THYROID ORIGINAL STUDIES, REVIEWS,

Volume 22, Number 6, 2012

ª Mary Ann Liebert, Inc. AND SCHOLARLY DIALOG
DOI: 10.1089/thy.2011.0279
THYROID FUNCTION AND DYSFUNCTION

Inverse Association Between

Serum Free Thyroxine Levels and Hepatic Steatosis:
Results from the Study of Health in Pomerania

Till Ittermann,1,2,* Robin Haring,2,* Henri Wallaschofski,2 Sebastian E. Baumeister,1 Matthias Nauck,2
Marcus Dörr,3 Markus Lerch,4 Henriette E. Meyer zu Schwabedissen,5
Dieter Rosskopf,5 and Henry Völzke1

Background: Associations between thyroid function and hepatic steatosis defined by enzymatic and sonographic
criteria are largely unknown in the general population. Thus, the aim of the present study was to investigate the
association between thyroid function tests and sonographic as well as enzymatic criteria of liver status in a large
population-based study, the Study of Health in Germany (SHIP).
Methods: Data from 3661 SHIP participants without a self-reported history of thyroid or liver disease were
analyzed. Hepatic steatosis was defined as the presence of a hyperechogenic ultrasound pattern of the liver and
increased serum alanine transferase concentrations. Serum thyroid-stimulating hormone (TSH), free triiodo-
thyronine (FT3), and free thyroxine (FT4) concentrations were associated with hepatic steatosis using multino-
mial regression models adjusted for sex, age, physical activity, alcohol consumption, waist circumference, and
food intake pattern.
Results: We detected no consistent association of serum TSH and FT3 concentrations with hepatic steatosis. In
contrast, serum FT4 concentrations were inversely associated with hepatic steatosis in men (odds ratio
(OR) = 0.04 [95% confidence interval (CI) = 0.01; 0.17]) and women (OR = 0.06 [95% CI = 0.01; 0.42]).
Conclusions: Results from the present cross-sectional study suggest that low FT4 concentrations are associated
with hepatic steatosis. Longitudinal and intervention studies are warranted to investigate whether hypothy-
roidism increases the risk of hepatic steatosis or vice versa.

Introduction among 10,292 outpatients (12) investigated associations of

serum thyrotropin (thyroid-stimulating hormone; TSH) and

T hyroid hormones play an important role in hepatic

lipid homeostasis (1) through increased expression of low-
density lipoprotein (LDL) receptors on the hepatocytes (2) and
free thyroxine (FT4) concentrations with serum c-glutamyl
transpeptidase (GGT) and alanine aminotransferase (ALT)
concentrations. In that study (12), serum TSH concentrations
the activation of lipid-lowering liver transaminases (3), which were positively associated with liver enzyme concentrations,
result in lower serum LDL concentrations. On the other hand, while serum FT4 concentrations were inversely related with
the liver metabolizes thyroid hormones and thereby influ- liver enzyme concentrations. Finally, a small case–control
ences the regulation of their systematic endocrine effects (4). study found significantly lower concentrations of FT4, but not
Several previous studies investigated thyroid function tests TSH, in men with alcoholic fatty liver disease (10).
in patients with liver cirrhosis (5–9), but only three studies The results of these studies (10–12) indicate that hypothy-
addressed the association between thyroid function tests and roidism might be related to hepatic steatosis. This seems to be
hepatic steatosis (10–12). One study that was conducted in biologically plausible, as overt hypothyroidism is associated
inpatients using a matched case–control design demonstrated with visceral obesity (13,14), metabolic syndrome (15), insulin
a twofold higher risk of hypothyroidism in patients with he- resistance (16), and lipid peroxidation (17), all of which are
patic steatosis compared with controls (11). Another study closely related to hepatic steatosis. Moreover, evidence from

1
Institute for Community Medicine; 2Institute of Clinical Chemistry and Laboratory Medicine; 3Department of Cardiology; 4Department of
Medicine A; 5Institute of Pharmacology; Ernst Moritz Arndt University Greifswald, Greifswald, Germany.
*These two authors contributed equally to this work.

568
THYROID FUNCTION TESTS AND HEPATIC STEATOSIS 569

population-based studies (18–21) suggests that sonographic liver pattern was defined as the presence of a bright pattern
and laboratory markers of hepatic steatosis are associated with evident density differences between hepatic and renal
with the risk of atherosclerotic endpoints, and are therefore in parenchyma (21,31,32).
line with the increased cardiovascular risk in hypothyroidism For laboratory examinations, nonfasting blood samples
(22–24). were drawn from the cubital vein in the supine position. The
Previous studies are limited by selected populations laboratory quarterly takes part in the official national German
(10–12), small sample sizes (10,11), and potential residual external proficiency testing programs. In addition, internal
confounding through lacking information on influential quality controls were analyzed daily. Serum ALT, aspartate
medications (e.g., thyroid hormone replacement or antithy- aminotransferase (AST), and GGT concentrations were mea-
roid drugs), lifestyle characteristics (e.g., alcohol consump- sured photometrically (Hitachi 704; Roche, Mannheim, Ger-
tion), and obesity status among the study participants (10–12). many). ALT, AST, and GGT concentrations were expressed as
In addition, previous reports were partly based on increased lmol/(L$s), which corresponds to (lmol/[L$s]) · 60 = IU/L.
transaminase concentrations to indicate fatty liver (12), but For ALT, concentrations exceeding the sex-specific 75th per-
transaminase concentrations might be increased due to vari- centile were termed as increased ALT—0.69 lmol/(L$s) in
ous reasons not necessarily related to hepatic steatosis (25). In men and 0.41 lmol/(L$s) in women. Hepatic steatosis was
contrast, the diagnosis of hepatic steatosis based on ultra- defined as a hyperechogenic liver pattern in ultrasound and
sound and increased transaminase concentrations is much increased ALT concentrations (33,34). Concentrations of TSH,
more specific and sensitive than a diagnosis based solely on free triiodothyronine (FT3), and FT4 were analyzed by im-
liver enzymes (26). munochemiluminescent procedures (FT3, LUMItest, Brahms,
To overcome the limitations of previous research, the aim Berlin, Germany; TSH and FT4, LIA-mat, Byk Sangtec Diag-
of the present study was to investigate the association be- nostica GmbH, Frankfurt, Germany). All assays were per-
tween thyroid function tests and hepatic steatosis identified formed according to the manufacturer’s recommendations.
by increased ALT concentrations and liver ultrasound. For The functional sensitivity of the TSH assay was 0.03 mIU/L.
this, we used data from a large-scale population-based study, Hyper- and hypothyroidism as well as their overt and sub-
the Study of Health in Germany (SHIP). clinical forms were defined using reference limits established
in the study region (35). Hyperthyroidism was defined by
decreased serum TSH concentrations; hypothyroidism was
Materials and Methods
defined by increased serum TSH concentrations. Subclinical
Study population hyper- or hypothyroidism was defined by decreased or in-
creased serum TSH concentrations and FT4 and FT3 concen-
SHIP is a population-based study in West Pomerania, a
trations being in the reference range. Overt hyper- or
region in Northeast Germany. Details on study design have
hypothyroidism was defined by decreased or increased
been published previously (27,28). In brief, from the 212,157
serum TSH concentrations and increased or decreased FT3 or
inhabitants living in the area, a representative random sample
FT4 concentrations.
of 7008 subjects aged 20 to 79 years was selected using pop-
The anthropometric measures included waist circumfer-
ulation registries where all German inhabitants are registered.
ence, measured to the nearest 0.1 cm using an inelastic tape
Only individuals with German citizenship and main resi-
midway between the lower rib margin and the iliac crest in the
dency in the study area were included. The net sample
horizontal plane, with the subject standing comfortably with
(without migrated or deceased persons) comprised 6267 eli-
weight distributed evenly on both feet. Lifestyle factors in-
gible subjects, whereof 4308 finally participated (response
cluded physical activity, food intake pattern, and alcohol
68.8%). All participants gave written informed consent. The
consumption. Subjects who participated in physical training
study conformed to the ethical guidelines of the Declaration of
during summer or winter for at least 1 hour a week were
Helsinki as reflected in an a priori approval by the local Ethics
classified as being physically active. Alcohol consumption
Committee of the University of Greifswald.
was evaluated as beverage-specific alcohol consumption
There were 647 subjects (452 women) excluded due to (over-
(beer, wine, and distilled spirits) on the last weekend and last
laps exist) known self-reported thyroid disease (n = 349), current
weekday preceding the examination, and the mean daily al-
thyroid-related medication according to the anatomical-
cohol consumption was calculated using beverage-specific
therapeutical-chemical code H03 (n = 285), positive hepatitis
pure ethanol volume proportions (36). Food intake pattern
antibodies or reported history of liver cirrhosis (n = 37), and
was selected from a validated food frequency questionnaire.
missing data in any of the considered variables (n = 131). The
The classifications were summarized to a dietary pattern score
final study population for the present analysis comprised
for each subject (37). Gender-specific tertiles of this score re-
3661 subjects (1741 women).
flected the food quality: lower tertile, unfavorable dietary
pattern ( < 12 for men, < 14 for women); medium tertile, in-
Measurements
termediate dietary pattern (12 to 14 for men, 14 to 16 for
All participants underwent an extensive standardized women); and upper tertile, favorable dietary pattern ( > 14 for
medical examination including the collection of blood sam- men, > 16 for women).
ples. Liver ultrasound was performed by trained physicians
using a 5 MHz transducer and a high-resolution instrument
Statistical analysis
(Vingmed VST Gateway, Santa Clara, CA). The sonographers
were unaware of the participant’s clinical and laboratory Selected baseline demographic, behavioral, and clinical
characteristics. In SHIP, ultrasound examinations and read- characteristics were compared by sex using v2 test (qualitative
ings have strict quality standards (29,30). A hyperechogenic data) or Mann–Whitney U-test (quantitative data). Serum
570 ITTERMANN ET AL.

Table 1. Selected Baseline Characteristics Stratified by Sex

Men (n = 1920) Women (n = 1741) pa

Age (years) 50.5 – 16.6 48.1 – 16.1 < 0.001

ALT (lmol/[L$s]) 0.58 – 0.37 0.36 – 0.21 < 0.001
AST (lmol/[L$s]) 0.42 – 0.27 0.32 – 0.15 < 0.001
GGT (lmol/[L$s]) 0.81 – 1.73 0.37 – 0.53 < 0.001
TSH (mU/L) 0.77 – 0.86 0.99 – 3.12 < 0.001
FT3 (pmol/L) 5.31 – 0.89 5.17 – 0.83 < 0.001
FT4 (pmol/L) 13.11 – 3.07 12.43 – 4.49 < 0.001
Hyperthyroidism 156 (8.1%) 123 (7.1%) 0.231
Subclinical 140 (7.3%) 109 (6.3%) 0.217
Overt 16 (0.8%) 14 (0.8%) 0.923
Hypothyroidism 34 (1.8%) 66 (3.8%) < 0.001
Subclinical 29 (1.5%) 50 (2.9%) 0.005
Overt 5 (0.3%) 16 (0.9%) 0.008
Hyperechogenic liver pattern 724 (37.7%) 363 (20.8%) < 0.001
Hepatic steatosis < 0.001
US - & ALT < 75th percentile 1014 (52.8%) 1107 (63.6%)
US - & ALT > 75th percentile 182 (9.5%) 271 (15.6%)
US + & ALT < 75th percentile 412 (21.5%) 167 (9.6%)
US + & ALT > 75th percentile 312 (16.3%) 196 (11.3%)
Waist circumference (cm) 95.4 – 11.7 82.5 – 13.1 < 0.001
Alcohol consumption (g/day) 19.8 – 23.1 5.5 – 9.2 < 0.001
Physically active 803 (41.8%) 772 (44.3%) 0.124
Food intake pattern 0.999
Unfavorable 683 (35.6%) 619 (35.6%)
Intermediate 448 (23.3%) 406 (23.3%)
Optimal 789 (41.1%) 716 (41.1%)

Data are absolute numbers (percentages) or means – standard deviation.

a
p-Values were calculated with v2 test for categorical and Mann–Whitney U-test for continuous variables.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, c-glutamyl transpeptidase; TSH, thyroid-stimulating hormone; FT3,
free triiodothyronine; FT4, free thyroxine; US + , hyperechogenic liver ultrasound pattern; US - , normoechogenic liver ultrasound pattern.

TSH, FT3, and FT4 concentrations were associated with liver association between serum TSH concentrations and hepatic
enzymes and hyperechogenic liver pattern by multivariable steatosis in men but not in women. No association was de-
regression models adjusted for age, physical activity, alcohol tected between serum TSH concentrations and ALT concen-
consumption, waist circumference, and food intake pattern. trations within the reference range combined with a
For categorical outcomes, multinomial logistic regression hyperechogenic liver pattern in men and women. Further,
models were applied. To weaken the impact of outliers in the there were no significant interaction terms between serum
exposure variable on the associations investigated, TSH, FT3, TSH concentrations and any of the confounders, which were
and FT4 concentrations were power-transformed before they significantly associated with hepatic steatosis.
were used in regression analyses (38). Since the interaction In both men and women, FT4 concentrations were signifi-
term between serum TSH concentrations and sex was signif- cantly associated with a hepatic steatosis (Table 2; Figs. 1 and 2).
icantly associated to the outcome variables, all analyses were The median of the serum FT4 concentration distribution,
done stratified for men and women. A value of p < 0.05 was lower, and upper reference limits established in the study
considered statistically significant in all calculations. All sta- region (35) are provided in Figures 1 and 2. The graphs indi-
tistical analyses were performed by Stata 11.0 (Stata Cor- cate the probability for the outcome at a specific serum FT4
poration, College Station, TX). level. For example, the probability for the combined presence
of ALT concentrations ‡ 75th percentile and a hyperechogenic
liver pattern decreases from 22% at a serum FT4 level of
Results
8.3 pmol/L to 8% at a serum FT4 level of 18.9 pmol/L. Results
Characteristics of the study population are presented in did not differ significantly when additionally adjusted for
Table 1. Compared with women, men were older, had a serum TSH concentrations.
higher waist circumference, and consumed more alcohol. FT3 concentrations were not associated with diagnostic
Men also had significantly higher concentrations of liver en- markers of hepatic steatosis with one exception. In men FT3
zymes (ALT, AST, and GGT), FT3, and FT4, but lower serum concentrations were positively associated with a hyperecho-
TSH concentrations. Prevalence of a hyperechogenic liver genic liver pattern in combination with normal ALT (Table 2).
pattern was nearly twice as high in men than in women. For sensitivity analysis we categorized serum TSH con-
Similarly, we found a higher prevalence of hepatic steatosis centrations according to hyper- and hypothyroidism. We
in men (16.3%) compared with women (11.3%). detected no association between hyper- or hypothyroidism
Table 2 shows the results of the multinomial logistic re- and hepatic steatosis in men or women (Table 3). Analyses
gression analyses, which revealed a borderline significant without adjustment for waist circumference revealed a
THYROID FUNCTION TESTS AND HEPATIC STEATOSIS 571

Table 2. Association Between Serum Thyroid Hormone Concentrations and Hepatic Steatosis
in Men and Women

Men (n = 1920) Women (n = 1741)

Exposure OR [95% CI] p-Value OR [95% CI] p-Value

Power-transformed TSH
US - ALT - REF REF
US - ALT + 3.15 [0.24; 41.92] 0.384 2.69 [0.61; 11.88] 0.192
US + ALT - 1.10 [0.13; 9.04] 0.928 0.47 [0.05; 4.46] 0.509
US + ALT + 7.00 [0.86; 56.90] 0.069 1.64 [0.27; 10.08] 0.594
Power-transformed FT4
US - ALT - REF REF
US - ALT + 0.66 [0.12; 3.52] 0.629 1.38 [0.35; 5.44] 0.644
US + ALT - 0.20 [0.06; 0.67] 0.009 0.04 [0.01; 0.30] 0.001
US + ALT + 0.04 [0.01; 0.17] < 0.001 0.06 [0.01; 0.42] 0.004
Power-transformed FT3
US - ALT - REF REF
US - ALT + 0.87 [0.24; 3.08] 0.827 0.62 [0.22; 1.75] 0.368
US + ALT - 2.60 [1.03; 6.56] 0.043 0.45 [0.12; 1.78] 0.258
US + ALT + 1.30 [0.43; 3.90] 0.640 1.19 [0.32; 4.44] 0.791

OR: outcome was analyzed by multinomial logistic regression adjusted for age, physical activity, alcohol consumption, waist
circumference, and food intake pattern.
ALT + , ALT ‡ sex-specific 75th percentile; OR, odds ratio; CI, confidence interval.

significant association between hypothyroidism and hepatic Our results partly agree with those from Targher et al. (12),
steatosis in women (odds ratio (OR) = 2.23 [95% confidence who reported positive or negative relations of serum TSH or
interval (CI) = 1.11; 4.46]) but not in men (OR = 1.89 [95% FT4 concentrations with serum ALT and GGT concentrations.
CI = 0.80; 4.53]). In our study there was only a significant inverse association
between FT4 concentrations and hepatic steatosis. Differences
Discussion between the study of Targher et al. (12) and our investigation
might have resulted from lack of consideration of major
This is the first study to investigate whether there is an
confounders and mediators, such as waist circumference, al-
association between thyroid function tests and hepatic stea-
cohol consumption, and food intake pattern, in the latter
tosis using a population-based design. We found an inverse
study (12). Further, the study by Targher et al. (12) was con-
association between serum FT4 concentrations and hepatic
ducted in inpatients with possible selection bias of more se-
steatosis, whereas serum TSH and FT3 concentrations were
vere forms of thyroid, liver, and comorbid conditions in the
not consistently associated with hepatic steatosis after ad-
analytical sample. In contrast, our study population was re-
justment for sex, age, physical activity, alcohol consumption,
cruited population-based where subjects with preclinical
waist circumference, and food intake.
states might be well presented. Moreover, the outcome in the
latter study (12) was only defined using serum transaminase

FIG. 1. Probability for hepatic steatosis at different serum-

free thyroxine (FT4) concentrations in men. Vertical dashed FIG. 2. Probability for hepatic steatosis at different serum
lines represent the lower reference limit (8.3 mIU/L), the FT4 concentrations in women. Vertical dashed lines repre-
median of the serum FT4 level distribution (11.8 mIU/L), sent the lower reference limit (8.3 mIU/L), the median of the
and the upper reference limit (18.9 mIU/L). ALT, alanine serum FT4 level distribution (11.8 mIU/L), and the upper
aminotransferase. reference limit (18.9 mIU/L).
572 ITTERMANN ET AL.

Table 3. Association Between Hyper- and Hypothyroidism and Hepatic Steatosis in Men and Women

Men (n = 1920) Women (n = 1741)

Exposure OR [95% CI] p-Value OR [95% CI] p-Value

Hyperthyroidism
US - ALT - REF REF
US - ALT + 0.66 [0.29; 1.50] 0.323 1.06 [0.62; 1.79] 0.838
US + ALT - 0.82 [0.54; 1.25] 0.364 0.79 [0.42; 1.48] 0.463
US + ALT + 0.73 [0.41; 1.29] 0.283 0.70 [0.36; 1.36] 0.293
Hypothyroidism
US - ALT - REF REF
US - ALT + 1.41 [0.48; 4.15] 0.533 1.36 [0.69; 2.67] 0.370
US + ALT - 0.75 [0.24; 2.39] 0.629 0.82 [0.31; 2.17] 0.695
US + ALT + 2.18 [0.84; 5.63] 0.109 1.30 [0.59; 2.86] 0.510

OR: outcome was analyzed by multinomial logistic regression adjusted for age, physical activity, alcohol consumption, waist
circumference, and food intake pattern.
Hyperthyroidism, TSH £ 0.25 mIU/L; hypothyroidism, TSH > 2.12 mIU/L.

concentrations, whereas hepatic steatosis was more sensi- hypothalamicpituitary-thyroid axis to increase TSH secretion
tively defined using both sonographic and laboratory criteria (38,39). In our analysis, hypothyroidism was associated with
in our study. Serum ALT concentrations are not only a marker hepatic steatosis in women but not in men without adjust-
for liver disease but also for general cell death. Thus, the re- ment for waist circumference. Thus, mediation by obesity
ported association between serum TSH and ALT concentra- might play a role in the association between hypothyroidism
tions might be referred to a relationship of serum TSH and hepatic steatosis even though we detected no consistent
concentrations with general cell death but not necessarily association between serum TSH concentrations and hepatic
with hepatic steatosis. steatosis.
Liangpunsakul and Chalasani (11) reported a higher Second, patients with hypothyroidism have not only an
prevalence of hypothyroidism in patients with a hepatic increased risk of hyperlipidemia (39) but also exhibit in-
steatosis than in controls. Our study did not confirm this creased fatty acid oxidation and hepatic output of triglycer-
finding, since hypothyroidism defined by increased serum ides. Consequently, hypothyroid patients might be prone to
TSH levels was not significantly associated with hepatic altered lipid peroxidation (17), which is one of the leading
steatosis. The results from the latter study (11), however, causes of liver cell damage (40). Third, decreased thyroid
might not be comparable to those from our study. Their study function is associated with insulin resistance, which appears
population was recruited from inpatients (11) and, in contrast to be a hallmark of hepatic steatosis (16), as well as features of
to our study, hypothyroidism was defined according to the metabolic syndrome (15), including visceral obesity, glu-
medication and previous diagnosis of hypothyroidism. Given cose intolerance, hypertriglyceridemia, hypertension, and
the clinical setting of the other study, the frequency of partici- low high-density lipoprotein cholesterol (41). Thyroid hor-
pants with overt hypothyroidism in the group of hypothyroid mones have pleiotropic effects on energy homeostasis (14,42),
individuals was higher than in our population-based study. lipid and glucose metabolism (43–45), and blood pressure
This suggestion is further supported by the fact that serum FT4 (46), relating thyroid hormone concentrations with parame-
but not serum TSH levels were significantly associated with ters of the metabolic syndrome.
hepatic steatosis in our study. Thus, overt but not subclinical On the other hand, the missing association between serum
hypothyroidism might be associated with hepatic steatosis. In TSH concentrations and hepatic steatosis along with the in-
our study population it was not possible to investigate the verse association between serum FT4 concentrations and he-
association between overt hypothyroidism and hepatic stea- patic steatosis in our study might argue for a relationship
tosis because the numbers were inadequately low. between hepatic steatosis and decreased thyroid function
Green et al. (10) reported decreased FT4 and T4 concen- tests in the alternative direction. In men, *80% of T3 is pro-
trations but normal FT3 and T3 concentrations in patients duced from T4 by conversion in liver and kidneys. It is well
with a fatty liver disease, whereas FT4 concentrations were known that liver fat accumulation influences the metabolism
normal and FT3 concentrations decreased in end-stage liver of different hormones. For example, increased 5b-reductase
disease. Our study not only confirms the results from that activities (47) result in fewer cortisol/cortison metabolites and
study, but extends current knowledge by providing evidence an altered negative feedback control of the hypothalamic–
that an association between FT4 and hepatic steatosis is not pituitary–adrenal axis, which in consequence enhances the
only present in clinically well-defined patients but also in the adrenocorticotropic hormone–dependent dehydroepiandroster-
general population. on sulfate production. The inverse association between serum
There are explanations, which may substantiate the rela- dehydroepiandrosteron sulfate levels and the risk of hepatic
tionship between thyroid function tests and hepatic steatosis. steatosis in young men has been suggested by our group
First, the association between thyroid function and hepatic previously (48). Regarding thyroid hormones, the conversion
steatosis might be related to the possible effect of hypothy- of T4 to T3 might be reduced by stress, starvation, or chronic
roidism on the development of obesity (13,14). Adipokines illness, which may result in the so-called low T3 syndrome or
like leptin, secreted by visceral adipose tissue, stimulate the nonthyroidal illness. This phenomenon has been explained by
THYROID FUNCTION TESTS AND HEPATIC STEATOSIS 573

the idea that the body tries to reduce its metabolism to con- 6. Faber J, Thomsen HF, Lumholtz IB, Kirkegaard C, Siersbaek-
serve energy. Early studies suggested that some fatty acids Nielsen K, Friis T 1981 Kinetic studies of thyroxine, 3,5,3¢-
are potent inhibitors of extrathyroidal T4 conversion (49). triiodothyronine, 3,3,5¢-triiodothyronine, 3¢,5¢-diiodothyronine,
Recently, it has been shown that impaired thyroid hormone 3,3¢-diiodothyronine, and 3¢-monoiodothyronine in patients
action itself may contribute to altered transcriptional changes with liver cirrhosis. J Clin Endocrinol Metab 53:978–984.
in human liver, which results in hepatic lipid accumulation 7. Guven K, Kelestimur F, Yucesoy M 1993 Thyroid function
and might be associated with insulin resistance in fatty liver tests in non-alcoholic cirrhotic patients with hepatic en-
(50). Therefore, an increase of fatty acids in hepatic steatosis cephalopathy. Eur J Med 2:83–85.
might inhibit T4 to T3 conversion, which in turn perpetuates 8. L’Age M, Meinhold H, Wenzel KW, Schleusener H 1980
Relations between serum levels of TSH, TBG, T4, T3, rT3 and
fat accumulation in the liver. Thus, the question of causality
various histologically classified chronic liver diseases. J En-
remains unanswered by our cross-sectional study. Long-
docrinol Invest 3:379–383.
itudinal and intervention studies to determine the direction of
9. Oren R, Sikuler E, Wong F, Blendis LM, Halpern Z 2000 The
causality are strongly needed in this respect. effects of hypothyroidism on liver status of cirrhotic pa-
In summary, results from the present cross-sectional study tients. J Clin Gastroenterol 31:162–163.
suggest that low FT4 concentrations are associated with hepatic 10. Green JR, Snitcher EJ, Mowat NA, Ekins RP, Rees LH,
steatosis. We detected no consistent association between serum Dawson AM 1977 Thyroid function and thyroid regulation
TSH concentrations and hepatic steatosis. Longitudinal and in- in euthyroid men with chronic liver disease: evidence of
tervention studies are warranted to investigate whether hypo- multiple abnormalities. Clin Endocrinol (Oxf) 7:453–461.
thyroidism is causally related to hepatic steatosis or vice versa. 11. Liangpunsakul S, Chalasani N 2003 Is hypothyroidism a risk
factor for non-alcoholic steatohepatitis? J Clin Gastroenterol
37:340–343.
Acknowledgments
12. Targher G, Montagnana M, Salvagno G, Moghetti P, Zop-
SHIP is part of the Community Medicine Net (www pini G, Muggeo M, Lippi G 2008 Association between serum
.medizin.uni-greifswald.de/cm) of the University of Greifs- TSH, free T4 and serum liver enzyme activities in a large
wald, which is funded by grants from the German Federal cohort of unselected outpatients. Clin Endocrinol (Oxf)
Ministry of Education and Research (BMBF, grant 01ZZ0403); 68:481–484.
the Ministry for Education, Research, and Cultural Affairs; 13. Knudsen N, Laurberg P, Rasmussen LB, Bulow I, Perrild H,
and the Ministry for Social Affairs of the Federal State of Ovesen L, Jorgensen T 2005 Small differences in thyroid
Mecklenburg–West Pomerania. The analyses were further function may be important for body mass index and the
supported by the German Research Foundation (DFG Vo occurrence of obesity in the population. J Clin Endocrinol
Metab 90:4019–4024.
955/5-2) and the GANI_MED consortium, funded by the
14. Nyrnes A, Jorde R, Sundsfjord J 2006 Serum TSH is posi-
German Federal Ministry of Education and Research. The
tively associated with BMI. Int J Obes (Lond) 30:100–105.
contributions to data collection made by field workers, study
15. Uzunlulu M, Yorulmaz E, Oguz A 2007 Prevalence of sub-
physicians, ultrasound technicians, interviewers, and com- clinical hypothyroidism in patients with metabolic syn-
puter assistants are gratefully acknowledged. Novo Nordisk drome. Endocr J 54:71–76.
provided partial grant support for the determination of 16. Maratou E, Hadjidakis DJ, Kollias A, Tsegka K, Peppa M,
plasma samples and data analysis. Alevizaki M, Mitrou P, Lambadiari V, Boutati E, Nikzas D,
Tountas N, Economopoulos T, Raptis SA, Dimitriadis G 2009
Disclosure Statement Studies of insulin resistance in patients with clinical and
The authors have nothing to disclose. subclinical hypothyroidism. Eur J Endocrinol 160:785–790.
17. Sundaram V, Hanna AN, Koneru L, Newman HA, Falko JM
1997 Both hypothyroidism and hyperthyroidism enhance
References
low density lipoprotein oxidation. J Clin Endocrinol Metab
1. Malik R, Hodgson H 2002 The relationship between the 82:3421–3424.
thyroid gland and the liver. QJM 95:559–569. 18. Lee DH, Jacobs DR Jr, Gross M, Kiefe CI, Roseman J, Lewis
2. Ness GC, Lopez D, Chambers CM, Newsome WP, Cornelius CE, Steffes M 2003 Gamma-glutamyltransferase is a pre-
P, Long CA, Harwood HJ Jr 1998 Effects of L-triiodothyro- dictor of incident diabetes and hypertension: the Coronary
nine and the thyromimetic L-94901 on serum lipoprotein Artery Risk Development in Young Adults (CARDIA)
levels and hepatic low-density lipoprotein receptor, 3-hydroxy- Study. Clin Chem 49:1358–1366.
3-methylglutaryl coenzyme A reductase, and apo A-I gene 19. Lee DH, Silventoinen K, Hu G, Jacobs DR Jr, Jousilahti P,
expression. Biochem Pharmacol 56:121–129. Sundvall J, Tuomilehto J 2006 Serum gamma-glutamyl-
3. Ness GC, Lopez D 1995 Transcriptional regulation of rat transferase predicts non-fatal myocardial infarction and fatal
hepatic low-density lipoprotein receptor and cholesterol 7 coronary heart disease among 28,838 middle-aged men and
alpha hydroxylase by thyroid hormone. Arch Biochem Bio- women. Eur Heart J 27:2170–2176.
phys 323:404–408. 20. Meisinger C, Doring A, Schneider A, Lowel H 2006 Serum
4. Reed L, Pangaro LN 1995 Physiology of the thyroid gland gamma-glutamyltransferase is a predictor of incident coro-
and the liver. In: Becker KL (ed) Principles and Practice of nary events in apparently healthy men from the general
Endocrinology and Metabolism. Vol. Lippincott Co., Phila- population. Atherosclerosis 189:297–302.
delphia, pp 285–291. 21. Volzke H, Robinson DM, Kleine V, Deutscher R, Hoffmann
5. Bianchi GP, Zoli M, Marchesini G, Volta U, Vecchi F, Iervese W, Ludemann J, Schminke U, Kessler C, John U 2005 He-
T, Bonazzi C, Pisi E 1991 Thyroid gland size and function in patic steatosis is associated with an increased risk of carotid
patients with cirrhosis of the liver. Liver 11:71–77. atherosclerosis. World J Gastroenterol 11:1848–1853.
574 ITTERMANN ET AL.

22. Biondi B, Klein I 2004 Hypothyroidism as a risk factor for consumption in a general population: results of the study of
cardiovascular disease. Endocrine 24:1–13. health in Pomerania (SHIP). J Stud Alcohol 64:75–82.
23. Fowler PB 2000 Risk in cardiovascular disease. Subclinical 37. Winkler G, Schwertner B, Döring A 1995 [Short methods for
hypothyroidism is risk factor for coronary heart disease. Bmj the characterization of a dietary pattern: use and evaluation
321:175. of a food-frequency questionnaire]. Ernährungsumschau 42:
24. Rodondi N, Newman AB, Vittinghoff E, de Rekeneire N, 289–291. (In German.)
Satterfield S, Harris TB, Bauer DC 2005 Subclinical hypo- 38. Royston P, Sauerbrei W 2008 Multivariable Model–Building:
thyroidism and the risk of heart failure, other cardiovascular A Pragmatic Approach to Regression Anaylsis based on
events, and death. Arch Intern Med 165:2460–2466. Fractional Polynomials for Modelling Continuous Variables.
25. Lewis JR, Mohanty SR 2010 Nonalcoholic fatty liver disease: John Wiley & Sons, Chichester United Kingdom.
a review and update. Dig Dis Sci 55:560–578. 39. Duntas LH 2002 Thyroid disease and lipids. Thyroid 12:287–
26. Springer F, Machann J, Claussen CD, Schick F, Schwenzer NF 293.
2010 Liver fat content determined by magnetic resonance im- 40. Poli G, Albano E, Dianzani MU 1987 The role of lipid per-
aging and spectroscopy. World J Gastroenterol 16:1560–1566. oxidation in liver damage. Chem Phys Lipids 45:117–142.
27. John U, Greiner B, Hensel E, Lüdemann J, Piek M, Sauer S, 41. Grundy SM 2008 Metabolic syndrome pandemic. Arter-
Adam C, Born G, Alte D, Greiser E, Haertel U, Hense HW, ioscler Thromb Vasc Biol 28:629–636.
Haerting J, Willich S, Kessler C 2001 Study of Health In 42. Krotkiewski M 2002 Thyroid hormones in the pathogenesis
Pomerania (SHIP): a health examination survey in an east and treatment of obesity. Eur J Pharmacol 440:85–98.
German region: objectives and design. Soz Praventivmed 43. Danese MD, Ladenson PW, Meinert CL, Powe NR 2000
46:186–194. Clinical review 115: effect of thyroxine therapy on serum
28. Volzke H, Ittermann T, Schmidt CO, Dorr M, John U, Wal- lipoproteins in patients with mild thyroid failure: a quanti-
laschofski H, Stricker BH, Felix SB, Rettig R 2009 Subclinical tative review of the literature. J Clin Endocrinol Metab
hyperthyroidism and blood pressure in a population-based 85:2993–3001.
prospective cohort study. Eur J Endocrinol 161:615–621. 44. Kutty KM, Bryant DG, Farid NR 1978 Serum lipids in hy-
29. Dorr M, Wolff B, Robinson DM, John U, Ludemann J, Meng pothyroidism—a re-evaluation. J Clin Endocrinol Metab
W, Felix SB, Volzke H 2005 The association of thyroid 46:55–56.
function with cardiac mass and left ventricular hypertrophy. 45. Torrance CJ, Devente JE, Jones JP, Dohm GL 1997 Effects of
J Clin Endocrinol Metab 90:673–677. thyroid hormone on GLUT4 glucose transporter gene ex-
30. Volzke H, Alte D, Schmidt CO, Radke D, Lorbeer R, pression and NIDDM in rats. Endocrinology 138:1204–1214.
Friedrich N, Aumann N, Lau K, Piontek M, Born G, Have- 46. Asvold BO, Bjoro T, Nilsen TI, Vatten LJ 2007 Association
mann C, Ittermann T, Schipf S, Haring R, Baumeister SE, between blood pressure and serum thyroid-stimulating
Wallaschofski H, Nauck M, Frick S, Arnold A, Junger M, hormone concentration within the reference range: a popu-
Mayerle J, Kraft M, Lerch MM, Dorr M, Reffelmann T, lation-based study. J Clin Endocrinol Metab 92:841–845.
Empen K, Felix SB, Obst A, Koch B, Glaser S, Ewert R, Fietze 47. Westerbacka J, Yki-Jarvinen H, Vehkavaara S, Hakkinen
I, Penzel T, Doren M, Rathmann W, Haerting J, Hannemann AM, Andrew R, Wake DJ, Seckl JR, Walker BR 2003 Body fat
M, Ropcke J, Schminke U, Jurgens C, Tost F, Rettig R, Kors distribution and cortisol metabolism in healthy men: en-
JA, Ungerer S, Hegenscheid K, Kuhn JP, Kuhn J, Hosten N, hanced 5beta-reductase and lower cortisol/cortisone me-
Puls R, Henke J, Gloger O, Teumer A, Homuth G, Volker U, tabolite ratios in men with fatty liver. J Clin Endocrinol
Schwahn C, Holtfreter B, Polzer I, Kohlmann T, Grabe HJ, Metab 88:4924–4931.
Rosskopf D, Kroemer HK, Kocher T, Biffar R, John U, 48. Volzke H, Aumann N, Krebs A, Nauck M, Steveling A,
Hoffmann W 2011 Cohort profile: the study of health in Lerch MM, Rosskopf D, Wallaschofski H 2010 Hepatic
Pomerania. Int J Epidemiol 40:294–307. steatosis is associated with low serum testosterone and high
31. Neuschwander-Tetri BA 2005 Nonalcoholic steatohepatitis serum DHEAS levels in men. Int J Androl 33:45–53.
and the metabolic syndrome. Am J Med Sci 330:326–335. 49. Chopra IJ, Huang TS, Beredo A, Solomon DH, Chua Teco
32. Volzke H, Schwarz S, Baumeister SE, Wallaschofski H, GN, Mead JF 1985 Evidence for an inhibitor of extra-
Schwahn C, Grabe HJ, Kohlmann T, John U, Doren M 2007 thyroidal conversion of thyroxine to 3,5,3¢-triiodothyronine
Menopausal status and hepatic steatosis in a general female in sera of patients with nonthyroidal illnesses. J Clin En-
population. Gut 56:594–595. docrinol Metab 60:666–672.
33. Haring R, Wallaschofski H, Nauck M, Dorr M, Baumeister 50. Pihlajamaki J, Boes T, Kim EY, Dearie F, Kim BW, Schroeder
SE, Volzke H 2009 Ultrasonographic hepatic steatosis in- J, Mun E, Nasser I, Park PJ, Bianco AC, Goldfine AB, Patti
creases prediction of mortality risk from elevated serum ME 2009 Thyroid hormone-related regulation of gene ex-
gamma-glutamyl transpeptidase levels. Hepatology pression in human fatty liver. J Clin Endocrinol Metab
50:1403–1411. 94:3521–3529.
34. Lau K, Lorbeer R, Haring R, Schmidt CO, Wallaschofski H,
Nauck M, John U, Baumeister SE, Volzke H 2010 The asso-
ciation between fatty liver disease and blood pressure in a Address correspondence to:
population-based prospective longitudinal study. J Hyper- Till Ittermann, M.D.
tens 28:1829–1835. Institute for Community Medicine
35. Volzke H, Alte D, Kohlmann T, Ludemann J, Nauck M, John Ernst Moritz Arndt University Greifswald
U, Meng W 2005 Reference intervals of serum thyroid Walther-Rathenau-Str. 48
function tests in a previously iodine-deficient area. Thyroid Greifswald 17487
15:279–285. Germany
36. Alte D, Ludemann J, Piek M, Adam C, Rose HJ, John U 2003
Distribution and dose response of laboratory markers to alcohol E-mail: [email protected]