Controversies in the management of pediatric liver disease: Hepatitis B, C and NAFLD: Summary of a single topic conference


  • Benjamin L. Shneider,

    Corresponding author
    1. Division of Pediatric Hepatology, Department of Pediatrics, Mount Sinai School of Medicine, New York, NY
    • Mount Sinai Medical Center, One Gustave L. Levy Place, Box 1656, New York, NY 10029
    Search for more papers by this author
    • fax: 212-241-8991

    • All authors contributed equally to this work.

  • Regino González-Peralta,

    1. Pediatric Hepatology and Liver Transplantation, Division of Pediatric Gastroenterology, Department of Pediatrics, University of Florida, Gainesville, FL
    Search for more papers by this author
    • All authors contributed equally to this work.

  • Eve A. Roberts

    1. Division of Gastroenterology, Hepatology and Nutrition, The Hospital for Sick Children and Departments of Paediatrics, Medicine and Pharmacology, University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
    • All authors contributed equally to this work.

  • Potential conflict of interest: Dr. Shneider was on the Speakers' Bureau for Axcan Scandipharm. He received grants from Schering-Plough and Gilead.

The AASLD Hepatitis Single Topic Conference (STC) entitled “Controversies in the Management of Pediatric Liver Disease: Hepatitis B,C and NAFLD,” was held at the Emory Conference Center on March 3 and 4, 2006. The conference was organized to juxtapose the current state-of-the-art in the management of pediatric liver disease with evidenced-based approaches that exist for the same problems in adults. The principal goal of the conference was to assess critically current management of hepatitis B and C, and NAFLD in children, since evidence-based approaches are generally lacking. The conference included formal didactic presentations by experts (Table 1A-C) and informal active discussions between faculty and conference participants. The informal discussions were very animated and educational including both question and answer sessions and case discussions. Herein, we synthesize some of the information presented during the proceedings. Our aim with this summary is to highlight salient issues that were discussed in the STC rather than present a comprehensive review of the topics or formulate consensus statements. Individual sections have been reviewed by the faculty for that particular section, although they are not co-authors per se. The success of the conference was due in large part to the active exchange of information that took place between faculty and the participants.

Table 1A. Faculty Participants for Hepatitis B (Moderator Benjamin L. Shneider)
Faculty MemberInstitutionTopic
Mei-Hwei ChangNational Taiwan UniversityNatural History
Giorgina Mieli-VerganiKing's College HospitalCurrent Management
Anna Suk-Fong LokUniversity of MichiganApproaches in Adults
Maureen JonasChildren's Hospital, BostonRationale for Treatment
Ronald SokolUniversity of ColoradoRationale against Treatment
Eugene SchiffUniversity of MiamiFuture Directions
Table 1B. Faculty Participants for Hepatitis C (Moderator Regino González-Peralta)
Faculty MemberInstitutionTopic
Latifa T.F. YeungUniversity of TaiwanNatural History
Karen F. MurrayUniversity of WashingtonCurrent Management
Michael W. FriedUniversity of North CarolinaApproaches in Adults
Kathleen B. SchwarzJohns Hopkins UniversityRationale for Treatment
William F. BalistreriChildren's, CincinnatiRationale against Treatment
John McHutchisonDuke UniversityFuture Directions
Table 1C. Faculty Participants for NAFLD (Moderator: Eve A. Roberts)
Faculty MemberInstitutionTopic
Christopher P. DayUniversity of NewcastlePathophysiology and Adults
David E. KleinerNational Cancer InstitutePathology
Jeffrey B. SchwimmerUniv. of Calif., San DiegoNatural History
Philip RosenthalUniv. of Calif., San FranciscoCurrent Management
Paul B. PencharzUniversity of TorontoRationale for Treatment
Keith D. LindorMayo ClinicFuture Directions

Hepatitis B

The complex biology and natural history of hepatitis B virus (HBV) coupled with recent advances in therapeutic approaches makes clinical decision-making in this common infection very problematic. A relatively limited published dataset related to HBV in children has left the clinician without clear-cut clinical guidelines for nearly all aspects of the care of children with chronic HBV infection. The presentations (Table 1A), case discussions and question and answer sessions at the STC highlighted marked differences in approaches to HBV infection in children by clinicians who have longstanding expertise in the management of childhood liver disease. In addition, it was also notable that the application of adult HBV guidelines to children is problematic and potentially not appropriate, since children are typically in a different stage of HBV infection and are at lower near-term risk of developing significant morbidity and mortality.1 Currently, there are National Institutes of Health (NIH)-funded multi-center studies of hepatitis C virus (HCV) and nonalcoholic fatty liver disease (NAFLD) in children and adults but there is a notable absence of similar investigations of HBV infection in children or adults.

A careful and critical understanding of the natural history of HBV infection in children is essential in making informed decisions regarding monitoring and treatment. One needs to be cognizant of the effects of HBV genotype and potential environmental effects in different regions of the world when analyzing the natural history of HBV. Thus, longitudinal data from Southeast Asia may not necessarily be applicable to Western Europe or North America. It is useful to subdivide chronic HBV infection in children into perinatal versus other modes of transmission. Perinatal transmission of HBV is typically characterized by an immunotolerant state that is presumed to be fostered by prenatal transplacental fetal exposure to the HBeAg. Children who have perinatally acquired HBV infection usually have normal or near normal serum aminotransferases, very high serum HBV DNA levels and persistent HBsAg and HBeAg positivity (Fig. 1, Box A). They typically maintain this pattern of laboratory test results for most of childhood and often into adulthood. Spontaneous HBeAg seroconversion may occur during the immunotolerant HBV phase, although the rates during early childhood are low.2 Seroconversion rates are generally less than 2% before the age of 3 years and then HBeAg seroconversion occurs at a rate of approximately 4% to 5% per year. HBeAg seroconversion rates are higher for individuals with elevated aminotransferase levels. In the United States, immunotolerant perinatally acquired HBV infection represents approximately 50% of children with chronic HBV infection. Childhood acquired HBV infection is usually characterized by somewhat lower circulating levels of HBV, more biochemical evidence of hepatitis and higher rates of spontaneous eAg seroconversion. Seroconversion rates are higher for genotype B relative to C.3 It is not widely appreciated that spontaneous HBeAg seroconversion may be associated with significant elevation of aminotransferases, generally ascribed to immune mediated cytolysis. This stage may persist for months to years during which substantial liver damage can be accrued before HBeAg seroconversion (Fig. 1, Box B).4 Thus, it can be very difficult to assess the stage of disease in a child with chronic HBV infection from a single set of laboratory values. In some children HBV precore mutants that prevent HBeAg production are selected. These children may have persistent elevation in aminotransferases and relatively high levels of HBV DNA in serum after HBeAg seroconversion.5 Most children who undergo HBeAg seroconversion will be in the inactive carrier state for years to decades.

Figure 1.

Natural History of HBV Infection and Potential Interventions. (Modified from DS Chen. From hepatitis to hepatoma : Lessons from type B viral hepatitis. Science 1993;262:369-370. Part or all of this figure originally modified from D-S Chen, J. Gastroenterol. Hepatol. 8 470 (1993), Blackwell Scientific Publications. Reprinted with permission.

With the implementation of universal newborn immunization against HBV in 1984 in Taiwan, the HBsAg seroprevalence rate was dramatically reduced.6 The incidence of hepatocellular carcinoma (HCC) in children aged 6 to 14 dropped more than 50% from 263 cases/48,764,799 (0.54 cases/100,000) to 35 cases/17,817,510 (0.20 cases/100,000). It is worth noting that even in the pre-vaccination era the incidence of HCC in childhood was less than 0.01% (assuming a 10% prevalence of HBVsAg). Annual mortality from HBV-related fulminant hepatitis decreased 3-fold after vaccination was implemented.7 Similar results have been described as a result of routine immunization efforts in Alaska.8 Routine immunization initially resulted in an increase in the percentage of infected children with HBV surface gene “a” determinant mutants, although the total number of children with HBV surface mutants has dropped in recent years due to the marked reduction in the number of HBV positive children.9 This should not be construed as an argument against immunization, which has led to remarkable improvements in HBV related morbidity and mortality. It is clear that routine HBV vaccination programs result in decreased disease burden and should be considered where financially feasible (Fig. 1, Box C).

There are only a limited number of long-term studies of HBV starting in childhood. One of the most common questions that arises relates to the prognosis for children with chronic HBV. The answer to that question is often derived based upon cross-sectional data from adult cohorts followed in tertiary care centers. A 25% lifetime risk of developing HCC is often given, although this risk estimate does not emanate from longitudinal analysis of HBV infection first diagnosed in childhood. In addition, these estimates are based upon cohorts from Southeast Asia and it is not certain whether they can be extrapolated to other patient populations. Since current screening and treatment approaches are based upon these estimates, accurate assessment of risk is of critical importance. In Taiwan, Chang prospectively analyzed the course of 415 children with chronic HBV for an average of nearly 8 years. In this cohort only two children developed cirrhosis, one of whom developed HCC.2 Long-term follow up of 174 children in Montreal revealed one child with cirrhosis and no cases of HCC.10 Tenyear follow up of 73 children in the United Kingdom revealed no cases of cirrhosis or HCC.11 Eighteen-year follow up of a cohort of children in Italy identified 3 children who had HCC, although it was identified at presentation in all three cases.12 Unfortunately, at present there are inadequate numbers of patients with sufficiently long follow-up to give an accurate assessment of the lifelong risk and/or predictors of serious morbidity and mortality in children with chronic HBV.

Hepatocellular carcinoma has been reported in children with chronic HBV.8, 13, 14 The case series typically are collected from a cross-sectional or retrospective analysis in a single center and are not typically the result of prospective analyses. There are biases toward the publication of series with larger numbers of cases. Although anecdotal cases of HCC in the first years of life exist.15, 16 when liver cancer occurs in childhood, it occurs much more commonly in the second decade of life. Since denominators are not available in most of the series published, it is nearly impossible to assess an accurate risk of development of HCC before the age of 20 years in children with HBV infection acquired early in life. Informal discussions in the STC revealed that in light of the reported cases of HCC in children, significant and costly efforts are often undertaken in both screening and treatment. Many clinicians obtain serum alpha fetoprotein levels biannually and surveillance abdominal ultrasonography on an annual basis. Interestingly, typical screening efforts in adults are not recommended until the age of 40 or 50 except for those with cirrhosis or a family history of HCC (Fig. 1, Box D).17 Thus children who transition to adult practices may have their surveillance discontinued for 20 or 30 years. It must be acknowledged that much of this screening in children stems from medicolegal concerns. There are no data to suggest that screening for HCC in children with HBV is cost-effective, nor is there evidence that it alters the natural history of the disease in childhood. Thus, it can be concluded that there is no standard of care that requires screening for HCC in children with chronic HBV, although one might consider it in children with advanced fibrosis or cirrhosis of a family history of HCC.

The STC highlighted significant controversies revolving around the issue of the indications for and methods of treatment of chronic HBV in children. There was general agreement that the ultimate goal of therapy is to reduce the morbidity and mortality related to the development of cirrhosis and HCC. Prevention of spread of HBV and social concerns have become less pronounced reasons for treatment due to effective active and passive immunization strategies. Although there is some data in selected adults that treatment may reduce the risk of development of HCC, no data in children show statistically and clinically relevant reductions in morbidity and mortality as a result of current treatments applied during childhood. This issue is further complicated by spontaneous HBeAg seroconversion that often occurs during childhood.4 Current trends of using nucleoside analogs in treating adults to reduce HBV viral loads are of limited relevance to the child with immunotolerant HBV infection.18 A major concern is long-term safety and efficacy of nucleoside analogs. Approved nucleoside analogs except for lamivudine have limited safety record even among adults, and resistance has been reported in association with all nucleoside analogs. There was active debate regarding the indications for treatment in children. If there was a general consensus it was that treatment might be considered for children with biochemically and histologically active disease who may be at higher risk to develop cirrhosis during childhood. Identifying these patients may be difficult since these findings are common in children undergoing spontaneous HBeAg seroconversion. Active disease in this setting for more than six months may be an indication to consider treatment both in HBV infected children and adults. Interferon (IFN) and lamivudine are both approved as single agents by the Food and Drug Administration for the treatment of HBV in children, based in part upon randomized clinical trials.19, 20 When therapy is indicated, the preferred choice of therapeutic agent is not clear. Durable response rates seem to be higher with IFN, although its use is associated with significant toxicity. By contrast, lamivudine is significantly safer than IFN but prolonged use of lamivudine is often associated with the emergence of resistant viral strains, which may limit its future use as part of combination therapy. This issue may be particularly problematic given the long expected lifespan for children.

Rapid and on-going advances in understanding the biology and treatment of HBV makes the future very bright for children with this infection. New approaches to the management of HBV in children are clearly indicated. Ultimately it would be most useful to interrupt this chronic infection at a time before there is irreversible progression to an increased risk of HCC (Fig. 1, Box E). Unfortunately, the timing of this stage is not well understood. A novel approach to the treatment of immunotolerant HBV was presented by Dr. Giorgina Mieli-Vergani.21 That approach, which was evaluated in only a limited number of children, utilized sequential and combination therapy with lamivudine and IFN. Significant numbers of patients underwent both HBeAg and HBsAg seroconversion, which is very different than previous experience with combination therapy in adults. If confirmed in subsequent studies, this could represent a tremendous advance in the approach to the management of HBV in children. Ultimately, if these findings can be replicated and/or improved in prospective multi-center studies, this approach may yield the greatest long-term reduction in the risk of development of HCC. As new therapeutic agents become available it will be critically important to reassess the end-points utilized in determining the effectiveness of specific therapies in children. Identification of predictors of future risk of cirrhosis and/or HCC is another priority in pediatrics. It is hoped that a future conference might focus on this important issue. Ideally this discussion should include active input from the Food and Drug Administration, the NIH, the pharmaceutical industry, the lay public and clinicians actively participating in research related to and care of children with HBV.


STC, single topic conference; HBV, hepatitis B virus; HCV, hepatitis C virus; NAFLD, nonalcoholic fatty liver disease; NIH, National Institutes of Health; Ag, antigen; Ab, antibody; HCC, hepatocellular carcinoma; PCR, polymerase chain reaction; RNA, ribonucleic acid; SVR, sustained virologic response; NASH, nonalcoholic steatohepatitis; ALT, alanine aminotransferase; AST, aspartate aminotransferase; IFN, interferon; Peg, pegylated; AT, aminotransferase.

Hepatitis C

Hepatitis C virus infection is a major global health problem. In the United States, there are approximately 7 million adults and 70-100,000 children infected with HCV.22–24 The financial burden of this viral infection is staggering with projected medical costs of $10.7 billion in adults in the years 2010-2019 and $199—426 million over the next 10 years in children.23, 25 While the incidence of HCV infection has significantly decreased since the implementation of reliable blood-bank screening strategies, maternal-infant transmission remains an important route of transmission.26

By contrast to HBV, there is no reliable strategy to prevent maternal-infant HCV transmission and screening all pregnant women for HCV infection does not appear to be cost-effective.27 Although young infants may not be suitable candidates to receive currently available antiviral therapies, they are at some risk of developing chronic liver disease. Therefore, early detection of HCV infection in young infants may be useful in selecting a unique group who would benefit from judicious medical monitoring. To screen infants born to mothers with HCV infection one reasonable approach is to test for the presence of anti-HCV in serum at 12-18 months of age, when maternally derived antibodies have disappeared. Advocates of this approach highlight its economical advantage as this strategy relies on relatively inexpensive serological assays and point out that little is gained by confirming HCV infection at an earlier age. However, as noted during the STC, PCR testing may be useful in selected cases where the potential of having passed an infection results in significant maternal anxiety, despite the low (≈6%) risk of transmission.26 In this setting excluding HCV infection earlier with PCR testing is reassuring and may be worth the added expense. Of note, infants with detectable HCV RNA in infancy should be monitored as spontaneous viral clearance may occur during childhood. A recent large study suggested that spontaneous viral clearance was more common with HCV genotype 3 infections than other genotypes and rarely occurred after 3 years of age.28

Formulating reliable disease-specific management guidelines depends on the availability of accurate information about its natural history. Throughout the STC there was a clear theme that additional data is needed regarding the long-term outcomes of persons infected with HCV during childhood and prognostic factors in this population. This paucity of pediatric-specific data may lead to potentially erroneous application of adult data to manage children infected with HCV. However, available information (based primarily on retrospective analyses) suggests that HCV infection is generally a benign condition in children but significant liver disease including chronic hepatitis, cirrhosis and HCC can occur. As in adults, HCV-infected children with abnormal liver biochemistry are more likely to have worse histological disease than those with normal liver tests but advanced liver disease may be seen even in those without biochemical aberrations. Liver disease does not appear to correlate with specific HCV genotypes or circulating levels of HCV RNA in children.28, 29

There are no specific guidelines to care for children with chronic HCV infection. Conference participants agreed that these young individuals should be monitored periodically for disease activity. The clinical evaluation should include a careful assessment for liver disease, including physical examination, liver biochemistry and determination of viral load. Other potential causes of liver disease should be excluded in children with abnormal liver tests. Another contentious issue which elicited animated debate was related to the need to report the diagnosis of chronic HCV infection to daycare, school, athletic programs or other similar organizations. Although the legality to report infectious diseases may vary between geographic locations, some argued that the social stigma associated with an infectious disease outweighed the potential benefits of disclosure, particularly in the current era of universal precautions. Screening for liver cancer in children with chronic HCV infection is particularly problematic because there are only four reported cases of HCC arising in adolescents and young adults with childhood-acquired infection and all had cirrhosis.30, 31 It appears then, that HCC is a rare complication in children with HCV infection but those with significant fibrosis or cirrhosis should undergo periodic surveillance. Because pediatric-specific data is unavailable, monitoring HCV-infected children with significant fibrosis or cirrhosis with serum concentrations of alpha-fetoprotein and abdominal sonograms as in adults seem reasonable.17

The need for and timing of a liver biopsy is another issue that led to considerable debate during the STC. Because significant histological abnormalities may be seen even in asymptomatic children with normal liver tests, some argued that a liver biopsy should be performed as part of the initial evaluation since this information was critically important to adequately assess the degree of hepatic injury. By contrast, others questioned this approach highlighting the potential risks associated with this procedure and the frequent mild histological abnormalities noted in most infected children. Non-invasive methods to assess liver fibrosis including serological markers and transient elastography appear promising in adults32 but they have not been evaluated in children. In general, the STC participants agreed that a liver biopsy, as for any other test, should be considered only if the results will influence medical decision-making. A liver biopsy may be specifically useful in children who are being considered for antiviral treatment. Even in that scenario, the need for histological assessment is controversial, particularly for those infected with HCV genotypes 2 or 3 who have a high (>70%) probability of achieving a sustained virologic response (SVR) with currently available treatments (Table 2).

Table 2. Treatment Trials in Children With Chronic HCV Infection
StudyNo. studied1Treatment RegimenSustained Response2
All typesHCV type 1HCV type 2/3
  • 1

    Number of HCV infected patients treated in the studies cited.

  • 2

    Undetectable serum HCV RNA 6 months after stopping treatment.

  • 3

    Numbers in parenthesis represent percent of treated children.

  • 4

    Only included patients with genotype 1

  • Adapted from RP González-Peralta. Treatment of chronic hepatitis C in children. Pediatr Transplantation 2004; 8:639-43.

Christensson et al.3711IFN-Ribavirin7/11 (64)32/5 (40)5/5 (100)
Lackner et al.3812IFN-Ribavirin6/12 (50)6/12 (50)N/A4
Wirth et al.3941IFN-Ribavirin25 (61)18/34 (53)7/7 (100)
González-Peralta et al.40118IFN-Ribavirin70 (46)33/82 (36)21/25 (84)
Schwarz et al.4314PEGIFN5/13 (38)5/13 (38)N/A
Wirth et al.4459PEGIFN-Ribavirin36/59 (59)28/46 (50)13/13 (100)

Although clearly important, very little is known about the pathogenesis of HCV infection in children. Extensive experimental evidence in adults suggests that immunological mechanisms play a prominent role in mediating HCV-related liver cell damage.33 Unraveling viral–host interactions have enhanced our understanding of the disease process and may lead to new therapies and vaccine strategies. Because there are distinct differences in both innate and adaptive immunological function between children and adults may explain observed dissimilar outcomes.34

Ideally, the decision to treat children should depend on a well-defined risk to develop severe liver disease. Children whose liver disease would be more likely to progress would be treated, while those expected to have a benign course would be spared treatment and thereby medication-related adverse events. Since, as mentioned above, the natural history of chronic HCV infection in children is poorly understood and risk factors associated with disease progression in this population are unknown, deciding if and when to treat children is currently a challenging and sometimes contentious proposition (Table 3). However, children should not be categorically denied antiviral treatment solely on the basis of their age. The use of heterogeneous treatment regimens in small, mostly uncontrolled, clinical trials of IFN monotherapy for childhood HCV makes it difficult to draw definitive conclusions, but generally, SVR appear better in children (36%) than in adults (≈10%).35, 36 Because of the observed synergy in adults, several trials have assessed IFN plus ribavirin in children.37–40 Overall, SVR occurred in 73 of 144 (51%) children treated with IFN plus ribavirin in these trials. Several randomized clinical trials in adults verify considerably better virologic response rates (50%-60%) with the use of pegylated (peg) interferons, particularly when combined with ribavirin.41, 42 To date, only two published reports have evaluated peg IFN in children with chronic HCV. In one pilot trial of children with HCV genotype 1 infection, 5 of 13 (38%) given pegIFN alfa-2a alone achieved SVR.43 In the other trial, 36 of 62 (58%) HCV-infected children treated with pegIFN alfa-2b with ribavirin attained an SVR, however virologic response was seen in only 22 of 46 (48%) of those with HCV genotype 1 infection.44 Factors associated with a favorable response to IFN alone or in combination with ribavirin in children are similar to those in adults and include infection with HCV genotypes 2 or 3 and low pretreatment HCV RNA levels. In one small study, children younger than 12 years of age appeared to fare better.40 As in adults, African-American children had poorer response rates to combination treatment than Whites but the small number of patients enrolled limited the analysis.40 Sustained virologic response rates are similar between HCV-infected children with normal and abnormal pre-treatment serum AT levels.36, 40, 41 Important factors that may forecast response early in the course of treatment and in individualizing therapy in adults, such as early virologic response and ‘duration of undetectability’ have not been studied in children.45–47 Adverse events are common in children during treatment with IFN products alone or in combination with ribavirin and include influenza-like symptoms, anorexia, weight loss, depression and hematological abnormalities.

Table 3. Children With HCV Infection: To Treat or Not to Treat?
In Favor of TreatmentAgainst Treatment
Prevent disease progressionAsymptomatic disease in most
Caregivers facilitate complianceSlow rate of disease progression
Remove social stigmaLow overall response rates
Avoid potential neurocognitive sequelaeExpensive therapy
Economic factorsSignificant side effects
 • Costs reduced (less drug used) • Flu-like syndrome
 • Reduce HCV liver disease costs • Bone marrow depression
 • Minimize Unemployment costs • Depression
Factors associated with favorable response in adults are common in childrenEvolving future therapies
 • Young age 
 • Mild histological disease 
 • Little co-morbidity 

In contrast to adults, the rates of SVR appear to be similar between HCV-infected children treated with IFN products alone or with ribavirin. However, the pediatric data derive primarily from small and uncontrolled studies, which currently hinders our ability to provide firm therapeutic recommendations. The results of the ongoing PEDS-C study,48 a randomized, placebo-controlled trial comparing pegIFN alone or with ribavirin, will better define optimal treatment for children with HCV infection.

In a relatively short period of time, significant strides have been made in our understanding of HCV infection, particularly in adults. It is clear from the foregoing discussion that there is a critical need for studies to better delineate the natural history, pathogenesis and outcome of HCV infection in children to help formulate evidence-based management strategies for this unique population. The establishment of multi-center consortiums such as PEDS-C48 and the European Pediatric Hepatitis C Virus Network49 will greatly facilitate the fulfillment of this goal.

Nonalcoholic Fatty Liver Disease (NAFLD)

By contrast to the situation with chronic hepatitis B and even chronic hepatitis C in childhood in 2006, comparatively little is known about the pathophysiology and natural history of nonalcoholic fatty liver disease (NAFLD) in children.50 Therapy for this disorder is only at the earliest stages of comprehensive and systematic investigation. With respect to childhood NAFLD the objective at this conference was to determine the current status of this disorder in children and project the direction(s) for future research relating to treatment options (Table 4).

Table 4. New Developments or Issues Relating to Childhood NAFLD Highlighted at the STC
IssueOld assessmentNew assessmentComment
Natural historyRather benignSerious liver disease (cirrhosis; recurrent cirrhosis after liver transplantation; hepatocellular carcinoma in early adulthood)Pediatric hepatologists are well-justified to take a more aggressive approach to diagnosis and treatment of this disorder.
Role of liver biopsyCan be difficult to perform: use clinical parameters only or delay depending on evolution of diseaseSometimed difficult but highly informative: use earlyThis is still highly controversial: objective criteria needed to guide selection of patients for liver biopsy.
Coincidental liver diseaseDefine NAFLD in the absence of other definable liver diseasesNAFLD may occur in children who also have chronic hepatitis B or CNAFLD poses important problems for treatment of the chronic viral hepatitis in children.

Nonalcoholic steatohepatitis (NASH) was first described in adults in the late 1970s on the basis of histopathological findings. Since then, NASH has been identified as being part of a spectrum of chronic liver disease associated with disordered insulin action on the liver and most often with overweight or obesity. NAFLD as a spectrum of liver disease comprises simple steatosis (large-droplet fat accumulation in hepatocytes without attendant inflammation), NASH itself (large-droplet fat accumulation in hepatocytes associated with inflammation and fibrosis), and finally cirrhosis, which in adults often looks cryptogenic. NASH was reported in children in the early 1980s, and nowadays NAFLD has proved to be extremely prevalent in both adults and children (Table 5). This is partly due to increased awareness of the diagnosis, but the exponential rise in overweight and obesity generally, and certainly in children, in the past 20 years has apparently contributed to its increased prevalence. Currently, the population prevalence of fatty liver in American children is estimated at 7% to 10%. A comprehensive review of clinical studies in pediatric NAFLD has been presented elsewhere.50 As in adults, simple steatosis represents the preponderance of NAFLD in children; the incidence of cirrhosis in childhood NAFLD is very low, but cirrhosis certainly occurs.51 A few instances of individuals developing HCC in adulthood after having NAFLD as children have recently been reported.52, 53

Table 5. Differences Between NAFLD in Adults and Children
IncidenceF>MM>FFemale preponderance no longer so clear in adults either
HistopathologyNASH: Looks like alcoholic hepatitisNASH: Portal distribution of inflammation and fibrosis ('type 2')'Type 1' pattern in childhood NASH looks like histological pattern in adult NASH: tends to occur in older adolescents, infrequent
Differential diagnosisMainly alcoholic liver disease, also Wilson disease, other drug hepatotoxicityMainly inherited metabolic diseases, including but not limited to Wilson disease 
Confounding disordersCardiovascular disease, diabetes mellitus, polypharmacyPhysiological changes associated with pubertyIncidence of type 2 diabetes, essential hypertension rising in children

NAFLD in adults is associated with visceral adiposity (sometimes termed “abdominal obesity”), hypertriglyceridemia, hypertension, and type 2 diabetes mellitus and is considered to be the hepatic manifestation of the metabolic syndrome. Nearly every adult with NAFLD has insulin resistance. Progress of disease from simple steatosis to NASH seems to be slow in most adult patients: estimates suggest that over 5-10 years 17% of patients with simple steatosis will develop chronic steatohepatitis with fibrosis, approximately one quarter of these will go on to develop severe fibrosis, and finally approximately 1% of the original cohort will develop cirrhosis. Those who develop cirrhosis have a poor prognosis and a high risk of HCC. NAFLD recurs in the hepatic graft after liver transplantation, and recurrence of NAFLD after transplantation for cryptogenic cirrhosis provides a cogent argument for regarding these patients as having NAFLD in the first place. The factors governing the rate of disease progression in adult NAFLD are not known. They may include the character of the diet itself, the intestinal flora especially if small bowel overgrowth is present, and as yet undetermined genetic factors.

Disease mechanism in adult NAFLD is complex. Elevated levels of free fatty acids appear to play the central, important role in the pathogenesis of the liver disease as they are the source of the accumulated hepatic fat, cause hepatic insulin resistance, injure hepatocytes through oxidative stress, and activate inflammatory pathways including those involving NF-κB, TNF-α and endoplasmic reticulum stress.54, 55 Oxidative stress itself is also an important component of the disease process. Adiponectin, produced by adipocytes, is anti-steatotic and anti-inflammatory. Leptin, norepinephrine, and angiotensinogen are adipocytokines which may enhance development of fibrosis, when levels are elevated; decreased levels of adiponectin favor fibrosis. Although the disease mechanism in NAFLD in children is not entirely clear, insulin resistance certainly plays a pivotal role.56 The degree of insulin resistance in childhood NASH is greater than that found with obesity alone.57

Children with NAFLD are often entirely free of symptoms, although many present with vague abdominal pain which may localize to the right upper quadrant. Some children are identified incidentally with either liver tests or a liver sonogram. NAFLD occurs more frequently in boys than in girls, and the average age for diagnosis is approximately 12 years, although NAFLD may be diagnosed at almost any age except in infancy. Most, but not all, children are obese. Some ethnic predisposition has been noted.58 Physical examination can be problematic since significant organomegaly may be difficult to appreciate.59, 60 Acanthosis nigricans is present in many patients. Laboratory tests in childhood NAFLD require critical evaluation. Children with simple steatosis are expected to have normal serum aminotransferases, but those with actual steatohepatitis may also have normal serum aminotransferases. When aminotransferases are elevated, the ALT is higher than the AST, usually twice as high. Gamma glutamyl transpeptidase may be a sensitive marker of NASH in children.57 Hyperlipidemia is typical, with prominent elevations in serum triglycerides, although serum cholesterol is frequently elevated as well. Fasting insulin levels are significantly elevated in children with NAFLD, but glucose intolerance is exceptional.

Histology is essential for the definitive diagnosis of NAFLD. Large-droplet (macrovesicular) steatosis is a requisite finding, although it may not be prominent in the cirrhotic phase of NAFLD. NASH is a specific pattern of liver injury: it is not simply the concurrence of hepatic steatosis and inflammation in the absence of a definable disease process.61 In adults the classic findings include ballooned, hydropic hepatocytes, which may contain Mallory bodies or megamitochondria or demonstrate other signs of oxidative injury. Inflammatory infiltrates can include lymphocytes and macrophages, and neutrophils may be identified around hepatocytes containing Mallory bodies. This hepatocellular injury occurs mainly in zone 3 of the acinus. Fibrosis is perisinusoidal, initially around terminal hepatic venules and later extending toward portal tracts. Distinctive features of NASH may be lost as cirrhosis develops. In children the histopathology of NASH frequently differs from that found in adult NASH.57 Inflammation and fibrosis tend to be portal and periportal. Inflammatory infiltrates are typically composed of lymphocytes; hepatocellular ballooning and Mallory bodies are not found. Portal tracts are expanded with fibrosis and bridging fibrosis between portal tracts may be found. The distribution of steatosis may also tend toward zones 1 and 2. Overall, the injury is not located in zone 3. The basis for this difference between adults and children is not known.

Current practice typically employs a clinical diagnosis of NAFLD based on imaging studies to identify hepatic steatosis, biochemical measures of liver dysfunction, elimination of competing diagnoses, and finally assessment of overweight/obesity (whether visceral adiposity or metabolic syndrome or both are present). Despite being the acknowledged “gold standard”, currently liver biopsy does not figure into the immediate diagnostic algorithm for most, or even many, patients. How best to utilize liver biopsy in the diagnostic process is still highly controversial, as was evident in discussions at the STC. The critical need for objective criteria (relating to hepatic or metabolic derangement) to guide this decision was apparent, but there was also a growing consensus that liver biopsy should be performed early in the diagnostic assessment, especially in children with significantly elevated aminotransferases. Assessment of obesity currently depends on determining the body mass index and evaluating it based on the age and gender of the child using Centers for Disease Control and Prevention or other well-validated reference values.62 Measurement of mid-arm circumference and triceps skin fold thickness provides data relating to lean mass (from calculated mid-arm muscles circumference) and body fat (triceps skin fold thickness), also well-validated for children. Use of direct or indirect measures of visceral adiposity remains a research endeavor at the present time.

Treatment is aimed at the gradual reduction of body weight. Surprisingly few well-designed clinical trials are available in children to show that weight loss is the right treatment strategy for childhood NAFLD.63 Since insulin resistance leads to defects in both carbohydrate and fat metabolism and most children with NAFLD are overweight or obese, a nutritional strategy for treatment should aim to eliminate hyperinsulinemia and manipulate carbohydrate and fat intake with the overall objective of achieving a healthy body weight. The low-glycemic index diet may be an effective and physiological nutritional intervention, and it has been effective in adolescents.64 Severe caloric restriction or extremely rapid weight loss may be harmful. Exercise programs have not been evaluated in childhood NAFLD, and while increasing exercise is ineffective by itself for effecting weight loss, in combination with the appropriate diet it enhances weight loss and normal insulin action. Conversely, inactivity generally due to screen-viewing (television, computer, handheld electronic games) is associated with increased obesity. Drug treatment for childhood NAFLD has not been examined rigorously: metformin57 and vitamin E are currently being investigated in an NIH-supported prospective multi-center trial.

Therapies aimed at weight reduction in the morbidly obese child have not been used in childhood NAFLD to any great extent. These treatment options include sibutramine, which affects perception of satiety by inhibiting cerebral uptake of 5-hydroxytryptamine and norepinephrine, and orlistat, a pancreatic lipase inhibitor. They have been used successfully in adolescents with severe obesity although issues relating to both safety and efficacy of these drugs in this age-bracket remain unresolved.65 Likewise, the role of bariatric surgery in children and adolescents has not been defined66 although it maybe effective in some adults with NASH.67 An important lesson from broad experience with the treatment of obesity is that all interventions, especially those with an important behavioral component, work best when family-based. Strategies for treating NAFLD in adults may have relevance to childhood NAFLD in the future. Relevant drugs might include pentoxifylline, lipid-lowering agents such as atorvastatin or the antioxidant probucol, and angiotensin II receptor antagonists. If gut-derived bacterial toxins have a role in the disease mechanism of NAFLD, then manipulation of this microenvironment by probiotics may be a promising novel approach to treatment.

Understanding disease mechanism, critical clinical and pathological features of NAFLD in children, and the overall natural history of the disease in children is absolutely critical for permitting effective design of clinical research to investigate this disease. Progress with developing rational therapy of this disorder absolutely requires well-designed randomized controlled trials. Forthcoming research initiatives should concentrate on these issues in childhood NAFLD. The importance of primary prevention, a prime consideration in pediatrics generally, needs strong emphasis. Raising children with a healthy body weight should obviate NAFLD in most of them. This requires reassessment of normal height/weight charts for infants and children, as for example recently achieved by the World Health Organization, and research relating to the socioeconomics of obesity. While these priorities seem somewhat removed from the domain of hepatology, they have direct relevance to this new important liver disease.


The authors appreciate faculty presenters for careful review of the manuscript and conference participants, whose enthusiastic participation made the symposium uniquely successful.