The objective of the study is to show that substitution of the usual lipid emulsion (Intralipid 20%®) at a dose between 0.5 and 1.0 g/kg/infusion of parenteral nutrition (n-6:n-3 ratio = 7:1) by an equivalent dose of 0.5 to 1 g/kg/infusion of another lipid emulsion, OMEGAVEN 10%® very rich in omega-3 (n-3) (n-6:n-3 ratio = 1:7) induces regression of PNALD due to the anti-inflammatory and anti-fibrotic effects of n-3 EFA.

Regression of liver disease will be defined by normalization of the five liver function tests (LFT): conjugated bilirubin, gamma GT, alkaline phosphatase, AST and ALT transaminases.

Background: Parenteral nutrition-associated liver disease (PNALD) in the context of home parenteral nutrition for non-neoplastic chronic intestinal failure is the main metabolic complication of HPN, as reflected by the combined liver and intestine transplantation rate of 45% in adults. Prior to the onset of severe fibrosis, the main basic histological lesions of cholangitis and steatosis evolve in parallel with abnormal liver function tests (LFT), the prevalence of which increases with the duration of HPN. An n-6 polyunsaturated fatty acid lipid supplement, based on soybean oil, has been shown to be a major and independent determinant of PNALD at doses greater than 1 g/kg/infusion.

Study objective: to demonstrate that replacement of the usual lipid emulsion (Intralipid 20%®) at a dose of 0.5 to 1.0 g/kg/PN infusion (n-6:n-3 ratio = 7) by an equivalent dose of a new lipid emulsion, OMEGAVEN 10%®, very rich in n-3 polyunsaturated fatty acids (n-3:n-6 ratio = 7) induces regression of PNALD via the anti-inflammatory and anti-fibrotic effects of n-3 polyunsaturated fatty acids.

Material and methods: The treated population presents with severe chronic non-neoplastic intestinal failure (rare disease), requiring HPN in an accredited centre. The median actuarial risk of PNALD is 50% after 5 years of HPN in adults. PNALD will be defined on inclusion by 2 out of 5 abnormal LFTs: ALT and AST transaminases, conjugated bilirubin, alkaline phosphatase and gamma-glutamyltranspeptidase. This multicentre prospective randomized double-blind study will include patients with no organ failure other than intestinal failure and especially no signs of decompensated cirrhosis on abdominal Doppler ultrasound. It will exclude unstable patients especially as a result of recent infection (<6 weeks). On inclusion, HPN must have been administered for at least 12 weeks and, after inclusion in the study, 6 weeks of HPN will be devoted to standardization of inter-centre practices. To be eligible for inclusion, patients must have a predictable duration of HPN of more than 18 weeks with a degree of dependence ≥ 2 nutritional infusions per week.

Other causes of liver disease will be excluded (viral, autoimmune, alcohol, hepatotoxic drugs, biliary obstruction).

The study duration per patient will be 22 weeks (6 weeks of standardization of HPN, 12 weeks of treatment in one of the two arms and 4 weeks of follow-up). HPN with Intralipid 20%® will be continued unchanged in the HPN control arm and Intralipid 20%® will be replaced by an equivalent dose (0.5 to 1.0 g/kg/Infusion) of OMEGAVEN 10%® up to a maximum dose of 40 mg per infusion due to formulation constraints, in the interventional arm.

The primary endpoint will be normalization of the 5 LFT parameters at the 12th week of treatment (W18). According to published data, the probability of normalization of LFT is less than 10% in the control arm and 50% in the treated arm. A total of 32 patients must therefore be recruited in each arm to demonstrate a significant difference of normalization of LFT at the 12th week of treatment (Chi-square or Fisher test, p<0.05) between the two groups. This population could be recruited over a period of 18 months from the population treated in the three centres participating in the study: Paris for the Ile-de-France region, Lyon for the Rhone-Alpes region and Lille for the Nord-Pas de Calais region, as these three centres follow 50% of the French adult HPN population, representing 250 patients.

A monthly safety study will be performed by clinical examination and determination of the usual laboratory parameters. Fibrosis evaluated by Fibroscan (in kilopascal) and steatosis evaluated by abdominal Doppler ultrasound (in three grades), the relevant laboratory parameters reflecting the lipid supplements such as n-3 and n-6 essential fatty acids, lipoperoxidation index, explanatory factors of the harmful effect of n-6 polyunsaturated fatty acids and the protective effect of n-3 polyunsaturated fatty acids such as proinflammatory and anti-inflammatory cytokines and serum markers of hepatic fibrosis, will be compared between the two arms at the beginning and at the end of treatment.

Expected result: This innovative prospective randomized study concerns PNALD which, at the stage of cirrhosis in adults, is responsible for the patient's death in more than 22% of cases and requires combined liver and small intestine transplantation in 45% of cases. The proposed therapeutic intervention, prior to onset of severe PNALD, is based on emerging rational hypotheses associated with a highly probable positive clinical expression.

Inclusion Criteria:

• Age greater than or equal to 18 years.
• Patients on HPN for chronic benign intestinal failure:

with a degree of HPN dependence ≥ two cycles of PN per week with at least one ternary infusion (comprising lipids) per week with a maximum lipid intake of 40 grams per ternary infusion

• Expected duration of HPN dependence greater than 18 weeks at the time of inclusion.
• Receiving HPN for at least 12 weeks in one of the three study centres, which is a sufficient period to allow resolution of any drug-induced or septic cholestasis and cytolysis related to a previous hospitalisation.
• Presence of PNALD, defined by an abnormality of at least two of the five liver function tests performed (conjugated bilirubin, gamma glutamyltransferase, alkaline phosphatase, AST, ALT).
• Stable patient with no infection during the six weeks preceding inclusion.
• Medical examination performed before inclusion.
• Written informed consent.
• Covered by French national health insurance.

Exclusion Criteria:

• Active cancer, regardless of the primary site.
• Uncontrolled cardiopulmonary insufficiency.
• Decompensated cirrhosis.
• Severe renal failure.
• Uncontrolled diabetes or endocrinopathy.
• Hyperlipoproteinaemia and hypertriglyceridaemia (≥ 3 mmol/L).
• Other causes of liver disease (biliary obstruction, alcohol, hepatitis B virus, hepatitis C virus, CMV, hepatotoxic drugs).
• Systemic corticosteroid therapy or biotherapy (anti-TNF).
• Pregnant women or nursing mothers.
• Inclusion in another study terminated or less than three months.
• Known allergy to fish or egg proteins.