State of the Problem
One of the most problematic comorbidities of obesity is the spectrum of disease known as nonalcoholic fatty liver disease (NAFLD).
The association between obesity and steatosis, inflammation, and fibrosis was first noted in 1958 (1). However, the impact of this finding was not recognized until obesity became epidemic in the adult and child populations. NAFLD describes a continuum of conditions that range from simple steatosis at one end of the spectrum, through nonalcoholic steatohepatitis (NASH), to cirrhosis and end-stage liver disease (2).
The presence of hepatic steatosis in the general population has been estimated to be between 10% and 25%, making it the most common liver disease in developed countries (3). The prevalence of NAFLD is predicted to rise, driven by the increases in obesity, insulin resistance, and diabetes, the major risk factors for the disease.
The natural history of progression from NAFLD through NASH to cirrhosis is not well delineated, particularly in childhood. In some adult series, fibrosis and even cirrhosis are present in 15% to 50% of patients on initial biopsy (4). In a retrospective study of liver biopsies in pediatric patients believed to have NASH, some degree of fibrosis was present at the time of biopsy in 14 of 14 specimens (5). In a prospective study, 24 obese pediatric patients with NASH had liver biopsies, 17 (71%) had fibrosis, and 1 child had cirrhosis at diagnosis (6).
The prevalence of NAFLD in adults is estimated to be as high as 23% (7), with estimates of NASH (presence of NAFLD with inflammation and hepatocellular necrosis) in the adult population ranging between 1% and 3% (8) and 18.5% in obese adults (9). Liver disease in obese children and adolescents is largely a silent disease, usually discovered incidentally or on screening with little in the way of signs and symptoms. Hepatic steatosis was found in 53% of obese children evaluated
by ultrasonography, with elevations of liver enzymes in 25% (10). Data from the National Health and Nutrition Examination Survey III (NHANES III) (1988–1994) showed that overall, 6% of overweight adolescents and 10% of obese adolescents already had elevated alanine aminotransferase (ALT) levels (11), an indication of hepatic inflammation and possible NASH (6).
Most patients with NAFLD have insulin resistance. Components of the metabolic syndrome, obesity, diabetes, and hyperlipidemia, are associated with NASH. In one study, 87% of patients with NASH fulfilled the criteria for the metabolic syndrome (12), and in studies of adult diabetic patients, more than one quarter have been found to have NASH (13). Type 2 diabetes, along with a higher body mass index (BMI), has also been associated with a greater rate of progression of fibrosis (14). When placed in the context of the escalating obesity epidemic and the increase in type 2 diabetes, the impact of this disease on affected children and adolescents is enormous.
The term NASH was first used by Ludwig and colleagues to describe hepatic histology consistent with alcoholic hepatitis in a group of obese, diabetic women who denied alcohol use (16,17).
The evolution of NASH is thought to be a two-stage process (18). The “first hit” is accumulation of fat in the liver, and the “second hit” is believed to involve increasing oxidative stress, which acts as a catalyst for the progression of simple steatosis to NASH (19). Free fatty acids (FFAs) accumulate within the hepatocytes and undergo oxidation, generating reactive oxygen species, which can react to form peroxides; the peroxides then cause membrane injury and release of tumor necrosis factor-α (TNF-α), which ultimately stimulates fibrosis (19) (Fig. 10.1).
Macrovesicular fat accumulates in the liver of patients with obesity and results from a combination of the following:
In a study of 144 adult patients (20), NASH was diagnosed by a persistent elevation of transaminases for more than 3 months, liver biopsy specimen with greater than 10% fat, lobular and/or portal inflammation and/or Mallory bodies, fibrosis, or cirrhosis with exclusion of other liver disease. In these adults, 26% had inflammation without abnormal fibrosis, 57% had some degree of fibrosis, and 17% had cirrhosis (20).
Most of our knowledge of the natural history of NAFLD and NASH comes from adult studies. Steatosis without inflammation or fibrosis has been followed up in adult patients for up to 16 years with no progression to cirrhosis (21). No similar study has been performed in children. There is, however, controversy over whether steatosis is a benign process.
NAFLD and NASH may be completely asymptomatic. In a series of adult patients with the metabolic syndrome but normal liver enzymes, positive findings of NASH were noted in 58 of 80 patients, 26 of 80 patients had fibrosis, and 8 of 80 patients had silent cirrhosis. Risk factors for fibrosis were female gender, long history of obesity, metabolic syndrome, and BMI greater than 45 (22).
Progression to NASH involves the development of inflammation and fibrosis, the suggested second hit. Elevation of liver transaminases is taken as an indication of inflammation, but fibrosis can be detected only by liver biopsy.
In one study, elevated ALT, hypertension, and insulin resistance were predictors of NASH in severely obese adults undergoing obesity surgery (23). Fifty percent of the patients identified with NASH also had diabetes. In those with diabetes, only 6%
had no evidence of NASH (23). Independent predictors of fibrosis in adult patients with NASH include age (>45 years), BMI greater than 31, and type 2 diabetes. An aspartate aminotransferase (AST)/ALT ratio of greater than 1 was associated with more severe fibrosis (20).
FIG. 10.1. Nonalcoholic fatty liver disease (NAFLD) to nonalcoholic steatohepatitis (NASH). (Reprinted from
Day CP, James OF. Steatohepatitis: a tale of two “hits”? Gastroenterology. 1998;114:842–845, with permission.
Hepatocellular necrosis on biopsy has been found to be a poor prognostic sign. In a series of 98 adults who were followed up for 10 years, 21% of patients with fat and ballooning degeneration and 28% of patients with fat and fibrosis developed cirrhosis compared with 4% of patients with fatty liver alone (24).
The first hit in the pathophysiology of NAFLD is the accumulation of fat in the hepatocyte. In the presence of hyperinsulinemia, the adipocyte responds by increasing lipolysis and FFA delivery to the liver. Higher levels of insulin also prompt the hepatocytes to increase fatty acid synthesis and decrease fatty acid oxidation. There also may be increased degradation of apolipoprotein B 100, which is responsible for transporting excess FFAs out of the liver. All of these responses increase fat accumulation in the hepatocyte. Peripheral and hepatic insulin resistance has been found in almost all patients with NAFLD, irrespective of the coexistence of impaired glucose tolerance or obesity. Individuals with NAFD show impaired insulin-mediated suppression of hepatic glucose production compared with controls (12,25). Insulin resistance is significantly more common in patients with NASH than in those with other causes of noncirrhotic chronic liver disease (12). In a retrospective review of patients diagnosed with NASH, 18% of patients had an affected first-degree relative, suggesting the contribution of an inherited defect (26).
Fatty livers are unusually vulnerable to injury (17) and at increased risk for damage from other factors such as viruses, endotoxin, alcohol, and toxins (27). Multiple factors have been associated with the second hit, triggering the progression from steatosis to NASH. These include the following (18,19):
Age is also a risk factor for cirrhosis, which may reflect the longer duration of risk for the second hit (20). Compared with controls, patients with NASH have
increased expression of TNF-α messenger ribonucleic acid (mRNA) in liver and adipose tissue. The degree of overexpression correlates with histologic severity (28). TNF-α and endotoxin activate stellate cells, which increases collagen type I, the major collagen in fibrotic hepatic tissue (29). Higher serum TNF-α levels correlate with increased severity of NASH as manifested by greater inflammation and fibrosis (28).
NASH may be a recurring disease; fatty liver followed by steatohepatitis has been noted to develop after liver transplantation for cryptogenic cirrhosis (30). Leptin, which is present in elevated levels in obese individuals, promotes insulin resistance and in animal models alters insulin signaling in hepatocytes, resulting in increasing hepatocellular fatty acid production (31). Thus, elevated leptin levels occurring in obese patients may contribute to a metabolic environment that sustains hepatosteatosis.
Clinical symptoms of NASH, when they occur, are subtle and may include mild abdominal pain and fatigue. In adults, generalized fatigue, lethargy, and mild epigastric or right upper quadrant pain have been noted in 30% of patients (19).
If liver disease has progressed, patients may experience pruritus, anorexia, and nausea. When cirrhosis is present, patients may develop anasarca, variceal hemorrhage, and/or symptoms of hepatic failure (25). In a study of adults with NASH, obesity predicted progression of fibrosis, with one third having progression of fibrosis on a second biopsy 4 years after presentation. There were no differences between both groups regarding age, gender, diabetes, hyperlipidemia, ALT levels, AST-to-ALT ratio levels, albumin levels, prothrombin activity, steatosis, or inflammation (32). In a study of 310 obese children in Japan, 24% had an elevation in ALT. In a subgroup of 77 obese children with ALT values greater than 30 IU/L, 83% had a fatty liver on ultrasound compared with 18% of 27 children with normal ALT levels (33). Other factors associated with elevated ALT levels in overweight and obese adolescents include increased age, elevated glycosylated hemoglobin, elevated triglycerides, and decreased levels of serum antioxidants—vitamin E, β-carotene, and vitamin C (11).
In a series of 36 children, seen between 1985 and 1995, who presented with unexplained liver enzyme elevations to a gastroenterology clinic, 30 of the 36 patients were obese, 16 had hepatomegaly, and 13 had acanthosis nigricans. The mean value of AST was 104 ± 16 U/L (normal <37 U/L; range 26–523 U/L); the mean ALT value was 179 ± 31 U/L (normal <40 U/L; range 10–644 U/L). Liver ultrasound was performed in the majority of patients and showed increased echogenicity suggestive of fatty infiltration. There was no correlation between the severity of obesity and liver transaminase elevations. Biopsy was obtained in 24 of 36 patients. On biopsy, all children had large droplet (macrovesicular) steatosis, 21 children (88%) had inflammation, and 17 children (71%) had fibrosis. Fibrosis was moderately severe in 7 children and occurred in 2 children without evidence of inflammation (6). In a study of 228 obese children in Japan, hyperinsulinemia was correlated with NASH (34).
All obese children and adolescents should be evaluated for NASH.
Hepatic steatosis is prevalent in the obese population. A finding highly suggestive of hepatic steatosis on physical examination is hepatomegaly; however, this can sometimes be difficult to detect because of the degree of obesity. In addition, the finding of acanthosis nigricans should increase one's suspicion that the child may be at risk for hepatic steatosis. Liver ultrasound is the modality most commonly used for detecting hepatic steatosis. Computed tomography (CT) and magnetic resonance imaging (MRI) are also able to detect hepatic steatosis, although these modalities are more expensive and the CT scan unnecessarily exposes the patient to radiation. The diagnosis of NAFLD can be made in the presence of hepatic steatosis after other causes of fatty liver are excluded (Table 10.1). Tests of liver inflammation should be performed and include AST, ALT, and γ-glutamyltransferase (GGT). Tests for alkaline phosphatase (ALK PO4), total bilirubin, and albumin can also be considered. Other
causes of liver inflammation should be eliminated before making the diagnosis of NASH (Table 10.2).
TABLE 10.1. Differential diagnosis of hepatic steatosis
Liver biopsy remains essential for diagnosing and evaluating the progression of NASH.
Imaging studies, as noted previously, are able to detect steatosis but are not accurate predictors of fibrosis. If the previously mentioned evaluation does not reveal a cause of elevated liver enzymes, a liver biopsy should be considered because this is
the most accurate means of truly assessing the degree of inflammation and fibrosis, which is essential for making a definitive diagnosis and evaluating the prognosis of NASH.
TABLE 10.2. Evaluation of elevation of liver enzymes in obese children
Indications for liver biopsy are not standardized in adults and are even more problematic in children and adolescents. Day (35) lists some indicators for deciding on biopsy in adults:
Biopsy findings of NASH include steatosis, predominantly macrosteatosis; ballooning of hepatocytes; perisinusoidal fibrosis; and mixed lobular inflammatory infiltrate (36).
The classification of NAFLD and NASH is histopathologic and is outlined in Table 10.3. The histologic appearance of NASH is illustrated inFigures 10.2, 10.3, and 10.4.
There is no single standardized treatment for patients with NAFLD or NASH. The treatment of NAFLD should focus on preventing or reversing events that provoke inflammation and insulin resistance.
FIG. 10.2. A liver biopsy from a 10-year-old with nonalcoholic steatohepatitis (NASH) shows fatty infiltration (clear) with intrahepatocellular fat, portal inflammation, and bridging fibrosis.
FIG. 10.3. A higher magnification view of extensive bridging fibrosis.
Because of the association of NASH with obesity, weight loss has been a mainstay of treatment. Early studies of adult postsurgical bariatric patients showed that drastic weight loss could lead to increased inflammation and fibrosis. This was thought to be due to increased production of FFAs and lipid peroxidation increasing oxidative stress. This finding has led to caution and a recommendation to avoid rapid weight loss in these patients (37). Gradual weight loss and control of diabetes will reduce hepatosteatosis as well as transaminase elevation (38,39).
FIG. 10.4. High-power view of panacinar steatosis.
TABLE 10.3. Brunt Classification of nonalcoholic fatty liver disease
In a small study of adults who used diet and exercise to reduce weight over a 3-month period, liver enzymes improved, as did liver histology, compared with no change in the control group (40). In another series of adults who underwent adjustable gastric band placement and weight loss, there were improvements in lobular steatosis, necroinflammatory changes, and fibrosis on a second biopsy 2 years after surgery. This improvement was more pronounced in patients with characteristics of the metabolic syndrome who had been more severely affected prior to surgery (41). In a small series of pediatric patients with elevated aminotransferases and fatty liver on ultrasound, those who lost at least 10% of their excess weight showed normalized ALT and AST values and decreased ultrasound evidence of fatty infiltration (39). In another series of obese children who underwent ultrasonography at the start of a diet and exercise program, 53% had evidence of liver steatosis noted on ultrasound and 25% had elevated transaminase values. Both steatosis and liver enzyme elevations resolved with weight loss (10).
Orlistat, a gastric and pancreatic lipase inhibitor, has been studied in small groups of adult patients as a treatment for NASH. In one study of 10 obese adults with biopsy-proven NASH, treatment with orlistat and dietary counseling for 6 months improved steatosis in more than one half and fibrosis in one third of the patients (42). No similar studies have been performed in children.
Metformin has been tried in the treatment of NASH in adults and, despite improvements in insulin resistance and a decrease in liver enzymes, there was no significant change in inflammation or fibrosis when compared with patients with NASH who underwent dietary treatment alone (43).
Treatment of NASH with peroxisome proliferator activated receptor agonists (thiazolidinediones) remains experimental and has been tried in adults. These drugs have not been tested in children. The initial studies were carried out with troglitazone, which has been withdrawn from the market because of hepatotoxicity. Improvement in steatosis, cellular injury, parenchymal inflammation, and fibrosis has been seen with rosiglitazone; however, weight gain and increased liver enzymes occurred after treatment (44).
Because oxidative stress is considered to play a role in the etiology of NASH, antioxidants such as vitamin E have been used in treatment. Results have been disappointing, with most studies showing no or only minor improvement in inflammation and fibrosis (45). A pediatric study using vitamin E showed normalization of liver enzymes but did not include liver biopsy in the study (46).
Currently, treatment of NASH in children should center on a nutrition and exercise plan to promote gradual weight loss, treatment of diabetes if present, and avoidance of other agents that are hepatotoxic. Orlistat is approved for weight loss in children but has not been evaluated in the pediatric age group for treatment of NASH. Inhibition of fibrosis remains a target of therapy for NASH, but no drugs are yet approved as antifibrotic drugs in humans.
Alcohol intake is associated with severe liver disease in obese adults (47). In one study, approximately 50% of obese adolescents who reported modest alcohol ingestion (four times per month or more) had elevated ALT levels (11). Ethanol has been found to induce cytochrome P450, which is responsible for activating metabolites to toxic intermediaries and is already increased in patients with NASH. There is concern about further insult to the liver in children with NASH, and agents associated with liver toxicity, especially alcohol, should be avoided.
All children with BMI greater than the 95th percentile should be screened for NASH with liver inflammation studies (ALT, AST, GGT). Children with elevated transaminases should have a thorough evaluation to rule out other causes of liver disease. Ultrasound may be performed to document steatosis. Children should be evaluated for diabetes and impaired glucose tolerance, which should be addressed if necessary, as well as for other components of the metabolic syndrome (i.e., hypertension, hyperlipidemia, elevated waist/hip ratio). Assessment for any additional obesity-related comorbidities should be completed. Using lifestyle modifications of nutrition and physical activity, weight loss is an important goal. Liver biopsy should be considered to definitively determine diagnosis and evaluate for inflammation and fibrosis as a means of assessing severity and prognosis. If fibrosis is present, lifestyle intervention should be intensified and liver disease closely monitored, with consideration being given to enrollment in promising therapeutic trials if available.
Hypothalamic and Pituitary Disease
Patients who have hypothalamic and/or pituitary dysfunction may be at increased risk for NAFLD.
In one series of 21 patients with hypothalamic/pituitary dysfunction caused by brain tumor, idiopathic hypopituitarism (6/21), hypophysitis (1/21), and Prader-Willi syndrome (1/21), NAFLD was diagnosed within 6 years of diagnosis. In follow-up, 2 underwent liver transplant, and 1 died of hepatocellular carcinoma (48). Children with hypothalamic injury or disease should be evaluated and followed up for the development of NAFLD and NASH; aggressive lifestyle therapy focusing on nutrition and activity should be pursued.
Ongoing Support: Case of MK
MK is a 13-year-old white boy who comes for a physical examination prior to playing fall sports. He is 5 ft 7 in. tall and weighs 195 lb, with a BMI of 30.5. His blood pressure is 123/78 mm Hg. He has had no serious illnesses since he was last seen 1 year ago. He has had some minor complaints of intermittent nausea and occasional headaches. He is not taking any medication. He has done well in school and participates in several sports. His parents are both overweight, and his maternal grandfather has just been diagnosed with type 2 diabetes and hypertension.
You perform a review of systems, paying particular attention to obesity-related comorbidities. He is having occasional headaches, which are self-limited, and he reports no difficulties with vision. He denies shortness of breath, wheezing, snoring, restless sleeping, or daytime tiredness. He does have some nausea, which he attributes to “what I ate.” He is not complaining of any hip or knee pain. He reports he is “feeling good” and has not been sad or unhappy. On physical examination you note that his blood pressure is slightly elevated, he has centripetal obesity, and he has mild acanthosis nigricans of his neck and axilla. He is Tanner stage III. Because he meets the criteria of obesity/overweight for an adolescent (BMI is at the 95th percentile for age and gender), you screen for obesity-related comorbidities with laboratory studies. You let MK and his mother know that you are ordering laboratory studies for diabetes and for dyslipidemia and liver disease.
You begin your discussion with MK and his mother by showing them his growth chart and BMI. Mrs. K comments that MK is a “big boy” just like the rest of the family and says that his weight does not worry her. MK denies any concern about his weight but does say that he wishes he could run a little faster. You review the family history and note that not only his maternal grandmother but also several aunts and
uncles have hypertension and diabetes; his mother says that MK's doctor had just mentioned that he had “borderline” diabetes at his last physical. You mention that MK has slightly elevated blood pressure and acanthosis nigricans and discuss these findings in light of the family history and the risk of type 2 diabetes. Mrs. K says that the family “should probably do something about it.”
With this lead you ask MK to review his daily eating and activity with you. He notes he is skipping breakfast, eating a school lunch, buying a soda and chips after school, and eating one to two portions of dinner. The family eats out two or three times per week, and Mrs. K notes that MK often brings snacks to his room after dinner to eat while playing computer games. MK estimates that he is drinking about five regular sodas per day. When you ask about screen time, Mrs. K gives a pointed look at MK and says, “he is playing on the computer all the time.” MK says that besides his 1 hour of homework, he plays on the computer from the time he gets home from school or sports until bedtime. You estimate about 4 to 5 hours per day.
You explain to MK and his mother that the five sodas a day that he is drinking represent about 750 kcal, the equivalent of 50 teaspoons of sugar. His mother is surprised and expresses willingness to buy water and diet soda for the family instead. MK shrugs and says that drinking diet instead of regular soda “would be ok.” You agree with his mother that this is a good place to start, and you schedule a follow-up visit to review test results and note progress on the change from regular soda.
Three Weeks Later
When MK returns in 3 weeks, you note that his weight is 192 lb, his height is 5 ft 7 in., and his BMI is 30.1. His blood pressure is 119/76 mm Hg. He and his mother report that the family is now drinking diet drinks or water between meals; MK says he has a regular soda “once in a while,” usually when he is with his friends. You congratulate them on making this important change and review the results of the laboratory tests with them (Table 10.4).
You begin by explaining that MK's elevated fasting insulin indicates insulin resistance and places him in the continuum of risk for type 2 diabetes, especially coupled with his high BMI, acanthosis nigricans, and positive family history. You note that the high values of cholesterol and especially triglycerides often accompany insulin elevations. In addition to these findings, his liver enzymes are elevated, which indicate
inflammation in his liver, possibly due to fatty infiltration. This alarms his mother because she never realized the liver could be affected by his weight gain.
TABLE 10.4. Case study laboratory test results
In consultation with a pediatric gastroenterologist, you order laboratory studies to rule out other causes of liver disease (Table 10.2) and order a liver ultrasound to detect fatty infiltration. You discuss liver biopsy, but the gastroenterologist points out that there are currently no pediatric criteria for biopsy (studies are in progress) and MK does not meet the adult criteria. You decide to pursue the laboratory work and ultrasound while you are working with MK and his family on a weight loss plan.
You explain your discussion with the pediatric gastroenterologist, the fact that you are ordering laboratory tests to rule out other causes of liver problems and an ultrasound of the liver. You emphasize that the only known treatment for NASH is weight loss and let MK and his mother know they have made a good start and you will be working with them to continue the process of change. You suggest that they may want to consider another change in eating and activity to decrease MK's energy intake. Together you, MK, and his mother decide that reducing the number of times he and the family eat out and substituting healthy snacks would be the next step. You also work with MK on a plan to exercise on the days he does not have sports practice and to try cut back 1 hour on his computer use.
You review the diagnostic and treatment plan and also caution MK that certain medications have liver-related side effects and he should check with you before taking any new medicine; emphasize that this is an additional reason not to drink alcohol.
You plan to see MK monthly to check on his weight and lifestyle changes, as well as to keep him updated on new diagnosis and treatment data from studies of NASH as they are completed.