| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Diabetes and Obesity Research Program, Garvan Institute of Medical Research, (J.R.G., K.S., D.J.C., L.V.C.), and Diabetes Center (L.V.C.) and Department of Cardiology (C.S.H.), St. Vincent’s Hospital, Sydney, Australia 2010
Address all correspondence and requests for reprints to: Prof. Lesley Campbell, Diabetes Center, St. Vincent’s Hospital, 372 Victoria Street, Darlinghurst 2010, Australia. E-mail: l.campbell{at}garvan.org.au.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Because alcohol is usually consumed with a meal, and humans spend the majority of their time in the postprandial state, it may influence cardiovascular risk via favorable effects on postprandial metabolic factors. Aberrant postprandial lipid metabolism, specifically postprandial hypertriglyceridemia, is an independent risk factor for atherosclerotic cardiovascular disease (6). The Atherosclerosis Risk in Communities study reported that subjects consuming 14 g or more of alcohol/d had a 27% lower postprandial triglyceride response to a nonalcohol-containing meal compared with subjects consuming less than 14 g/d (7). Although this study examined the chronic effect of moderate drinking on postprandial lipemia, few have addressed the acute effect of a small amount of alcohol (similar to that associated with cardioprotection in chronic studies) on postprandial lipids and cardiovascular risk factors.
Inflammatory processes are thought to play an important role in the development and progression of atherosclerotic plaques (8). Alcohol may reduce cardiovascular risk in moderate drinkers via effects on antiinflammatory pathways. Consistent with this hypothesis, moderate alcohol consumers are reported to have lower levels of circulating inflammatory markers, such as C-reactive protein (CRP), than abstainers (9, 10, 11). However, the acute effect of a small amount of alcohol on postprandial CRP levels has not yet been reported. There has also been interest in adiponectin, a novel adipocyte-derived hormone reported to have antiinflammatory and antiatherogenic properties (12). It is possible that moderate alcohol consumption down-regulates endothelial cell inflammatory processes through effects on adiponectin secretion.
Arterial stiffness, or reduced compliance, is an independent predictor of cardiovascular mortality (13, 14). The augmentation index (AIx) is a validated noninvasive measure of arterial stiffness. It is considered an important marker of coronary artery disease (15, 16) and is an independent predictor of total and cardiovascular mortality (14). Reductions in arterial stiffness may therefore contribute to lower cardiovascular risk in moderate alcohol consumers.
Obesity, particularly central abdominal obesity, is a strong predictor of dyslipidemia, insulin resistance, type 2 diabetes, and cardiovascular disease (17, 18). We recently reported that moderate alcohol consumers have lower total and central adiposity than abstainers (5) and that lower body fat partly mediates some of the metabolic benefits of moderate alcohol consumption (4). Although alcohol-induced changes in energy expenditure may contribute to these observations, conclusions regarding the effect of moderate alcohol consumption on postprandial energy balance cannot be made from previous studies, because most have used a large quantity of alcohol, significantly greater than that associated with metabolic benefits and cardiovascular protection in epidemiological studies.
The primary aim of this study of postmenopausal women was to examine the effect of a small amount of alcohol on 1) postprandial glucose, insulin, lipid, CRP, and adiponectin excursions; 2) AIx; and 3) energy expenditure and substrate oxidation. A major limitation of previous studies of postprandial lipid metabolism is failure to consider the confounding effect of adding carbohydrate to a high fat meal. Our group recently reported that postprandial glucose and insulin levels were higher, and triglyceride and fatty acid levels were lower, when a large vs. a small amount of carbohydrate was added to a high fat meal (19). An additional confounding factor in meal studies is insulin resistance, because it has been reported that postprandial metabolic responses are influenced by insulin sensitivity (20). Therefore, secondary aims were to determine whether the acute effect of alcohol on these metabolic factors was affected by the addition of carbohydrate to a high fat meal and whether postprandial responses were different in insulin-sensitive and insulin-resistant subjects.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Twenty healthy postmenopausal women were recruited through advertisements placed in local newspapers, on the hospital campus, and at public seminars. Smokers, subjects with known type 2 diabetes or coronary artery disease, and subjects taking lipid-lowering medications, multiple antihypertensive agents, or hormone replacement therapy were excluded. All subjects had normal liver function tests. Median habitual alcohol intake was 4.7 g/d (range, 0–23 g/d). Subjects were classified as either relatively insulin-sensitive (n = 11) or insulin-resistant (n = 9) based on their homeostasis model assessment insulin resistance (HOMA-IR) score (21) being less than (mean HOMA-IR, 2.0 ± 0.2) or greater than (mean HOMA-IR, 3.8 ± 0.3) the mean of the whole group, respectively. Subjects with fasting plasma glucose levels of 6.1 mmol/liter or more (n = 3) underwent a 75-g oral glucose tolerance test after completion of the study. All three subjects had impaired glucose tolerance. All subjects provided written informed consent. The study was approved by the research and ethics committee at St. Vincent’s Hospital.
Study design
Four meal studies [two low carbohydrate (LC) and two high carbohydrate (HC) high fat meals] were conducted on 4 separate days at least 1 wk apart. One subject only attended for the two LC meals because of time constraints. Subjects were asked to fast from 2200 h the night before each study and to refrain from vigorous physical activity and alcohol for 48 h before the study. Four subjects were taking one antihypertensive agent and omitted medication on each morning of study. Subjects attended the Clinical Research Facility at the Garvan Institute of Medical Research at 0800 h after minimum physical activity. Weight, height, and waist circumference were measured with the subject in a hospital gown. Using a nonalcohol-containing swab, an iv cannula was inserted into a large antecubital vein for repeated blood sampling and was kept patent by infusion of 0.9% saline (
120 ml/h). Subjects remained supine for the duration of the study. After a 15-min rest, baseline fasting samples were taken for glucose, insulin, total and high density lipoprotein (HDL) cholesterol, triglycerides, nonesterified fatty acids (NEFAs), blood alcohol, liver function tests, CRP, and adiponectin. In 11 subjects, fasting indirect calorimetry was then performed for 30 min, followed by two baseline measurements of AIx, which were averaged in analyses. Repeat samples for fasting plasma glucose, insulin, total and HDL cholesterol, triglycerides, and NEFAs were taken before the meal was consumed; the two baseline samples were averaged when calculating postprandial responses for these biochemical variables. The test meal was then consumed over 20 min. No other food or drink was allowed after the meal except for less than 600 ml water. Blood samples, indirect calorimetry, and duplicate AIx measurements were repeated at regular intervals for 360 min, as outlined in Fig. 1
.
|
High fat meals were prepared on the day of study and matched for fat and protein contents, but differed in carbohydrate content and, consequently, calories. LC high fat meals (1084 kcal, consisting of 85 g fat, 45 g protein, and 23 g carbohydrate) were consumed during visits 1 and 2. Fifteen grams of alcohol (90%, vol/vol) in orange juice was randomly added to 1 of the study days (LC-alc meal) and was ingested over 15 min, 5 min after the first mouthful of food. On the day when alcohol was not consumed (LC-con meal), a similar volume of orange juice alone was provided. HC high fat meals (1385 kcal, consisting of 88 g fat, 37 g protein, and 104 g carbohydrate) were consumed during visits 3 and 4, with (HC-alc meal) and without (HC-con meal) alcohol.
Indirect calorimetry
Indirect calorimetry (Deltatrac Metabolic Monitor, Datex Instrumentarium, Helsinki, Finland) was performed for four 30-min periods as outlined in Fig. 1
. Gas exchange rates were recorded at minutely intervals. An equilibrium period of 10 min was allowed, and the last 20 min were averaged for calculation of energy expenditure (kilocalories per day), respiratory quotient (RQ; ratio of CO2 production to O2 consumption) and fat and carbohydrate oxidation, based on equations previously described (22). Urinary nitrogen excretion was estimated by a constant based on body weight (0.14 gN/kg·d), as previously reported by our group (23). Measured O2 consumption and CO2 production rates were adjusted for alcohol oxidation and were used to calculate RQ and substrate oxidation on the study days during which alcohol was consumed. For reasons discussed in detail previously (24, 25, 26), we assumed that 100% of the ingested alcohol dose was uniformly oxidized between the end of the meal and the time when blood alcohol levels reached 0 (120 min). Rates of O2 consumption and CO2 production measured at 30–60 min were therefore adjusted for alcohol oxidation, assuming that oxidation of 1 mol alcohol uses 67.2 liter O2 and produces 44.8 liter CO2 (25). The day to day coefficients of variation (CVs) for repeated fasting measurement of energy expenditure and RQ were 4.3% and 3.0%, respectively.
AIx
Brachial blood pressure (BP) was recorded using an Automatic Oscillometric Digital Blood Pressure Monitor (OMRON HEM-705CP, OMRON Corp., Tokyo, Japan), with the subject sitting at approximately 45° in bed. Applanation tonometry of the radial artery was used to measure AIx (SyphgomoCor, AtCor Medical, Sydney, Australia). A high fidelity transducer was placed over the radial artery, compressing, but not occluding, the artery against the underlying radius. The central arterial waveform was derived from the peripheral waveform using a validated transfer function (27). AIx was calculated as the ratio of the difference between the second systolic peak and the diastolic pressure and the difference between the first systolic peak and the diastolic pressure x 100% (28). At least two readings of central systolic and diastolic BP, heart rate, and AIx were averaged at each time point. AIx was measured throughout the postprandial period as outlined in Fig. 1. The day to day CV for repeated fasting measurements of AIx on 4 separate days was 5.3%.
Dual energy x-ray absorptiometry
Whole body dual energy x-ray absorptiometry (Lunar DPX GE-Lunar, Lunar Corp., Madison, WI) was used to analyze body composition. As previously described (18), a central abdominal window was outlined manually extending from the upper border of L2 to the lower border of L4 and laterally to the outer margin of the rib cage. The fat in this window was measured and expressed as a mass [central abdominal fat (CAF) in kilograms] and as a percentage of the total soft tissue content in this area (%CAF). Although this window contains both intraabdominal fat and sc abdominal fat, it excludes 30% of the latter, has a relatively high intraabdominal fat and a low sc abdominal fat content, and is strongly inversely related to insulin sensitivity (18). We have previously demonstrated a CV of less than 6% for %CAF, based on data from 10 female subjects scanned on four separate occasions (18).
Biochemical variables
Plasma glucose was determined immediately by the glucose oxidase method using a YSI glucose analyzer (model 2300 STAT PLUS 230V, YSI, Inc., Yellow Springs, OH). The remaining plasma and serum were stored at –20 and –80 C and assayed at a later date. Commercial RIAs (Linco Research, Inc., St. Charles, MO) were used to analyze serum insulin and adiponectin. The sensitivity of the adiponectin assay was 0.5 µg/ml, and the intra- and interassay CVs were 1.8–6.2% and 6.9–9.3%, respectively. Serum total cholesterol, HDL cholesterol, and triglycerides were determined spectrophotometrically at 490 nm by enzymatic colorimetry (Roche, Basel, Switzerland). NEFAs were determined spectrophotometrically at 550 nm, also by enzymatic colorimetry (Wako, Inc., Osaka, Japan). Inter- and intraassay CVs were less than 10% for these assays in our laboratory. CRP was measured using a Synchron LX System Chemistry Analyzer (Beckman Coulter, Inc., Sydney, Australia). The highly sensitive assay is based on the near-infrared particle immunoassay rate methodology. The CV for this assay was 2.2–3.7%, with a sensitivity of 0.2 mg/liter. Results below 0.2 mg/liter were entered as 0. As the 99th percentile of the CRP distribution is healthy subjects is 10 mg/liter (29), data for subjects with baseline CRP values greater than 10 mg/liter at any visit (n = 4) were excluded from CRP analyses. Liver function tests were determined using routine methods on an Olympus AU2700 analyzer (Integrated Sciences, Sydney, Australia). The plasma alcohol concentration was measured using an enzymatic method optimized for measurement of low alcohol concentrations, also on an Olympus AU2700 analyzer.
Statistical methods
Data are the mean ± SD for descriptive data and the mean ± SEM for comparisons between groups. Analyses were performed using StatView 5.0 (SAS Institute, Inc., Cary, NC). The postprandial incremental area under the curve (iAUC) was calculated by the trapezoidal method, inclusive of the basal period (–30–0 min), by subtracting baseline values extrapolated over 390 min from the total postprandial area. Comparisons between results were performed using the paired t test (for normally distributed data) and the Wilcoxon signed rank test (for skewed data). The ratio between insulin and glucose AUC (AUCINS/AUCGLUC) was calculated as an indicator of insulin secretion (30). Correlations between variables were expressed as Pearson’s or Spearman’s correlation coefficients. P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. Alcohol intake was not significantly related to baseline fasting metabolic factors (data not shown). Baseline CRP concentrations were strongly related to weight (r = 0.57; P = 0.02), body mass index (r = 0.74; P = 0.002), total body fat (r = 0.59; P = 0.01), percent total body fat (r = 0.57; P = 0.02), CAF (r = 0.72; P = 0.002), and %CAF (r = 0.76; P = 0.001).
|
Effects of carbohydrate on postprandial glucose, insulin, and lipid responses
As expected, postprandial glucose and insulin iAUC were higher after the HC-con vs. LC-con meal (P < 0.0001). In contrast, NEFA iAUC was lower (P = 0.01). There were no differences between the control meals in triglyceride (P = 0.85), total cholesterol (P = 0.43), or HDL cholesterol (P = 0.70) responses.
Postprandial blood alcohol levels
No subject had alcohol detected in the blood at baseline on any study day. Postprandial blood alcohol levels are shown in Fig. 2. Blood alcohol levels peaked at 15 min and reached baseline values by 120 min.
|
Plasma glucose levels peaked at 30 min after the LC-con, LC-alc, and HC-alc meals and at 60 min after the HC-con meal (Fig. 3). Postprandial insulin responses followed a similar pattern (Fig. 3). Postprandial glucose iAUC was lower (P = 0.04), and insulin iAUC was higher (P = 0.07) after the LC-alc vs. LC-con meal (Fig. 3). AUCINS/AUCGLUC was also higher after the LC-alc meal (P = 0.006). In contrast, alcohol had no significant effect on these variables after the HC meal (Fig. 3).
|
). Alcohol augmented the postprandial increment in triglyceride levels after both the LC-meal [iAUC, 33,038 ± 3,986 vs. 22,764 ± 3,366 mg/dl·min (373 ± 45 vs. 257 ± 38 mmol/liter·min); LC-alc vs. LC-con, respectively, P = 0.002] and HC-meal [iAUC, 33,038 ± 3,543 vs. 24,092 ± 4,074 mg/dl·min (373 ± 40 vs. 272 ± 46 mmol/liter·min); HC-alc vs. HC-con, respectively, P = 0.008; Fig. 3]. Total and HDL cholesterol levels decreased after both control meals, with no significant difference between meals with and without alcohol (data not shown). NEFA levels decreased after both control meals and were not significantly affected by alcohol ingestion (Fig. 3). Modifying effect of insulin sensitivity on alcohol-induced changes in postprandial glucose, insulin, and lipid levels
When postprandial responses were examined in insulin-sensitive and insulin-resistant subjects separately, only insulin-sensitive subjects had lower glucose iAUC (P = 0.001) and higher insulin iAUC (P = 0.05) after the LC-alc vs. LC-con meal. Differences between the LC-alc and LC-con meals were not significant in insulin-resistant subjects (P > 0.31). AUCINS/AUCGLUC was also increased by alcohol in insulin-sensitive subjects after the LC high fat meal only (P = 0.002), suggesting increased insulin secretion. There were no differences in postprandial glucose or insulin responses between the HC-con and HC-alc meals in either insulin-sensitive or insulin-resistant subjects (P > 0.20).
Alcohol consumption increased postprandial triglyceride iAUC after both meals in both groups; however, the difference between the HC meals with and without alcohol was not statistically significant in insulin-sensitive subjects (P = 0.14). In insulin-sensitive subjects, the ingestion of alcohol tended to enhance the postprandial suppression of NEFA levels after the LC-meal [NEFA iAUC, –1772 ± 635 vs. –979 ± 450 mg/dl·min (–67 ± 24 vs. –37 ± 17 mmol/liter·min); LC-alc vs. LC-con, respectively, P = 0.07] and HC meal [–2566 ± 476 vs. –1587 ± 476 mg/dl·min (–97 ± 18 vs. –60 ± 18 mmol/liter·min); HC-alc vs. HC-con, respectively, P = 0.08], but had no effect in insulin-resistant subjects (P > 0.77). Total cholesterol iAUC was increased by alcohol consumption after both the LC (P = 0.05) and HC (P = 0.04) meals in insulin-sensitive subjects. Alcohol had no effect on HDL responses in either group (P > 0.85).
Effects of alcohol on CRP and adiponectin levels
Circulating CRP levels decreased slightly after both control meals (Fig. 4). Alcohol enhanced postprandial reductions in CRP levels after the HC meal, but had an attenuating effect after the LC meal (Fig. 4).
|
. As previously reported by our group in another cohort (31), adiponectin levels did not change significantly from baseline after the LC-con meal. Although adiponectin levels increased 30 min after consumption of the HC-con meal, values throughout the remainder of the postprandial period were not significantly different from fasting values. Although this early rise in adiponectin was not observed when alcohol was consumed with the HC meal, addition of alcohol to the control meals had no sustained effect on total postprandial incremental adiponectin responses (iAUC: LC-con vs. LC-alc, P = 0.90; HC-con vs. HC-alc, P = 0.26). Effects of alcohol on postprandial BP, heart rate, and AIx
Baseline brachial BP averaged over four visits was 128 ± 15/79 ± 8 mm Hg. Brachial systolic and diastolic BP decreased 30 min after consumption of both control meals (LC-con, –9 ± 2/–13 ± 2 mm Hg; HC-con, –14 ± 3/–8 ± 3 mm Hg) and remained significantly lower than baseline for 180 min (P > 0.28 for difference in BP iAUC between LC-con and HC-con meals). Postprandial changes in BP were not significantly affected by alcohol consumption (iAUC LC-con vs. LC-alc and HC-con vs. HC-alc, P > 0.20).
Heart rate increased postprandially after both control meals (0 min: LC-con, 9 ± 2 beats/min; HC-con, 7 ± 2 beats/min) and remained higher than baseline throughout the entire postprandial period. The postprandial increment in heart rate was similar after the LC-con and HC-con meals (P > 0.55) and was not significantly affected by alcohol consumption (P > 0.20).
Baseline AIx averaged over all four visits was 158 ± 7%. AIx decreased after both control meals, but the reduction was significantly greater after the HC-con meal (iAUC, –3184 ± 781% vs. –6119 ± 1038%·min; LC-con vs. HC-con meal respectively, P = 0.02). Alcohol ingestion enhanced the reduction in AIx after the LC meal (iAUC, –3184 ± 781% vs. –6555 ± 984%·min; LC-con vs. LC-alc meal, respectively, P = 0.007), but had no significant effect after the HC meal (HC-con vs. HC-alc meal, P = 0.65; Fig. 5). Although a difference between the LC-con and LC-alc meals persisted after normalizing results to a heart rate of 75 beats/min, it was not statistically significant (P = 0.14).
|
As expected, energy expenditure increased after both control meals (Fig. 6). Alcohol ingestion significantly increased the postprandial increment in energy expenditure at 30–60 min after both the LC meal (P = 0.02) and the HC meal (P = 0.0009), despite attenuating fat and carbohydrate oxidation (Fig. 6). After the HC meal, fat oxidation rebounded from suppression during alcohol oxidation (30–60 min) to stimulation once alcohol had been completely oxidized (150–180 min; Fig. 6). When the increase in energy expenditure at 30–60 min was expressed as a percentage of baseline, meal-induced thermogenesis was greater when alcohol was added to both the LC meal (26.9 ± 3.4% vs. 17.6 ± 2.6%; P = 0.03) and the HC meal (26.2 ± 2.4% vs. 14.2 ± 2.3%; P = 0.0005).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. The finding that alcohol had differential metabolic effects depending on insulin sensitivity and the amount of carbohydrate in the meal is important and highlights the need to standardize meals and carefully phenotype subjects in studies of postprandial metabolism.
|
The difference in triglyceride responses between HC meals with and without alcohol was not significant in insulin-sensitive subjects. Because insulin suppresses hepatic VLDL secretion postprandially, this suggests that in insulin-sensitive subjects, the tendency for alcohol to increase VLDL levels was restrained by the increased insulin response induced by carbohydrate. In insulin-resistant subjects, the increased insulin response observed after the HC meal was not sufficient to counteract alcohol-induced increases in triglyceride levels. Thus, our results suggest that the postprandial triglyceride response after a meal is dependent not only on the carbohydrate content of the meal (and, hence, the magnitude of the insulin response it elicits), but also on the sensitivity of the liver to insulin-mediated inhibition of triglyceride (VLDL) secretion.
NEFA levels normally decrease after a meal due to inhibition of hormone-sensitive lipase by insulin (42). Although the ingestion of larger amounts of alcohol (35–50 g) has been shown to enhance postprandial reductions in NEFA levels (33, 35, 36), the current and previous studies of healthy subjects (37) and patients with type 2 diabetes (43) report no effect of small doses of alcohol (15–25 g). Interestingly, we found a tendency toward greater suppression of postprandial NEFA levels after alcohol consumption in insulin-sensitive subjects, with no significant effect in insulin-resistant subjects. This evidence suggests that the effect of alcohol on postprandial NEFA levels is dependent on both the amount of alcohol consumed and the subject’s insulin sensitivity.
The effect of alcohol on postprandial glucose and insulin levels was limited to insulin-sensitive subjects after the LC meal. Previous studies investigating the effect of alcohol on postprandial insulinemia in healthy subjects and in subjects with diabetes report variable results (33, 35, 37, 44, 45). Although there is evidence that alcohol potentiates the normal postprandial rise in insulin levels (37), as suggested by the increased AUCINS/AUCGLUC response in our study, this is not a consistent finding (33, 35). It has been suggested that alcohol enhances glucose-stimulated insulin release (26); others argue that ß-cell function is unaffected by alcohol consumption (46). Our findings suggest that the ß-cell may not adequately respond to alcohol in insulin resistance, which may explain the previous contradictory findings in different study populations.
Few studies have examined the effects of fat and carbohydrate on circulating levels of inflammatory markers and cytokines. A recent study reported that consumption of a meal temporarily decreased postprandial inflammatory activity, although CRP was not measured (47). We found that circulating CRP levels decreased slightly after both high fat meals. Interestingly, this reduction was enhanced by alcohol after the HC meal, but was attenuated after the LC meal. Although the acute effect of a small amount of alcohol on postprandial CRP levels has not been previously investigated, two intervention studies recently examined the effect of short-term alcohol consumption on fasting CRP levels. In one, the addition of 30–40 g alcohol to the evening meal decreased CRP levels by 35% after 3 wk compared with those during an alcohol-free diet, although this was only statistically significant in women (48). The lack of a significant effect in men confirms an earlier report (49). Given the long half-life of CRP, it is possible that the short-term changes in circulating levels of CRP in our study reflect dilutional effects or are a consequence of postprandial changes in blood supply.
An interesting observation was that AIx decreased significantly after all meals. Because humans spend most of the day in the postprandial state, this indicates that measurements of AIx should be performed under standardized conditions in relation to meals. However, it is not known whether cardiovascular risk is more accurately predicted by the fasting or postprandial measurement of arterial stiffness. Consistent with our results, AIx has been reported to decrease after a 75-g oral glucose tolerance test (50) and after insulin infusion (51). In our study the reduction in AIx was significantly greater after the HC-con meal, which also elicited a greater postprandial insulin response than the LC-con meal. Our results are therefore consistent with the possibility that insulin reduces arterial stiffness after a meal.
Alcohol enhanced the reduction in AIx after the LC meal, although this was attenuated after accounting for meal-induced changes in heart rate. Interestingly, the reduction in AIx was still evident at the end of the 360-min postprandial period, long after blood alcohol levels had returned to baseline, raising the possibility that metabolites of alcohol (e.g. acetate) may be responsible for this effect. In contrast, alcohol had no effect on the reduction in AIx after the HC meal, possibly due to carbohydrate per se having a significant AIx-lowering effect. Although regular consumers of moderate or large amounts of alcohol are reported to have similar or even higher AIx values than abstainers (52, 53, 54), little is known about the acute effect of a small amount of alcohol on AIx. In the only other study to examine the acute effect of alcohol on AIx, 56 g alcohol consumed in the fasting state without a meal reduced AIx over 90 min (53). The apparently contrasting effects of acute and chronic moderate alcohol consumption on AIx may be explicable, because AIx approached fasting values by the end of our study and therefore would not be expected to be different from baseline many hours later (e.g. the next morning) when AIx is most commonly measured in epidemiological studies.
A small amount of alcohol (equivalent to <10% of the total energy of the meal) stimulated postprandial thermogenesis after 30–60 min. This result is consistent with a recent study using 20–24 g alcohol, which reported that postprandial energy expenditure after an alcohol-rich meal was greater than that after isoenergetically dense meals rich in fat or carbohydrate (37). It is noteworthy that the magnitude of this short-term increase in energy expenditure in our study was disproportionate to the calories consumed as alcohol (110 kcal), because the consumption of an additional 300 kcal as carbohydrate (in the HC-con meal) had no significant effect on energy expenditure at 30–60 min (LC-con vs. HC-con, P = 0.23; Fig. 6). It could be speculated that changes in sympathetic activation contribute to this increase in energy expenditure, because alcohol ingestion increased plasma catecholamines in a study of Japanese men (45), although others report no effect of alcohol on plasma adrenaline and noradrenaline concentrations (55). Our results agree with previous suggestions that alcohol stimulates energy expenditure, although the magnitude and duration of this effect remain controversial (56, 57, 58, 59, 60). However, most investigations have used significantly larger quantities of alcohol than would usually be consumed with a meal, and results cannot be extrapolated to the effect of a small amount of alcohol, similar to that associated with maximal cardiovascular protection in previous studies of women (61).
The relative suppression of fat oxidation after alcohol consumption has been attributed to reduced fatty acid supply due to inhibition of lipolysis by acetate (62). Previous evidence (33) and the lack of a significant overall effect of alcohol on fatty acid levels in our study argue against this mechanism playing a major role. Furthermore, it is unlikely that this temporary inhibition of fat oxidation is likely to have a significant metabolic impact in the face of increased energy expenditure. Our findings suggest the regular consumption of small amounts of alcohol may favor a small energy loss and could partly explain our previous report that chronic moderate alcohol consumers have lower total and central abdominal adiposity than alcohol abstainers (5).
In conclusion, a small amount of alcohol enhanced the postprandial increment in insulin and attenuated the postprandial rise in glucose after the LC meal, particularly in insulin-sensitive women. In addition, alcohol had potentially beneficial effects on arterial stiffness when consumed with the LC meal. However, this amount of alcohol increased postprandial triglyceride levels regardless of the carbohydrate content of the meal, particularly in insulin-resistant subjects. We also observed that alcohol led to an immediate and disproportionate increase in energy expenditure despite inhibition of fat and carbohydrate oxidation. Our findings provide some potential mechanisms for the protective metabolic effects of moderate alcohol intake and suggest that these may be dependent on the subject’s insulin sensitivity and the amount of carbohydrate consumed with the meal. Because our study was limited to postmenopausal women, a group at significant cardiovascular risk, our findings need to be confirmed in younger women and men. Whether the acute effects observed in our study contribute to the lower risk of cardiovascular disease and type 2 diabetes in individuals who regularly consume small amounts of alcohol with meals is speculative. The apparently beneficial effect of alcohol on postprandial energy expenditure and the possibility that this may contribute to lower body fatness in moderate alcohol consumers merit additional investigation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: AIx, Augmentation index; BP, blood pressure; CAF, central abdominal fat; %CAF, percentage of CAF; CRP, C-reactive protein; CV, coefficient of variation; 
, change; HC, high carbohydrate; HDL, high-density lipoprotein; HOMA-IR, homeostasis model assessment insulin resistance; iAUC, incremental area under the curve; LC, low carbohydrate; NEFA, nonesterified fatty acid; RQ, respiratory quotient; VLDL, very low-density lipoprotein.
This work was supported by a Postgraduate Medical Scholarship from the National Health and Medical Research Council (ID 152502; to J.R.G.).
Received July 30, 2004.
Accepted October 25, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
B activation in peripheral blood mononuclear cells of healthy volunteers during postprandial lipemia. Circulation 102:1020–1026This article has been cited by other articles:
 |
|
 |
|
  M. G. Burt, G. Johannsson, A. M. Umpleby, D. J. Chisholm, and K. K. Y. Ho Impact of Growth Hormone and Dehydroepiandrosterone on Protein Metabolism in Glucocorticoid-Treated Patients J. Clin. Endocrinol. Metab., March 1, 2008; 93(3): 688 - 695. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Burt, G. Johannsson, A. M. Umpleby, D. J. Chisholm, and K. K. Y. Ho Impact of Acute and Chronic Low-Dose Glucocorticoids on Protein Metabolism J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3923 - 3929. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Stirban, M. Negrean, B. Stratmann, C. Gotting, J. Salomon, K. Kleesiek, and D. Tschoepe Adiponectin Decreases Postprandially Following a Heat-Processed Meal in Individuals With Type 2 Diabetes: An effect prevented by benfotiamine and cooking method Diabetes Care, October 1, 2007; 30(10): 2514 - 2516. [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. O'Keefe, K. A. Bybee, and C. J. Lavie Alcohol and Cardiovascular Health: The Razor-Sharp Double-Edged Sword J. Am. Coll. Cardiol., September 11, 2007; 50(11): 1009 - 1014. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. G. Burt, J. Gibney, and K. K. Y. Ho Protein metabolism in glucocorticoid excess: study in Cushing's syndrome and the effect of treatment Am J Physiol Endocrinol Metab, May 1, 2007; 292(5): E1426 - E1432. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Djousse and J. M. Gaziano Alcohol Consumption and Risk of Heart Failure in the Physicians' Health Study I Circulation, January 2, 2007; 115(1): 34 - 39. [Abstract] [Full Text] [PDF]
|
|
| ||||
, Blood alcohol was checked before all meals and at the other time points indicated when alcohol was consumed with the meal.
, P < 0.0005 (compared with the meal without alcohol). (Figure continues on next page.)
