An Observational Study of Smoflipid vs Intralipid on the Evolution of Intestinal Failure–Associated Liver Disease in Infants With Intestinal Failure

Abstract

Background:

SMOFlipid has a more diverse lipid profile than traditional Intralipid and has become the standard lipid for patients in our intestinal rehabilitation program. Our objective was to compare outcomes in neonates with intestinal failure (IF) who received SMOFlipid against those receiving Intralipid.

Methods:

This was a retrospective cohort study of infants with IF with a minimum follow-up of 12 months in 2008–2016. Patients were stratified into 2 groups: group 1 received SMOFlipid; group 2 was a historical cohort who received Intralipid. The primary outcome was liver function evaluated using conjugated bilirubin (CB) levels. Statistical analysis included the Mann-Whitney U and χ² tests, with an α value < 0.05 considered significant. Approval was obtained from our institutional review board.

Results:

Thirty-seven patients were evaluated (17 = SMOFlipid, 20 = Intralipid). SMOFlipid patients were less likely to reach CB of 34 (24% vs 55%, P = 0.05), 50 μmol/L (11.8% vs 45%; P = 0.028), and did not require Omegaven (0% vs 30%; P = 0.014). CB level at 3 months after initiation of parenteral nutrition (PN) was lower in patients receiving SMOFlipid (0 vs 36 μmol/L; P = 0.01). Weight z-scores were improved for patients receiving SMOFlipid at 3 months (−0.932 vs −2.092; P = 0.028) and 6 months (−0.633 vs −1.614; P = 0.018). There were no differences in PN-supported patients or demographics between the groups.

Conclusion:

Use of SMOFlipid resulted in decreased development of IF-associated liver disease in patients with IF when assessed using biochemical tests.

Introduction

Intestinal failure (IF) is a complicated medical condition that results in malabsorption of fluids and nutrients and requires partial or total parenteral support to sustain growth and hydration.^1,2 Pediatric IF is caused by conditions that include short-bowel syndrome (SBS), dysmotility disorders, and mucosal enteropathies, with SBS being the most common cause.^1–4

Traditionally, IF-associated liver disease (IFALD) was a significant cause of mortality in pediatric IF.^5–7 Cholestasis has been reported in several studies with estimates ranging from 30% to 90% of pediatric IF patients developing some degree of cholestasis, with a smaller proportion progressing to end-stage liver disease.^8–12 With the implementation of a variety of strategies both medical and surgical, the development of cholestasis has decreased over the past decade, and more recent reports show only 22% of pediatric IF patients developing cholestasis, with 5% progressing to end-stage liver disease.¹³ The etiology of IFALD is multifactorial, and several factors have been suggested to contribute to its development, including prematurity, sepsis, parenteral nutrition (PN) composition, and absence of enteral feeding.^14–19

PN is a significant factor in the etiology of IFALD with much of the focus on intravenous (IV) lipid emulsions (ILEs).^20–22 Conventional ILEs were primarily composed of soybean lipid or plant-based lipids that have been demonstrated to contribute to IFALD in a variety of ways, including increased phytosterol load, promotion of a proinflammatory state due to predominance of ω−6 fatty acids, accumulation of phospholipids, and inadequate provision of α-tocopherol.^23–27

SMOFlipid, a third-generation composite lipid emulsion, has been licensed in Canada since June 2013. Composite lipid emulsions are thought to have an improved lipid profile over traditional soybean lipid emulsions and have become the standard lipid emulsion for patients managed in our intestinal rehabilitation program (IRP). Our objective was to assess outcomes related to cholestasis and growth in neonates with IF who received a composite lipid emulsion compared with those who received the traditional soy-based lipid emulsion.

Methods

A retrospective cohort study was completed in infants (<12 months of age at start of PN therapy) with IF with a minimum follow-up of 12 months. All patients originated from the neonatal intensive care unit and were PN naïve at admission. Patients were recruited between January 1, 2008, and December 31, 2015, with the observation period ending on December 31, 2016. Patients were stratified into 2 groups: group 1 received SMOFlipid (2013–2015), and group 2 was a historical cohort of IF patients who received Intralipid (2008–2013). Our center does not practice lipid minimization. All IV lipids were dosed at conventional doses of 2–3 g/kg/d. Infants who were septic at admission were excluded.

Data Collection

Data were collected from our electronic patient registry. Data collection included demographics (ie, gestational age, gender, birth weight, and etiology of IF) and intestinal anatomy measured in absolute length (cm) and adjusted to the percentage of expected small bowel remaining.²⁸ We also collected anatomy related to the presence of a stoma and the ileocecal valve.

Outcome data collection included evaluation for IFALD (serum conjugated bilirubin [CB] level at 3, 6, 9, and 12 months after PN initiation; development of CB level of 34, 50, and 100 μmol/L; and administration of Omegaven). CB levels reported were based on commonly reported values of 34 μmol/L (2 mg/dL) in the literature and commonly used values within our program (50 and 100 μmol/L). Our program has used Omegaven traditionally for hepatic salvage for patients who had a CB > 100 μmol/L for >2 weeks without recent surgery or sepsis. Nutrition outcomes were collected including duration of PN therapy (days) and growth parameters (weight and height z-scores) at 3, 6, 9, and 12 months after PN initiation.

Other outcomes evaluated included sepsis (ie, number of septic events per 1000 central venous catheter days [positive blood culture treated with course of IV antibiotics]) and patient disposition (ie, duration of follow-up [days], mortality, and listed and/or receipt of intestinal/liver or multivisceral transplantation). Statistical analysis was completed using nonparametric tests based on the normality of the date. A Mann-Whitney U test was utilized for continuous variables (median and interquartile range) and χ² test for categorical variables (frequencies and proportions) with an α value of <0.05 considered significant, using IBM SPSS Statistics Version 23 (2015). Research ethics board approval was obtained for data collection and analysis.

Results

A total of 37 patients were evaluated, out of which 17 patients received SMOFlipid and 20 patients received Intralipid. Table 1 displays the patient characteristics with no significant differences between groups in gestational age, birth weight, proportion who were premature, or sex. There was also no significant difference in their underlying IF diagnosis, residual bowel anatomy, or percentage of PN support over time (Table 2).

Table 1.

Patient Characteristics.

Variable	Composite Lipid Emulsion (n = 17)	Soy-Based Lipid Emulsion (n = 20)	P-Value
Gestational age, weeks	36 (33–37)	36 (35–39)	0.311
Birth weight, kg	2.52 (2.03–2.97)	2.24 (1.98–3.24)	0.940
Premature, %	12 (70.6)	11 (55)	0.330
Male, %	11 (64.7)	10 (50)	0.368
Diagnosis			0.141
Atresia	4 (23.5)	5 (25)
Abdominal wall defect	9 (52.9)	10 (50)
Volvulus	0	1 (5)
Hirschsprung’s	0	3 (15)
Other	4 (23.5)	1 (5)

Values are medians with interquartile ranges or percentage of frequencies with proportions.

Table 2.

Patient Anatomy and PN Support.

Variable	Composite Lipid Emulsion (n = 17)	Soy-Based Lipid Emulsion (n = 20)	P-Value
IF category			0.700
Short-bowel syndrome	11 (64.7)	15 (75)
Dysmotility	4 (23.5)	4 (20)
Mucosal enteropathy	2 (11.8)	1 (5)
Percentage of small bowel remaining	28 (12–95)	57 (24–96)	0.460
Percentage of large bowel remaining	90 (63–100)	100 (50–100)	0.845
ICV resected (yes)	9 (52.9)	6 (30)	0.157
Ostomy (yes)	7 (41.2)	7 (35)	0.699
Percentage of overall energy provided by PN
PN at 3 months	100 (47–100)	100 (63–100)	0.845
PN at 6 months	57 (8–96)	35 (0–87)	0.311
PN at 9 months	38 (0–85)	0 (0–66)	0.357
PN at 12 months	24 (0–88)	0 (0–68)	0.517

Values represent frequencies (proportions) and medians (interquartile range).

ICV, ileocecal valve; PN, parenteral nutrition.

Evaluation of their median CB levels over the first 12 months of PN therapy showed a significant difference between those on SMOFlipid and those on Intralipid at 3 months (13 vs 37 μmol/L, P = 0.023) (Figure 1) with no difference at other time points. Patients who received SMOFlipid were less likely to reach a CB level of 34 μmol/L (23.5% vs 57.1%, P = 0.050) and CB level of 50 μmol/L (11.8 vs 42.9%, P = 0.036) than those who received Intralipid (Figure 2). The proportion of patients who reached a CB level of 100 μmol/L was lower in SMOFlipid patients but did not reach statistical significance (0 vs 4 [19%], P = 0.170) (Figure 2). Median peak CB levels were lower in SMOFlipid (37 vs 63 μmol/L, P = 0.056) (Figure 3).

Figure 1.

Serum conjugated bilirubin levels at 3, 6, 9, and 12 months. This figure represents the serum conjugated bilirubin levels at 3, 6, 9, and 12 months after the initiation of parenteral nutrition between patients receiving SMOFlipid compared with traditional soybean-based lipid emulsion.

Figure 2.

Proportion of patients who reached conjugated bilirubin levels of 34, 50, and 100 μmol/L. This figure compares the proportion of patients receiving SMOFlipid with those receiving traditional soybean-based lipid emulsion who reach conjugated bilirubin levels of 34, 50, and 100 μmol/L.

Figure 3.

Peak serum conjugated bilirubin levels. Comparison of peak serum conjugated bilirubin levels in patients receiving SMOFlipid with those receiving traditional soybean-based lipid emulsion.

No patients in the SMOFlipid group received Omegaven for salvage therapy of advanced liver disease compared with 30% of the Intralipid group (0 vs 30%, P = 0.014). The number of PN days in patients receiving SMOFlipid was actually longer at last follow-up (421 vs 213 days, P = 0.242) but not statistically significant. The proportion of PN energy was also greater in the SMOFlipid group, with the majority of Intralipid patients weaning off PN by 9 and 12 months. One patient in the Intralipid group no patients in the SMOFlipid group were listed for transplantation, and there was no statistical difference in central venous catheter days. Importantly, cholestasis was not confounded by sepsis. There was no difference in septic episodes per 1000 catheter days between groups. Outcome data are presented in Table 3.

Table 3.

Clinical Status at Last Follow-Up.

Variable	Composite Lipid Emulsion (n = 17)	Soy-Based Lipid Emulsion (n = 20)	P-Value
Parenteral nutrition, d	421 (203–822)	213 (104–364)	0.242
Omegaven therapy, %	0	6 (30.0)	0.014
Listed for transplant, %	0	1	1
Central venous line days	364 (211–774)	313 (126–578)	0.357
Sepsis/1000 catheter days	2.4 (0–4.2)	2.7 (0–4.8)	1

Values are medians with interquartile ranges and percentage of frequencies with proportions.

Patients on SMOFlipid had greater weight z-scores compared with patients on Intralipid at 3 months (−0.932 vs −2.092; P = 0.028) and 6 months (−0.633 vs −1.614; P = 0.018), with no difference at 9 and 12 months. There was no difference between groups for height or head circumference between our 2 groups. It is important to note anthropometrics for both groups (weight, height, and head circumference) remained within normal parameters of ±2 z-scores (Figure 4). This is relevant because no baby was managed by a lipid-restriction protocol.

Figure 4.

Growth parameters over first 12 months from initiation of PN. Comparison of patients who received SMOFlipid with those who received traditional soybean-based lipid emulsion, concerning growth parameters of height, length, and head circumference over the first 12 months after initiation of PN. PN, parenteral nutrition.

Discussion

In this retrospective cohort study, we demonstrated that neonates who were administered SMOFlipid, a composite lipid emulsion, at conventional doses, showed a lower incidence and decreased severity of IFALD over the first 12 months of administration compared with a historical cohort of patients who received traditional soybean-based lipid emulsions. Patients receiving SMOFlipid were less likely to reach CB levels of 34 and 50 μmol/L, and we were able to demonstrate that this was not confounded by septic events, which is often implicated in contributing to IFALD. No patients required Omegaven salvage of their liver function. It is important to note that the only difference in CB level was noted in the first 3 months of PN therapy. There are a few reasons for this finding. At 3 months, most patients in both groups were receiving a significant proportion of energy parenterally, and the elevated CB at that time may be a reflection of that exposure to soybean oil. Subsequently, we observed a shorter duration of PN therapy in the Intralipid group, with the majority of Intralipid patients weaning off PN by 9 and 12 months. Also, Omegaven was used for salvage therapy in the Intralipid group when CB levels reached 100 μmol/L, which would also promote resolution of CB. Interestingly, although SMOFlipid patients received a higher proportion of energy parenterally over the first year, their CB remained low. Growth parameters (weight, height, and head circumference) also remained within normal parameters, and no significant difference existed between the patients on SMOFlipid or soybean-based lipid emulsion.

The pathophysiology of IFALD is not completely understood, but it is related to the relationship between specific risk factors (PN composition including ω−6 lipid emulsions, phytosterols, reduced enterohepatic circulation, recurrent infections, prematurity) and the impact on bile flow. Reduced bile flow impacts the liver through the development of cholestasis and inflammation, which can lead to fibrosis from bile acid toxicity and inflammation.^29,30 Several strategies to improve the development and progression have been proposed, including manipulation of PN, primarily related to lipid management, early introduction of enteral feeding, and prevention of central line–associated bloodstream infections. Our center has moved from traditional soybean-based lipid emulsions to SMOFlipid, which provides a more balanced approach to lipid administration, and early results have demonstrated its efficacy in reducing the development of IFALD.^22,31

Soybean lipid emulsions have been implicated as a major contributor to the development and progression of IFALD.^29,32 Historically, 30%–90% of patients developed some degree of IFALD, with liver disease being the most significant cause of mortality in children with IF.^{3,8,9,12,14,33} Advancement in the management of children with IF has resulted in a significant reduction in mortality. A time series analysis published by Oliveira et al in 2016 evaluated significant treatment interventions that had been introduced for the management of pediatric IF and found that establishment of an IRP and introduction of ω−3 lipid emulsions had the most significant impact on the reduction of mortality in this patient population.³⁴ Within the modern era of pediatric IF management, alternative lipid-management strategies have included lipid-dose minimization or change in lipid composition.

Lipid minimization involves significant reduction of lipid dosage. Colomb et al in 2000 described their experience in 10 patients, with 30% developing an essential fatty acid deficiency and all children demonstrating inadequate growth.³² A subsequent study by Cober et al in 2012 was published comparing a cohort of patients using lipid minimization with a historical control group who received conventional lipid dosing.³⁵ Forty-two percent of lipid-minimization patients showed resolution of cholestasis compared with 10% in the historical control. Eight lipid-minimization patients developed elevated triene:tetrene ratios but did not meet the threshold for an essential fatty acid deficiency, and all children were within normal parameters for weight, height, and head circumference. Although lipid minimization is effective in resolution or prevention of cholestasis, the greater concern is the long-term impact of limiting fat on the neonatal brain. The long-term impact on brain, cognitive, and retinal development remains unknown and will be difficult to tease out in such a heterogeneous group of patients with multiple comorbidities. The other aspect of this approach that is rarely discussed is what adjustments are made to the PN components to provide adequate energy for growth. With the reduction of the fat source, other components such as dextrose and protein need to be increased. Increased dextrose (>50 kcal/kg/d) can lead to hepatic abnormalities and can increase the portal insulin:glucagon ratio with other implications for the patient.³⁶

The use of Omegaven, an ω−3 lipid emulsion, as sole lipid support at 1 g/kg/d was first reported by the Children’s Hospital of Boston, and several case series publications since that time have discussed its role in the treatment of IFALD.^19,37–39 The ω−3 polyunsaturated fatty acid, eicosapentoenoic acid, and docosahexaenoic acid content leads to a decreased leukocyte response aiding the resolution of the inflammatory response, with a number of small studies demonstrating that majority of patients experience resolution of cholestasis and inflammatory changes.^19,37–41 Original publications employed Omegaven as salvage therapy for advanced liver disease. The preferred approach is obviously prevention of IFALD rather than salvage; therefore, alternative lipid strategies are being introduced earlier in the course of illness. Lipid restriction by either minimizing the dose of soybean lipid to 1 g/kg/d or using Omegaven as sole lipid source at 1 g/kg/d will prevent cholestasis, but there is concern about the long-term impact on somatic growth and neurocognitive outcome in premature infants.

Development of third-generation, composite lipid emulsions such as SMOFlipid permits delivery of lipid energy at conventional doses while still allowing for hepatic protection. The ω−6:ω−3 ratio is 2.5:1, which is in the range to optimize bile flow.^42,43 Our program philosophy has always advocated for a balanced approach. We originally utilized fish oil emulsions in combination with standard soybean lipid emulsions for preserving adequate growth, as progression of IFALD often occurs during the neonatal period.⁴⁴ Development of composite lipid emulsions allows for intervention before the development of IFALD with the hopes of delaying or preventing the process, as well as providing adequate fat during a period of rapid growth. Results to date in studies using composite lipid emulsions have been based on short-term or small sample sizes but have demonstrated decreased CB levels or a relatively smaller increase when compared with standard lipid emulsions.^22,45–47 Goulet et al was the first to evaluate the use of SMOFlipid in pediatric patients receiving home PN in a short-term 29-day study.⁴⁵ The primary outcome was safety, but secondary outcomes were related to liver function. SMOFlipid patients demonstrated a reduction in their total bilirubin concentrations (−1.5 ± 2.4 μmol/L) compared with those on standard lipid emulsions, who demonstrated an increase from baseline (2.3 ± 3.5 μmol/L).⁴⁵ A pilot randomized multicenter trial of infants with established IFALD that had a CB between 17 and 50 μmol/L (1–3 mg/dL), over a 12-week period, demonstrated that patients receiving SMOFlipid had lower serum CB levels at end of trial and were less likely to have a CB level > 50 μmol/L, compared with those receiving traditional lipid emulsions (27% vs 69%, P = 0.04). There was also a higher proportion of patients that had a CB level of 0 μmol/L at end of trial while receiving IV lipid at conventional doses of 2.5 g/kg/d.²² Although results from previous SMOFlipid publications showed promise in the reduction of IFALD, they did not answer the question related to those who continue to receive PN for a prolonged period of time and whether the initial prevention or reduction of cholestasis would be sustained.

Our study highlights the first long-term evaluation of SMOFlipid in infants with IF and demonstrates that there is a sustained effect for children requiring prolonged PN. In our setting, SMOFlipid was initiated from admission rather than after IFALD was already established. All patients were PN naïve without established IFALD, based on biochemical monitoring. No patients required hepatic salvage with Omegaven when started on SMOFlipid during the neonatal period. In fact, no patient has been initiated on Omegaven at our institution since SMOFlipid became licensed in Canada in June 2013. Our study focused on the infant population, who are often at the greatest risk of developing IFALD because of their immature liver, high PN requirements, and growth and developmental needs.

Recent reports have highlighted the concern that although patients may exhibit biochemical resolution of cholestasis and inflammation after the introduction of Omegaven for hepatic salvage, there may be persistence or progression of fibrosis that may lead to the development of noncholestatic liver disease with portal hypertension necessitating the eventual need for liver, intestinal, or multivisceral transplantation.^48–50 Primary prevention of IFALD could be key for the prevention of noncholestatic liver disease that may result in late transplantation compared with children who developed IFALD early in their disease course.

This study has limitations that need to be acknowledged. First, although we were able to demonstrate reduced cholestasis in patients on SMOFlipid compared with standard soybean lipid, the retrospective design and small sample size did not allow us to account for all confounding variables, but we hypothesized this was related to shorter duration of PN therapy in the Intralipid group, which results in normalization of CB, as well as the greater use of Omegaven for salvage therapy. We were also unable to explain the difference in length of PN therapy between groups. Although the duration of PN therapy was longer in the SMOFlipid group, cholestasis is minimal over the first year of treatment. We also found that outcomes related to peak bilirubin levels trended toward significance, but we were likely underpowered. With respect to nutrition data, we collected information related to total parenteral energy. Data were presented as a percentage of total energy intake to determine the PN contribution. Our approach is to provide adequate energy to maintain normal growth, delivered in a balanced approach. The purpose of presenting the growth data (weight, height, head circumference) is because there has been significant controversy in the literature about lipid-management strategies to protect the liver and the impact on somatic growth. We included these growth data over the course of the first year to show that patients displayed positive growth within normal parameters. Our center does not practice lipid restriction, so lipid is delivered in conventional doses of 2–3 g/kg/d and weaned proportionally with other PN constituents as enteral feeds are advanced. With the exception of the median z-score for weight at 3 months in the Intralipid group, all patients were growing within 2 SD of z-score over the course of the 1-year study. Our findings actually show that growth in both groups was maintained.

Conclusion

The use of a composite lipid emulsion, delivered at conventional doses, resulted in a lower incidence and decreased severity of IFALD in infants receiving PN over a 12-month period. Preventing or delaying the onset of liver complications allows time for adaptation and can have a significant impact on the outcomes of children with IF. Next steps include following patients on SMOFlipid long-term to see whether the incidence of IFALD remains reduced or whether there is later progression, as well as the evaluation of neurocognitive outcomes.

Clinical Relevancy Statement.

Pediatric intestinal failure is a chronic medical condition resulting in patients remaining chronically dependent on parenteral nutrition (PN). Intestinal failure–associated liver disease (IFALD) is a common challenge in the management of pediatric intestinal failure and often results in prolonged time receiving PN and progression of liver disease. SMOFlipid has been approved for use in Canada since 2013 and is a first-line therapy for patients on prolonged PN at our institution. This study represents our early evaluation of patients using SMOFlipid compared with a historical cohort of patients who had been receiving Intralipid to evaluate early outcomes and development of IFALD.