Continuation of Metformin Use After a Diagnosis of Cirrhosis Significantly Improves Survival of Patients With Diabetes
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
LIVER FAILURE/CIRRHOSIS/PORTAL HYPERTENSION
Continuation of Metformin Use After a Diagnosis of
Cirrhosis Significantly Improves Survival of Patients
With Diabetes
Xiaodan Zhang,1 William S. Harmsen,2 Teresa A. Mettler,1 W. Ray Kim,1 Rosebud O. Roberts,3
Terry M. Therneau,2 Lewis R. Roberts,1 and Roongruedee Chaiteerakij1,4
The risks and benefits of metformin use in patients with cirrhosis with diabetes are
debated. Although data on a protective effect of metformin against liver cancer develop-
ment have been reported, metformin is frequently discontinued once cirrhosis is diagnosed
because of concerns about an increased risk of adverse effects of metformin in patients
with liver impairment. This study investigated whether continuation of metformin after cir-
rhosis diagnosis improves survival of patients with diabetes. Diabetic patients diagnosed
with cirrhosis between 2000 and 2010 who were on metformin at the time of cirrhosis
diagnosis were identified (n 5 250). Data were retrospectively abstracted from the medical
record. Survival of patients who continued versus discontinued metformin after cirrhosis
diagnosis was compared using the log-rank test. Hazard ratio (HR) and 95% confidence
interval (CI) were calculated using Cox’s proportional hazards analysis. Overall, 172
patients continued metformin whereas 78 discontinued metformin. Patients who continued
metformin had a significantly longer median survival than those who discontinued metfor-
min (11.8 vs. 5.6 years overall, P < 0.0001; 11.8 vs. 6.0 years for Child A patients, P 5
0.006; and 7.7 vs. 3.5 years for Child B/C patients, P 5 0.04, respectively). After adjusting
for other variables, continuation of metformin remained an independent predictor of bet-
ter survival, with an HR of 0.43 (95% CI: 0.24-0.78; P 5 0.005). No patients developed
metformin-associated lactic acidosis during follow-up. Conclusion: Continuation of metfor-
min after cirrhosis diagnosis reduced the risk of death by 57%. Metformin should there-
fore be continued in diabetic patients with cirrhosis if there is no specific contraindication.
(HEPATOLOGY 2014;60:2008-2016)
See Editorial on Page 1818 leading cause of mortality both in the United States and
globally.1,2 The prevalence of type 2 diabetes in patients
with cirrhosis has been reported as 37%, five times
greater than in those without cirrhosis.3 Patients with
C
irrhosis is responsible for over 29,000 deaths
annually in the United States and 1 million cirrhosis with diabetes have a greater risk of liver-related
deaths annually worldwide, making it the 12th complications and death than patients with cirrhosis
Abbreviations: AFP, alpha-fetoprotein; ALD, alcoholic liver disease; ALT, alanine aminotransferase; BMI, body mass index; CI, confidence interval; CLD,
chronic liver disease; HbA1c, glycated hemoglobin; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HR, hazard ratio; INR, inter-
national normalized ratio; MELD, model for end-stage liver disease; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; PH, portal
hypertension; PT, prothrombin time; SD, standard deviation.
From the 1Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, MN; 2Division of Biomedical Statistics and Informatics,
Department of Health Sciences Research, Mayo Clinic College of Medicine and Mayo Clinic Cancer Center, Rochester, MN; 3Division of Epidemiology, Depart-
ment of Health Sciences Research, Mayo Clinic College of Medicine, and Mayo Clinic Cancer Center, Rochester, MN; and 4Department of Medicine, Faculty of
Medicine, Chulalongkorn University and King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bangkok, Thailand.
Received January 24, 2014; accepted April 29, 2014.
Additional Supporting Information may be found at onlinelibrary.wiley.com/doi/10.1002/hep.27199/suppinfo.
This work was supported by National Institutes of Health (NIH; grant nos.: CA128633 and CA165076; to L.R.R.), the Mayo Clinic Center for Cell Signaling
in Gastroenterology (NIDDK P30DK084567), the Mayo Clinic Cancer Center (CA15083), and the Mayo Clinic Center for Clinical and Translational Science
(NIH/NCRR CTSA grant no.: UL1 TR000135). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the
NIH.
2008HEPATOLOGY, Vol. 60, No. 6, 2014 ZHANG ET AL. 2009
without diabetes.4-9 Similarly, the risk of death from cir- Given the potential beneficial effects of metformin
rhosis among diabetic patients was greater than in the on liver inflammation and carcinogenesis, we hypothe-
general population, with a standardized mortality ratio sized that continuation of metformin use after the
of 2.5, which is even higher than the mortality ratio of diagnosis of cirrhosis would improve the survival out-
1.3 conferred by cardiovascular disease in patients with come of diabetic patients. The primary aim of our
diabetes.10 Therefore, it is important to improve study was to assess the survival difference between dia-
cirrhosis-related mortality in diabetic patients. betic patients who continued taking metformin versus
Metformin, a commonly prescribed oral hypoglyce- those who discontinued metformin after cirrhosis diag-
mic agent, has a protective effect against cancer devel- nosis. We also investigated other factors determining
opment and cancer mortality in type 2 diabetic survival outcomes in patients with cirrhosis with
patients.11-13 Metformin use was associated with a diabetes.
decreased risk of hepatocellular carcinoma (HCC)
development in diabetic patients with chronic liver dis-
ease (CLD).14 Metformin was also independently asso-
Patients and Methods
ciated with a reduction in HCC incidence and liver- Patients. All diabetic patients diagnosed with cir-
related death in patients with type 2 diabetes and hep- rhosis between January 1, 2000 and December 31,
atitis C virus (HCV)-induced cirrhosis.15 2010 and who were on metformin at cirrhosis diagno-
Nonalcoholic fatty liver disease (NAFLD) appears sis were included in this study (n 5 250). Patients
to partially mediate an increased risk of liver-related were categorized into two groups, that is, (1) those
death among patients with diabetes. The presence of who continued metformin after cirrhosis diagnosis and
NAFLD increased deaths in diabetic patients, with a (2) those who discontinued metformin after cirrhosis
hazard ratio (HR) of 2.2, with 19% of deaths being diagnosis. Continuation of metformin use was defined
liver related.16 Metformin prevented and reversed stea- as taking metformin for at least 3 months after cirrho-
tosis and inflammation in a nondiabetic mouse model sis diagnosis, and discontinuation of metformin use
of nonalcoholic steatohepatitis (NASH)17 and was defined as cessation of metformin within a 3-
improved liver histology and alanine aminotransferase month period after the diagnosis of cirrhosis. The 3-
(ALT) levels in patients with NASH.18,19 month period was arbitrarily chosen as a cutoff to
Because of theoretical concerns about the increased allow time for health care providers to decide whether
risk of lactic acidosis in diabetic patients with CLD, to continue or discontinue metformin. Metformin was
many clinicians are reluctant to prescribe metformin typically discontinued within this time frame after cir-
for diabetic patients with CLD and some recommend rhosis diagnosis in our cohort.
discontinuation of metformin after the diagnosis of The diagnosis of cirrhosis was ascertained by histol-
cirrhosis. However, the evidence that metformin indu- ogy (n 5 124) and/or by clinical features, namely,
ces liver injury is weak.20 The incidence of lactic aci- portal hypertension (PH), morphologic characteristics
dosis in diabetic patients treated with metformin is consistent with cirrhosis in cross-sectional radiologic
low, with approximately 0.03-0.5 cases per 1,000 images (small-sized nodular liver 6 caudate lobe
patient-years, and the incidence of lactic acidosis hypertrophy as well as PH indicated by the presence
among diabetic patients who take metformin does not of collateral vessels, varices, and/or splenomegaly),
differ from the incidence in diabetic patients not and/or thrombocytopenia (platelet count 6.5% or random glucose of 200 mg/dL with pres-
tion with cirrhosis and diabetes. ence of symptoms). Exclusion criteria were as follows:
Address reprint requests to: Lewis R. Roberts, M.B., Ch.B., Ph.D., Division of Gastroenterology and Hepatology, College of Medicine, Mayo Clinic, 200 First
Street Southwest, Rochester, MN 55905. E-mail: roberts.lewis@mayo.edu; fax: 507-284-0762.
Copyright VC 2014 by the American Association for the Study of Liver Diseases.
View this article online at wileyonlinelibrary.com.
DOI 10.1002/hep.27199
Potential conflict of interest: Nothing to report.2010 ZHANG ET AL. HEPATOLOGY, December 2014 (1) a history of malignancy except for nonmelanoma Variables with P < 0.1 in the univariate model were skin cancer at cirrhosis diagnosis; (2) incomplete clini- included in the multivariate model. MELD score was cal information; (3) age
HEPATOLOGY, Vol. 60, No. 6, 2014 ZHANG ET AL. 2011
Table 1. Characteristics of Patients With Cirrhosis With Diabetes
Variables Continued Metformin (n 5 172) Discontinued Metformin (n 5 78) P Value
Age,* years 61.7 6 9.8 60.2 6 9.9 0.24
(range) (38-87) (36-81)
Male, n (%) 100 (58.1) 42 (53.9) 0.52
Race, n (%) 0.96
White 150 (87.2) 69 (88.5)
Nonwhite 17 (9.9) 8 (10.2)
Unknown 5 (2.9) 1 (1.3)
BMI, kg/m2† 34.3 (20.7-63.5) 34.8 (20.1-72.7) 0.55
Obesity, n (%) 112 (65.1) 54 (69.2) 0.56
Etiology of cirrhosis, n (%) 0.80
HBV 6 (3.5) 0
HCV 16 (9.3) 14 (18.0)
ALD 19 (11.0) 10 (12.8)
NASH 98 (57.0) 44 (56.4)
Others 8 (4.7) 6 (7.7)
Unknown 25 (14.5) 4 (5.1)
MELD score* 8.9 6 3.0 10.6 6 4.42012 ZHANG ET AL. HEPATOLOGY, December 2014 Fig. 2. Survival of Child-Pugh class A (n 5 181) (A) and Child- Fig. 3. Survival of 142 patients with NASH-related cirrhosis (A) and Pugh class B/C (n 5 66) (B) patients with cirrhosis with diabetes 108 with non-NASH-related cirrhosis (B) who continued metformin ver- who continued metformin versus those who discontinued metformin. sus those who discontinued metformin. the benefit of continuation of metformin, we com- 30; 7.7 vs. 3.5 years; P 5 0.04) with an HR of 0.46 pared the survival of the diabetic patients who contin- (95% CI: 0.21-0.98; P 5 0.04; Fig. 2B). ued (n 5 121) versus discontinued metformin (n 5 When patients were categorized based on the etiol- 129) 1 year after cirrhosis diagnosis. The median sur- ogy of cirrhosis, the beneficial effect of metformin on vival of the patients who continued metformin survival was observed only in the NASH-related cir- remained significantly greater than that of the patients rhosis group. The median survival of patients with who discontinued metformin within 1 year after cir- NASH-related cirrhosis who continued metformin (n rhosis diagnosis (11.8 vs. 6.0 years; P < 0.0001; Sup- 5 98) versus discontinued metformin (n 5 44) was porting Fig. 2). 12.1 versus 5.1 years (P 5 0.0004) with an HR of The benefit of continuation of metformin on sur- 0.33 (95% CI: 0.17-0.63; P < 0.0001; Fig. 3A). Sur- vival was found regardless of severity of cirrhosis, as vival of those who continued versus those who discon- defined by Child-Pugh classification. Of the 181 Child tinued metformin was not statistically different in the A patients with cirrhosis, those who continued metfor- alcohol-, HCV-, and HBV-related cirrhosis groups, min (n 5 133) had a significantly greater median sur- likely a result of the small number of patients in each vival than those who discontinued metformin (n 5 group (29, 30, and 6 patients, respectively). Thus, to 48; 11.8 vs. 6.0 years; P 5 0.006) with an HR of obtain a larger number, we combined patients with death of 0.47 (95% CI: 0.27-0.82; P 5 0.009; Fig. cirrhosis from alcohol, HCV, HBV, other, and 2A). Similarly, in a pooled analysis, Child B (n 5 61) unknown etiologies into a non-NASH cirrhosis group and C (n 5 5) patients with cirrhosis who continued (n 5 108); no significant difference was observed in metformin (n 5 36) had a significantly greater median survival between the continued metformin and discon- survival than those who discontinued metformin (n 5 tinued metformin subgroups (Fig. 3B).
HEPATOLOGY, Vol. 60, No. 6, 2014 ZHANG ET AL. 2013
Eighty-one patients (43 of 172 [25%] in the contin- 1.07-1.18; P < 0.0001; Table 2). Given that the infor-
ued metformin group and 38 of 78 [49%] in the dis- mation on AFP value was missing in 52 of 250
continued group) died during a median follow-up time (20.8%) patients, we also performed another multivar-
of 5.2 years (range, 0.6-14.6). The median durations of iate analysis excluding the AFP value, and the primary
follow-up were 4.8 (range, 0.6-14.6) and 5.4 years finding of an association between the continuation of
(range, 0.6-13.6) in the continued and discontinued metformin use and survival remained unchanged (HR,
groups, respectively. At the last follow-up visit, 129 0.38; 95% CI: 0.23-0.64; P 5 0.0002; Supporting
(75%) and 40 (51%) patients in the continued and dis- Table 1).
continued groups were alive, of whom 54 (41.9%) and
15 (37.5%) patients had been followed for at least 5
years, respectively. Of the 81 deceased patients, 39
Discussion
(48%) had information on causes of death available in We conducted a clinic-/hospital-based cohort study
the medical record (21 and 18 in the continued and to determine the outcomes and safety of metformin in
discontinued metformin groups, respectively). The diabetic patients with cirrhosis. Continuation of met-
causes of death included liver-related deaths (n 5 32; formin use after cirrhosis diagnosis was associated with
82% of those with available cause of death informa- significantly improved survival in diabetic patients,
tion), pulmonary disease (n 5 3), and heart disease, regardless of severity of liver impairment. Our data
renal failure, accident, and leukemia (n 5 1 for each). suggest that metformin is safe in diabetic patients with
Predictors of Survival of Patients With Cirrhosis cirrhosis because no patients developed lactic acidosis
With Diabetes. Continuation of metformin use was while receiving metformin in our cohort.
significantly associated with better survival in patients In this study, metformin use was significantly associ-
with cirrhosis with diabetes. In univariate analysis, con- ated with a 57% reduction in risk of all-cause mortality
tinuation of metformin (HR, 0.43; 95% CI: 0.28-0.66; in diabetic patients with all stages of cirrhosis. Our find-
P 5 0.0001) and high serum albumin level (HR, 0.53; ing suggested that metformin may prevent liver-related
95% CI: 0.36-0.77; P 5 0.001) were significantly associ- mortality because most patients (82%) died from liver-
ated with better survival. By contrast, every 10-year related diseases. The mechanism by which metformin
increase in age (HR, 1.59; 95% CI: 1.25-2.02; P 5 reduces mortality in patients with cirrhosis is currently
0.0001), male sex (HR, 1.77; 95% CI: 1.11-2.82; P 5 unknown. Given that metformin use had a protective
0.02), increased MELD score (HR, 1.08; 95% CI: 1.02- effect against death only in the subgroup of NASH-
1.13; P 5 0.007), loge bilirubin (HR, 1.39; 95% CI: related patients with cirrhosis, it is biologically plausible
1.07-1.82; P 5 0.015), serum creatinine (HR, 2.7; 95% that metformin may slow progression of liver fibrosis by
CI: 1.51-4.87; P 5 0.0009), and every 10-ng/mL attenuating steatohepatitis. This hypothesis is supported
increase in serum AFP (HR, 1.13; 95% CI: 1.09-1.17; P by evidence from recent studies.17-19 Metformin has
< 0.0001) were significantly associated with worse sur- been demonstrated to reverse steatosis and inflammation
vival. Etiology of cirrhosis was also associated with sur- in a mouse model with NASH, reduce hepatocellular
vival; compared to the reference group with HBV- injury and improve ALT levels in NASH patients, and
induced cirrhosis, patients with HCV, NASH, or other significantly decrease hepatic fat, necroinflammation,
causes of cirrhosis had HRs in the 1.3-1.5 range, whereas and fibrosis in nondiabetic NASH patients.17-19
those with cirrhosis resulting from ALD had an HR of We did not find a significant survival benefit in sub-
3.8; however, none of these associations was statistically groups of patients with cirrhosis induced by HCV,
significant. Ethnicity, ALT, platelet count, INR, PT, and HBV, or alcohol, possibly because of the small num-
fasting plasma glucose were not associated with survival bers of patients, or the entire group of patients with
(Table 2). non-NASH-induced cirrhosis. Additional studies in
The benefit of metformin use on survival outcome larger cohorts are needed to determine whether met-
remained significant after adjusting for other variables formin confers a survival benefit for diabetic patients
(i.e., age, gender, albumin, MELD score, AFP level, with cirrhosis from causes other than NASH.
and etiology of cirrhosis; Table 2). Continuation of Baseline liver function, as indicated by MELD score
metformin use was independently associated with a and Child-Pugh classification, is known to be a strong
57% decrease in the risk of death with a HR of 0.43 predictor of survival in patients with cirrhosis. In our
(95% CI: 0.24-0.78; P 5 0.005). By contrast, every study, MELD score was not associated with survival of
10-ng/mL increase in AFP value was associated with a patients with cirrhosis, likely because most patients
13% increased risk for mortality (HR, 1.13; 95% CI: had low MELD scores. We found that a significantly2014 ZHANG ET AL. HEPATOLOGY, December 2014
Table 2. Cox Proportional Hazards Analysis of Variables Associated With Death of Patients With Cirrhosis With Diabetes
Univariate Analysis Multivariate Analysis*
Variables HR (95% CI) P Value Adjusted HR (95% CI) P Value
Age, per 10 years 1.59 (1.25-2.02) 0.0001 1.29 (0.91-1.81) 0.15
Male 1.77 (1.11-2.82) 0.02 1.51 (0.78-2.92) 0.22
Race
Nonwhite 1.00 (reference)
White 2.33 (0.85-6.40) 0.10
BMI 0.25 (0.04-1.54) 0.14
Obesity 1.35 (0.85-2.10) 0.20
Etiology of cirrhosis 0.01 0.03
HBV 1.00 (reference) 1.00 (reference)
HCV 1.37 (0.17-10.77) 0.39 (0.04-3.52)
ALD 3.86 (0.51-29.25) 2.29 (0.29-18.27)
NASH 1.32 (0.18-9.69) 0.97 (0.13-7.53)
Others 1.42 (0.16-12.77) 0.71 (0.07-7.57)
MELD score 1.08 (1.02-1.13) 0.0067 1.01 (0.93-1.09) 0.82
Laboratory results
AST level 1.0 (0.996-1.004) 0.97
ALT level 0.998 (0.994-1.002) 0.41
Loge bilirubin 1.39 (1.07-1.82) 0.015
Albumin 0.53 (0.36-0.77) 0.001 0.53 (0.28-1.01) 0.06
Platelets 1.001 (0.998-1.004) 0.56
INR 1.27 (0.54-2.98) 0.59
Creatinine 2.70 (1.51-4.87) 0.0009
Fasting plasma glucose, 0.998 (0.994-1.003) 0.47
AFP, per 10 ng/mL† 1.13 (1.09-1.17)HEPATOLOGY, Vol. 60, No. 6, 2014 ZHANG ET AL. 2015
fluctuations in blood sugar levels; hypoglycemia is a large enough to sustain the multivariate model that
particular concern of physicians when prescribing oral would be needed to investigate the effect of metformin
hypoglycemic agents. Therefore, some physicians prefer and, at the same time, provide valid estimates of the
insulin to avoid hypoglycemia in diabetic patients with concomitant effects of other medications. We were not
cirrhosis. However, metformin does not cause hypogly- able to confirm a benefit of metformin use on reduc-
cemia and can be used in combination with insulin ing HCC incidence as a result of the limited number
without increasing the risk of hypoglycemia. of patients who developed HCC in our cohort. More-
Interestingly, AFP was an independent risk factor over, although we adjusted for MELD score to account
for death of patients with cirrhosis with diabetes. for the influence of more-severe comorbidities in the
Given that a rising AFP value is used as a predictor of decision to discontinue metformin, MELD score may
HCC development, we investigated whether metfor- not completely account for the effects of all possible
min decreased the risk of HCC development in our comorbidities.
cohort. We did not find a significant association A number of research questions should be further
between continuation of metformin and reduced HCC investigated. It is worth examining the effect of metfor-
risk (data not shown). This result was inconsistent min on survival of patients with non-NASH cirrhosis in
with the study by Chen et al., which reported that a larger cohort. Whether initiating metformin subsequent
metformin reduced risk of HCC development by 7% to diagnosis of cirrhosis would improve survival of dia-
per year.13 This inconsistency may result from the lim- betic patients needs to be answered. It will also be inter-
ited number of patients who developed HCC in our esting to further investigate whether metformin has an
cohort (n 5 10 [5.8%] in continued and 5 [6.4%] in antifibrotic effect by slowing down progression or revers-
discontinued groups), because this cohort study was ing the degree of liver fibrosis. Given that the prevalence
powered to address the difference in survival, instead of impaired glucose tolerance in patients with cirrhosis is
of HCC incidence, in the patients with cirrhosis. The as high as 50%-60%, it will be of interest to investigate
information on lack of apparent HCC risk reduction whether metformin also improves the outcomes of
was provided as an underpowered secondary analysis. patients with cirrhosis with impaired glucose tolerance.
The major strength of our study lies in the large In conclusion, we report on the novel observation of
cohort of patients with cirrhosis with diabetes who an association between continuation of metformin use
have various causes of cirrhosis and a wide range of after cirrhosis diagnosis and improved survival of dia-
degrees of severity of liver impairment. To our knowl- betic patients. Our findings strongly suggest that contin-
edge, this is the largest cohort of unselected patients uation of metformin improves the survival of NASH-
with cirrhosis with diabetes in which the effect of met- related patients with cirrhosis with diabetes. Validation
formin use has been investigated. We also provided of these results in a larger cohort would support contin-
data on the safety of metformin use in patients with uation of metformin use in NASH-related cirrhosis.
cirrhosis. Therefore, this study has contributed signifi-
cantly to filling the knowledge gap in the debate on Acknowledgment: The authors thank Ms. Jennifer
whether metformin can be safely prescribed to patients L. Rud for her secretarial assistance.
with cirrhosis. Importantly, our study revealed the
novel observation of an association between metformin
use and improved survival in a cohort of patients with References
cirrhosis with diabetes. 1. Mini~ no AM, Murphy SL, Xu J, Kochanek KD. Deaths: final data for
Our study had a number of limitations. We were 2008. Natl Vital Stat Rep 2011;59:1-126.
not able to take into consideration the effect of dose 2. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V,
et al. Global and regional mortality from 235 causes of death for 20
or duration of metformin treatment on survival.
age groups in 1990 and 2010: a systematic analysis for the Global Bur-
Because of the retrospective study design, detailed den of Disease Study 2010. Lancet 2012;380:2095-2128.
information on use of antidiabetic agents was not 3. Wlazlo N, Beijers HJ, Schoon EJ, Sauerwein HP, Stehouwer CD,
always available. Thus, we were not able to determine Bravenboer B. High prevalence of diabetes mellitus in patients with
liver cirrhosis. Diabet Med 2010;27:1308-1311.
the minimum effective dose of metformin that resulted 4. Wlazlo N, van Greevenbroek M, Curvers J, Schoon EJ, Friederich P,
in improved survival of patients with cirrhosis. Metfor- Twisk J, et al. Diabetes mellitus at the time of diagnosis of cirrhosis is
min is the major antidiabetic agent with a biologically associated with higher incidence of spontaneous bacterial peritonitis,
but not with increased mortality. Clin Sci 2013;125:341-348.
plausible mechanism for improving liver function. 5. El-Serag HB, Tran T, Everhart JE. Diabetes increases the risk of
Other medications may also benefit patients with cir- chronic liver disease and hepatocellular carcinoma. Gastroenterology
rhosis; however, the sample size of the study was not 2004;126:460-468.2016 ZHANG ET AL. HEPATOLOGY, December 2014
6. El-Serag HB, Everhart JE. Diabetes increases the risk of acute hepatic 18. Bugianesi E, Gentilcore E, Manini R, Natale S, Vanni E, Villanova N,
failure. Gastroenterology 2002;122:1822-1828. et al. A randomized controlled trial of metformin versus vitamin E or
7. Butt Z, Jadoon NA, Salaria ON, Mushtaq K, Riaz IB, Shahzad A, prescriptive diet in nonalcoholic fatty liver disease. Am J Gastroenterol
et al. Diabetes mellitus and decompensated cirrhosis: risk of hepatic 2005;100:1082-1090.
encephalopathy in different age groups. J Diabetes 2013;5:449-455. 19. Garinis GA, Fruci B, Mazza A, De Siena M, Abenavoli S, Gulletta E,
8. Veldt BJ, Chen W, Heathcote EJ, Wedemeyer H, Reichen J, Hofmann WP, et al. Metformin versus dietary treatment in nonalcoholic hepatic stea-
et al. Increased risk of hepatocellular carcinoma among patients with hepati- tosis: a randomized study. Int J Obes 2010;34:1255-1264.
tis C cirrhosis and diabetes mellitus. HEPATOLOGY 2008;47:1856-1862. 20. Brackett CC. Clarifying metformin’s role and risks in liver dysfunction.
9. Quintana JO, Garcia-Compean D, Gonzalez JA, Perez JZ, Gonzalez J Am Pharm Assoc 2010;50:407-410.
FJ, Espinosa LE, et al. The impact of diabetes mellitus in mortality of 21. Brown JB, Pedula K, Barzilay J, Herson MK, Latare P. Lactic acidosis
patients with compensated liver cirrhosis-a prospective study. Ann Hep- rates in type 2 diabetes. Diabetes Care 1998;21:1659-1663.
atol 2011;10:56-62. 22. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal and
10. de Marco R, Locatelli F, Zoppini G, Verlato G, Bonora E, Muggeo M. nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus.
Cause-specific mortality in type 2 diabetes. The Verona Diabetes Study. Cochrane Database Syst Rev 2010;(4):CD002967.
Diabetes Care 1999;22:756-761. 23. Chan NN, Brain HP, Feher MD. Metformin-associated lactic acidosis:
11. Bo S, Benso A, Durazzo M, Ghigo E. Does use of metformin protect against a rare or very rare clinical entity? Diab Med 1999;16:273-281.
cancer in type 2 diabetes mellitus? J Endocrinol Invest 2012;35:231-235. 24. Edwards CM, Barton MA, Snook J, David M, Mak VH, Chowdhury
12. Chaiteerakij R, Yang JD, Harmsen WS, Slettedahl SW, Mettler TA, TA. Metformin-associated lactic acidosis in a patient with liver disease.
Fredericksen ZS, et al. Risk factors for intrahepatic cholangiocarci- QJM 2003;96:315-316.
noma: association between metformin use and reduced cancer risk. 25. Renda F, Mura P, Finco G, Ferrazin F, Pani L, Landoni G. Metformin-
HEPATOLOGY 2013;57:648-655. associated lactic acidosis requiring hospitalization. A national 10 year
13. Chen HP, Shieh JJ, Chang CC, Chen TT, Lin JT, Wu MS, et al. Met- survey and a systematic literature review. Eur Rev Med Pharmacol Sci
formin decreases hepatocellular carcinoma risk in a dose-dependent 2013;17(Suppl 1):45-49.
manner: population-based and in vitro studies. Gut 2013;62:606-615. 26. Misbin RI, Green L, Stadel BV, Gueriguian JL, Gubbi A, Fleming GA.
14. Donadon V, Balbi M, Mas MD, Casarin P, Zanette G. Metformin and Lactic acidosis in patients with diabetes treated with metformin. N
reduced risk of hepatocellular carcinoma in diabetic patients with Engl J Med 1998;338:265-266.
chronic liver disease. Liver Int 2010;30:750-758. 27. Kamber N, Davis WA, Bruce DG, Davis TM. Metformin and lactic
15. Nkontchou G, Cosson E, Aout M, Mahmoudi A, Bourcier V, Charif I, acidosis in an Australian community setting: the Fremantle Diabetes
et al. Impact of metformin on the prognosis of cirrhosis induced by Study. Med J Aust 2008;188:446-449.
viral hepatitis C in diabetic patients. J Clin Endocrinol Metab 2011; 28. Seidowsky A, Nseir S, Houdret N, Fourrier F. Metformin-associated
96:2601-2608. lactic acidosis: a rognostic and therapeutic study. Crit Care Med 2009;
16. Adams LA, Harmsen S, St Sauver JL, Charatcharoenwitthaya P, Enders 37:2191-2196.
FB, Therneau T, et al. Nonalcoholic fatty liver disease increases risk of
death among patients with diabetes: a community-based cohort study.
Am J Gastroenterol 2010;105:1567-1573. Supporting Information
17. Kita Y, Takamura T, Misu H, Ota T, Kurita S, Takeshita Y, et al. Met-
formin prevents and reverses inflammation in a non-diabetic mouse
Additional Supporting Information may be found at
model of nonalcoholic steatohepatitis. PloS One 2012;7:e43056. onlinelibrary.wiley.com/doi/10.1002/hep.27199/suppinfo.You can also read
