Original Articles
Prevalence of Obesity and Diabetes in Patients With Cryptogenic
Cirrhosis: A Case-Control Study
ASMA POONAWALA, SATHEESH P. NAIR, AND PAUL J. THULUVATH
It has recently been suggested that nonalcoholic steato-
hepatitis (NASH) is an under-recognized cause of crypto-
genic cirrhosis (CC) on the basis of higher prevalence of
obesity and type II diabetes among these patients. To test
this hypothesis, we studied 65 consecutive patients with
advanced cirrhosis (Child-Pugh Score > 7) of undetermined
etiology (CC) from our active waiting list for liver transplan-
tation in January 1993, 1996, and 1999. For each patient, we
selected 2 age- and sex-matched controls from the corre-
sponding lists. The prevalence of obesity (defined as body
mass index [BMI] > 30) and diabetes were compared be-
tween the groups. Sixteen patients (and their 32 controls)
with CC were excluded as further review of records sug-
gested other possible etiologies. Thus, the final analysis in-
cluded 49 patients and 98 controls. The etiology of cirrhosis
in the control group was alcohol in 16.3%, chronic viral
hepatitis in 30.6%, autoimmune hepatitis in 8.2%, and pri-
mary biliary cirrhosis (PBC) or primary sclerosing cholan-
gitis in 35.7%. The prevalence of obesity (55% vs. 24%) and
type II diabetes (47% vs. 22%) was significantly higher in
patients with CC compared with controls. Twenty-three
percent of patients with CC had both obesity and diabetes
compared with 5% among controls (P .002). There was no
difference in the prevalence of hypercholesterolemia (serum
cholesterol > 200 mg/dL) between the groups. In conclu-
sion, patients with advanced CC are more likely to be obese
and diabetic compared with age- and sex-matched patients
with advanced cirrhosis. This supports the hypothesis that
NASH may be an etiological factor in some of the patients
with CC. (HEPATOLOGY 2000;32:689-692.)
Cryptogenic cirrhosis (CC), a term applied to unexplained
chronic liver disease, has been reported to account for 3% to
31% of patients with end-stage liver diseases.1,2 Approxi-
mately 7% to 14% of liver transplantations are performed for
CC, and it is among the common 5 indications for orthotopic
liver transplantation.3 Several conditions have been proposed
Abbreviations: CC, cryptogenic cirrhosis; NASH, nonalcoholic steatohepatitis; HCV,
hepatitis C virus; PBC, primary biliary cirrhosis; ALT, alanine transaminase; ALP, alka-
line phosphatase; AST, aspartate transaminase; BMI, body mass index.
From the Department of Medicine, The Johns Hopkins University School of Medicine,
Baltimore, MD.
Received April 18, 2000; accepted June 29, 2000.
Address reprint requests to Paul J. Thuluvath, M.D., The Johns Hopkins Hospital,
Room 429, 1830 E. Monument Street, Baltimore, MD 21205. E-mail: pjthuluv@
welch.jhu.edu; fax: 410-614-9612.
Copyright 2000 by the American Association for the Study of Liver Diseases.
0270-9139/00/3204-0003$3.00/0
doi:10.1053/jhep.2000.17894
689
as a possible etiology of CC, including occult alcohol abuse,
unknown viral infections (non-A, non-B, non-C hepatitis), or
burnt-out autoimmune hepatitis. Recently, it was suggested
that nonalcoholic steatohepatitis (NASH) may be a common
cause of CC. Although NASH has certain clinical and histo-
logical characteristics, most of the histological markers disap-
pear when the disease, usually asymptomatic and therefore
undiagnosed, progresses to cirrhosis. Therefore, it is almost
impossible to make a diagnosis of NASH with confidence
when patients present with advanced cirrhosis. Risk factors
for NASH are type II diabetes, obesity, and hyperlipidemia.
Typical NASH patients are middle-aged females with the
above clinical profiles.4-10
In a recent study, a group of patients with CC were com-
pared with patients with known NASH, hepatitis C virus
(HCV), and primary biliary cirrhosis (PBC); patients with CC
and NASH were found to have a higher prevalence of obesity
and diabetes compared with PBC and HCV groups.11 On the
basis of this study, it was suggested that NASH may be a
predominant cause of CC. However, the groups used as con-
trols in this study were not age- or sex-matched. In addition, it
was not clear whether the severity of liver disease was compa-
rable between the groups. Moreover, because PBC and HCV
patients have certain well-defined, clinical characteristics, an
ideal disease control group should be representative of all
common liver diseases. The aim of our study was to compare
the prevalence of diabetes and obesity in patients with CC to
age- and sex-matched controls with a broad spectrum of other
chronic liver diseases, excluding patients with NASH.
PATIENTS AND METHODS
Patients. Patients with CC were identified from the Johns Hopkins
Hospital liver-transplantation registry. All patients who had been on
the active waiting list for liver transplantation in January 1993, 1996,
and 1999 were screened for the diagnosis of CC. Among 686 patients
active on the registry during the above time period, 65 were initially
found to have a diagnosis of CC.
Case-Control Group. The same liver-transplantation registry was
used for selection of case-controls. Patients who had NASH were
excluded. For each patient, 2 contemporary patients, age- (within
0-3 years), sex-, and race-matched (after excluding NASH), were
identified from the registry in a consecutive manner, with their di-
agnosis concealed. A total of 130 patients were selected as case-
controls in this manner. The patients were included only if sufficient
clinical data were available and if the diagnosis was confirmed after
reviewing updated information.
Data Collection. All patients had undergone thorough evaluation
by a hepatologist and had extensive investigations as a part of the
diagnostic work-up or during the evaluation for liver transplanta-
tion. Thus, data were adequate in accurately identifying the patients
690 POONAWALA, NAIR, AND THULUVATH HEPATOLOGY October 2000

with CC. All patients were evaluated by a substance counselor, a TABLE 1. Demographics of Patients With CC and Case-Controls
psychologist, and a social worker for ongoing or previous use of Case-
drugs or alcohol as a part of the liver-transplantation evaluation. The CC Control
following information was recorded for each patient: demographics, (N 49) (N 98) P
history of alcohol or intravenous drug use, family history of liver
disease, and history of blood transfusions. The laboratory data col- Age (yr) 54 10.1 52.6 9.6 NS
lected included serum alanine transaminase (ALT), alkaline phos- Sex (female) 55% 55% NS
phatase (ALP), aspartate transaminase (AST), hepatitis B surface Obesity (BMI 30)* 47% 24.4% .008
antigen, hepatitis B core antibody, HCV reverse-transcriptase poly- Severe obesity (BMI 35) 22% 10% .05
merase chain reaction, hepatitis C antibody, ferritin, transferrin sat- Type II diabetes mellitus 47% 22% .002
uration, antinuclear antibody titers, anti-smooth muscle antibody Obesity diabetes mellitus 23% 5% .001
titers, and antimitochondrial antibody titers. Serum ceruloplasmin High cholesterol (200 mg/dL) 21% 32% NS
levels, 1-antitrypsin phenotype or titers, and quantitative immuno- AST 60.4 57 86.1 58 .02
globulin levels (IgG, IgA, IgM) were recorded. Pretransplantation ALT 42.1 34 76 66 .03
biopsy reports were reviewed, and evidence for fatty liver was re- AST/ALT ratio 1.6 0.6 1.4 0.6 NS
corded. Height and weight were recorded for all patients during ALP 179 196 279 267 .02
pretransplantation evaluation, and these values were used to calcu- Steatosis on liver histology 10% 7% NS
late the body mass index (BMI). Obesity was defined as BMI 30, Abbreviation: NS, not significant.
and severe obesity as BMI 35.12 Patients with BMI between 25 and *BMI weight in kg/(height in meters)2.
29.9 were considered overweight. We determined the prevalence of
diabetes mellitus in patients with CC and case-controls by reviewing
all pertinent records. The type of diabetes and duration of diabetes
were noted; the type of diabetes was determined from the records of ritin and iron saturation were measured in all patients and
the patients primary care physician. All patients, except 1, had type were within normal limits. All patients had advanced liver
II diabetes. However, the duration of diabetes was not obtained for disease (Child B or C) with 1 or more complications, and all
many patients. The cholesterol and triglyceride values were obtained had satisfied minimal listing UNOS criteria for liver transplan-
from records of either primary care physicians or Johns Hopkins tation.
Hospital. The fasting cholesterol levels were available for 117 (of
147) patients, but fasting triglyceride levels were available in only 31
Table 1 shows the ALT, AST, and ALP values of patients at
patients. the time of listing for transplantation. The AST was signifi-
Statistical Analysis. The continuous variables were compared using cantly higher than ALT, with a mean AST/ALT ratio of 1.6.
the 2-tailed Students t test. Categorical variables were compared The etiology of liver disease in 98 case-controls (2 controls for
with Fishers exact test. All analyses were performed using the SPSS each patient with CC) was alcohol in 16.3%, HCV or HBV in
(SPSS Inc., Chicago, IL) program on a personal computer. 30.6%, autoimmune in 8.2%, PBC or primary sclerosing
cholangitis in 35.7%, and others in 9.2%. As defined by our
RESULTS protocol, none of the case-controls had NASH or CC. The
After careful review of all medical records, including the average age of case-controls was 52.6 9.6 years, similar to
work-up for liver transplantation, 16 patients who were orig- patients with CC. Because controls were also sex-matched,
inally classified as CC in the liver-transplantation registry the male:female ratio was identical to that of CC. Control
were found to have other causes of liver disease. The main patients had significantly higher AST, ALT, and ALP values
reason for this discrepancy was because of incomplete inves- compared with patients with CC (Table 1).
tigations or erroneous interpretations of test results when the The mean BMI of patients with CC was 30.2 5.9, and for
patients were referred to our center for liver transplantation. case-controls, it was 27.8 5.6. Prevalence of obesity in the
These patients were initially listed as CC, but the diagnosis CC group was 47% compared with 24% in controls (P
was not corrected in the registry when new information was .008); 22% of patients with CC were severely obese compared
available. Other less-common reasons for exclusions were in- with 10% among case-controls. Diabetes (all except 1 patient
complete medical information and indeterminate test results. had type II diabetes) was more common in patients with CC
The age- and sex-matched controls who were selected for the (47% vs. 22% in controls; P .002); all patients had diabetes
16 excluded patients were also discarded. For final analysis, before the development of liver failure. Twenty-three percent
we had 49 patients with CC and 98 case-controls. The clinical of patients with CC had both diabetes and obesity compared
profile of the 49 patients with CC and the 98 controls is shown with 5% in the case-control group (P .001). Because only a
in Table 1. The mean age of patients with CC was 54 10.1 few patients had triglyceride values, these results were not
years; 55% of patients with CC were female. Three patients analyzed. However, 84% of controls and 69% of patients with
had moderate alcohol consumption (2 drinks/day), but this CC had cholesterol values. The amount of patients with CC
was not considered to be the cause of liver disease or signifi- who had high cholesterol levels was 20.6% compared with
cant either by the hepatologists or their primary care physi- 32.5% of the control group (P not significant).
cians. Only 1 patient had a history of remote intravenous drug
use. However, serology for HCV and hepatitis B virus was DISCUSSION
negative in this patient. Eleven patients had a history of blood In this study, we compared the prevalence of diabetes and
transfusions, but none of them had past or present evidence of obesity in patients with CC and a carefully matched case-
exposure to HCV or hepatitis B virus. Seven patients had a control group. The prevalence of obesity and diabetes was
family history of liver disease. A positive antinuclear antibody higher in patients with CC compared with age- and sex-
was seen in 14 patients (28%), but no evidence of autoim- matched case-controls. Our observations support the hypoth-
mune hepatitis was seen in their liver biopsies. 1-antitrypsin esis that a significant proportion of patients with CC may have
phenotype was normal in all patients with CC. Similarly, fer- advanced from occult NASH.
HEPATOLOGY Vol. 32, No. 4, 2000
In an important study, Caldwell et al. characterized the
clinical profile of patients with CC, NASH, PBC, and HCV.11
Diabetes and obesity were more common in patients with CC
and NASH compared with either PBC or HCV. On the basis of
their observation, the authors suggested that NASH is an un-
recognized cause of CC. The results of our study corroborate
their observations. However, our study was distinct in several
aspects, the most important of which was the rigid criteria that
we used for the selection of case-controls. We believe that this
is the most important aspect of any retrospective study that
compares one group of patients with another. In the study by
Caldwell et al., patients with PBC were at least 10 years
younger than patients with CC.11 Because the incidence of
obesity and diabetes increases with age, this difference alone
could have explained the differences in the prevalence of obe-
sity and diabetes, especially in a group of predominantly mid-
dle-aged females. In addition, it was not clear from their study
whether the patients and controls had comparable severity of
liver disease, which could have had an influence on muscle
mass as well as glucose homeostasis.
The second control group Caldwell et al. used was a group
of patients with cirrhosis caused by HCV who were older than
age 50. By selecting a subgroup of relatively older patients
with HCV, the authors could have skewed their observations.
It is equally probable that these patients may have had more
advanced liver disease. To reduce these variables, we selected
a case-control group with comparable severity of liver disease
(Child-Pugh score 7) who were matched for age and sex.
The case-control group was selected after blinding the diag-
noses, thus avoiding the confounding effect of etiology of liver
disease, age, and sex in the prevalence of obesity and diabetes.
Despite these rigorous criteria, our observations were similar
to that of Caldwell et al. This supports the hypothesis that
many patients with CC may have progressed from undiag-
nosed NASH.
There is a broad consensus that diabetes and obesity are risk
factors for NASH. It is variously estimated that at least 10% to
16% of NASH patients ultimately progresses to cirrhosis.13-15
In a retrospective study with long-term follow-up, it has been
shown that liver-related mortality was higher in patients with
histologically advanced NASH.16 In the United States, 25% of
the population is obese.17 The prevalence of obesity in our
case-control group was 25%, which is similar to the national
average. In contrast, 47% of patients with CC were obese. The
prevalence of obesity in our patients with CC was lower than
that reported in Caldwells series, and this is possibly the
result of the fact that our patients were at least 10 years young-
er.11 A more striking finding was the difference in the preva-
lence of diabetes between patients with CC and patients with
cirrhosis caused by other etiologies. Type II diabetes is seen in
12% of the American population between the ages of 45 and
74.18 The presence of diabetes in our cryptogenic patients was
4 times the expected national rate. The prevalence of diabetes
was higher than the national average among case-controls,
which also may well be related to the impaired glucose ho-
meostasis seen in patients with moderate to severe cirrhosis.
Liver histology in our patients with CC did not reveal ex-
cessive fat accumulation (steatosis). In fact, there were no
significant histological findings that separated obese from
nonobese CC. This was not an unexpected observation, be-
cause fat often disappears as the disease progresses from
POONAWALA, NAIR, AND THULUVATH 691
NASH to cirrhosis.14,19 This could be part of the natural evo-
lution or adaptation, as seen in alcoholic liver disease and
hepatitis C, of advanced liver disease. In patients with NASH,
shunting of portal blood, decreased delivery of dietary fat into
the liver, reduced calorie intake, and sinusoidal capillariza-
tion are possible mechanisms by which steatosis disappears as
liver disease progresses.20-22
The diagnosis of CC was made after exclusion of all known
etiologies. In our study, all patients were seen by experienced
hepatologists, and all relevant investigations were performed
to exclude known causes of liver diseases. It is also important
to note that the hepatologists in our institution have a rela-
tively low threshold for diagnosing NASH in patients with
cirrhosis if the clinical profile was consistent with that diag-
nosis even in the absence of histological confirmation. Only
8% of our patients had a clinical diagnosis of CC. This makes
our observations even more significant.
In summary, there was a higher prevalence of obesity and
diabetes in patients with CC compared with age- and sex-
matched patients with other causes of cirrhosis. The preva-
lence was much higher than the national prevalence of obesity
and diabetes. Our findings support the hypothesis that NASH
is one of the important causes of CC.
REFERENCES
1. Byron D, Minuk GY. Clinical hepatology: profile of an urban, hospital-
based practice. HEPATOLOGY 1996;24:813-815.
2. Greeve M, Ferrell L, Kim M, Combs C, Roberts J, Ascher N, Wright TL.
Cirrhosis of undefined pathogenesis: absence of evidence for unknown
or autoimmune processes. HEPATOLOGY 1993;17:593-598.
3. Belle SH, Beringer KC, Detre KM. An update of liver transplantation in
the United States: recipient characteristics and outcome. In: Cecka JM,
Terasaki PI, eds. Clinical Transplants 1995. Los Angeles: UCLA Tissue
Typing Laboratory, 1996.
4. Ludwig J, Viggiano TR, McGill DB, Oh BJ. Nonalcoholic steatohepatitis.
Mayo Clin Proc 1980;55:434-438.
5. Diehl AM, Goodman Z, Ishak KG. Alcohol-like liver disease in non alco-
holics. A clinical and histopathological comparison with alcohol-induced
liver injury. Gastroenterology 1988;95:1056-1062.
6. Adler M, Schaffner F. Fatty liver hepatitis and cirrhosis in obese patients.
Am J Med 1979;67:811-816.
7. Hay JE, Czaja AJ, Rakela J, Ludwig J. The nature of unexplained chronic
aminotransferase elevations of a mild to moderate degree in asymptom-
atic patients. HEPATOLOGY 1989;9:193-197.
8. Pinto HC, Baptista A, Camilo ME, Valente A, Saragoca A, de Moura MC.
Nonalcoholic steatohepatitis. Clinicopathological comparison with alco-
holic hepatitis in ambulatory and hospitalized patients. Dig Dis Sci 1996;
41:172-179.
9. Sheth SG, Gordon FD, Chopra S. Nonalcoholic steatohepatitis. Ann In-
tern Med 1997;126:137-145.
10. Wanless IR, Lentz JS. Fatty liver hepatitis (steatohepatitis) and obesity:
an autopsy study with analysis of risk factors. HEPATOLOGY 1990;12:1106-
1110.
11. Caldwell SH, Oelsner DH, Iezzoni JC, Hespenheide EE, Battle EH,
Driscoll CJ. Cryptogenic cirrhosis: clinical characterization and risk fac-
tors for underlying disease. HEPATOLOGY 1999;29:664-669.
12. National Research Council. Diet and health, implications for reducing
chronic disease risk. National Research Council. Washington, DC: Na-
tional Academy Press, 1989:99-135.
13. Lee RG. Nonalcoholic steatohepatitis. A study of 49 patients. Hum Pathol
1989;20:594-598.
14. Powell EE, Cooksley GE, Hanson R, Searle J, Halliday JW, Powell LW.
Natural history of nonalcoholic steatohepatitis: a follow up study of 42
patients followed for up to 21 years. HEPATOLOGY 1990;11:74-80.
15. Bacon B, Farahvash MJ, Janney CG, Neuschwander-Tetri BA. Non alco-
holic steatohepatitis: an expanded clinical entity. Gastroenterology 1994;
107:1103-1106.
16. Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough
AJ. Nonalcoholic fatty liver disease: a spectrum of clinical and patholog-
ical severity. Gastroenterology 1999;116:1413-1419.
692 POONAWALA, NAIR, AND THULUVATH
17. Kral JG, Heymsfield S. Morbid obesity definitions, epidemiology and
methodological problems. Gastroenterol Clin North Am 1987;16:197-205.
18. Kenny SJ, Aubert RE, Geiss LS. Prevalence and incidence of non-insulin
dependant diabetes. In: Diabetes in America. 2nd ed. Bethesda, MD:
National Institutes of Health, 1995:47-67.
19. Abdelmalek M, Ludwig J, Lindor KD. Two cases from the spectrum of
nonalcoholic steatohepatitis. J Clin Gastroenterol 1995;20:127-130.
20. Schaffner F, Popper H. Capillarization of hepatic sinusoids. Gastroenter-
ology 1963;44:239-242.
HEPATOLOGY October 2000
21. Nosadini R, Avogaro A, Mollo F, Maraescotti C, Tiengo A, Duner E,
Merkel C, et al. Carbohydrate and lipid metabolism in cirrhosis. Evi-
dence that hepatic uptake of gluconeogenic precursors and of free fatty
acid depends on effective hepatic flow. J Clin Endocrinol Metab 1984;58:
1125-1132.
22. Matsui O, Kadoya M, Takahashi S, Yoshikawa J, Gabata T, Takashima T,
Kitagawa K. Focal sparing of segment IV in fatty livers shown by sonog-
raphy and CT: correlation with aberrant gastric venous drainage. AJR
Am J Roentgenol 1995;164:1137-1140.