Liver Fat Accumulation in Response to Overfeeding With a High-fat Diet

A Comparison Between South Asian and Caucasian Men

Siti N. Wulan; Vera B. Schrauwen-Hinderling; Klaas R. Westerterp; Guy Plasqui


Nutr Metab. 2015;12(18) 

In This Article


Study Subjects

Subjects were 10 healthy adult non-diabetic South Asian men and 10 Caucasian men, matched for body fat percentage. The number of subjects was determined based on previous studies showing an increase in liver fat content after short-term high fat feeding in Caucasian subjects.[21,23] Subject characteristics are presented in Table 1. Asian subjects had 4 grandparents from South Asia, while Caucasian subjects were non-Hispanic Europeans. South Asian subjects were students who temporarily living in Europe, and were not recruited from the European population. Subjects were selected based on the following inclusion criteria: healthy, not having metabolic diseases (diabetes or cardiovascular diseases), not using medication, aged between 20 and 40 years old with a body mass index between 18–29 kg/m2 for South Asians and 22–33 kg/m2 for Caucasians, having a stable body weight for the last three months, having no-low-or moderate alcohol intake, not being on a diet and no athletes. All subjects received verbal and written information before giving their consent to participate in the study. The study was approved by Medical Ethics Committee of Maastricht University Medical Centre, MEC No. 10-3-013 and registered in the public trial registry ( No. NL31217.068.10).

Study Design

This study was a diet-intervention study in free-living conditions. Body composition was measured at baseline to have a matched body fat percentage between two ethnic groups. Cardio-respiratory fitness was measured prior to the diet intervention. Daily physical activity was measured for 7 consecutive days with an accelerometer. A diet to maintain energy balance for 3 days was prepared on the basis of fat-free mass and the daily physical activity counts of each subject. Baseline measurement of hepatic fat content was performed afterwards. The overfeeding with a high fat diet for an interval of 4 days was started right after the baseline measurement and hepatic fat content was measured again after the overfeeding interval. All measurements were carried out at the Metabolic Research Unit Maastricht (MRUM), Maastricht University, Maastricht, The Netherlands.

Body Composition

Body composition was determined according to a 3-compartment model based on body weight, body volume and total body water. Body weight and body volume were determined in the morning, in the fasting state. Body volume was determined by hydro-densitometry with simultaneous measurement of residual lung volume using the helium dilution technique. Total body water was determined with deuterium dilution according to the Maastricht protocol.[25] Body composition was calculated from body density and total body water using the equation of Siri.[26]


Anthropometric measures were performed at the same time as body composition. Waist circumference was measured using a circumference measuring tape (Seca 201, United Kingdom) at the umbilical, while subject standing in very light clothing. Hip circumference was measured as the largest circumference between waist and thighs.[27] Skinfold thickness of the biceps, triceps, subscapular and suprailiac was determined using Harpenden skinfold caliper (Body Care, England). Anthropometry measures were performed using The NHANES body measurement guidelines.[28]

Cardio-respiratory Fitness

Physical fitness was assessed with an incremental test on a bicycle ergometer using the protocol of Kuipers et al..[29] During the test, O2-consumption and CO2-production were measured continuously and heart rate was monitored (Polar heart rate monitor, Polar Electro Oy, Kempele, Finland). After a warming up of 5 min at 100 Watt (W) for men, workload was increased with 50 W every 2.5 min. When heart rate (HR) reached a value of 35 beats per min (bpm) below the age predicted maximal HR (220 bpm - age) or the respiratory quotient (RQ = CO2-production/O2-consumption) exceeded 1, workload was increased with 25 W every 2.5 min until exhaustion. VO2 max was determined by averaging the last few points of maximum O2 consumption. Cardio-respiratory fitness was defined as VO2 max divided by FFM.

Daily Physical Activity Level

The daily physical activity level (PAL) was measured using a Direct Life triaxial accelerometer for movement registration (TracmorD) (Philips New Wellness Solutions; The device is a small (3.2 × 3.2 × 0.5 cm), light-weight (12.5 g) instrument. The accelerometer was attached to the lower back by means of an elastic belt. It registered accelerations minute by minute, in the mediolateral (x-axis), longitudinal (y-axis) and anterioposterior (z-axis) of the trunk as described elsewhere.[30] Subjects were instructed to wear the accelerometer for 7 consecutive days, during waking hours except during water activities. Subjects were advised to maintain their habitual physical activity level and not to perform any strenuous physical activity during the diet intervention. TracmorD output was expressed as activity counts/min. The TracmorD activity counts/min were summed over the entire monitoring period and divided by the number of monitoring days to determine the average TracmorD counts per day (counts/d). Daily PAL was calculated based on the activity counts/d with the formula, PAL = 1.354 + 256 × 10−9× counts/d.[30] Daily total energy expenditure was calculated with the formula of Bonomi et al., TEE accelerometer = 0.04 + 0.17 FFM + 1.67 × 10−6 × counts/d[30] by including activity counts/d (from the accelerometer) and fat-free mass (FFM, from the body composition measurement).

Energy Intake

The weight maintenance diet to be consumed at home for 3 days before the baseline measurement was calculated on the basis of TEE. TEE was calculated by the formula of Bonomi et al.[30] as described above. The macronutrients distribution of the diet prior to baseline measurement was 30 % fat, 55 % carbohydrate and 15 % protein.

The overfeeding with a high fat diet was prepared with 50 % excess energy above the requirement.[31] The macronutrients composition of the high fat diet was 60 % fat, 25 % carbohydrate and 15 % protein.[24,32] Fatty acids composition of the diet was 40 % saturated fatty acids and 60 % unsaturated fatty acids.

A written instruction was given to prepare the diet at home. During the weight maintenance, subjects were provided with the diet in an excess amount than TEE and were allowed to eat more or less from the diet prescribed, according to what they needed (ad libitum). Any additional intake from those prescribed foods was recorded. All unfinished foods were collected and returned to the university, to calculate actual energy intake. During the overfeeding period, subjects were asked to finish all the foods prescribed, but otherwise noted down and returned the unfinished foods. The diet consisted of normal ready-to-eat foods combining a typical Western and Asian diet. Foods were selected by reviewing the ingredients content to ensure there was no/limited effect of certain ingredients on fat oxidation (such as spices). During the high fat diet period, subjects were also provided with decaffeinated coffee and fruit tea, as caffeine was also reported to increase fat oxidation. Alcohol intake was limited during the diet intervention (only one serving per day if needed) and was not allowed within 2 days prior to the liver fat measurement.

Abdominal Fat Measurement

Fat in the abdominal compartment was measured using a 3.0 Tesla MRI scanner (Achieva, Philips Healthcare, Best, The Netherlands) with a body coil. A single MRI slice on the umbilical level[33] was acquired using an axial T1 weighted spin-echo sequence in breath hold with following parameters: echo time of 15 ms, repetition time of 500 ms, matrix size of 180 × 96 and slice thickness of 8 mm. Analysis of the abdominal MR images was performed using OsiriX DICOM viewer software. A threshold value was chosen for separation of lean and adipose tissue. Manual segmentation by the region-growing tool was used to segment the area of total abdominal fat (TAT) and region of interest (ROI) borderline for subcutaneous abdominal fat (SAT), whereas visceral fat (VAT) was calculated by subtracting SAT from TAT.

Hepatic Lipid Content

Lipid accumulation in the liver was measured before and after overfeeding with a high fat diet. All measurements were performed on a 3.0 T Philips Achieva scanner (Philips Healthcare, Best, The Netherlands) using a SENSE-cardiac coil.[34] A single voxel of 20 × 20 × 20 mm3 was positioned in the right liver lobe, avoiding large biliary or vascular structures.[34] Spectra were acquired using a point-resolved spectroscopy sequence PRESS[34] with repetition time of 4000 ms, echo time of 33 ms, and number of averages of 64. To minimize the motion artifacts, subjects were asked to breathe in the rhythm of the measurement and to be at end-expiration during acquisition of spectra.[35] To determine the intensity of the lipid peak, the water signal was suppressed using frequency-selective prepulses. The unsuppressed water resonance was used as internal reference (number of averages = 16). The spectra were fitted with AMARES[36] in the jMRUI software.[37] Values are given as T2-corrected ratios of the CH2 peak, relative to the unsuppressed water resonance (as percentage) according to Hamilton et al..[38]

Statistical Analysis

Data are presented as means and SDs. Data were first tested for normal distribution by using normality test Kolmogorov-Smirnov and Shapiro-Wilk. When normality was met, statistical comparison of the subjects' baseline characteristics and body fat distribution between ethnic groups was performed using independent sample t-test. Non-normally distributed data were compared using the non-parametric Mann-Whitney U test. ANOVA repeated measures were performed to compare differences in changes in liver fat accumulation before and after overfeeding with a high fat diet within and between groups. ANCOVA analysis and multiple regression analysis were applied to assess the main effect of ethnicity on parameters of interests by including potential covariates (determinants). The SPSS program version 20 (SPSS, Chicago, IL) was used for statistical analysis, and statistical significance was set as p < 0.05.