Transcript Risk of Hepatocellular Carcinoma in Diabetic Patients and

Am J Gastroenterol 2012; 107:46–52
Drmohammad Sadrkabir
 Hepatitis B and hepatitis C are well-known etiological
factors that lead to hepatocellular carcinoma (HCC) in
Taiwan. Cirrhosis, obesity, diabetes mellitus (DM), fatty
liver disease, hereditary hemochromatosis, alcohol,
smoking, and other dietary and environmental exposures
are also factors that contribute to the development of HCC.
 A recent case–control study that consists of 2,061 patients
with HCC and 6,183 noncancer controls has shown a
significant association between DM and the risk of HCC
with an odds ratio of 2.87 .
 A study of site-specific cancer mortality in Asian
populations has also reported that DM is associated with
higher risk of mortality from liver and pancreatic cancer
(1.51 and 1.78, respectively) .
 Metformin, a widely used anti-diabetic drug, has
recently attracted great attention for lowering cancer
risk. It has been found to inhibit cancer cell growth in
vitro and in vivo.
 Epidemiological studies have also shown that
metformin therapy is associated with reduced risks of
breast cancer and HCC .
 Another case–control study conducted in the United
States also showed that treatment with metformin or
thiazolidinediones is associated with a 70% reduction
in HCC risk among diabetic patients .
 Both the case–control studies are limited, with small
clinical samples of HCC.
 To the best of our knowledge, only one cohort study,
also from Taiwan, has observed the effect of DM
medication in reducing the risk of HCC among Asian
populations, but only metformin was studied .
 To clarify the role of diabetes in the risk of developing
HCC, a population-based cohort study, taking
advantage of a large-size data set available from the
National Health Insurance program in Taiwan, was
conducted.
 The present study investigates further whether the risk
of HCC increases with the presence of hepatitis B
and/or hepatitis C. Furthermore, it examines whether
the HCC risk is reduced with DM therapies, including
metformin and thiazolidinediones.
METHODS
 The National Health Insurance program in Taiwan is a
universal health insurance system covering more than
99% of the country's population of 23 million .
 Data used in this analysis came from the Longitudinal
Health Insurance Database, which contains the claims’
history of 1 million subjects randomly selected from the
entire population. This database contains registration of
insurance, inpatient and outpatient claims, prescribed
drugs, and basic sociodemographic information,
including sex and birth date.
 The retrospective cohort study was composed of two
study groups: a diabetic patients group and a matched
non-diabetic comparison group.
 The diabetic group included patients with an initial




diagnosis of DM who had been prescribed anti-diabetic
drugs (e.g., metformin, sulfonylureas, thiazolidinediones,
α-glucosidase inhibitors, D-phenylalanine derivatives,
dipeptidyl peptidase 4 inhibitors, and incretin mimetic
agents or insulins) in 2000–2005.
All subjects younger than 20 years old on the day of
diagnosis were excluded.
For each diabetic patient, four subjects without medical
claims for diabetes who were frequency matched with sex
and age (per 5 years) in the same period were randomly
selected.
Subjects diagnosed with cancer before the index date were
excluded from the present study.
Other comorbidities presented before the index date were
defined as follows: obesity, cirrhosis , alcoholic liver
damage, nonalcoholic fatty liver disease , hereditary
hemochromatosis ; hepatitis B, and hepatitis C .
 Both diabetic and non-diabetic groups were followed up to
determine the incidence of HCC until the end of 2008 or
censored because of death, withdrawal from the insurance
program, or loss to follow-up consultations.
 All types of personal identification on files connected with
the present study were scrambled using surrogate
identification numbers to secure patient privacy. The
present study was exempted from ethical review.
Statistical analysis
 χ2-Tests and Student's t-tests were used to compare the
differences between the DM group and the non-diabetic group
regarding sociodemographic characteristics and comorbidities.
 Cox proportional hazard models were used to estimate the HR
with 95% CI, which determined the association between
diabetes and the risk of developing HCC. The risk of developing
HCC associated with comorbidities such as DM, cirrhosis,
hepatitis B, and hepatitis C were also estimated. Further analysis
investigated medications available for the DM treatment. The
Cox proportional hazard analysis was also used to estimate
whether there were reduced HCC risks associated with DM
medications.
 All analyses were performed using the SAS software version 9.1
(SAS Institute, Cary, NC), and the statistical significance level
was set at two-sided P<0.05.
RESULTS
The multivariate
Cox proportional
hazard regression
analysis with sex,
age, and
comorbidities:
 As shown in Table5, the association between anti-diabetic drugs
and the risk of HCC was further analyzed. The median duration
of taking metformin was 2.1 years (mean, 2.5 years), similar to
that of sulfonylureas (mean, 2.4 years), but more than 1 year
longer than that of other anti-diabetic drugs (P<0.001). After
adjusting for sex, age, and comorbidities, the patients taking
metformin had the lowest HCC HR at 0.49 (95% CI=0.37–0.66),
followed by patients taking thiazolidinediones (HR=0.56, 95%
CI=0.37–0.84). Taking insulin, sulfonylurea, and α-glucosidase
inhibitors also reduced the HCC risk; however, the reductions
were not statistically significant.
DISCUSSION
 Diabetic patients have an incidence of HCC twice higher
than non-diabetics, indicating that approximately 11
additional cases of HCC develop annually per 10,000
diabetic patients. The incidence of HCC is higher in
patients with DM regardless of sex, age, or follow-up
period.
 The adjusted HR for developing HCC in diabetics in the
current study is 1.73, which is lower than the risk found in
other studies. In the systematic review by El-Serag et al.,
DM was associated with an increased risk of developing
HCC (risk ratio=2.5, 95% CI=1.9–3.2) independent of
alcohol use or viral hepatitis. Another systematic review
showed that patients with DM are 3.64 times (95%
CI=2.61–5.07) more likely to develop HCC compared with
non-diabetics.
 In line with previous findings, the present study observed a
synergistic effect between DM and liver comorbidities
regarding the development of HCC.
 The HCC risk was strongly associated with these three
comorbidities, particularly cirrhosis (HR=8.65).
 Patients with hepatitis C are at twice higher HCC risk than
those with hepatitis B (HR=5.61 vs. HR=2.52).
 A new and significant finding is that HCC increased
markedly to an HR of 72.4 for diabetics with cirrhosis and
hepatitis C compared with patients without those
disorders.
 Several studies have assessed the association between anti-
diabetic drugs and the risk of developing HCC; all found
reduced risk associated with metformin treatment . The
results of the present study confirm the effect of
metformin.
 Furthermore, we found that thiazolidinedione treatment is
also significantly associated with a reduced incidence of
HCC. To our knowledge, only one hospital-based case–
control study in the United States has shown that
thiazolidinediones may reduce the risk of HCC, which is
consistent with our observation.
 Evidence from in vivo studies has shown that
thiazolidinediones inhibited tumor formation in the liver.
 Previous case–control studies have shown that sulfonylurea
users have an increased risk of developing HCC compared
with non-users . In contrast, the present study shows a 25%
risk reduction in patients using sulfonylureas, but the
association was not statistically significant. We found that
92.5% of sulfonylurea users have switched to other antidiabetic drugs, which might, to some extent, offset the
unfavorable effect of sulfonylurea on risk of HCC. Further
prospective studies may be helpful for clarifying the
association of sulfonylureas with HCC.
 A hospital-based case–control study in the United States found
that patients that have had diabetes for >10 years have a 2.2-fold
increased risk of developing HCC (95% CI=1.2–4.8) compared
with those that have had diabetes for 2 to 5 years. Thus, the risk
may increase with the increasing duration of DM.
 We assessed the risk of developing HCC in terms of the duration
of taking anti-diabetic drugs. The risk of developing HCC
decreased as the duration of taking the medications increased.
The trend is most obvious for patients who have been taking
metformin for at least 1 year. The HR was 0.49 (95% CI=0.31–
0.78) after taking metformin for 12–23 months, which decreased
to 0.26 (95% CI=0.18–0.39) after taking it for ≥24 months
compared with non-users of metformin (data not shown). This
trend was less pronounced among patients taking α-glucosidase
inhibitors.
 The strength of the present study is its large sample size.
Although the concept is not novel, the population-based data set
with a large sample size allows the demonstration of risk factors
for HCC with a minimal tendency for selection bias in Taiwan.
 However, the present study has a number of limitations. First, a
number of suspected risk factors for HCC were not available,
such as cigarette smoking, aflatoxin exposure, and body mass
index. Second, the claims’ data do not contain laboratory test
results. Thus, the extent of DM control among the study subjects
was not accounted for because hemoglobin A1c values are not
available. Third, this observational study does not explore the
mechanism by which DM is related to HCC. Finally,
misclassification and measurement errors in drug exposure
might have occurred if the patients failed to take the prescribed
drugs. Non-compliance is likely to cause underestimation of the
drug effect.
CONCLUSION
 The current study suggests that patients with DM have a
higher risk of developing HCC.
 Comorbidities such as cirrhosis, hepatitis B, and hepatitis
C significantly aggravate the risk of developing HCC.
 The markedly elevated risk of developing HCC associated
with hepatitis C and its synergism with cirrhosis provides
new insights into HCC prevention. This observation may
prompt the screening of high-risk patients.
 On the other hand, patients taking metformin or
thiazolidinediones have reduced risks of developing HCC.
Thank you for your attention