Low digestible carbohydrates

Polyols are quite closely related to sugars, with the reducing group of the sugar being replaced by a hydroxyl group. Hydrogenated monosaccharides include sorbitol, mannitol, xylitol, and erythritol, and hydrogenated disaccharides include maltitol, lactitol, and isomalt, whereas Lycasin® is a mixture of hydrogenated saccharides and polysaccharides. In the EC, all polyols and HGS are considered as food additives (Le Bot, 1993). Lycasin® (HGS), maltitol, mannitol, and sorbitol are manufactured directly or indirectly from starch, the other polyols being manufactured from other raw materials such as milk, beet, cane sugar, and corn cobs (Billaux et al., 1991). Further details of these processes are to be found elsewhere in this volume.

Polyol content of natural foods

Polyols (mainly sorbitol, mannitol, and xylitol) occur naturally in a large number of fruits and vegetables. In fresh fruits up to approximately 8% polyols can be found, but in certain dried fruits the polyol content increases up to 21 % (Table 1). The polyol content of vegetables is lower than that of fruits but concentrations up to 0.8% have been recorded in some fresh vegetables with up to 1.5% in certain mushrooms (Table 2). Based on data published by the Food and Agriculture Organisation (FAO), Vincent (1989) estimated the theoretical total annual consumption of polyols in fruits as 1.162 X 10⁶ tonnes of sorbitol, 4.8 X 10⁵ tonnes of xylitol, and 2.4 X 10³ tonnes mannitol. He also estimated the intake of polyols from added sweeteners in food and pharmaceutical products to be 2.75 X 10⁵ tonnes, meaning that the intake of polyols added to food is less than 10% of the total polyol intake.

Fruit	Polyol content (mg 100 g-1)	Polyol content (%)
Rowanberry, dried	20984	21.0
fresh	8534	8.5
jam	4080	4.1
Plums, dried	14872	14.9
fresh	2975	3.0
jam	5366	5.4
juice	1511	1.5
Pear, dried	8417	8.4
fresh	1624	1.6
juice	877	0.9
Apricot, dried	4296	4.3
fresh	792	0.8
juice	437	0.4
Apple, dried	3202	3.2
fresh	585	0.6
Peach, dried	1787	1.8
fresh	267	0.7
juice	182	0.2
Orange, dried	1070	1.1
fresh	190	0.2
juice	109	0.1
Wild fruits, fresh	1261	1.3

Vegetable	Polyol content (mg 100 g-1 wet weight)	Polyol content (%)
Ringed boletus	1440	1.4
Mushrooms	1150	1.2
Honey mushrooms	1090	1.1
Parsley (root), fresh	820	0.8
Sweetcorn, fresh	200	0.2
Carrot, fresh	187	0.2
Leek, fresh	150	0.2
Celery, fresh	110	0.1
Cucumber, fresh	70	0.07
Cauliflower, fresh	26	0.03
Pumpkin, fresh	26	0.03

Polyol content of no-sugar-added or energy-reduced products in the marketplace

Polyols and HGS are used in the food, cosmetic and pharmaceutical industries, their main usage being as bulk sweeteners (giving both volume and sweetness) in the production of sugar-free, calorie-reduced, or tooth-friendly foods and foods for diabetics. The technological handling properties of many polyols used in the manufacture of foods are comparable with those of sugar, enabling increased use of polyols as alternative bulk sweeteners to take advantage of their hypocaloric and hypo-cariogenic properties. They are also used in small amounts as humectants. The most important food category for the use of polyols is confectionery (chewing gum, candies, compressed tablets, and chocolate). Table 3 gives a review of the typical polyol content of various sugar-free/ no-added-sugar and some energy-reduced products.

Product	Typical polyol content (%)
Candies
– hard	98
– soft	70-75
Compressed tablets	78-99
Chewing gum	up to 70
– coated chewing gum	up to 80
Chocolate
– bar	39-44
– light recipe	30
Baked goods
– biscuits	20-30
– sponge cake	20-30
Ice cream	15
– light	10
– light (other recipes)	4

Digestion and absorption of hydrogenated glucose syrups and polyols

Hydrogenated glucose syrups

These are first hydrolyzed by α-amylase producing di- and oligosaccharides, after which intestinal disaccharidases split dimers containing only glucose to free glucose which is easily absorbed and utilized by the body. However, disaccharides containing sorbitol linked to glucose are hydrolyzed slowly with little resultant absorption (Dahlqvist and Telenius, 1965). Lycasin 80/55® is a hydrogenated glucose syrup with a particular, defined composition (Sicard and Leroy, 1983) and more than 96% of it is broken down to glucose and sorbitol, the remainder being maltitol (Verwaerde and Dupas, 1982). Most of the glucose is absorbed but some of the sorbitol is not and is presented intact to the large intestine where it is fermented. Maltitol may be further hydrolyzed by α-glycosidase and isomaltase of the brush border but again, some remain intact and enter the colon where fermentation occurs. Very little faecal excretion of polyols occurs after the ingestion of Lycasin®. The sorbitol content of Lycasin® is, therefore, a major determinant of its properties and the digestion of sorbitol and the other polyols are dealt with further below.

Sorbitol

This is absorbed across the small intestine mucosa passively by facilitated diffusion (Lauwers et al., 1985), after which it is oxidized slowly to D-fructose by sorbitol dehydrogenase in the liver then entering the fructose-1-phosphate pathway without stimulating a glycaemic response (see below). However, sorbitol has a relative molecular mass of 182 D which is near the limit for passive diffusion (Dwivedi, 1978) and it is clear that absorption is incomplete (especially at low concentrations) and that residual sorbitol enters the large bowel where it is fermented, this becoming apparent by the appearance of hydrogen in exhaled breath. For example, a significant increase in the amount of hydrogen excreted in breath could be detected when as little as 5 g sorbitol was eaten (Hyams, 1983). Interestingly, Kruger et al. (1992) showed that in rats the digestion of sorbitol lacked dose dependence although the physiological response, in the form of intestinal water movement strongly correlated with the dose administered.

Beaugerie et al. (1990) demonstrated that 79% of a 30 g dose of sorbitol remained unabsorbed when ingested as an oral solution, although the degree of malabsorption was affected by several factors including, importantly, the composition of the solution in which the dose was administered. The mean percentage of sorbitol absorbed in the small intestine was significantly higher in pure sorbitol doses than in those containing maltitol and Lycasin® 80/55. (In contrast, the mean percentage of total maltitol metabolized was not significantly different for pure maltitol and maltitol contained in Lycasin® 80/55.) The vehicle of sorbitol ingestion was also considered by Zumbe and Brinkworth (1992) who studied the acute consumption of 20 g sorbitol incorporated into milk chocolate. They demonstrated incomplete sorbitol absorption that was manifest as osmotic diarrhoea; however, the subjects in their study showed increased tolerance if the dose was spread out over the day. Studies in which [U-¹⁴C]sorbitol was given either to normal rats (Ertel et al., 1983) or germ-free rats (Wursch et al., 1990) demonstrated 55% and 51% residual activity respectively in the gut contents, most being in the caecum. Beaugerie et al. (1990) were able to further show that the faecal excretion of sorbitol is negligible indicating that it is almost completely digested by the colonic flora.

Mannitol

Mannitol, like its isomer sorbitol, is only partially and slowly absorbed in the small intestine. Unabsorbed mannitol passes into the colon where it is fermented by bacteria. Nasrallah and Iber (1969) reported that on average 35% of an ingested dose of [U-¹⁴C]mannitol is excreted in the stools and 20% in the urine. The fact that very little ¹⁴CO₂ was expired by subjects in the study referred to above suggests that very little mannitol is metabolized in the tissues and what is absorbed is excreted in the urine.

Xylitol

This differs from sorbitol and mannitol in that it is a pentitol. Absorption from the small intestine occurs by diffusion after which it is mostly metabolized in the liver to D-xylulose, which is then phosphorylated and converted to hexose-6-phosphate and eventually glucose (Wang and Van Eys, 1981).

Erythritol

This differs from all the other polyols so far considered in that it is a four-carbon sugar alcohol. It occurs naturally (Shindou, 1989) but can also be produced in large amounts by fermentation of glucose and sucrose with Aureobasidium sp. (Goossens and Roper, 1993). Using [U-¹³C]-labelled erythritol Hiele et al. (1993) reported that erythritol is rapidly absorbed in the small intestine by passive diffusion but that no ¹³CO₂ is excreted in the exhaled breath. Oku and Noda (1990) found that when erythritol was administered to humans at 0.1 g kg-1 body weight, 80% was excreted in the urine and 6% in the expired air. Presumably, the colonic bacteria of Oku and Nodas’ subjects had adapted to metabolize whatever polyol entered the colon, whereas in Heile’s subjects they had not? Erythritol has 75% of the sweetness of sucrose so it might be useful as a sweetener which because of its rapid absorption will not stimulate gastrointestinal symptomatology and, because it is mostly excreted in the urine without a post-prandial insulin response, might be suitable for diabetics. However little work is available in the public domain concerning this sugar alcohol and diuretic effects might prove to be a problem.

Lactitol

Lactitol ( 4-O-β-D-galactopyranosyl-D-sorbitol) is a disaccharide alcohol that requires hydrolysis prior to absorption and if completely hydrolyzed yields equimolar amounts of galactose and sorbitol. However, Nilsson and Jagerstad (1987) used human intestinal biopsies in vitro, to demonstrate that lactitol is a poor substrate for lactase which showed an activity of only 1 % of that towards lactose which itself is around 25% of that of sucrose. Furthermore, lactitol is not absorbed intact from the small intestine to any great extent (Metzger et at., 1988; Patil et at., 1987). Beaugerie et al. (1991) showed that up to 84% of a 20 g dose of lactitol given in an iso-osmotic solution remained unabsorbed. Using [U-¹⁴C)lactitol Grimble et al. (1988) showed that 62.9% of a 20 g dose was excreted as ¹⁴CO₂ via the lungs whilst only 6.5% and 2% were recovered from the urine and faeces, respectively. This suggests that lactitol is extensively fermented in the human colon and that a significant proportion of the bacterial metabolites are available for colonic absorption and ultimately metabolized in the tissues.

Maltitol

Maltitol (4-O-α-D-glucopyranosyl-D-sorbitol) is hydrolyzed by enzymes of the small intestine brush border (Lian-Loh et al., 1982; Rosiers et al., 1985; Wursch and Del Vedovo, 1981; Wursch et al., 1990; Ziesenitz and Siebert, 1987), particularly by the α-glycosidase of the sucrase-isomaltase complex (Zunft et al., 1983). This enzyme has a much lower affinity for maltitol than it does for either sucrose or maltose (Kamoi, 1975), releasing equimolar amounts of glucose and sorbitol. However, maltitol is also hydrolyzed by isomaltase (Zeisenitz, 1986). In their in vitro study, Nilsson and Jagerstad (1987) reported that maltase has 10% activity towards maltitol compared to maltose. Nevertheless, Beaugerie et al. (1991) calculated that about 90% of ingested maltitol is digested with subsequent absorption of approximately 70% of the sorbitol released and nearly all the glucose. Hence they suggested that approximately 75% of the ingested dose is digested with the remaining 25% entering the colon as maltitol and sorbitol (2:3). According to Beaugerie et al. (1990) any maltitol reaching the small intestine is almost completely digested by the colonic bacteria and almost none appears in the faeces. On the other hand, Rennhard and Bianchine (1976) gave doses of [U-¹⁴C)-labelled maltitol to volunteers and detected 3.6% and 4.6% in the urine and stools, respectively.

Most of the other research concerning the digestion and absorption of maltitol comes from non-human studies but throws some light on what might occur in the human gut. Wursch et al. (1990) showed that in rats and mice 40% of ingested maltitol reached the colon, rather more than is proposed for humans. Contrary to this, Oku et al. (1971) claimed that although 5% of [U-¹⁴C]maltitol ingested by rats was absorbed directly, being rapidly excreted in the urine, the greater part passed into the colon where it was fermented rather than being hydrolyzed by digestive enzymes in the small intestine. Similar results were also obtained in rat studies by Lian-Loh et al. (1982). Kruger et al. (1992) studied the kinetics of gastrointestinal transit of different loads of maltitol in rats and showed that its digestibility and absorption in the small intestine depends inversely on the dose ingested and also on the maltitol cleavage product sorbitol being very slowly absorbed. In general, however, overall support can be given to the conclusion of Nilsson and Jagerstad (1987) that, unlike lactitol and isomalt, significant amounts of maltitol can be digested and utilized by humans.

Isomalt

Isomalt (D-glucosyl-α (1-1)-D-mannitol and D-glucosyl-α-(1-6)-D-sorbitol) is cleaved very slowly by the small intestinal hydrolases (Grupp and Siebert, 1978). The α-glycosidase of the sucrase-isomaltase complex hydrolyses isomalt to yield 50% glucose and 25% each of sorbitol and mannitol. Grupp and Siebert concluded that, in humans, the relative rate of isomalt cleavage by α-glycosidase was 2% that of maltose. Nilsson and Jagerstad (1987) used human intestinal biopsies in vitro to demonstrate that the hydrolysis of isomalt is about 10% of the expected rate for palatinose but only 1 to 3% of that towards isomaltose. Using mucosal homogenates, purified brush border enzymes and everted segments of rat jejunum, Goda et al. (1988) demonstrated that hydrolysis of isomalt was 6 to 7% of that of sucrose. More recently, Kruger et al. (1991) demonstrated that in rats isomalt hydrolysis was more complete when administered at lower doses than at higher doses. In pigs, about 60% of ingested isomalt passes the end of the small intestine (Van Weerden and Huisman, 1993a, b) but none appears in the faeces, indicating that it is completely digested by bacterial fermentation in this species. In humans, Grupp and Siebert (1978) found that very little isomalt was excreted in either the faeces or urine, levels being 0.2% and 0.1 %, respectively.

Insulin response following ingestion of low digestible carbohydrates

Rising plasma glucose levels following glucose absorption from the alimentary β-cells of the islets of Langerhans via the mediator GIP. Sucrose is known to elicit a higher plasma glucose response and thus a higher insulin response than complex carbohydrates (Crapo et at., 1976). The rate of absorption into the blood of glucose from glucose syrups is in the same order as that from glucose per se, but compared to sucrose there are some advantages in that glucose syrups fail to stimulate the rise in plasma lipids seen after high sucrose diets and consequently a smaller amount of fat is laid down in the liver (Lian-Loh et al., 1982). It was subsequently determined that the total serum insulin response over the 2 h period following the consumption of Lycasin® and maltitol syrups was significantly smaller than that following the consumption of D-glucose. Hydrogenated glucose syrups thus behaved differently from normal glucose syrups resulting in a smaller insulin response, possibly because of the non-availability of some covalently bound glucose in HGS (Lian-Loh et al., 1982).

Lactose intolerant people are also known to have a flat blood glucose curve after the ingestion of lactose because they lack lactase, the enzyme which is necessary to hydrolyze lactose. For similar reasons polyols are also absorbed minimally from the small intestine and they typically lead to glycaemic and insulin responses which are characteristically lower than those occurring after the consumption of either glucose or sucrose. For this reason, HGS and polyols have both been used for several years to sweeten products for diabetics who need to avoid large post-prandial rises in glucose and insulin.

In a prospective double-blind controlled crossover study on ten healthy males aged 21 to 30 years, Thiebaud et al. (1984) investigated the acute effects of oral ingestion of 30 g loads of isomalt and sucrose on plasma glucose and insulin. After sucrose, plasma glucose and insulin increased markedly in the first 60 min whereas after isomalt, plasma glucose and insulin levels increased only slightly during the first 90 min. The maximum increment for plasma glucose was 2.78 ± 0.29 versus 0.33 ± 0.29 mmol l^-1 after isomalt. For plasma insulin, the maximum increment was 22.4 ± 7.2 after glucose versus 3.4 ± 1.7 μU ml^-1 after isomalt showing a sevenfold to eightfold difference in maximum values for both glucose and insulin levels.

Maltitol also causes a lower rise in blood glucose and insulin levels than a corresponding oral dose of glucose or sucrose. Felber et al. (1987) conducted a study in which eight normal subjects consumed 30 g of either maltitol or sucrose on two separate days with an interval of at least a week. Plasma glucose and insulin levels both peaked 30 min after ingestion of the load for both sugars, but the oral load of maltitol produced markedly less stimulation of carbohydrate oxidation than an equivalent load of sucrose. The peak of the glucose concentration was significantly smaller after maltitol than after sucrose (2.1 ± 0.4 versus 3.8 ± 0.4 mmol l^-1 ) as was the peak of the insulin concentration after maltitol and sucrose, these being 9.3 ± 2.7 versus 25.5 ± 5.0 μU ml^-1 respectively. Recently, Tsuji et al. (1990) also showed that after the intake of 10 g of either glucose or maltose by each of five subjects, their serum glucose and insulin levels reached a maximum at 30 min post-consumption. However, there was only a slight glycaemic response after taking maltitol and neither serum glucose nor insulin levels showed any response after taking sorbitol.

As in the previous examples, plasma glucose levels did not increase following the consumption of lactitol (0.5 g kg^-1) by six healthy subjects aged 19 to 38 years and eight patients with liver cirrhosis (Metzger et al., 1988). Mean basal plasma glucose levels were 5.8 and 4.4 mmol l^-1 and the maximum increments following lactitol were +1.4 and +0.6 mmol l^-1 for cirrhotic and healthy subjects, respectively.