The impact of fasting on IHL (intrahepatic lipid) content in human subjects has not been investigated previously, but results indicate that it may change rapidly in response to metabolic cues. The aim of the present study was to measure IHL content after fasting and to correlate this with circulating lipid intermediates. A total of eight healthy non-obese young males were studied before and after 12 or 36 h of fasting. IHL content was assessed by 1H-magnetic resonance spectroscopy, and blood samples were drawn after the fasting period. IHL content increased significantly after the 36 h fasting period [median increase 156% (range, 4–252%); P<0.05]. Furthermore, a significant positive correlation between this increase and 3-hydroxybutyrate concentration was detected (P=0.03). No significant change in IHL content was demonstrated after the 12 h fasting period. The baseline median inter-individual variation in IHLs was 0.51% (range, 0.25–0.72%). The coefficient of variation of IHL measurements was 11.6%; 25–30% of the variation was of analytical origin and the remaining 70–75% was attributed to repositioning. In conclusion, IHL content increases in healthy male subjects during fasting, which demonstrates that nutritional status should be accounted for when assessing IHLs in clinical studies. Moreover, the increase in IHLs was positively correlated with the concentration of 3-hydroxybutyrate.

INTRODUCTION

NAFLD (non-alcoholic fatty liver disease) is prevalent in obese subjects and is considered, by many, to be an important feature of the metabolic syndrome. Although the majority of such cases are asymptomatic, several reports underline an association between NAFLD and dyslipidaemia, hypertension, diabetes and cardiovascular disease [1]. The correlation between NAFLD and insulin resistance is close, and is independent of BMI (body mass index), gender and age [25]. Furthermore, the morbidity and mortality in patients with NAFLD is increased [6], which is attributed to cardiovascular disease, diabetes and liver cirrhosis [79].

Elevation of NEFAs (non-esterified fatty acids) appears to be pivotal in the development of NAFLD, but the amount of IHLs (intrahepatic lipids) is also determined by factors such as the capacity of hepatocytes to uptake and turnover NEFAs, the release of TAG [triacylglycerol (triglyceride)] in VLDL (very-low-density-lipoprotein) particles and TAG storage in spherical droplets [10]. Donnelly et al. [11] have calculated that, in patients with NAFLD, 59% of TAG in the liver derives from NEFAs, 26.2% from de novo lipogenesis and 14.9% from the diet.

Intervention studies reveal that the pool of IHLs may change rapidly, especially in response to dietary changes. In patients with Type 2 diabetes, 2 weeks on a hypocaloric diet or an 8% weight reduction reduce IHLs by 27 and 81% respectively [12,13]. Furthermore, 2 weeks on a low-fat diet has been shown to decrease IHLs by 20% in overweight non-diabetic subjects [14]. In rats, as little as 3 days of high-fat feeding causes a 3-fold increase in IHLs [15].

It is well-described that fasting induces an increase in circulating levels of NEFAs and a decrease in both hepatic and peripheral insulin sensitivity. This is associated with an increase in both intramyocellular lipid content [16,17] and hepatic production of ketone bodies [18], both of which reflect mobilization and oxidation of lipid intermediates. Therefore fasting is probably associated with significant changes in IHLs in human subjects, but this remains to be verified experimentally.

The aim of the present study, therefore, was to measure changes in IHLs during fasting in normal subjects by1H-MRS (magnetic resonance spectroscopy) and to correlate this with circulating levels of ketone bodies.

MATERIALS AND METHODS

Subjects

The study population comprised eight healthy non-obese young males [age, 23.6±0.7 years; BMI, 22.8±0.4 kg/m2; fasting plasma glucose, 4.3±0.1 mmol/l; HbA1c (glycosylated haemoglobin A), 5.5±0.1%; ALT (alanine aminotransferase), 23.4±2.8 units/l; TAG, 0.73±0.08 mmol/l; total cholesterol, 4.2±0.2 mmol/l; HDLs (high-density lipoproteins), 1.5±0.1 mmol/l; and LDLs (low-density lipoproteins), 2.4±0.2 mmol/l]. All subjects were students from the local university.

The exclusion criteria included a family history of Type 2 diabetes, use of any medication, alcohol consumption >21 units/week, known liver disease, claustrophobia and the presence of magnetic implants.

All subjects were instructed to consume a diet with no major deviations from the national recommendations (i.e. a maximum of 30% of energy from fat, 50–60% from carbohydrates and 10–20% from protein), to avoid high-fat meals and to abstain from alcohol for 3 days before each study period. During the fasting period they were allowed to drink tap or mineral water and to perform normal ambulatory activities, excluding any kind of exercise.

Study design

The subjects were randomly assigned to one of two groups: group 1, 1H-MRS examination at baseline and after 36 h of fasting (n=6); and group 2, 1H-MRS examination at baseline and after 12 h of fasting (n=3). Blood samples were drawn from both groups immediately after the fasting period. One of the subjects participated in both groups separated by 2 months.

The protocol was approved by the regional Ethics Committee, and the nature and potential risks were explained before participants gave written informed consent. The study was conducted according to the declaration of Helsinki (2000) of the World Medical Association.

1H-MRS

1H-MRS was performed using a Signa Excite 1.5 telsa twin speed scanner (GE Medical Systems). The subjects were allowed to eat and drink until the beginning of the study period. The baseline examination was performed at 20.00 hours. Group 1 was then fasted for 36 h and group 2 was fasted overnight (12 h), before re-examination at 08.00 hours.

During each visit, three 1H-MRS measurements (spectra 1–3) were made to enable calculations of the inter-individual variation in baseline IHLs and the intra-day CV (coefficient of variation) to be made. First, spectra 1 and 2 were obtained (with the same shimming and spatial position), and the subject was taken out of the machine, allowed to move around for a few minutes and then repositioned before spectra 3 was made. Subjects were positioned feet first in the supine position. A belt was strapped around the lower part of the thorax and upper part of abdomen to minimize respiratory movement of the diaphragm and liver. To generate the spectra, a standard whole-body coil was used for radiofrequency transmission and signal reception. The exact orientation of the liver was verified in a three-plane T2 weighted localizer pulse sequence.

An oblique plane T1 weighted gradient echo pulse sequence using a single breath-holding technique with a TR (repetition time) of 140 ms and a TE (echo time) of 2.2/4.4 ms was performed to enable identification of the area of interest.

The volume of interest (2 cm3×2 cm3×3 cm3) was carefully positioned in the lower posterior part of the liver (area 6), avoiding the inclusion of costae, visible vessels and the bile duct. Autoshimming was performed to optimize the magnetic field homogeneity. A water-suppressed point resolved spectroscopy sequence during free breathing (TE, 30 ms; TR, 2000 ms; number of acquisitions, 128) was applied, using water as the autocentre frequency. FWHM (full width at half maximum) of the unsuppressed water peak was 10.2±1.6 Hz. Each session lasted approx. 40 min.

Analytical procedures and calculations

RLC (relative lipid content)

The spectra were analysed using SAGE (version 7; GE Medical Systems). The height of the suppressed water signal intensity peak (Swater) was measured at approx. 4.8 p.p.m. and the lipid/(-CH2-)n signal intensity (Slipid) peak at approx. 1.4 p.p.m. Representative spectra from subject number 4, before and after the 36-h fast, are shown in Figure 1. Peak height rather than the AUC (area under the curve) was used, as our FWHM was relatively low and constant (10.2±1.6); moreover, using AUC resulted in a higher CV.

Spectra from subject number 4 before (baseline) and after the 36-h fast
Figure 1
Spectra from subject number 4 before (baseline) and after the 36-h fast

(A) Suppressed water peak; (B) lipid peak.

Figure 1
Spectra from subject number 4 before (baseline) and after the 36-h fast

(A) Suppressed water peak; (B) lipid peak.

RLC was calculated using the following formula:

 
formula

where WS% is the percentage of water suppression.

Validation of the data

The CV was calculated by dividing the S.D. with the mean (of the three RLCs), and is presented as a percentage. To demonstrate the intra-day CV, the median CV of RLCs was determined.

Blood analysis

A blood sample was drawn at the end of the fasting period. Plasma glucose was measured in duplicate immediately after sampling on a Beckman Glucoanalyser (Beckman Instruments). Serum samples were frozen immediately and stored at −20 °C. Insulin, growth hormone and cortisol were analysed using a time-resolved fluoroimmunoassay (AutoDELFIA; PerkinElmer), C-peptide was analysed by ELISA (DakoCytomation), and NEFAs were analysed using a commercial kit (Wako Chemicals). ALT was determined using a commercial method (Cobas Integra 800; Roche Diagnostics). Glycerol, lactate, alanine and 3-hydroxybutyrate were measured using a Cobas biocentrifugal analyser with fluorimetric attachment (Roche Diagnostics) [20].

Statistical analysis

Intercooled Stata 9.0 was used for the statistical analysis. To analyse the changes in IHLs, the mean of the three RLCs, at baseline and after the fasting period, were used as the best estimates of the ‘true’ values, and a paired Student's t test was used to determine statistical significance. For estimation of the different components of variation in the RLC measurements, two Bland–Altman analyses were made followed by a two-component ANOVA on the S.D. For statistical significant results, a P value <0.05 was required. As the distribution of the RLC data were skewed, data were log-transformed before applying the relevant statistical tests, and only the medians are shown. Unless otherwise stated, results are presented as means±S.E.M.

RESULTS

Baseline variation in RLCs

The inter-individual variation in RLCs before the fasting periods ranged from 0.25 to 0.72% (5th to the 95th percentile) with a median of 0.51%. As one of the subjects participated in both the 12h- and 36-h fast, a mean of his baseline RLC was used.

Validation of RLC measured by 1H-MRS

The median CV for RLC measured by 1H-MRS was 11.6%. None of the spectra were significantly different, as shown in Figure 2. After performing a two-component analysis, 25–30% of the variation was found to be due to differences between spectra 1 and 2, attributed to elements of uncertainty in the analysis of data, equipment errors and respiration movement, with the remaining 70–75% being attributed to repositioning before the third spectra was made. This underlines the importance of precision when the voxel is placed.

Reproducibility of the three measurements of RLC obtained at the 18 magnetic resonance sessions

Figure 2
Reproducibility of the three measurements of RLC obtained at the 18 magnetic resonance sessions

The broken line depicts the optimal correlation (slope=1). (A) Correlation between RLC 1 and 2 (r=0.97, P<0.01); (B) correlation between RLC 1 and 3 (r=0.95, P<0.01); (C) correlation between RLC 2 and 3 (r=0.93, P<0.01).

Figure 2
Reproducibility of the three measurements of RLC obtained at the 18 magnetic resonance sessions

The broken line depicts the optimal correlation (slope=1). (A) Correlation between RLC 1 and 2 (r=0.97, P<0.01); (B) correlation between RLC 1 and 3 (r=0.95, P<0.01); (C) correlation between RLC 2 and 3 (r=0.93, P<0.01).

Changes in IHLs during the fasting periods

The IHL increased significantly during the 36-h fast. Median RLC increased from 0.42 to 0.74% (95% confidence interval, 1.34–3.54%; P=0.009) (Figure 3). The median RLC% (percentage change in RLC) during the 36-h fast was 156% (range, 4–252%; P<0.05) (Table 1). No significant change in the IHL was observed after the 12-h fast (RLC, 0.72% before compared with 0.42% after fasting; P=0.45).

RLC at baseline (●) and after the 36-h fast (○) from each subject in group 1 (n=6)

Figure 3
RLC at baseline (●) and after the 36-h fast (○) from each subject in group 1 (n=6)

Median RLC values at baseline (■) and after the 36-h fast (□) are shown. A significant difference (P=0.009) was observed between baseline and after the 36-h fast.

Figure 3
RLC at baseline (●) and after the 36-h fast (○) from each subject in group 1 (n=6)

Median RLC values at baseline (■) and after the 36-h fast (□) are shown. A significant difference (P=0.009) was observed between baseline and after the 36-h fast.

Table 1
RLC% after 12 and 36 h of fasting in the study subjects
RLC%
Subject12-h fast36-h fast
 151.7 
 160.1 
 167.9 
 260.8 
 76.6 
−8.7 −4.2 
10.1  
−41.1  
RLC%
Subject12-h fast36-h fast
 151.7 
 160.1 
 167.9 
 260.8 
 76.6 
−8.7 −4.2 
10.1  
−41.1  

Analytes in blood

The levels of the measured analytes at the end of each fasting period are shown in Table 2. Owing to the small sample size, a statistical comparison between the levels after the two fasting studies was not performed, but, as expected, the concentrations of all of the lipid intermediates increased with more prolonged fasting. There was a significant positive correlation between the increase in RLC% and the 3-hydroxybutyrate concentration after the 36-h fast (r=0.85, P=0.03), as shown in Figure 4. The RLC% as well as RLC after the 36-h fast did not correlate with either NEFA levels (P=0.31, P=0.60) or ALT levels (P=0.68, P=0.20).

Table 2
Blood samples drawn after 12 and 36 h of fasting

Values are means±S.E.M.

Time of fasting
12 h36 h
n 
NEFAs (μmol/l) 493.3±59.5 1115.7±87.1 
Glucose (mmol/l) 4.9±0.1 3.8±0.1 
Insulin (pmol/l) 13.6±3.8 13.8±1.7 
C-peptide (pmol/l) 325.5±36.9 207.8±67.9 
Glucagon (pg/ml) 36.7±7.1 88.7±13.9 
Cortisol (nmol/l) 447.7±114.5 473.7±56.8 
Growth hormone (μg/l) 2.9±1.6 8.3±3.5 
Alanine (μmol/l) 216.7±40.5 636.7±38.2 
Lactate (μmol/l) 466.7±53.4 194.2±15.8 
Glycerol (μmol/l) 53.3±14.8 76.7±2.5 
3-Hydroxybutyrate (μmol/l) 215±87.6 2455±233.9 
ALT (units/l) 8.6±3.0 9±1.4 
Time of fasting
12 h36 h
n 
NEFAs (μmol/l) 493.3±59.5 1115.7±87.1 
Glucose (mmol/l) 4.9±0.1 3.8±0.1 
Insulin (pmol/l) 13.6±3.8 13.8±1.7 
C-peptide (pmol/l) 325.5±36.9 207.8±67.9 
Glucagon (pg/ml) 36.7±7.1 88.7±13.9 
Cortisol (nmol/l) 447.7±114.5 473.7±56.8 
Growth hormone (μg/l) 2.9±1.6 8.3±3.5 
Alanine (μmol/l) 216.7±40.5 636.7±38.2 
Lactate (μmol/l) 466.7±53.4 194.2±15.8 
Glycerol (μmol/l) 53.3±14.8 76.7±2.5 
3-Hydroxybutyrate (μmol/l) 215±87.6 2455±233.9 
ALT (units/l) 8.6±3.0 9±1.4 

Correlation between RLC% during the 36-h fast and 3-hydroxybutyrate immediately after the 36-h fast

Figure 4
Correlation between RLC% during the 36-h fast and 3-hydroxybutyrate immediately after the 36-h fast

r=0.85, P=0.03. 3-OHB, 3-hydroxybutyrate.

Figure 4
Correlation between RLC% during the 36-h fast and 3-hydroxybutyrate immediately after the 36-h fast

r=0.85, P=0.03. 3-OHB, 3-hydroxybutyrate.

DISCUSSION

The present study demonstrates for the first time that fasting is associated with an increase in IHLs in human subjects. Apart from extending our knowledge about substrate metabolism in humans, the present findings also demonstrate that food intake must be taken into consideration when interpreting IHL results.

Fasting is associated with distinct and dynamic changes in substrate metabolism, and Cahill [18] has indicated that approx. 40% of NEFAs are metabolized via hepatic ketogenesis after long-term fasting, reflecting that ketone bodies become increasingly important as fuel for the brain. Our present study has shown that the increase in IHLs after 36 h of fasting was positively correlated with the hepatic 3-hydroxybutyrate production. Although high levels of IHLs in steady-state conditions such as the metabolic syndrome appear to be predictive of disease activity, the observed increase during fasting is more likely to reflect a physiological adaptation to an alteration in substrate metabolism in general and the increase in hepatic ketogenesis in particular.

In several studies, ALT levels have been used as a surrogate marker of IHLs [21,22], but, in the present study, ALT did not correlate with any indices of IHLs. This may reflect the short duration of the present study, but it is also plausible that the mechanisms underlying the increase in IHLs during fasting differ from those of NAFLD.

As the present study population consisted of lean and healthy adult males, without any evidence of hepatic disease or steatosis, one would anticipate IHLs to be low. We found a baseline median RLC of 0.51%, with inter-individual variation ranging from 0.25 to 0.72% (5th to 95th percentile). Compared with the median of 1.9% recorded by Szczepaniak et al. [23] in a group with a low risk of steatosis, IHLs in our present subjects is indeed low. Several factors, including dietary habits, physical fitness, age and gender, are known to influence IHLs in otherwise healthy subjects. Westerbacka et al. [14], who measured IHLs by 1H-MRS in obese non-diabetic women before and after 2 weeks of an isocaloric high-fat or low-fat diet, recorded a 20% decrease in IHLs in the low-fat diet group and a 35% increase in the high-fat diet group; however, a single high-fat meal does not appear to influence IHLs [23]. A high level of habitual physical activity is associated with low IHLs [24]. Cross-sectional studies indicate that age is inversely related to IHLs [25,26], whereas the impact of gender remains controversial.

In conclusion, the present study shows for the first time that IHLs increase during fasting in healthy human subjects and that this increase correlates positively with circulating levels of 3-hydroxybutyrate. This contributes to our knowledge about the regulation of substrate metabolism during alterations in nutritional supply. It also emphasizes that nutritional status should be standardized when assessing IHLs in patients.

Abbreviations

     
  • ALT

    alanine aminotransferase

  •  
  • AUC

    area under the curve

  •  
  • BMI

    body mass index

  •  
  • CV

    coefficient of variation

  •  
  • FWHM

    full width at half maximum

  •  
  • IHL

    intrahepatic lipid

  •  
  • MRS

    magnetic resonance spectroscopy

  •  
  • NAFLD

    non-alcoholic fatty liver disease

  •  
  • NEFA

    non-esterified fatty acid

  •  
  • RLC

    relative lipid content

  •  
  • RLC%

    percentage change in RLC

  •  
  • TAG

    triacylglycerol

  •  
  • TE

    echo time

  •  
  • TR

    repetition time

We thank Lisa Erika Dalman for assistance during the magnetic resonance sessions. The study was sponsored by Pfizer via an unrestricted research grant.

References

References
1
Cortez-Pinto
 
H.
Camilo
 
M. E.
Baptista
 
A.
De Oliveira
 
A. G.
De Moura
 
M. C.
 
Non-alcoholic fatty liver: another feature of the metabolic syndrome?
Clin. Nutr.
1999
, vol. 
18
 (pg. 
353
-
358
)
2
Marchesini
 
G.
Brizi
 
M.
Morselli-Labate
 
A. M.
, et al 
Association of nonalcoholic fatty liver disease with insulin resistance
Am. J. Med.
1999
, vol. 
107
 (pg. 
450
-
455
)
3
Westerbacka
 
J.
Corner
 
A.
Tiikkainen
 
M.
, et al 
Women and men have similar amounts of liver and intra-abdominal fat, despite more subcutaneous fat in women: implications for sex differences in markers of cardiovascular risk
Diabetologia
2004
, vol. 
47
 (pg. 
1360
-
1369
)
4
Roberts
 
E. A.
 
Pediatric nonalcoholic fatty liver disease (NAFLD): a ‘growing’ problem?
J. Hepatol.
2007
, vol. 
46
 (pg. 
1133
-
1142
)
5
Marchesini
 
G.
Brizi
 
M.
Bianchi
 
G.
, et al 
Nonalcoholic fatty liver disease: a feature of the metabolic syndrome
Diabetes
2001
, vol. 
50
 (pg. 
1844
-
1850
)
6
Angulo
 
P.
 
Obesity and nonalcoholic fatty liver disease
Nutr. Rev.
2007
, vol. 
65
 (pg. 
S57
-
S63
)
7
Schindhelm
 
R. K.
Diamant
 
M.
Dekker
 
J. M.
Tushuizen
 
M. E.
Teerlink
 
T.
Heine
 
R. J.
 
Alanine aminotransferase as a marker of non-alcoholic fatty liver disease in relation to type 2 diabetes mellitus and cardiovascular disease
Diabetes Metab. Res. Rev.
2006
, vol. 
22
 (pg. 
437
-
443
)
8
Schindhelm
 
R. K.
Diamant
 
M.
Heine
 
R. J.
 
Nonalcoholic fatty liver disease and cardiovascular disease risk
Curr. Diabetes Rep.
2007
, vol. 
7
 (pg. 
181
-
187
)
9
Bugianesi
 
E.
Vanni
 
E.
Marchesini
 
G.
 
NASH and the risk of cirrhosis and hepatocellular carcinoma in type 2 diabetes
Curr. Diabetes Rep.
2007
, vol. 
7
 (pg. 
175
-
180
)
10
Bradbury
 
M. W.
 
Lipid metabolism and liver inflammation. I. Hepatic fatty acid uptake: possible role in steatosis
Am. J. Physiol. Gastrointest. Liver Physiol.
2006
, vol. 
290
 (pg. 
G194
-
G198
)
11
Donnelly
 
K. L.
Smith
 
C. I.
Schwarzenberg
 
S. J.
Jessurun
 
J.
Boldt
 
M. D.
Parks
 
E. J.
 
Sources of fatty acids stored in liver and secreted via lipoproteins in patients with nonalcoholic fatty liver disease
J. Clin. Invest.
2005
, vol. 
115
 (pg. 
1343
-
1351
)
12
Tamura
 
Y.
Tanaka
 
Y.
Sato
 
F.
, et al 
Effects of diet and exercise on muscle and liver intracellular lipid contents and insulin sensitivity in type 2 diabetic patients
J. Clin. Endocrinol. Metab.
2005
, vol. 
90
 (pg. 
3191
-
3196
)
13
Petersen
 
K. F.
Dufour
 
S.
Befroy
 
D.
Lehrke
 
M.
Hendler
 
R. E.
Shulman
 
G. I.
 
Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes
Diabetes
2005
, vol. 
54
 (pg. 
603
-
608
)
14
Westerbacka
 
J.
Lammi
 
K.
Hakkinen
 
A. M.
, et al 
Dietary fat content modifies liver fat in overweight nondiabetic subjects
J. Clin. Endocrinol. Metab.
2005
, vol. 
90
 (pg. 
2804
-
2809
)
15
Samuel
 
V. T.
Liu
 
Z. X.
Qu
 
X.
, et al 
Mechanism of hepatic insulin resistance in non-alcoholic fatty liver disease
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
32345
-
32353
)
16
Wietek
 
B. M.
Machann
 
J.
Mader
 
I.
, et al 
Muscle type dependent increase in intramyocellular lipids during prolonged fasting of human subjects: a proton MRS study
Horm. Metab. Res.
2004
, vol. 
36
 (pg. 
639
-
644
)
17
Stannard
 
S. R.
Thompson
 
M. W.
Fairbairn
 
K.
Huard
 
B.
Sachinwalla
 
T.
Thompson
 
C. H.
 
Fasting for 72 h increases intramyocellular lipid content in nondiabetic, physically fit men
Am. J. Physiol Endocrinol. Metab.
2002
, vol. 
283
 (pg. 
E1185
-
E1191
)
18
Cahill
 
G. F.
 
Fuel metabolism in starvation
Annu. Rev. Nutr.
2006
, vol. 
26
 (pg. 
1
-
22
)
19
Kawamitsu
 
H.
Kaji
 
Y.
Ohara
 
T.
Sugimura
 
K.
 
Feasibility of quantitative intrahepatic lipid imaging applied to the magnetic resonance dual gradient echo sequence
Magn. Reson. Med. Sci.
2003
, vol. 
2
 (pg. 
47
-
50
)
20
Harrison
 
J.
Hodson
 
A. W.
Skillen
 
A. W.
Stappenbeck
 
R.
Agius
 
L.
Alberti
 
K. G.
 
Blood glucose, lactate, pyruvate, glycerol, 3-hydroxybutyrate and acetoacetate measurements in man using a centrifugal analyser with a fluorimetric attachment
J. Clin. Chem. Clin. Biochem.
1988
, vol. 
26
 (pg. 
141
-
146
)
21
Schindhelm
 
R. K.
Diamant
 
M.
Dekker
 
J. M.
Tushuizen
 
M. E.
Teerlink
 
T.
Heine
 
R. J.
 
Alanine aminotransferase as a marker of non-alcoholic fatty liver disease in relation to type 2 diabetes mellitus and cardiovascular disease
Diabetes Metab. Res. Rev.
2006
, vol. 
22
 (pg. 
437
-
443
)
22
Ryan
 
M. C.
Abbasi
 
F.
Lamendda
 
C.
Carter
 
S.
McLaughlin
 
T. L.
 
Serum alanine aminotransferase levels decrease further with carbohydrate than fat restriction in insulin-resistant adults
Diabetes Care
2007
, vol. 
30
 (pg. 
1075
-
1080
)
23
Szczepaniak
 
L. S.
Nurenberg
 
P.
Leonard
 
D.
, et al 
Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population
Am. J. Physiol. Endocrinol. Metab.
2005
, vol. 
288
 (pg. 
E462
-
E468
)
24
Perseghin
 
G.
Lattuada
 
G.
De Cobelli
 
F.
, et al 
Habitual physical activity is associated with intrahepatic fat content in humans
Diabetes Care
2007
, vol. 
30
 (pg. 
683
-
688
)
25
Lonardo
 
A.
Lombardini
 
S.
Scaglioni
 
F.
, et al 
Fatty liver, carotid disease and gallstones: a study of age-related associations
World J. Gastroenterol.
2006
, vol. 
12
 (pg. 
5826
-
5833
)
26
Bedogni
 
G.
Miglioli
 
L.
Masutti
 
F.
Tiribelli
 
C.
Marchesini
 
G.
Bellentani
 
S.
 
Prevalence of and risk factors for nonalcoholic fatty liver disease: the Dionysos nutrition and liver study
Hepatology
2005
, vol. 
42
 (pg. 
44
-
52
)