Article Text

Download PDFPDF

Intestinal failure-associated liver disease in hospitalised children
  1. Judith Pichler1,4,
  2. Venetia Horn2,
  3. Sarah Macdonald3,
  4. Susan Hill4
  1. 1Department of Paediatric and Adolescent Medicine, University Clinic, Vienna, Austria
  2. 2Department of Pharmacy, Great Ormond Street Hospital for Children NHS Trust, London, UK
  3. 3Department of Dietetics, Great Ormond Street Hospital for Children NHS Trust, London, UK
  4. 4Department of Paediatric Gastroenterology, Great Ormond Street Hospital for Children NHS Trust, London, UK
  1. Correspondence to Dr Susan Hill, Department of Paediatric Gastroenterology, Great Ormond Street Hospital for Children NHS Trust, London WC1N 3JH, UK; susan.hill{at}gosh.nhs.uk

Abstract

Objective and aim Liver disease is a potentially life-threatening complication of intravenous/parenteral nutrition (PN). Our aim was to determine the incidence, aetiology and outcome of intestinal failure-associated liver disease (IFALD) in hospitalised children treated with long-term PN (>27 days).

Methods Over 4 years all long-term intestinal failure (IF) patients were reviewed for the possible predisposing factors of age, diagnosis, PN lipid, sepsis, length of PN treatment and length of hospitalisation. Outcome measures were IFALD incidence, severity and prognosis.

Results Of 60/279 (22%) children aged 0–18 years who developed IFALD, 13 (5%) progressed to type 3/end stage disease. IFALD was associated with younger age (p=0.03), longer treatment (p<0.001), longer hospitalisation (p=0.01), surgical diagnosis (p=0.005) and prematurity (p=0.03). IFALD was not associated with sepsis. Intestinal surgery was associated with IFALD independently of age (p=0.03). Survival was 86%, with three deaths attributed to IFALD (1% of all cases), all of which were surgical.

Conclusion IFALD incidence was lower than previously reported in paediatric patients, with surgical neonates at greatest risk.

View Full Text

Statistics from Altmetric.com

Introduction

Parenteral nutrition (PN) is essential supportive treatment for severe intestinal failure (IF). Long-term PN is defined as treatment for >27 days.1 Potentially life-threatening intestinal failure-associated liver disease (IFALD) may affect 40–60% of paediatric patients on long-term PN.2,,4 Risk factors include prematurity, lack of enteral intake, PN lipid and sepsis.2 3 This study reviewed the incidence, aetiology, severity and outcome of IFALD in hospitalised children on long-term PN.

Methods

All children treated for IF in our tertiary hospital from 2006 to 2010 were reviewed. Dates of commencing and stopping PN, gender, age, diagnosis (categorised as primary digestive disease (PDD) or primary non-digestive disease (PNDD)), prematurity, duration of hospitalisation, PN lipid, liver histology, sepsis and survival/cause of death were recorded.

Laboratory tests recorded at the start, end and every 4 weeks included blood plasma bilirubin concentration, alanine transferase (ALT), alkaline phosphatase (ALP) and γ-glutamyl transferase (γ-GT).

Type 1 IFALD was defined as ALT, ALP and/or γ-GT 1.5× the reference range for ≥6 weeks, and total bilirubin <50 µmol/l; type 2 as type 1+total bilirubin 50–100 µmol/l (≥50% conjugated); and type 3 as end-stage liver disease with clinical manifestations and total bilirubin >100 µmol/l.5

What is already known on this topic

  • Great awareness of intestinal failure-associated liver disease (IFALD) in children and in adults has developed over recent decades.

  • The reported incidence of IFALD with long-term parenteral nutrition in children is 40–60%.

  • The development of IFALD is related to age at starting parenteral nutrition, prematurity and, in neonates, abdominal surgery.

What this study adds

  • IFALD was lower than previously reported in children with only 22% of patients on long-term parenteral nutrition affected.

  • Liver disease was reversible in all but the sickest children with other major organ failure.

  • In contrast to previous studies, liver disease was not associated with sepsis, perhaps because of early broad spectrum antibiotic treatment.

Sepsis was defined as a positive peripheral/central venous blood culture from a symptomatic child with no other infection focus. Blood cultures were always taken before starting antibiotics.

Parenteral nutrition

PN was individually formulated and prescribed according to ESPGHAN/ESPEN guidelines.6 Soya Intralipid (Fresenius KABi, Uppsala, Sweden) 20% was changed to either medium chain triglyceride (MCT)/long chain triglyceride (LCT) Lipofundin (B Braun, Melsungen, Germany) or soya/MCT/olive/fish oils (SMOF, Fresenius KABi, Uppsala, Sweden) if ALP, ALT or γ-GT were 1.5× the reference range or bilirubin was >50 µmol/l for ≥2 weeks (<1 week in infants). PN was ‘cycled’ and enteral feed introduced early. PN was lipid-free 28 days after bone marrow transplantation.

The Local Research Ethics Committee Chair confirmed ethics approval was not required as this was a retrospective study.

We used non-parametric methods to analyse data together with logistic regression to define determinants of IFALD. All analyses were conducted using SPSS V.18.

Results

Patient demographic characteristics

In total, 279/1628 children (male:female ratio: 147:132, 53%:47%) had long-term IF (see table 1 for details). Median age was 1.3 years with 125 children under 1 year of age, including 46 (17%) neonates and 36 born preterm. Neonates started PN within a few days of birth, some following preadmission at another hospital. Median PN duration was 54 days and median hospitalisation was 92 days at our hospital.

Table 1

Characteristics of children with and without IFALD

Intestinal failure-associated liver disease

Sixty (22%) of the 279 children developed IFALD (male:female ratio: 25:35, 42%:58%). Please see table 1 for IFALD details and table 2 for liver function tests. Children with IFALD started treatment at a median age of 0.6 years (0.3 years, type 3) versus 1.5 years for those with non-IFALD (p=0.03). Thirteen (36%) of 36 premature neonates developed IFALD.

Table 2

Change in liver function tests on PN treatment

Thirty one children (52% of cases) had type 1 IFALD, which progressed to type 2 in 29 and deteriorated to type 3 in 13 (22% of cases). In 41/60 (69%), IFALD commenced within 6 weeks of beginning PN treatment.

IFALD was associated with longer PN treatment (median 77 days) (p<0.001) and hospitalisation (median 119 days) (p=0.01) versus non-IFALD. Significantly more IFALD cases were treated for >90 days (p=0.007).

See table 1 for IFALD incidence according to diagnosis. Twenty four of 154 (16%) patients with PNDD and 36 (29%) of 125 with PDD developed IFALD (p=0.01). IFALD-affected PDD was associated with prematurity (p<0.001) and younger age at starting PN (p<0.001) versus PNDD. Intestinal surgery presented the greatest risk (p=0.005; 25/75, 32% of those affected). With IFALD as the dependent variable, logistic regression analysis found surgical diagnosis to be a significant IFALD predictor independent of prematurity (OR 0.31; 95% CI 0.06 to 0.95; p=0.03).

IFALD with lipid-containing PN was significantly higher in PDD (p=0.02; 36/88 vs 20/103 in PNDD). Twenty eight of 60 (47%) changed lipid solution. Of the 33 (12%) children out of 279 (31 after a bone marrow transplant) who received lipid-free PN, four developed IFALD.

A total of 153 (55%) of the 279 children developed catheter-related sepsis; 38 of the 60 with IFALD and 115 of the 219 with non-IFALD (p=0.1).

Twelve children with IFALD and two with non-IFALD underwent liver biopsy. Histology was consistent with IFALD in six and showed mild cholestasis (one child with type 2 IFALD), fibrosis and mild cholestasis (one with type 1, one with type 3), porto-portal bridging fibrosis and moderate cholestasis (one with type 2), and cirrhosis and severe cholestasis (two with type 3). Histology was related to underlying diagnosis in three cases of graft-versus-host-disease and one of haemophagocytic lymphohistiocytosis. Four cases (two IFALD, two non-IFALD) had non-specific histology.

Outcome and mortality

The survival rate was 86%, with 40 deaths (23 attributed to underlying disease and 17 to PN: 14 sepsis, three IFALD) and 239 overall survivors. All three IFALD children who died (1% of all cases) could not be fed enterally due to severe intestinal dysmotility, had intestinal surgery, had other life-threatening disease and were too small and sick for liver and possible intestinal transplant listing.

The survival rate was similar for IFALD (47/60) and non-IFALD (193/219) (p=0.1). Forty out of 60 (67%) children with IFALD were weaned off PN, as were 166 of the 216 with non-IFALD. Overall, 12% of children with IFALD and non-IFALD were discharged home on PN.

Children with type 3 IFALD had a significantly worse outcome with only four out of 13 being weaned off PN (p=0.001) and eight out of 13 dying (p<0.001) compared to all other cases.

Liver enzymes normalised in 18 (38%) of 47 IFALD survivors (12 type 1, five type 2 and one type 3, who was removed from transplant listing) and improved in the remaining 26.

Discussion

This study provides data on the epidemiology, risk and outcome of IFALD in infants and children with PNDD and PDD.

The 279 IF children who required long-term treatment had a wide variety of PNDD and PDD aetiologies, demonstrating that long-term IF co-exists with many underlying diseases. The higher proportion of older children than described previously probably reflects the improved survival of children with complex PNDD diagnoses and co-morbidities referred for specialist tertiary centre management who survive with IF as a complication of aggressive treatment with, for example, chemotherapy, or bone marrow transplant.

IFALD incidence at 22% was lower than the 40–60% reported previously.2,,4 Aetiological factors were prematurity, intestinal surgery, longer PN treatment and longer inpatient stay.

The highest risk patients were surgical neonates, but even this group had a lower incidence of 32% compared with the 67% described previously in neonates on long-term PN after intestinal resection.7 The immature neonatal liver and bile metabolism are particularly vulnerable to PN-related toxicity.2 3 Intestinal surgery was an independent risk factor.

The better outcome might be due to our management strategies of changing the PN lipid source at liver disease onset, early enteral feeding, cycling PN, early suspected sepsis treatment and prescribing according to ESPGHAN/ESPEN guidelines.6

Improved outcome associated with reducing intravenous soya lipid was in keeping with previous reports that substituting omega-3 fish oil-based lipid for intravenous soya lipid reversed severe cholestasis in IFALD infants.3 8

The lack of association of septicaemia with IFALD may have been partly due to the higher median age of patients compared to previous studies (more mature livers may be less vulnerable to damage from sepsis) as well as rapid treatment.2 3

Children with liver disease of uncertain aetiology underwent liver biopsy. Histological findings confirmed clinical diagnoses and the importance of considering aetiologies other than IFALD, particularly in PNDD.

Two-thirds of cases of IFALD developed within 6 weeks of commencing treatment, in keeping with the conjugated bilirubin increase within 2 weeks and elevated liver enzymes within 4–6 weeks of starting PN previously described.2 3

Children with IFALD may have had more severe underlying disease since PN was used for longer and median inpatient stay was about 4 months as compared to under 2 months for non-IFALD. The IFALD association with prolonged PN use has been recognised previously.2 3

Most patients (74%) were weaned off PN. The early introduction of enteral nutrition is key to achieving intestinal autonomy by stimulating and enhancing intestinal adaptation.2 3 Weaning is usually gradual; the proportion of PN required versus the enteral feed should be assessed daily even when only a minimal amount is tolerated, for example 10 ml/day.

The overall survival of 86% was similar in children with and without IFALD, indicating that IFALD is usually reversible. Infants with the most severe/type 3 IFALD had the worst outcomes, with three deaths due to IFALD. Every effort should be made to improve enteral function and wean children with type 3 IFALD from PN for the best possible chance of survival.

A similar proportion (12%) of children with IFALD and with non-IFALD had persistent IF despite treatment, were stable on PN treatment and were successfully discharged home on overnight treatment.

A limitation of the study was that it was retrospective. However, management inconsistencies were minimised since patients were treated in one centre with PN management guidelines. Our suggestions that changes in practices, for example the provision of an alternative lipid, cycling PN and early enteral feeding, may have improved outcome need randomised case–control studies for confirmation.

In summary, there was a lower incidence of IFALD and better outcome than described previously in hospitalised children even when similar disease groups were compared. Improved outcome was associated with newer management strategies including use of mixed intravenous lipid, cycling PN, early enteral feed introduction and urgent treatment of suspected sepsis. Premature neonates undergoing intestinal surgery are at the greatest risk.

References

View Abstract

Footnotes

  • Competing interests None.

  • Ethics approval The Local Research Ethics Committee approved this study.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.