Nonalcoholic steatohepatitis (NASH) occurs most commonly in adults with obesity, hyperlipidemia, and maturity-onset diabetes (1–5). It has also been described in association with pregnancy, starvation, or chronic treatment with drugs, notably corticosteroids. It may also occur, with elevated serum aminotransferases, in adults in the absence of these risk factors (6–8). Sufficient evidence has accrued to indicate that NASH, as such, occurs in the absence of obvious causative factors such as starvation and that affected patients are not abusing ethanol secretly. Recent observations suggest, however, that the stereotype of the typical NASH patient as middle-aged, massively obese, and female may not be accurate. Moreover, although initially considered a relatively benign disorder, NASH can progress to cirrhosis in adults.
Nonalcoholic steatohepatitis has been reported in children (9–15). Most of these series are small or do not include data from liver biopsies. The purpose of this study was to describe and characterize clinically NASH in children. We report a consecutive prospective series of children with NASH diagnosed at the Hospital for Sick Children during approximately 10 years. This is the one of the largest series of NASH in children reported to date.
PATIENT AND METHODS
All children diagnosed with NASH at the Hospital for Sick Children, Toronto, from December 1985 through April 1995 were included. Patients were excluded if they were receiving concomitant corticosteroid therapy or if they were found to have a metabolic disorder known to cause fatty liver (e.g., Wilson disease, neutral lipid storage disease, or fatty acid oxidation disorders). The first child was diagnosed as having NASH in December 1985 on the basis of clinical obesity and abnormal serum aminotransferase levels, results of laboratory studies excluding other known liver diseases, and results of a percutaneous liver biopsy showing steatosis and inflammation. Thereafter, NASH was included in the differential diagnosis of children with unexplained elevation of aminotransferases, especially if obese. In total, 36 patients were identified. By way of comparison, during this same period, 11 children with symptomatic liver disease were diagnosed as having Wilson disease.
Age, sex, weight, height, initial symptoms, other medical conditions, and medications were noted. Laboratory studies included liver function tests, serology for hepatitis B and C, serum copper and ceruloplasmin, immunoglobulins, nonspecific tissue autoantibodies, random serum glucose, and a fasting lipid profile. A random ethanol level was obtained in one patient.
Thirty-one patients underwent hepatic sonography. A percutaneous liver biopsy was obtained in 24 patients at the time of diagnosis. Follow-up data were available in 21 patients (mean follow-up, 1.5 years; range, 0.2–5.5 years). Statistical analysis was performed with Student's t-test, with P < 0.05 taken as statistically significant.
The median age at diagnosis was 12 years. Thirteen children were less than 11 years old. Most patients had initially reported nonspecific abdominal pain and were referred because of elevated serum aminotransferases or abnormal hepatic sonograms. Two patients had diabetes mellitus at diagnosis (one had been on insulin therapy for many years) and two had later development of insulin-dependent diabetes mellitus. Two brothers had Bardet–Biedl syndrome. One patient had a history of recurrent pancreatitis, one had pustular psoriasis (not treated with methotrexate or corticosteroids), and another had dermatomyositis that had been in remission for several years after anti-inflammatory treatment. Two children had some jaundice. Most children were in good health without other medical illness. No patient had a history of significant drug or medication use, and none used ethanol, according to careful history. None had received a blood transfusion. There was no family history of liver disease; only two patients had a family history of maturity-onset diabetes.
Most patients were obese: 30 of 36 patients had weight in higher than the 97th percentile for age and body weight, more than 120% of ideal weight for height. The mean weight was 147% of ideal body weight (mean weight 71 kg, range 32–138 kg). Sixteen patients had hepatomegaly with the liver edge palpable below the right costal margin, and one of these had hepatosplenomegaly. Thirteen patients had acanthosis nigricans (Fig. 1) observed around the nape of the neck and/or in the axillae; all but one of these were obese. None had typical cutaneous stigmata of chronic liver disease. One had erythema nodosum.
One or both serum aminotransferases were elevated at diagnosis in all but one patient. In 35 patients, the mean value for aspartate aminotransferase (AST) was 104 ± 16 U/l (mean ± SEM; normal range, <37 U/l); AST was normal or near normal in 4 patients (range, 26–523 U/l) and was not recorded in 1 patient. In 33 patients tested, alanine aminotransferase (ALT) was 179 ± 31 U/l (normal range, <40 U/l), and 3 patients had normal or near-normal ALT (range, 10–644 U/l). There was no correlation between the severity of obesity and the degree of serum aminotransferase elevation. Alkaline phosphatase was normal for age in all patients. One patient had mild, persistent unconjugated hyperbilirubinemia, and another had mild conjugated hyperbilirubinemia. Serum albumin and prothrombin time were normal in all patients. Serum copper and ceruloplasmin, measured in 34 of 36 patients, were normal. Nonspecific autoantibodies were tested in 21 patients: these were generally not detected. One patient had positive anti-smooth muscle (anti-actin) and anti-mitochondrial antibodies with normal total immunoglobulins; she was also heterozygous for α1-antitrypsin deficiency with phenotype PI MZ. Two patients, including the patient with dermatomyositis, had positive anti-nuclear antibodies. Hepatitis B serology obtained in 33 patients and anti-hepatitis C virus antibody (since availability of the test) in 21 patients were uniformly negative.
Fasting blood lipid profiles were determined in 20 patients and found to be abnormal in 18. Seven patients had hypercholesterolemia and 11 had increased serum triglycerides. Four patients had elevation of both. Fasting serum cholesterol level was 4.43 ± 0.91 mmol/l (m ± SD; normal, 3.20–4.40 mmol/l), and triglyceride was 2.05 ± 1.06 mmol/l (normal, 0.34–1.58 mmol/l). Except in the two patients with diabetes mellitus, random blood glucose was normal in all patients.
Hepatic sonograms were obtained in 31 of 36 patients. Twenty-four showed abnormalities including hepatomegaly and increased echogenicity suggestive of fatty infiltration.
Percutaneous liver biopsy was obtained in 24 patients. All patients had large-droplet steatosis. Most had inflammation (21, or 88%), and many had fibrosis (17, or 71%). There was inflammatory infiltration and fibrosis of varying severity (Fig. 2); the majority showed some degree of both inflammation and fibrosis. Fibrosis was moderately severe in seven patients, including two patients without inflammatory activity. One patient, in addition to these seven, had cirrhosis at diagnosis. None of the biopsy analyses showed any Mallory hyaline. One patient had capillarization of the sinusoids, observed in electron microscopic examination. A comparison of patients with no fibrosis and patients with severe fibrosis-cirrhosis is shown in Table 1. There were no statistically significant differences between these two groups of patients, although, on average, serum aminotransferases were higher in those with severe fibrosis. The patient with normal aminotransferases, however, had extensive fibrosis suggestive of early cirrhosis.
Follow-up data were available on 21 patients. The mean duration of follow-up was 18 months (range, 2–65 months). Six patients lost weight, and all had improvement in serum aminotransferase levels. In two of these patients AST and ALT completely normalized. Most patients had great difficulty losing weight. The AST and ALT levels fluctuated in those patients who were not able to achieve weight loss. One patient subsequently had severe hypothyroidism.
At the time this study began, whether chronic liver disease resembling alcoholic liver disease but occurring in nonalcoholics actually existed was controversial. Identifying this disease in children provided evidence that NASH as such actually existed. Finding a large number of such patients now suggests that it constitutes a significant problem in pediatric hepatology. The diagnosis of NASH rests in part on histologic findings, and we have accepted steatosis with evidence of necroinflammatory activity as the essential histologic criteria for this diagnosis. In many patients, fibrosis is also present. We believe that the term “nonalcoholic steatohepatitis” is unduly restrictive and prefer the term “nonalcoholic fatty liver disease.” However, the former terminology appears entrenched.
Despite many case series describing NASH in adults, there have been only a few reports of this disorder in children. Moran et al. (9) described three children with obesity and steatohepatitis. In a study of 299 obese children, Kinugasa et al. (10) found 36 (13%) to have abnormal aminotransferases. Liver biopsy was performed in 11 of these children, and results confirmed steatohepatitis. One patient had cirrhosis, along with maturity-onset diabetes mellitus and hyperlipidemia. Vajro et al. (11) reported a series of seven obese children with persistently elevated aminotransferases. These children may have had NASH; however, liver biopsy was obtained in only one patient, and the results showed steatohepatitis. Recently Baldridge et al. (12) reported a series of 14 children with idiopathic hepatic steatosis, identified by retrospective review of results of all liver biopsies performed in a tertiary-care pediatric hospital. All were obese, and most had abnormal AST and ALT. Two patients with test results showing normal liver function were discovered incidentally at laparotomy. In a screening study of 310 obese Japanese children Tazawa et al. (13) found that 24% had elevated serum ALT and 83% had a fatty fibrotic pattern observed on hepatic sonography. This same sonographic appearance was found in 19% of children studied who had normal ALT. Although elevated ALT and fatty liver by ultrasound testing appeared to be somewhat more common in older children with more severe obesity, no statistically significant differences were found in different age groups or with longer duration of obesity. An Italian screening study of 195 obese children found fatty liver by sonography in 55%, elevated serum AST or ALT in 20%, and both features in 15%(14).
In our prospective series of 36 children with NASH, findings were heterogeneous. Obesity was the most common clinical denominator. The average patient had a body weight approximately 50% higher than the ideal for height. Although the mean weight overall was 71 kg, of particular concern was the very obese younger child, specifically, 13 children less than 11 years of age whose average weight was 56 kg. However, six patients were not obese (childhood obesity defined as weight >120% of ideal weight for height). Typically, these children were tall with large bones and proportionately heavy body weight. Unlike some adult series, boys were more common than girls in this series of children. Male predominance was also noted in a Japanese survey of fatty liver detected by sonography in a cohort of 810 children between the ages of 4 and 12 years (16). Approximately half of all patients in the present study had hepatomegaly, and most had elevated serum aminotransferases. Other liver function test results were normal. Eighteen of 20 patients studied had hyperlipidemia; hypertriglyceridemia was more common and more severe.
Long-term follow-up in adults with NASH has shown that the disease is usually slowly progressive but can ultimately lead to cirrhosis in some patients (17,18). Other studies in adults have documented cirrhosis complicating NASH (3–5,8). In this series one patient, a 10 year-old girl with hepatosplenomegaly, had cirrhosis at diagnosis, the second case of cirrhosis in childhood NASH thus far reported. We speculate that some patients with “cryptogenic” cirrhosis occurring in adulthood have in fact had NASH since childhood.
One third of the children in this series had acanthosis nigricans. This association has not been described in other reports of children with NASH. Acanthosis nigricans may be subtle and can be missed without careful examination. Although acanthosis nigricans may occur in simple childhood obesity, it has also been shown to be a cutaneous marker of hyperinsulinemia (19,20). Keratinocytes have receptors for insulin, epidermal growth factor, and insulin-like growth factors. In hyperinsulinemia, circulating insulin, because of its structural similarity to insulin-like growth factor, binds to these receptors and stimulates cell division, leading to acanthosis. Richards et al. (21) reported 22 children with obesity, insulin resistance, acanthosis nigricans, and hyperandrogenemia. Obesity was always the first component of the syndrome to appear. Unfortunately, liver function tests were not obtained in this study.
Other features of the patients in this series suggest that abnormalities of carbohydrate metabolism may be important in the pathogenesis of NASH in children. Two of the patients in our series were known to have diabetes, and two other patients had development of diabetes later. A further patient had insulin-dependent diabetes that developed beyond the period of follow-up for this study at a time when his obesity had resolved and serum aminotransferases were normal. Two patients had Bardet–Biedl syndrome, an inherited syndrome characterized by retinal dystrophy, polydactyly, obesity, renal abnormalities, and male hypogenitalism. Non–insulin-dependent diabetes mellitus (NIDDM) frequently develops, with insulin resistance due to abnormal insulin receptor function (22). Alström syndrome is a rare condition, phenotypically similar to Bardet–Biedl, but distinguished by sensorineural deafness and absence of mental retardation (23). Obesity, hyperlipidemia, NIDDM, and acanthosis nigricans commonly occur in Alström syndrome. Hepatic involvement with mild steatosis, portal inflammation, and moderate fibrosis has been described in one patient (24). Yet another report of a rare syndrome in this Bardet–Biedl/Alström spectrum has been published (25), in which one patient had hepatic steatosis and both patients had decreased insulin receptor binding.
We speculate that abnormal carbohydrate metabolism plays an important role in the pathogenesis of childhood NASH. The high prevalence of acanthosis nigricans suggests that abnormalities in insulin metabolism may be important. Wanless et al. (26) reported hepatic steatosis in a unique subcapsular distribution in patients receiving continuous ambulatory dialysis who took insulin intraperitoneally. In hyperinsulinemia, free fatty acids are preferentially esterified into triglycerides. The resultant hepatic steatosis may incite an inflammatory response, leading to fibrosis. Hypertriglyceridemia was prominent in many of the children in this series. In a few patients tested, hyperinsulinemia was detected. Hyperinsulinemia is characteristic of puberty (27), but recent studies with standardized measurement of insulin levels indicate that obese preadolescent children already have hyperinsulinemia with insulin resistance affecting glucose and lipid metabolism (28). Racial differences, especially relating to decreased hepatic clearance of insulin, may be relevant (29). In a cohort of 228 obese Japanese children in a single prefecture, hyperinsulinemia was found in 32% and by regression analysis appeared to be the major contributor to occurrence of elevated serum ALT, used as a marker for NASH (30). A recent study in obese adolescents documented hyperinsulinemia as well as perturbations to insulin-like growth factor I and its binding proteins (31).
Although chronic hyperinsulinemia coupled with hepatocellular insulin resistance may be important for development of NASH, genetic differences unrelated to carbohydrate and lipid metabolism may account for the heterogeneity of clinical findings in NASH. These include heterozygosity for the HFE gene (32), α1-antitrypsin phenotype (33), adequacy of anti-oxidant defenses such as hepatic glutathione levels, and hepatic cytochrome P450. Hepatocellular concentrations of CYP2E1, an ethanol-inducible cytochrome P450 that activates some xenobiotics to toxic intermediates, were found to be increased in adults with NASH (34). Genes regulating body weight, such as the ob gene (35), may also play a role in the clinical diversity of NASH.
The long-term prognosis of children with NASH remains unknown. In adults, weight reduction can lead to improvement in serum aminotransferase levels (36,37). All six patients in our series who lost weight had a decrease in serum aminotransferases. Although the number was small, the trend was toward improvement. In the series of obese children reported by Vajro et al. (11) persistently elevated aminotransferases normalized in all seven patients after weight reduction, and liver size decreased in those with hepatomegaly. In the only patient in that series with steatohepatitis detected in liver biopsy at diagnosis, inflammation resolved after weight reduction. A more recent report of 38 obese children with fatty liver diagnosed by sonography documents that the liver sonogram improved or became normal in 26 (79%) of the 33 children who lost weight when assessed at 3 months of age. All children who had had elevated aminotransferases had normal aminotransferases at that time (15). Whether improvement in aminotransferase levels always parallels improvement in liver histology is not known, because serial liver biopsies are rarely performed. However, until results of long-term follow-up studies are available in children, it is prudent to recommend a weight loss program for these obese children. We use a combination of energy intake restriction, mainly by reducing fat intake, and increased energy expenditure through aerobic exercise. Other possible treatments include ursodeoxycholic acid (38) or antioxidants such as vitamin E (39), but such treatments have not been evaluated rigorously.
We conclude that NASH is not limited to adults. Indeed, it may be fairly common in children. It should certainly be suspected in obese children with mildly elevated serum aminotransferases. Moreover, it can occur in nonobese children. Sonography may suggest the diagnosis of massive fatty infiltration in the liver, but a liver biopsy is needed to assess the degree of chronic damage. Steatohepatitis can be found in other diseases affecting the liver. Wilson disease must be excluded. Chronic viral hepatitis and drug-induced hepatitis must also be considered in the differential diagnosis. Not necessarily a benign disorder, NASH can progress to cirrhosis during childhood and may account for some cases of cirrhosis in adults in whom the cause is obscure. Although there may be diverse causes for NASH in children, we believe that many children have specific and potentially definable disorders of hepatocellular metabolism.
Dr. Rashid was supported by a Duncan Gordon Fellowship from the Hospital for Sick Children Foundation and a grant from Janssen Pharmaceutica. This report is dedicated to the memory of James L. Weber, MD, FRCPC, distinguished paediatrician and liver specialist, in appreciation of his mentorship.
1. Ludwig J, Viggiano TR, McGill DB, Ott BJ. Nonalcoholic steatohepatitis
: Mayo Clinic experience with a hitherto unnamed disease. Mayo Clin Proc 1980; 55:434–8.
2. Adler M, Schaffner F. Fatty liver
hepatitis and cirrhosis in obese patients. Am J Med 1979; 67:811–6.
3. Diehl AM, Goodman Z, Ishak KG. Alcohol-like disease in nonalcoholics. Gastroenterology 1988; 95:1056–62.
4. Lee RG. Nonalcoholic steatohepatitis
: A study of 49 patients. Hum Pathol 1989; 20:594–8.
5. Wanless IR, Lentz JS. Fatty liver
) and obesity
: An autopsy study with analysis of risk factors. Hepatology 1990; 12:1106–10.
6. Schaffner F, Thaler H. Nonalcoholic fatty liver
disease. Prog Liver Dis 1986; 8:283–6.
7. Alvarez SZ, Billote J, Ongkingco–Obillos C. Chronic hepatitis with fatty change. J Gastroenterol Hepatol 1990; 5:223–7.
8. Bacon BR, Farahwash MJ, Lanney CG, Neuschwander–Tetri BA. Nonalcoholic steatohepatitis
: An expanded clinical entity. Gastroenterology 1994; 107:1103–9.
9. Moran JR, Ghishan FK, Halter SA, Greene HL. Steatohepatitis
in obese children: A cause of chronic liver dysfunction. Am J Gastroenterol 1983; 78:374–7.
10. Kinugasa A, Tsunamoto K, Furukawa N, Sawada T, Kusunoki T, Shimada N. Fatty liver
and its fibrous changes found in simple obesity
of children. J Pediatr Gastroenterol Nutr 1984; 3:408–14.
11. Vajro P, Fontanella A, Perna C, Orso G, Tedesco M, De Vincenzo A. Persistent hypertransaminasemia resolving after weight reduction in obese children. J Pediatr 1994; 125:239–41.
12. Baldridge AD, Perez–Atayde AR, Grame–Cooke F, Higgins L, Lavine JE. Idiopathic steatohepatitis
in childhood: A multicenter retrospective study. J Pediatr 1995; 127:700–4.
13. Tazawa Y, Noguchi H, Nishiomiya F, Takada G. Serum alanine aminotransferase activity in obese children. Acta Paediatr 1997; 86:238–41.
14. Bergomi A, Lughetti L, Corciulo N, et al. Italian multicenter study on liver damage in pediatric obesity
(abstract). Int J Obes Relat Metab Disord 1998; 22(suppl 4):S22.
15. Franzese A, Vajro P, Argenziano A, et al. Liver involvement in obese children: Ultrasonography and liver enzyme levels at diagnosis and during follow-up in an Italian population. Dig Dis Sci 1997; 42:1428–32.
16. Tominaga K, Kurata JH, Chen YK, et al. Prevalence of fatty liver
in Japanese children and relationship to obesity
: An epidemiological ultrasonographic survey. Dig Dis Sci 1995; 40:2002–9.
17. Powell EE, Cooksley WGE, Hanson R, Searle J, Halliday JW, Powell LW. The natural history of nonalcoholic steatohepatitis
: A follow-up study of forty-two patients for up to 21 years. Hepatology 1990; 11:74–80.
18. Propst A, Propst T, Judmaier G, Vogel W. Prognosis in nonalcoholic steatohepatitis
. Gastroenterology 1995; 108:1607.
19. Brockow K, Steinkraus V, Rinninger F, Abeck D, Ring J. Acanthosis nigricans
: A marker for hyperinsulinemia. Pediatr Dermatol 1995; 12:323–6.
20. Stuart CA, Gilkison CR, Smith MM, Bosma AM, Keenan BS, Nagamani M. Acanthosis nigricans
as a risk factor for non-insulin dependent diabetes mellitus. Clin Pediatr (Phila) 1998; 37:73–9.
21. Richards GE, Cavallo A, Meyer WJ, et al. Obesity
, acanthosis nigricans
, insulin resistance and hyperandrogenemia: Pediatric perspective and natural history. J Pediatr 1985; 107:893–7.
22. Green JS, Parfrey PS, Harnett JD, et al. The cardinal manifestations of Bardet–Biedl syndrome, a form of Laurence–Moon–Biedl syndrome. N Engl J Med 1989; 321:1002–9.
23. Dyer DS, Wilson ME, Small KW, Pai GS. Alström syndrome: A case misdiagnosed as Bardet–Biedl syndrome. J Pediatr Ophthalmol Strabismus 1994; 31:272–4.
24. Connolly MB, Jan JE, Couch RM, Wong LTK, Dimmick JE, Rigg JM. Hepatic dysfunction in Alström disease. Am J Med Genet 1991; 40:421–4.
25. Boor R, Herwig J, Schrezenmeir J, Pontz BF, Schönberger W. Familial insulin resistant diabetes associated with acanthosis nigricans
, polycystic ovaries, hypogonadism, pigmentary retinopathy, labyrinthine deafness, and mental retardation. Am J Med Genet 1993; 45:649–53.
26. Wanless IR, Bargman JM, Oreopoulos DG, Vas SI. Subcapsular steatonecrosis in response to peritoneal insulin delivery: A clue to the pathogenesis of steatonecrosis in obesity
. Mod Pathol 1989; 2:69–74.
27. Amiel SA, Caprio S, Sherwin RS, Plewe G, Haymond MW, Tamborlane WV. Insulin resistance of puberty: A defect restricted to peripheral glucose metabolism. J Clin Endocrinol Metab 1991; 72:277–82.
28. Caprio S, Bronson M, Sherwin RS, Rife F, Tamborlane WV. Co-existence of severe insulin resistance and hyperinsulinaemia in pre-adolescent obese children. Diabetologia 1996; 39:1489–97.
29. Jiang X, Srinivasan SR, Radhakrishnamurthy B, Dalferes ER, Jr, Berenson GS. Racial (black-white) differences in insulin secretion and clearance in adolescents: The Bogalusa Heart Study. Pediatrics 1996; 97:357–60.
30. Kawasaki T, Hashimoto N, Kikuchi T, Takahashi H, Uchiyama M. The relationship between fatty liver
and hyperinsulinemia in obese Japanese children. J Pediatr Gastroenterol Nutr 1997; 31:317–21.
31. Attia N, Tamborlane WV, Heptulla R, et al. The metabolic syndrome and insulin-like growth factor I regulation in adolescent obesity
. J Clin Endocrinol Metab 1998; 83:1467–71.
32. George DK, Goldwurm S, MacDonald GA, et al. Increased hepatic iron concentration in nonalcoholic steatohepatitis
is associated with increased fibrosis
. Gastroenterology 1998; 114:311–8.
33. Czaja AJ. Frequency and significance of phenotypes for alpha1-antitrypsin deficiency in type 1 autoimmune hepatitis. Dig Dis Sci 1998; 43:1725–31.
34. Weltman MD, Farrell GC, Hall P, Ingelman–Sundberg M, Liddle C. Hepatic cytochrome P450 2E1 is increased in patients with nonalcoholic steatohepatitis
. Hepatology 1998; 27:128–33.
35. Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM. Positional cloning of the mouse obese gene and its human homologue. Nature 1994; 372:425–32.
36. Palmer M, Schaffner F. Effect of weight reduction on hepatic abnormalities in overweight patients. Gastroenterology 1990; 99:1408–13.
37. Teli MR, James OF, Burt AD, Bennett MK, Day CP. The natural history of nonalcoholic fatty liver
: A follow-up study. Hepatology 1995; 22:1714–9.
38. Laurin J, Lindor KD, Crippin JS, et al. Ursodeoxycholic acid or clofibrate in the treatment of non-alcohol-induced steatohepatitis
: A pilot study. Hepatology 1996; 23:1464–7.
39. Lavine JE. Treatment of obesity
with vitamin E (abstract). Gastroenterology 1998; 114:A1284–5.