Yeop, I.*; Taylor, C.J.§; Narula, P.§; Johnson, L.†; Bowen, C.‡; Gupte, G.L.*
‡Department of Histopathology, Birmingham Children's Hospital, Birmingham
§Department of Paediatric Gastroenterology, Sheffield Children's Hospital, Sheffield, UK.
Address correspondence and reprint requests to Dr Girish Gupte, Consultant Paediatric Hepatologist, Birmingham Children's Hospital, Steelhouse Lane, Birmingham B4 6NH, UK (e-mail: email@example.com).
Received 13 January, 2011
Accepted 8 July, 2011
The authors report no conflicts of interest.
A male child born at 25 weeks of gestation to nonconsanguineous parents (with no history of recreational drug misuse) with gastroschisis developed short bowel syndrome with a residual bowel length of 45 cm without ileocaecal valve. He subsequently underwent a laparotomy for feeding intolerance and had serial transverse enteroplasty, following which his gut was lengthened to 75 cm.
The patient continued to have difficulties establishing full enteral feeds and was dependent on parenteral nutrition (PN) throughout his life. His clinical course was also complicated by 8 episodes of line-related Enterococcus sepsis, which were treated promptly. Enteral feeding was continued but limited because of recurrent vomiting; however, he developed intestinal failure–associated liver disease (IFALD) and was referred for transplant assessment.
During the transplant assessment, at the age of 9 months (5 months corrected for gestational age) he was receiving 30 mL · kg−1 · day−1 of PN. An abdominal ultrasound scan showed coarse echotexture of liver and significant splenomegaly extending to the left iliac fossa, but no other abnormal lesions were identified. A grade 1 varix and portal gastropathy were seen on endoscopy, thus suggesting advanced IFALD. In view of the findings of established portal hypertension, a liver biopsy was not performed. The relevant blood results during this period are as follows: platelet count 109 × 109cells/mm3, total bilirubin 256 (unconjugated 36) μmol/L, aspertate aminotransferase 488 U/L, alanine aminotransferase 310 U/L, γ-glutamyl transferase 202 U/L, albumin 33 g/L, and prothrombin time 11 seconds. Routine human immunodeficiency virus, hepatitis B, and hepatitis C screening were negative. Following the assessment, the patient was listed for a combined liver and small bowel transplant. He was discharged home on long-term PN at 10 months of age. Although he was on the small bowel transplant list, intestinal rehabilitation continued with advancing enteral feeds as per our unit's policy. Because of feeding intolerance and faltering growth, he continued to be dependent on long-term PN, as detailed in Table 1.
At the age of 20 months or 16 months corrected, the patient underwent the recommended transplant and had an unexplained intraoperative death. A postmortem examination was performed and the explanted liver was cirrhotic with a biliary pattern consistent with PN-associated liver disease. A nodule was seen in an isolated section, associated with a lack of portal tracts and expanded, thickened reticulin fibres. An increased sinusoidal pattern was also seen with CD34 staining indicative of a well-differentiated hepatocellular carcinoma (HCC) (Fig. 1) (1).
HCC is a rare primary tumour in children seen secondary to chronic liver diseases and following long-term PN in 2 cases. Vileisis et al (2) reported fatal HCC from PN-associated biliary cirrhosis in a 26-month-old child who was born at 27 weeks of gestation. He experienced 2 episodes of necrotising enterocolitis, underwent 3 laparotomies, and was receiving PN until age 16 months, after which the liver dysfunction gradually improved. He subsequently died of a severe gastrointestinal haemorrhage. At autopsy, an 8 × 9 cm polypoid tumour from the undersurface of the right hepatic lobe was seen extending to and causing erosion of the posterior duodenal wall. A separate 5 × 7 cm tumour was embedded in the left hepatic lobe parenchyma. Microscopically, there was widespread biliary cirrhosis, which was probably the factor contributing to HCC.
Patterson et al (3) reported a 6-month-old infant, born at 30 weeks of gestation, who died of respiratory failure. At autopsy, the liver was enlarged, firm, and green with no gross evidence of cirrhosis. Microscopically, there was cholestasis with bridging portal fibrosis. A single focus of neoplasia of a maximum diameter of 1 mm was found to be consistent with HCC. Although this patient had received prolonged PN, there was no evidence of cirrhosis, unlike the 26-month-old child reported by Vileisis et al. Any known possible predisposing factor of HCC was ruled out, thus suggesting PN as a potential carcinogenic agent and raising the possibility of HCC developing earlier in the pathological process, that is, with fibrosis.
HCC was also an unexpected finding in the explanted liver of our patient at autopsy. At the age of 20 months, our patient was PN dependent for most of his life. It is possible that the multiple factors (prematurity, recurrent infections, and sepsis) predisposing him to IFALD also may have contributed to the development of HCC. With negative viral routine screening and a cirrhotic liver with the typical appearance of PN-associated liver disease, it seems likely that PN had a major role in the development of HCC in our patient. The possibility of the lesion being de novo is unlikely considering the rare occurrence of the tumour in children and the associated cirrhosis seen on the explanted liver.
It is known that PN can be hepatotoxic, containing ingredient(s) that are carcinogenic or carcinogenic in premature infants with prolonged administration. Despite better understanding of PN-associated liver disease and recent advances, significant liver damage occurred in our patient. The previously reported patients received intralipid as part of their PN regimen, whereas our patient received clinoleic (20% soybean oil, 80% olive oil) lipid, then SMOF (30% soybean oil, 30% medium-chain triglycerides, 25% olive oil, and 15% fish oil) until he died. Whether the particular lipid preparations used contributed to the severity of liver disease and whether SMOF had a beneficial effect in this case is difficult of which to be certain. Other potential carcinogens such as iron were not part of his PN regimen. Because PN composition continues to be developed with ongoing research and monitoring of administration improves, the hepatotoxic effect can be avoided or minimised, but the potential carcinogenic effect of PN ingredient(s) require further investigation and research.
In addition to PN, concurrent medications with known hepatotoxic effects could contribute to liver damage. Our patient received spironolactone and octreotide and high-dose intravenous omeprazole for 15 months, all of which could be significant.
Chronic inflammation in the liver caused by infections, carcinogenic chemicals, autoimmunity, and metabolic conditions has been recognised as the most important factor in the development of HCC (4,5). Recurrent infections causing chronic inflammation lead to DNA damage from the generation of free radical oxygen and nitrogen species produced by leucocytes and other phagocytic cells as part of the defence mechanism. During the inflammatory process, hypoxia may occur, potentially leading to decreased DNA repair (6). Thus, inflammatory cells appear to promote, influence, and regulate tumour growth.
In conclusion, HCC appears to be a rare complication of the long-term use of PN. Better understanding of the carcinogenic effect of PN would greatly improve the profile of this commonly used treatment. Vigilance with regular ultrasound and awareness regarding the possibility of HCC need to be considered in children receiving long-term PN.
1. Cotran R, Kumar V, Robbins S. Robbins. Pathologic Basis of Disease
, 5th ed. Philadelphia: WB Saunders; 1994.
2. Vileisis RA, Sorensen K, Gonzalex-Crussi F, et al. Liver malignancy after parenteral nutrition. J Pediatr
3. Patterson K, Kapur S, Chandra R. Hepatocellular carcinoma in a noncirrhotic infant after prolonged parenteral nutrition. J Pediatr
4. Umeda T, Hino O. Molecular aspects of human hepatocarcinogenesis mediated by inflammation: from hypercarcinogenic state to normo- or hypocarcinogenic state. Oncology
2002; 62 (suppl 1):38–42.
5. Barash H, Gross E, Edrei Y, et al. Accelerated carcinogenesis following liver regeneration is associated with chronic inflammation-induced double-strand DNA breaks. Proc Natl Acad Sci U S A
6. Coussens LM, Werb Z. Inflammation and cancer. Nature