Diabetes, Liver Cancer, and Cirrhosis: What Next?

Abstract

Potential conflict of interest: Dr. Yki‐Järvinen consults and received grants from Novo Nordisk. She consults for MSD and Lilly. See article on page 1308 Large prospective studies in the United States1 and Europe2 have established that the incidence and risk of mortality from liver cancer and cirrhosis are approximately 2‐fold higher in patients with type 2 diabetes as compared with nondiabetic subjects. This is true for both men and women and the risks remain significant even after adjustment for age, obesity, and alcohol intake. Less is known of China, where the speed of “coca‐colonization” and its metabolic consequences have outrun the rest of the world. China had 114 million patients with diabetes in 2017, which was more than in any other country.3 The burden of hepatitis B virus (HBV) infection in China is also greater than anywhere else (75 million people in 2016), although rapidly decreasing as a result of a successful HBV vaccination program (http://www.wpro.who.int/hepatitis/data/hepatitis_data_statistics/en/#). In this issue of Hepatology, Pang et al. provide the most comprehensive analysis of the association between diabetes, plasma glucose, and incidence of fatty liver, cirrhosis, and liver cancer in China.4 The prospective China Kadoorie Biobank study recruited 500,000 people from five urban and five rural geographic locations during 2004 to 2008. The estimated population response rate was 30%. Data collection included questionnaires as well as on‐site measurements of body weight and height, plasma glucose, and hepatitis B surface antigen (HBsAg). The incidence of liver diseases and liver cancer was analyzed using disease registries and national health insurance databases during a median follow‐up of 10 years, by which time 8% of the participants had died and 1% were lost to follow‐up. Individuals with prior history of cancer or cirrhosis or hepatitis at baseline were excluded. Diabetes was associated with increased risk of liver disease with adjusted hazard ratios of 1.49 for liver cancer, 1.81 for cirrhosis, 1.76 for hospitalized NAFLD, and 2.26 for hospitalized alcoholic liver disease (ALD). All of these diseases were linearly correlated with blood glucose, even within the nondiabetic range. Adjustment for body mass index attenuated the association for NAFLD but not for other liver diseases. Interestingly, diabetes also increased the risk of viral (HBV) cirrhosis 1.82‐fold. Of all the cases with cirrhosis in patients with diabetes (n = 224), 35% were HBsAg positive. Among the hospitalized patients with diabetes, NAFLD was approximately 6 times as common as ALD. The increased risk of both hospitalized ALD and viral cirrhosis in persons with diabetes as compared to people without diabetes is interesting. The data in ALD are in keeping with recent data showing a 6.79‐fold increase in severe liver events in persons with diabetes as compared with people without diabetes among alcohol risk users.5 Although classification of the etiology of liver disease is often dichotomous (nonalcoholic versus alcoholic, HBsAg positive versus negative), in real life the patient might have double or triple trouble. Neither excess alcohol consumption nor HBsAg positivity excludes harmful hepatic consequences of metabolic factors in a patient with diabetes. Thus, NAFLD may coexist in patients with ALD and HBV and facilitate the development of cirrhosis and liver cancer as it does in patients lacking the latter two conditions (Fig. 1).Figure 1: Schematic diagram of risk factors for cirrhosis and liver cancer based on data generated in U.S.,1 European,2 and Chinese4 cohorts. Diabetes increases the risk of cirrhosis and liver cancer because of NAFLD. If diabetes coexists with ALD or HBV, the risk of cirrhosis and liver cancer is doubled (denoted as “double trouble”) compared with nondiabetic subjects with ALD and HBV. The incidence numbers for liver cirrhosis and cancer are based on the study by Pang et al.4 In addition, common genetic risk factors, such as the PNPLA3 I148M gene variant,10 modulate the risk of severe liver disease. Abbreviation: DM, diabetes mellitus.The study by Pang et al. provides some limited clues as to why diabetes increases the risk of liver disease even independent of obesity. The positive correlation between glucose concentrations and liver disease could reflect the combined effects of insulin resistance and relative insulin deficiency, which are both are hallmarks of type 2 diabetes.6 Both factors increase delivery of free fatty acids, the main substrate for synthesis of intrahepatocellular triglycerides, to the liver. In addition, glucose can be converted to saturated fat in the liver through de novo lipogenesis, which is the relatively most increased pathway in NAFLD. How significant is the absolute increase in risk of death due to liver cirrhosis and cancer in diabetes? Only approximately 1% of all noncancer deaths are attributable to liver cirrhosis in diabetes in both men and women.7 Regarding liver cancer, according to a prospective cohort of 1 million U.S. adults followed for 26 years from 1982 until 2008, 16% of men and 14% of women with diabetes died of cancer. Liver cancer accounted for approximately 4% of cancer deaths in men and 2% in women. The absolute contribution of liver cancer to total mortality is therefore low, although rapidly increasing. A recent study estimated the incidence of liver cancer attributable to diabetes and obesity to be 47% higher in 2025 than 2002.8 The doubling in the relative risk of liver cancer by diabetes is more than that of any other cancer. However, the absolute increase in the number of liver cancer deaths attributable to diabetes is comparable to that in colon and pancreatic cancer in men with diabetes, and to that in colon, pancreatic, breast, and endometrial cancer in women with diabetes.7 Screening for merely liver cancer does not seem justified in light of these data unless better tools to identify individuals at risk can be developed. A limitation of the study of Pang et al. is the lack of data on family history and established genetic markers of risk of liver cirrhosis and cancer. The risk of NAFLD cirrhosis was recently shown to be 16‐fold higher in first‐degree relatives of patients with NAFLD cirrhosis.9 Among the genetic markers, the PNPLA3 I148M gene variant has been consistently shown to increase the risk of cirrhosis. According to a meta‐analysis, carriers with this gene variant had a 1.67‐fold higher risk of NASH and alcohol‐related cirrhosis than noncarriers.10 The PNPLA3 I148M gene variant also predisposes to steatosis in HBV. Studies estimating the risk of severe liver disease based on family history and genetic risk markers in addition to obesity and diabetes status would be interesting and potentially important as they might provide simple means of identifying individuals at particularly high risk. In conclusion, epidemiological data from large cohorts are available from the United States, Europe, and now also from China, which show diabetes to approximately double the risk of cirrhosis and liver cancer. The data from China convincingly document that these increased risks are not restricted to NAFLD but pertain to patients with diabetes and ALD or HBV as well. Judging from pathophysiology, the most likely reason is co‐existing NAFLD, although this diagnosis is often excluded if the patient has ALD or HBV. The presence of diabetes or components of the metabolic syndrome could alert the clinician of the possibility of double or triple trouble steatosis or cirrhosis. These patients could benefit from intensive lifestyle intervention or bariatric surgery even if diagnosed as having ALD or HBV.