Histopathology and pathogenesis of obesity-related liver disease

Obesity is associated with a spectrum of fatty liver diseases, the most common of which is hepatic steatosis (fatty liver). Population-based studies that relied on sensitive abdominal imaging techniques to detect hepatic triglyceride demonstrate that about a third of the United States adult population has hepatic steatosis. In those studies, fatty liver was strongly associated with obesity. Other risk factors include insulin resistance and dyslipidemia, conditions that are also extremely prevalent in the general population. Although fatty liver disease may also occur in nonobese individuals, including those with lipoatrophy or who consume excessive alcohol, overweight-obesity remains one of the greatest risk factors for fatty liver disease.

Fatty liver is associated with obesity because the liver responds to caloric excess by synthesizing triglyceride. Normally, these triglycerides are then exported from the liver in lipoproteins and stored in peripheral adipose depots. Triglycerides are retained in the liver when their assimilation into peripheral adipose tissues is impaired or when lipoprotein export mechanisms become inefficient. Systemic insulin resistance is a common cause of the former, whereas various conditions that interfere with efficient packaging and secretion of lipoproteins from hepatocytes (eg, abetalipoproteinemia, choline deficiency) cause the latter.

Based on analysis of case series at referral centers, it is estimated that about a quarter of the individuals with fatty livers have steatohepatitis. Steatohepatitis is a more serious form of liver damage than simple steatosis because hepatocyte injury or death is significantly more prominent in steatohepatitis than in simple steatosis, and this hepatic injury is often accompanied by inflammatory cell accumulation and some degree of fibrosis. Obesity-related steatohepatitis resembles alcohol-induced steatohepatitis histologically, and obesity seems to increase the risk of developing alcohol-induced fatty liver. Hence, a clinical diagnosis of nonalcoholic steatohepatitis (NASH) can only be made when it is certain that there has been less than or equal to only social levels of alcohol consumption (ie, ≤2 drinks per day).

Analysis of groups of nonalcoholic patients who have undergone sequential liver biopsies suggests that as many as a third of NASH patients may progress to cirrhosis within a decade of follow-up, whereas progressive fibrosis rarely occurs in patients who merely have steatosis on index liver biopsy. In NASH, as in most other causes of chronic hepatitis, the risk of progressing to cirrhosis is greatest in patients who demonstrate fibrosis on their initial liver biopsy. Like patients with other types of chronic fibrosing liver diseases, patients with NASH develop bridging fibrosis as cirrhosis evolves. NASH is also characterized by a unique pattern of collagen deposition around hepatocytes and along sinusoids, however, even relatively early stages of liver injury. This pericellular-perisinusoidal fibrosis is presumably part of a wound-healing response to ongoing liver injury. It is typically most prominent in the areas of the liver lobule that are near terminal hepatic venules (zone 3), where hepatocyte injury begins in NASH, and has been dubbed “chicken wire fibrosis.” Although a subset of individuals with NASH develop progressive liver fibrosis and ultimately become cirrhotic over time, others remain stable and still others improve. At present, the factors that determine the outcomes of NASH are not well understood.

Emerging evidence also suggests that primary liver cancers, both hepatocellular carcinoma and intrahepatic cholangiocarcinoma, can occur during NASH. Fortunately, this seems to be a relatively rare occurrence until cirrhosis is well-established. Cirrhosis remains the more common underlying liver disorder in obese patients with primary hepatic neoplasms. It has been estimated that the risk of developing primary liver cancer is about 1% per year once cirrhosis has developed in NASH.

Clinical features and laboratory abnormalities of obesity-related liver diseases

Liver disease often goes undetected for decades in obese individuals because serum aminotransferase values may be normal or only mildly elevated on an intermittent basis. As in other liver diseases, hyperbilirubinemia (jaundice), hypoalbuminemia, and coagulopathy develop only when liver damage is very advanced. In addition, abdominal adiposity may compromise early diagnosis of hepatomegaly or more subtle abnormalities related to portal hypertension (ie, splenomegaly, ascites). Individuals with obesity-related liver disease typically exhibit an android pattern of adiposity (ie, relative overexpansion of abdominal adipose depots) and clinical features of insulin resistance (eg, acanthosis nigricans, polycystic ovary syndrome). They also often have laboratory abnormalities that are characteristic of the metabolic syndrome, including hypercholesterolemia; low levels of high-density lipoproteins; hypertriglceridemia; hyperuricemia; hyperinsulinemia; insulin resistance on glucose tolerance testing; and even fasting hyperglycemia (ie, diabetes). Serum levels of autoantibodies, such as antinuclear antibody or anti–smooth muscle antibody, may be increased, but elevated levels of total serum proteins rarely occur and hyperglobulinemia is not typical unless cirrhosis with portal-systemic shunting has already developed. Hyperferritinemia in the absence of other evidence of iron overload (ie, transferrin saturation or increased serum iron concentration) has been reported to occur in some patients with obesity-related fatty liver disease. Markers for chronic viral hepatitis, Wilson disease, and α ₁ -antitrypsin deficiency, however, are typically negative. Abdominal imaging studies generally demonstrate hepatic steatosis, although different approaches differ in their sensitivity for detecting hepatic fat. Proton nuclear magnetic resonance spectroscopy is the most sensitive technique, but this is not widely applied outside of research settings. Comparison of liver-spleen density on either standard MRI or CT is generally useful. Routine ultrasonography (the least sensitive test) is capable of detecting hepatic fat when accumulation is present in more than a third of the liver cells. Because there is no one serum marker that is diagnostic of fatty liver disease and fatty liver is an extremely prevalent condition, testing must be done to exclude other causes of chronic aminotransferase elevations before concluding that steatosis-NASH is the explanation for abnormal liver enzymes when fatty liver has been demonstrated by some type of abdominal imaging study.

It is also important to emphasize that the magnitude of aminotransferase elevation is not reliable for distinguishing NASH from steatosis. Serum liver enzymes can be normal in individuals with NASH or NASH-related cirrhosis and may be increased in individuals with simple steatosis. Recently, it has been suggested that serum levels of cytokeratin 8/18 cleavage products (which are produced during cellular apoptosis) are generally greater in NASH than in simple steatosis and may be helpful in differentiating individuals with NASH (who have a more ominous prognosis) from those with steatosis (who are likely to follow a benign clinical course). Combinations of clinical tests (eg, the AST/ALT ratio plus platelet count) or commercially available fibrosis markers are also somewhat helpful in identifying patients who are developing liver fibrosis. At present, however, liver biopsy remains the gold standard test for confirming the clinical suspicion of NASH, for grading the severity of resultant liver injury and inflammation, and for establishing the stage of fibrosis. As in other types of chronic liver disease, in NASH the accuracy of liver biopsy for staging the extent of liver fibrosis depends on the size of the biopsy core, and sampling artifacts may be problematic in biopsy specimens that are less than 2 cm in length.

Clinical features and laboratory abnormalities of obesity-related liver diseases

Liver disease often goes undetected for decades in obese individuals because serum aminotransferase values may be normal or only mildly elevated on an intermittent basis. As in other liver diseases, hyperbilirubinemia (jaundice), hypoalbuminemia, and coagulopathy develop only when liver damage is very advanced. In addition, abdominal adiposity may compromise early diagnosis of hepatomegaly or more subtle abnormalities related to portal hypertension (ie, splenomegaly, ascites). Individuals with obesity-related liver disease typically exhibit an android pattern of adiposity (ie, relative overexpansion of abdominal adipose depots) and clinical features of insulin resistance (eg, acanthosis nigricans, polycystic ovary syndrome). They also often have laboratory abnormalities that are characteristic of the metabolic syndrome, including hypercholesterolemia; low levels of high-density lipoproteins; hypertriglceridemia; hyperuricemia; hyperinsulinemia; insulin resistance on glucose tolerance testing; and even fasting hyperglycemia (ie, diabetes). Serum levels of autoantibodies, such as antinuclear antibody or anti–smooth muscle antibody, may be increased, but elevated levels of total serum proteins rarely occur and hyperglobulinemia is not typical unless cirrhosis with portal-systemic shunting has already developed. Hyperferritinemia in the absence of other evidence of iron overload (ie, transferrin saturation or increased serum iron concentration) has been reported to occur in some patients with obesity-related fatty liver disease. Markers for chronic viral hepatitis, Wilson disease, and α ₁ -antitrypsin deficiency, however, are typically negative. Abdominal imaging studies generally demonstrate hepatic steatosis, although different approaches differ in their sensitivity for detecting hepatic fat. Proton nuclear magnetic resonance spectroscopy is the most sensitive technique, but this is not widely applied outside of research settings. Comparison of liver-spleen density on either standard MRI or CT is generally useful. Routine ultrasonography (the least sensitive test) is capable of detecting hepatic fat when accumulation is present in more than a third of the liver cells. Because there is no one serum marker that is diagnostic of fatty liver disease and fatty liver is an extremely prevalent condition, testing must be done to exclude other causes of chronic aminotransferase elevations before concluding that steatosis-NASH is the explanation for abnormal liver enzymes when fatty liver has been demonstrated by some type of abdominal imaging study.

It is also important to emphasize that the magnitude of aminotransferase elevation is not reliable for distinguishing NASH from steatosis. Serum liver enzymes can be normal in individuals with NASH or NASH-related cirrhosis and may be increased in individuals with simple steatosis. Recently, it has been suggested that serum levels of cytokeratin 8/18 cleavage products (which are produced during cellular apoptosis) are generally greater in NASH than in simple steatosis and may be helpful in differentiating individuals with NASH (who have a more ominous prognosis) from those with steatosis (who are likely to follow a benign clinical course). Combinations of clinical tests (eg, the AST/ALT ratio plus platelet count) or commercially available fibrosis markers are also somewhat helpful in identifying patients who are developing liver fibrosis. At present, however, liver biopsy remains the gold standard test for confirming the clinical suspicion of NASH, for grading the severity of resultant liver injury and inflammation, and for establishing the stage of fibrosis. As in other types of chronic liver disease, in NASH the accuracy of liver biopsy for staging the extent of liver fibrosis depends on the size of the biopsy core, and sampling artifacts may be problematic in biopsy specimens that are less than 2 cm in length.

Comorbid conditions commonly associated with obesity-related liver disease

Obesity-related liver disease is strongly associated with certain other medical conditions. These include insulin resistance and type 2 diabetes; dyslipidemias hypercholesterolemia, low high-density lipoprotein, hypertriglyceridemia, and hypertension (ie, the metabolic syndrome); polycystic ovary syndrome; cardiovascular disease; cancers; sleep apnea; and hypothyroidism. It has been difficult to determine which, if any, of these conditions contribute to the pathogenesis of nonalcoholic fatty live and NASH and which comorbidities are actually consequences of the liver pathology. Preclinical and clinical research suggests that insulin resistance is involved in the pathogenesis of nonalcoholic fatty liver and NASH. It has also been postulated that sleep apnea causes hypoxia and stress-related hormonal responses that may contribute to the severity of fatty liver injury. Chronic liver injury may provoke activation of tissue deiodinases that inactivate thyroid hormones, however, thereby promoting hypothyroidism. Recently, hypothyroidism was identified as an independent risk factor for hepatocellular carcinoma in women. Chronic liver injury probably also contributes to systemic inflammation that is now believed to fuel progressive cardiovascular disease, exacerbate insulin resistance, and increase the risk for various types of malignancy. Sex hormones also seem to influence the prevalence and severity of fatty liver disease. Hepatic steatosis is generally more common in men than women, but nonalcoholic fatty liver disease (NAFLD)-related cirrhosis seems to occur more often in women. Among women with fatty liver disease, disease severity seems to worsen after menopause. Certain ethnic groups are also more vulnerable to fatty liver damage than others. For example, studies suggest that risk is greatest in Asians and Native Americans (including Hispanics and Eskimos), intermediate in whites, and lowest in African Americans.

Prognosis of obesity-related liver disease

Population-based studies demonstrate that fatty liver disease increases both overall and liver-related mortality. As in the general population, cardiovascular disease is the leading cause of death in patients with NAFLD. Moreover, the severity of NAFLD-related liver injury seems to influence the risk for mortality from cardiovascular disease: death from cardiovascular disease is lower in individuals who do not have fatty livers than in those who do, and higher in patients with NASH than in those with simple steatosis. Cancer is the second leading cause of death in NAFLD patients, as it is in the general adult population. Liver disease is the third leading cause of mortality in NAFLD patients, however, whereas it ranks as the eighteenth leading cause of death in adults without fatty livers.

Because NAFLD is more than an order of magnitude more prevalent in the United States than chronic viral hepatitis, NAFLD is the most common cause of cirrhosis in American adults, despite the fact that progression to cirrhosis probably occurs in fewer than 5% of the NAFLD population. Head-to-head comparison of outcomes in patients with cirrhosis related to either chronic hepatitis C or NAFLD has revealed that both types of liver disease have similarly ominous outcomes, including significant morbidity and mortality related to portal hypertension, liver failure, or primary liver cancer. Indeed, because the diagnosis of cirrhosis tends to be delayed in NAFLD compared with chronic hepatitis C, NAFLD patients are more likely to present to physicians with complications of advanced liver disease (eg, ascites, variceal hemorrhage, hepatocellular carcinoma) than patients with hepatitis C virus–related cirrhosis, who are typically diagnosed before overt symptoms and signs of liver disease emerge. Also, comorbidities and older age often preclude consideration of liver transplantation as a treatment option for patients with NAFLD-cirrhosis, rendering many of these patients dependent on medical management of hepatic decompensation.

Effects of weight loss and treating comorbidities on liver disease outcomes

The ultimate objective of therapy for obesity-related fatty liver disease is to prevent progression to cirrhosis. In addition, management must include efforts to prevent adverse outcomes from associated cardiovascular disease and cancer, which are the top two causes of death in this patient population. Because cardiovascular disease, cancer, and cirrhosis are all ultimately complications of obesity, weight reduction is a primary therapeutic target. Success requires consistent reductions in energy intake relative to energy expenditure. This can be accomplished either by caloric restriction (diet); increased exercise; or combinations of these two approaches. At present, there is not consistent evidence that restricting one particular type of macronutrient (ie, dietary fats or carbohydrate) provides benefit beyond what is realized when total calorie ingestion is reduced. Similarly, because the major benefits of exercise seem to be related to its effects on net energy expenditure with resultant reduction in overall adiposity, the level of physical activity must be sufficiently vigorous and sustained to provoke catabolism. This generally necessitates at least 30 minutes of moderately aerobic exercise five times per week. Obese patients with NAFLD often find it difficult to achieve these goals because they tend to be very physically deconditioned. Hence, diet generally becomes the mainstay of therapy, although an exercise program that includes a combination of aerobic and resistance work should be encouraged.

Weight reduction medications, such as drugs that enhance fat malabsorption, have not been proved to provide benefits beyond that achieved alone by caloric restriction. Recently, clinical trials with cannabinoid receptor 1 (CB1) antagonists were initiated because in addition to their anorexogenic actions, these drugs seemed to improve insulin sensitivity and prevent liver fibrosis in preclinical studies. The trials in humans were halted, however, because of major neuropsychiatric adverse events that were associated with chronic use of CB1 antagonists.

Bariatric surgery is another approach that is often considered to improve obesity. It is extremely effective in accomplishing long-term reductions in body mass index and improves many of the comorbidities that accompany obesity, including hypertension, dyslipidemia, sleep apnea, and insulin resistance. Indeed, the latter improves dramatically within days of surgery, long before significant weight loss is observed, suggesting that some of the benefits of bariatric surgery may be mediated by changes in gut-derived neuropeptides. Noncirrhotic patients with NAFLD generally tolerate bariatric surgery without added morbidity or mortality and seem to respond with reductions in hepatic steatosis and necroinflammation. Long-term effects on liver fibrosis remain uncertain because follow-up liver biopsies have not been done routinely. Cirrhosis remains a general contraindication to bariatric surgery, as it is to most types of elective intra-abdominal surgical procedures, because of the increased risk for perioperative hepatic decompensation and liver-related mortality. At present, comparative data on the relative merits of different types of bariatric surgical procedures on liver histology are scant, but it seems likely that the greatest reductions in hepatic steatosis will result from those procedures that produce the greatest and most sustained reductions in body mass index.

Obese patients with NAFLD typically have comorbid illnesses. Evidence proves that treating these conditions is generally not harmful to the liver. Indeed, in some circumstances, it may actually help to improve fatty liver disease. Postmarketing surveillance of statin treatment for hypercholesterolemia, for example, demonstrates that (contrary to expectations) statin therapy is associated with improvements in liver enzyme elevations in NAFLD patients. Prospective, controlled trials to evaluate the efficacy of statins as a therapy for NASH have not been reported. Small case series suggest, however, that statins may improve steatosis and hepatic inflammation. Effects on liver fibrosis have not yet been established. Fibrate therapy for hypertriglycerdemia also seems to be harmless for the liver, but there is no evidence that these agents actually improve NASH. Data from a growing number of clinical trials suggest, however, that several different insulin-sensitizing agents that are used to treat diabetes may also improve NASH (see later). Certain cardiovascular therapeutics, such as angiotensin receptor blockers, have also been evaluated as treatments for NASH, and seem to reduce both hepatic necroinflammatory activity and liver fibrosis in small, nonrandomized studies. Replacing thyroid hormones to normalize serum levels of free T4 is not known to harm (or independently benefit) the underlying fatty liver disease. Treatment of sleep apnea with continuous positive airway pressure has been shown to improve nocturnal hypoxemia and reduce related stress responses that may contribute to weight gain, hypertension, and cardiac dysfunction, but it remains to be proved that these apparent benefits of continuous positive airway pressure prevent or improve NASH.