Acta Diabetol (2014) 51:455460

DOI 10.1007/s00592-013-0538-y

ORIGINAL ARTICLE

Impact of liver disease severity and etiology on the occurrence

of diabetes mellitus in patients with liver cirrhosis
Jean Michel Petit Samia Hamza Fabien Rollot Vanessa Sigonney
Elodie Crevisy Laurence Duvillard Jean Jacques Raab Jean Pierre Bronowicki
Brigitte Bernard-Chabert Vincent Di Martino Michel Doffoel Helene Barraud
Carine Richou Jean Louis Jouve Patrick Hillon

Received: 25 September 2013 / Accepted: 27 November 2013 / Published online: 19 December 2013
Springer-Verlag Italia 2013

Abstract The association between liver cirrhosis (LC)

and diabetes mellitus (DM) is well known. However, the
impact of the severity or etiology of LC on the occurrence
of DM is relatively unknown. We aimed to determine the
prevalence and clinical correlates of DM in a large cohort of
patients with cirrhosis. A total of 1,068 patients with LC
were included in this cross sectional study (CIRCE study).
The diagnosis of cirrhosis irrespective of its etiology was
based on histological confirmation by liver biopsy or, in the
absence of biopsy, on typical clinical, morphological and
biological data. Data related to the cirrhosis etiology:
alcohol, viral markers of hepatitis B, C, iron load parameters and autoimmune markers were collected for each
patient. Venous blood samples were taken in the morning
after 12-h overnight fasting. There were 383 patients with

cirrhosis associated with hepatocellular carcinoma (HCC).

DM was found in 412 (39.7 %) patients. Patients with DM
were older and more likely to be overweight and male, with
a family history of DM and a diagnosis of HCC. DM was
not associated with a history of stroke or myocardial
infarction. Cirrhosis secondary to hepatitis infection was
less strongly associated with DM than with NASH or
alcoholic cirrhosis. The severity of LC was not associated
with DM. In multivariate analysis, the factors associated
with DM were age, BMI, a family history of DM, and statin
use. There was a significant interaction between HCC and
cirrhosis etiology for the risk of DM. Cirrhosis secondary to
hepatitis was associated with a lesser presence of DM only
in patients with HCC (interaction p = 0.0015). LC was
strongly associated with DM, with around 40 % of diabetic
patients. In the group of patients with LC without HCC,
diabetes was not associated with the etiology of cirrhosis.

Communicated by Massimo Porta.

This study was conducted for the CIRCE study group.
Please refer the Appendix section for the CIRCE study group
members.
J. M. Petit
S. Hamza
F. Rollot
L. Duvillard

J. L. Jouve
P. Hillon
University Hospital Dijon, Dijon, France
J. M. Petit
S. Hamza
V. Sigonney
E. Crevisy

L. Duvillard
P. Hillon
INSERM U 866, Dijon, France
J. M. Petit
S. Hamza
L. Duvillard
P. Hillon
University of Burgundy, Dijon, France
J. M. Petit (&)
Service de Diabetologie et dEndocrinologie, CHU du Bocage,
BP 77908, 21079 Dijon Cedex, France
e-mail: jean-michel.petit@chu-dijon.fr

Keywords Diabetes mellitus
Cirrhosis

Non-alcoholic fatty liver disease
Viral hepatitis

Alcoholic liver disease

J. J. Raab
General Hospital, Metz, France
J. P. Bronowicki
H. Barraud
INSERM 954, University Hospital Nancy, Nancy, France
B. Bernard-Chabert
University Hospital Reims, Reims, France
V. Di Martino
C. Richou
University Hospital Besancon, Besancon, France
M. Doffoel
University Hospital Strasbourg, Strasbourg, France

123

456

Abbreviations
LC
Liver cirrhosis
DM
Diabetes mellitus
HCC Hepatocellular carcinoma
MI
Myocardial infarction
HCV Hepatitis C virus

Introduction
Diabetes mellitus (DM) is very common in patients with
liver cirrhosis. Almost one-third of people with cirrhosis
have diabetes mellitus [1]. The main mechanism responsible for glucose abnormalities in cirrhotic subjects is a
defect in glucose uptake that produces marked and sustained hyperglycemia [2]. In these patients, diabetes mellitus influences the natural history of the liver disease and
increases the risk of developing hepatocellular carcinoma
(HCC), [3]. The impact of the severity of the liver disease,
the presence of HCC or the etiology of the liver cirrhosis
on the occurrence of DM is relatively unknown. For
instance, insulin resistance occurs in the early stages of
HCV infection and before the development of liver cirrhosis, but we do not known whether cirrhosis secondary to
hepatitis C infection is associated with a higher risk of DM
[4]. We aimed to determine the prevalence and clinical
correlates of DM in a large cohort of patients with
cirrhosis.

Acta Diabetol (2014) 51:455460

absence of biopsy, on typical clinical, morphological and

biological data. All cases of HCC evolving in a cirrhotic
liver were included. Criteria for the diagnosis of HCC were
those defined by the European Association for Study of the
Liver [5]. The absence of HCC in controls with liver cirrhosis at inclusion was assessed through high-quality
imaging examinations (abdominal US, CT scan or MRI)
and AFP below 100 ng/ml within the 2 months preceding
inclusion. Alcoholic cirrhosis was defined as alcohol consumption [14 units/week for the women and [28 units/
week for men or participant reports of current/previous
alcohol abuse. Diabetes was diagnosed using the 1997
American Diabetes Association fasting criteria (fasting
plasma glucose [7.0 mmol/l).
Exclusion criteria
Patients under 35 years old with progressive extrahepatic
cancer, human immunodeficiency virus infection, acute
alcoholic hepatitis, major somatic or psychiatric illness not
compatible with inclusion in the study, or with non-HCC
primary liver cancer were excluded.
Analytical procedures
Venous blood samples were taken in the morning after
12-h overnight fasting. Plasma glucose, HA1c, serum HDL
cholesterol, LDL cholesterol, serum triglycerides and
plasma liver enzymes were determined by standard
procedures.
Statistical analysis

Materials and methods

Study design and population
This investigation is an ancillary study of CIRCE, a multicenter case-control study of the etiological, pathophysiological and predictive aspects of liver disease in order to
investigate the mechanisms of liver carcinogenesis. Cirrhotic patients with or without HCC treated in the university hospitals of the French Canceropole du Grand Est
(Besancon, Dijon, Metz, Nancy, Reims, Strasbourg) have
been included since November 2008. Patients gave their
written informed consent to participate in the study. The
aims were to assess the role of metabolic and environmental factors (alcohol, tobacco, viruses and diet) and
drugs in the risk of HCC among cirrhotic patients.
Inclusion criteria
The diagnosis of cirrhosis irrespective of its etiology was
based on histological confirmation by liver biopsy or, in the

123

Proportions were used to describe categorical data, and

means and standard deviations to describe continuous data.
Patients characteristics were compared according to DM
occurrence by the Khi2 test for categorical data and the
nonparametric WilcoxonMannWhitney test for continuous data. Associations between DM and other covariates
were studied using unconditional logistic regression.
Interactions between each covariate and the occurrence of
DM were systematically tested by the likelihood ratio. A
p value below 0.05 was considered significant. The statistical analyses were done using STATA 12 software.

Results
A total of 1,068 patients with LC were included in this
study. There were 383 patients with cirrhosis associated
with hepatocellular carcinoma. The mean duration of cirrhosis was 3.5 4.7 years. DM was found in 412 (38.6 %)
patients. There were 307 (74.5 %) patients treated with

Acta Diabetol (2014) 51:455460

457

Table 1 subjects characteristics according to the presence of diabetes mellitus

Without DM

With DM

656

412

453 (69.0 %)

333 (80.8 %)

\65

453

242

C65

203 (30.9 %)

170 (41.2 %)

371
282 (43.1 %)

162
250 (60.6 %)

\0.001

No

448

194

\0.001

Yes

148 (24.8 %)

169 (46.5 %)

n
Gender (male)

p value

Impact of liver conditions on the occurrence of diabetes

mellitus
\0.001

Age (years)
0.001

BMI (kg/m2)
\27
C27
Family history of diabetes

History of myocardial infarction

No

624

380

Yes

31 (4.73 %)

28 (6.86 %)

No

634

390

Yes

20 (3.05 %)

18 (4.41 %)

No

163

100

Yes

490 (75.0 %)

312 (75.7 %)

of DM (Table 1). DM was not associated with a history of

stroke or myocardial infarction.

0.14

DM was present in 78 of the 263 (29.6 %) patients with

hepatitis, and in 256 of the 633 (40.6 %) patients with
alcoholic cirrhosis (Table 1). Cirrhosis secondary to hepatitis infection was less strongly associated with DM than
with alcoholic cirrhosis or the group of NASH and others
liver diseases (p = 0.001). There was no difference
between patients infected by hepatitis C and by hepatitis B
regarding the occurrence of DM. The severity of LC
(CHILD classification) was not associated with DM.
Patients with DM were more likely to have hepatocellular
carcinoma (Table 1).
Multivariate analysis

History of stroke
0.25

Current smoker
0.80

HDL cholesterol (mmol/l)

0.78 0.52

0.77 0.46

0.69

LDL cholesterol (mmol/l)

Triglycerides (mmol/l)

1.96 1.30
1.14 0.6

1.80 1.08
1.17 0.6

0.15
0.08

No

603

318

Yes

53 (8.07 %)

94 (22.8 %)

No

187

122

Yes

464 (71.2 %)

285 (70.0 %)

No

444

241

Yes

212 (32.5 %)

171 (41.6 %)

335 (52.6 %)

216 (53.7 %)

197 (30.9 %)

135 (33.5 %)

104 (16.3 %)

51 (12.7 %)

Hepatitis C
Hepatitis B

159 (71.3 %)
26 (65.0 %)

64 (28.7 %)
14 (35 %)

Alcoholic

377 (59.5 %)

256 (40.4 %)

NASH and othersa

71 (50.3 %)

70 (38.9 %)

Statin therapy
\0.001

Alcohol abuse
0.66

Hepatocellular carcinoma
0.001

Child classification
0.24

Etiology of liver cirrhosis

In multivariate analysis, the factors associated with DM

were sex (ORfemale 0.59; 95 % CI 0.430.83; p = 0.002),
age ([65 years old: OR 1.39, 95 %CI 1.041.87;
p = 0.027), BMI ([27 kg/m2, OR 1.70, 95 %CI
1.292.23, p \ 0.001), family history of DM (OR 2.88,
95 %CI 2.133.85, p \ 0.001) and statin use (OR 2.61,
95 %CI 1.763.85; p \ 0.001) (Table 2). There was a
significant interaction between hepatocellular carcinoma
and the cirrhosis etiology for the risk of DM. Only in
patients with hepatocellular carcinoma, alcoholic cirrhosis
and NASH were associated with greater presence of DM
(interaction p = 0.0015) (Table 2).
When HCC variable is taken as the dependent variable
in the multivariate analysis, factors significantly associated
with HCC were sex (ORfemale 0.38 [0.270.55], p \ 0.001),
age (ORC65 2.79 [2.103.70], p \ 0.001) and the etiology
of liver cirrhosis (ORhepatitis B versus hepatitis C 2.03
[0.994.12], p = 0.05; ORNASH and other versus hepatitis C 0.58
[0.350.95], p = 0.031). The association between diabetes
and HCC was borderline to significance (OR 1.32
[0.991.76], p = 0.059).

0.001

Including autoimmune hepatitis, hemochromatosis and primary

biliaris cirrhosis

anti-diabetic drugs (metformin 59.6 %, insulin 50.5 % and

sulfonylurea 50.5 %). Patients with DM were older and
more likely to be overweight and male with family history

Discussion
The presence of liver cirrhosis is associated with significant
impairment in glucose homeostasis [1, 6]. The reported
prevalence of DM in people with cirrhosis ranges from 20
to 40 % [1, 6, 7]. The strength of our study lies in the
number of subjects included, with more than a thousand
cirrhotic patients. We found that 38.5 % of patients with
liver cirrhosis had DM. This number is slightly higher than
those described in previous studies [1, 6, 7]. One

123

458

Acta Diabetol (2014) 51:455460

Table 2 Independent factors associated with diabetes mellitus (multivariate logistic regression)
OR

95 % CI

Sex
Male

Female

0.59

0.002
[0.430.83]

Age
\65

C65

1.39

0.027
[1.041.87]

BMI
\27

C27

1.70

\0.001
[1.292.23]

family history of DM
No

Yes

2.88

\0.001
[2.133.89]

Statin therapy
No

Yes

2.61

Etiology of liver cirrhosis

\0.001
[1.763.85]

Hepatocellular carcinoma
No

Yes

Hepatitis C

Hepatitis B

1.79 [0.635.14] p = 0.276

0.96 [0.303.05]

p = 0.952

Alcoholic
NASH and othera

1.12 [0.731.72] p = 0.601

2.61 [1.394.88]

p = 0.003

1.33 [0.762.33] p = 0.316

4,36 [1.7510.89]

p = 0.002

There was a significant interaction between hepatocellular carcinoma and cirrhosis etiology for the risk of diabetes mellitus. Cirrhosis secondary
to hepatitis was associated with a lesser risk of diabetes mellitus only in the group of patients with hepatocellular carcinoma (interaction
p = 0.0015)
a
Including autoimmune hepatitis, hemochromatosis and primary biliaris cirrhosis

explanation could be the high prevalence of alcoholic cirrhosis in our study with more than 70 % of patients
reporting current or previous alcohol abuse; alcohol abuse
is a well known risk factor of DM [8]. DM is associated
with a poor prognosis in liver cirrhosis with 23-fold
increase in the risk of hepatocellular carcinoma [3, 9]. DM
is an independent risk factor for hepatocellular carcinoma,
regardless of the presence of HCV, HBV, alcoholic liver
disease, or non-specific cirrhosis [9]. Though the impact of
DM on the natural history of liver disease is widely
accepted, the impact of the characteristics of the liver
disease on the risk of DM is relatively unknown. Several
studies have suggested a link between HCV infection and
DM. Previously, we reported that insulin resistance
occurred in the early stages of HCV infection, before the
development of liver cirrhosis [4]. HCV infection leads to
insulin signaling defects in hepatic IRS-1 tyrosine phosphorylation that contribute to insulin resistance, which
leads to the development of DM in patients with HCV
infection [10]. These data suggest that DM could be more
common in patients with cirrhosis secondary to hepatitis C
infection than in other cirrhotic patients. Interestingly, our

123

study did not find a higher risk of DM in patients with

hepatitis infection. In contrast, we found a lower risk of
DM only in the group of patients with hepatocellular carcinoma. This was probably because the impact of alcohol
abuse and NASH on the risk of DM is stronger than that of
hepatitis infection. Chronic alcohol abuse is considered a
potential risk factor for the incidence of type 2 diabetes
mellitus. Compared with moderate consumption, heavy
consumption may be associated with up to a 43 % higher
incidence of diabetes [8]. In the same way, NASH is
strongly associated with DM [11]. Our study did not rule
out the diabetogenic effect of HCV infection, but only
suggested that this effect is no greater than that of alcohol
or NASH. In patients with cirrhosis, the main factors linked
to the presence of DM are the classical factors: age,
overweight and a family history of DM. Our study found
that the severity of the liver cirrhosis and the presence of
HCC were not associated with the occurrence of DM. The
association between HCC and DM is dependent on the
etiology of the cirrhosis, the gender of the patient and
glycemic control, which could explain some of the controversial results of different studies [12, 13]. However,

Acta Diabetol (2014) 51:455460

when we performed a logistic regression with HCC as

outcome, we found that the association between DM and
HCC was border significant (p = 0.06). These results
reflect a limitation of our study regarding the sample size of
subjects included and the significant interaction between
HCC and the cirrhosis etiology for the risk of DM.
Recently, Tseng demonstrated the importance of taking
into account these potential biases to evaluate the association between HCC and diabetes status [13]. The significant association between statin therapy and DM observed
in our study was probably linked to the medical habit of
prescribing statins for patients with DM. However, we
cannot exclude the possibility that statin therapy increases
the risk of diabetes as previously demonstrated in the
general population [14].
In this large cohort, we found no association between a
history of macrovascular complications (myocardial
infarction and stroke) and DM. This finding is probably
explained by the already known tendency to a low prevalence of ischemic events in cirrhotic patients [15]. Interestingly, Marchenisi et al. [15] had previously observed
that cirrhotic patients, even in the presence of overt diabetes, were at a low risk of cardiovascular disease. The low
prevalence of a history of myocardial infarction or stroke in
our cirrhotic patients with diabetes may be related to a
shorter duration of diabetes and to protective factors
(reduced blood pressure and low cholesterol levels), [15].
These data confirm that in patients with DM and cirrhosis,
the prognosis is more strongly related to the liver disease
than to DM and cardiovascular diseases.
Several limitations of the current study need to be considered. First, it was a cross sectional study. Second, we do
not have the data regarding the diabetes duration, in consequence, we are unable to distinguish patients with hepatogenous diabetes and patients with type 2 diabetes and cirrhosis.
In conclusion, liver cirrhosis is strongly associated with
DM, with around 40 % of people with LC suffering from
diabetes. Even though DM is very common in patients with
cirrhosis, in the group of patients with LC without HCC,
diabetes was not associated with the characteristics of the
liver disease. In comparison with cirrhosis secondary to
alcoholic disease or NASH, cirrhosis secondary to hepatitis
was associated with a lesser presence of DM only in
patients with hepatocellular carcinoma.
Acknowledgments Institut National du Cancer (INCA), Ligue
Nationale contre le Cancer, Fondation de France, Canceropole Grand
Est. We thank Dr C Bonithon-Kopp of the Centre dInvestigation
Clinique INSERM CHU de Dijon for assistance with this study and
Philip BASTABLE for his help in reviewing the manuscript.
Conflict of interest The authors who have taken part in this study
declare that they have no conflicts of interest to disclose with respect
to this paper.

459

Appendix
CiRCE Study Group includes (1) CiRCE Coordination
France: Bronowicki JP, Di Martino V, Doffoel M, Hillon P
(CiRCE coordinator), Thiefin G, CiRCE Coordination
China: Wen H, He FP, Lu XM, Hillon P and Vuitton D (2)
Circe Scientific Board: Faivre J (President), Cercueil JP,
Cottet V, Delmas D, Ducoroy P, Duvillard L, Guenneugues
M, Gueant JL, Habersetzer F, Latruffe N, Manfait M,
Oudet P, Sockalingum G, CiRCE pharmacologists : Sgro
C, Gillet P, Kantelip JP, Trenque T and Welsch M.
xydyfywsb@sina.com; hefp5577@126.com; luxiaomei88@163.com;
dominique.vuitton@univ-fcomte.fr;
jean.faivre@chu-dijon.fr; jean-pierre.cercueil@chu-dijon.fr;
vanessa.cottet@u-bourgogne.fr;
dominique.delmas@
u-bourgogne.fr;
patrick.ducoroy@efs.sante.fr;
Laurence.duvillard@chu-dijon.fr;
marc.guenneugues@
canceropole-ge.org;
Jean-Louis.Gueant@medecine.uhpnancy.fr;
Francois.Habersetzer@chru-strasbourg.fr;
latruffe@u-bourgogne.fr; michel.manfait@univ-reims.fr;
Pierre.Oudet@chru-strasbourg.fr; ganesh.sockalingum@
univ-reims.fr;; pierre.gillet@chu-nancy.fr; jpkantelip@
chu-besancon.fr; ttrenque@chu-reims.fr; marie.welsch@chrustrasbourg.fr.

References
1. Garcia-Compean D, Jaquez-Quintana JO, GonzalezGonzalez
JA, Maldonado-Garza H (2009) Liver cirrhosis and diabetes: risk
factors, pathophysiology, clinical implications and management.
World J Gastroenterol 21(15):280288
2. Nielsen MF, Caumo A, Aagaard NK, Chandramouli V, Schumann WC, Landau BR, Schmitz O, Vilstrup H, Nielsen MF
(2005) Contribution of defects in glucose uptake to carbohydrate
intolerance in liver cirrhosis: assessment during physiological
glucose and insulin concentrations. Am J Physiol Gastrointest
Liver Physiol 288:G1135G1143
3. Veldt BJ, Chen W, Heathcote EJ, Wedemeyer H, Reichen J,
Hofmann WP, de Knegt RJ, Zeuzem S, Manns MP, Hansen BE,
Schalm SW, Janssen HL (2008) Increased risk of hepatocellular
carcinoma among patients with hepatitis C cirrhosis and diabetes
mellitus. Hepatology 47:18561862
4. Petit JM, Bour JB, Galland-Jos C, Minello A, Verges B, Guiguet
M, Brun JM, Hillon P (2001) Risk factors for diabetes mellitus
and early insulin resistance in chronic hepatitis C. J Hepatol
35:279283
5. Bruix J, Sherman M, Llovet JM, Beaugrand M, Lencioni R,
Burroughs AK, Christensen E, Pagliaro L, Colombo M, Rodes J
(2001) Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL conference. European
Association for the Study of the Liver. J Hepatol 35:421430
6. Zein NN, Abdulkarim AS, Wiesner RH, Egan KS, Persing DH
(2005) Prevalence of diabetes mellitus in patients with end-stage
liver cirrhosis due to hepatitis C, alcohol, or cholestatic disease.
J Hepatol 32:209217

123

460
7. Wlazlo N, Beijers HJ, Schoon EJ, Sauerwein HP, Stehouwer CD
(2010) High prevalence of diabetes mellitus in patients with liver
cirrhosis. Diabet Med 27:13081311
8. Howard AA, Arnsten JH, Gourevitch MN (2004) Effect of
alcohol consumption on diabetes mellitus: a systematic review.
Ann Inter Med 140:211219
9. Davila JA, Morgan RO, Shaib Y, McGlynn KA, El-Serag HB
(2005) Diabetes increases the risk of hepatocellular carcinoma in
the United States: a population based case control study. Gut
54:533539
10. Aytug S, Reich D, Sapiro LE, Bernstein D, Begum N (2003)
Impaired IRS-1/PI3-kinase signaling in patients with HCV: a
mechanism for increased prevalence of type 2 diabetes. Hepatology 38:13841392
11. Kotronen A, Yki-Jarvinen H (2008) Fatty liver: a novel component of the metabolic syndrome. Arterioscler Thromb Vasc Biol
28:2738

123

Acta Diabetol (2014) 51:455460

12. Arase Y, Kobayashi M, Suzuki F, Suzuki Y, Kawamura Y, Akuta
N, Kobayashi M, Sezaki H, Saito S, Hosaka T, Ikeda K, Kumada
H, Kobayashi T (2013) Effect of type 2 diabetes on risk for
malignancies includes hepatocellular carcinoma in chronic hepatitis C. Hepatology 57:964973
13. Tseng CH (2013) Diabetes is not an independent risk factor for
hepatocellular carcinoma. Diabetes Metab Res Rev 29:515524
14. Wang KL, Liu CJ, Chao TF, Huang CM, Wu CH, Chen SJ, Chen
TJ, Lin SJ, Chiang CE (2012) Statins, risk of diabetes, and
implications on outcomes in the general population. J Am Coll
Cardiol 60:12311238
15. Marchesini G, Ronchi M, Forlani G, Bugianesi E, Bianchi G,
Fabbri A, Zoli M, Melchionda N (1999) Cardiovascular disease
in cirrhosisa point-prevalence study in relation to glucose tolerance. Am J Gastroenterol 94:655662