Type 2 diabetes and hepatocellular carcinoma: A case–control study in patients with chronic hepatitis B
Type 2 diabetes has been suggested as an independent risk factor for the development of hepatocellular carcinoma (HCC). However, the role of Type 2 diabetes on the development of HCC in the presence of chronic hepatitis B (CHB) remains inconclusive. We conducted this hospital-based case–control study to evaluate the roles of Type 2 diabetes in HCC development in patients with CHB. From January 2004 to December 2008, a total of 6,275 eligible consecutive patients with chronic hepatitis B virus (HBV) infection were recruited. A total of 1,105 of them were patients with HBV-related HCC and 5,170 patients were CHB but without HCC. We used multivariate logistic regression models to investigate the association between Type 2 diabetes and HCC risk. The prevalence of Type 2 diabetes is higher among HCC patients without cirrhosis than among those with cirrhosis (12.1% vs. 6.7%, p = 0.003). Type 2 diabetes was associated with a significantly high risk of HCC in female patients after adjusting for age, family history of HCC, city of residence, hepatitis B e antigen and cirrhosis with an odds ratio (95% confidence interval, CI) of 1.9 (1.1–3.4). Restricted analyses among female patients without cirrhosis indicated that Type 2 diabetes was strongly associated with HCC risk with adjusted odds ratio (95% CI) of 5.6 (2.2–14.1). In conclusion, Type 2 diabetes is independently associated with the increased risk of HCC in female CHB patients. Female CHB patients with Type 2 diabetes are of a high HCC risk population and should be considered for HCC close surveillance program.
Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, with 55% occurring in China alone.1, 2 In China, nearly 80% of HCC cases have been linked to hepatitis B virus (HBV) infection and approximately 60–90% of these develop in patients with cirrhosis.3, 4 Other potential risk factors, such as diabetes mellitus, alcohol abuse and obesity, may also play a role in the development of HCC.5
A number of cohort and case–control studies have investigated the relationship between diabetes mellitus and HCC risk.6–15 Type 2 diabetes has been suggested as an independent risk factor for the development of HCC. However, the role of Type 2 diabetes in the development of HCC in the presence of chronic hepatitis B (CHB) has not been well documented. First, only a few cohort studies have followed a population with chronic HBV infection. In addition, most of the case–control studies used a normal population or cancers other than HCC as controls, and chronic HBV infection status was not well matched between cases and controls.16 Second, none of these studies followed a cohort of patients with CHB or matched cirrhosis status between cases and controls with CHB. The majority of HBV-related HCC develops in patients with cirrhosis,17 in which the prevalence of Type 2 diabetes is higher than in the general population, and in CHB patients without cirrhosis.18 So, these studies may also inappropriately estimate the role of Type 2 diabetes in HCC development in the presence of CHB. Third, whether Type 2 diabetes plays the same role in HCC development between men and women in the presence of CHB is not clear. It has been speculated that in the absence of other strong etiological risk factors for HCC, such as cirrhosis and male gender, Type 2 diabetes plays an additional role on its development.
Thus, we have conducted this hospital-based case–control study to more precisely evaluate the role of Type 2 diabetes in the development of HCC in the presence of CHB and to explore whether Type 2 diabetes plays the same role in HCC development in males and females and in cirrhotic and noncirrhotic patients with CHB.
Study design and population
This case–control study is part of an ongoing hospital-based prospective investigation19 that was conducted in Jinan Infectious Disease Hospital, a tertiary hospital in Shandong, China. The study was approved by the hospital ethics committee, and written informed consent for participation was obtained from each study participant.
Subjects who fulfilled the following criteria were recruited into the study: hospitalized for HCC or CHB, age ≥ 30 years, positive for hepatitis B surface antigen (HBsAg), negative for anti-HCV, without a history or other evidence of cancer other than HCC, without a history or other evidence of hepatitis other than hepatitis B, no cancer treatment, and no treatment with nucleotide/nucleosides or interferon, residence of Shandong Province. From January 2004 and December 2008, a total of 8,027 eligible consecutive patients (1,410 HBV-related HCC and 6,617 CHB without HCC) were enrolled. A total of 882 eligible patients (9.9%) were not recruited because of patient refusal or patient sickness. Statistical analyses indicated that the eligible patients who were not recruited did not differ from the recruited patients in terms of demographic and clinical features (retrieved from patients' medical records). As non-HCC patients with CHB requiring hospital care may be more likely to have heavy alcohol use than HCC patients, 305 HCC cases and 1,447 CHB patients with alcohol consumption were excluded from subsequent analysis.
A total of 1,105 HBV-related HCC patients were used as the case. The 5,170 hospital cross-sectional CHB patients without HCC were used as the control.
The variables analyzed in our study included age, sex, city of residence, family history of liver cancer, hepatitis B e antigen (HBeAg), cirrhosis, platelet count, Child-Pugh grade and Type 2 diabetes.
On entry to the hospital, all subjects were interviewed in person by trained physicians. Demographic data, family histories and medical histories were collected. Physical examinations, blood counts, serum biochemical parameters, prothrombin time, serum alpha-fetoprotein (AFP), anti-HCV (AxSYM HCV, version 3.0, Abbott Laboratories, Abbott Park, IL), HBsAg, anti-HBs, HBeAg, anti-HBe and anti-HBc (Abbott Laboratories) were all measured. HBV DNA quantification (Roche COBAS HBV Amplicor Monitor assay), ultrasound examinations and gastrointestinal barium meal X-ray examinations were also performed. First-degree relatives (parents, siblings and children) with liver cancer were considered to have a positive family history of liver cancer. A structured questionnaire was used to record the information collected.
At the hospital, a standardized questionnaire was used to interview the patients about their history of alcohol drinking, including drinking frequency, average monthly intake of each type of alcoholic beverage. We assessed total alcohol intake according to the average ethanol content, by volume, of beer (4–5%), wine (grape wine, rice wine 8–12%) and liquor (38–60%).20 Subjects with an ethanol intake of 30 g/day or more for men and 20 g/day or more for women for longer than 10 years were considered to have a positive history of alcohol consumption.
Diagnosis of HCC and cirrhosis
The criteria for diagnosis of HCC were positive histology or cytology, two imaging findings showing HCC from different sources (ultrasonography, enhanced computed tomography scan or magnetic resonance imaging), or serum AFP level greater than 400 ng/mL in combination with one positive image finding. Of these 1,105 cases, 298 cases were diagnosed by histological pathology, 364 cases with an enhanced computed tomography scan and magnetic resonance imaging and 443 cases with a finding of AFP higher than 400 ng/mL in combination with an abnormal finding from an enhanced computed tomography scan or ultrasonography.
Liver cirrhosis was histologically or clinically diagnosed. In patients with a biopsy, the diagnosis of cirrhosis was based on liver histology according to the criteria of Desmet et al.21 In patients without biopsy specimens, the diagnosis of cirrhosis was based on clinical and morphological criteria, ultrasound or computed tomography, according to standard definitions.22 These included the presence of clinical manifestations of portal hypertension (e.g., esophageal varices, encephalopathy or ascites), biochemical abnormalities (e.g., decreased serum albumin and platelets or prolonged prothrombin time) and obvious morphologic changes in the liver detected by hepatic imaging. Minor signs were also noted clinically, such as palmar erythema, spider angioma and clubbing of the fingers. Ultrasonography was performed by experienced radiologists for every patient on entry to the hospital.23 The severity of cirrhosis was evaluated using Child-Pugh score.24
Diagnosis of Type 2 diabetes
Diagnosis of diabetes mellitus was based on the 1999 World Health Organization criteria.25 Subjects who were found to have a fasting blood sugar level between 5.6 and 6.1 mmol/L were defined as having impaired fasting glucose and were referred for confirmation of their diabetes status. Those with a fasting blood sugar level of at least 7.0 mmol/L on at least two occasions or those diagnosed with Type 2 diabetes before entering the hospital were defined as having Type 2 diabetes and were referred for further diabetic care. For patients with a history of diabetes mellitus, the type of diabetes, the age at diagnosis and the duration of diabetes were also recorded. Patients with prior diagnosis of Type 1 diabetes were excluded from the study.
Clinical parameters were evaluated using the chi-squared test for categorical variables and the Mann–Whitney U test or Kruskal–Wallis test for continuous variables. Possible confounding effects among the variables were adjusted using a multivariate logistic regression model, and adjusted odds ratios (AORs) and 95% confidence intervals (CI) were calculated. Effect modifications were evaluated by stratification, statistical interaction was assessed by including main effect variables and their product terms in the logistic regression model. Considering platelet count and Child-Pugh grade which reflect the severity of liver disease have been reported to be associated with HCC risk,12, 26 these two variables were also adjusted, respectively, in the logistic regression models stratified by cirrhosis. For all tests, a p value of less than 0.05 was considered significant. All data analyses were performed using SPSS v. 13.0 (SPSS, Chicago, IL).
The demographic and clinical characteristics of HCC patients and their hospital cross-sectional CHB controls are summarized in Table 1. Most HCC patients were male (84.7%), HBeAg negative (76.8%) and cirrhotic (68.7%). Cases and controls had a similar distribution of family history of liver cancer and city of residence (Jinan and other cities of Shandong Province). HCC patients were older than controls; the mean age ± standard error was 53.8 ± 9.3 for patients with HCC and 44.9 ± 10.7 for controls. The mean age ± standard error of male and female subjects with HCC was 53.4 ± 9.1 years and 56.0 ± 10.0 years (p = 0.001), respectively. The mean age of HCC patients with cirrhosis and without cirrhosis was similar (53.9 ± 9.0 vs. 53.6 ± 9.8, p = 0.54). After adjusting age, sex, city of residence, family history of liver cancer, HBeAg status and cirrhosis, multiple logistic regression analyses indicate the AOR (95% CI) of HCC risk for male sex, old age (10-year increment) and cirrhosis are 2.6 (2.2–3.2), 2.0 (1.8–2.1) and 2.8 (2.4–3.3), respectively.
Table 1. Demographic and clinical characteristics of hepatitis B virus-related HCC cases and hospital cross-sectional CHB controls
Prevalence of Type 2 diabetes
The overall prevalence of Type 2 diabetes in HCC patients and the cross-sectional controls was 8.4% and 6.3%, respectively (p = 0.01) (Table 1). However, as shown in Table 2, in every age group, the prevalence of Type 2 diabetes among HCC patients is not higher than among controls. On the contrary, in the 50- to 59-year-old age group, the prevalence of Type 2 diabetes is higher among CHB controls than among HCC patients. In male patients, the prevalence of Type 2 diabetes was similar between HCC cases and controls. However, in female patients and in patients without cirrhosis, Type 2 diabetes was more frequent among HCC patients than among controls. Unexpectedly, the prevalence of Type 2 diabetes is higher among HCC patients without cirrhosis than among those with cirrhosis (12.1% vs. 6.7%, p < 0.001).
Table 2. Prevalence of Type 2 diabetes in different age, sex and cirrhosis status groups and duration of diabetes among hepatitis B virus-related HCC cases and hospital cross-sectional CHB controls
Type 2 diabetes and HCC risk
As shown in Table 3, in multivariate logistic regression comparing all HCC cases with the hospital cross-sectional controls, no association between Type 2 diabetes and HCC risk was observed. However, in subgroup analysis in female subjects, a significant association between Type 2 diabetes and HCC risk was observed with an AOR of 1.9 (95% CI, 1.1–3.4). In contrast, subgroup analysis in male subjects, no association between Type 2 diabetes and HCC risk was observed. These indicate effect modification (heterogeneity) of gender. The estimated X2 for homogeneity between AORs in men and women was 5.7. In addition, including an interaction term of gender and Type 2 diabetes in the logistic regression model along with the main effects of gender and Type 2 diabetes revealed a significant interaction between gender and Type 2 diabetes on HCC development (p = 0.008). The AOR (95% CI) for the interaction term of diabetes and sex was 0.42 (0.22–0.79), and the β coefficient was −0.88. Besides, including an interaction term of cirrhosis and Type 2 diabetes in the logistic regression model along with the main effects of cirrhosis status and Type 2 diabetes revealed a significant interaction between cirrhosis status and Type 2 diabetes on HCC development (p = 0.004). The AOR (95%) for the interaction term of diabetes and cirrhosis was 0.45 (0.27–0.77), and the β coefficient was −0.79. Restricted multivariate logistic regression analyses among female patients without cirrhosis indicated a strong association between Type 2 diabetes and HCC risk, with an AOR of 5.6 (95% CI, 2.2–14.1). In female patients with cirrhosis, no association between Type 2 diabetes and HCC risk was observed.
Table 3. Prevalence of Type 2 diabetes and AOR for the association between Type 2 diabetes and the development of HCC in the presence of CHB: Multivariate logistic regression analyzes
The proportion of patients with a duration of Type 2 diabetes > 5years is significantly higher among HCC cases than among controls (p < 0.001) (Table 1). Restricted analyses among diabetic cases and controls showed the AOR of developing HCC were 2.2 (95%CI, 2.2–4.2) for patients diagnosed with a duration of diabetes > 5 years after adjusting age, sex, family history of liver cancer, HBeAg and cirrhosis.
In this large case–control study in patients with CHB, Type 2 diabetes was a significant risk factor for HCC in women. The study revealed a high prevalence of Type 2 diabetes among female patients with HCC, but not among male patients with HCC, yielding a two-fold higher risk of HCC for female patients with Type 2 diabetes than for those without Type 2 diabetes. To our knowledge, this finding has not been reported in previous studies.
It is possible that a true relationship exists between Type 2 diabetes and HCC risk in female patients. A gender-based dimorphic pattern for leptin in patients with diabetes mellitus may partially explain the association. There is a strong link between leptin and cancer growth and development,27 with increasing evidence for the involvement of leptin in HCC development. Higher plasma leptin levels in females with diabetes mellitus may increase HCC risk. Additionally, higher serum ferritin and iron overload in women with diabetes mellitus may also account, in part, for the association between diabetes and HCC risk.28, 29 Liver iron overload is known to be a risk factor for HCC in patients with hemochromatosis, noncirrhotic liver, alcoholic cirrhosis or nonalcoholic steatohepatitis.30–32 Thus, iron overload might also be a possible link between diabetes mellitus and HCC.
Another important finding of our study is that we observed a significantly higher prevalence of Type 2 diabetes among HCC patients without cirrhosis than among those with cirrhosis (12.1% vs. 6.7%, respectively, p < 0.001). Accordingly, we noted a marginal significant AOR 1.44 (95% CI, 0.96–2.16) of HCC risk for those with Type 2 diabetes among subjects without cirrhosis. Moreover, restricted multivariate logistic regression analyses stratified by sex and cirrhosis status indicated that for female CHB patients without cirrhosis, those who had Type 2 diabetes had an approximately six times higher risk of HCC than did those without Type 2 diabetes (p < 0.001). Our finding indicates a synergistic association between Type 2 diabetes and cirrhosis status on HCC risk. Because of the small number of female HCC patients without cirrhosis in our study, this result should be considered with caution. However, this result is in consistent with a Taiwan cohort study, which reported the synergistic association between Type 2 diabetes and HCC risk among patients with both diabetes and hepatitis B infections, compared to those with the viral infections but without diabetes.33 Interestingly, several recent studies showed that HCC in metabolic syndrome and nonalcoholic fatty liver disease often develop in the absence of apparent cirrhosis.34–36 These results also support our finding that Type 2 diabetes is an additional HCC risk in the absence of cirrhosis in patients with CHB. High insulin concentration and insulin resistance preceding cirrhosis and HCC in Type 2 diabetes might play a key role in the HCC carcinogenesis. Thus, the discovery of a high prevalence of Type 2 diabetes in noncirrhotic HBV-related HCC patients should not be surprising.
We also noted a significant association between diabetes duration and HCC risk in our study population, a two-fold higher HCC risk for subjects with diabetes duration > 5 years than for those with a shorter duration. This finding is in agreement with newly published study by Hassan et al.,37 which reported that the magnitude of association between diabetes and HCC increased as the duration of diabetes increased. The result suggested that the association between diabetes and HBV-related HCC risk was not a temporal relationship.
Our study has several advantages over previous studies. First, to the best of our knowledge, this is the largest case–control study published to date that investigates the impact of Type 2 diabetes on the development of HCC in the presence of CHB. Second, cross-sectional inpatient CHB controls in our study largely represents the population from which the HBV-related HCC cases arose. In areas of high HBV endemicity, persons with cirrhosis have an approximately three-fold higher risk for HCC than those with chronic hepatitis but without cirrhosis and a 16-fold higher HCC risk than the inactive carrier.17 So, most HBV-related HCC arose from CHB with or without cirrhosis but not from inactive HBsAg carrier. Third, some possible confounding factors, such as alcohol consumption, family history of HCC, city of residence, HBeAg and cirrhosis status were evaluated in our study.
There are certain limitations to the study. First, overweight and obesity were not included in our study because of the overweight and duration was difficult to ascertain, subject to recall bias and be confounded by ascites in subjects with cirrhosis. Second, not all HCC and cirrhosis subjects were diagnosed histologically. However, we used a standard clinical criteria combined with well-defined hepatic imagine method for the diagnosis of HCC and cirrhosis patients without a liver biopsy.23 Moreover, all of the patients in our study were followed and revaluated every 1–3 months in the outpatient division of our hospital.
In conclusion, we report the first independent association between Type 2 diabetes and an increased HCC risk in female patients with CHB. This indicates besides male CHB patients and patients with cirrhosis, female CHB patients with Type 2 diabetes are of a high HCC risk population and should be considered for HCC close surveillance program. Further clinical and basic studies are necessary to confirm the gender difference in HCC risk and to elucidate the underlying mechanisms of this risk.