High-fat (HF) diets of 2 weeks have been shown to accelerate gastric emptying (GE). To date, no studies have shown any alteration in GE following shorter HF diets. The aim of this study was to assess if an HF, high-energy diet of 3 days can adapt gastrointestinal (GI) transit, blood lipids and satiety.
Eleven male volunteers participated in a study consisting of three, 3-day interventions each separated by a test day. During the first intervention, volunteers recorded their diet. In the second and third interventions, volunteers repeated their food diary plus either a low-fat yogurt or HF yogurt supplement in randomized order. Test days involved measurement of GE using the 13C octanoic-acid breath-test, mouth-to-caecum transit time (MCTT) using the inulin H2 breath test and satiety using visual analogue scales. Blood samples for measurement of lipaemia were taken using a venous cannula.
MCTT was different between the three test days (P=0.038), with the shortest MCTT following the HF intervention. GE was shortest following the HF intervention. There were no differences in satiety between the interventions. The HF intervention reduced triglycerides, total cholesterol and low-density lipoprotein cholesterol, and increased high-density lipoprotein cholesterol.
This study shows that changes in GI transit owing to an HF diet can occur in a time period as short as 3 days.
The obesity epidemic escalated from about 1980 and has been rising relentlessly ever since (James, 2008). Overconsumption of high-energy-dense food is among the contributors to positive energy balance (Rolls, 2000). Our current environment is characterized by an unlimited supply of convenient, relatively inexpensive, highly palatable, energy-dense foods (Hill and Peters, 1998).
It is well documented that eating a high-fat (HF) meal in comparison to a low-fat (LF) meal reduces gastric emptying (GE) rates (Cecil et al., 1999) to the extent that it can delay GE of the following meal (Clegg and Shafat, 2010a). HF diet intervention studies have shown the opposite effects in that GE, especially of HF food, is accelerated. However, the time frame over which this occurs has never been closely examined. Cunningham et al. (1991) showed that 14 days on an HF diet significantly decreased GE Thalf (the time necessary for half the meal to empty from the stomach) from 147 (88–206) (mean (range)) to 98 (80–116) min and mouth-to-caecum transit time (MCTT) from 360 (130–350) to 240 (200–520) min. Similar findings by Castiglione et al. (2002) showed an HF diet for 14 days increased GE rate of HF food (0.36±0.05 pre-diet vs 0.47±0.03% per min post-diet (mean±s.e.)), but not LF food. Boyd et al. (2003) examined the effect of a duodenal lipid infusion on antropyloroduodenal pressures, following 14 days on either an HF or LF diet. Results showed that the amplitude and tone of the isolated pyloric pressure waves was less following the HF diet compared with the LF diet. This decrease in pyloric contractile force is likely to contribute to acceleration in GE.
The presence of food in the stomach causes both satiation and satiety (Geliebter, 1988), and delaying the emptying of food leads to a prolonged satiety period. There is a school of thought that this increased GE caused by an HF diet can reduce the ability of the stomach to retain chyme causing reduced satiety, increased food intake (French et al., 1995; Boyd et al., 2003) and subsequent obesity. This may explain the accelerated GE that has been observed in obese individuals (Cardoso-Junior et al., 2007; Clegg and Shafat, 2009). Following the 2 weeks HF diet, satiety recorded through food diaries and visual analogue scales (VAS) showed tendencies towards decreased fullness, increased hunger and increased food intake after 2 weeks on an HF diet (French et al., 1995). Following a 2-week HF or LF diet, hunger was greater during an oral fat tolerance test following the HF diet (Boyd et al., 2003) and 2 weeks on an HF diet is also associated with increased maximum-tolerated stomach volume (Park et al., 2007). Taken together, these observations suggest that 2 weeks HF diet can reduce the satiating effect of fat. It is unclear as to whether these effects occur after a shorter adaptation period.
Accelerated GE has significant consequences for the development of diabetes and obesity, independent from dysregulation of appetite (Clegg and Shafat, 2009). Accelerated GE may also result in larger and earlier peaks in both plasma triglycerides and glucose concentrations (Liddle et al., 1988; Darwiche et al., 2001). These changes can lead to the development of insulin insensitivity (Temelkova-Kurktschiev et al., 2000; Ridker, 2008) and atherosclerosis.
No studies to date have examined the effect of an HF diet on GE over a time period of only 3 days. The aim of this study was to examine the effects of short-term HF dietary intervention on GE, MCTT, satiety and post-prandial lipid absorption.
Materials and methods
Following ethical approval from University Research Ethics Committee, 11 healthy male volunteers (24.7±3.1 years; 182±8 cm; 81.7±9.3 kg) were recruited to take part in this study. All procedures were in accordance with the ethical standards of the institution on human experimentation and written informed consent was obtained. Before participation, all volunteers completed a health history questionnaire to ensure that they had no medical ailments that would compromise their participation. None had any history of gastrointestinal (GI) disorder or suffered from GI upset before or during the study.
Volunteers participated in a randomized, single-blind, crossover design, attending the laboratory for three test sessions with 3 days between each test session. During the 3 days, before the first trial (control), volunteers followed their usual diet and completed a weighed food diary. Before the remaining two trials volunteers repeated their food diary with the addition of either an LF or HF yogurt given in randomized order.
The volunteers’ diet was supplemented with either a yogurt only (LF) for 3 days, or yogurt and oil combination for 3 days (HF and high energy). The LF and HF interventions were completed in randomized order. The LF intervention was completed to control for any possible effects of the yogurt itself. The LF intervention consisted of 260 g of yogurt; the HF intervention consisted of 260 g of yogurt and 90 g of sunflower oil combined to form a homogenate mixture; per day for a 3-day period. The LF yogurt consisted of 858 kJ (205 kcal), 8 g of fat, 15 g of protein and 20 g carbohydrate. The HF yogurt consisted of 4250 kJ (1015 kcal), 98 g of fat, 15 g protein and 20 g carbohydrate. Yogurts could be eaten at any time during the day as desired by the volunteer, except during the 12 h before testing.
On the test days, volunteers arrived at the laboratory following a 12 h overnight fast. Upon arrival at the laboratory, measurements of body mass and stature were taken from each of the volunteers. Baseline breath H2 and breath 13CO2 samples were collected. Volunteers were given an HF breakfast meal and 15 min in which to consume it. If the meal was finished before the allocated 15 min, the clock was reset to zero and all subsequent measurements were taken from this point onwards. All times presented here are relative to the end of meal consumption. At 6 h, volunteers were free to leave the laboratory and could eat ad libitum for the remainder of the day while weighing and recording all food consumed.
The breakfast test meal consisted of three pancakes made from 50 g egg (free range large eggs, Dunnes Stores, Dublin, Ireland), 37 g plain white flour (plain flour, Dunnes Stores, Ireland), 65 g whole milk (Dawn Dairies, Limerick, Ireland), 40 g of sunflower oil (St Bernard sunflower oil, Dunnes Stores, Ireland), 12 g of inulin (Raftiline HP, Orafti, Tienen, Belgium) and 100 mg 13C octanoic acid (Euriso-top, Saint-Aubin, France). This was served with 30 g of chocolate spread (Panda, Boyne Valley Foods, Couty Louth, Ireland) and 200 ml water (Kerry Spring Water, Dingle, Ireland). The meal consisted of 2504 kJ (599 kcal), 40.4 g of fat, 14.5 g of protein and 48.2 g of carbohydrate.
MCTT, GE and satiety
Breath H2 (H2 meter, Micro Medical, Chatham, UK) and breath air samples were taken at baseline and every 10 min throughout the 6 h following breakfast. These were used for the analysis of MCTT and GE, respectively. The test meal contained 10 g non-digestible substrate, inulin. Inulin is present as a plant storage carbohydrate in a number of vegetables and plants, including bananas, onion, garlic, wheat and chicory. High-performance inulin is manufactured to remove the shorter chain molecules so that it has an average degree of polymerization of 25 (Niness, 1999). Upon reaching the caecum, inulin is metabolized by colonic bacteria; hydrogen gas is released that can be detected in end exhalation breath. MCTT was defined as a consecutive increase in breath hydrogen over three consecutive readings of at least a cumulative 10 p.p.m. (Bond et al., 1975; Geboes et al., 2003). Geboes et al. (2003) found that MCTT correlated best with lactose 13C-ureide when high-performance raftilin was used and that there was no proportional bias between the two techniques. Inulin has also been shown to effect GE to much lesser extent than other substrates such as lactulose (Clegg and Shafat, 2010b).
Octanoic acid is firmly retained in a standard solid test meal in the gastric environment. However, in the duodenum octanoic acid is rapidly absorbed from the chyme and carried via the portal venous system to the liver. Here it is rapidly and completely oxidized to labelled CO2, which is exhaled into the breath and can be used to give a measure of GE (Ghoos et al., 1993). Following the addition of 100 mg 13C octanoic acid to the test meal, breath samples were taken every 15 min for 6 h. Breath samples for measurement of 13CO2 were analysed using isotope ratio mass spectrometry (ABCA, Sercon Ltd, Chesire, UK) and results were expressed relative to V-PDB, an international standard for known 13C composition. 13CO2 values were expressed as the excess amount in the breath above baseline and converted into moles. Data are displayed as percentage of 13CO2 dose recovered per hour and cumulative percentage 13CO2 recovered over time. Carbon dioxide production was assumed to be 300 mmol m−2 body surface area per hour. Body surface area was calculated using a validated weight–height formula (Haycock et al., 1978). This was then fitted to a GE model developed by Ghoos et al. (1993). For all the data, r2 coefficient between the modelled and raw data was calculated and r2>0.95. From this model, several parameters were measured. Lag phase (Tlag) and half-time (Thalf) were calculated using the formulae derived by Ghoos et al. (1993). Tlag is the time taken to maximal rate of 13CO2 excretion and is equivalent to the time of the inflection point. Thalf is the time it takes 50% of the 13C dose to be excreted. Latency phase (Tlat) and ascension time (Tasc) were derived from formulae developed by Schommartz et al. (1998). Tlat is the point of intersection of the tangent at the inflection point of the 13CO2 excretion curve representing an initial delay in the excretion curve. Tasc is the time course between the Tlat and Thalf, representing a period of high 13CO2 excretion rates. Further graphical representation of these time points can be found in Clegg and Shafat (2010b). Scintigraphic half-time (ThalfS) and lag phase (TlagS) are scintigraphic equivalent values developed by Ghoos et al. (1993) to be comparable to data from scintigraphic GE measurements. The total percentage dose recovered, the peak in percentage dose recovered per hour and the time at which the peak (Tpeak) occurred were obtained from the raw (unmodelled) data.
Satiety was measured using a 150 mm VAS to detect changes in hunger, thirst, desire to eat, tiredness, fullness and cold every 30 min throughout the 6 h. Variables thirst, tiredness and cold were used to distract volunteers from analysis of their satiety.
Blood samples were collected at baseline and following breakfast every 10 min for the first 30 min followed by every 30 min until 6 h. Serum separating clot activator tubes were allowed to clot at room temperature (for 15 min) before centrifugation began at 3500 r.p.m. for 5 min in 4 °C. Serum was removed and transferred to 1.5 ml plastic vials, and these were stored at −70 °C for subsequent biochemical analysis.
Methods for biochemical analysis of blood glucose, total cholesterol, high-density lipoprotein cholesterol (HDL-C) and triglycerides has been described previously (Clegg et al., 2007). Estimates of low-density lipoprotein (LDL)-cholesterol concentration were calculated using the Friedewald formula (Friedewald et al., 1972). Coefficient of variation was less than 2.5% for all blood glucose and lipid concentrations.
Food diaries were analysed for macronutrient content and energy intake using Compeat Pro Version 5.8 using McCance and Widdowson's 6th Edition food tables (Nutrition Systems, Grantham, UK).
Statistical analysis was performed using the SPSS version 15.0 (Surrey, UK). All data except VAS data were tested for normal distribution. Dependent on the results of these tests, data were examined using either a repeated measures analysis of variance or the Friedman test for non-parametric data. Trials were compared by examining the contrasts within the analysis of variance or by the Wilcoxon's signed rank test. VAS data were transformed by natural log and analysed using a two-way repeated measures analysis of variance with two within-subject factors—time and condition. Comparisons between the trials were carried out by examining the contrasts within the analysis of variance. All data are expressed as mean±s.d., unless otherwise stated and the significance level was set at P<0.05.
There were significant differences in dietary intake between the interventions for energy (F(2,10)=255.792, P<0.001), fat (F(2,10)=2232.310, P<0.001) and carbohydrate (F(2,10)=18.612, P<0.001), but not for protein (F(2,10)=3.989, P=0.061). Intakes of energy, carbohydrate and fat were greatest following the HF intervention. Daily intake of dietary fat was 88 g higher on the HF intervention, indicating that volunteers were compliant with the dietary manipulation (Table 1).
There were significant differences between the parameters Tlat (χ2(2,10)=11.091, P=0.004) and Tpeak (F(2,10)=3.523, P=0.049) between the three interventions. The LF intervention did not affect GE. The HF intervention accelerated the initial stages of GE as measured by Tlat, and Tpeak compared to the control and LF intervention (Table 2). There were no differences between the other parameters of GE (P>0.05).
Differences existed in MCTT between the three interventions (F(2,10)=4.186, P=0.038). MCTT decreased significantly between the control and HF intervention (control 280±60 min; HF intervention 226±84 min; P=0.025); no differences existed in MCTT between the LF intervention (291±47 min) either in the control or the HF intervention (both P>0.05). An example of one volunteers’ breath hydrogen concentrations can be seen in Figure 1.
Plasma triglyceride concentrations differed following the interventions (F(2,10)=5.176, P=0.015) and across time (F(2,10)=16.745, P<0.001). The differences exist between control and HF (P=0.004) and between LF and HF (P=0.025), but not between control and LF (P>0.05). Triglycerides were lower following the HF intervention, both in the fasted state and post-prandially.
For total cholesterol, there were differences between interventions (F(2,11)=19.299, P<0.001) and across time (F(2,11)=3.206, P<0.001). The differences exist between base and HF (P<0.001) and control and HF (P<0.001), but not between LF and HF (P>0.05). Total cholesterol was lowest following the HF intervention.
For LDL-cholesterol, there were differences between interventions (F(2,10)=4.165, P=0.031) and across time (F(2,10)=16.068, P<0.001). HF LDL was lower than control (P=0.008), but no differences existed between control and LF or LF and HF (P>0.05).
For HDL-C, there were no differences between the interventions (F(2,10)=2.597, P=0.99), but there was a difference across time (F(2,10)=5.324, P<0.001). HF HDL was significantly higher than LF (P=0.045), but there were no differences between the other tests (Figure 2).
For glucose, there were no differences between the interventions (P>0.05), but there were differences across time (F(2,10)=24.745, P<0.001). Glucose increased over the first 30 min post-prandially and then declined in all three tests. Plasma glucose concentration remained relatively stable around fasting values from 60 min to 6 h in all tests.
Results from the VAS showed that there was no difference between the three interventions for any of the parameters of hunger, desire to eat and fullness (P>0.05). However, there was an effect of time for all three parameters, with volunteers becoming less satiated over the 6-h period for hunger (F(2,10)=39.263, P<0.001), desire to eat (F(2,10)=42.919, P=0.001) and fullness (F(2,10)=29.897, P=0.005). There were no significant differences in energy or macronutrient dietary intakes on the evening of each test session (P>0.05; Table 2).
In summary, 3 days HF diet accelerated the initial phase of GE, reduced MCTT, and decreased fasting and post-prandial lipaemia without affecting hunger sensation or food intake.
The results of this study showed that GE Tlat and the Tpeak were shorter following the HF intervention. These results extend findings from studies that have examined the effects of a 2-week HF diet (Cunningham et al., 1991; Castiglione et al., 2002). There are two factors that may explain why current data demonstrate acceleration in GE, whereas others (Cunningham et al., 1991) found little or no differences over a similar time period: first, the intensity of nutritional change, and second, fatty acid-specific adaptation. The percentage energy provided from fat in the Cunningham et al. (1991) study was 55% and energy intake was 19.26MJ. In the current study the energy intake from fat increased by 15% from 31 to 46% and daily energy intake during the HF intervention was 15.46 MJ. In the Cunningham et al. (1991) study, the entire diet of the volunteers was changed compared with this study where the volunteers’ diet was supplemented with yogurt and oil. Hence, the volunteers in Cunningham et al. (1991) study were subjected to a wide variety of fatty acids. In this study, the volunteers were subjected to a narrow range of fatty acids from sunflower oil, primarily linoleic acid (LA; 18:2). It is known that the GI tract responds very specifically to different fatty acids (Robertson et al., 2002; Maljaars et al., 2009). Any adaptation is hence going to be specific to the effects of that fatty acid. It is interesting to speculate on the potential mechanisms underlying the adaptation of GI transit to chronic fat intake. Fat sensing in the gut has, for many years, been an established observation without a mechanistic explanation. Recently, G-protein-coupled receptor 120 was cloned (Moore et al., 2009), shown to be sensitive to fatty acids (Burns and Moniri, 2010) and to induce CCK secretion in vivo (Tanaka et al., 2008). Several steps in this sensory cascade may explain desensitization to fat, including, but not limited to, receptor expression downregulation, expression of different isoforms of the receptor (Burns and Moniri, 2010) and further intracellular steps in the enteroendocrine cells. Furthermore, G-protein-coupled receptor 120 is just one member of a large family of luminal receptors sensitive to fat and it may well be the anatomical distribution and relative expression of these receptors that determine GI sensitivity to ingested fat. In this study, the major component of fat in the test meal was also in the form of the same oil—sunflower oil. In this way, the volunteer's GI system only had to become accustomed to a narrow range of fatty acids, primarily LA (Cummins et al., 1967). The smaller dietary change in this study may allow faster adaptation of the GI tract as it only had to adapt to a specific set of fatty acids in this controlled HF diet.
As Tlat represents the first portion of the meal to empty from the stomach, it corresponds with evidence, which suggests that the feedback mechanisms from the small intestine have been disrupted or in this case perhaps delayed (French et al., 1995). The ileal brake and intestinal transit time have been proposed as targets for the modulation of appetite, food intake and adiposity (Maljaars et al., 2008). MCTT was accelerated in the HF group compared with the control group (P=0.025). As MCTT represents the time that the head of the meal reaches the caecum, this is in keeping with the GE data that found that the first portion of the test meal, the latency phase, emptied faster following the HF intervention. The shortening of MCTT is also in agreement with animal studies showing desensitization of the ileal brake to chronic infusion of fat (Brown et al., 1994). Other mechanisms such as duodenal brake may also explain these observations (Shahidullah et al., 1975). The number and distribution of fatty acid-sensitive receptors along the GI tract (Engelstoft et al., 2008) means that classifications into gross anatomical structures may be less useful than actual nutrient sensitivity. In humans, Cunningham et al. (1991) found that MCTT was shorter after 14 days on an HF diet. These data indicate that transit as far as the large intestine is faster after only 3 days on an HF diet. What implication does this have for post-prandial lipaemia?
Surprisingly, fasted and post-prandial triglycerides and cholesterol (total and LDL) were reduced by 3 days on an HF diet. The main source of fat added to the volunteer's diet was sunflower oil, whose primary fatty acid is LA (Cummins et al., 1967). LA is a polyunsaturated fatty acid that has been shown to lower serum levels of LDL cholesterol. The substitution of polyunsaturated fatty acid (the vast majority being LA, varying from 0.6 to 28.8% energy) for carbohydrates has more favourable effects on the total:HDL-C ratio than any class of fatty acids (Mensink et al., 2003; Harris, 2008). Although GE was accelerated following the HF intervention, the lack of change in satiety and lipaemia implies that HF diet is unlikely to impact negatively on an individual's health.
The effects of HF diets and the intestinal mechanisms leading to dysregulation of appetite have been reviewed recently (Little et al., 2007). In this study, satiety values and food intake were no different on the test day following each intervention. Castiglione et al. (2002) similarly found no difference in satiety following either HF or LF test meals or before and after a 2-week HF diet. Boyd et al. (2003) found that following a 2-week HF or LF diet, hunger was greater during an oral fat tolerance test following the HF diet; however, this did not affect subsequent food intake. Similarly, ileal infusion of fatty acids 18:1 or 18:2, but not 18:0, reduced appetite but not food intake (Maljaars et al., 2009). Park et al. (2007) found no changes in GI transit or appetite following an HF diet intervention. Using food diaries, French et al. (1995) found increases in food intake following a 2-week HF diet. However, in this study there were no changes in satiety and food intake and it appears that although 3 days is sufficient to change GE, it is not sufficient an adaptation period to alter satiety.
These data show that a 3-day HF diet is sufficient to accelerate GE and MCTT of an HF breakfast. This, however, did not affect satiety and food intake. This is the first study to show changes in GI transit over only 3 days and indicates that short-term changes in diets can impact GI processing. However, the lipaemia data indicate that in the short term the 3-day diet was not detrimental to cardiovascular health and cholesterol and triglyceride levels were improved. A longer adaptation may influence lipaemia and satiety as has been previously shown. Like many studies on this topic, the current study consisted of a high energy diet, as well as a HF diet. There is scope here to further understand if the acceleration in GI transit occurs to such an extent with a solely HF (not high-energy) diet in such a short time span. However, this study provides a valuable insight into how GI processing is altered following 3 days dietary modification.
We gratefully acknowledge the Irish Research Council for Science, Engineering and Technology (IRCSET) for their support of this project. Inulin was donated by Orafti, Belgium.
About this article
Effect of cinnamon on gastric emptying, arterial stiffness, postprandial lipemia, glycemia, and appetite responses to high-fat breakfast
Cardiovascular Diabetology (2011)