Effect of cinnamon on gastric emptying, arterial stiffness, postprandial lipemia, glycemia, and appetite responses to high-fat breakfast

Nine apparently healthy subjects (3 male, 6 female; age 26.2 ± 3 years; mass 66.5 ± 11.3 kg; BMI 22.4 ± 2.5 kg/m²; body fat 22.2 ± 6.5%) consented to participate in the study, which was approved by the Local Research Ethics Committee. Subjects had no history of GI-related complaints, CVD or diabetes and were not currently taking antioxidant or lipid-lowering medication. Subjects were non-smokers and were recreationally active (performing < 60 min of moderate to intense exercise per day, on every day of the week) determined using a 3 month exercise questionnaire. Fasting blood lipid, glucose and blood pressure (BP) levels were all within the normal limits. Subjects were required to rate at least 50% of the food items that would be presented in the buffet meal as 5 or higher in a food preference questionnaire for study inclusion. Two subjects were identified as restrained eaters (scoring > 12 on the eating restraint section (factor 1) of the Three-Factor Eating Questionnaire [22]).

Each subject completed two separate 1-day trials separated by 28 days: cinnamon and placebo (wheat flour) supplementation. Trials were conducted in a single-blind, placebo controlled randomized fashion and were identical apart from the content of the capsules. During the 3 d before the first trial, subjects recorded all of the food and drink that they consumed and repeated this diet before the second trial. Energy, macronutrient and fatty acid intake were calculated using CompEat Pro Nutritional Analysis software (Version 5.8; Nutrition Systems, Grantham, UK). Physical activity was also logged during this time period. Subjects were instructed to minimize their consumption of foods with naturally high ¹³C abundance on the day preceding each trial and were asked to refrain from alcohol consumption and vigorous exercise. Apart from these requirements, subjects ate and exercised ad libitum between the two testing periods.

Subjects reported to the laboratory after a 12 h overnight fast. Following 15 min interval of supine rest, measurements of arterial stiffness (see measurements) and BP were obtained and a baseline blood sample was collected. Subjects returned to the seated position and after a 10 min interval, baseline breath samples were taken and visual analogue scale (VAS) questionnaires were completed (t = -15 and -10 min). Once baseline measurements were taken, subjects consumed the test meal within 15 min (see below). Upon completion of the meal (t = 0 min), sequential postprandial measurements of GE, appetite sensations, arterial stiffness, BP and plasma glucose and lipids responses were taken. Expired breath samples were collected at t = 0 min, every 5 min for the first half hour after meal consumption and thereafter in 15-min intervals from t = 30 until 360 min for the detection of ¹³CO₂ (see measurements). VAS questionnaires were administered after consumption of breakfast (t = 0 min) and every 30 min until t = 360 min. Arterial stiffness, BP measurements and blood samples were collected hourly from t = 60 to t = 270 min (t = 240 min for blood samples). At the end of postprandial assessments (t = 360 min), subjects were presented with an assortment of cold lunch-type buffet foods. The meal comprised of six slices of wholegrain bread (235 g), six slices of white bread (235 g), 115 g sliced ham, 115 g sliced chicken, 100 g grated cheddar cheese, 60 g lettuce, 125 g cherry tomatoes, 80 g sliced cucumber, 90 g sweet corn, 100 g coleslaw, 30 g butter, 30 g mayonnaise, 30 g relish, 415 g strawberry yoghurt, 6 biscuits (75 g), 120 g chocolate Swiss roll, 60 g crisps and 300 ml water. The total energy content of the buffet meal was 16504 kJ. The selection presented was in excess of anticipated consumption. Subjects were given 30 min (i.e. t = 360 - 390 min) to eat ad libitum until 'comfortably full'. After ingestion of the buffet meal, subjects completed another VAS questionnaire and were then free to leave the laboratory. Additional VAS questionnaires were administered after consumption of the test meal and buffet meal (t = 0 and 390 min) to evaluate meal palatability as well as sensations of nausea and well-being. They recorded their subsequent food intake for the remainder of the test day.

The test meal consisted of three pancakes (36 g flour, 44 g egg, 70 g whole milk, 30 g sunflower oil) served with 20 g chocolate spread and 300 ml of water. The test meal was enriched with 150 μl ¹³C octanoic acid (Cambridge Isotope Laboratories, Andover, MA, USA), which was solubilized in the egg yolk prior to cooking. After homogenizing the yolk, it was mixed with the other ingredients to ensure uniform distribution of the label throughout the pancake batch. The test meal provided 2646 kJ of energy and consisted of 42 g carbohydrate, 46 g fat and 13 g protein. Subjects were instructed to ingest 8 gelatin capsules (four directly before and after meal) totaling 3 g Cinnamon powdered spice (Cinnamomum zeylanicum; Schwartz, UK) or a wheat flour placebo (Odlums, Ireland).

Data were checked for normal distribution before statistical analysis was preformed. Appetite sensation data was transformed by natural log. Fasting measurements of arterial stiffness, blood pressure and plasma lipid and glucose concentrations were calculated as the mean of the values collected at baseline (~t = -30 to -20 min). Baseline GE and appetite sensation scores were calculated as the mean of values collected at t = -15 and -10 min. Blood biomarkers, appetite sensations and arterial stiffness data were analyzed using a two-way (time × supplement) repeated-measures ANOVA. Incremental area under the glycemic curve (IAUC) was calculated using the trapezoidal rule by subtracting baseline values from measured plasma glucose concentrations. Relationships between variables were compared using Pearson correlations. Paired-sample t tests were used to compare GE parameters, habitual diet, physical activity intensity and duration as well as food intake at the buffet meal (quantity, energy consumed and macronutrient distribution). Statistical significance was established at the p < 0.05 level and the mean values ± SD are reported. All statistical analyses were carried out using SPSS-version 16.0 (SPSS, Inc., Chicago, IL, USA). Power calculation was conducted for gastric emptying T_half the primary endpoint evaluation. A sample size of 9 subjects was necessary to detect a 15.8% change in GE rate [13] in a two-sided paired Students' t-test with alpha set at 5% and a power of 80%. Trial registration number: at http://​www.​clinicaltrial.​gov: NCT01350284.

No significant effect of cinnamon supplementation was observed on gastric emptying parameters (Table 1).

We were unable to induce significant changes in GE using 3 g of cinnamon. The test meal (65% of energy from fat) was mainly from sunflower oil, which contains approximately 70% linoleic acid (C18:2n-6, a PUFA [31]). Long-chain fatty acids have a potent inhibitory effect on GE rate [20, 32] and have also been shown to increase CCK and GLP-1 concentrations [21]. We propose that cinnamon does not delay GE over and above the effects of the fat content of the meal. Furthermore, we found similar postprandial glycemic and lipemic responses under both conditions. However, it should be noted that we were unable to measure a hyperglycemic or hyperlipidemic state. Studies which employed the largest doses of cinnamon relative to carbohydrate in the test meal (carbohydrate/cinnamon ratio of 15 or lower [9, 13]) appear to have had the most potent effects on reducing postprandial glycemia [33]. In spite of the current high ratio of 14, we did not achieve a significant blood glucose-lowering effect. This is possibly due to glucose absorption from the small intestine being affected by the fat content of a meal [34].

Recent data indicates that the addition of 3 g cinnamon to a low-fat rice pudding test meal had no significant effect on GE rate or postprandial glycemia in healthy individuals [35]. However, cinnamon did significantly lower serum insulin levels and increase GLP-1 concentrations, a GI peptide which has been shown to increase glucose-dependent secretion of insulin, delay GE and reduce glucose absorption and postprandial glycemia [36, 37]. When added to the same test meal, 6 g cinnamon significantly delayed GE and reduced postprandial glycemia but the decrease in blood glucose concentration was more apparent than the delay in GE rate suggesting that GE cannot be the sole mechanism explaining lower blood glucose responses following cinnamon ingestion [13]. Agreeing with the findings of others [13, 33, 34], we found that cinnamon did not influence appetite sensations or subsequent food intake, probably as a result of similar GE rates between conditions [38]. Together with data presented in the current study, cinnamon is unlikely to be relevant in affecting the postprandial response to HF meals.

Assessment of previous day dietary intake indicated that a higher intake of C16:1, C20:1 and total n-3 was associated with a shorter GE T_half of the HF meal supplemented with the wheat flour placebo. To our knowledge, this is the first observation of specific dietary fatty acids from preceding diet affecting GI transit in humans. A single meal, supplemented with n-3 PUFAs, was less capable of triggering GLP-1 and CCK compared to other fats, resulting in a more rapid GE of a HF breakfast [39] while others showed that n-3 PUFA fish oil reduced CCK release and gallbladder contraction without affecting GE [40]. Both GLP-1 and CCK are putative mediators of the ileal brake, a feedback mechanism responsible for delaying transit and facilitating digestion, in response to lipids in the distal GI tract [41]. Our current findings extend these observations to illustrate that even short-term intake of n-3 fatty acids is associated with faster GE rates, in a population who were not eating a HF diet. This means that mechanisms apart from acute release of GLP-1 and CCK, must mediate the effects of specific fatty acids on GI transit. Recently, a 3-day HF yoghurt supplementation, rich in C18:2n-6 accelerated the GE rate of a test meal rich in the same fatty acid [42]. It is interesting to note that background intake of C18:2n-6, which was high in the test meal, did not show a strong association with GE of the meal. Our observations suggest that GE T_half of a HF meal is not just specifically affected by a background intake of that specific fatty acid and that the process of adaptation to a HF diet may involve mechanisms other than desensitization to a specific fatty acid. It is likely that different adaptations are continuously taking place in the gut, in response to the balance of fatty acids in the diet. It is tempting to speculate about the potential mechanisms for fat sensing and adaptation following the recently sequenced GPR120 protein, expressed on intestinal cells, and demonstrated to be differentially sensitive to different fatty acids [43].

One of the principle findings of this study was that no changes were observed in relation to the measures of vascular function. Interestingly, there was a main effect for time indicating a decrease in SI which at face value appears paradoxical given the transient impairment in vessel function following the ingestion of a HF meal [44, 45]. This apparent contradiction might be explained by the fact the test meal we used may not have been of sufficient energy and, in particular, fat content to evoke a change in vessel function. In most of the related literature the postprandial TAG concentration associated with vascular dysfunction (~2.0 mmol/l; [17, 18, 44]) is double than that presently observed. Moreover, the HF meal in the current study contained 297 - 1742 kJ less energy (and 14 - 34 g less fat) than other similarly designed investigations [17, 45‐47]. Such discrepancies in meal composition further highlight the need for a standardized, physiologically relevant HF meal [48] to be used in future corresponding studies, similar to the OGTT.

Due to its polyphenolic nature, cinnamon is thought to exhibit antioxidant properties [3, 49] which may be anti-atherogenic. It is proposed that the impairments in blood vessel function following the ingestion of HF loads are perpetrated via an oxidative stress mechanism that can increase the unwanted consumption of NO and favor the formation of further ROS, such as peroxynitrite (ONOO^-) [17‐19, 28, 44]. In the present study cinnamon ingestion had no apparent effect on indices of oxidative stress as a main effect for time was observed for LOOHs.

It has been documented that even high-normal fasting glucose levels can aggravate arterial stiffness [50] but when cinnamon was added to rats fed a high-fat high-fructose diet hepatic glycogen, hepatic insulin receptors and glut4 transporter in muscle tissue all increased [51]. Despite no reported change in arterial stiffness using our physiologically relevant high-fat meal, the possibility therefore exists that cinnamon could modulate stiffness by affecting (hepatic) glycemic control and thus this relationship merits further scrutiny.

Emerging research postulates the existence of a diurnal variation in endothelial function [52]. Fluctuations in the competitive balance between intrinsic local vasodilator function and sympathetic nervous system (SNS) α-adrenoreceptor-mediated vasoconstriction have recently been proposed as one potential mechanism to explain such findings [53]. Given the conceptual relationship between endothelial function and arterial stiffness, and the observed main effect for a decrease in HR (as an indirect measure of SNS activity), it is tempting to speculate that this explanation may in some way account for the SI data in the current investigation. Conversely, this was tempered by the fact that no changes in RI were recorded.

This study may have been underpowered and therefore a small effect below the detection threshold of the study cannot be ruled out. Three grams of cinnamon was used as because it was shown to have a similar chronic effect on fasting serum glucose and lipid profiles as 6 g [8]. However, recent evidence suggests a dose-dependent relationship for cinnamon consumed and the delay in GE [13, 35].