particular regard to EN and PN support and the serum bilirubin concentrations.
Statistical analyses were performed using SPSS version 16.0 (SPSS Inc, Chicago, IL), using descriptive statistics including 95% confidence intervals and parametric and nonparametric comparative statistics as appropriate. Ethical approval was not required for the present study because it was a retrospective cohort study.
The individual characteristics of the 8 children at entry to the study are presented in Table 1. Four had SBS and 4 had severe protracted diarrhoea of infancy (phenotypic diarrhoea of infancy (15) n = 2; autoimmune enteropathy n = 1; idiopathic n = 1). At the time of commencing SMOFlipid, the median age of the group was 30 weeks (range 16–164). Each had a persistently elevated serum bilirubin concentration (>70 μmol/L) based on weekly estimates during a period of at least 4 weeks. All of the children were receiving PN, including lipid each day, with the exception of patient 1. He had persistent jaundice for several months, and although he continued receiving daily PN, the lipid infusions were reduced to 3 times weekly in a dose of 1.5 g/kg (average 0.64 g · kg−1 · day−1) 5 weeks before changing to SMOFlipid. The SMOFlipid was then given using that same regimen for 8 weeks before being increased to 4 nights at 2.0 g/kg (average 1.1 g · kg−1 · day−1).
Table 2 summarises the serum bilirubin concentrations and nutritional support at entry to the study and at 6 months. At the time of the change to SMOFlipid, the median bilirubin was 143 μmol/L (range 71–275). At this time, the median intake of energy from PN was 95 kcal/kg (range 64–108). Subjects were receiving EN in accordance with individual tolerance (median 21 kcal/kg; range 0–50). Each child was receiving most of his or her energy intake as PN (median 81.5%, range 65.5–100). The median amount of lipid being administered was 2.0 g · kg−1 · day−1 (range 0.6–3.5). On commencing SMOFlipid, the total amount of lipid being administered was not changed and no other alterations were made to the PN prescription.
During the next 6 months, PN was reduced stepwise if improved EN tolerance permitted. Table 3 presents the changes in nutritional support during the 6-month follow-up period. At the 6-month time point, all of the children remained substantially dependent on PN for their energy requirements (median 74.5 kcal · kg−1 · day−1, range 48–90), and all but 1 were still receiving most of their energy intake as PN (median 68.4%, range 36.6–100). As a result of improving enteral tolerance, there was a significant reduction in PN and a significant reduction in the proportion of energy provided as PN (Table 3).
The serum bilirubin concentrations during the 6-month period following the change to SMOFlipid are presented in Figure 1 and the final concentrations in Table 2. In 5 of the 8 children, jaundice had completely resolved by the end of this follow-up period. In another child, there was a marked reduction in serum bilirubin with a progressive fall from 267 to 92 μmol/L. In 1 child, the bilirubin rose to 313 μmol/L during a 3-month period after commencing SMOFlipid but then fell progressively to 170 μmol/L during the next 3 months. In 1 case, there was no improvement, the bilirubin increasing gradually from an initial value of 150 μmol/L to a maximum value of 220 μmol/L. Overall, the median change in serum bilirubin in the 6-month follow-up period was −99 μmol/L (range −176 to 6).
The individual characteristics of the 9 children at entry to the study are presented in Table 4. All of the children in this group had intestinal failure due to SBS. The median age at which they fulfilled the bilirubin inclusion criterion was 24 weeks (range 8–64). This was not significantly different from the age of the SMOFlipid group (P = 0.074). All of the patients received PN from the first days of life. The total duration of PN administration in this group was not significantly different from the SMOFlipid group (P = 0.13).
Table 2 summarises the data in this group at entry to the study and at 6 months. At the point of entry, the median bilirubin concentration was 91 μmol/L (range 78–176). At this point, they were each receiving most of their energy intake as PN (median 96 kcal · kg−1 · day−1, range 65–116). Like the SMOFlipid group, they received EN to the limits of tolerance with a median intake of 8 kcal · kg−1 · day−1 (range 0–58). The median proportion of energy intake received as PN was 92.2% (range 60.3–100). The median amount of lipid being administered was 3.5 g · kg−1 · day−1 (range 2.5–3.5).
As with the SMOFlipid group, during the 6-month follow-up period, PN was reduced stepwise if improved EN tolerance permitted. Table 3 presents the changes in nutritional support during the 6-month follow-up period. At the end of the 6-month period, although the energy intake as PN had fallen significantly (P = 0.014), they nevertheless remained substantially dependent on it, receiving a median PN energy intake of 74.5 kcal · kg−1 · day−1 (range 25–95). By 6 months, there was also a significant increase in EN energy intake (median 49 kcal · kg−1 · day−1; range 30–111; P = 0.009). In contrast to the SMOFlipid group, the amount of PN lipid being administered in this group had fallen significantly (P < 0.001).
The serum bilirubin concentrations in these children during the 6-month period of follow-up are presented in Figure 2 and the final concentrations in Table 2. At 6 months, hyperbilirubinemia had resolved in only 2 of the 9 children. Overall, there was a nonsignificant increase in the bilirubin concentration during the 6-month interval (median 79 μmol/L, range −36 to 127; P = 0.193).
Comparison of SMOFlipid and Intralipid Groups
With regard to nutritional support, there was no significant difference between the 2 groups in PN or EN energy intake or in the relative proportion of energy received as PN in the 2 groups at the outset or at 6 months (Table 2). The amount of PN lipid being administered was greater in the Intralipid group at the outset (P = 0.005), but by 6 months, there was no difference between the groups (Table 2). Comparing the changes in both groups during the 6-month period of follow-up, the Intralipid group experienced a greater reduction in the proportion of energy being given as PN (P = 0.008) (Table 3).
There was no significant difference in the serum bilirubin concentrations in the SMOFlipid and Intralipid groups at the outset (P = 0.194). At 6 months, the median serum bilirubin in the SMOFlipid group was 19 (range 6–213) compared with 185 (range 11–269) in the Intralipid group, this difference being of borderline statistical significance (P = 0.058) (Table 2). There was a significant difference between the groups with regard to the change in serum bilirubin during the 6-month follow-up period (P = 0.02) (Table 3), the median bilirubin concentration having fallen by 99 μmol/L in the SMOFlipid group while increasing by 79 μmol/L in the Intralipid group.
We observed total resolution of jaundice in 5 of 8 children with PN-associated jaundice on changing to SMOFlipid. Clinical experience indicates that unless enteral feeding can be increased and PN can be markedly reduced or discontinued, such an event is uncommon (12). The known tendency for cholestasis to persist or worsen in those who remain significantly dependent on PN was observed in the Intralipid group, in whom the median serum bilirubin concentration increased during the 6-month follow-up period. In contrast, the median serum bilirubin concentration fell in the SMOFlipid group, and this was a significant change when compared with the Intralipid group.
Clearly, the 2 groups of children were not comparable in every respect. As with all historic control comparisons, the findings should be interpreted with caution. It could be argued that the Intralipid group was managed differently from the SMOF lipid group in some respect other than the choice of lipid emulsion; however, identical management protocols were used through this time period and no treatment innovations were introduced.
The different disorders responsible for intestinal failure are potentially important variables requiring careful consideration. SBS was the underlying cause in all of the Intralipid group and in half of the SMOFlipid group, the remainder experiencing severe protracted diarrhoea of infancy. At entry to the study, however, the estimated median serum bilirubin in the SMOFlipid group was higher than that in the Intralipid group, suggesting that the degree of hyperbilirubinemia was at least as severe in the SMOFlipid group. Both groups were heavily dependent on PN at the outset. During the 6-month follow-up period, the Intralipid group reduced the proportion of energy they were receiving as PN when compared with the SMOFlipid group (P = 0.008) and the amount of PN lipid received (P < 0.001) when compared with the SMOFlipid group. This increase in enteral tolerance reflected the process of intestinal adaptation in SBS. Increasing enteral tolerance may have been expected to favour improved liver function. Despite this advantage, the Intralipid group fared less well in terms of resolution of jaundice.
Lipid emulsions are a standard constituent of present PN regimens. A normal human diet contains fat as an important macronutrient, so inclusion of lipid in PN has seemed appropriate. Moreover, lipids have advantageous properties, being of low osmolality and having a high energy concentration. PN lipid emulsions were first introduced approximately 50 years ago (10). The use of intravenous lipid gradually became widespread because of the problems of glucose as a sole nonprotein energy source, with adverse effects including hyperglycemia, hepatic steatosis, essential fatty acid deficiencies, and fat-soluble vitamin deficiencies (10,16).
It is increasingly recognized that specific lipids can exert potent biological effects, particularly modulating inflammatory and immune processes. Most PN lipid emulsions have been derived from vegetable oils (soybean or safflower). These contain a high concentration of the ω-6 polyunsaturated fatty acid (PUFA) linoleic acid and relatively low concentrations of the ω-3 PUFA α-linolenic acid (ratio of 7:1) (10). Excessive linoleic acid may have adverse effects because it is converted to arachidonic acid, a precursor of eicosanoids that are both proinflammatory and thrombogenic (17–19). α-Linolenic acid is a precursor of the ω-3 PUFAs eicosapentaenoic acid and docosahexanaeoic acid. Eicosapentaenoic acid is a precursor of eicosanoids that are believed to have anti-inflammatory and antithrombogenic properties (20). In addition, soybean-derived lipids contain high concentrations of γ-tocopherol and are low in α-tocopherol, which is a major lipophilic antioxidant, and this may increase the risk of oxidative stress (19,21). Another concern relates to the phytosterol content of soybean lipids. Phytosterols are cholesterol-like substrates that have been shown in animal experiments to impair bile flow (22). Intralipid, the lipid emulsion used in the control group in the present study, is derived solely from soybean oil.
Fish oil is a source of ω-3 PUFA. A fish oil–based lipid emulsion (Omegaven; Fresenius Kabi) has been commercially manufactured for intravenous infusion. It is generally advised that it should be administered with other lipids so as to provide not more than 20% of the infused emulsion (10). It has a low ratio (1,8) of ω-6 to ω-3 PUFAs (10). Animal studies have suggested benefits with fish oil in terms of endotoxemia and sepsis survival and immune modulation (23–25). Studies in adult patients undergoing surgery have suggested beneficial effects in relation to inflammation and sepsis (26).
A group in Boston reported that in children with PNALD discontinuation of PN soybean oil lipid emulsion and administration of a markedly reduced dose of intravenous lipid in the form of a fish oil–derived fat emulsion was associated with improvement or resolution of cholestasis (11–14). In an open-label trial, 42 infants with cholestasis were changed from Intralipid to Omegaven (Fresenius Kabi) with a “goal” lipid dose of 1 g · kg−1 · day−1. The comparison group was a historic cohort of 59 infants with PNALD and cholestasis who remained on Intralipid in doses of up to 4 g · kg−1 · day−1. There were significantly more deaths and liver transplants in the Intralipid cohort. Among the survivors, cholestasis was reversed in 19 of 38 treated with Omegaven compared with 2 of 36 Intralipid controls.
In a retrospective study from Toronto, 12 children with advanced PNALD were treated with a combination of Intralipid (1 g/kg) and Omegaven (1 g/kg), the dose of Intralipid being reduced if no improvement of deterioration in liver function was seen (27). Of these, 9 had complete resolution of hyperbilirubinemia, although this only occurred with the discontinuation of Intralipid in 5 of these cases.
SMOFlipid is a complex mixed emulsion of 20% lipid containing 30% soybean oil, 30% MCT, 25% monounsaturated fatty acids, and 15% fish oil. The rationale for this composition was that it was believed to be more comparable with the varied fat composition that would have been present in foods available during the course of human evolution, and consequently may be regarded as optimal. MCTs are rapidly metabolised and cleared and are not subject to peroxidation (10), and their inclusion in the mixture would reduce the amount of ω-6 PUFAs. Human and animal studies suggest that monounsaturated fatty acids have immunomodulatory properties, although the clinical importance of such effects is unknown. The ratio of ω-6 to ω-3 PUFAs in SMOFlipid is 2.5:1 (10). In a randomised trial in adults undergoing major surgery compared with a soybean oil–based emulsion, the use of SMOFlipid was associated with lesser disturbances to liver enzymes (19).
In our patients changed to SMOFlipid, a pattern of recovery from jaundice was seen that was similar to that reported with Omegaven. It could be argued that the response to the change in lipid regimen in the Omegaven studies could be accounted for by the reduction in the total amount of lipid administered. Our patients received on average 0.3 g · kg−1 · day−1 of fish oil. This was probably significantly less than the Omegaven study, in which the intended goal was 1 g · kg−1 · day−1, although the actual amount their patients received was not reported (14). In our patients, the total amount of lipid was not reduced, despite that a marked beneficial change occurred. This is important because caution has been expressed about monotherapy with Omegaven (10). Omegaven is low in the essential fatty acid linoleic acid, and although it has been used alone, it is suggested that it should be used in combination with other lipid sources.
We conclude that SMOFlipid may have important liver-protecting properties in children receiving PN, and its introduction may constitute a significant advance in PN formulation. Randomized controlled trials are needed to study the efficacy and safety of SMOFlipid in the prevention of PNALD in children.
1. Spencer AU, Neaga A, West B, et al. Pediatric short bowel syndrome: redefining predictors of success. Ann Surg
2. Gupte GL, Beath SV, Kelly DA, et al. Current issues in the management of intestinal failure. Arch Dis Child
3. Buchman A. Total parenteral nutrition-associated liver disease. J Parenter Enteral Nutr
4. Beath SV, Davies P, Papadopoulou A, et al. Parenteral nutrition-related cholestasis in postsurgical neonates: multivariate analysis of risk factors. J Pediatr Surg
5. Kelly DA. Intestinal failure-associated liver disease: what do we know today? Gastroenterology
2006; 130 (suppl 1):S70–S77.
6. Sokol RJ, Taylor SF, Devereaux MW, et al. Hepatic oxidant injury and glutathione depletion during total parenteral nutrition in weanling rats. Am J Physiol
7. Cavicchi M, Beau P, Crenn P, et al. Prevalence of liver disease and contributing factors in patients receiving home parenteral nutrition for permanent intestinal failure. Ann Intern Med
8. Chessex P, Lavoie JC, Rouleau T, et al. Photooxidation of parenteral multivitamins induces hepatic steatosis in a neonatal guinea pig model of intravenous nutrition. Pediatr Res
9. Colomb V, Jobert-Giraud A, Lacaille F, et al. Role of lipid emulsions in cholestasis associated with long-term parenteral nutrition in children. J Parenter Enteral Nutr
10. Wanten GJ, Calder PC. Immune modulation by parenteral lipid emulsions. Am J Clin Nutr
11. de Meijer VE, Gura KM, Le HD, et al. Fish oil-based lipid emulsions prevent and reverse parenteral nutrition-associated liver disease: the Boston experience. JPEN J Parenter Enteral Nutr
12. Gura KM, Lee S, Valim C, et al. Safety and efficacy of a fish-oil-based fat emulsion in the treatment of parenteral nutrition-associated liver disease. Pediatrics
13. Gura KM, Duggan CP, Collier SB, et al. Reversal of parenteral nutrition-associated liver disease in two infants with short bowel syndrome using parenteral fish oil: implications for future management. Pediatrics
14. Puder M, Valim C, Meisel JA, et al. Parenteral fish oil improves outcomes in patients with parenteral nutrition-associated liver injury. Ann Surg
15. Hartley JL, Zachos NC, Dawood B, et al. Mutations in TTC37 cause trichohepatoenteric syndrome (phenotypic diarrhoea of infancy). Gastroenterology
16. Wretlind A. Development of fat emulsions. J Parenter Enteral Nutr
17. Adolph M. Lipid emulsions in parenteral nutrition. Ann Nutr Metab
18. Furst P, Kuhn KS. Fish oil emulsions: what benefits can they bring? Clin Nutr
19. 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. J Parenter Enteral Nutr
20. Ton MN, Chang C, Carpentier YA, et al. In vivo and in vitro properties of an intravenous lipid emulsion containing only medium chain and fish oil triglycerides. Clin Nutr
21. Pitkanen OM. Peroxidation of lipid emulsions: a hazard for the premature infant receiving parenteral nutrition? Free Radic Biol Med
22. Clayton PT, Whitfield P, Iyer K. The role of phytosterols in the pathogenesis of liver complications of pediatric parenteral nutrition. Nutrition
23. Mascioli E, Leader L, Flores E, et al. Enhanced survival to endotoxin in guinea pigs fed IV fish oil emulsion. Lipids
24. Mascioli EA, Iwasa Y, Trimbo S, et al. Endotoxin challenge after menhaden oil diet: effects on survival of guinea pigs. Am J Clin Nutr
25. Lanza-Jacoby S, Flynn JT, Miller S. Parenteral supplementation with a fish-oil emulsion prolongs survival and improves rat lymphocyte function during sepsis. Nutrition
26. Calder PC. Use of fish oil in parenteral nutrition: rationale and reality. Proc Nutr Soc
27. Diamond IR, Sterescu A, Pencharz PB, et al. Changing the paradigm: omegaven for the treatment of liver failure in pediatric short bowel syndrome. J Pediatr Gastroenterol Nutr