INTRODUCTION — A substantial proportion of infants who depend upon parenteral nutrition (PN) because of intestinal failure develop cholestatic liver disease, known as intestinal failure-associated liver disease (IFALD) or PN-associated liver disease (PNALD) [1]. The disorder is particularly common among infants with bowel resection (short bowel syndrome) and/or prematurity and is a major cause of morbidity, intestinal and/or liver transplantation, and death in these patients. It is less common and generally less severe among older children and adults but occasionally develops in those undergoing long-term treatment with PN.
The pathogenesis, clinical features, and management of IFALD are discussed in this topic review. Related discussions about the management of PN and enteral nutrition are discussed in separate topic reviews:
●(See "Parenteral nutrition in infants and children".)
●(See "Parenteral nutrition in premature infants".)
●(See "Overview of enteral nutrition in infants and children".)
●(See "Approach to enteral nutrition in the premature infant".)
DEFINITIONS
●Intestinal failure-associated liver disease (IFALD) – IFALD is defined as liver disease that arises as a consequence of the medical and surgical management strategies for intestinal failure. This replaces the term parenteral nutrition-associated liver disease (PNALD), which recognized the important role of parenteral nutrition (PN) in the pathogenesis of the disease. However, the broader term IFALD is now preferred, in recognition that there are multiple other patient-dependent and treatment-related contributors and risk factors to the pathogenesis of IFALD, as discussed below [2]. (See 'Pathogenesis' below.)
●Cholestasis – Cholestasis is typically defined as an elevated serum direct bilirubin ≥2 mg/dL (34.2 micromol/L). This threshold is used in most of the research studies on IFALD. Measurements of conjugated bilirubin are sometimes used (with the same threshold), although the normal range for conjugated bilirubin is somewhat lower than for direct bilirubin. (See "Approach to evaluation of cholestasis in neonates and young infants", section on 'Definitions'.)
EPIDEMIOLOGY — Severe and progressive IFALD is most common in infants and neonates, particularly those with short bowel syndrome [3,4]. In a systematic review, the overall incidence of cholestasis in preterm infants receiving prolonged parenteral nutrition (PN; ≥14 days) was 25 percent and the incidence of IFALD in pediatric patients with intestinal failure was approximately 50 percent, although the incidences varied widely across studies [5]. Moreover, the incidence of PN-associated cholestasis was directly proportional to the duration of PN (15.7 percent for infants receiving PN for ≤1 month and 60.9 percent for those receiving PN for ≥2 months). (See "Chronic complications of short bowel syndrome in children", section on 'Intestinal failure-associated liver disease'.)
IFALD is also seen in 40 to 60 percent of children who receive long-term PN and 15 to 40 percent of adults on home PN [1,3,5,6].
PATHOGENESIS — The mechanism of IFALD is uncertain but is probably multifactorial. Components of the parenteral nutrition (PN) probably act as causative factors, but bacterial endotoxins and lack of enteral feeding also play significant roles [7]. The risk factors described below are closely correlated. For example, premature infants are also more likely to have complications, such as necrotizing enterocolitis, that result in short bowel syndrome.
●Immature liver function – Risk factors for IFALD include prematurity, low birth weight, and intrauterine growth restriction, suggesting that immaturity of the liver is a predisposing factor [2,7-12]. As examples:
•In one report, birth weight <500 g was associated with a 30-fold increase in risk for IFALD as compared with birth weight >750 g [8].
•In a group of 62 premature infants with birth weight <2000 g, cholestasis developed in 50 percent of those with birth weight <1000 g but only in 7 percent of those with birth weight >1500 to 2000 g [11]. The incidence of cholestasis approached 90 percent for infants on PN for more than 90 days, combining both birth weight categories.
The immature liver appears to be more sensitive to the other pathogenic factors discussed in the following sections [13].
●Infection and inflammatory mediators – Episodes of sepsis, including catheter-related sepsis, are associated with the severity of liver disease in IFALD [14,15]. Bacterial translocation from overgrowth in the small intestine also may contribute [16]. Other types of infections unrelated to central lines or postsurgical dilated intestines with bacterial translocation may contribute to IFALD. These include various viral infections or, classically, Escherichia coli urinary tract infections. Endotoxins and inflammatory mediators have been shown to decrease activity of key bile acid transporters in the liver, which is one explanation of the mechanism of sepsis-associated cholestasis [17].
●Short bowel syndrome – Cholestatic liver disease is more common in patients with short bowel syndrome than in patients who require PN for other reasons, suggesting a role for factors unique to short bowel syndrome in the pathogenesis of cholestatic liver disease [18]. The following factors have been proposed, but their relative importance is uncertain:
•An exaggerated proinflammatory cytokine response [19]
•Increased bacterial translocation and sepsis [16]
•Altered gut hormone secretion and bile flow
•Nutrient deficiencies, including essential fatty acids, carnitine, choline, taurine, vitamin E, and selenium [20]
The duration of PN and extent of bowel loss are important determinants of IFALD in infants with short bowel syndrome [21]. Infants with extensive small bowel resections (eg, residual length ≤10 percent of expected) are more likely to develop cholestasis as compared with those with less extensive bowel resections or intestinal dysmotility [22,23].
●Type of parenteral lipid – Soybean lipid emulsions (eg, Intralipid) are the most common form of lipid used for PN in the United States and many other countries and have been implicated in the pathogenesis of IFALD. These emulsions are rich in omega-6 fatty acids.
Accumulating evidence suggests that soy-based lipid emulsions are important contributors to the pathogenesis of IFALD [2]. Studies in animal models showed that a soy-based lipid emulsion is associated with liver injury as compared with a fish oil-based emulsion, which is rich in omega-3 fatty acids [24,25]. Moreover, studies in adults have shown that soybean lipid infusions of >1 g/kg are one of several risk factors for developing cholestatic liver disease [6]. Studies of limiting the lipid dose in pediatric populations have mixed results (see 'Efficacy' below). Finally, studies in infants demonstrate a reversal of IFALD after changing from soy-based to fish oil-based fat emulsions. (See 'Fish oil-based lipid emulsions' below.)
A proposed mechanism through which soy-based lipid emulsions might cause liver injury is that excessive omega-6 fatty acids have proinflammatory effects and impair triglyceride export [26]. By contrast, omega-3 fatty acids tend to have antiinflammatory and insulin-sensitizing effects, acting through the GPR120 receptor [27]. Soy-based lipid emulsions also contain phytosterols (eg, stigmasterol, beta-sitosterol, and campesterol) that reduce bile transport and hepatocyte secretion and may contribute to liver injury [28-30].
●Other PN components – Other PN components that may contribute to IFALD include:
•Amino acid dose and composition – A role for amino acids in the pathogenesis of IFALD in humans has been suggested but not established. In animals, amino acid infusion impairs bile secretion and some amino acids (eg, glycine, methionine) are hepatotoxic [31,32]. Potential mechanisms are:
-Amino acid dose – Clinical evidence is inconclusive for whether the total amino acid is associated with IFALD. One study found that short-term amino acid infusions in infants cause biochemical evidence of cholestasis [33]. However, other clinical studies have failed to confirm an association between the total dose of amino acids and IFALD. A randomized trial examined the effects of two doses of amino acid supplementation as part of PN given to premature infants [34,35]. There was no difference in the rates of cholestasis at 28 days of life (12.5 percent in patients given the lower dose of amino acids versus 8.6 percent in those given a higher dose), but the study was underpowered to detect such an effect. Similarly, a retrospective study did not find a difference in the protein: nonprotein calorie ratio among infants with cholestasis as compared with those without cholestasis [36].
-Photodegradation – A possible mechanism for adverse effects of amino acids are toxic effects from products of photodegradation [37,38]. This is based on indirect evidence from animal models and clinical evidence of potentially toxic byproducts in infants given light-exposed PN [39,40]. When exposed to light, amino acids and other components of PN can form free radicals, lipid peroxides, or other degradation products. In particular, photooxidation generates hepatotoxic free radicals from the amino acids tryptophan and tyrosine, and riboflavin acts as a photosensitizing agent to accelerate this process [41].
To reduce this risk, PN solutions ideally should be protected from light before and during the infusion, as recommended by the European Medicines Agency, typically by putting opaque covers over the bags and infusion set [42].
-Amino acid imbalance – Another possible mechanism for a toxic effect of amino acids is an imbalance in several conditionally essential amino acids. In particular, taurine deficiency, L-cysteine deficiency, and methionine excess have been implicated as causes of IFALD. Low concentrations of taurine relative to glycine promote glycine-conjugation of bile acids, which are hepatotoxic [43,44]. To prevent taurine deficiency, amino acid products for neonates contain higher concentrations of taurine compared with amino acid preparations designed for adults because neonates have limited ability to convert methionine into taurine. Thus, neonates receiving the "adult" product may develop taurine deficiency, which may contribute to the development of IFALD [45]. (See "Parenteral nutrition in premature infants", section on 'Amino acids'.)
•Carbohydrate excess – Although dextrose is not considered hepatotoxic, excessive carbohydrate in PN may contribute to the development of IFALD [2]. The proposed mechanism is that excessive carbohydrates enhance insulin release, leading to upregulation of enzymes involved in fatty acid synthesis, which causes hepatic steatosis [46,47]. Hepatic steatosis is thought to predispose to inflammation via oxidative stress from reactive oxygen species.
•Toxicity from trace elements – Concerns have been raised that long-term exposure to trace elements added to PN solutions may contribute to IFALD or have other toxic effects, based on clinical inference. As a result, many institutions routinely reduce or eliminate trace element infusions in PN after cholestasis develops (eg by reducing the dose of trace elements in the PN by 50 percent) [48]. If this approach is taken, careful monitoring for copper deficiency is imperative, with periodic measurements of copper and ceruloplasmin. One case series discussed the development of clinical symptoms of copper deficiency after the removal of copper from their PN solutions [49]. In addition, a zinc supplement should be added to maintain the previous level of zinc supplementation. The recommended dose for parenteral zinc is [50]:
-<10 kg – 50 to 150 micrograms/kg/day
-10 to <40 kg – 50 to 125 micrograms/kg/day (maximum daily dose 5000 micrograms/day)
-≥40 kg – 2000 to 5000 micrograms/kg/day
(See "Overview of dietary trace elements", section on 'Copper' and "Overview of dietary trace elements", section on 'Manganese'.)
The potential for toxic effects of copper and manganese arises because they are primarily excreted through the biliary tract, so they tend to accumulate in individuals with cholestasis. Some studies and case reports document elevated serum or liver levels of these elements in patients on long-term PN, raising concerns about the possibility of hepatic injury or neurotoxicity [51-54]. However, there was no direct evidence that this had adverse effects in these patients.
•Aluminum toxicity – Aluminum toxicity can predispose to cholestasis in infants. Neonates and premature infants receiving long-term PN therapy are at the highest risk. Aluminum is found in raw materials and may become incorporated into products during the manufacturing process. In addition, aluminum can contaminate solutions as a result of leaching from glass containers during storage. Although the aluminum content of PN solutions has been substantially reduced during the past 30 years, several commonly used solutions that are used for PN, including calcium gluconate and phosphate salts, contain clinically significant levels of aluminum. Aluminum promotes hepatocellular uptake of transferrin, which allows aluminum to enter the hepatocyte bound to transferrin, an effect that can theoretically lead to cholestasis [55].
There is no consensus to define "safe" levels of parenteral aluminum intake. Since 2004, the US Food and Drug Administration (FDA) and the American Society for Parenteral and Enteral Nutrition have recommended limiting parenteral aluminum exposure to less than 4 to 5 micrograms/kg/day [56,57]. However, it is often difficult to achieve exposures under these limits using available products, particularly for preterm infants who typically have high calcium needs. Even when the solutions with the lowest-available aluminum concentrations are used, the aluminum exposure from PN solutions exceeds the FDA recommendation for infants and children weighing less than 50 kg [58,59].
•Toxic effects from intravenous tubing – Use of intravenous tubing that contains di(2-ethylhexyl) phthalate (DEHP) has been implicated in the pathogenesis of IFALD, based on limited evidence. Nonetheless, we suggest using DEHP-free tubing for PN in at-risk infants, if it is available.
In one study, changing to DEHP-free intravenous tubing was associated with a marked reduction in the incidence of cholestasis among neonates receiving PN for two or more weeks (from 50 to 18 percent) [60]. Determining whether DEHP has a causal role in the pathogenesis of IFALD will depend on whether these findings are reproduced in other care settings, as well as analysis for potential confounders.
CLINICAL FEATURES
Presentation — Infants with IFALD typically develop jaundice and elevated levels of conjugated (or direct) bilirubin approximately two weeks after starting parenteral nutrition (PN), but the onset may be later.
The primary laboratory finding is elevated conjugated bilirubin. Direct bilirubin ≥1 mg/dL (17.1 micromol/L) or conjugated bilirubin ≥0.5 mg/dL (8.6 micromol/L) is abnormal and warrants evaluation and general measures to prevent progression. Direct or conjugated bilirubin ≥2 mg/dL (34.2 micromol/L) indicates significant cholestatic liver disease and warrants further intervention. This threshold is often used for either direct or conjugated bilirubin, although the normal range for direct bilirubin is somewhat higher. (See "Approach to evaluation of cholestasis in neonates and young infants", section on 'Definitions'.)
Aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyl transpeptidase (GGTP) may be mildly elevated [13,61].
Evaluation — Cholestasis should be attributed to PN only after other major causes have been excluded (table 1). Common causes of transient conjugated hyperbilirubinemia in neonates include sepsis or infection. All infants with sustained conjugated hyperbilirubinemia should be evaluated for biliary obstruction (eg, biliary atresia), infection, and several metabolic and genetic liver diseases that require specific therapy. Several of these disorders require urgent management. (See "Causes of cholestasis in neonates and young infants" and "Approach to evaluation of cholestasis in neonates and young infants".)
Histopathology — Histologic changes in IFALD range from steatosis to cirrhosis. Steatosis (fatty liver without inflammation) is often the first histologic change seen on liver biopsy and is particularly common in adults on PN [62]. Continued exposure to PN leads to progressive changes, including hepatocellular ballooning with steatosis, portal inflammation, canalicular and intracellular cholestasis (bile plugs), and bile duct proliferation that may mimic biliary obstruction [63]. The degree of fibrosis can range from minimal portal fibrosis to cirrhosis. It is important to recognize that subcapsular wedge biopsies obtained at operation can overestimate the degree of fibrosis; percutaneous needle biopsy usually provides more accurate staging of fibrosis.
Clinical course — The liver disease usually is progressive as long as PN is continued. It may progress more slowly if oral feeds are advanced and PN is decreased and if recurrent infections are avoided.
The cholestasis and aminotransferase abnormalities tend to improve after the parenteral therapy is discontinued but may persist in some cases [64-66]. As an example, among 31 infants with IFALD who were successfully weaned from PN, conjugated bilirubin normalized at a median of 13 weeks after weaning (95% CI 8-14 weeks) and ALT levels normalized at a median of 35 weeks (95% CI 24-80 weeks) [64]. One year after weaning from PN, all patients had normal conjugated bilirubin, but 42 percent still had abnormal ALT.
PN-induced portal fibrosis often occurs early in the course of IFALD and may progress rapidly, sometimes within weeks of starting PN [63,67,68]. Fibrotic changes are common even after cholestasis has resolved. As an example, in a study of 66 children with short bowel syndrome and a history of IFALD who underwent liver biopsy, nearly 90 percent had some degree of fibrosis. Even among the children in whom serum bilirubin had normalized, 55 percent of biopsies revealed some fibrosis. Furthermore, three of the eight patients with cirrhosis did not have biochemical evidence of cholestasis at the time of the biopsy [69].
DIAGNOSIS — The diagnosis of IFALD is usually made on clinical grounds in children with intestinal failure, long-term dependence on parenteral nutrition (PN), and cholestasis that is otherwise unexplained.
Although diagnostic criteria have not been established, a typical case definition in infants includes:
●Cholestasis – Usually defined as an elevated serum direct or conjugated bilirubin ≥2 mg/dL (34.2 micromol/L)
●Duration of PN >14 days
●Exclusion of other causes of liver injury (table 1) (see "Approach to evaluation of cholestasis in neonates and young infants")
GENERAL MEASURES — For all infants who require parenteral nutrition (PN), each of the following steps helps to reduce the risk of IFALD. For those who have developed IFALD, these measures should be continued and intensified. All of these measures should be used concurrently.
Maximize enteral feeding — The most important step in management of IFALD is to promote enteral feedings to reduce the dependence upon PN [2]. Enteral feeding exposes the gastrointestinal tract to nutrient and hormonal stimuli, which are not present when the bowel is kept empty [70]. Luminal substrates and intraluminal nutrients are essential for development of intestinal adaptation [71]. Even with minimal enteral nutrition (ie, trophic feeds), human intestinal epithelial cell growth, brush-border enzyme activity, and motility are enhanced [72].
Measures that help to promote enteral feeding and weaning of the PN solution include (see "Management of short bowel syndrome in children", section on 'Advancement of enteral feeds'):
●Initiate enteral feeds early in the postoperative period.
●Encourage nonnutritive suckling for all infants, including those who cannot tolerate even trophic feedings, because this facilitates the development of sucking behavior, may improve tolerance to enteral feedings, and assists in the transition from tube to bottle feedings [73].
●Initially, use continuous feeds because these are often better tolerated than bolus feeds and may hasten weaning from PN. As enteral feeding progresses (eg, when enteral feeds provide one-half of energy requirements), convert to intermittent bolus feedings if tolerated.
●Use continuous rather than cycled PN for neonates to reduce the risk of hypoglycemia. However, as the infant matures, cycle the PN when they able to tolerate it, typically after 44 to 48 weeks postmenstrual age. There is some evidence that this limits PN hepatotoxicity [74].
With careful management, many children can be weaned from PN before their liver becomes irreversibly damaged, thus avoiding the substantial short- and long-term risks associated with transplantation procedures [75,76]. In infants with mild cholestasis, liver dysfunction usually improves after discontinuation of the PN. (See 'Clinical course' above.)
Limit soybean oil lipids and optimize parenteral nutrition macronutrient balance — Due to accumulating concerns that certain lipid emulsions may be important contributors to the pathogenesis of IFALD, we suggest limiting the dose of pure soybean oil lipids to 1 g/kg/day upon the initiation of PN if a prolonged PN course is anticipated (ie, >21 days), provided that adequate calories for growth can be provided through other PN components [54]. This is also an appropriate first step for infants with mild cholestasis (eg, direct or conjugated bilirubin between 1 and 2 mg/dL [8.6 to 34.1 micromol/L]). This approach is consistent with recommendations from the American Society for Parenteral and Enteral Nutrition [76]. Lipids should contribute no more than 25 percent of the nonprotein energy content of the PN [2]. (See 'Pathogenesis' above.)
This suggested dosing is based on low-quality clinical evidence for benefit:
●For treatment of infants with mild established cholestasis, a lower lipid dose of soybean oil lipids (eg, 1 g/kg/day) is associated with reversal or delayed progression of cholestasis (one small randomized study and several observational studies) [77-79]. Most of these studies found little or no effect on growth parameters. Note that this degree of lipid restriction should not be used with composite lipid emulsions that contain soybean oil, as it will increase the risk of essential fatty acid deficiency [80]. (See 'Composite lipid emulsions' below.)
●For preventing or delaying the onset of IFALD, a benefit of soybean oil lipid reduction has not been established. A randomized trial of 136 neonates without IFALD compared the effects of low-dose soy-based lipid (1 g/kg/day) with the standard dose (3 g/kg/day) [81]. After two weeks of PN, similar proportions of infants in each group had developed cholestasis (69 versus 63 percent, respectively). There were no differences in growth velocity or major neonatal morbidities at 28 days of life or at hospital discharge. Similar conclusions were reached by a smaller observational study [82].
These studies suggest that modest limitations on pure soy-based lipids are generally feasible and may be beneficial, but the evidence is not sufficient to establish a threshold of soy-based lipid emulsion that consistently reverses or prevents IFALD. More stringent soybean oil lipid restrictions (<1 g/kg/day) are sometimes used but may have adverse consequences on growth and essential acid deficiency (EFAD). (See 'Stringent lipid-reduction strategies' below.)
Changing to a different lipid preparation with reduced amounts of soy oil (eg, fish oil-based lipid emulsion or a composite lipid emulsion) is generally reserved for infants at risk for IFALD, eg, those with congenital intestinal defects, history of necrotizing enterocolitis, or severe feeding intolerance. (See 'Fish oil-based lipid emulsions' below and 'Composite lipid emulsions' below.)
Excess total energy as well as excessive contributions from carbohydrates or protein have also been associated with increased risks for IFALD [33,47,83-85]. As a result, it is important to prescribe PN within recommended ranges for each macronutrient as appropriate to the age and nutritional needs of each infant. Premature infants have especially high needs for energy and protein, as discussed in a separate topic review (see "Parenteral nutrition in premature infants"). Nonetheless, PN should be adjusted to provide the minimum level of total energy and macronutrients required for adequate growth [2]. (See 'Pathogenesis' above.)
Measures to prevent sepsis — Prevention of catheter-related sepsis is an important step in prevention and management of IFALD. Recurrent episodes of sepsis are an important risk factor for IFALD, and sepsis is one of the leading causes of death in patients with advanced IFALD [14]. The risk for sepsis may be attributed to the risk for catheter-related infection in patients on PN and to bacterial translocation as a result of bacterial overgrowth in patients who are not enterally fed [86]. The effects of sepsis on the liver may be amplified by the exaggerated inflammatory response in patients with short bowel syndrome. (See 'Pathogenesis' above.)
To minimize the risk for catheter-related sepsis:
●The PN catheter should be always managed with a strict aseptic technique by experienced clinicians or carefully trained caregivers.
●When catheter-related or other infections are suspected, they should be identified and treated promptly. Antibiotic and ethanol locks may be helpful in treating catheter-related sepsis. Recommendations for managing central venous catheters in children with chronic intestinal failure are provided in a position statement [87]. Details about techniques for catheter management are reviewed separately. (See "Intravascular non-hemodialysis catheter-related infection: Clinical manifestations and diagnosis" and "Intravascular non-hemodialysis catheter-related infection: Treatment".)
●Where available, we suggest initiating ethanol locks in patients with a history of catheter-related sepsis, consistent with a consensus statement [87]. Ethanol locks help to prevent recurrent infections in patients on chronic PN, based on limited evidence [76,88-91]. We typically use 70% ethanol daily, with a dwell time between 2 and 12 hours. The dose is tailored to the volume of the patient's catheter.
Typically, the use of ethanol locks has been limited to infants over three months of age and weighing more than 5 kg [92]. An advantage of ethanol locks is that they do not induce antibiotic resistance, unlike antibiotic lock therapy. However, ethanol lock therapy also is associated with greater risk of thrombus and shortens the life of the catheter, so we generally reserve ethanol lock therapy for infants who have experienced one or more septic episodes. Moreover, there are supply and cost issues associated with the use of ethanol lock therapy that are prompting clinicians to reevaluate this practice. (See "Lock therapy for treatment and prevention of intravascular non-hemodialysis catheter-related infection", section on 'Alternative lock solutions'.)
Rotating, short courses of antibiotics are used at many centers for various reasons including reducing the risk of bacterial overgrowth and thus the risk of sepsis. However, we do not take this approach, due to the potential for promoting antimicrobial resistance. Institutions employing this strategy have reported reductions in rates of IFALD, but the reports are unable to separate the effects of this intervention from other measures used simultaneously to manage IFALD [93-95]. For this purpose, metronidazole is most commonly used, given its broad anaerobe coverage and its own inherent antiinflammatory properties [96,97]. Some institutional protocols have alternated metronidazole with enteral gentamicin or kanamycin, with or without one or more weeks off of antibiotics between courses [94,95]. Other medications sometimes used for bacterial overgrowth prophylaxis include trimethoprim-sulfamethoxazole and rifaximin.
Other measures — In addition, the following interventions are used in an effort to prevent development or progression of IFALD:
●Review the medication list for potentially hepatotoxic medications that could be contributing to the development of liver disease. As an example, drugs such as famotidine (which is often added directly to PN) can cause elevations of serum aminotransferase levels. Cholestasis may be associated with certain antiinfectives, such as fluconazole, trimethoprim-sulfamethoxazole, erythromycin, and penicillins (eg, oxacillin or ampicillin). (See "Drug-induced liver injury".)
●For all infants and other individuals at risk for IFALD, take the following precautions with PN:
•Protect the PN solution from ambient light using ultraviolet-resistant covers for the PN bag and infusion set [42,98]. Protection from light will minimize photo-oxidation, which creates free radicals and lipid peroxides that may contribute to liver injury.
•Minimize the amount of aluminum contamination whenever possible.
•Avoid intravenous tubing that contains di(2-ethylhexyl) phthalate (DEHP), if possible.
(See 'Pathogenesis' above.)
TREATMENT OF PROGRESSIVE INTESTINAL FAILURE-ASSOCIATED LIVER DISEASE
Fish oil-based lipid emulsions — Accumulating evidence suggests that changing from a conventional soybean oil-based lipid emulsion to a fish oil-based lipid emulsion helps to reverse IFALD in infants who require ongoing parenteral nutrition (PN). The proposed mechanism is that soybean-based emulsions are thought to promote hepatic triglyceride synthesis and inflammation because of their high content of omega-6 fatty acids and phytosterols, compared with fish oil, which is rich in omega-3 fatty acids and contains no phytosterols. (See 'Pathogenesis' above.)
Indications and implementation
●Indications – For infants with marked progressive IFALD (direct or conjugated bilirubin >2 mg/dL), despite optimization of the general measures described above, and who are predicted to require PN for at least an additional 30 days, we suggest changing the lipid source to a fish oil-based lipid emulsion (Omegaven). As of 2018, this lipid emulsion was approved by the US Food and Drug Administration (FDA) for use in pediatric patients with IFALD (defined by a direct bilirubin level >2 mg/dL after exclusion of other causes of cholestasis) [99]. This preparation is also available in many countries outside of the United States.
We do not mix fish oil lipid emulsions with conventional soybean oil emulsions or use fish oil-containing composite emulsions when treating IFALD. Monotherapy with a fish oil-based lipid emulsion is supported by a position paper by the American Society for Parenteral and Enteral Nutrition [100].
●Dose – We initiate the fish oil-based lipid emulsion at a dose of 1 g/kg/day. Higher doses of this lipid emulsion (ie, 1.5 g/kg/day) may be necessary in patients with preexisting malnutrition or if there is a need to reduce the proportion of calories from carbohydrates (eg, due to hyperglycemia); these higher doses are an off-label use [101].
●Monitoring – Because of hypothetical risks for essential acid deficiency (EFAD) in infants maintained on fish oil as the exclusive source of lipids, we routinely monitor the triene:tetraene ratio at least once monthly. Triene:tetraene ratios >0.2 suggest EFAD. Clinical symptoms of EFAD include poor weight gain, rash, and thrombocytopenia. (See 'Safety' below and "Chronic complications of short bowel syndrome in children", section on 'Common deficiencies'.)
●Duration – Duration of therapy remains controversial. In one study of 48 children with IFALD who were treated with six months of fish oil monotherapy, cholestasis resolved in 71 percent [102]. Among those who resumed soybean oil (27 subjects), cholestasis recurred in 26 percent (median follow-up 16 months, range 3 to 51 months). This study suggests that long-term fish oil therapy may be warranted in some children to prevent recurrence of cholestatic liver disease, but further studies are needed.
Efficacy — Available evidence indicates that changing from soy-based to fish oil-based lipid emulsions is effective for treating IFALD, based on one small randomized trial, pair-matched and multicenter integrated studies, and many observational studies [103-108].
Evidence includes a randomized study in 16 infants with IFALD that evaluated the effectiveness of fish oil-based emulsions in comparison with soybean oil emulsions in reversing IFALD [103]. Patients were treated with either soybean oil- or fish oil-based emulsions at a dose of 1.5 g/kg/day. This small study was unable to detect a significant difference in the primary outcome for the study, which was IFALD reversal when observed at the four-month mark. However, significant differences were seen in secondary outcomes for the study: In the group given fish oil-based emulsions, there was a slower rate of increase in conjugated bilirubin levels (0.6 versus 13.5 micromol/L per week) and alanine aminotransferase (ALT) levels (1.1 versus 9.1 international units/L per week) as compared with the group given soybean oil-based emulsions. In addition, the rate of cholestasis improvement with advancement of enteral nutrition was significantly greater in the fish oil-based emulsions group.
Further evidence comes from a multicenter prospective observational study of 189 infants <2 years old with PN-related cholestasis (direct bilirubin >2 mg/dL) who were treated by changing to a fish oil-based lipid emulsion, compared with 73 pair-matched historical controls who continued on a soybean oil-based lipid emulsion [104]. Infants treated with the fish oil emulsion were more likely to achieve cholestasis resolution (65 versus 16 percent), have improvement in biomarkers of liver injury (including aspartate aminotransferase to platelet ratio index [APRI], a marker for hepatic fibrosis), and were less likely to undergo liver transplantation (4 versus 12 percent). Mortality was similar between the two groups, although the group treated with fish oil emulsion were more premature and had more severe liver disease at baseline. Infants with more severe elevations of direct bilirubin at baseline were more likely to progress to liver transplantation or death. These findings suggest important benefits of a fish oil-based lipid emulsion for PN-dependent infants with cholestasis and probably an advantage for initiation of this therapy early in the course of cholestatic liver disease [105]. The study was the basis for approval of a fish oil-based lipid emulsion by the FDA in 2018 [99]. These findings are consistent with several prior observational studies, as summarized in a meta-analysis [106]. Since most of these observational studies used a historical control group, it is possible that the improved outcomes for groups treated with fish oil may have been related to changes in management methods and/or reduction in total lipid dose. (See 'Stringent lipid-reduction strategies' below.)
Possible benefits on extrahepatic comorbidities were suggested by a retrospective study in which preterm infants who received fish oil-based lipid emulsion had a lower incidence of bronchopulmonary dysplasia and retinopathy of prematurity compared with historical controls maintained on soy-based lipid emulsion, despite greater baseline risk factors (lower gestational age and higher baseline direct bilirubin) [109].
Safety — Potential concerns about the safety of fish oil-based lipid emulsions are listed below. Existing data suggest that these do not present significant safety problems, provided that the fish oil lipid emulsion is given at a dose of at least 1 g/kg/day.
●EFAD – There is a minimal risk for EFAD if at least 1 g/kg/day of fish oil is used [54,110,111]. Similar results were seen in a study of 10 infants who were receiving no enteral nutrition [112]. One study found no EFAD during three years of follow-up of patients treated with a fish oil-based lipid emulsion, for whom PN contributed approximately 80 percent of total energy [113].
EFAD is also a risk when transitioning from PN to full enteral feedings. Given that patients with intestinal failure are prone to fat malabsorption, careful monitoring is imperative. In some cases, the use of pancreatic enzymes may be warranted to facilitate the transition. (See "Chronic complications of short bowel syndrome in children", section on 'Common deficiencies'.)
●Growth – Growth appears to be adequate in infants treated with fish oil lipid emulsions at 1 g/kg/day. A multicenter, retrospective, pair-matched study compared 82 children with IFALD treated with fish oil monotherapy (1 g/kg/day) with 41 historical controls receiving soybean oil monotherapy (up to 3 g/kg/day) [114]. Pair matching was based on baseline serum conjugated bilirubin levels and gestational age. Although growth measures (changes in body weight, height/length, and head circumference) were similar for both groups, children receiving the fish oil lipid emulsion showed an overall improved growth trajectory compared with baseline and, by 28 weeks, their mean body weight was appropriate for their age.
Similar results were seen in a prospective observational study utilizing an overlapping cohort, in which infants with IFALD who were treated with a fish oil-based lipid emulsion (dose 1 g/kg/day) experienced catch-up growth after 12 months of age so that mean growth parameters were normal by 24 months of age (weight-for-age Z-score 0.13, 95% CI -0.18 to 0.45), unlike the historical cohort treated with soy-based lipid emulsion [115].
●Bleeding – Theoretical concerns have been raised that use of fish oil lipid emulsions might increase bleeding risk because the high omega-3 fatty acid content may reduce platelet aggregation. However, a retrospective observational study found that infants treated with fish oil emulsion had a lower incidence of bleeding compared with historical controls treated with a soy-based lipid emulsion [109]. Although the study was limited by its retrospective design and likely confounders, it provides some reassurance that fish oil-based emulsions probably do not substantially increase bleeding risk.
Composite lipid emulsions — A composite emulsion consisting of soybean, medium-chain triglyceride, olive, and fish oil lipid emulsion (SMOF) has been used in hopes of delaying the progression of IFALD, but the evidence is insufficient to determine whether SMOF is useful for either treatment or prevention of IFALD [116-120]. Despite this lack of evidence for benefit compared with soy-based lipid emulsions, SMOF is often used as a first-line lipid emulsion in Europe for infants with mild cholestasis. SMOF is approved in the United States as a parenteral lipid source for infants but is not specifically approved for treatment of IFALD [121].
A few studies suggest possible benefits of SMOF [116,122-125]. In a small randomized trial, infants with cholestasis who were treated with SMOF were less likely to have disease progression compared with those who continued on a soy-based lipid emulsion (conjugated bilirubin exceeding 50 micromol/L, 27 versus 69 percent) and more likely to have resolution of cholestasis (27 versus 9 percent) [122]. However, other studies and a meta-analysis did not find a clinically significant benefit of SMOF in infants [116-120,126].
Limited evidence suggests that SMOF may be useful for children requiring long-term PN after hospital discharge. In one study of children on long-term home PN (mean age seven years), changing to SMOF permitted an improved macronutrient balance in PN (increased dose of lipids and corresponding decrease in glucose), with no biochemical or clinical evidence of essentially fatty acid deficiency or other adverse effects [127]. Prior to this intervention, most participants had been maintained on a soy-based lipid emulsion at a low dose (1 g/kg/day) to reduce the risk for IFALD.
An important concern is that SMOF predisposes to EFAD because of its relatively low content of essential fatty acids. Several case reports describe infants who developed IFALD while on SMOF lipid emulsion at doses between 0.5 and 2 g/kg/day, which improved when therapy was switched to an emulsion containing only fish oil or an enteral source of essential fatty acids [80,128]. A minimum SMOF dose of 2.5 to 3 g/kg/day may be needed to provide sufficient essential fatty acids. However, these higher doses may attenuate or even nullify any hepatoprotective effect of SMOF [80,119,126].
Stringent lipid-reduction strategies — As noted above, general measures to prevent development and progression of IFALD include limiting the dose of soy-based lipid to 1 g/kg/day, provided that adequate calories for growth can be provided through other PN components [54]. (See 'Limit soybean oil lipids and optimize parenteral nutrition macronutrient balance' above.)
More severe lipid restrictions (eg, 1 g/kg given twice weekly) have been explored. Limited evidence from a small observational study suggests that this strategy helps to improve cholestasis and did not interfere with growth but was associated with markers of EFAD in 8 of 24 subjects [129]. The EFAD was detected by routine monitoring of the triene:tetraene ratio and effectively treated by liberalizing the lipid dose (eg, 1 g/kg given three times weekly).
Ursodeoxycholic acid — For infants with IFALD who are able to tolerate some enteral intake, we do a trial of ursodeoxycholic acid (UDCA; ursodiol), ie, 10 mg/kg/dose twice or three times daily [54]. A few studies suggest UDCA may have some benefit in normalizing biochemical markers of IFALD, although more robust and controlled clinical trials are necessary to provide sufficient evidence to make strong recommendations for routine use [130-132]. No parenteral form of UDCA exists. Its use is limited to patients able to tolerate enteral medications. Concerns regarding efficacy in patients with ileal resections have been raised. However, some studies suggest UDCA may be absorbed throughout the entire small intestine and, possibly, colon [133]. Of note, infants lacking distal small bowel may experience increased stool output on UDCA. In addition, some preparations of UDCA include sorbitol, an inert ingredient which can exacerbate diarrhea in patients with short bowel syndrome. These causes of increased stool output may be misinterpreted as feeding intolerance.
Although phenobarbital also has some choleretic properties, it is not helpful in preventing or treating IFALD and may delay more definitive therapy [134].
Other interventions — The mainstays of treatment for IFALD are optimization of progression to full enteral feedings and intravenous fish oil monotherapy. No other specific therapies to either prevent or reduce the severity of IFALD has been found. Strategies with some merit include the following:
●Cysteine supplementation has been suggested to counteract the effects of excessive methionine, which is thought to be a mechanism of oxidative hepatotoxicity in IFALD, as discussed above. In a small case series, N-acetylcysteine was given as a source of cysteine to two infants and one older child with IFALD [135]. Treatment for 3 to 18 months with N-acetylcysteine was associated with apparent improvement in cholestatic liver disease, as suggested by reductions in serum bilirubin, aminotransferase activities, and ferritin in all three patients. Although these results are promising, considerably more study is required before this can be generally recommended as a treatment for IFALD.
●Antioxidant therapy has been suggested as a therapeutic option in treating IFALD. This approach is based on the hypothesis that oxidative stress acts as a second "hit," leading to the cell injury and death in the setting of hepatic steatosis (see 'Pathogenesis' above). Although treatment with vitamin E appears to attenuate hepatic injury in animal models, human data are still lacking. Proponents of this hypothesis suggest that the addition of alpha-tocopherol to parenteral lipid emulsions, either at the time of manufacture or exogenously, can minimize risks for IFALD [136,137]. Other experimental evidence in animal models refutes this hypothesis [138].
●Choline deficiency is an established risk factor for developing IFALD. Choline is a precursor for phospholipid biosynthesis. In long-term PN patients, plasma free choline concentrations are significantly lower [139]. Choline deficiency has been linked with histologic and biochemical hepatic abnormalities. Studies suggest that supplementing PN with choline may be beneficial in reversing these complications. A parenteral form of choline is under investigation as a treatment for IFALD and prevention of choline deficiency in PN-dependent patients [140].
PROGNOSIS — IFALD in patients with short bowel syndrome is a major indication for combined liver-small bowel transplantation [93].
Historically, the prognosis was poor for infants with short bowel syndrome and severe and prolonged hyperbilirubinemia. As examples:
●Among infants with total bilirubin >6 mg/dL at three to six months of age, 36 percent progressed to liver failure [141].
●In another study, among infants with conjugated bilirubin >2 mg/dL before two months of age, 17 percent died or went on to liver transplantation, and among those with a maximum conjugated bilirubin >10 mg/dL, 38 percent died or went on to liver transplantation [142].
The use of fish oil monotherapy and other modern management techniques have dramatically improved these outcomes. As an example, in a cohort of 189 children <2 years of age with IFALD treated with fish oil compared with pair-matched historical controls treated with soybean oil lipids [104]:
●The risk of liver transplantation was lower with fish oil (4 versus 12 percent overall, 8 versus 24 percent at 36 weeks, and 8 versus 49 percent at 68 weeks)
●In regression analysis, mortality was lower with fish oil (10 versus 14 percent at baseline direct bilirubin of 2 mg/dL, 17 versus 23 percent at baseline direct bilirubin of 12.87 mg/dL)
Other outcomes of fish oil monotherapy are discussed above. (See 'Efficacy' above.)
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Nutrition support (parenteral and enteral nutrition) for neonates including preterm infants" and "Society guideline links: Nutrition support (parenteral and enteral nutrition) in infants and children" and "Society guideline links: Pediatric liver disease" and "Society guideline links: Short bowel syndrome".)
SUMMARY AND RECOMMENDATIONS
●Definition – Intestinal failure-associated liver disease (IFALD) is defined as liver disease that arises as a consequence of intestinal failure or its medical and surgical management. The diagnosis of IFALD is usually made on clinical grounds in children with intestinal failure, long-term dependence on PN, and cholestasis, if other specific causes of liver injury have been excluded. Cholestasis is usually defined as an elevated serum direct or conjugated bilirubin ≥2 mg/dL (34.2 micromol/L). (See 'Definitions' above.)
●Risk factors and pathogenesis – Risk factors for developing IFALD are young age (especially premature birth), short bowel syndrome, intercurrent and recurrent infections, lack of enteral feedings, and duration of PN. In addition, increasing evidence suggests that the phytosterols present in soybean oils that are used in conventional intravenous lipid emulsions may contribute to the risk for developing IFALD. (See 'Epidemiology' above and 'Pathogenesis' above.)
●Clinical features – The clinical features of IFALD are conjugated hyperbilirubinemia, with mild or moderate elevations of serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyl transpeptidase (GGTP). The liver disease is usually progressive unless parenteral therapy can be reduced or discontinued. (See 'Clinical features' above.)
●Preventive measures
•Maximize enteral feeding – For all infants requiring PN, the risk for IFALD can be reduced by adjusting PN to provide the minimum level of total energy and macronutrients required for adequate growth, transitioning to enteral feeds as rapidly as is tolerated, cycling PN when possible. (See 'Maximize enteral feeding' above and "Management of short bowel syndrome in children".)
•Limit soy-based lipids – For infants requiring PN for more than 21 days who are on a pure soybean oil lipid emulsion, we suggest administering these lipids at no more than 1 g/kg/day (Grade 2C). This suggestion assumes that adequate calories for growth can be safely provided through other PN components. Limiting soybean oil lipid emulsion to this degree may delay progression of IFALD. (See 'Limit soybean oil lipids and optimize parenteral nutrition macronutrient balance' above and 'Stringent lipid-reduction strategies' above.)
•Avoid infection – Take stringent measures to avoid and promptly treat catheter-related sepsis. (See 'Measures to prevent sepsis' above.)
●Treatment – The optimal management of IFALD has not been established, and practice varies based on expert opinion and local resources. At our institution, we take the following approach:
•General measures – For infants who develop cholestasis (direct or conjugated bilirubin ≥2 mg/dL), we ensure that each of the general measures discussed above are in place, particularly reduction of the lipid dose to no more than 1 g/kg/day and adjusting PN to provide the minimum level of total energy and macronutrients required for adequate growth. (See 'Stringent lipid-reduction strategies' above.)
•Fish oil-based lipid emulsion – For infants with progressive IFALD despite the above measures who require ongoing PN, we suggest treating with a fish oil-based lipid emulsion rather than a conventional soybean oil-based lipid emulsion (Grade 2C). Changing to a fish-oil based lipid emulsion resolves or substantially improves cholestasis in a majority of infants and is associated with lower risk of liver transplantation and death compared with matched controls treated with soybean oil lipid emulsions. If fish oil is used as the exclusive source of lipids, at least 1 g/kg/day should be given and the triene:tetraene ratio and clinical symptoms should be monitored to detect essential fatty acid deficiency (EFAD). (See 'Fish oil-based lipid emulsions' above and 'Prognosis' above.)
•Other steps
-Trace elements may contribute to the pathogenesis of IFALD, based on limited evidence. Therefore, if the direct or conjugated bilirubin is ≥2 mg/dL for 30 days despite reduction of lipid dose and other preventative measures, we reduce the dose of the trace elements in the PN by 50 percent. To avoid trace mineral deficiencies, we add a zinc supplement and monitor for copper deficiency. (See 'Pathogenesis' above.)
-For infants who develop cholestasis and are able to tolerate some enteral intake, we suggest a trial of treatment with ursodeoxycholic acid (UDCA), ie, 10 mg/kg/dose given twice or three times daily (Grade 2C). UDCA may improve biochemical markers of liver injury and has minimal adverse effects. (See 'Ursodeoxycholic acid' above.)
Do you want to add Medilib to your home screen?