|
|
||||||||
Department of Food Science & MicrobiologyHuman Nutrition UnitUniversity of Milan (C.C., M.G., G.T.)
Department of Public HealthHuman Nutrition UnitUniversity of Parma (F.B.), ITALY
Address reprint requests to: Cristina Casiraghi, Department of Food Science & Microbiology-Human Nutrition Unit-University of Milan, Via Celoria N.2, 20133 Milan, ITALY. E-mail: maria.casiraghi{at}unimi.it
| ABSTRACT |
|---|
|
|
|---|
Objective: To investigate the acute postprandial response on plasma glucose, insulin and lipids after consumption of two experimental products made from barley flour enriched with ß-glucan in comparison with similar products made from whole-wheat flour.
Methods: A group of 10 healthy volunteers (5 males, age 25.4 ± 0.5 y, BMI 22.6 ± 0.7 Kg/m2) received at breakfast, in random order and in different days, portions (40g of available carbohydrate) of different cereal products or white bread consumed together with a load of 90000 UI retinol. Products were crackers and cookies made either from barley or whole-wheat flour in a 2 x 2 design, where the two factors were the cereal source of dietary fiber (DF), and the food processing. Barley products supplied 12 g DF, 50% soluble, with 3.5 g of ß-glucan per portion. Whole-wheat products supplied about 14 g of dietary fiber, mainly in the insoluble form, with negligible amount of ß-glucan. Fasting and post-prandial glucose and insulin were evaluated for 180 min after the meals; retinyl-palmitate (RP) and triacylglycerol (TAG) were evaluated hourly over 8 hours. Glycemic (GI) and Insulinemic (II) indexes of products were also assessed, using white bread as reference.
Results: Glucose curves were significantly different between types of food processing (p < 0.01) but not between cereal sources of DF (p = 0.07). On the contrary, the effect of fiber but not of processing was evident when glucose response was expressed as Glycemic Index (effect of DF p < 0.01, effect of processing p = 0.69). Individual GI values were 78, 81, 49 and 34 for whole-wheat crackers (WWCr), whole-wheat cookies (WWc), barley crackers (BCr) and barley cookies (Bc) respectively. Insulin curves were significantly different both between type of processing and fiber source (p < 0.001 for both effects). Again, insulin indices were different between fiber but not between processing (p < 0.5 and p = 0.174 respectively). RP and TAG daily profiles were not significantly different between the factors studied.
Conclusions: Products prepared from barley flour enriched with ß-glucan exhibit favourable responses on glucose metabolism, and particularly on insulinemic responses. In general, cookies responded better to the addition of barley fiber than crackers. Our results highlight the complexity of the effect that barley fiber may exert when added to different food products in reducing postprandial metabolic responses.
Key words: barley ß-glucan, functional foods, glycemic and insulinemic indices, post-prandial lipemia
Abbreviations: ANOVA = analysis of variance Bc = barley cookies BCr = barley crackers BMI = Body Mass Index CM = chylomicra DF = dietary fiber FP = food processing GI = Glycemic Index HSD = honest significant differences II = Insulinemic Index Mw = molar weight OGTT = oral glucose tolerance test RM-ANOVA = repeated measures analysis of variance RP = retinyl-palmitate SEM = standard error of the mean TAG = triacylglycerol VLDL = very low density lipoprotein WBr = white bread reference WWc = whole-wheat cookies WWCr = whole-wheat crackers
| INTRODUCTION |
|---|
|
|
|---|
However, most of the conventional cereal products eaten at breakfast or as a snack have medium/high GI, ranging from 60 to 120 (bread as reference) [4].
Viscous fibers have been shown to reduce post-prandial glycemia and insulinemia [57], and have also been recognized since the 1960s as having lipid-lowering effects. Because barley grain has a high concentration of soluble fiber, and especially of ß-glucan (linear chains of ß-glucosyl residues joined through (1
3) and (1
4) linkages), there is an emerging interest in barley as a functional food ingredient [810]. Indeed this has promoted interest in high- ß-glucan cultivars and in technological processing to concentrate these components in barley flour.
The purpose of this study was to evaluate glucose and insulin responses of two different products obtained from barley flour enriched in ß-glucan (8.5%) in comparison with similar products prepared with whole wheat flour and to calculate their Glycemic and Insulinemic Indices. We used a concentrated ß-glucan fraction from a commercially variety (Hordeum vulgare, Cv. Aliseo) of barley, obtained by air classification technique, to prepare the food products, cookies and crackers. Moreover, if long-term beneficial effects of dietary fiber on lipid metabolism seem established [11], the mechanism whereby dietary fiber affects lipid digestion and absorption is not completely understood. This is particularly important because it is now recognized that elevated post-prandial lipemia is an independent risk factor for coronary heart disease [1213]. Therefore, we investigated the effects of barley ß-glucan on post-prandial lipemia, in order to verify if barley fiber specifically affects post-prandial lipids absorption.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Chemical Analysis of Experimental Products
Products were analysed by ICC standard procedures [15] for moisture (method 110/1), crude protein (N x 6.25; method 105/2) and total fat (method 136). Soluble and insoluble dietary fiber was assessed by the enzymatic-gravimetric procedure [16]; ß-glucan content was determined using the approved enzymatic method 32.23 (Megazyme International, Ireland, Bray, Ireland) [17]. Carbohydrates were evaluated as simple sugars [18], total and resistant starch [19].
Composition of experimental products is reported in Table 1.
|
2 week apart. Test meals were served with 500 mL of unsweetened tea and contained 40 g available carbohydrates, either from the four experimental products or from white bread, the standard food used for GI calculation that was administered three times [20]. The carbohydrate load was lower than the generally accepted 50 g used for GI calculation but this was due to the portion volume, that otherwise would have been too high to allow products to be eaten in about ten minutes. However it has been demonstrated that GI values can be reliably calculated from available carbohydrate loads ranging from 25 to 50 g and over [21]. Portion sizes were 95 g for crackers, 85 g for cookies and 90 g for white bread and provided 1553, 1659, 1592 and 1687 KJ respectively for WWCr, WWc, BCr and Bc. A water-soluble vitamin A preparation containing 90000 UI retinol (Arovit ®, Roche S.p.A., Milan) was administered just before the start of test meals, diluted in 50 ml of tea. The use of Vit A has been shown to be a convenient and pertinent method to label chylomicra (CM) for tracking fat absorption [2223].
Meals were administered between 0800 and 0830 and were eaten within 10 min. Zero time was set at the time of the first morsel of food. Blood samples, obtained through an intravenous catheter inserted into a forearm vein, were taken shortly before and at 15, 30, 45, 60, 90, 120, 180 min and then at hourly intervals for the following 5 h after the test meals. A standard light lunch, the same for each of the test days, composed of white bread (50 g), roast-beef (70 g) and apple (200 g) (27.0 g protein, 2.3 g lipid, 59.2 g carbohydrate; 1481 KJ) was served 4 h after breakfast. Subjects were asked to abstain from cigarette smoking and were allowed free access to water.
Blood samples were drawn into vacutainer tubes (Venoject II, VP-054SAHL, Terumo Europe, Leuven, Belgium) containing Li-Heparine 75 USP, covered with aluminum foil in order to protect the samples from the light. Plasma was then recovered within 30 min by centrifugation (1000 g, for 15 min at room temperature), divided into portions and stored at 20°C for later analysis of glucose, insulin, triacylglycerol (TAG) and retinyl-palmitate (RP). Glucose was assayed using an automatic analyser (YSI Stat 2300, Yellow Springs, OH, USA) and insulin by Microparticle Enzyme Immunoassay [24] using a dedicated analyser (MEIA Insulin, IMX System, Abbott Laboratories). The Glycemic and Insulinemic Indices (II) were calculated from the 120 and 180 min incremental areas under the post-prandial glucose and insulin responses respectively, ignoring any area beneath the baseline [25]. White wheat bread was used as reference food: the mean IAUC of three trials was used to calculate both GI and II. Plasma concentrations of RP were estimated by reverse-phase HPLC after plasma total lipids extraction with Folchs solution, as reported by Lemieux et al [26]. Triacylglycerols were measured by enzymatic colorimetric method, using a clinical analyser (ILab 600 Analyzer, Instrumentation Laboratory, Lexington, MA, USA).
Data Analysis
Results are expressed as mean ± SEM. Data were submitted to two ways Repeated Measures Analysis of Variance (RM-ANOVA), with a 2 x 2 design, were the two factors were the source of dietary fiber, i.e. barley vs wheat, and the food processing (FP), i.e. crackers vs cookies. The effect of each factor, the interaction between each factor and time and the interaction between the two factors was examined. If significant effects of interactions (p < 0.05) were found, the significance of differences between products and/or time were checked by Tukeys Honest Significant Differences post-hoc test. GI and II were analysed by one way RM-ANOVA. The analyses were performed using the StatSoft Statistica Package for Windows (release 4.5, Statsoft Inc., Tulsa, OK, USA).
| RESULTS |
|---|
|
|
|---|
In Fig. 1 are reported the blood glucose responses induced by wheat and barley products, according to the type of product (crackers and cookies). A statistically significant effect was found for FP (p < 0.01), and for the interaction DF x FP (p < 0.05), with cookies having lower glycemic responses than crackers and Bc being significantly lower than all other products (p < 0.01, Tukey post hoc test). Interaction time x DF was also statistically significant (p < 0.05) with Bc showing the lowest post-prandial glycemia, significantly different from that of WWc at 30 and 45 min after breakfast (p < 0.01 and p < 0.05 respectively). On the contrary, although BCr tended to give a lower glycemic profile than WWCr, their difference was not statistically significant.
|
|
|
|
|
| DISCUSSION |
|---|
Barley products elicited more favourable metabolic responses than whole-wheat products when looking at post-prandial glucose and insulin profiles, resulting in significantly (p < 0.05) lower Glycemic indices for cookies and in significant (p < 0.05) reduced Insulinemic indices for both crackers and cookies. The GI values observed for barley cookies (mean GI = 34; range 1065) tend to be lower than the range reported elsewhere for cookies (57106) [4]. The GI values calculated in this study are affected by a high variability, with a mean CV% of 63 (range 5574). Such an high CV value, however, is comparable with other results from the literature [20] and could be partially related to the use of venous blood, often associated with greater within-subjects variation of both glycemic responses and GI values and non-normal distribution of GI values [20]. However, the elevated number of plasma variables simultaneously evaluated in our experimental protocol did not permit the use of capillary blood, which probably would have been resulted in a lower variability of our results. The role of viscous fiber in reducing the rate of absorption is well documented in literature [27]. Barley is high in soluble viscous fiber and the consumption of products such as pasta, bread and porridge enriched with barley ß-glucan has been repeatedly shown in literature to blunt glycemic and insulinemic responses [2831].
The extent of reduction has been investigated by Jenkins [32], who demonstrated that each gram of ß-glucan can be expected to lower the GI by 4 GI units, and that this effect is not attenuated by the food process. The GI reduction induced by barley ß-glucan compared to whole-wheat observed in our study was even more marked, ranging from 8.5 (crackers) to 15.2 (cookies) units of GI reduction per gram of added ß-glucan.
Battilana et al. [33] suggested that the lowered post-prandial glucose concentrations which are observed after a single meal containing ß-glucan are essentially due to a delayed and somewhat reduced carbohydrate absorption from the gut, and do not result from the effects of fermentation in the colon.
That the effect is related to the impairment in transit and absorption due to meal viscosity is suggested by the studies of Wood et al [34], who has shown a highly significant linear inverse relationship between log[viscosity] of ß-glucan mixtures and the glucose and insulin responses in healthy subjects. The relationship showed that 7996% of the changes in plasma glucose and insulin was attributable to viscosity. Recently, the same author [35] reported a significant relationship between the peak of blood glucose and a combination of the logarithm of the concentration and the logarithm of the molar weight (Mw) of purified oat ß-glucan given in a 50 gOGTT at doses from 1.8 to 14.5 g/500 ml. Unfortunately, we did not measure the Mw of ß-glucan used in our products; a difference in Mw, and thus in the viscosity induced in the intestinal lumen, can be advocated to partially explain the different potential in blunting glycemic responses of barley ß-glucan used in our study compared to ß-glucan used by Jenkins.
In the present study, despite a similar composition in lipids, protein and amount of fiber, cookies showed lower glycemic and insulinemic IAUCs, in comparison with crackers, even if this trend reached statistical significance only for whole-wheat insulinemic response (p < 0.05).
Processing may have an effect on glucose and insulin responses. Water available for starch gelatinisation, temperature, baking time and pressure, are all factors that influence the rate of starch digestibility [36] and thus the glycemic impact of a product. A combination of 510% dough moisture in cookies, as opposite to 1020% in crackers, and the presence of simple sugars that further reduce water activity, might have resulted in a lower degree of starch gelatinisation in cookies.
This effect could have been strengthened by the presence of barley ß-glucans, which can trap large amounts of water themselves, thus explaining the significant (p < 0.05) reduction in glucose IAUCs and in GI evaluated for barley cookies in comparison with whole-wheat cookies.
In the present study neither fiber source nor type of processing seemed to affect plasma TAG response to tested products. Similarly, Cara et al. (1992) [37] demonstrated that 10 g of dietary fiber added to a meal containing 70 g of fat had the same effect in reducing post-prandial TAG response, irrespectively of the type of fiber fed (oat bran, rice bran, wheat fiber or wheat germ).
Measurement of plasma triacylglycerols concentrations after meals is a simple way to assess post-prandial lipemia, but this approach does not differentiate triacylglycerol-rich lipoproteins of intestinal origin, i.e. chylomicra and their remnants, from endogenous triacylglycerol-rich lipoproteins (VLDL) derived from the liver. This distinction can be made by administrating vitamin A with the test meals. Vitamin A is incorporated into lipoproteins as retinyl-palmitate (RP) during fat absorption and is thought to remain associated with the same lipoprotein from its secretion by the intestine [38,39], until uptake by the liver without being resecreted as lipoprotein component [40]. Therefore RP constitutes, even though with some limitations [26,41], a useful marker of chylomicra secretion and clearance [2223]. In this study RP post-prandial responses were similar among products suggesting that neither fiber source nor food structure had an effect in the modulation of fat absorption. It cannot be excluded that the effect of fiber in a meal containing higher amounts of fat could have been different, although the type of food product tested is likely to be preferentially consumed alone as a snack or within a light breakfast, without substantial amount of added fats. Effective low GI high-fiber products could play an important role in increasing fiber intakes and reducing the glycemic load of the diet. The availability of functional cereal products naturally enriched with dietary fiber and slightly processed in order to prevent starch gelatinisation, could help the public to increase dietary fiber intake and to maintain protective post-prandial metabolic patterns. However, our results highlight the complexity of the effect that barley fiber may exert when added to different food products in reducing postprandial metabolic responses.
| ACKNOWLEDGMENTS |
|---|
|
|
|---|
Received March 10, 2005. Accepted February 11, 2006.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
D. A. Timm and J. L. Slavin Dietary Fiber and the Relationship to Chronic Diseases American Journal of Lifestyle Medicine, June 1, 2008; 2(3): 233 - 240. [Abstract] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |