Chemical structure

Oat β-glucan is a linear polymer of d-glucose bonded by β-(1 → 4) and β-(1 → 3) glucosidic linkages. About 90 % of the glucose units are arranged in blocks of β-(1 → 3)-linked cellotriosyl and cellotetraosyl units, with the rest being the longer cellulosic segments (Fig. 1). The ratio of trisaccharides–tetrasaccharides in oat β-glucan is typically 2:1, which is distinctively different from that found in barley (3:1) and wheat (4:1) β-glucans⁽ Reference Cui, Wood and Blackwell ^{8 ,} Reference Cui, Wood and Nishinari ^{9 )}. This structural variation leads to differences in some of the physical properties of these β-glucans, which will be discussed in the following sections. A small variation in the trisaccharide–tetrasaccharide ratio has been reported for oat β-glucan, which may arise from several factors including differences between species⁽ Reference Miller, Wood and Pietrzak ^{10 )}, growing conditions, and extraction and analytical methods. A number of reports have suggested the presence of amino-acid residues⁽ Reference Autio, Myllymäki and Suortti ^{11 ,} Reference Vårum and Smidsrød ^{12 )} and inner C-6 carbon-bound phosphomonoesters⁽ Reference Ghotra, Vasanthan and Wettasinghe ^{13 )} in the oat β-glucan molecule. However, these additional structural features have not always been observed.

Fig. 1

Chemical structure of oat β-glucan. G represents a glucose unit and the numbers indicate the linkage sites; the ratio of m/n is approximately 2.

Molecular weight and conformation

The molecular weight (MW) and conformation of oat β-glucan are important because they can largely determine many other physical properties, including the solubility, solution viscosity and viscoelasticity and gelation properties. For instance, high MW oat β-glucan does not form a gel, but gels can be produced from samples with reduced MW under certain conditions⁽ Reference Lazaridou and Biliaderis ^{14 ,} Reference Lazaridou, Biliaderis and Izydorczyk ^{15 )}. All the studies so far have indicated that oat and other cereal β-glucans adopt overall an extended random coil conformation in aqueous solution⁽ Reference Vårum, Martinsen and Smidsrød ^{16 –} Reference Roubroeks, Mastromauro and Andersson ^{18 )}. Oat β-glucan, like all other natural polysaccharides, has a MW distribution that is polydisperse. This means that, instead of one unique value (as in proteins), a wide MW distribution exists in each polymer sample. Therefore, depending on the method used for the measurement, various MW averages are obtained to describe a polymeric material. The weight-average MW and number-average MW (M _w and M _n) are the most frequently encountered values for β-glucans, which are often measured, for example, by static light-scattering and osmotic pressure measurements, respectively⁽ Reference Vårum and Smidsrød ^{12 ,} Reference Vårum, Smidsrod and Brant ^{19 ,} Reference Li, Wang and Cui ^{20 )}. High-performance size exclusion chromatography has been extensively used in recent years for characterisation of β-glucan. When connected with multi-detectors (using reflective index, light scattering and viscometry), the high-performance size exclusion chromatography method can provide a range of molecular parameters in addition to the MW distribution and average MW⁽ Reference Wang, Wood and Cui ^{21 )}. The method of specific dye binding of calcofluor to cereal β-glucans offers a unique way of determining the MW of β-glucan by high-performance size exclusion chromatography. The post-column detection of calcofluor binding to β-glucan makes it possible to measure β-glucan in the presence of other polysaccharides that do not bind this fluorescent stain⁽ Reference Wood ^{22 ,} Reference Rimsten, Stenberg and Andersson ^{23 )}. This is especially useful in making comparisons between different food products and extracts containing β-glucan, because purification of β-glucan is not necessary before the measurement.

Factors that affect the molecular weight

The MW of β-glucan in common oat grain (Avena sativa L.), which is the main commercial species, is influenced by the variety and growing conditions⁽ Reference Andersson and Börjesdotter ^{24 ,} Reference Ajithkumar, Andersson and Åman ^{25 )}. It has not been possible to directly measure the MW of β-glucans in the cell-wall matrix of oat grain. The MW of β-glucan in groats or flour is subject to changes caused by endogenous hydrolytic enzymes, and various processing and storage factors. Because endo-β-glucanases are present in oat seeds⁽ Reference Wood, Siddiqui and Paton ^{26 )}, inactivation of these enzymes (such as by IR heating, steaming or boiling in aqueous ethanol) is considered essential to obtain high MW β-glucan extracts from the oat-grain cell walls⁽ Reference Wood and Webster ^{27 ,} Reference Doehlert, Zhang and Moore ^{28 )}. The reported MW of β-glucan is also therefore subject to significant variations, as a result of differences in the conditions used for extraction and purification of the polysaccharide. The MW obtained for purified oat β-glucans through a well-designed process to minimise molecular depolymerisation is in the range of 2 × 10⁶ to 3 × 10⁶Da. It is known that the MW of β-glucan can be decreased as a result of food preparation such as bread-making⁽ Reference Andersson, Armö and Grangeon ^{29 )}, baking of muffins or cookies⁽ Reference Beer, Wood and Weisz ^{30 ,} Reference Kerckhoffs, Hornstra and Mensink ^{31 )}, and extensive extrusion⁽ Reference Tosh, Brummer and Miller ^{32 )}. Åman et al. ⁽ Reference Åman, Rimsten and Andersson ^{33 )} reported that the MW of β-glucan was preserved in rolled oats, oat bran and different types of oat-bran concentrate, porridge and pancakes, whereas the degradation of β-glucan was observed in bread, pasteurised apple juice, pasta and teacake (see Decker et al. ⁽ Reference Decker, Rose and Stewart ^{34 )} in this Supplement for more examples).

Content and location in oat grain

The β-glucan contents in oat groats vary considerably with cultivars and growing conditions. It has been shown that a 2- to 3-fold variation in the β-glucan content exists among oat cultivars⁽ Reference Miller, Wood and Pietrzak ^{10 ,} Reference Ajithkumar, Andersson and Åman ^{25 ,} Reference Peterson, Wesenberg and Burrup ^{35 )}. A number of surveys from different parts of the world revealed that the β-glucan contents were in the range of 1·8–5·5 % of the total dry weight of the oat groat, and frequently in the range of 4·5–5·5 %⁽ Reference Miller, Wood and Pietrzak ^{10 ,} Reference Saastamoinen, Hietaniemi and Pihlava ^{36 ,} Reference Dvončová, Havrlentová and Hlinková ^{37 )}. Higher β-glucan contents of up to 7 % have also been reported for some oat varieties⁽ Reference Decker, Rose and Stewart ^{34 )}.

Oat β-glucan is primarily located in the cell walls of the starchy endosperm (Fig. 2). The endosperm cell walls of oats⁽ Reference Miller, Fulcher, Webster and Wood ^{38 )} have a layered structure. In this layered structure, there is a relatively thin outer layer, consisting of mainly water-insoluble polysaccharides such as cellulose, glucomannan and arabinoxylan, and a thick inner layer of water-soluble polysaccharide, most of which consist of the mixed linkage β-glucan plus a small amount of soluble arabinoxylan. The cell wall of the subaleurone layer is usually much thicker than that of the inner endosperm⁽ Reference Wood and Fulcher ^{39 )} and is rich in β-glucan. A relatively thin layer of β-glucan is also found in the inner layer of the aleurone cell walls. This inner layer is surrounded by a thick insoluble outer layer, which in turn makes the β-glucan less readily soluble compared with the endosperm β-glucan⁽ Reference Miller, Fulcher, Webster and Wood ^{38 ,} Reference Wood and Fulcher ^{39 )}.

Fig. 2

β-Glucan in the cell wall of oat stained with calcofluor. a, aleurone; s, subaleurone; e, endosperm. (Courtesy of S. Shea Miller, Agriculture and Agri-Food Canada.)

Solubility and its determination

In the literature, the term ‘solubility’ for a polysaccharide does not represent the true thermodynamic equilibrium solubility. Rather, it is used to refer to the property of a polysaccharide in a solid form or contained in a solid food matrix to disperse in a liquid medium (often water) and form a homogeneous dispersion under specified conditions. It is sometimes used interchangeably with the word ‘extractability’ of the polysaccharides. Solubility is usually expressed by the percentage of the dissolved fraction relative to the total amount of the polysaccharide in the original solid matrix under specific conditions (such as temperature).

The solubility of β-glucan in water has been tested recently under different conditions of processing and laboratory measurements⁽ Reference Johansson, Tuomainen and Anttila ^{40 )}. The choice of the method depends on the objective of the investigation. Clearly, it is only meaningful to compare the solubility of different β-glucans, and indeed other polysaccharide preparations, under the same conditions. In order to establish reliable methods for the preparation of β-glucan concentrates with maximum yields and minimal depolymerisation, several studies have been carried out to compare extractability of β-glucan under various conditions. Different extraction schemes have been developed based on these studies for laboratory or commercial production of β-glucan, and the methods have been reviewed previously⁽ Reference Wood ^{41 ,} Reference Stevenson, Webster and Wood ^{42 )}. The extractability of β-glucan was found to be strongly dependent on many factors, such as temperature, pH and the presence of digestive enzymes. When the health benefits of β-glucan as a soluble fibre are of primary interest, the solubility properties of β-glucan under conditions close to the gut environment are highly relevant. Because animal and human trials are expensive and time-consuming, having an in vitro method as a pre-screening tool to evaluate the properties of β-glucan in oat products is very useful.

There are two main types of in vitro methods used to evaluate the solubility of β-glucan in foods. One of these is designed to simulate physiological conditions of digestion in the upper GIT⁽ Reference Beer, Wood and Weisz ^{30 ,} Reference Lebet, Arrigoni and Amado ^{43 )} of human subjects. For a typical example of such a procedure, see the online supplementary materials. This method involves incubating test foods with a series of digestive enzymes at human body temperature (37°C) and appropriate pH levels for defined periods to simulate the acidic and neutral conditions in the stomach and small intestine, respectively. The solubilised β-glucan in the supernatant can then be quantitatively measured by the method of Jørgensen & Aastrup⁽ Reference Jørgensen and Aastrup ^{44 )}, and the total β-glucan content in the same sample can be measured by the method of McCleary & Glennie-Holmes⁽ Reference McCleary and Glennie-Holmes ^{45 )}. A similar procedure has been developed by directly measuring the viscosity developed during the simulated digestion process. This allows the contribution of solubility and MW of β-glucan to the viscosity to be accounted for simultaneously⁽ Reference Gamel, Abdel-Aal and Wood ^{46 )}. Another procedure, which adapts essentially the method of Asp et al. ⁽ Reference Asp, Johansson and Hallmer ^{47 )}, was developed initially for the isolation of water-soluble NSP⁽ Reference Johansson, Tuomainen and Anttila ^{40 )}. This method involves a de-starching process during which the samples are treated at boiling temperature to allow starch gelatinisation. The starch and protein are then hydrolysed by various enzymes, and β-glucan is precipitated from the supernatants by alcohols. A similar approach to this is the use of hot-water extraction at 100°C with only a heat-stable α-amylase and 0·28 mg/ml CaCl₂ added⁽ Reference Rimsten, Stenberg and Andersson ^{23 ,} Reference Ajithkumar, Andersson and Åman ^{25 ,} Reference Åman, Rimsten and Andersson ^{33 )}. Few direct comparisons have been made among the results from these methods. Because it is known that increasing extraction temperature increases the solubility of the polymer⁽ Reference Johansson, Tuomainen and Anttila ^{40 )}, the amount of solubilised β-glucan is likely to be lower using in vitro physiological extraction than that obtained using the hot-water extraction method⁽ Reference Beer, Wood and Weisz ^{30 )}. Further work is required, therefore, to investigate the precise relationship between in vitro measurements of β-glucan solubility and polysaccharide dissolution in vivo.

Factors that influence solubility

Sources

As mentioned earlier, most β-glucan in the oat groat is located in the inner layer of the cell wall that is enclosed by an insoluble cellulosic and hemicellulosic outer layer. It is generally less water soluble than extracted β-glucan material. Because the cell walls in the inner endosperm are thinner than in the aleurone and subaleurone layers, it appears that β-glucan in the milled whole groat is more easily extracted than that in oat bran⁽ Reference Beer, Wood and Weisz ^{30 ,} Reference Miller, Fulcher, Webster and Wood ^{38 )}. For example, Beer et al. ⁽ Reference Beer, Wood and Weisz ^{30 )} has found that the proportion of soluble β-glucan extracted from oat brans and rolled oats, by hot-water extraction in conjunction with a heat-stable α-amylase treatment, was 30–65 and 70 %, respectively (Table 1). Under physiological conditions, the percentage of extractable β-glucan was also noticeably higher in rolled oats than in bran samples.

Table 1

The β-glucan content and the percentage and molecular weight of extracted β-glucan from oat bran and rolled oats using hot water (90°C for 2 h with Termamyl) or physiological in vitro digestion method⁽ Reference Beer, Wood and Weisz ^{30 )}

ND, not determined.

* Samples cooked before extraction.

Processing and food preparation

The solubility of β-glucan from oat bran varies largely because of different processing conditions in addition to varietal differences⁽ Reference Mälkki, Autio and Hanninen ^{48 )}. Using an in vitro digestion model to simulate human digestion, Beer et al. ⁽ Reference Beer, Wood and Weisz ^{30 )} found that only 12–33 % of total β-glucan in bran and rolled oats was solubilised, whereas Tosh et al. ⁽ Reference Tosh, Brummer and Miller ^{32 )} reported a slightly higher solubility value of 39 % in a commercial oat bran. When the oat bran was incorporated into an extruded cereal, the solubility of β-glucan increased markedly from 39 to 67 % in breakfast cereals prepared under mild extrusion conditions. Further increases in the extrusion temperature and mechanical disruption resulted in a complete solubilisation of β-glucan, although the process was accompanied by a >10-fold decrease in MW. Microscopy observations clearly showed that extrusion disrupted the cell-wall structure from which β-glucan labelled by calcofluor was released and dispersed into the food matrix (Fig. 3). Zhang et al. ⁽ Reference Zhang, Bai and Zhang ^{49 )} also reported that extrusion improved the solubility of β-glucan in oat bran. Johansson et al. ⁽ Reference Johansson, Tuomainen and Anttila ^{40 )} have shown that the baking of wheat bread with oat bran and subsequent drying of the bread decrease the amount of β-glucan solubilised, whereas the cooking of oat flakes as porridge increases it. Also, fermentation of oat bran by rye sourdough was shown to increase the solubility of oat β-glucan⁽ Reference Johansson, Tuomainen and Anttila ^{40 ,} Reference Degutyte-Fomins, Sontag-Strohm and Salovaara ^{50 )}.

Fig. 3

Cryosections of test cereals, stained with calcofluor and mounted in acid fuchsin, showing that extrusion disrupted the cell wall structure, during which β-glucan (blue stain) was released and dispersed into the food matrix. (a) Premix (control); (b) after moderate extrusion and (c) after extensive extrusion. AL, aleurone cells; SAL, subaleurone cell walls. (Adapted from Tosh et al. ⁽ Reference Tosh, Brummer and Miller ^{32 )}, with permission).

The presence of endogenous enzymes, mainly β-glucanase, can have both detrimental and beneficial effects, the former being a depolymerisation of β-glucan and the latter being an increase in polymer solubility⁽ Reference Zhang, Doehlert and Moore ^{51 )}. In a study by Tosh et al. ⁽ Reference Tosh, Brummer and Wolever ^{52 )}, β-glucanase at different concentrations was added to the wheat dough containing oat bran to produce muffins containing different MW β-glucans. A decrease in MW of β-glucan from 2 200 000 to 400 000 Da led to an increase in the solubility of β-glucan in oat-bran muffins from 44 to 57 %. This phenomenon is in agreement with dissolution studies showing that the hydrolysis of a similar water-soluble polysaccharide, such as guar galactomannan, significantly increased solubilisation of the polymer⁽ Reference Wang, Ellis and Ross-Murphy ^{53 )}. However, when the MW was reduced further to 120 000 Da, the solubility of β-glucan decreased to as low as 26 %. The authors suggested that the subsequent diminution of β-glucan solubility at very low MW (120 000 Da) was due to a stronger self-association of the depolymerised β-glucan, thus leading to the formation of insoluble aggregates. It is also worth speculating that the interaction of β-glucan with other food components, especially proteins via the Maillard reaction, might have contributed to the low solubility. Lower MW polysaccharides seem to possess an increased reactivity towards proteins⁽ Reference Wong, Day and Augustin ^{54 )} and other components found in foods⁽ Reference Lazaridou, Biliaderis and Izydorczyk ^{15 ,} Reference Tosh, Wood and Wang ^{55 )}.

Viscoelastic properties of β-glucan solution

A freshly prepared and fully hydrated oat β-glucan solution behaves as a typical neutral random-coil polysaccharide solution. The viscosity of a polysaccharide solution depends on the structure, MW and polymer concentration, as well as conditions of measurement, such as shear rate and temperature. The shear rate dependence of viscosity for oat β-glucan has been extensively studied⁽ Reference Doublier and Wood ^{2 ,} Reference Ren, Ellis and Ross-Murphy ^{60 )}. In terms of conditions in the GIT, shear rate is related to the degree of mixing of fluid (digesta) caused by peristalsis, although flow patterns in the gut are obviously much more complex. Fig. 4 shows the steady shear-flow profiles for solutions of a high MW (2 × 10⁶Da) oat β-glucan at different concentrations⁽ Reference Ren, Ellis and Ross-Murphy ^{60 )}. At low β-glucan concentrations (approximately lower than the overlap concentration c*), the solution viscosity does not change with respect to shear rate, a form of behaviour that is typical of a Newtonian fluid. At high concentrations, however, shear-thinning behaviour is observed, meaning that the viscosity decreases with an increasing shear rate. Because of the shear-rate dependence of viscosity, the solution viscosities of different samples have to be compared at the same shear rate, most usefully the value at ‘zero’ shear rate is preferred. In heterogeneous systems containing a polysaccharide gum and insoluble particulates, which are common in food suspensions or ingested food mixed with digestive fluids (i.e. digesta) in the gut, more complex rheological behaviour is seen. For instance, if there is a high fractional volume of insoluble particulates present, the suspensions exhibit shear-thinning even at low shear rates (i.e. loss of the Newtonian plateau) and also substantial increases in viscosity are observed⁽ Reference Rayment, Ross-Murphy and Ellis ^{61 ,} Reference Rayment, Ross-Murphy and Ellis ^{62 )}. For calculation of the zero-shear viscosity from the flow profiles, see online supplementary material. As seen in the literature, most researchers, in an effort to simplify the method and obtain reproducible data, remove the particulate material by centrifugation before taking viscosity readings, because the presence of insoluble material can seriously interfere with measurements. Moreover, a single measurement of viscosity is often performed at an arbitrary shear rate (see Table 2). This is because the shear rates that occur in the intestinal lumen of human subjects and experimental animals following a meal containing a polysaccharide gum are not known and may vary at different regions of the gut and at different postprandial times. Also, zero shear viscosity is not always readily obtainable due to instrumental limitations. Caution is therefore needed when comparing viscosity data from different laboratories.

Fig. 4

Apparent viscosity (η) v. shear rate ( $$ \gamma \bullet $$ ) for different concentrations of oat β-glucan solutions. (Re-printed, with permission, from Ren et al. ⁽ Reference Ren, Ellis and Ross-Murphy ^{60 )}).

Table 2

Examples of viscosity (η) measured at specific shear rates ( $$ \gamma \bullet $$ ) and temperature (temp.) of oat-containing meals tested in human studies, which have demonstrated positive postprandial blood-glucose lowering (PGL) and cholesterol-lowering (CL) effects

M _p, peak molecular weight; M _w, weight average molecular weight; values with *indicate M_w only; ND, not determined.

Solution viscosity is affected by the polymer concentration (c) and MW in a power-law relationship: η~(c[η]) ⁿ . The intrinsic viscosity, [η], is a measure of hydrodynamic volume of the individual molecule in solution and is related to MW by, for instance, the Mark–Houwink relationship ([η] = k MW^α, where k and α are constants). Therefore, the parameter c[η] represents the space occupancy of the polymer molecules in the solution by taking into account both concentration and MW factors. Fig. 5 is a double logarithmic plot of η_sp,0 (the specific viscosity at zero shear rate) v. the parameter cη for oat β-glucan solutions. This plot consists of two linear regions connected by a continuous transition region starting at c[η] approximately 1. The concentration at this point is designated c*, which corresponds to the concentration at which individual β-glucan molecules begin to overlap. For example, for two β-glucans of MW 1 million Da and 0·1 million Da, the c* values are approximately 0·1 and 0·5 %, respectively. At c< c*, the solution is classified as being in a dilute regime and the relationship η_sp,0~c ^1·1 holds; whereas at c>c*, the solution is classified as being in a semi-dilute regime and η_sp,0~c ^4·4. It is well known that the viscosity of β-glucan solution changes more markedly with concentration in the semi-dilute solution than it does in dilute solution. The zero-shear viscosity at the critical concentration is approximately 1 mPa·s (Fig. 5). In most clinical trials that showed positive blood-glucose and cholesterol-lowering effects, the viscosity (as estimated in vitro in the shear rate range 10–30 s^− 1) is considerably higher than this value (Table 2), although the in vivo viscosity may be different from that measured in vitro ⁽ Reference Wood, Braaten and Scott ^{63 )}. According to Wood⁽ Reference Wood ^{41 )}, the (in vitro viscosity) dose–response for the glycaemic response of β-glucan is in the range of 20–2000 mPa·s (at shear rate 30 s^− 1). When the viscosity is lower than 20 mPa·s, no glucose-lowering effect is observed, whereas when the viscosity is higher than 2000 mPa·s, no further reductions in blood glucose are seen. In other words, the metabolic response appears to reach a plateau above 2000 mPa·s. This highlights the importance that changes in MW have on the viscosity of β-glucan products, which in turn determines their physiological activity.

Fig. 5

Log η_sp,0 (the specific viscosity at zero shear rate) v. log c[η] for oat β-glucan. ●: data of Doublier & Wood⁽ Reference Doublier and Wood ^{2 )}; Δ: data from Ren et al. ⁽ Reference Ren, Ellis and Ross-Murphy ^{60 )}. (Reprinted, with permission, from Ren et al. ⁽ Reference Ren, Ellis and Ross-Murphy ^{60 )}).

It must be emphasised that in vitro viscosities of polysaccharides might not accurately reflect the viscosities of these polymers when present in the GIT. The level of viscosity required to achieve a specific physiological effect has not been carefully reviewed and systematically studied. There are no data available on the direct measurements of luminal viscosity of oatmeals. However, studies by Marciani et al. ⁽ Reference Marciani, Manoj and Hills ^{64 ,} Reference Marciani, Gowland and Spiller ^{65 )} on locust bean gum, a source of water-soluble galactomannan with rheological properties similar to high MW oat β-glucan, provides some useful comparative information. In these studies, the gastric-content samples were obtained after ingestion of meals containing different doses of locust bean gum using nasogastric intubation and aspiration, and the viscosity was measured in vitro by a viscometer. This group also used an echo-planar MRI technique to measure the in vivo gastric lumen viscosity, which allowed viscosity to be measured non-invasively⁽ Reference Marciani, Manoj and Hills ^{64 )}. A good correlation was obtained between the rheological data from in vitro viscometry and echo-planar MRI. Table 2 lists the viscosity of test meals that have demonstrated positive physiological functions. Taking these data into consideration, along with the recommended minimum dose of oat β-glucan used for achieving significant physiological effects, the viscosity of the luminal digesta seems to fall within the lower end range of a semi-dilute regime (see Fig. 5).

From the analysis of data from Lan-Pidhainy et al. ⁽ Reference Lan-Pidhainy, Brummer and Tosh ^{56 )}, it was found that the viscosity was related to concentration exponentially by an exponent of 1·51, although the soluble β-glucan concentration was well within the range of the dilute regime. This is due, at least in part, to the existence of particulate materials, which are known to raise the viscosity of a solution/suspension as explained by Rayment et al. ⁽ Reference Rayment, Ross-Murphy and Ellis ^{61 )}. It is well documented that cereal β-glucan tends to form macromolecular aggregates in aqueous solutions⁽ Reference Li, Wang and Cui ^{20 ,} Reference Wang, Wood and Cui ^{21 ,} Reference Tosh, Wood and Wang ^{66 )}. Moreover, the low-MW oat β-glucan aggregates more easily than their high-MW counterparts⁽ Reference Cui, Wood and Blackwell ^{8 ,} Reference Lazaridou and Biliaderis ^{14 ,} Reference Lazaridou, Biliaderis and Izydorczyk ^{15 )}. Low-MW β-glucan solutions have been shown to have a ‘yield stress flow’ behaviour at very low shear rates giving higher than expected apparent viscosity values, which is attributed to the presence of β-glucan aggregates (Fig. 4)⁽ Reference Doublier and Wood ^{2 ,} Reference Ren, Ellis and Ross-Murphy ^{60 )}.

Conclusions and future research

Based on the results of physiological and human studies to date, there is strong evidence to show that oat β-glucan lowers total and LDL-cholesterol and attenuates postprandial glycaemia and insulinaemia when it is ingested in sufficiently high doses and at a suitably high MW. The physico-chemical properties of oat β-glucan, namely MW and solubility, should be considered when assessing the physiological properties. Oat β-glucan concentrates or isolates have shown similar beneficial effects to oat bran and rolled oats, provided that detrimental processing conditions are avoided.

The current recommended intake of ≥ 3 g/d β-glucan by US Food and Drug Administration does not ensure a significant decrease in blood cholesterol, because the quantity of β-glucan ingested does not solely determine the physiological efficacy. It is the amount and MW of β-glucan that become solubilised in the GIT that determines, to a large extent, the blood-glucose and cholesterol-lowering properties. Clear inverse linear relationships have been demonstrated between the in vitro log (viscosity) of drink or extracts from solid foods and postprandial blood-glucose or fasting blood-cholesterol concentrations. It appears that there is a range of viscosity values over which β-glucan exhibits such biological activities, although the minimum and maximum values for this range need to be defined. There is little doubt that in vitro evaluation can be effectively used to screen oat products in order to increase the likelihood of success in achieving beneficial biological effects in human subjects. How the viscosity of drinks or extracts measured in vitro corresponds to the in vivo viscosity developed in the GIT is not straightforward. A combination of factors, including the dilution effects of digestive fluids, enzymic degradation of β-glucan and interaction of this polymer with other food components, is likely to influence the in vivo viscosity. Direct measurements of viscosity in the human gut have not yet been carried out on ingested oat products, as has been done with other polysaccharides such as the leguminous galactomannans (locust bean gum)⁽ Reference Marciani, Manoj and Hills ^{64 ,} Reference Marciani, Gowland and Spiller ^{65 )}. Accordingly, the minimum in vivo viscosity that is required to achieve desirable metabolic effects is currently unknown.

The development of inexpensive and palatable oat-containing products, which are still physiologically effective, is needed in the future. To facilitate this, it is necessary to establish a standardised protocol to characterise β-glucan products, which provides sufficient physico-chemical data to allow useful comparisons and interpretations to be made of the results obtained in various human studies.