PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cancer. Author manuscript; available in PMC 2017 August 8.
Published in final edited form as:
PMCID: PMC5548177
NIHMSID: NIHMS886809

Population attributable fractions of risk factors for hepatocellular carcinoma in the United States

Abstract

Objectives

Hepatocellular carcinoma (HCC) incidence has been increasing in the United States for several decades. As the incidence of hepatitis C virus (HCV) infection declines and the prevalence of metabolic disorders rises, the proportion of HCC attributable to various risk factors may be changing.

Methods

Data from the Surveillance, Epidemiology, and End Results (SEER)-Medicare linkage were used to calculate population attributable fractions (PAFs) for each risk factor over time. HCC cases (n=10,708) diagnosed during the years 2000–2011 were compared to a 5% random sample of cancer-free persons (n=332,107) residing in the SEER areas. Adjusted odds ratios (ORs) and PAFs were calculated for hepatitis C virus (HCV), hepatitis B virus (HBV), metabolic disorders, alcohol-related disorders, smoking, and genetic disorders.

Results

Overall, the PAF was greatest for metabolic disorders (32.0%), followed by HCV (20.5%), alcohol (13.4%), smoking (9.0%), HBV (4.3%) and genetic disorders (1.5%). The PAF for all factors combined was 59.5%. PAFs differed by race/ethnicity and gender. Metabolic disorders had the largest PAF among Hispanics (39.3%, CI=31.9–46.7%) and whites (34.8%, CI=33.1–36.5%), while HCV had the largest PAF among blacks (36.1%, CI=31.8–40.4%) and Asians (29.7%, CI=25.9–33.4%). Between 2000 and 2011, the PAF of metabolic disorders increased from 25.8% (CI=22.8–28.9%) to 36.0% (CI=33.6–38.5%). In contrast, the PAFs of alcohol-related disorders and HCV remained stable.

Conclusions

Among U.S. Medicare recipients, metabolic disorders contribute more to the burden of HCC than any other risk factor and the fraction of HCC due to metabolic disorders has increased in the last decade.

Keywords: hepatocellular carcinoma, metabolic disorders, hepatitis C virus, hepatitis B virus, population attributable fractions

INTRODUCTION

Primary liver cancer is the sixth most commonly occurring malignancy worldwide and the second leading cause of cancer mortality.1 In the United States, the incidence of liver cancer has been increasing since 1975,2 and liver cancer is projected to be among the top three causes of cancer mortality by 2030.3 The predominant histologic form of primary liver cancer is hepatocellular carcinoma (HCC). In contrast to many other types of cancer, HCC frequently occurs among persons with known risk factors who have underlying liver disease.

Risk factors for HCC in the U.S. include chronic infection with either hepatitis C virus (HCV) 4 or hepatitis B virus (HBV),5 excessive alcohol consumption,5 cigarette smoking and rare genetic disorders (porphyrias,6 hemochromatosis,7 Wilson’s disease,8 alpha-1 antitrypsin deficiency,9 glycogen storage diseases 9, 10). Recent data suggest that a constellation of metabolic disorders, which include diabetes,11 obesity,12, 13 impaired glucose tolerance,14 metabolic syndrome,15 and non-alcoholic fatty liver disease (NAFLD),16, 17 are also important HCC risk factors.

The extent to which each risk factor contributes to the overall HCC burden on a population level can be determined by calculation of the population attributable fraction (PAF), which is an important measure for cancer control policy development. Previous studies have examined the PAFs of HCC risk factors in the United States; 18 but relative contributions of HCC risk factors to the burden of disease may have changed over time. Recent reports have shown that the prevalence of metabolic disorders is rising: 35.4% of U.S. residents aged 60 years and older are obese,19 26.9% of persons 65 years and older have diabetes20 and approximately one-fifth of the adult population have metabolic syndrome.21 In contrast, the prevalence of HCV infection has declined due to the elimination of HCV from the nation’s blood supply in the early 1990s.22 As a result of these concurring trends PAFs may have changed in recent years. To examine this hypothesis, we estimated PAFs of the HCC risk factors for the period 2000–2011, using the SEER-Medicare Linked Database.23, 24

METHODS

Data Source

The SEER-Medicare Linked Database is a linkage of cancer registry data from the National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) program with the Medicare medical claims database. The SEER 18 database excluding the Alaska Native Registry covers approximately 30% of the U.S. population.25 Medicare is the national health insurance program administered by the U.S. government that covers approximately 97% of persons aged 65 years and above.26 Almost all Medicare beneficiaries are enrolled in Part A, which covers inpatient medical services, and 96% of Part A beneficiaries are also enrolled in Part B, which covers outpatient services. The SEER-Medicare linkage matches approximately 93% of Medicare aged individuals in SEER cancer registries to the Medicare medical claim file.24, 27 To identify risk factors, International Classification of Diseases, version 9 (ICD-9) codes were used to capture medical conditions and key health risk behaviors (Table 1). Because obesity, diabetes, impaired glucose metabolism (IGM), metabolic syndrome and non-alcoholic fatty liver disease (NAFLD) are highly correlated disorders, these conditions were combined into one group, referred to as ‘metabolic disorders.’ Rare genetic disorders (porphyrias, hemochromatosis, Wilson, alpha-1 antitrypsin deficiency and glycogen storage diseases) known to be associated with HCC were combined into a single group, referred to as ‘genetic disorders.’ Because the SEER-Medicare database contains limited information on alcohol consumption, a collection of alcohol-related medical conditions in addition to reported history of alcohol abuse were used as the proxy variable for heavy alcohol consumption under the variable heading ‘alcohol.’28 Similarly, the combination of chronic obstructive pulmonary disease (COPD)29 and tobacco use were used as a proxy for heavy smoking under the variable heading ‘smoking’. Thus, the following factors were analyzed for odds ratios (ORs) and PAFs: metabolic disorders, HCV, HBV, genetic disorders, alcohol, and tobacco.

Table 1
International Classification of Diseases, Ninth Edition Codes for Hepatocellular Carcinoma Risk Factors

Cases and Controls

HCC cases were identified using ICD for Oncology, third edition [ICD-O-3] topography code C22 and morphology codes 8170–8175. The case inclusion criteria included: diagnosis year between 2000 and 2011; age 68–100 years old; enrolled in Medicare Parts A and B continuously during the 36 months before diagnosis. A minimum age of 68 years was required to allow a 3-year Medicare history for risk factor identification. A 3-year minimum was selected in order to allow a sufficiently long interval for exposures to be identified but not so long as to limit the number of cases which could be included. Exclusion criteria included: a history of enrollment in an HMO at any time in the 36 months before diagnosis; and HCC diagnosis noted solely on autopsy or death certificate. A 5% random sample of cancer-free Medicare recipients residing in the SEER areas of the cases was used as the comparison population. Using a random number generator, index dates assigned to eligible controls based on birth year were used for PAF calculations in each of 4-year diagnosis intervals. Control selection was based on the same eligibility criteria as was case selection. The race/ethnicity variable was that used by Medicare.

Statistical Analysis

Population attributable fractions were calculated using the formula described in Bruzzi et al.30 Multivariable logistic regression analyses that adjusted for the state buy-in-status (as an indication of lower socioeconomic status), age, gender and race/ethnicity and the other five risk factors were used to calculate ORs, PAFs and corresponding 95% confidence intervals (CI) in the overall study population; see SAS code in.31 ORs and PAFs were also calculated by gender, race/ethnicity and diagnosis time period (2000–2003, 2004–2007, 2008–2011). Wald Chi-square tests were calculated as global tests of significance difference in ORs by year of diagnosis. Data analyses were conducted using SAS, v.9.4 (SAS, Cary, NC).

RESULTS

The analysis included 10,708 persons with HCC and 332,107 cancer-free persons. Table 2 shows that the mean age of cases was 76.9 years compared to 76.6 among controls. The majority of cases were male (66.3%), while the majority of the controls were female (61.4%). More cases were diagnosed in the most recent time period 2008–2011 (38.3%) than in the prior two. Most cases (70.9%) and controls (83.5%) were white.

Table 2
Distribution of Demographic and Risk Factors in Patients With Hepatocellular Carcinoma (Cases) and Controls, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

Among cases, 50.1% had a history of metabolic disorders, compared to 24.8% of controls. There was evidence of HCV infection among 20.8% of cases and 0.4% of controls. Similarly, 4.5% of cases had a history of HBV, compared to 0.1% of controls. An alcohol-related condition was documented in 15.5% of cases and 1.6% of controls. A history of smoking was identified among 33.1% of cases and 18.7% of controls. Genetic disorders were found among 1.7% of cases and 0.2% of controls.

The ORs associated with each of the six risk factors, adjusted for sex, age at diagnosis, race/ethnicity, state buy-in status and other risk factors are shown in Table 3. The risk of developing HCC was highest in association with HCV (OR=59.9, 95%CI=55.1–65.1), followed by HBV (OR=21.6, 95%CI=17.9–26.0), alcohol (OR=7.3, 95%CI=6.8–7.8), genetic disorders (OR=7.3, 95%CI=6.0–8.9), metabolic disorders (OR=2.8, 95%CI=2.7–2.9) and finally, smoking (OR=1.4, 95%CI=1.3–1.4). The OR associated with HCV was approximately twice as high among women (OR=81.8 95%CI=72.9–91.9) than men (OR=42.1, 95%CI=37.6 – 47.3). In contrast, the OR associated with HBV was more than twice as high among men (OR=28.2, 95%CI=22.3–35.6) as women (OR=13.2, 95%CI=9.5–18.3) and the OR associated with genetic disorders was three times higher among men (OR=9.7, 95%, 95%CI=7.7–12.3) than women (OR=3.2, 95%CI=2.0–5.1). The ORs associated with metabolic conditions, alcohol and smoking did not vary greatly by gender.

Table 3
Odds Ratios and 95% Confidence Intervals for Hepatocellular Carcinoma Risk Factors, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

The ethnic/racial examination of risk factors is shown in Table 4. As with the gender-specific analyses, risks were ordered in roughly the same way as in the overall analysis. In all four groups, the ORs were highest in association with HCV and lowest in association with smoking. The ORs associated with HBV were the second highest in all groups except Hispanics and varied widely, with Asians having the greatest risk (OR=31.2, 95%CI=23.2–42.2) and Hispanics having the lowest risk, which did not attain statistical significance (OR=2.5, 95%CI=0.4–15.5). The ORs associated with alcohol-related conditions also varied widely, with Hispanics having the highest risk (OR=9.5, 95%CI=6.8–13.2) and blacks having the lowest risk (OR=3.6, 95%CI=2.7–4.7). Increased risk associated with metabolic disorders was found in all racial/ethnic groups with the greatest ORs found among whites (OR=3.1, 95%CI 2.9–3.3) and Hispanics (OR=2.8, 95%CI=2.2–3.5). The OR associated with genetic conditions was statically significant only among whites (OR=8.2, 95%CI=6.7–10.2, p<0.001). The HCC risk associated with smoking was similar among whites, blacks and Asians. Among Hispanics, there was no significant association between smoking and risk.

Table 4
Odds Ratios and 95% Confidence Intervals for Hepatocellular Carcinoma Risk Factors by Race/Ethnicity, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

The risks associated with each factor by period of diagnosis (2000–2003, 2004–2007, 2008–2011) are shown in Table 5. Global tests of the odds ratios associated with HCV (p<0.0001), genetic conditions (p=0.002), alcohol (p=0.04), and smoking (p<0.0001) found significant differences by time interval. In contrast, global tests of the odds ratios associated with HBV (p=0.26) and metabolic disorders (p=0.17) found no significant differences by time interval.

Table 5
Odd ratios and 95% Confidence Intervals for Hepatocellular Carcinoma Risk Factors by Years of Diagnosis, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

The overall and gender-specific PAFs of each factor are shown in Table 6. Overall, the greatest PAF was for metabolic disorders (PAF=32.0%, 95%CI=30.5–33.5), followed by HCV (PAF=20.5%, 95% CI=19.0–22.0), alcohol (PAF=13.4%, 95% CI=11.7–15.0), smoking (PAF=9.0% for 95% CI= 6.9–11.1), HBV (PAF= 4.3%, 95%CI 2.5–6.1) and genetic disorders (PAF=1.5%, 95% CI=−0.4–3.3). The PAF of all six factors combined was 59.5% (95%, CI=58.5–60.5). Metabolic disorders were associated with the greatest PAFs among both genders (men: PAF 32.5%, 95%CI=30.7–34.4; women: PAF 30.8%, 95%CI=28.1–33.4). The PAFs of HCV and alcohol-related disorders differed by gender. The PAF of HCV was higher among women (PAF=26.2, 95%CI =23.8–28.6) than men (PAF=17.5, 95%CI=15.7–19.4), while the PAF of alcohol-related disorders was higher among men (PAF=17.0%; 95%CI=15.1–18.9) than women (PAF=6.2%; 95%CI=3.2–9.3). There was no significant difference in the PAFs by gender for smoking, HBV or genetic disorders.

Table 6
Overall and Sex-Specific Hepatocellular Carcinoma Population Attributable Fractions, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

Table 7 presents PAFs by racial/ethnic group. While metabolic disorders had the greatest PAF among Hispanics (39.3%, 95%CI=31.9–46.7) and whites (34.8%, 95%CI= 33.1–36.5), HCV had the greatest PAF among blacks (36.1%, 95%CI=31.8–40.4) and Asians (29.7%, 95%CI=25.9–33.4). The third greatest contributor to HCC among all groups except Asians was alcohol-related disorders, while the third greatest contributor among Asians was HBV (17.8%, 95%CI=13.4–22.2). The PAF of smoking was similar in whites (10.5%, 95%CI=8.0–13.0) and blacks (10.6%, 95%CI=2.8–18.5), but was lower among Asians (5.2%, 95%CI=−0.7–11.1) and Hispanics (1.1, 95%CI=−11.1–13.2).

Table 7
Race and Ethnicity-Specific Hepatocellular Carcinoma Population Attributable Fractions, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

Between 2000–2003 and 2008–2011, the PAF of all risk factors combined increased from 52.2 to 63.7% (Table 8). Increases in PAF were most evident in association with metabolic disorders (25.8% to 36.0%) and smoking (5.1% to 12.2%). The PAF associated with HCV modestly increased until 2004–2007, but was then stable. In contrast, the PAFs of HBV (3.9% to 4.7%), alcohol (12.3% to 13.8%) and genetic disorders (1.9% to 1.1%) were fairly stable over time.

Table 8
Hepatocellular Carcinoma Population Attributable Fractions by Years of Diagnosis, Surveillance, Epidemiology, and End Results-Medicare, 2000–2011

DISCUSSION

The current U.S. study found that 59.5% of HCC diagnosed between 2000 and 2011 can be attributed to six known HCC risk factors. One factor alone, metabolic disorders, accounted for 32% of the total HCC burden, increasing from 26% to 36% over the period of study. In contrast, the attributable fraction associated with HCV remained stable throughout at approximately 20%. The attributable fraction of the various factors differed by racial/ethnic group and by gender. While metabolic disorders were the greatest contributor to HCC among whites and Hispanics, HCV was the greatest contributor among blacks and Asians. By gender, the contribution of HCV was higher among women than men, but the contribution of alcohol was higher among men than women.

The 32% PAF associated with metabolic disorders in the present study is similar to that reported for obesity/diabetes (PAF=37%) in a prior SEER-Medicare study that covered the years 1973–2007.18 Other studies of HCC attributable risks5, 3235 have not reported the contribution of more than one metabolic disorder at a time. Examinations of obesity alone have reported PAFs of 7.0% in Italy 35 and 16.1% in a large European cohort from several countries.5 Examinations of diabetes alone have reported PAFs ranging from 8% to 14% in Italy,32, 35 20% in Texas,33 and between 6% and 27% in Hawaii, depending on racial/ethnic group.34 Comparison of the current results with those of other studies, however, must await the reporting of PAF results for metabolic disorders as a group.

In the present study, HCV had the second highest PAF (20.5%) of any factor after metabolic disorders. This PAF is consistent with most prior estimates, both from the U.S.18, 33 and Europe,5, 36 but notably lower than a prior estimate from Italy (PAF=65%).35 Females had a higher PAF (26.2%) than males (17.5%). Although the reason for this is unclear, it may be related to higher HCV-related mortality among men at younger ages. Among the racial/ethnic groups in the current study, blacks had the highest PAF associated with HCV (36.1%), while Asians had the second highest (29.7%). Among black persons, the high burden of HCV-related HCC is consistent with race-specific HCV prevalence estimates from U.S. surveys.37 The large fraction of HCV-related HCC cases among Asians in the current analysis may be due to the high percentage of persons of Japanese ancestry in SEER catchment areas (i.e., Hawaii and California). Although HBV is the dominant HCC-related virus in most Asian countries, an HCV epidemic after World War II in Japan resulted in an epidemic of HCV-related HCC in the ensuing decades.38

The PAF associated with HBV (4.3%) was similar to the report of the prior SEER-Medicare study, but lower than the PAFs reported from Europe (7.9–16.0%) 5, 35, 36 and from a Texas case-control study (16.0%).33 The differences in PAF between the U.S. and Europe are not unexpected as the U.S. is a low-endemicity region for HBV, while some European countries are of intermediate endemicity.39 The difference between the current study and the U.S. case-control study may be related to the difference in ascertainment methods of the populations. The case-control study used hospital controls with cancer, thus their exposures may have differed from those of the general population.33 In the current study, the highest PAF for HBV was observed among Asians (17.8%), with PAFs of less than 3% seen among all other groups. As chronic HBV infection is more common among Asians than among other racial/ethnic groups, particularly Asians not born in the U.S., the higher PAF among Asians is in line with expectations.40 The prevalence of HBV is declining in most Asian countries, however, because HBV vaccination of neonates is now widespread.41, 42 As a result, the PAF associated with HBV will almost certainly decline in coming generations.

The PAF associated with smoking (9%) was similar to the PAF of 12% reported from Italy,35 but strikingly lower than two estimates from a European multi-country cohort study (34.9–47.6%).5, 43 The large difference in PAF associated with smoking may be due to the unusually low prevalence of smokers in the European cohort comparison group.35, 44 In the current study, the PAF for smoking increased over time, although not significantly so, from 5.1% to 12.2%. Whether this indicates a real increase in the contribution of smoking to HCC risk, however, is not clear. The variable used in the current study was a combination of codes for COPD and tobacco use. If physicians were more likely over time to indicate that HCC cases were smokers, a larger proportion of HCC cases would appear to be attributable to smoking.

The PAF due to alcohol (13.4%) in the current study was similar to that of a large European cohort (10.2%), but lower than PAFs of 18–33% reported by other European studies and by one hospital-based U.S. case-control study (PAF=32%).33 As the alcohol variable in the current study was based on codes for alcohol abuse and alcohol-related conditions, it is likely that only persons who consumed large quantities of alcohol were identified. Thus the PAF of the current study may only reflect the contribution of excessive alcohol consumption to HCC risk. The difference between the current study and the U.S. study could have been due to the use of hospitalized controls with cancer in the case-control study.33

The small fraction of HCC cases attributable to genetic disorders (PAF=1.5%) in the current study was consistent with the prior SEER-Medicare analysis.18 As the HCC-related genetic disorders are more common among persons of European ancestry, the PAF among whites (1.9%) was considerably higher than the PAFs among other racial/ethnic groups. Other studies have yet to report PAFs due to genetic disorders, although the rarity of the disorders among all populations indicates that the PAFs will be small.

Our findings challenge assumptions that HCV is the primary factor responsible for the rising incidence of HCC in the U.S.45 In persons of age 68 years and greater, metabolic disease was responsible for an increasing and greater fraction of HCC cases than was HCV infection, which was stable over time. These PAFs could change in the future, however, as the individuals most likely to be infected with HCV are members of the 1945–1965 birth cohort,46 which was still too young to be included in the current analysis. Although some persons in this birth cohort would have become eligible for Medicare in 2010, they would not have reached the minimum age (68 years) for study inclusion. As members of the 1945–1965 birth cohort age, however, a larger proportion of cases could be attributable to HCV if the use of effective anti-HCV drugs does not become widespread.

The current study had strengths and limitations. A major strength was that the SEER-18 registries included over 30% of the population, allowing robust analyses within population strata and across time periods. Furthermore the study population is comprised of members of the age group with the highest incidence rates of HCC.47 In addition, cancer case ascertainment in SEER areas is estimated to be 98% 24 and Medicare covers up to 97% of persons aged 65 years old and older.26 Limitations include that the results of the current analysis may not be generalizable to persons younger than 68 years of age. The identification of risk factors using medical claims data also has a potential for underreporting of behavioral risk factors such as smoking and alcohol 48 and underreporting of conditions that are only ascertained based on medical indication (i.e., HBV and HCV). Obesity is likely underreported as well.48

In conclusion, the finding that metabolic disorders have the highest PAF of all HCC risk factors in the U.S., and that their PAF has been increasing, suggests that these conditions should receive more attention as modifiable risk factors for HCC.

Acknowledgments

Research supported by: The Intramural Research Program of the NIH, NCI.

Andrew J. Muir, Department of Medicine, Duke Clinical Research Institute, Duke University; Huiman Xie Barnhart, Department of Biostatistics and Bioinformatics, Duke Clinical Research Institute, Duke University

Footnotes

Conflicts of interest: All authors of the manuscript declare that they have no conflicts of interest to report.

References

1. Parkin DM, Bray F, Ferlay J, Pisani P. Estimating the world cancer burden: Globocan 2000. Int J Cancer. 2001;94:153–156. [PubMed]
2. Altekruse SF, McGlynn KA, Reichman ME. Hepatocellular carcinoma incidence, mortality, and survival trends in the United States from 1975 to 2005. J Clin Oncol. 2009;27:1485–1491. [PMC free article] [PubMed]
3. Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian LM. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res. 2014;74:2913–2921. [PubMed]
4. Bruno S, Silini E, Crosignani A, et al. Hepatitis C virus genotypes and risk of hepatocellular carcinoma in cirrhosis: a prospective study. Hepatology. 1997;25:754–758. [PubMed]
5. Trichopoulos D, Bamia C, Lagiou P, et al. Hepatocellular carcinoma risk factors and disease burden in a European cohort: a nested case-control study. J Natl Cancer Inst. 2011;103:1686–1695. [PMC free article] [PubMed]
6. Lang E, Schafer M, Schwender H, Neumann NJ, Frank J. Occurrence of Malignant Tumours in the Acute Hepatic Porphyrias. JIMD Rep. 2015 [PMC free article] [PubMed]
7. Fracanzani AL, Conte D, Fraquelli M, et al. Increased cancer risk in a cohort of 230 patients with hereditary hemochromatosis in comparison to matched control patients with non-iron-related chronic liver disease. Hepatology. 2001;33:647–651. [PubMed]
8. Walshe JM, Waldenstrom E, Sams V, Nordlinder H, Westermark K. Abdominal malignancies in patients with Wilson’s disease. Qjm. 2003;96:657–662. [PubMed]
9. Hamed MA, Ali SA. Non-viral factors contributing to hepatocellular carcinoma. World J Hepatol. 2013;5:311–322. [PMC free article] [PubMed]
10. Dragani TA. Risk of HCC: genetic heterogeneity and complex genetics. J Hepatol. 2010;52:252–257. [PubMed]
11. Polesel J, Zucchetto A, Montella M, et al. The impact of obesity and diabetes mellitus on the risk of hepatocellular carcinoma. Ann Oncol. 2009;20:353–357. [PubMed]
12. Regimbeau JM, Colombat M, Mognol P, et al. Obesity and diabetes as a risk factor for hepatocellular carcinoma. Liver Transpl. 2004;10:S69–73. [PubMed]
13. Karagozian R, Derdak Z, Baffy G. Obesity-associated mechanisms of hepatocarcinogenesis. Metabolism. 2014;63:607–617. [PubMed]
14. Khan MM, Saito S, Takagi S, et al. Relationship between hepatocellular carcinoma and impaired glucose tolerance among Japanese. Hepatogastroenterology. 2006;53:742–746. [PubMed]
15. Welzel TM, Graubard BI, Zeuzem S, El-Serag HB, Davila JA, McGlynn KA. Metabolic syndrome increases the risk of primary liver cancer in the United States: a study in the SEER-Medicare database. Hepatology. 2011;54:463–471. [PMC free article] [PubMed]
16. Baffy G, Brunt EM, Caldwell SH. Hepatocellular carcinoma in non-alcoholic fatty liver disease: an emerging menace. J Hepatol. 2012;56:1384–1391. [PubMed]
17. Ertle J, Dechene A, Sowa JP, et al. Non-alcoholic fatty liver disease progresses to hepatocellular carcinoma in the absence of apparent cirrhosis. Int J Cancer. 2011;128:2436–2443. [PubMed]
18. Welzel TM, Graubard BI, Quraishi S, et al. Population-attributable fractions of risk factors for hepatocellular carcinoma in the United States. Am J Gastroenterol. 2013;108:1314–1321. [PMC free article] [PubMed]
19. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, 2011–2012. Jama. 2014;311:806–814. [PMC free article] [PubMed]
20. National Center for Chronic Disease Prevention and Health Promotion DoDT. [accessed 28 June 2015];Diabetes in Older Adults. Available from URL: http://www.cdc.gov/diabetes/risk/age/olderadults.html.
21. Beltran-Sanchez H, Harhay MO, Harhay MM, McElligott S. Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999–2010. J Am Coll Cardiol. 2013;62:697–703. [PMC free article] [PubMed]
22. Denniston MM, Jiles RB, Drobeniuc J, et al. Chronic hepatitis C virus infection in the United States, National Health and Nutrition Examination Survey 2003 to 2010. Ann Intern Med. 2014;160:293–300. [PMC free article] [PubMed]
23. [accessed June 28, 2015];SEER-Medicare Linked Database. Available from URL: http://healthcaredelivery.cancer.gov/seermedicare/
24. Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care. 2002;40:Iv-3–18. [PubMed]
25. Overview of the SEER Program. Available from URL: http://seer.cancer.gov/about/overview.html.
26. SEER-Medicare: Medicare Enrollment & Claims Data. Available from URL: http://healthcaredelivery.cancer.gov/seermedicare/medicare/
27. SEER-Medicare: How the SEER & Medicare Data are Linked. Available from URL: http://healthcaredelivery.cancer.gov/seermedicare/overview/linked.html.
28. Efird LM, Miller DR, Ash AS, et al. Identifying the risks of anticoagulation in patients with substance abuse. J Gen Intern Med. 2013;28:1333–1339. [PMC free article] [PubMed]
29. Mannino DM, Buist AS. Global burden of COPD: risk factors, prevalence, and future trends. Lancet. 2007;370:765–773. [PubMed]
30. Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C. Estimating the population attributable risk for multiple risk factors using case-control data. Am J Epidemiol. 1985;122:904–914. [PubMed]
31. Graubard BI, Fears TR. Standard errors for attributable risk for simple and complex sample designs. Biometrics. 2005;61:847–855. [PubMed]
32. Braga C, La Vecchia C, Negri E, Franceschi S. Attributable risks for hepatocellular carcinoma in northern Italy. Eur J Cancer. 1997;33:629–634. [PubMed]
33. Hassan MM, Hwang LY, Hatten CJ, et al. Risk factors for hepatocellular carcinoma: synergism of alcohol with viral hepatitis and diabetes mellitus. Hepatology. 2002;36:1206–1213. [PubMed]
34. Setiawan VW, Hernandez BY, Lu SC, et al. Diabetes and racial/ethnic differences in hepatocellular carcinoma risk: the multiethnic cohort. J Natl Cancer Inst. 2014:106. [PMC free article] [PubMed]
35. Polesel J, Montella M, Dal Maso L, Crispo A, Serraino D, Talamini R. Re: hepatocellular carcinoma risk factors and disease burden in a european cohort: a nested case-control study. J Natl Cancer Inst. 2012;104:1681–1683. author reply 1683–1684. [PubMed]
36. Donato F, Gelatti U, Limina RM, Fattovich G. Southern Europe as an example of interaction between various environmental factors: a systematic review of the epidemiologic evidence. Oncogene. 2006;25:3756–3770. [PubMed]
37. McQuillan GM, Kruszon-Moran D, Kottiri BJ, Curtin LR, Lucas JW, Kington RS. Racial and ethnic differences in the seroprevalence of 6 infectious diseases in the United States: data from NHANES III, 1988–1994. Am J Public Health. 2004;94:1952–1958. [PubMed]
38. Chung H, Ueda T, Kudo M. Changing trends in hepatitis C infection over the past 50 years in Japan. Intervirology. 2010;53:39–43. [PubMed]
39. Ott JJ, Stevens GA, Groeger J, Wiersma ST. Global epidemiology of hepatitis B virus infection: new estimates of age-specific HBsAg seroprevalence and endemicity. Vaccine. 2012;30:2212–2219. [PubMed]
40. Screening for Chronic Hepatitis B Among Asian/Pacific Islander Populations - New York City 2005. MMWR. 2006;55(18):505–9. [PubMed]
41. [accessed June 5, 2015];Immunization coverage. Available from URL: http://www.who.int/mediacentre/factsheets/fs378/en/
42. Wade N. Anabolic Steriods. Science. 1972;176:1399–1403. [PubMed]
43. Agudo A, Bonet C, Travier N, et al. Impact of cigarette smoking on cancer risk in the European prospective investigation into cancer and nutrition study. J Clin Oncol. 2012;30:4550–4557. [PubMed]
44. Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980–2012. Jama. 2014;311:183–192. [PubMed]
45. Kohler BA, Sherman RL, Howlader N, et al. Annual report to the nation on the status of cancer, 1975–2011, featuring incidence of breast cancer subtypes by race/ethnicity, poverty, and state. J Natl Cancer Inst. 2015:107. [PMC free article] [PubMed]
46. Yuan JM, Ross RK, Stanczyk FZ, et al. A cohort study of serum testosterone and hepatocellular carcinoma in Shanghai, China. Int J Cancer. 1995;63:491–493. [PubMed]
47. Altekruse SF, Henley SJ, Cucinelli JE, McGlynn KA. Changing hepatocellular carcinoma incidence and liver cancer mortality rates in the United States. Am J Gastroenterol. 2014;109:542–553. [PMC free article] [PubMed]
48. Measures that are Limited or not Available in the Data. Available from URL: http://healthcaredelivery.cancer.gov/seermedicare/considerations/measures.html#2.