Hepatocellular carcinoma (HCC) is the most rapidly increasing cause of cancer death in the United States.1 Although many risk factors for HCC are well defined, including hepatitis B virus (HBV), hepatitis C virus (HCV), and alcohol, most series have indicated that 5% to 30% of patients with HCC lack a readily identifiable risk factor for their cancer.2 The majority of “cryptogenic” HCC in the United States is attributed to nonalcoholic fatty liver disease (NAFLD),3 which is a hepatic manifestation of the metabolic syndrome (Fig. 1).
Metabolic syndrome is a constellation of problems that includes obesity, dyslipidemia, diabetes, and insulin resistance.4 The prevalence of metabolic syndrome is increasing, paralleling the obesity epidemic in the United States. Nearly 25% of the US population meets criteria for the metabolic syndrome,5 and US obesity rates (body mass index [BMI] >30 kg/m2) currently also exceed 25% in most regions of the country. Up to 75% of obese adults will develop fatty liver disease. Unlike the HCV epidemic, which is estimated to peak in 2010, the obesity/metabolic syndrome epidemic shows no signs of abating.6
NAFLD comprises a spectrum of disorders from fatty liver disease to progressive inflammation and cirrhosis. The prevalence of NAFLD varies widely, depending on the method of assessment. Approximately 2% to 5% of the US population has “cryptogenic” elevated liver enzymes consistent with NAFLD,7 and up to 90% of individuals with obesity have some degree of fatty liver disease.8 Ultrasound and magnetic resonance imaging (MRI) studies from the United States and other Western countries suggest that from 20% to 30% of the population has evidence of fatty liver disease attributed to NAFLD.9-11 Approximately 10% of patients with NAFLD progress to nonalcoholic steatohepatitis (NASH), and 8% to 26% of individuals with NASH progress to cirrhosis.12
Retrospective data suggest that, after cirrhosis develops, 4% to 27% of cases of NASH transform to HCC.13 These figures lead to theoretical HCC incidence rates that ranges from 0.6 in 100,000 population to 210 in 100,000 population (Fig. 2). The obesity/metabolic syndrome epidemic is relatively recent, and it is likely that several decades will be required before NASH develops into cirrhosis. Thus, to date, the NASH-HCC “epidemic” may not have fully established itself. Currently, individuals who have NASH but no underlying cirrhosis are not screened routinely for HCC at most centers because of their low risk of developing HCC. Approximately 10% to 25% of individuals with NASH go on to develop cirrhosis, but it is not yet clear what predisposes patients to this progression of disease. The ability to determine which patients will progress to cirrhosis will have important screening implications for the future.
Nonalcoholic Fatty Liver Disease and Nonalcoholic Steatohepatitis
The gold standard for the diagnosis of both NAFLD and NASH is tissue biopsy. NAFLD is characterized by hepatic steatosis in the absence of a history of significant alcohol use or other known liver disease. Alcohol intake as low as 20 g daily in women, and 30 g daily in men may be sufficient to cause alcohol-induced liver disease; 12 ounces of beer (350 mL), 4 ounces (120 mL) of wine, and 1.5 ounces (45 mL) of hard liquor each contain 10 g of alcohol.14 NASH is a progressive form of NAFLD and includes inflammatory components on pathology. The NASH Clinical Research Network designed and validated a pathologic scoring system for NASH that is the sum of measures of steatosis, lobular inflammation, and hepatocellular ballooning (a feature of cellular injury characterized by large hepatocytes).15 Sampling error may be a problem using biopsy samples to diagnose NAFLD and NASH.16 In addition, once cirrhosis has developed, NASH pathology may be difficult to evaluate, because fat often disappears as NASH transforms into cirrhosis.17 Thus, accurate nonpathologic markers that distinguish NASH from other underlying disorders would be helpful to study this entity, but none have been identified that are diagnostic. Imaging also has not been a useful tool to date for the diagnosis of NASH.18
Most cryptogenic cirrhosis in the United States is believed to be related to risk factors associated with NASH and the metabolic syndrome.19 Several studies have now demonstrated that patients with cryptogenic cirrhosis are more likely to be obese and are up to 4 times as likely to be diabetic as those with other forms of cirrhosis.20, 21 Hispanics have the highest rates of both cryptogenic cirrhosis and NASH in many studies, and they also have the most rapidly increasing incidence of HCC in the United States.21, 22 One study that defined rates of the metabolic syndrome in the United States by ethnicity indicated that Mexican Americans have the highest rate (32%), whereas 24% of European Americans and 22% of African Americans meet criteria for the metabolic syndrome.23
It is not clear why African Americans may have lower rates of NASH than other racial/ethnic groups, despite a high prevalence of diabetes and obesity. Some investigators have suggested that African American have less insulin resistance despite the presence of other risk factors of the metabolic syndrome. For example, African Americans with adult-onset diabetes have human leukocyte antigen serotype DQ differences, which may lead to beta cell dysfunction rather than insulin resistance. In addition, adiponectin, a cytokine negatively related to obesity and diabetes, does not seem to correlate with insulin resistance in African Americans. Others have suggested differences in body fat distribution among African Americans.19, 21, 24-26
It also currently remains unclear whether diabetes, obesity, and the metabolic syndrome are risk factors for HCC independent of the presence of NAFLD. However, it appears likely that NAFLD usually mediates the relation between metabolic syndrome and HCC based on the high correlation between features of metabolic syndrome and NAFLD. Up to 70% of patients with type II diabetes and up to 90% of obese patients have some degree of fatty liver disease.8 Although NAFLD can occur in the absence of metabolic syndrome, this appears to be relatively uncommon. Marchesini et al assessed the prevalence of the metabolic syndrome in 304 patients who had NAFLD but no overt diabetes, and 67% of those with NAFLD and obesity had metabolic syndrome compared with only 18% of normal-weight patients.27 Similarly, it is likely that the majority of those with metabolic syndrome who develop HCC also have cirrhosis before their diagnosis, although this also is unproven. Case reports have described patients with NASH who developed cirrhosis and then HCC,17, 28 and animal models have demonstrated a clear progression from NASH, to cirrhosis, to cancer.29 Large prospective studies are needed to address both of these issues more definitively.
Metabolic syndrome risk factors also may be modified by other underlying liver diseases with respect to HCC risk. For instance, diabetes appears to be synergistic with both virally mediated HCC and alcohol-related HCC.30 In a large prospective cohort study from Taiwan, obesity led to a 4-fold overall increased risk of HCC in patients with HCV. In patients without underlying viral infection, there was an approximately 2-fold increased risk. However, for those with HBV, BMI was not found to be associated with HCC risk. Diabetes led to a 2-fold to 3-fold increase in the risk of HCC regardless of the underlying viral etiology and also had a synergistic effect with obesity, leading to a >100-fold increased cancer risk.31 It is interesting to note that those investigators demonstrated an inverse correlation between serum levels of HBV DNA and triglyceride levels. They noted that, in culture models, the HBV X protein inhibited the secretion of apolipoprotein B, which is a component of very-low-density lipoprotein.32
Yu et al recently reported on a prospective study of 2903 men (government employees from Taiwan) who were positive for HBV surface antigen. At a mean follow-up of nearly 15 years, the authors observed that the hazard ratio among overweight men for HCC was 1.48 (95% confidence interval [95% CI], 1.04-2.12), whereas the hazards ratio among obese men was 1.96 (95% CI, 0.72-5.38).33 Those authors observed no significant effect of diabetes on the development of HCC after adjusting for quartiles of BMI and other confounders; however, they noted that only 2.5% of their sample reported diabetes at enrollment. Those with higher BMI were significantly more likely to have underlying fatty liver disease (assessed by ultrasound) as well as cirrhosis, leading to the question of whether these results may have been confounded by ascites in the patients with higher BMI. However, as the authors noted, the presence of chronic HBV may provide a synergistic effect with obesity on hepatic lipid accumulation and the development of steatosis.
Thus, cancer risks related to underlying liver disease from either HBV/HCV or alcohol may be worsened by features of metabolic syndrome. In addition, patients with insulin resistance and NASH often have hepatic iron deposition, which is distinct from hemochromatosis.34 Furthermore, those with NASH-related cirrhosis who have more stored iron in their livers (and do not have hemochromatosis) appeared to have a higher risk of cancer in a large retrospective study.35 Given the heterogeneity of underlying risk factors in patients with HCC, more study will be necessary to dissect out the correlations between these risk factors and the influence on HCC risk.
Obesity and Metabolic Syndrome
Metabolic syndrome may be defined as an increased waist circumference or BMI >30 kg/m2 and any 2 of the following: 1) triglycerides >150 mg/dL or treatment for elevated triglycerides; 2) high-density lipoprotein (HDL) cholesterol <40 mg/dL in men or <50 mg/dL in women or treatment for low HDL; 3) systolic blood pressure >130 mm Hg, diastolic blood pressure >85 mm Hg, or treatment for hypertension; and 4) fasting plasma glucose >100 mg/dL or previously diagnosed type II diabetes.36 Obesity, a major component of the metabolic syndrome, also can be defined in several ways. BMI is an individual's weight in kilograms divided by the square of height in meters. Both the World Health Organization and the US Department of Health and Human Services define “overweight” as a BMI between 25 kg/m2 and 29.9 kg/m2, “obese” as a BMI >30 kg/m2, and “normal” as a BMI between 18.5 kg/m2 and 24.9 kg/m2.37, 38 Others have defined important correlates of obesity, such as truncal obesity (measuring either waist circumference or the waist/hip ratio) or visceral fat (often measured by computed tomography or MRI). These anthropomorphic measurements may be more sensitive measures of the sequelae of obesity, such as metabolic syndrome and NASH, compared with BMI.39 These definitions have their own complexity, because different ethnic groups have different amounts of visceral fat accumulation for given amounts of total body fat,40, 41 and women and older individuals also tend to have higher percentages of body fat than men or younger individuals with the same BMI because of differences in body composition.42 Similarly, the normal BMI range for Asians is lower than that for other racial/ethnic groups, with a BMI of 23 kg/m2 considered an “action point” for public health interventions in Asian patients.43, 44
The Centers for Disease Control's National Center for Health Statistics has conducted a series of cross-sectional surveys, known as National Health and Nutrition Surveys (NHANES). Beginning in 1960, these surveys recorded height and weight in nationally representative samples of Americans. During 2003 and 2004, NHANES data indicated that 33% of adults ages 20 years to 74 years were obese, compared with 11% of men and 16% of women in the 1960s, and that much of the increase occurred after 1980.45 There also were differences in age, with older individuals, African Americans, and Hispanic women and children more likely to be obese than the general population.45 Obesity increases the risk for many types of health problems, including diabetes, hypertension, and cardiovascular disease, as well as the overall risk of death.46, 47
Overall mortality clearly is related to obesity, but many have suggested that the relation may be “U-shaped,” with optimal survivals at intermediate BMIs,48, 49 or “J-shaped,” with higher mortality for thin patients than for individuals in the middle BMI range.50 This may be because smokers and individuals with chronic illness tend to be thin, and both tend to have high mortality rates.
Obesity and Cancer Risk
Several studies suggest that obese patients also are at increased risk for several types of cancer, both in the United States and in other countries.51-54 A large meta-analysis published in The Lancet indicated that increased BMI was associated strongly with the risk of esophageal, thyroid, colon, and renal cell carcinomas in men. In women, endometrial, gallbladder, esophageal, and renal cell carcinomas were increased in those with a 5 kg/m2 increase in BMI.54 The Million Woman Study, a large prospective cohort of women in England and Scotland, also reported significant increases in several cancers among women with increasing BMI, including postmenopausal breast cancer, pancreatic and ovarian cancers, and several hematologic malignancies.55
Explanations for these increases in cancer incidence have focused on correlations between metabolic syndrome, adipokines, and hormone levels. For instance, postmenopausal breast cancer and endometrial cancer risk in obese women may be mediated by increased estrogen levels. In addition, the insulin growth factor axis has been implicated in the risk of several types of cancer among both men and women. High levels of peripheral insulin-like growth factor 1 have been associated with an increased risk of developing several types of cancer, including prostate, colon, and breast cancers.56-58 Adipokines related to obesity, such as leptin, also may mediate cancer risk through their effects on angiogenesis.59
Furthermore, increasing evidence suggests that obesity may lead to a state of chronic inflammation. Excess consumption of fatty acids and glucose can lead to the increased expression of several signaling molecules with known importance in carcinogenesis, including nuclear factor κB, epidermal growth factor, and fibroblast growth factor.60, 61
Obesity and Cancer Outcome
Individuals with features of metabolic syndrome, such as obesity, may have worsened outcomes from many different types of cancer, particularly HCC. An article by Calle et al in the New England Journal of Medicine indicated that obesity is associated with significantly increased cancer death rates, particularly from HCC (Fig. 3).52
These findings were not replicated in a study of men who were Korean government employees and teachers published by Park et al. In their report, 14,578 men were followed from 1996 to 2004. Unlike the results reported by Calle et al,52 Park et al reported that a BMI ≥ 25 kg/m2 did not lead to a statistically significant increase in mortality for patients with HCC (hazard ratio, 1.03; 95% CI, 0.92-1.14). This was much less than the cutoff used by Calle et al of ≥40 kg/m2 for the highest BMI, which may explain the discordant findings.62
Reasons for these possibly worsened outcomes remain unclear and may include comorbidities, such as coronary artery disease, and/or under dosing of chemotherapy in some types of cancers in obese individuals.63 Unfavorable tumor characteristics also may be associated with obesity because of delayed screening or biologic characteristics of the tumor itself. However, several studies have suggested that the increased mortality observed with obesity is not related to later diagnosis, because differences often persist after adjusting for disease stage and tumor size.64, 65
Another possible mechanism connecting obesity to worsened clinical outcomes may be related to dysregulated angiogenesis. Adipose tissue induces the expression of vascular endothelial growth factor and other adipokines in both human and animal models.66 In 58 patients with renal cell carcinoma, high serum leptin levels were associated significantly with venous invasion in pathology samples and aggressive clinical features.67 Adipose tissue induces expression of leptin, a hormone that regulates body mass. It has been demonstrated that leptin, in turn, promotes angiogenesis and mediate the progression of NASH to HCC in animal models.68 Leptin also up-regulates the signal transduction pathways involved in cancer progression, such as the Janus kinase/signal transducer and activator of transcription pathway, the protein kinase B or AKT pathway, and the extracellular signal-regulated kinase pathway, in HCC cells.69 These correlations suggest a possible association between the metabolic syndrome and worsened clinical outcomes that may be mediated by adipokines, which lead to increased vascular invasion.
Metabolic Syndrome and HCC: Risk, Outcome, and NASH Management
To our knowledge to date, the mechanism that mediates the interaction between NAFLD, NASH, and HCC has not been elucidated completely. Some of the hypothetical mechanisms are illustrated in Figure 4. For HCC, the relation with obesity appears to be mediated primarily by factors related to metabolic syndrome, NAFLD, and NASH. Insulin modulates intracellular signaling through the tyrosine kinase activity of the insulin receptor. It is believed that defects in these signaling pathways contribute to insulin resistance, which, in turn, can lead to hepatic fat accumulation by lipolysis.14 Then, hepatic fat accumulation can produce inflammatory changes in the liver.14 In particular, free fatty acids may lead to hepatic inflammation through the production of cytokines, such as tumor necrosis factor α (TNF-α). In addition, it is believed that mitochondrial dysfunction also leads to free radical production and oxidative stress, which may provide the “second hit” that allows progression from steatosis to steatohepatitis and cirrhosis.14, 70 Moreover, it has been demonstrated that leptin levels are increased in patients with NASH,71 pointing to a possible role for increased angiogenesis and vascular invasiveness in HCC in the setting of metabolic syndrome.
With respect to obesity as a risk factor for HCC, the exact relation between HCC and risk still is being defined. Nair et al used the United Network of Organ Sharing, a database on all liver transplantations performed in the United States, and demonstrated that obesity was an independent predictor of HCC in patients with alcoholic cirrhosis and cryptogenic cirrhosis but not in patients with cirrhosis of other etiologies.72 A meta-analysis of 11 cohort studies conducted in Europe, the United States, and Asia indicated that those who were overweight had a significantly increased relative risk of developing HCC (1.07; 95% CI, 1.01-1.15), whereas for those who were obese, the relative risk was even higher at 1.85 (95% CI, 1.44-2.37). The analysis excluded studies that included patients with cirrhosis to avoid confounding by ascites.73
Although earlier studies did not demonstrate a clear correlation between diabetes and HCC risk, more recent epidemiologic data suggest that diabetes likely is associated with a 2-fold to 4-fold increased risk of HCC.74-76 It is not known whether insulin resistance causes NASH. Patients with cirrhosis of all types may become insulin resistant within the liver, because insulin is not cleared properly.77, 78 However, it is believed that peripheral insulin resistance is “primary” and leads to hepatic steatosis, which then can contribute to both peripheral and hepatic insulin resistance.79 Steatosis, or fatty liver, also can be observed in patients with HCV infection, and those with both HCV and fatty liver changes have a greater risk of HCC than those with HCV alone.80 Thus, mechanisms of carcinogenesis may relate at least in part to the intermediate steps of fatty change in the liver and insulin resistance rather than the “final outcome” of cirrhosis. Available data suggest that diabetes precedes liver disease and subsequently increases HCC risk.76 By using data from a Veterans Administration population and excluding those with chronic liver disease at baseline, El-Serag et al demonstrated that the incidence of HCC was doubled among patients with diabetes and was higher among those with longer follow-up. Similar findings were reported recently in a Japanese cohort study.76, 81 Those reports strongly suggest that insulin resistance precedes cirrhosis and HCC.
We performed a preliminary retrospective analysis of patients who underwent surgery for HCC and observed a dose-response relation between increasing BMI and the percentage of patients with microvascular invasion in their specimens,82 supporting the idea that the obese milieu may contribute to angiogenesis. Patients with other etiologies of fatty liver disease, as noted earlier, also may be at higher risk of cancer. A study of 99 patients who underwent liver transplantation with underlying HCV indicated that those with steatosis had a higher risk of having underlying HCC.80
Given the likely relations between metabolic syndrome and HCC risk, several strategies have been attempted in the management of NASH (Table 1). The most straightforward are interventions that lead to weight loss. Bariatric surgery has produced clear improvements in liver histology.83 In children, a 2-year lifestyle intervention that included diet and 45 minutes per day of aerobic exercise improved metabolic parameters and liver histology.84
Table 1. Selected Therapeutic Strategies for Nonalcoholic Steatohepatitis
One pharmacologic intervention that has received attention as a potential treatment for NASH is pioglitazone, a thiazolidinedione that improves insulin resistance and has anti-inflammatory effects in patients with type II diabetes. In a proof-of-concept study, 55 patients with impaired glucose tolerance or type II diabetes received either a hypocaloric diet plus pioglitazone (45 mg daily) or a hypocaloric diet plus placebo. Patients who were assigned to the diet plus pioglitazone group had improved glycemic control and also had reduced necroinflammation on pathology after 6 months of treatment.85 Unfortunately, these results reportedly reversed quickly when the medication was withdrawn, and weight gain was a major side effect of the drug.86 Other insulin-sensitizing drugs, such as metformin, have been used, and improvements have been observed in metabolic parameters.87
Because it is believed that oxidative stress mediates the progression of simple steatosis to steatohepatitis, antioxidants also have been tried in patients with fatty liver disease. In the lifestyle intervention in children described earlier, the addition of α-tocepherol (600 IU daily) and ascorbic acid (500 mg daily) did not improve the results compared with placebo (all children received the exercise and lifestyle intervention).84 In adults, a randomized trial of 49 patients who were randomized receive to a combination of α-tocepherol and ascorbic acid or placebo led to improvements in fibrosis after 6 months in the treatment group, but the placebo group unexpectedly did just as well, which rendered the interpretation of results unclear.88, 89 A pilot study of pentoxifylline, a drug that inhibits TNF-α, also led to improvements in biochemical markers, but the drug led to nausea in many individuals.90
The outcomes of patients with metabolic syndrome and HCC have not been well studied to date. For example, little is known regarding how those with underlying NASH respond to systemic treatments for their HCC. Leung et al evaluated 149 patients who received combined cisplatin, interferon-α, doxorubicin, and 5-fluorouracil and reported that the presence of HCV serology was an independent predictor of a worsened response rate. It is difficult to conclude that their results were definitive, because only 4 patients were positive for HCV. Furthermore, because nearly all of the remaining patients had underlying HBV, the authors noted that it was impossible to determine outcomes in those with other etiologies of liver disease. The presence of cirrhosis also led to worsened overall survival in patients who were treated on this regimen, similar to results reported in past trials with single-agent doxorubicin and data from the recently published phase 2 trial of sorafenib in HCC.92-94 Neither the Sorafenib HCC Assessment Randomized Protocol (SHARP) trial nor the recently published Asia-Pacific trial of HCC using sorafenib addressed the outcomes of patients nonviral disease.95, 96 In the SHARP trial, those with vascular invasion demonstrated the suggestion of a better response to sorafenib, although it was not statistically significant. Thus, data indicating a higher degree of vascular invasion in tumors from patients with NASH may have implications for response to antiangiogenic agents such as sorafenib.
Outcomes in obese patients with HCC after surgery may also be worse. One Japanese study demonstrated no effect on overall survival or disease recurrence for patients undergoing an initial resection but a significantly worsened overall survival and disease recurrence was noted in obese patients who underwent repeat hepatectomy.97 Patients with hepatic steatosis may be at increased risk for tumor recurrence after resection.98 Finally, diabetics also may have an increased risk of HCC recurrence. Patients who underwent resection for HCC with hepatitis C and were diabetic were found to have significantly worse survival compared with patients who were not diabetic. In multiple regression analysis, receiving insulin was an independent predictor of recurrence, underscoring the possibility that insulin may have carcinogenic properties in vivo.99
Summary and Future Directions
Obesity and the metabolic syndrome are growing epidemics in the United States and worldwide. These diseases are associated with both an increased risk of and worsened outcomes for many types of cancer. In the liver, inflammatory and angiogenic changes because of underlying insulin resistance and fatty liver disease likely will lead to an increase in the number of patients with HCC in the near future. Much work is needed to define more clearly the risks of developing HCC in individuals with underlying metabolic syndrome, the best methods for screening those at risk, and, ultimately, the best treatments targeting the underlying mechanisms of pathogenesis.
Conflict of Interest Disclosures
Supported in part by a K12 award from the National Institutes of Health (KL2 RR024157-03) and by the Steven J. Levinson Medical Research Foundation.