Mercer, David F.*; Hobson, Brandy D.†; Fischer, Ryan T.‡; Talmon, Geoffrey A.§; Perry, Deborah A.||; Gerhardt, Brandi K.*; Grant, Wendy J.*; Botha, Jean F.*; Langnas, Alan N.*; Quiros-Tejeira, Ruben E.†
Intestinal failure–associated liver disease (IFALD) is one of the most clinically important sequelae of intestinal failure in children. Although the causes of IFALD are multifactorial (1), one important component appears to be the provision of parenteral lipids. Traditionally, soy-derived lipid emulsions (LE) have been provided at a typical recommended dose range of 2 to 4 g · kg−1 · day−1(2). In the absence of enteral feeds, infants will begin to manifest signs of cholestasis within weeks of initiating parenteral support at these levels. Additional factors important in the development of IFALD, however, include but are not limited to prematurity of hepatic enzyme systems, loss of bile acid enterohepatic circulation, repeated episodes of sepsis, and deficiencies of conditionally essential nutrients such as taurine, choline, and glutamine (1).
An important development in the management of IFALD has been the introduction of fish oil emulsions. First described in reports in 2005 (3) and 2006 (4) and examined in subsequent clinical series (5–7), these novel emulsions have gained widespread use, and have become integrated into parenteral nutrition (PN) management strategies for children with intestinal failure across North America. Although there does appear to be an evidence of a reduction in total bilirubin levels in children receiving fish oil emulsion (FOE), it is unclear whether this is an effect of the addition of omega-3 fatty acids (Ω-3 FA), lipid reduction alone, or a combination of both. Furthermore, it is not evident whether a reduction in bilirubin levels correlates with an overall improvement in survival or the achievement of enteral independence, arguably among the most important goals of intestinal rehabilitation. A report by Soden et al (8) of 2 children receiving FOE therapy showed failure of regression of fibrosis despite biochemical improvement, suggesting the pathologic process may not be receding as anticipated based on serum bilirubin levels.
At the University of Nebraska Medical Center, we have been using lipid minimization and FOE together as components of a comprehensive rehabilitation strategy. In our clinical use of FOE, we noted that in several cases, although serum bilirubin levels fell to normal levels, there did not appear to be an accompanying reversal in progression of hepatic fibrosis. To investigate this observation further, we performed blinded pathologic examination of samples from all children enrolled in our clinical trial of Omegaven (www.clinicaltrials.gov identifier NCT00826020) who had >1 liver biopsy while receiving FOE therapy, all of whom experienced complete reversal of hyperbilirubinemia.
We began enrolling patients in a Food and Drug Administration– and an institutional review board–approved prospective open-label evaluation of FOE therapy in children with intestinal failure in May 2009. After obtaining informed consent from parents, children enrolled in the protocol had baseline laboratory evaluations performed, and then had their standard parenteral lipids (Intralipid, Fresenius Kabi, Bad Homburg, Germany) stopped and replaced with FOE (Omegaven; Fresenius-Kabi, Bad Homburg, Germany) at a dose of approximately 1 g · kg−1 · day−1. Basic laboratories including liver function tests were performed by protocol at least weekly over the duration of the study, and additional laboratories specific to FOE therapy including serum triglycerides and essential fatty acid profiles were performed by protocol at predefined timepoints during the course of treatment.
In the normal course of their management, if any child required an open abdominal operation (eg, for restoration of intestinal continuity or serial transverse enteroplasty), additional parental consent was obtained for procurement of an open liver biopsy. This was done as our standard of care with children both enrolled and not enrolled in the FOE protocol. Of the 16 biopsies examined, 4 were percutaneous liver biopsies obtained in the course of evaluation for transplantation, to determine need for a simultaneous liver-intestine transplant. Although we had many patients with a single liver biopsy while receiving FOE, for the present study we selected children who had >1 biopsy during the course of their therapy. There were no other inclusion or exclusion criteria for the study.
For open liver biopsies, at the time of operation and under direct vision, a liver biopsy (∼1 cm3) was excised sharply, and bleeding from the biopsy surface was controlled with electrocautery. Percutaneous biopsies were obtained with ultrasound guidance and monitored sedation, using a Tru-Cut biopsy needle (Cardinal Health, Dublin, OH) to obtain a core biopsy 1 mm in diameter and 2 to 3 cm in length. Liver biopsies were immediately sent to pathology for formalin fixation and paraffin blocking. Representative H&E- and Masson-stained slides from each case were reviewed blindly and independently by a dedicated pediatric pathologist and a dedicated liver pathologist, and were scored based on published scoring systems for cholestasis (9) and fibrosis/inflammation (10). The level of steatosis was graded from 0 to 3 on each biopsy, and additionally, a “yes/no” score was given for ductal proliferation. When wedge biopsies were examined, the deepest levels were used to avoid overestimation of fibrosis near the liver edge. The criteria for scoring are summarized in Table 1.
All children were placed on continuous enteral feeds using an elemental formula at the earliest opportunity following surgery, supplemented with parenteral energy including FOE as required to maintain weight gain and nutritional parameters. Children were seen every 1 to 2 weeks in a dedicated clinic and reviewed by a surgeon, gastroenterologist, dietitian, feeding therapist, and nurse coordinator. As enteral tolerance advanced, parenteral energy was reduced, with an attempt to minimize lipids as tolerated, in general by progressively dropping nights of FOE. When it was determined clinically that enteral absorption of fats was sufficient to prevent essential fatty acid deficiency, FOE was stopped.
Six patients enrolled in the FOE study had >1 liver biopsy available for review. Their baseline clinical data and outlined FOE study data are presented in Table 2. The median gestational age was 35 weeks, with median birth weight of 2064 g. The common most reason for intestinal loss was gastroschisis (5/6 children), with a median intestinal length of 26 cm beyond the ligament of Treitz (measured directly at operation). Because we included any child treated with FOE having >1 liver biopsy, there was no particular selection of children with gastroschisis, and no diagnosis was excluded. Most children had lost their ileum and ileocecal valve and had roughly 2 of 3 of their colonic length, with 1 child having a small amount of terminal ileum and all of the colon.
The individual scores for each of the biopsies examined in the study are shown in Table 3, accompanied by concurrent tests of hepatocellular injury (aspartate aminotransferase/alanine aminotransferase), cholestasis (alkaline phosphatase, bilirubin), and hepatic synthetic function (albumin, prothrombin time-international normalized ratio). Triglyceride levels were also followed serially. Growth parameters and percent enteral and parenteral energy at each point are also shown, based on an average estimated total energy need of 85 kcal · kg−1 · day−1.
All patients maintained adequate growth and development during the course of the study. Total bilirubin levels normalized after a median of 6 weeks, and children received FOE for a median of 40.5 weeks (range 21–78 weeks). After obtaining 2 consecutive measurements of direct bilirubin >2 mg/dL required to enroll patients, for convenience we subsequently followed total bilirubin levels serially rather than direct bilirubin levels; as such, the true time to normalization of direct bilirubin could be less than that observed with the total values. Persistent mild-to-moderate elevations in transaminases were observed along with mild rises in alkaline phosphatase, and hepatic synthetic function was well preserved throughout. No hypertriglyceridemia was noted and there were no episodes of hypocoagulability, both potential adverse effects of FOE therapy. All patients were stable during the study, with the only major clinical events typically being central-line infections without sepsis; there was no increased frequency of these events noted over that of the general pediatric IFALD population in our program.
No patients had overt clinical signs of portal hypertension during the course of observation, such as edema, ascites, or gastrointestinal bleeding. Platelet counts were followed as an indirect marker of portal hypertension, and were found in general to fluctuate over time with half of the children showing slight rises and half showing slight drops. Ultrasounds performed around the times of biopsies were examined to determine potential changes in spleen size over time. Patients 1, 2, and 5 had only a single ultrasound evaluation, with spleens noted to be at the 50th, 99th, and 93rd percentiles, respectively, and with all having platelet counts in the normal range. Patient 3 was at the 94th percentile at 8 weeks and rose to >99th percentile at both weeks 48 and 67 with normal platelet counts throughout. Patient 4 was thrombocytopenic throughout with platelet counts of approximately 100 to 120 × 103 cells/μL and a spleen that rose from the 25th to the 97th percentile between 8 and 38 weeks of life. Interestingly, patient 6 had a drop in spleen size from the 77th to the 25th percentile between 28 and 83 weeks of life, with a rise in platelet count from 273 × 103 cells/μL to 454 × 103 cells/μL. This was the only patient in the series to show a drop in fibrosis score, detailed below.
During the timepoints studied, steatosis was either absent or mild in all biopsies, with no discernable pattern emerging. Ductal proliferation was variable during the course of study, with no pattern noted. Both cholestasis and inflammation scores were highest at the beginning of observation, and dropped during the course of treatment, with scores of either 0 or 1 in all cases at the final timepoint studied. Fibrosis scores were significantly elevated (grade 2 or grade 3) in 5 of 6 patients, and were either stable (2 patients) or rose (3 patients) during the course of observation. In only 1 case (patient 6), the fibrosis score was seen to drop from grade 2 to grade 1 between 28 and 83 weeks of life. Representative biopsies from patient 3 (Fig. 1) show progression of fibrosis during the timepoints examined, despite clinical stability and normalized bilirubin and transaminases.
All children are alive and well presently. Patients 1, 2, and 3 have made progress in weaning off PN, and presently receive 40%, 25%, and 30% (respectively) of their energy requirements parenterally. Patient 5 was taken off PN at 76 weeks of age, and continues to do well on oral diet and supplemental enteral feeds. Patients 4 and 6 went on to undergo isolated small intestine transplants; in both cases, the indication for transplantation was recurrent life-threatening central-line infections, and neither was felt to require a liver transplant. Patient 4 was making progress in reducing parenteral requirements at the time of listing, being down to ∼50% of energy intake. Patient 6, who had only 7 cm of jejunum was not able to make significant reductions in parenteral requirements despite aggressive attempts at weaning.
The management of liver disease in children with intestinal failure has changed significantly during the last decade, and has led to a reduction in the number of children developing end-stage liver disease requiring early transplantation, as reflected in the significantly reduced number of combined liver-intestine and liver-pancreas-intestine grafts performed in the United States during the last 3 years (OPTN data, accessed August 31, 2011). The increasing recognition that development of liver disease is the result of not only PN, but also of additional processes (immature enzyme systems, bacterial translocation, repeated episodes of sepsis, and so on) associated with the short bowel syndrome, outlined in a review by Goulet et al (1) has led to the use of the term IFALD, rather than PN-associated liver disease. One important element of the change in management has been increased attention to lipid management, either through strategies of lipid minimization or alternate parenteral lipid sources. The most important, and at times contentious, of these alternate lipid sources has been parenteral fish oil emulsions. Because evidence of cholestasis appears in the neonatal liver within weeks of starting PN, there is a growing drive among clinicians to initiate alternate lipid therapy as early as possible in children with intestinal failure.
Relative to traditional soy-based LE, FOE have high levels of the Ω-3 FA eicosapentaenoic acid and docosahexaenoic acid, lower levels of Ω-6 FA, higher levels of α-tocopherol, and no phytosterols (being animal-based rather than plant-based) (11). In published case series, their administration seems to be associated with clinical regression of cholestasis (6,7). Proposed mechanisms for these effects include the enhanced production of eicosanoids from Ω-3 FA, which are “less inflammatory” than those derived from the Ω-6 FA, altered intracellular lipid peroxidation, and prostaglandin-mediated enhancement of bile flow. Although all are theoretic possibilities, there is no direct evidence to support any one as clinically meaningful as yet. Although limited, there is in vitro evidence that certain phytosterols may inhibit key hepatocyte membrane transporters (12), with additional clinical evidence showing an association between serum levels of phytosterols and degree of cholestasis (13,14). The fact that lipid minimization strategies are also effective in reducing cholestasis (15) argues that the effect is related to a reduction in level of some “toxic” component of soy-based lipids, perhaps phytosterols, rather than a specific positive effect of FOE.
These data suggest that although children show biochemical improvement in cholestasis after treatment with fish oil–based LE, there does not appear to be similar histologic improvement in hepatic fibrosis, at least during the timepoints covered by this series. The degree of cholestasis and grade of inflammation of the biopsies generally reduced over time, and ductal proliferation and steatosis were variable both within and between patients. These data support what was recently reported by Soden et al (8), and expand further by showing in repeated cases not only the lack of regression of fibrosis, but clear progression despite bilirubin remaining normal. Additionally, they are in accordance with older data showing no relation between liver function tests and fibrosis in adults with fatty liver disease (16) or cirrhosis (17), or in animal models of PN-induced steatosis (18). It should also be noted that all children had been receiving PN and conventional lipid therapy, with minimal or no enteral nutrition, since birth. Many of the initial biopsies were taken right at the time of FOE initiation, or shortly thereafter, and thus reflect the antecedent therapy and not FOE treatment. This speaks to how quickly fibrosis can develop in children on conventional PN and lipids because 3 children had fibrosis greater than grade 1 at the first biopsy, with 1 child having stage 2 fibrosis at 14 weeks of life.
It was also noted in the data that during the time course of the study, most patients had persistent elevations of aspartate aminotransferase, alanine aminotransferase, and γ-glutamyl transpeptidase, although these elevations were variable to some extent both within and between patients. The scores for inflammation in the biopsies fell to zero in 3 of 6 patients, 0 to 1 in 2 of 6 and remained at a score of 1 in only 1 patient. These data suggest that there is a persistent low level of hepatocyte injury, which in some cases is beneath what is possible to observe histologically. This ongoing injury could easily contribute to fibrosis progression, and underscores the conclusion that biochemical resolution of cholestasis does not mean IFALD is arrested.
Despite persistence or progression of fibrosis, all children had well-preserved hepatic synthetic function, and although there was evidence of splenomegaly radiologically, did not manifest any clinical symptoms of liver failure or portal hypertension. This could be interpreted to mean that even the presence of significant fibrosis or cirrhosis is clinically inconsequential if sufficient progress is made in advancing enteral feeds. Because fibrosis within the liver occurs in response to a chronic ongoing insult to the hepatic parenchyma, it is possible that the reduction in histologic evidence of cholestasis and inflammation may manifest in durable stability of the degree of fibrosis, or even its eventual regression, in a timeframe outside that of the present study. Although one patient appeared to show regression during this timeframe, others could easily take longer to see the same effect. Although we did strive to look at the deepest levels of surgical biopsies to avoid artifact, it is possible that some of our results could be explained by sampling variability within the liver, with other areas perhaps showing lesser degrees of fibrosis. Because 5 of 6 children in this series purely by chance had an underlying diagnosis of gastroschisis, it is possible that there could be an undiscovered mechanism inherent to this condition, which contributes to fibrosis progression; if so, this may limit the generalizability of our findings to other disease states. Finally, as with all other presently published studies of FOE in IFALD, the present study can only show association between treatment with FOE and either alterations in liver function tests or histology. There are many factors changing in these children over time, such as improved adaptation, advancement of enteral feeds, and reduction in parenteral lipid doses, all of which are likely to have effects on overall outcomes.
We hope these histologic data will stimulate further research in the field of IFALD, both to substantiate or refute the findings of our own study and to determine the mechanism or mechanisms responsible for its development. Perhaps most importantly, we would like to see further work to determine the significance of these histologic findings on clinically relevant long-term sequelae such as enteral independence or development of complications of end-stage liver disease.
We believe the importance of these data is not to impugn fish oil emulsions as a tool in the treatment of children with intestinal failure, but rather to caution that the biochemical resolution of cholestasis is at best a weak surrogate marker for the outcomes parents and clinicians are most interested in, namely enteral independence and overall survival. There is clearly ongoing damage in most of the livers while children are anything other than enterally independent, and we believe these findings make a strong case for early referral of children with short bowel syndrome to specialized intestinal rehabilitation centers, where a comprehensive and life-long approach can be taken to optimize care. Modifying the lipid source in isolation may well not be enough. That liver disease appears to progress while biochemical tests are improving should prompt caution in accepting the potentially false security of a normalized bilirubin. Furthermore, these data may suggest potential benefit of periodic evaluation of liver fibrosis, either by biopsy or a validated noninvasive method, in any child who remains on PN.
1. Goulet O, Joly F, Corriol O, et al. Some new insights in intestinal failure-associated liver disease. Curr Opin Organ Transplant
2. Mehta NM, Compher C. A.S.P.E.N. Clinical Guidelines: nutrition support of the critically ill child. JPEN J Parenter Enteral Nutr
3. Gura KM, Parsons SK, Bechard LJ, et al. Use of a fish oil-based lipid emulsion to treat essential fatty acid deficiency in a soy allergic patient receiving parenteral nutrition. Clin Nutr
4. 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
5. 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
6. Lee SI, Valim C, Johnston P, et al. Impact of fish oil-based lipid emulsion on serum triglyceride, bilirubin, and albumin levels in children with parenteral nutrition-associated liver disease. Pediatr Res
7. Puder M, Valim C, Meisel JA, et al. Parenteral fish oil improves outcomes in patients with parenteral nutrition-associated liver injury. Ann Surg
8. Soden JS, Lovell MA, Brown K, et al. Failure of resolution of portal fibrosis during omega-3 fatty acid lipid emulsion therapy in two patients with irreversible intestinal failure. J Pediatr
9. Dixon LR, Crawford JM. Early histologic changes in fibrosing cholestatic hepatitis C. Liver Transpl
10. Batts KP, Ludwig J. Chronic hepatitis. An update on terminology and reporting. Am J Surg Path
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. Carter BA, Taylor OA, Prendergast DR, et al. Stigmasterol, a soy lipid-derived phytosterol, is an antagonist of the bile acid nuclear receptor FXR. Pediatr Res
13. Clayton PT, Bowron A, Mills KA, et al. Phytosterolemia in children with parenteral nutrition-associated cholestatic liver disease. Gastroenterology
14. Iyer KR, Spitz L, Clayton P. BAPS prize lecture: new insight into mechanisms of parenteral nutrition-associated cholestasis: role of plant sterols. J Pediatr Surg
15. Cober MP, Teitelbaum DH. Prevention of parenteral nutrition-associated liver disease: lipid minimization. Curr Opin Organ Transplant
16. Charatcharoenwitthaya P, Lindor KD, Angulo P. The spontaneous course of liver enzymes and its correlation in nonalcoholic fatty liver disease. Dig Dis Sci
17. Carlisle R, Galambos JT, Warren WD. The relationship between conventional liver tests, quantitative function tests, and histopathology in cirrhosis. Dig Dis Sci
18. Sax HC, Talamini MA, Brackett K, et al. Hepatic steatosis in total parenteral nutrition: failure of fatty infiltration to correlate with abnormal serum hepatic enzyme levels. Surgery