Carbohydrates represent the most important dietary energy source for humans quantitatively, and the main determinant of postprandial blood glucose response( Reference Lafiandra, Riccardi and Shewry 1 ), as the ingestion of carbohydrates is accompanied by an increase in blood glucose levels. On the basis of their effect on postprandial glycaemia, carbohydrate-rich foods are classified into three categories of glycaemic index (GI) (low GI ≤55, medium GI 55–69 or high GI ≥70)( 2 ). GI indicates the extent to which a carbohydrate containing food affects postprandial blood glucose levels compared with a reference food (glucose or white bread) which contains the same amount of available carbohydrate( Reference Jenkins, Wolever and Taylor 3 ). In the same context, insulinaemic index (II) is a marker that indicates the elevation of blood insulin levels in a 2-h postprandial period. The ingestion of high GI foods cause a high increase in postprandial blood glucose concentrations, whereas the increase is less pronounced for foods with a low GI. A high postprandial blood glucose response, in turn, triggers a disproportionately high insulin response which is associated with the development of hyperinsulinaemia and insulin resistance( Reference Jenkins, Kendall and Augustin 4 ).
Among carbohydrate rich-foods, bread is a staple food consumed in all parts of the world in different forms originating from variations in the ingredients traditionally used and processing techniques applied( Reference Lau, Soong and Zhou 5 ). Even in the same country, bread baking is in continuous development and various methods of baking and different kinds of bread are available. As a consequence, bread’s GI values can vary a lot and bread can fall in the category of either low, medium or high GI( Reference Fardet, Leenhardt and Lioger 6 ). Lowering the GI of bread is of scientific interest especially for populations who consume large amounts of bread, such as the Nordic European countries( Reference Goletzke, Atkinson and Ek 7 ). Attenuation of GI can be achieved by different approaches, including the addition of intact structures not accessible to human amylases, viscous and non-viscous fibre( Reference Scazzina, Siebenhandl-Ehn and Pellegrini 8 ), legume flours, fruit-based ingredients and enrichment with specific micronutrients( Reference Stamataki, Yanni and Karathanos 9 ).
The aim of this review was to examine another significant determinant of glucose response to bread, namely bread making procedures. In particular, the influence of leavening techniques and baking method of bread on postprandial glucose and insulin responses are discussed in this article. Insulin response was also examined in order to get a wider picture of the postprandial metabolic effects caused by processing conditions. Pubmed, Scopus, Sciencedirect and Springer databases were searched for interventional studies regarding GI and/or acute postprandial glucose response to bread using relevant indexing terms. Cited references of identified articles were also revised for appropriate missed studies. To include a study in this review, human subjects should have acutely consumed different types of breads, differing in leavening techniques, or other processing parameters, compared with a reference food which was either white wheat bread or glucose, after overnight fast. There were no restrictions in terms of participants’ characteristics. Dates of publication were limited to 2000 onward, in an attempt to present updated literature, and only human studies in English language papers were eligible. Identified studies were classified in two categories according to whether breads differed in leavening technique or other processing conditions, and are summarised in Tables 1 and 3, respectively.
Human studies examining the effect of sourdough fermentation on postprandial glucose (G) and insulin response to bread (Mean values with their standard errors; mean values and standard deviations)
iAUC, incremental AUC; GI, glycaemic index; II, insulinaemic index; T2D, type 2 diabetes; AC, available carbohydrates; WB, white bread; IR, insulin resistant; WWF, white wheat flour.
* Statistical significance, P<0·05.
† Subjects are healthy individuals, if not, further details are given.
‡ Unless otherwise stated, total dietary fibre content corresponds to g/100 g of bread.
§ More information about how AUC, iAUC, iAUCcut and net iAUC are calculated can be found in the article by Brouns et al. (44). GI and II values are expressed with G as the reference, being assigned the value of 100.
││ Time point −15 was used as the baseline, and values below the baseline were considered to be negative peaks. Data were log transformed before statistical analysis.
Leavening techniques
Sourdough fermentation
Sourdough has traditionally been used as leavening agent in bread making, whereas nowadays only few bakeries work with sourdough at an industrial scale. Its use in bread making has a great impact on bread characteristics, including texture, flavour, shelf life and nutritional quality( Reference Arendt, Ryan and Dal Bello 10 ). This fermentation technique has primarily been used for wheat and rye baking. Wheat and rye sourdoughs do not exhibit characteristic differences in fermentation microbiota or their metabolic activity( Reference De Vuyst and Neysens 11 ). Table 1 summarises human trials investigating the effect of sourdough fermentation on postprandial glucose and insulin response to bread, and Table 2 presents selected properties of sourdough breads.
Selected sourdough details of sourdough breads (Mean values and standard deviations)
PBF, partially baked frozen.*Sourdough addition levels correspond to g of sourdough per 100 g of bread dough.
† ml NaOH 0·1 m used for titration until pH 8·5.
‡ ml NaOH 0·1 m used to titrate 10 g of bread.
§ Milliequivalents of H+/kg.
Sourdough fermentation has been proved an effective strategy for lowering the metabolic response to bread due to the production of organic acids( Reference Fardet, Leenhardt and Lioger 6 ), whereas the exact mechanisms have not been fully elucidated. Breads enriched with organic acids, either physiologically produced during fermentation or even artificially added, have been shown to improve postprandial glucose and insulin responses in healthy subjects( Reference Liljeberg, Lönner and Björck 12 ). A large number of studies have confirmed the above results not only in healthy individuals( Reference De Angelis, Damiano and Rizzello 13 – Reference De Angelis, Rizzello and Alfonsi 15 ) but also in subjects with impaired glucose tolerance( Reference Lappi, Selinheimo and Schwab 16 – Reference Maioli, Pes and Sanna 17 ). Rye sourdough fermented breads have exhibited lower acute postprandial insulin responses( Reference Rosén, Silva and Andersson 18 – Reference Juntunen, Laaksonen and Autio 19 ) and occasionally improved glycaemic profiles( Reference Rosén, Silva and Andersson 18 , Reference Bondia-Pons, Nordlund and Mattila 20 ).
A sourdough (Lactobacillus plantarum P1 and Lactobacillus brevis P2) wholemeal bread enriched with oat fibre (3·9 g β-glucan/50 g of available carbohydrates) was found to have low GI (53·7) compared with glucose( Reference De Angelis, Rizzello and Alfonsi 15 ), and a wheat sourdough bread enriched with oat and rye fibre (total dietary fibre content 8·13 g/100 g bread) was found to have GI 41 (v. GI 100 of glucose) in another study( Reference De Angelis, Damiano and Rizzello 13 ). The clear effect of sourdough fermentation on the GI was revealed when same breads, which differed only in the leavening technique, were compared( Reference Scazzina, Del Rio and Pellegrini 21 ). Postprandial glucose response was reduced in both sourdough white bread and sourdough wholemeal bread in comparison to breads made with Saccharomyces cerevisiae.
Juntunen et al.( Reference Juntunen, Laaksonen and Autio 19 ) studied the acute effect of three rye breads prepared with sourdough fermentation on glucose and insulin responses in comparison with refined wheat bread. The test breads were endosperm rye bread (100 % commercial endosperm rye flour, sourdough), traditional rye bread (100 % wholemeal rye flour, sourdough) and high-fibre rye bread (60 % wholemeal rye flour, 40 % rye bran and sourdough). Sourdough was prepared from either commercial endosperm flour for the endosperm rye bread or wholemeal rye flour for the other two breads, along with L. brevis and L. plantarum, fresh yeast and water. Postprandial AUC for glucose did not significantly differ between the rye breads and refined wheat bread, whereas postprandial insulin AUC was significantly decreased after the ingestion of the endosperm rye bread and the traditional one. The present study showed that in healthy subjects less insulin is needed for the control of glucose levels after the consumption of sourdough rye breads compared with refined wheat bread. However, the lower insulin response cannot be explained by the fibre content, as an insignificant decrease in insulin response caused by the high fibre rye bread (29 g total dietary fibre/50 g of available carbohydrates). In addition, sourdough fermented endosperm rye bread was also shown to induce lower postprandial insulin response compared with white wheat bread without improving blood glucose profile in healthy subjects in the study by Bondia-Pons et al.( Reference Bondia-Pons, Nordlund and Mattila 20 ). In another study, a rye wheat sourdough bread was found to have medium GI (GI=62) and caused lower postprandial insulin response compared with a soft pretzel( Reference Goletzke, Atkinson and Ek 7 ).
The postprandial metabolic response of sourdough fermented wheat bread was tested in overweight and obese subjects, who represent a target group with increased risk for developing type 2 diabetes (T2D)( Reference Najjar, Parsons and Duncan 14 ). Glucose AUC (180 min) for sourdough bread (100 % white wheat flour, 37 % sourdough starter from white flour) was significantly lower than for whole wheat bread (71 % whole wheat flour, 29 % whole grain flour), whereas insulin response did not differ between the test breads. The more modest glucose profile without differences in insulin response, may indicate that sourdough increased insulin sensitivity in these individuals. In another study by Lappi et al.( Reference Lappi, Selinheimo and Schwab 16 ) the ingestion of sourdough fermented wholemeal wheat bread (100 % wheat flour from peeled grains) by subjects with the metabolic syndrome resulted in retarded postprandial glucose response, and significantly lower maximum insulin increase compared with white bread. AUC for insulin did not differ between the test breads, but a trend for lower AUC after the ingestion of the wholemeal wheat bread produced with sourdough than that of white bread was observed.
The effect of acute postprandial glucose and insulin response to breads varying in carbohydrate quality due to sprouted grains and sourdough fermentation in adult patients with T2D was examined in a study, which showed that breads made of sprouted grains and produced with sourdough fermentation exhibit lower blood glucose incremental AUC (iAUC), compared with whole grain or white wheat sourdough fermented breads( Reference Tucker, Vandermey and Robinson 22 ). The particular type of bread also caused significantly lower insulin response but only in the second meal. The sprouting treatment of cereal grains causes changes in the grain structure and composition. It is reported to decrease starch content and increase the content and availability of vitamins, minerals, and antioxidants. Acute consumption of sprouted-grain sourdough bread had previously shown to improve glycaemia by lowering glucose response in overweight or obese males( Reference Mofidi, Ferraro and Stewart 23 ). In this randomised single-blind cross-over design trial, overweight or obese subjects consumed the following bread types, eleven-grain bread (whole grain with sourdough culture), sprouted-grain bread (whole grain with sourdough culture), sourdough white bread, twelve-grain bread (whole grain) and white wheat bread. The main finding was that sprouted grain bread lowered glucose responses, which could be attributed to the increased dietary fibre content and increased availability of mg, vitamin E, phenolic compounds and phytoestrogens possibly able to act synergistically to lower glycaemia.
Suggested mechanisms of sourdough action
In an attempt to explore the mechanisms underlying the lower metabolic responses to sourdough bread consumption, gastric emptying rate has been examined by several studies( Reference Liljeberg and Björck 24 – Reference Darwiche, Ostman and Liljeberg 25 ). Darwiche et al.( Reference Darwiche, Ostman and Liljeberg 25 ) using ultrasonography showed a decrease in the gastric emptying rate of barley bread containing sodium propionate, which caused lower glucose and insulin levels compared with plain barley bread. Similar results were obtained by Liljeberg & Bjorck( Reference Liljeberg, Åkerberg and Björck 26 ) when bread was enriched with sodium propionate but not with lactic acid. However, Najjar et al.( Reference Najjar, Parsons and Duncan 14 ), who examined gastric emptying rate by adding paracetamol in the flour of the breads, did not detect any differences between white bread and sourdough white bread. In agreement to the previous study, no difference in gastric emptying rate was reported in the study by Bondia-Pons et al.( Reference Bondia-Pons, Nordlund and Mattila 20 ) among endosperm sourdough rye bread and white bread. The involved mechanisms remain to be elucidated.
Sourdough fermentation is hypothesised to increase the resistant starch (RS) content of breads. Breads made of white wheat flour and wholemeal flour and fermented with selected sourdough L. plantarum and L. brevis strains contain higher concentrations of RS (approximately 5 %) than breads started with baker’s yeast alone (RS: 1·4–1·7 %)( Reference De Angelis, Rizzello and Alfonsi 15 ). In accordance, breads prepared with white or wholemeal wheat flour and leavened with sourdough were found to have higher amounts of RS compared with same breads leavened with S. cerevisiae. Liljeberg et al.( Reference Liljeberg, Åkerberg and Björck 26 ) observed an increase of RS content in breads containing increasing concentrations of lactic acid, leading to the hypothesis that the presence of organic acids in bread may increase starch retrogradation and thus RS content.
It has been also reported that sourdough fermented breads may result in lower rate of starch hydrolysis( Reference Ostman, Nilsson and Liljeberg-Elmstahl 27 ). De Angelis et al.( Reference De Angelis, Rizzello and Alfonsi 15 ) reported that the rate of starch hydrolysis of breads containing lactic and acetic acids was lower compared with non-acidified breads in vitro. Furthermore, comparing microbial acidification by selected sourdough lactobacilli with chemical acidification in terms of rate of starch hydrolysis, breads fermented with sourdough lactobacilli cause a significantly greater effect. Opposite results are demonstrated by Scazzina et al.( Reference Scazzina, Del Rio and Pellegrini 21 ), who supported that the leavening technique does not influence starch digestibility or availability to hydrolytic enzymes. In this study no difference in rate of starch digestion in vitro was found between white or wholemeal breads fermented either by sourdough microflora or by yeast.
Another approach suggests that proteolysis, which takes place during sourdough fermentation( Reference Gänzle, Loponen and Gobbetti 28 ), may contribute to lowered postprandial glucose responses. Lappi et al.( Reference Lappi, Selinheimo and Schwab 16 ) analysed the state of protein in an attempt to explain the mechanisms underlying the beneficial postprandial responses to sourdough fermented bread. Although in vitro analysis revealed no difference in the hydrolysis rate of protein among the sourdough and other wholemeal wheat breads, the content of soluble proteins was higher and their molecular weights was lower in sourdough fermented breads, suggesting that more solubilized proteins are ingested in the case of sourdough fermented wholemeal wheat bread. Although a higher insulin response could be expected after the sourdough bread consumption due to stimulation of insulin secretion from the protein degradation by sourdough fermentation, a trend for lower insulin response was observed.
It is possible, that the sourdough fermentation increases the interaction between starch and gluten proteins of cereals, resulting in the creation of a barrier which limits starch bioavailability and enzyme accessibility. The effect is positive if the acid is present during the gelatinisation of starch but not after the thermal treatment( Reference Ostman, Nilsson and Liljeberg-Elmstahl 27 ). Other effects of sourdough leavening appear to be the synthesis of free phenolic compounds which are thought to improve glucose tolerance( Reference Gobbetti, Rizzello and Di Cagno 29 – Reference Katina, Liukkonen and Kaukovirta-Norja 31 ).
Sourdough fermentation in combination with partial baking-freezing technology
Sourdough fermentation technology was recently combined with partially baking freezing technology for the production of low GI breads and gluten-free breads( Reference Novotni, Cukelj and Smerdel 32 – Reference Novotni, Curić and Bituh 33 ). Partial-baking freezing technology involves initial baking equivalent to 75 % of total baking time, followed by freezing, storage and final baking, and makes freshly baked bread available at any time. Novotni et al. ( Reference Novotni, Curić and Bituh 33 ) used four different lactic acid bacteria starters to ferment sourdough for the production of partially baked frozen wholemeal wheat bread. Five partially baked frozen breads were prepared and tested for postprandial glucose response by ten healthy volunteers in a cross-over design study (glucose solution was used as reference). Breads included a control wholemeal bread without sourdough, a bread prepared with L. plantarum sourdough (10 %), another one with PL1 (Lactobacillus fermentum) sourdough (10 %), a bread with PL3 (L. fermentum with phytase) sourdough (10 %) and finally a bread prepared with LV4 (L. brevis combined with S. cerevisiae var. chevalieri) sourdough (10 %). Their GI values were calculated 70, 60, 56, 56 and 50, respectively. The bread prepared with L. plantarum sourdough did not result in a significantly lower GI compared with control partially baked bread without sourdough. These results suggest that partially baked frozen sourdough bread with a reduced GI can be prepared using L. brevis and L. fermentum starters. This combined technology was also used for the development of low GI gluten-free breads( Reference Novotni, Cukelj and Smerdel 32 ). Commercial starter PL3 (L. fermentum combined with phytase) was used in four levels, 7·5, 15, 22·5 and 30 g of sourdough/100 g of bread and a control also gluten-free bread without sourdough was prepared. In all, eleven healthy subjects participated in this cross-over design trial and consumed in six separate occasions the five test breads and a glucose solution as reference food. The results showed that sourdough decreased bread’s GI by 9, 16, 14 or 7 units, respectively. Interestingly, the highest amount of sourdough gave the smallest reduction in bread’s GI. The researchers attributed this result to the observed decrease of bread viscosity that could have caused an increased postprandial glucose response, and recommended that sourdough should be added in moderate amounts in order to deliver the desired outcome.
Other processing conditions
Apart from the leavening technique used for bread making, alteration of other bread processing conditions is another strategy that could be used for the improvement of white bread’s GI. Processing can modify the microstructure, composition and availability of different compounds, including starch. The reviewed studies are presented on Table 3.
Human studies examining the effect of other processing conditions on postprandial glucose (G) and insulin response (Mean values and standard errors; mean values and standard deviations; mean values and 95 % confidence intervals)
iAUC, incremental AUC; GI, glycaemic index; II, insulinaemic index; WBB, Western baked bread; OSB, oriental steamed bread; T2D, type 2 diabetes; AC available carbohydrates; WB, white bread; FBNF, fresh baked non-frozen; PBF, partially baked, frozen; WWF, white wheat flour.
* Statistical significance, P <0·05.
† Subjects are healthy individuals, if not, further details are given.
‡ Unless otherwise stated, total dietary fibre content corresponds to g/100 g of bread.
§ More information about how AUC, iAUC, iAUCcut and net iAUC are calculated can be found in the article by Brouns et al.(44). GI and II values are expressed with G as the reference, being assigned the value of 100.
││ Short corresponds to 30 min proving time, whereas long to 50 min. Desem is the Dutch word for leaven and means a fermented dough.
**Breads (50 g of AC) were consumed as part of a breakfast.
It has been recently shown by Lau et al.( Reference Lau, Soong and Zhou 5 ), that even when the same ingredients are used, the application of varied processing procedures including mixing time, mixing intensity, proofing period and method of cooking can significantly influence postprandial glucose response to white bread. In particular, four types of white bread were prepared, a Western baked bread (high protein wheat flour, proofing time 70 min, baked at 210°C for 11 min), an oriental steamed bread, popular staple in Asia (medium protein wheat flour, proofing time 40 min, baked at 100°C for 10 min), a modified baked bread and a modified steamed bread (both with the recipe of Western bread and oriental processing conditions). The ingestion of these four breads by healthy subjects allowed the calculation of their GI values, which were seventy-five for modified baked bread, seventy-one for Western baked bread, sixty-eight for oriental steamed bread and sixty-five for modified steamed bread. Differences in processing procedures caused differences in physical structure, which in turn resulted in starch digestibility differences. It was assumed that steaming resulted in a more compact structure which could inhibit the accessibility of amylase to starch granules leading to a slower rate of glucose release and therefore a lower GI.
Pumpernickel breads are commonly consumed in Germany and Scandinavia and constitute a healthy alternative to wheat bread for T2D patients. They are usually baked for 20 h at 120°C, and as a result they contain high amounts of RS and a mechanically firmer and denser structure. Low processing temperature can block the hydration and swelling of starch granules hindering starch gelatinisation and facilitating the formation of RS( Reference Jenkins, Wolever and Jenkins 34 ). Indeed, in the study by Breen et al.( Reference Breen, Ryan and Gibney 35 ) a pumpernickel rye bread, was shown to promote lower postprandial glucose and insulin responses in T2D patients, compared both with white bread and whole wheat grain bread.
The manipulation of dough proofing, resulting in different loaf volume, was examined in the study by Burton & Lightowler( Reference Burton and Lightowler 36 ). In all, ten healthy subjects participated in a randomised trial and consumed four slices of four breads differing in volume but not in macronutrient composition (glucose was used as reference). The volumes of bread loaves were 1100, 1700, 2400 and 3000 ml. GI values were shown to be significantly reduced by lowering loaf volume, and were found 38, 72, 86 and 100, respectively. This study suggests that reducing bread volume is accompanied by a drop in bread’s GI. The postprandial glucose response to a compact and dense bread made from candeal-flour (poor in gluten), prepared with short fermentation time and commonly consumed in Spain was evaluated in the study by Gonzalez-Anton et al.( Reference Gonzalez-Anton, Rico and Sanchez-Rodriguez 37 ). The GI was calculated 86 (with glucose as reference) showing no difference compared with white bread (GI=76), but II was found lower compared with reference.
The effect of different rising methods and leavening agents on bread’s GI was studied in the study by Fredensborg et al.( Reference Fredensborg, Perry and Mann 38 ). Long fermentation (50 min rise and proofing time) breads were compared with short fermented breads (30 min proofing time), but no difference in GI value was detected. In addition, the use of different leavening agents were compared, yeast, desem (fermented dough, in Dutch) and sourdough, but again they did not significantly affect the GI of bread. These results suggest that increasing the fermentation time by approximately 20 min does not sufficiently influence the GI of bread, and possibly other factors are required for the design of low GI breads. Rizkalla et al.( Reference Rizkalla, Laromiguiere and Champ 39 ) demonstrated a comparison between postprandial plasma glucose and insulin responses after the ingestion of a variety of French breads differing in baking process. Test breads were classic baguette, traditional baguette, loaf of wholemeal bread, loaf of bread fermented with yeast or with leaven and a glucose challenge as reference, and consumed in amounts that contained 50 g of available carbohydrates. Results showed that GI values ranged from 57, for the traditional baguette, to 85 for the wholemeal bread, whereas no significant difference was detected among them. However, the II of the traditional baguette and of the bread fermented with leaven were lower than the other breads. These results were attributed to the long fermentation time and production of organic acids for the bread fermented with leaven, and to the artisanal preparing method that prevented the dough to rise at its maximal capacity, for the traditional baguette.
The application of frozen storage on baking technology has been shown to favourably alter the GI of white bread. A randomised controlled cross-over design trial by Borczak et al.( Reference Borczak, Sikora and Sikora 40 ) studied the effect of both fibre addition and freezing treatment in white-flour wheat rolls on postprandial blood glucose in human volunteers. Healthy subjects consumed four types of wheat rolls which included (1) fully baked, non-frozen, (2) fully baked, non-frozen with dietary fibre (10 %), (3) partially baked and frozen and (4) partially baked and frozen with dietary fibre (10 %), in amounts that yielded 50 g of available carbohydrates. The results showed that both the addition of dietary fibre and freezing applied to wheat rolls significantly reduced the GI by 34 % (GI of partially baked frozen with dietary fibre was 53 with glucose as reference food) compared with fully baked without dietary fibre wheat roll (GI=87). Another study by the same research group, examined the effect of sourdough addition to partially baked and frozen wheat rolls on bread’s GI( Reference Borczak, Sikora and Sikora 41 ). For this purpose, four test wheat rolls were tested by healthy subjects: fully baked non-frozen, fully baked non-frozen with 3 % dehydrated sourdough, partially baked frozen, partially baked frozen with 3 % dehydrated sourdough. The GI values of the products were calculated 87, 63, 67 and 43, respectively. Both factors, freezing and sourdough, significantly reduced the GI of wheat rolls. Heating-cooling cycles that are present during partially baking freezing technology result in the formation of retrograded starch, RS3, which is not susceptible to enzymatic digestion. Unlike the above studies, no difference was detected in terms of postprandial glycaemia and GI values between an ordinary white bread and a precooked frozen white bread, which differed only in the baking procedure( Reference Gonzalez-Anton, Rico and Sanchez-Rodriguez 37 ).
Burton & Lightowler( Reference Burton and Lightowler 42 ) studied the effect of storage and preparation conditions on postprandial glucose response to white bread. In this randomised cross-over design trial, ten healthy volunteers participated. In both homemade and commercial white breads, freezing and defrosting, toasting from fresh and toasting following freezing and defrosting resulted in significantly lower glucose response. The effect of adding extruded chickpea flour to white bread on postprandial glucose and insulin response in healthy subjects was examined by Johnson et al. ( Reference Johnson, Thomas and Hall 43 ), but results did not show any favourable metabolic effects compared with native chickpea flour.
Discussion
The present review aimed to give a run down on the human trials that have focused on the effect of bread making technology on postprandial glucose response to bread. Studies examining the postprandial glucose and insulin response after the ingestion of breads differing in the leavening technique used, proofing time or baking process applied are included in this article, aiming to designate effective strategies for the modulation of bread’s GI. It is noteworthy that there is an ongoing debate about how reliable the measurements of GI values, and II values, are. Consequently, some variation in measurements on the GI of breads could be attributed to that factor.
Sourdough fermentation of either white or whole meal breads has consistently been shown to attenuate the GI( Reference Goletzke, Atkinson and Ek 7 , Reference De Angelis, Damiano and Rizzello 13 , Reference De Angelis, Rizzello and Alfonsi 15 , Reference Novotni, Cukelj and Smerdel 32 – Reference Novotni, Curić and Bituh 33 ). Postprandial glucose response following the consumption of breads leavened with sourdough lactobacilli also results in lower iAUC values, compared with reference foods( Reference Najjar, Parsons and Duncan 14 , Reference Lappi, Selinheimo and Schwab 16 , Reference Scazzina, Del Rio and Pellegrini 21 – Reference Mofidi, Ferraro and Stewart 23 ). Wholemeal wheat and white wheat sourdough breads do not exhibit differences in glucose response, as the presence of insoluble fibre does not seem to influence glycaemic potential of breads( Reference Scazzina, Del Rio and Pellegrini 21 ). According to the accumulated data, the lowest GI value has been demonstrated by De Angelis et al.( Reference De Angelis, Damiano and Rizzello 13 ), and is a wheat sourdough bread enriched with dietary oat and rye fibre manufactured in industrial plant. Sprouting treatment of whole-grain wheat, which is known to increase the content and availability of vitamins, minerals, antioxidants and phytochemicals, in combination with sourdough fermentation led to the production of breads which exerted lower AUC for glycaemia in patients with T2D( Reference Tucker, Vandermey and Robinson 22 ) and overweight or obese subjects( Reference Mofidi, Ferraro and Stewart 23 ). The amount of sourdough used for the manufacture of gluten-free breads influences GI value, with moderate amounts (15 and 22·5 g of sourdough) to exhibit the highest drops in GI, whereas higher amount of sourdough resulted in the lowest drop in GI possibly due to a decrease in viscosity( Reference Novotni, Cukelj and Smerdel 32 ). However these results remain to be confirmed in non-gluten-free breads as well. Different starters also lead to different concentration of lactic and acetic acids, thus different GI values. A lower ratio of lactic:acetic acid has been proposed to be associated with greater GI reduction in gluten-free breads( Reference Novotni, Curić and Bituh 33 ).
The evidence regarding the effect of sourdough breads consumption on postprandial insulin response is less clear. Among the reviewed studies, II of a sourdough rye bread was evaluated in the study by Goletzke et al.( Reference Goletzke, Atkinson and Ek 7 ) and found 70 (with glucose as reference), whereas Rizkalla et al.( Reference Rizkalla, Laromiguiere and Champ 39 ) reported II 59 for a white bread leavened with sourdough. Lower postprandial insulin response to breads fermented with sourdough, compared with reference foods, has been reported in some studies( Reference Lappi, Selinheimo and Schwab 16 , Reference Juntunen, Laaksonen and Autio 19 – Reference Bondia-Pons, Nordlund and Mattila 20 , Reference Mofidi, Ferraro and Stewart 23 ), whereas there are also studies showing no difference( Reference Najjar, Parsons and Duncan 14 , Reference Tucker, Vandermey and Robinson 22 ). The consumption of breads prepared with endosperm rye flour has been associated with improved postprandial insulin profile without affecting glucose response( Reference Rosén, Silva and Andersson 18 ). In accordance to this observation, rye sourdough breads do not consistently result in attenuated glucose response, as there are data proving beneficial impact( Reference Goletzke, Atkinson and Ek 7 ) as well as data demonstrating no difference in postprandial glycaemia( Reference Juntunen, Laaksonen and Autio 19 , Reference Bondia-Pons, Nordlund and Mattila 20 ).
The fermentation of wheat and rye flours by lactic acid bacteria results to the formation of organic acids, especially lactic acid, which are mainly responsible for the acute metabolic benefits. Among the mechanisms reported in the literature slower gastric emptying rate( Reference Liljeberg and Björck 24 , Reference Liljeberg, Åkerberg and Björck 26 ), lower rate of starch digestibility( Reference De Angelis, Rizzello and Alfonsi 15 ), formation of RS( Reference De Angelis, Rizzello and Alfonsi 15 , Reference Scazzina, Del Rio and Pellegrini 21 ), interactions between starch and proteins( Reference Ostman, Nilsson and Liljeberg-Elmstahl 27 ), firmer structure especially in rye breads( Reference Juntunen, Laaksonen and Autio 19 ), release of amino acids, peptides and free phenolic compounds in the gut( Reference Katina, Laitila and Juvonen 30 – Reference Katina, Liukkonen and Kaukovirta-Norja 31 ), and proteolysis( Reference Lappi, Selinheimo and Schwab 16 ) are included. Available data do not allow drawing any safe conclusions. Fermenting bread with sourdough lactobacilli instead of baker’s yeast is potentially an effective strategy for lowering the GI and possibly insulin response of bread products. Besides the beneficial effects of sourdough biotechnology on sensory, structural and shelf life properties, a large body of evidence shows its potential to improve nutritional quality by increasing levels of vitamins and bioactive compounds and improving mineral bioavailability( Reference Brouns, Bjorck and Fryan 44 ).
The manipulation of processing procedures can result in physical structure differences and in turn starch digestibility variation of breads with the same macronutrient composition. Steamed compared with baked white bread has been shown to exert low GI regardless the bread recipe used( Reference Lau, Soong and Zhou 5 ). When ascertaining whether macronutrient composition or processing parameters has a greater impact on postprandial glucose response, the latter was found to play a more pivotal role( Reference Lau, Soong and Zhou 5 ). Similarly the application of pumpernickel bread baking conditions, which are known to hinder the swelling of starch granules thus starch gelatinisation, also lowers postprandial glucose response to bread( Reference Breen, Ryan and Gibney 35 ). Reducing bread volume by modifying proofing conditions resulted in significant decrease in bread’s GI in the study by Burton & Lightowler( Reference Burton and Lightowler 36 ). Dense and compact bread structure possibly results in limited accessibility of amylase to gelatinised starch granules. When no resting period is applied (preventing the dough to rise at its maximum capacity) or the gluten content of flour is low, the high porous bread structure cannot be created, which in turn significantly affects glucose response( Reference Gonzalez-Anton, Rico and Sanchez-Rodriguez 37 ). The increase of proofing time from 30 to 50 min did not significantly alter the GI of bread in another study( Reference Fredensborg, Perry and Mann 38 ). Another way of achieving a high RS content, thus low GI of baked products, is by applying multiple heating-cooling cycles, which promote the retrogradation of starch. Partial-baking freezing technology combined with either the addition of 10 % dietary fibre( Reference Borczak, Sikora and Sikora 40 ) or sourdough in low amount (3 %)( Reference Borczak, Sikora and Sikora 41 ) has resulted in the design of low GI breads. The application of simple household methods such as freezing and toasting of white bread has also shown to favourably alter glucose response to white bread( Reference Burton and Lightowler 42 ).
As far as postprandial insulin response is concerned, data has shown lower II for a dense bread (Candeal-flour bread) compared with reference food (glucose)( Reference Gonzalez-Anton, Rico and Sanchez-Rodriguez 37 ), lower iAUC after the ingestion of a pumpernickel rye bread compared with white bread( Reference Breen, Ryan and Gibney 35 ) and lower II to a traditional white baguette( Reference Rizkalla, Laromiguiere and Champ 39 ), which artisanal method of preparation did not allow the dough to rise properly. The above initial data suggest that there can be an effect of physical structure of bread on postprandial insulin response, resulting from the retarded glucose release and absorption. However, more randomised controlled trials are needed to understand the potential links of bread microstructure and postprandial insulin response.
Conclusions
Although bread is daily consumed in almost every part of the world and could be one of the main determinants of dietary glycaemic load, there is great variation in postprandial glucose and insulin responses, coming from baking technology differences. Currently, baked goods are mainly produced by highly industrialised processes, and the traditional long-time fermentation of the dough has been replaced by the use of baker’s yeast leavening agents. Use of sourdough fermentation is a challenging technology for attenuating the GI of bread. It constitutes not only a tool to exploit the potential of wheat, rye and whole-grain flours, but also an alternative and effective way to decrease glucose and insulin response to bread, whereas the mechanism behind these pronounced effects are still to be fully elucidated. The alteration of other processing conditions could also constitute an innovative route to beneficially alter the postprandial glucose response to bread products, but more research is required in order to examine the potential effects of this strategy on postprandial insulin response as well. Continuous research for novel baking processes that promote consumers health, through lowering postprandial responses, emphasizes the importance of considering alternative baking technologies for bread production.
Acknowledgements
N. S. S. was awarded a fellowship by the Greek State Scholarship Foundation.
The authors’ contributions are as follows: all authors contributed to the design of the article. N. S. S. contributed to the conception of the literature search strategy, undertook the literature search and wrote the manuscript; A. E. Y. contributed to the conception of the literature search strategy and critically evaluated the document at all stages. V. T. K. had the responsibility for producing the last version of the article.
The authors declare that there are no conflicts of interest.
