952
ORIGINAL CONTRIBUTIONS
nature publishing group
LIVER
see related editorial on page 959
Relationship of Non-alcoholic Fatty Liver Disease With
Cholecystectomy in the US Population
Constance E. Ruhl, MD, PhD1 and James E. Everhart, MD, MPH2
OBJECTIVES:
Other than weight-related conditions, risk factors for non-alcoholic fatty liver disease (NAFLD) are
not well defined. We investigated the association of gallstones and cholecystectomy with NAFLD
in a large, national, population-based study.
METHODS:
Among adult participants in the third US National Health and Nutrition Examination Survey,
1988–1994, ultrasonography for gallstone disease was performed, and videotapes were subsequently
evaluated for NAFLD. Odds ratios (ORs) for the association of gallstone disease with NAFLD were
calculated using logistic regression analysis to adjust for common associated factors.
RESULTS:
Among 12,232 participants without viral hepatitis or significant alcohol intake, the prevalence of
gallstones was 7.4%, cholecystectomy 5.6%, and NAFLD 20.0%. Participants with cholecystectomy
had higher age–sex-adjusted prevalence of NAFLD (48.4%) than those with gallstones (34.4%) or
without gallstone disease (17.9%) (P < 0.01 for all comparisons). Controlling for numerous factors
associated with both NAFLD and gallstone disease, multivariate-adjusted analysis confirmed the
association of NAFLD with cholecystectomy (OR = 2.4; 95% confidence interval (CI): 1.8–3.3),
but not with gallstones (OR = 1.1; 95% CI: 0.84–1.4).
CONCLUSIONS: The association of NAFLD with cholecystectomy, but not with gallstones, indicates that cholecystec-
tomy may itself be a risk factor for NAFLD.
SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg
Am J Gastroenterol 2013; 108:952–958; doi:10.1038/ajg.2013.70; published online 2 April 2013
INTRODUCTION
Cholecystectomy is a frequent surgical procedure that contributes
significantly to health-care costs. However, long-term medical
consequences of cholecystectomy are believed to be uncommon
(1–3). Non-alcoholic fatty liver disease (NAFLD) is an increasingly common condition whose risk factors, other than weightrelated conditions, are not well defined. Although cholelithiasis
was noted to be common among the original case series of nonalcoholic steatohepatitis patients published in 1980 (4), few subsequent studies have investigated a possible association between
NAFLD and gallstone disease. Even fewer have examined cholecystectomy and gallstones separately in relation to NAFLD (5–13).
It is possible that having had a cholecystectomy might increase
the risk of NAFLD through the metabolic effects of the loss of a
gallbladder. Biologically plausible mechanisms exist in the alteration of the enterohepatic circulation of bile acids following cholecystectomy (14) and in the loss of the metabolic activity of the
gallbladder mucosa in the absence of a gallbladder (15). An effect
of loss of the gallbladder on triglyceride metabolism has been
demonstrated in mice in whom cholecystectomy resulted in an
increase in hepatic fat (16). Therefore, we examined the association of gallstone disease with NAFLD in a large, national, population-based study with an emphasis on whether cholecystectomy
had a stronger association than gallstones with NAFLD.
METHODS
The third National Health and Nutrition Examination Survey
(NHANES III) was conducted in the United States from 1988 to
1994 by the National Center for Health Statistics of the Centers
for Disease Control and Prevention (17). The survey consisted
of cross-sectional interview, examination, and laboratory data
collected from a complex multistage, stratified, clustered probability sample representative of the civilian, non-institutionalized
1
Social and Scientific Systems, Inc., Silver Spring, Maryland, USA; 2Department of Health and Human Services, National Institute of Diabetes and Digestive and
Kidney Diseases, National Institutes of Health, Bethesda, Maryland, USA. Correspondence: Constance E. Ruhl, MD, PhD, Social and Scientific Systems, Inc.,
8757 Georgia Avenue, 12th Floor, Silver Spring, Maryland 20910, USA. E-mail: [email protected]
Received 26 June 2012; accepted 27 December 2012
The American Journal of GASTROENTEROLOGY
VOLUME 108 | JUNE 2013 www.amjgastro.com
NAFLD and Cholecystectomy
© 2013 by the American College of Gastroenterology
Sampled persons 20–74 years
18,738
Interviewed
16,115
No interview
2,623
Mobile Examination Center exam
14,645
No exam (1,318) or home exam (152)
No viral hepatitis or
iron overload or
significant alcohol consumption
13,317
Hepatitis B or C (428) or
transferrin saturation >50% (422) or
alcoholic drinks/day >3 for men or >2
for women (478)
No type 1 diabetes,
men & non-pregnant women
13,021
Type 1 diabetes (21) or
pregnant woman (275)
Gallbladder ultrasound
12,653
Ultrasound missing (318) or
inconclusive for gallstone disease (50)
Analysis sample
Hepatic steatosis image
12,232
(5,569 men, 6,663 women)
Image missing (307) or
ungradable for hepatic steatosis (114)
Figure 1. The third National Health and Nutrition Examination Survey
(NHANES III) sample for analysis of non-alcoholic fatty liver disease
(NAFLD) and gallstone disease.
(cups/day; 0, > 0– < 1, 1–2, and > 2), body mass index (weight
(kg)/height (m2)), waist and hip circumferences (cm), systolic
and diastolic blood pressure (mm Hg), and serum concentrations
of hemoglobin A1C (%; < 6.5, ≥6.5) (22), total and high-density
lipoprotein cholesterol (mg/dl), and C-reactive protein (mg/dl;
0–0.3, > 0.3) (23–26). Participants were asked about right upper
quadrant or epigastric pain in the past 12 months. Among the subgroup of participants who attended a morning examination after
an overnight fast, concentrations of plasma glucose (mg/dl) and
serum triglycerides (mg/dl), insulin (pmol/l), C-peptide (pmol/ml),
and leptin (μg/l) were measured. Insulin resistance and insulin
secretion indices were calculated using the homeostasis model
assessment (27). Metabolic syndrome was defined according to the
National Cholesterol Education Program Adult Treatment Panel
III criteria (28).
Statistical analysis
Characteristics and prevalence of NAFLD age and sex standardized to the 1990 US population distribution were generated for
participants with normal gallbladder, all gallstone disease, gallstones, and cholecystectomy. The unadjusted prevalence of NAFLD
by gallstone disease status was compared using a χ2-test. Odds
ratios (ORs) for NAFLD, comparing persons with gallstone disease
with those without, were calculated using logistic regression
analysis (SUDAAN, PROC LOGISTIC, SUDAAN User’s Manual,
Release 10.0, 2008; Research Triangle Institute, Research Triangle
The American Journal of GASTROENTEROLOGY
LIVER
population with oversampling of persons aged 60 years and older,
African Americans, and Hispanics. The survey was approved
by the Centers for Disease Control and Prevention Institutional
Review Board, and all participants provided written informed
consent to participate.
Gallstone disease was ascertained among NHANES participants aged 20–74 by ultrasound and defined as ultrasounddocumented gallstones or evidence of a cholecystectomy (a right
upper quadrant or epigastric scar and the absence of a gallbladder) by standard criteria (18). The diagnosis of gallstones required
finding an echogenic region within the gallbladder and acoustic
shadowing reproducible in two views (longitudinal and transverse)
(19). Based on videotaped recordings of ultrasound examinations,
there was excellent agreement on gallbladder disease diagnosis
between the ultrasonographer and reviewing radiologist (agreement of 99% with a kappa statistic of 0.97).
NAFLD was ascertained by ultrasound. Archived videotapes of
gallbladder ultrasounds were reviewed in 2009–2010 to assess the
presence of fat within the hepatic parenchyma using the standard
criteria (20,21). Evaluation of hepatic steatosis was performed
using five criteria: parenchymal brightness, liver to kidney contrast,
deep beam attenuation, bright vessel walls, and gallbladder wall
definition. Readers under the supervision of a radiologist who was
an expert in ultrasonography evaluated each scan. Hepatic steatosis was initially categorized as normal, mild, moderate, or severe
and then recoded as normal-mild or moderate-severe, which was
the classification used for the current analysis. For the two-level
hepatic steatosis categorization, the percent agreement for intraand inter-rater reliability was 91% (with a kappa statistic (95%
confidence interval (CI)) of 0.77 (0.73–0.82)) and 89% (kappa,
0.70 (0.64–0.76)), respectively.
Among the 14,645 participants who attended an examination at
a mobile examination center, we excluded those with hepatitis B
(serum hepatitis B surface antigen positive; n = 68) or hepatitis C
(serum hepatitis C antibody positive; n = 360), iron overload (serum
transferrin saturation > 50%; n = 422), significant alcohol consumption ( > 3 drinks/day for men or > 2 drinks/day for women; n = 478),
type 1 diabetics (n = 21), and pregnant women (n = 275) (Figure 1).
We also excluded persons who did not undergo a gallbladder ultrasound (n = 318) or whose gallbladder lumen could not be adequately
visualized on ultrasound (n = 50), and those who had no ultrasound
available for assessment of hepatic steatosis (n = 307) or whose
ultrasound was ungradable (n = 114), leaving an analysis sample
of 12,232 participants. A secondary analysis of fasting laboratory
measures was conducted among 5,012 participants without known
diabetes who were randomly assigned to be examined in the morning after an overnight fast. Excluded were 919 participants who
missed the morning examination or who fasted for < 9 or ≥24 h.
Data were collected, as previously described, on factors believed
to be related to NAFLD or gallstone disease: age (years), sex, raceethnicity (non-Hispanic white, non-Hispanic black, Mexican
American, and other), cigarette smoking (never or former, < 1 pack/
day, ≥1 pack/day), alcohol intake (never or former, < 1 drink/day, 1–2
drinks/day, or > 2 drinks/day), doctor-diagnosed diabetes, physical activity intensity (metabolic equivalents), coffee consumption
953
LIVER
954
Ruhl and Everhart
Park, NC) to adjust for common associated factors. Men and
women were examined both combined and separately because of
sex differences in the prevalence of NAFLD and gallstone disease.
Multivariate-adjusted analyses excluded persons with missing values for any factor included in the model. A P-value of < 0.05 was
considered to indicate statistical significance. All analyses utilized
sample weights and design effects of the survey (29).
RESULTS
Among the 12,232 participants with ultrasound readings for both
NAFLD and gallstone disease, the prevalence (±s.e.) of gallstone
disease was 13.0% (±0.66%); 7.4% (±0.46%) for ultrasound-diagnosed
gallstones and 5.6% (±0.31%) for a history of a cholecystectomy.
Because participants with gallstone disease were more likely to be
older and women (Table 1), relationships with other characteristics were examined standardized for age and sex. In contrast to a
normal gallbladder, gallstone disease was associated with MexicanAmerican ethnicity, less education, higher body mass index and
waist circumference, diabetes, lower total and high-density lipoprotein cholesterol levels, elevated C-reactive protein, lower alcohol
intake, less physical activity, right upper quadrant or epigastric
pain, higher concentrations of C-peptide, leptin, and triglycerides,
insulin resistance, and the metabolic syndrome. Compared with
the presence of gallstones, having had a cholecystectomy was associated with female sex, less education, higher body mass index and
waist circumference, diabetes, higher systolic and diastolic blood
pressure, elevated C-reactive protein, right upper quadrant or epigastric pain, higher concentrations of total cholesterol, C-peptide
and leptin, insulin resistance, and the metabolic syndrome.
The overall prevalence (±s.e.) of NAFLD was 20.0% (±0.81%)
and was twice as high with gallstone disease (35.7% (±2.0%)) compared with no gallstone disease (17.7% (±0.82%), P < 0.001). This
relationship held with adjustment for age (Table 2). After adjusting
for additional factors associated with NAFLD or gallstone disease,
NAFLD was found almost 60% more often among persons with
gallstone disease (Table 2).
With gallstone disease split between ultrasound-detected gallstones or cholecystectomy, the unadjusted prevalence (±s.e.) of
NAFLD was 27.1% (±1.9%) with gallstones and 47.1% (±3.4%)
with cholecystectomy (P < 0.001). For both sexes, the age-standardized prevalence of NAFLD was higher among those who had
had cholecystectomy than those who had ultrasound-diagnosed
gallstones (Figure 2). Men with a history of cholecystectomy had a
particularly high prevalence of NAFLD (68%). With adjustment for
age, persons with gallstones had modestly higher odds of NAFLD,
whereas those with cholecystectomy were over three times as likely
to have NAFLD, compared with persons without gallstone disease
(Table 2). In multivariate-adjusted analyses, gallstones were no
longer independently related to NAFLD. In contrast, persons with
a cholecystectomy were still almost two and half times as likely
to have NAFLD (Table 2). In sex-specific analyses, stronger associations were also found for NAFLD with cholecystectomy than
with gallstones (Table 2). To evaluate the robustness of our findings, we conducted a sensitivity analysis in which we included
The American Journal of GASTROENTEROLOGY
participants with viral hepatitis, iron overload, or significant alcohol
consumption ( > 3 drinks/day for men or > 2 drinks/day for
women). The results changed little with these modifications to the
analysis sample (data not shown). An additional analysis was conducted in which we limited the sample to 985 participants with
more than moderate alcohol intake ( > 2 drinks per day for men
or > 1 drink per day for women). Among this relatively small subgroup, there was no overall association of gallstone disease with
hepatic steatosis: OR 1.01 (95% CI: 0.31–3.27; P = 0.99). However,
consistent with the primary analysis, there was a trend for an
association of cholecystectomy with steatosis (OR = 2.59; 95% CI:
0.44–15.21, P = 0.29) that was not seen for ultrasound-diagnosed
gallstones (OR 0.51; 95% CI: 0.16–1.63, P = 0.25).
A secondary analysis was conducted among a random subgroup
of 5,012 participants examined in the morning after an overnight
fast, adjusting for insulin-resistance index in place of glucose status and for serum triglyceride concentrations in addition to the
factors adjusted for in the main analysis. There was little change
in the relationship of NAFLD with gallstone disease with adjustment for these additional factors (Supplementary Table 1 online).
The results were similar when a metabolic syndrome variable was
substituted for its components (data not shown).
Because persons with right upper quadrant or epigastric pain
associated with NAFLD might be more likely to seek medical care
and, if gallstones were present, to undergo a cholecystectomy,
multivariate models were adjusted for pain. In addition, we evaluated the relationship of NAFLD with gallstones and cholecystectomy stratified by the presence or absence of right upper quadrant
or epigastric pain. If nonspecific pain associated with NAFLD
increased the risk of cholecystectomy, then a stronger association
for cholecystectomy and NAFLD would have been expected among
participants with pain than among participants without pain.
We found the reverse; that is, a high OR for cholecystectomy, but
not gallstones among participants without pain and comparable
ORs for cholecystectomy and gallstones among participants with
pain (Table 3). These divergent associations were further evaluated by adding interaction terms of pain with gallstones or with
cholecystectomy to the models for gallstone disease. A negative
interaction was found (P = 0.001 for both age-adjusted analysis
and multivariate-adjusted analysis), meaning that participants
with cholecystectomy and pain had a lower OR for NAFLD than
participants with cholecystectomy and no pain.
Participants with a history of cholecystectomy who provided an
age at surgery were further categorized based on time since surgery
of < 10 years (n = 336) or ≥10 years (n = 416). The odds of NAFLD
did not differ based on time since cholecystectomy in either ageadjusted (P = 0.98) or multivariate-adjusted analyses (P = 0.93).
DISCUSSION
In this large, national, population-based study, we found that
NAFLD was associated with a history of cholecystectomy, but not
ultrasound-diagnosed gallstones. Specifically, in multivariate analysis, NAFLD was unrelated to gallstones independent of other factors. In contrast, study participants with a cholecystectomy were two
VOLUME 108 | JUNE 2013 www.amjgastro.com
NAFLD and Cholecystectomy
955
No gallstone disease
(N=10,417)
Total population (%)
All gallstone disease
(N=1,815)
Gallstones
(N=1,020)
Cholecystectomy
(N=795)
87.0
13.0
7.4
5.6
Non-alcoholic fatty liver disease (%±s.e.)
17.9±0.79
40.8±2.8*
34.4±2.3*
48.4±4.0*,†
Age (years; mean±s.e.)
41.1±0.39
53.6±0.61*
53.3±0.73*
54.1±0.68*
Women (%±s.e.)
50.3±0.52
69.9±1.8*
63.1±2.3*
79.0±2.5*,†
Non-Hispanic white
75.4±1.2
75.9±2.5
75.6±2.7
77.7±3.2
Non-Hispanic black
11.3±0.68
9.4±1.3
9.8±1.5
9.1±1.9
Race ethnicity (%±s.e.)
5.2±0.43
7.9±1.5*
8.5±1.9*
7.3±1.5
Education (years; mean±s.e.)
Mexican American
12.5±0.084
12.2±0.14*
12.5±0.17
11.6±0.16*,†
BMI (kg/m2; mean±s.e.)
26.3±0.10
28.9±0.60*
28.0±0.38*
30.8±0.93*,†
Waist circumference (cm; mean±s.e.)
90.9±0.22
98.0±1.3*
95.7±0.79*
102.2±1.9*,†
Glucose status abnormal (%±s.e.)b
5.3±0.40
8.1±0.87*
6.6±0.91
10.7±1.5*,†
Serum total cholesterol (mg/dl; mean±s.e.)
203±0.79
199±1.7*
197±2.0*
205±2.0†
Serum HDL cholesterol (mg/dl; mean±s.e.)
50.2±0.37
46.6±0.74*
47.1±0.73*
47.4±1.0*
Systolic blood pressure (mm Hg; mean±s.e.)
121±0.22
122±0.88*
121±0.73
125±1.1*,†
Diastolic blood pressure (mm Hg; mean±s.e.)
74.4±0.19
75.0±0.80
74.7±0.63
76.6±0.83*,†
C-reactive protein > 0.3 mg/dl (%±s.e.)
23.2±0.90
34.9±2.6*
29.5±2.1*
43.5±3.4*,†
Never/former
72.1±0.87
72.0±3.6
75.2±3.0
69.4±4.2
< 1 pack/day
12.8±0.59
7.8±1.1*
7.9±1.4*
10.0±2.2
≥1 pack/day
15.1±0.73
20.2±3.3
16.9±2.4
20.7±3.8
Never/former
44.0±1.2
51.8±2.9*
50.6±3.4*
56.2±3.5*
Cigarette smoking (%±s.e.)
Alcohol drinking (%±s.e.)
< 1 drink/day
43.6±1.1
41.1±2.7
38.5±3.3
39.5±3.4
1–2 drinks/day
9.9±0.61
6.6±1.9
10.2±1.9
4.1±1.4*,†
> 2 drinks/day
2.5±0.22
0.52±0.23*
0.72±0.39*
0.32±0.33*
0
34.7±0.91
34.1±3.3
34.6±2.6
35.7±4.1
> 0– < 1
15.2±0.74
11.4±1.7
10.9±2.0
11.5±1.8
1–2
25.9±0.81
29.1±2.4
28.6±2.4
28.2±1.9
>2
24.2±0.71
25.5±4.0
26.0±2.1
24.6±4.8
Physical activity (METs; mean±s.e.)
115.2±3.6
74.8±4.9*
78.2±5.0*
85.2±8.4*
Right upper quadrant or epigastric pain (%±s.e.)
15.6±0.71
34.0±3.0*
27.1±2.5*
40.8±3.2*,†
Fasting plasma glucose (mg/dl; mean±s.e.)c
96.5±0.33
98.0±0.79
97.9±0.91
99.9±1.1*
667±9.2
846±36.3*
821±36.4*
959±35.0*,†
10.8±0.25
14.4±0.69*
13.2±0.72*
19.3±1.1*,†
2.4±0.059
3.1±0.15*
2.8±0.13*
3.9±0.14*,†
113±2.4
136±5.9*
124±4.3*
163±5.0*,†
Coffee drinking (cups/day; %±s.e.)
Fasting serum C-peptide (pmol/ml; mean±s.e.)
c
Fasting serum leptin (μg/l; mean±s.e.)c
c
Fasting insulin resistance index (mean±s.e.)
Fasting insulin secretion index (mean±s.e.)c
c
Fasting serum triglycerides (mg/dl; mean±s.e.)
Metabolic syndrome (%±s.e.)c
131±2.4
144±4.1*
136±5.3
147±5.0*
19.2±0.76
34.1±2.9*
23.5±2.8
43.7±3.2*,†
BMI, body mass index; HDL, high-density lipoprotein; MET, metabolic equivalent.
*P < 0.05 compared with no gallstone disease; †P < 0.05 compared with gallstones.
a
Mean age and % women were not standardized.
b
Doctor-diagnosed diabetes or hemoglobin A1C≥6.5%.
c
Measured in non-diagnosed diabetics who were examined in the morning after an overnight fast (N=5,012; 4,308 with no gallstone disease, 704 with gallstone disease,
400 with gallstones, and 304 with cholecystectomy).
© 2013 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
LIVER
Table 1. Age- and sex-standardizeda characteristics of participants by gallstone disease status
956
Ruhl and Everhart
Age-adjusted (N=12,232; 5,569 men, 6,663 women)
OR
95% CI
P-value
Multivariate-adjusteda (N=10,752; 4,852 men, 6,052 women)
OR
95% CI
P-value
All
Gallstone diseaseb
2.06
1.71–2.48
< 0.001
1.56
1.27–1.91
< 0.001
Gallstones
1.38
1.11–1.71
0.004
1.09
0.84–1.42
0.49
Cholecystectomy
3.29
2.50–4.34
< 0.001
2.43
1.81–3.28
< 0.001
2.45
1.92–3.14
< 0.001
2.24
1.66–3.02
< 0.001
Gallstones
1.72
1.24–2.40
0.002
1.73
1.21–2.48
0.003
Cholecystectomy
5.54
3.59–8.55
< 0.001
4.07
2.63–6.30
< 0.001
2.14
1.68–2.72
< 0.001
1.35
1.05–1.74
0.022
Men
Gallstone disease
Women
Gallstone disease
Gallstones
1.27
0.98–1.63
0.067
0.83
0.59–1.16
0.26
Cholecystectomy
3.38
2.44–4.70
< 0.001
2.11
1.50–2.97
< 0.001
CI, confidence interval; NAFLD, non-alcoholic fatty liver disease; OR, odds ratio.
a
Adjusted for age, sex, race-ethnicity, education, BMI, waist circumference, doctor-diagnosed diabetes or hemoglobin A1C≥6.5%, HDL cholesterol, systolic blood pressure,
diastolic blood pressure, cigarette smoking, alcohol intake, coffee intake, physical activity, and right upper quadrant or epigastric pain.
b
Gallstone disease included both gallstones and cholecystectomy. Models included either all gallstone disease or gallstones and cholecystectomy.
90
Men
Prevalence of NAFLD (%)
LIVER
Table 2. Logistic regression odds ratios for the association of NAFLD with gallstone disease
a,b
Women
80
70
a
60
50
a,b
40
67.6
30
49.4
20
10
34.3
20.7
15.3
19.7
0
No GSD Gallstones Cholecystectomy
No GSD Gallstones Cholecystectomy
Figure 2. Age-standardized prevalence (with 95% confidence interval)
of non-alcoholic fatty liver disease (NAFLD) by gallstone disease (GSD)
status and sex. aP < 0.05 compared with no gallstone disease; bP < 0.05
compared with gallstones.
and a half times as likely to have NAFLD compared with persons
without gallstone disease after adjusting for multiple factors. The
association with cholecystectomy was particularly strong among
men. About two-thirds of men with a history of cholecystectomy
also had NAFLD, making this an easily identifiable group who
might be considered for evaluation of the presence of NAFLD.
Few studies of the relationship between NAFLD and gallstone
disease have been conducted. Among clinical series, gallstone disease was found in 20% of patients with NAFLD (5); NAFLD was
found in 55% of patients with gallstone disease (6); and both conditions were quite common among patients with metabolic risk
factors (41% had NAFLD, 34% had gallstone disease) (7). Among
population studies, associations of ultrasound-detected gallstone
disease and fatty liver were found among predominantly low-weight
The American Journal of GASTROENTEROLOGY
residents of Okinawa (9) and Taiwan (11), and Hispanic residents
of Santiago, Chile (10). No association was found in a Taiwanese
population with either fatty liver or liver enzymes (12). There was
a borderline association (P = 0.049) of ultrasound-detected fatty
liver and gallstones in an Italian study among persons randomly
selected based on drinking and obesity status from a population
sample, but adjusted analysis was not performed (30). All of these
studies enrolled far fewer participants than were in NHANES III.
Only the Chilean study investigated relationships with cholecystectomy and gallstones separately. They found a stronger association
of fatty liver with cholecystectomy than with gallstones (10). Using
elevated serum liver enzyme levels rather than hepatic steatosis, a
previous US study found an independent association with cholecystectomy, but not with gallstones (13). Cholecystectomy was also
associated with the development of cirrhosis in that report.
The reasons for the stronger relationship of NAFLD with cholecystectomy than with the presence of gallstones are uncertain.
Symptomatic gallstones leading to a cholecystectomy and asymptomatic gallstones might differ in relation to NAFLD, although
there is little evidence for such a relationship. Alternatively, the
absence of a gallbladder could increase the risk of NAFLD through
a metabolic effect. Following cholecystectomy, bile is continuously secreted into the small intestine, and the bile acid pool circulates more quickly, exposing the liver to a greater flux of bile
acids (14,31–33). This is potentially significant, given the role bile
acids are now known to have in the regulation of hepatic lipid and
glucose metabolism. In addition to their well-established roles in
absorption of dietary lipids and cholesterol homeostasis, bile acids
also act as signaling molecules with systemic endocrine functions
(34–36). Through the activation of nuclear receptors, such as farnesoid X receptor (NR1H4), they have a role in hepatic lipid and
VOLUME 108 | JUNE 2013 www.amjgastro.com
NAFLD and Cholecystectomy
957
Age-adjusted (N=12,220; 10,009 without pain,
2,211 with pain)
Multivariate-adjusteda (N=10,752; 8,808
without pain, 1,944 with pain)
% With NAFLD
ORb
No GSD
17.1
1.0
Gallstones
26.8
1.35
1.03–1.77
1.08
0.79–1.46
Cholecystectomy
52.5
4.09
2.97–5.65
3.02
2.19–4.18
95% CI
OR
95% CI
Without pain
1.0
With pain
No GSD
21.1
1.0
Gallstones
28.3
1.36
0.88–2.11
1.08
1.0
0.66–1.77
Cholecystectomy
36.5
1.95
1.26–3.02
1.66
0.98–2.81
CI, confidence interval; GSD, gallstone disease; NAFLD, non-alcoholic fatty liver disease; OR, odds ratio.
a
Adjusted for age, sex, race-ethnicity, education, BMI, waist circumference, doctor-diagnosed diabetes or hemoglobin A1C≥6.5%, HDL cholesterol, systolic blood pressure,
diastolic blood pressure, cigarette smoking, alcohol intake, coffee intake, and physical activity.
b
From logistic regression analysis. OR for NAFLD, comparing persons with gallstones or with cholecystectomy with those without GSD.
glucose metabolism. As ligands for the G-protein-coupled receptor,
TGR5/Gpbar-1, they have an effect on energy expenditure and glucose homeostasis. Consequently, alterations in bile acid transport
and activation of nuclear receptors and the cell signaling pathways
may contribute to the pathogenesis of NAFLD.
The gallbladder mucosa is metabolically active, secreting and
absorbing compounds to and from the bile. The loss of this metabolic activity in the absence of a gallbladder provides another
possible mechanism for an increased risk of NAFLD following
cholecystectomy. An important example is secretion of fibroblast
growth factor 19 (FGF19). As recently reported, the gallbladder
is its main source of enterohepatobiliary expression (15). FGF19
has a role in the negative feedback regulation of bile salt synthesis
through its signaling function in the portal circulation. FGF19 has
also been shown to inhibit hepatic fatty acid synthesis (37), and
hepatic response to it is impaired among patients with NAFLD and
insulin resistance (38). In a preliminary report, cholecystectomy
led to a decline in serum FGF19 levels coupled with enhanced bile
acid synthesis (39). Although the molecular mechanisms are not
yet clear, an effect of cholecystectomy on triglyceride metabolism
has recently been demonstrated in an animal model (16). Among
mice, cholecystectomy increased serum and hepatic triglyceride
levels and hepatic very low density lipoprotein production.
An important strength of the current study was the use of ultrasonography for unbiased ascertainment of gallstones. Because
ultrasound-detected gallstones were not independently associated
with NAFLD, it is unlikely that the association of cholecystectomy
with NAFLD was due to unmeasured factors common to gallstone
disease and NAFLD. It is possible, however, that NAFLD could have
increased the risk of cholecystectomy. Nonspecific abdominal pain
related to NAFLD could have led to diagnosis of gallstones and
then to cholecystectomy. But if that were true, the association of
cholecystectomy with NAFLD would have been expected to have
been stronger among participants with abdominal pain. In fact,
the opposite was found: the association of cholecystectomy with
© 2013 by the American College of Gastroenterology
NAFLD was stronger among participants without pain (Table 3),
even though participants with cholecystectomy were more likely to
report such pain (Table 1). Thus, it seems unlikely that nonspecific
pain associated with NAFLD was the reason for the association of
cholecystectomy with NAFLD.
NHANES III is a cross-sectional survey; therefore, we were unable
to establish a temporal association between gallstone disease and
NAFLD. We were able to examine the timing of prior cholecystectomy
and did not find an association with NAFLD. Thus, the increased risk
of NAFLD appeared to be with cholecystectomy itself, not with when
cholecystectomy was performed. Another limitation of the study was
the use of ultrasound to detect NAFLD. The accuracy of ultrasonography is excellent for the diagnosis of gallstone disease, but not as
good for NAFLD (40). These limitations are balanced by the benefits of using a large, national, population-based sample, particularly
the avoidance of ascertainment bias that occurs in studies of selected
patients and the ability to generalize the results to a national population. In addition, NHANES III provided the first data on hepatic
steatosis in a large sample representative of the US population.
In conclusion, in the US population, NAFLD is common among
persons with gallstone disease, but the relationship is confined to
persons who had cholecystectomy rather than persons with ultrasound-diagnosed gallstones. Therefore, cholecystectomy might be
considered a risk factor for NAFLD. Because cholecystectomy is so
commonly performed and its long-term medical consequences are
believed to be uncommon, it is important that the findings of this
study be confirmed or rejected.
CONFLICT OF INTEREST
Guarantor of the article: James E. Everhart, MD, MPH.
Specific author contributions: Study concept and design, statistical analysis and interpretation of data, and drafting of manuscript:
Constance E. Ruhl; study concept and design, statistical analysis and
interpretation of data, and critical revision of manuscript: James E.
Everhart. Both authors approved the final draft submitted.
The American Journal of GASTROENTEROLOGY
LIVER
Table 3. Association of NAFLD with gallstone disease by right upper quadrant/epigastric pain status
LIVER
958
Ruhl and Everhart
Financial support: This work was supported by a contract from the
National Institute of Diabetes and Digestive and Kidney Diseases
(HHSN267200700001G and HHSN276201200161U).
Potential competing interests: None.
Study Highlights
WHAT IS CURRENT KNOWLEDGE
3Non-alcoholic fatty liver disease (NAFLD) is an increasingly
common condition, and cholecystectomy is a frequent
surgical procedure.
3Long-term medical consequences of cholecystectomy are
believed to be uncommon.
3Possible risk factors for NAFLD, other than nutritional and
weight-related conditions, have received little attention.
3Few studies have investigated a possible association
between NAFLD and cholecystectomy.
WHAT IS NEW HERE
3NAFLD is associated with cholecystectomy, but not with
gallstones, when controlling for multiple factors common
to both.
3This suggests that cholecystectomy may itself be a risk
factor for NAFLD.
3Further research is needed to confirm or reject these
findings.
REFERENCES
1. Konsten J, Gouma DJ, von Meyenfeldt MF et al. Long-term follow-up after
open cholecystectomy. Br J Surg 1993;80:100–2.
2. Ure BM, Troidl H, Spangenberger W et al. Long-term results after laparoscopic cholecystectomy. Br J Surg 1995;82:267–70.
3. Victorzon M, Lundin M, Haglund C et al. Short and long term outcome
after laparoscopic cholecystectomy. Ann Chir Gynaecol 1999;88:259–63.
4. Ludwig J, Viggiano TR, McGill DB et al. Nonalcoholic steatohepatitis:
Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clin Proc
1980;55:434–8.
5. Loria P, Lonardo A, Lombardini S et al. Gallstone disease in non-alcoholic
fatty liver: prevalence and associated factors. J Gastroenterol Hepatol
2005;20:1176–84.
6. Ramos-De la Medina A, Remes-Troche JM, Roesch-Dietlen FB et al.
Routine liver biopsy to screen for nonalcoholic fatty liver disease (NAFLD)
during cholecystectomy for gallstone disease: is it justified? J Gastrointest
Surg 2008;12:2097–102; discussion 2102.
7. Koller T, Kollerova J, Hlavaty T et al. Cholelithiasis and markers of nonalcoholic fatty liver disease in patients with metabolic risk factors.
Scand J Gastroenterol 2012;47:197–203.
8. Lonardo A, Lombardini S, Scaglioni F et al. Fatty liver, carotid disease
and gallstones: a study of age-related associations. World J Gastroenterol
2006;12:5826–33.
9. Nomura H, Kashiwagi S, Hayashi J et al. Prevalence of gallstone disease in a
general population of Okinawa, Japan. Am J Epidemiol 1988;128:598–605.
10. Nervi F, Miquel JF, Alvarez M et al. Gallbladder disease is associated
with insulin resistance in a high risk Hispanic population. J Hepatol
2006;45:299–305.
11. Chen CH, Huang MH, Yang JC et al. Prevalence and risk factors of gallstone
disease in an adult population of Taiwan: an epidemiological survey.
J Gastroenterol Hepatol 2006;21:1737–43.
12. Lu SN, Chang WY, Wang LY et al. Risk factors for gallstones among
Chinese in Taiwan. A community sonographic survey. J Clin Gastroenterol
1990;12:542–6.
13. Ioannou GN. Cholelithiasis, cholecystectomy, and liver disease. Am J
Gastroenterol 2010;105:1364–73.
14. Kullak-Ublick GA, Paumgartner G, Berr F. Long-term effects of cholecystectomy on bile acid metabolism. Hepatology 1995;21:41–5.
The American Journal of GASTROENTEROLOGY
15. Zweers SJ, Booij KA, Komuta M et al. The human gallbladder secretes
fibroblast growth factor 19 into bile: towards defining the role of fibroblast
growth factor 19 in the enterobiliary tract. Hepatology 2012;55:575–83.
16. Amigo L, Husche C, Zanlungo S et al. Cholecystectomy increases hepatic
triglyceride content and very-low-density lipoproteins production in mice.
Liver Int 2011;31:52–64.
17. NCHS. National Health and Nutrition Examination Survey (NHANES).
Available at: http://www.cdc.gov/nchs/nhanes.htm (cited 12 December
2012).
18. Everhart JE, Khare M, Hill M et al. Prevalence and ethnic differences in
gallbladder disease in the United States. Gastroenterology 1999;117:632–9.
19. NCHS. Third National Health and Nutrition Examination Survey. Gallbladder ultrasonography procedure manual. Available at: http://www.cdc.
gov/nchs/data/nhanes/nhanes3/cdrom/nchs/manuals/gallblad.pdf (cited
12 December 2012) Westat, Inc.: Rockville, MD, 1988.
20. Lazo M, Hernaez R, Bonekamp S et al. Non-alcoholic fatty liver disease and
mortality among US adults: prospective cohort study. BMJ 2011;343:d6891.
21. NCHS. Third National Health and Nutrition Examination Survey: hepatic
steatosis ultrasound images assessment procedure manual. Available at:
http://www.cdc.gov/nchs/data/nhanes/nhanes3/Hepatic_Steatosis_
Ultrasound_Procedures_Manual.pdf (cited 12 December 2012).
22. American Diabetes Association. Executive summary: standards of medical
care in diabetes--2010. Diabetes Care 2010;33 (Suppl 1): S4–S10.
23. Ruhl CE, Everhart JE. Association of diabetes, serum insulin, and
C-peptide with gallbladder disease. Hepatology 2000;31:299–303.
24. Ruhl CE, Everhart JE. Association of coffee consumption with gallbladder
disease. Am J Epidemiol 2000;152:1034–8.
25. Ruhl CE, Everhart JE. Determinants of the association of overweight with
elevated serum alanine aminotransferase activity in the United States.
Gastroenterology 2003;124:71–9.
26. Ruhl CE, Everhart JE. Elevated serum alanine aminotransferase and
gamma-glutamyltransferase and mortality in the United States population.
Gastroenterology 2009;136:477–85 e11.
27. Matthews DR, Hosker JP, Rudenski AS et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose
and insulin concentrations in man. Diabetologia 1985;28:412–9.
28. National Cholesterol Education Program (NCEP) Expert Panel on
Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults
(Adult Treatment Panel III). Third Report of the National Cholesterol
Education Program (NCEP) Expert Panel on Detection, Evaluation, and
Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)
final report. Circulation 2002;106:3143–421.
29. Breslow NE, Day NE. Statistical methods in cancer research: the design and
analysis of cohort studies. International Agency for Research on Cancer:
Lyon, France, 1987, pp. 48–79.
30. Bellentani S, Saccoccio G, Masutti F et al. Prevalence of and risk factors for
hepatic steatosis in Northern Italy. Ann Intern Med 2000;132:112–7.
31. Almond HR, Vlahcevic ZR, Bell CC Jr et al. Bile acid pools, kinetics and
biliary lipid composition before and after cholecystectomy. N Engl J Med
1973;289:1213–6.
32. Shaffer EA, Small DM. Biliary lipid secretion in cholesterol gallstone
disease. The effect of cholecystectomy and obesity. J Clin Invest 1977;59:
828–40.
33. Roda E, Aldini R, Mazzella G et al. Enterohepatic circulation of bile acids
after cholecystectomy. Gut 1978;19:640–9.
34. Hylemon PB, Zhou H, Pandak WM et al. Bile acids as regulatory molecules.
J Lipid Res 2009;50:1509–20.
35. Trauner M, Claudel T, Fickert P et al. Bile acids as regulators of hepatic lipid
and glucose metabolism. Dig Dis 2010;28:220–4.
36. Wagner M, Zollner G, Trauner M. Nuclear receptors in liver disease.
Hepatology 2011;53:1023–34.
37. Bhatnagar S, Damron HA, Hillgartner FB. Fibroblast growth factor-19,
a novel factor that inhibits hepatic fatty acid synthesis. J Biol Chem
2009;284:10023–33.
38. Schreuder TC, Marsman HA, Lenicek M et al. The hepatic response to FGF19
is impaired in patients with nonalcoholic fatty liver disease and insulin
resistance. Am J Physiol Gastrointest Liver Physiol 2010;298:G440–5.
39. Barrera F, Molina HR, Azocar L et al. Cholecystectomy changes the diurnal
rhythm and reduces serum levels of FGF19 coupled with a raise in bile acid
synthesis. Hepatology 2011;54:728A.
40. Hernaez R, Lazo M, Bonekamp S et al. Diagnostic accuracy and reliability
of ultrasonography for the detection of fatty liver: a meta-analysis.
Hepatology 2011;54:1082–90.
VOLUME 108 | JUNE 2013 www.amjgastro.com