Journal of Human Nutrition and Dietetics

RESEARCH PAPER
Dietary sorbitol and mannitol: food content and distinct
absorption patterns between healthy individuals and
patients with irritable bowel syndrome
C. K. Yao, H.-L. Tan, D. R. van Langenberg, J. S. Barrett, R. Rose, K. Liels, P. R. Gibson &
J. G. Muir
Departments of Gastroenterology, Eastern Health and Central Clinical Schools, Monash University, Melbourne, Victoria, Australia

Keywords Abstract
hydrogen breath tests, intestinal absorption,
irritable bowel syndrome, mannitol, polyols, Background: Sorbitol and mannitol are naturally-occurring polyol isomers.
sorbitol malabsorption. Although poor absorption and induction of gastrointestinal symptoms by
sorbitol are known, the properties of mannitol are poorly described. We
Correspondence aimed to expand data on food composition of these polyols, and to com-
C. K. Yao, Department of Gastroenterology,
pare their absorptive capacities and symptom induction in patients with
Central Clinical School, Monash University, Level
irritable bowel syndrome (IBS) and healthy individuals.
6, The Alfred Centre, 99 Commercial Road,
Melbourne, VIC 3004, Australia. Methods: Food samples were analysed for sorbitol and mannitol content.
Tel.: +61 3 9903 0270 The degree of absorption measured by breath hydrogen production and gas-
Fax: +61 3 9903 0556 trointestinal symptoms (visual analogue scales) was evaluated in a rando-
E-mail: [email protected] mised, double-blinded, placebo-controlled study in 21 healthy and 20 IBS
subjects after challenges with 10 g of sorbitol, mannitol or glucose.
How to cite this article Results: Certain fruits and sugar-free gum contained sorbitol, whereas man-
Yao C.K., Tan H.-L., van Langenberg D.R., Barrett
nitol content was higher in certain vegetables. Similar proportions of
J.S., Rose R., Liels K., Gibson P.R. & Muir J.G.
patients with IBS (40%) and healthy subjects (33%) completely absorbed
(2013) Dietary sorbitol and mannitol: food
content and distinct absorption patterns between sorbitol, although more so with IBS absorbed mannitol (80% versus 43%;
healthy individuals and patients with irritable P = 0.02). Breath hydrogen production was similar in both groups after
bowel syndrome. J Hum Nutr Diet. lactulose but was reduced in patients with IBS after both polyols. No differ-
doi:10.1111/jhn.12144 ence in mean (SEM) hydrogen production was found in healthy controls
after sorbitol [area-under-the-curve: 2766 (591) ppm 4 h–1] or mannitol
[2062 (468) ppm 4 h–1] but, in patients with IBS, this was greater after sor-
bitol [1136 (204) ppm 4 h–1] than mannitol [404 (154) ppm 4 h–1;
P = 0.002]. Overall gastrointestinal symptoms increased significantly after
both polyols in patients with IBS only, although they were independent of
malabsorption of either of the polyols.
Conclusions: Increased and discordant absorption of mannitol and sorbitol
occurs in patients with IBS compared to that in healthy controls. Polyols
induced gastrointestinal symptoms in patients with IBS independently of
their absorptive patterns, suggesting that the dietary restriction of polyols
may be efficacious.

ety of food. Recently published food composition data by

Introduction
our group indicated that sorbitol tends to be more com-
Sorbitol and mannitol are polyols frequently used as sugar mon in fruits, whereas mannitol is found more commonly
substitutes by the food industry to produce ‘low-calorie’ in vegetables (Muir et al., 2009). Both polyols appear to
products (Zumbe et al., 2001). It is less commonly known have primary roles in these plants as energy reserves and
that both polyols are also present naturally in a wide vari- as osmoregulatory agents (Lewis & Smith, 1967).

ª 2013 The British Dietetic Association Ltd. 1

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

Sorbitol and mannitol are six-carbon polyol isomers capacity of sorbitol and mannitol in patients with IBS com-
with a similar molecular weight and size, differing only in pared to those in healthy individuals and assessing symp-
the orientation of one of their hydroxyl group (Le & tomatic responses in relation to patterns of absorption.
Mulderrig, 2001; see Figure S1). They appear to be partly
absorbed via passive diffusion across the small intestinal
Materials and methods
epithelium (Beaugerie et al., 1990; Krugliak et al., 1994).
Breath hydrogen studies indicate a high prevalence of sor- Measurement of sorbitol and mannitol content
bitol malabsorption in healthy adults, where 71% show A total of 73 foods from several food groups were screened
malabsorption after consumption of 10 g of sorbitol and for quantities of sorbitol and mannitol. As previously
20% report gastrointestinal symptoms of bloating, flatu- described in detail by Muir et al. (2009), the sampling and
lence and abdominal pain (Hyams, 1983). Some studies processing methods of the foods complied with the guide-
have extrapolated these findings to the absorption of lines of Food Standards Australia New Zealand. Triplicates
mannitol, and have assumed that a similar proportion of of the extracts from food samples were analysed for sorbi-
mannitol is absorbed; however, this has not been com- tol and mannitol content using high-performance liquid
pared directly (Wursch et al., 1989; Le & Mulderrig, chromatography (HPLC) with an evaporative light scatter-
2001). ing detector (ELSD) analytical technique. Sugar standards,
Short-chain carbohydrates called FODMAPs (Ferment- HPLC apparatus (consisting of ELSD Waters 2424, HPLC
able, Oligo-, Di-, Monosaccharide and Polyols) that are pump Waters 515, Waters auto sampler 717 plus and
slowly or poorly absorbed in the small intestine increase Waters column heater; Waters Corp, Milford, MA, USA)
luminal water content of the small and large intestine by and chromatographic procedure were performed as
virtue of their osmotic effect (Barrett et al., 2010; Marci- described in Muir et al. (2009). Samples were reanalysed if
ani et al., 2010) and are rapidly fermented by intestinal any of the triplicate measurement values differed by >5%
bacteria, resulting in increased gas production (Ong et al., from the mean of other samples (Muir et al., 2009). The
2010). The resultant luminal distension has been hypoth- relative SD for using HPLC measurement technique has
esised to trigger abdominal symptoms in patients with previously been calculated for sorbitol (3.76%) and man-
functional bowel disorders because of the visceral hyper- nitol (2.85%) (Muir et al., 2009). Food composition of
sensitivity and dysmotility that characterises such patients sorbitol and mannitol were constructed using standard
(Chey et al., 2001; Agrawal et al., 2008). Because sorbitol serving sizes and weights from FOODWORKS, version 6 (Xyris
and mannitol are also poorly or slowly absorbed in the Software Australia Pty Ltd, Highgate Hill, Australia).
small intestine and readily fermented by bacteria, it is
assumed that they will also induce symptoms in a similar
Subjects
way. However, it is not well established whether the mal-
absorption of sorbitol with or without fructose and, sub- Patients with IBS and no other co-existing gastrointestinal
sequently, the induction of symptoms occurs to a greater disorders were recruited from an outpatient breath testing
extent in patients with irritable bowel syndrome (IBS) centre at Box Hill Hospital (Box Hill, Victoria, Australia).
compared to controls because one study attributes a clear Healthy participants with no previous history of gastroin-
difference (Fernandez-Banares et al., 1993) and another testinal illness or symptoms were also recruited through
does not (Nelis et al., 1990). Regardless of the inconsis- advertisements in Box Hill Hospital. All participants were
tent findings, it is appropriate to restrict the intake of aged ≥18 years and not pregnant. None of them had been
sorbitol and mannitol-rich foods in symptomatic individ- treated with antibiotics in the previous 6 weeks, or pre- or
uals. However, knowledge of the sorbitol and mannitol probiotics in the 2 weeks leading up to the study. The
content in foods is limited and needs to be expanded to intake of nonsteroidal anti-inflammatory drugs and alcohol
prevent unnecessary restrictions. was recorded. During screening, a detailed history was
It was hypothesised that, first, as isomeric molecules, obtained to confirm that IBS patients met the Rome III
sorbitol and mannitol are absorbed similarly in the small diagnostic criteria (Longstreth et al., 2006). Written
intestine of healthy subjects and patients with IBS; second, informed consent was obtained from all participants prior
both would induce gastrointestinal symptoms predomi- to commencing the study. The study protocol was approved
nantly in patients with IBS; and third, symptom induction by the Eastern Health Research and Ethics Committee.
would be associated with the malabsorption of the polyol.
The present study therefore aimed to address the knowl-
Experimental design
edge gaps by quantifying the content of sorbitol and
mannitol naturally-occurring in various foods and also to Prior to entering the study, all subjects were initially chal-
address the hypothesis by examining patterns of absorptive lenged with 15 g of lactulose (15% w/v) for 2 h to ensure

2 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

that they produced adequate breath hydrogen, denoted as carbohydrate) (Rumessen et al., 1990). Responses were
a rise in breath hydrogen ≥10 ppm in two consecutive arbitrarily categorised as marked (61–100%), moderate
breath samples (Bate et al., 2010). Subjects who had a rise (31–60%), low (6–30%) or no (0–5%) malabsorption.
in breath methane ≥15 ppm from baseline were classified Oro-caecal transit time (OCTT), defined as the time (min)
as ‘methane producers’. This was followed by a routine from sugar ingestion of first breath hydrogen rise
breath hydrogen test with 35 g of fructose (17% w/v) to ≥10 ppm in two consecutive samples, was measured for
determine the presence of fructose malabsorption. lactulose (Bate et al., 2010).
The effects of a single dose (10 g) of sorbitol, mannitol
or glucose (as placebo) were investigated using a rando-
Assessment of gastrointestinal symptoms
mised, double-blinded, placebo-controlled, cross-over
design. Randomisation of the test sequence was per- Gastrointestinal symptoms were assessed by completing
formed with a computer-generated list of random num- bowel symptom questionnaires before and 4 h after sugar
bers and the sequence was kept by an independent ingestion. Symptoms were rated on 100-mm visual
investigator. The dose of 10 g was chosen to reflect a analogue scales (VAS) according to severity from 0 (no
realistic but high intake of the individual polyol that is symptoms) to 100 mm (worst it has been) for overall gas-
achievable through the diet (Muir et al., 2009; Barrett & trointestinal symptoms, abdominal pain/discomfort, bloat-
Gibson, 2010) and is also not likely to induce osmotic ing, wind, as previously applied (Francis et al., 1997). A
diarrhoea in healthy individuals (Hyams, 1983). The sug- composite score for abdominal symptoms was calculated by
ars were given as a colourless solution in 100 mL of water combining scores for abdominal pain, bloating and wind.
(10% w/v), made up in identical unmarked containers to Scores were corrected for the baseline level of symptoms.
ensure blinding of both participants and researchers. The
absorption of these sugars was then determined through
Statistical analysis
breath hydrogen testing. Each sugar challenge was fol-
lowed by a washout period of 1–7 days before crossing Because a per-protocol analysis was planned, only partici-
over to the next sugar challenge to allow symptoms pants who completed all three interventions were included
(if any) to return to baseline prior to re-challenge. in the final analysis. Primary end-points were the propor-
tion of individuals with malabsorption and their breath
hydrogen response to the sugar challenges. Secondary end-
Breath hydrogen testing
points included gastrointestinal symptom response to
Breath hydrogen testing was performed as previously out- polyol ingestion, the correlation of breath hydrogen-
lined in detail (Ong et al., 2010). Briefly, participants producing ability and OCTT for lactulose with polyol
were provided with a diet low in fibre and poorly- absorption/malabsorption, and the proportion of individu-
absorbed short-chain carbohydrates for 24 h. After an 8-h als with concomitant fructose and polyol malabsorption. A
fast, a baseline breath sample was taken before ingesting sample size of 20 participants in each group was required
the sugar solution. Thereafter, breath samples were col- to detect a 30% change in breath hydrogen response to the
lected at 15-min intervals into collection bags (Quintron polyol challenges, with a power of 80%. P ≤0.05 (two-
Instrument Co., Milwaukee, WI, USA) for 4 h. Subjects tailed) was considered statistically significant.
were instructed to refrain from smoking and strenuous The frequency of sugar malabsorption was compared
exercise during the test to minimise potential confound- using Fisher’s exact or a chi-squared test. Comparisons
ing of these variables on breath hydrogen responses between subject groups were performed with unpaired
(Thompson et al., 1985). Breath hydrogen and methane t-tests and the Mann–Whitney U-test for nonparametric
concentrations were analysed immediately using a gas data. Comparisons between sugar challenges were also
chromatograph (Microlyzer Model DP Plus; Quintron analysed using paired t-tests and the nonparametric Wil-
Instrument Co., Milwaukee, WI, USA). coxon signed-rank tests, respectively. Correlations were
A rise in breath hydrogen concentrations ≥10 ppm performed using Pearson’s and Spearman’s correlation for
above baseline in two consecutive breath samples was used nonparametric data. Effect sizes (eta squared statistic, r2)
to define malabsorption. Total breath hydrogen and were calculated and expressed as the mean difference
methane production for lactulose and test sugars were (MD) and 95% confidence interval (CI). Results are
estimated using the trapezoid rule to approximate the expressed as the mean (SEM) for parametric data, median
area-under-the-curve (AUC) against time (ppm h–1) for [interquartile range (IQR)] for nonparametric and dis-
each sugar (Kotler et al., 1982). The degree of sugar mal- continuous variables, and proportions as percentages
absorption was semi-quantitatively determined relative to (95% CI). Statistical tests for data analysis were computed
AUC after ingestion of 15 g of lactulose (a non-absorbable using SPSS, version 19.0 (IBM Corp., Armonk, NY, USA).

ª 2013 The British Dietetic Association Ltd. 3

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

2 h–1; P = 0.84; unpaired t-test]. In the six healthy con-

Results
trols and six patients with IBS who were classified as
The composition of sorbitol and mannitol in a range of ‘methane-producers’, breath methane production after
foods is listed in Table 1. Of 73 foods analysed, sorbitol lactulose was similar between controls [805 (206) ppm
was detected in 12, namely in fruits and sugar-free chew- 2 h–1] and IBS patients [897 (359) ppm 2 h–1; P = 0.85,
ing gum. Sorbitol content (per average portion eaten) unpaired t-test].
was the highest in prunes > dried apricot > dried
pear > sugar-free chewing gum > plum > apricot > apple
Frequency of polyol malabsorption
juice > dried apple > wasabi > cherries > coconut milk >
dried coconut. Mannitol was present in seven foods and The proportion of patients with IBS and healthy con-
these were predominantly vegetables, six of which con- trols with sugar malabsorption is shown in Fig. 2. Mal-
tained moderate amounts when a standard portion size is absorption of sorbitol occurred at a similar frequency in
consumed. Celery contained the highest amount (per both groups (P = 0.67; Fisher’s exact test). By contrast,
average portion eaten) of mannitol, followed by butternut mannitol malabsorption occurred less frequently in IBS
pumpkin > pomegranate > celeriac. Neither sorbitol, nor patients compared to controls (P = 0.02). No subject
mannitol was detected in nuts and legumes. malabsorbed glucose. As shown in Fig. 3, a greater pro-
portion of patients with IBS were classed as having a
low degree or no malabsorption of mannitol compared
Subject characteristics
to healthy controls (P = 0.02; chi-squared). By contrast,
Forty-one participants completed all sugar challenge tests. no difference was observed between the two subject
This included 20 subjects with IBS and 21 healthy con- groups for sorbitol (P = 0.24). No differences were
trols. Eleven IBS subjects were constipation-predominant, observed across IBS subtypes according to bowel habits
seven were diarrhoea-predominant and two had alternat- (data not shown).
ing bowel habits. Table 2 compares the baseline charac- Fructose malabsorption occurred in 63% of patients
teristics between the two groups. No differences in with IBS and 42% of controls. Its frequency was not asso-
characteristics were observed. ciated with having malabsorption of both polyols, one of
the polyols or none in patients with IBS (16% versus
21% versus 26%, respectively; P = 0.84) or in healthy
Gastrointestinal symptoms
controls (21% versus 11% versus 11%; P = 0.21).
Severity of gastrointestinal symptoms for healthy and IBS
subjects as scored by the VAS after all three sugar chal-
Relationship between oro-caecal transit time and polyol
lenges are shown in Fig. 1. In the healthy controls, overall
malabsorption
symptoms were generally minimal compared to IBS sub-
jects (sorbitol, P = 0.02; mannitol, P = 0.004) with no Median (range) OCTT in patients with IBS was 105
differences in scores across both polyols compared to glu- (75–135) min, which was significantly longer than 75 (45–
cose (not significant, Wilcoxon signed-rank test) and 120) min observed in healthy controls (P = 0.03, Mann–
composite score (not significant). For patients with IBS, Whitney U-test). However, no correlations were found in
the development of symptoms after the ingestion of glu- the whole group between median OCTT and total AUC
cose was uncommon and very mild (Fig 1). Compared to (sorbitol: r = 0.20, P = 0.22; mannitol: r = 0.28, not
glucose, significantly greater overall gastrointestinal symp- significant) for either sorbitol or mannitol.
toms developed after sorbitol (P = 0.05) and mannitol
(P = 0.02). Other symptoms were more frequent after the
Breath hydrogen production
polyols, although the composite scores did not reach
statistical significance compared to those after glucose Breath hydrogen profiles over 4 h after the ingestion of
(P = 0.12). There were no differences in the development sorbitol, mannitol and glucose for the 27 subjects who
of symptoms after sorbitol compared to those after man- malabsorbed sorbitol and/or mannitol are shown in
nitol (not significant). Fig. 4. In healthy controls, total breath hydrogen pro-
duction after sorbitol [2766 (591) ppm 4 h–1] and man-
nitol [2062 (468) ppm 4 h–1] were similar (P = 0.07,
Breath responses to lactulose
paired t-test) (Fig. 4a), whereas that after glucose was
There were no differences in mean breath hydrogen markedly less [211 (113) ppm 4 h–1; P = 0.01 for both].
produced after lactulose between patients with IBS The mean (95% CI) difference in AUC between sorbitol
[1334 (282) ppm 2 h–1] and controls [1412 (250) ppm and glucose was 3452 (2064–4839) ppm 4 h–1. Similarly,

4 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

Table 1 Content of sorbitol and mannitol in various food groups shown as grams per 100 g edible portion and grams per serving

Sorbitol Mannitol

Amount according to Present Published Present Published

Food serving sizes study data USDA study data USDA

Vegetables Artichoke hearts g/100 g 0 0

g/1 cup (204 g) 0 0
Carrot g/100 g 0 0 0.2
g/1 cup (116 g) 0 0 0.2
Celeriac g/100 g 0 0.1
g/1 celeriac (72 g) 0 0.1
Celery g/100 g 0 0 1.5 0.2
g/1 cup (127 g) 0 0 1.9 0.2
Leek g/100 g 0 0
g/1 leek (83 g) 0 0
Lettuce, rocket g/100 g 0 0
g/1 cup (35 g) 0 0
Sugar snap peas g/100 g 0 0
g/1 cup (120 g) 0 0
Pumpkin, butternut g/100 g 0 0.4
g/1 cup (120 g) 0 0.5
Parsnip g/100 g 0 0
g/1 cup (124 g) 0 0
Tomatoes, sundried g/100 g 0 0
g/1 TB (7 g) 0 0
Radish g/100 g 0 0 0 0.1
g/1 radish (20 g) 0 0 0 0.0
Seaweed g/100 g 0 Trace
g/0.25 cup (40 g) 0 Trace
Water chestnuts g/100 g 0 0
g/1 cup (120 g) 0 0
Asparagus g/100 g 0* 0.1*
g/4 spears (60 g) 0 0.1
Bok choy g/100 g 0.2* 0*
g/1 cup (85 g) 0.2 0
Brussel sprouts g/100 g 0.2* 0*
g/1 cup (164 g) 0.3 0
Broccoli g/100 g 0.4* 0*
g/1 cup (94 g) 0.4 0
Cabbage, common g/100 g 0.2* 0*
g/1 cup (94 g) 0.2 0
Capsicum, green g/100 g 0.4* 0*
g/1 cup (102 g) 0.4 0
Cauliflower g/100 g 0* 2.6*
g/1 cup (132 g) 0 3.4
Corn, sweet g/100 g 0.5* 0*
g/1 cob (85 g) 0.4 0
Cucumber g/100 g 0* 0 0* 0.1
g/1 cup (128 g) 0 0 0 0.1
Mushrooms g/100 g 0.1* 2.6*
g/1 cup (74 g) 0.1 1.9
Snow peas g/100 g 0* 1.2*
g/10 pods (33 g) 0 0.4
Sweet potato g/100 g 0* 0.3*
g/1 cup (140 g) 0 0.4
Legumes Beans, continental g/100 g 0 0
g/0.5 cup (90 g) 0 0
Baked beans g/100 g 0 0
g/0.5 cup (90 g) 0 0

ª 2013 The British Dietetic Association Ltd. 5

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

Table 1 (Continued)

Sorbitol Mannitol

Amount according to Present Published Present Published

Food serving sizes study data USDA study data USDA

Lima beans g/100 g 0† 0.1†

g/0.5 cup (90 g) 0 0.1
Fruits Plum g/100 g 2.4 0.6 0 0
g/1 fruit (66 g) 1.6 0.4 0 0
Apple g/100 g 1.2* 0.3 0* 0
g/1 fruit (165 g) 1.9 0.5 0 0
Apricot g/100 g 1.2 0.8 0 0
g/1 fruit (112 g) 1.3 0.9 0 0
Blackberries g/100 g 4.1* 0*
g/10 berries (50 g) 2.1 0
Boysenberries g/100 g 0 0
g/10 berries (80 g) 0 0
Carambola g/100 g 0 0
g/1 fruit (116 g) 0 0
Cherries g/100 g 0.7 1.0–2.1 0 0
g/5 cherries (35 g) 0.3 0.3–0.7 0 0
Grapes g/100 g 0* 0.1 0* 0
g/10 grapes (40 g) 0 0.0 0 0
Longon g/100 g 0.7* 0*
g/5 fruits (15 g) 0.1 0
Nectarine g/100 g 1.0* 0.6 0* 0
g/1 fruit (151 g) 1.5 0.9 0 0
Peach g/100 g 0.9* 0.2 0.5* 0
g/1 fruit (145 g) 1.3 0.3 0.7 0
Figs g/100 g 0 0
g/1 fruit (50 g) 0 0
Pear g/100 g 2.3* 2.3 0* 0
g/1 fruit (166 g) 3.8 3.8 0 0
Pomegranate g/100 g Trace 0.3
g/1 fruit (76 g) Trace 0.2
Rhubarb g/100 g 0 0
g/1 stalk (104 g) 0 0
Strawberries g/100 g 0 0
g/5 berries (60 g) 0 0
Tamarillo g/100 g 0 0
g/1 fruit (44 g) 0 0
Dried apple g/100 g 1.9 0
g/0.5 cup (45 g) 0.9 0
Dried apricot g/100 g 6.0 0
g/0.5 cup (67 g) 4.1 0
Dried, shredded g/100 g 0.6 0
coconut
g/0.5 cup (37 g) 0.2 0
Dried pear g/100 g 8.1 0
g/6 pieces (27 g) 2.2 0
Prunes g/100 g 10.8 12.0 0 0.0
g/0.25 cup (80 g) 8.7 9.6 0 0.0
Grains and Pumpernickel g/100 g 0 0
cereals bread
g/1 slice (50 g) 0 0
Spelt bread, organic g/100 g 0 0
g/2 slices (52 g) 0 0

6 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

Table 1 (Continued)

Sorbitol Mannitol

Amount according to Present Published Present Published

Food serving sizes study data USDA study data USDA

Raisin toast g/100 g 0 0

g/1 slice (75 g) 0 0
Naan bread g/100 g 0 0
g/1 piece (46 g) 0 0
Oat bran g/100 g 0 0
g/1 cup (137 g) 0 0
Wheat bran, processed g/100 g 0 0
g/1 cup (63 g) 0 0
Wheat bran, unprocessed g/100 g 0 0
g/1 cup (63 g) 0 0
Basmati rice g/100 g 0 0
g/1 cup (137 g) 0 0
Polenta/cornmeal g/100 g 0 0
g/1 cup (254 g) 0 0
Nuts Almonds g/100 g 0 0
g/10 nuts (12 g) 0 0
Cashew g/100 g 0 0
g/10 nuts (12 g) 0 0
Peanuts g/100 g 0 0
g/10 nuts (12 g) 0 0
Beverages Apple juice g/100 g 0.5 1.0 0 0
g/1 cup (250 mL) 1.1 2.1 0 0
Orange juice g/100 g 0 0
g/1 cup (250 mL) 0 0
Coconut milk g/100 g 0.1 0
g/1 cup (250 mL) 0.3 0
Beer g/100 g 0 0
g/1 can (375 mL) 0 0
Gin g/100 g 0 0
g/30 mL 0 0
Rum g/100 g 0 0
g/30 mL 0 0
Vodka g/100 g 0 0
g/30 mL 0 0
Whiskey g/100 g 0 0
g/30 mL 0 0
Wine, red g/100 g 0 0
g/100 mL 0 0
Wine, white g/100 g 0 0
g/100 mL 0 0
Other food ‘Sugarless’ chewing gum g/100 g 41.9 0
products
g/2 gum strips (4 g) 1.7 0
Horseradish g/100 g 11.1 0.3
sauce (wasabi)
g/1 tsp (5 g) 0.6 0.0

Average serving sizes were obtained from FOODWORKS, version 6 (Xyris Software Australia Pty Ltd).
Shaded boxes indicate: not analysed or no data available.
Comparisons were made with previously published composition data taken from Muir et al. (2009)* and Biesiekierski et al. (2011)†, as well as the
USDA Nutrition Composition Database (US Department of Agriculture Agricultural Research Service, 2011).

ª 2013 The British Dietetic Association Ltd. 7

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

Table 2 Baseline characteristics of irritable bowel syndrome (IBS)

subjects and healthy controls

IBS patients Healthy controls

(n = 20) (n = 21) P-value

Ratio of men: 5 : 15 3 : 18 0.45†

women, n
Age (years)* 34 (27–47) 25 (22–48) 0.10‡
Alcohol intake 3 (0–7) 1 (0–2) 0.22‡
(g day–1)*
Regular NSAID 5 (25) 1 (5) 0.18†
use [n (%)]

NSAID, nonsteriodal anti-inflammatory drug. Figure 2 Proportion (%) of subjects with malabsorption of sorbitol or
*Values are expressed as the median (interquartile range). mannitol in patients with irritable bowel syndrome (IBS) (n = 20) and
†
Fisher’s exact test; ‡Mann–Whitney U-test. healthy controls (n = 21). Bars represent 95% confidence intervals.
*Significantly different (P = 0.02) compared to healthy controls.

the mean (95% CI) difference between AUC mannitol

and AUC glucose was 2519 (1410–3628) ppm 4 h–1. A hydrogen production was significantly lower in the IBS
strongly positive correlation between the hydrogen group compared to that in the healthy controls after sor-
production after sorbitol and mannitol ingestion was bitol (P = 0.02, unpaired t-test) and mannitol
also demonstrated (r = 0.78, P < 0.0001; Pearson’s (P = 0.003) but not after lactulose (P = 0.68).
correlation).
Conversely, the patterns of breath hydrogen production
Breath methane production
differed between sorbitol and mannitol in the patients
with IBS (Fig. 4b). Total breath hydrogen production In methane-producers, methane production was similar
after mannitol [601 (228) ppm 4 h–1] was significantly between challenges with sorbitol and mannitol in both
lower than that for sorbitol [1629 (210); P = 0.003, healthy controls [n = 6; AUC sorbitol: 595 (155)
paired t-test], although it was significantly higher in com- ppm 4 h–1 versus. mannitol: 476 (147) ppm 4 h–1; P =
parison than that after glucose [102 (42) ppm 4 h–1; 0.56, paired t-tests] and IBS patients [n = 6; sorbitol:
P ≤ 0.05]. The mean (95% CI) difference in AUC 538 (210) ppm 4 h–1 versus mannitol: 231 (76) ppm 4
between mannitol and glucose was 499 ( 10.2–1008) h–1; P = 0.09].
(r2 = 0.29). The mean (95% CI) difference was also sub-
stantial between AUC after sorbitol and that after glucose
Relationship between polyol malabsorption and
[1526 (1048–2004) ppm 4 h–1, r2 = 0.82]. Consistently,
gastrointestinal symptoms
no correlation was found between hydrogen production
in response to sorbitol or mannitol (r = 0.42, P = 0.07). Polyol-associated overall abdominal symptoms appeared
Within this same cohort of malabsorbers, total breath to be independent of the absorptive pattern of the

Figure 1 Comparison of 4-h overall gastrointestinal symptoms and composite scores of abdominal pain, wind and bloating, assessed by the
visual analogue scale (VAS) (Montalto et al.), after sorbitol, mannitol and glucose challenges in healthy (n = 18) and irritable bowel syndrome
(IBS) subjects (n = 20). Horizontal lines represent median VAS scores. Significant differences are shown: *P = 0.05 for overall symptoms after
sorbitol and **P ≤ 0.03 for overall symptoms after mannitol compared to glucose (Wilcoxon sign-ranked test). Significant results for IBS patients
compared to healthy controls are shown: †P < 0.01 and ‡P < 0.02.

8 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

polyols. Thus, three of 12 versus five of eight (P = 0.17,

Fisher’s exact) with or without malabsorption of sorbitol,
respectively, and three of four versus five of 16 (P = 0.25)
of mannitol developed abdominal symptoms. Figure 5
shows the association of overall abdominal symptom
scores with the semi-quantitative classification of extent
of malabsorption in patients with IBS. No significant
associations were found between the extent of malabsorp-
tion and symptom scores for either mannitol (r = 0.07,
P = 0.97, Spearman’s correlation) or sorbitol (r = 0.15,
P = 0.55).

Discussion
As 6-carbon isomers that are absorbed by passive diffu-
sion across the intestinal barrier, sorbitol and mannitol
might be anticipated to show similar absorption patterns
in any individual. Indeed, in agreement with a previous
study (Wursch et al., 1989), the absorption patterns in
healthy volunteers were similar; approximately 40% com-
Figure 3 Frequency (%) of subjects in the different categories of pletely absorbed each polyol and the patterns were con-
polyol malabsorption estimated using semi-quantitative analyses cordant in almost all of them. By contrast, the findings in
between patients with irritable bowel syndrome (IBS) (n = 20) and
a demographically-matched population of patients with
healthy controls (n = 21). *Statistically significant differences between
the proportion of patients with mannitol malabsorption and that in
IBS were very different. First, double the proportion
the healthy controls within each category (P = 0.02). (80%) completely absorbed mannitol. Second, the absorp-
tion patterns were discordant in that three out of five

(a)

(b)

Figure 4 Breath hydrogen response over 4 h to the

ingestion of sorbitol, mannitol and glucose in (a)
healthy (n = 15) and (b) irritable bowel syndrome
(IBS) subjects (n = 12) with sorbitol and/or mannitol
malabsorption. Values are plotted as the
mean (SEM).

ª 2013 The British Dietetic Association Ltd. 9

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

Figure 5 Severity of overall gastrointestinal symptoms according to visual analogue scale (VAS) in relation to the degree of malabsorption for
sorbitol, mannitol and glucose in patients with irritable bowel syndrome (IBS) (n = 20). No correlation was observed for sorbitol (Spearman’s
r = 0.15, P = 0.55) or mannitol (r = 0.07, P = 0.97).

patients had different patterns of absorption for mannitol Perhaps the most intriguing observation was the discor-
and sorbitol. Third, the ability to absorb both sorbitol dance between sorbitol and mannitol absorption in
and mannitol is enhanced in patients with IBS, as shown patients with IBS but not in healthy controls. As shown
by lower total hydrogen production in those with malab- by the 4-h hydrogen production, mannitol absorption was
sorption, despite similar hydrogen production to lactu- almost three-fold greater than that of sorbitol, a finding
lose. Finally, both polyols appeared to trigger symptoms that has not been observed previously. This is unlikely to
independently of its absorption pattern. be artefactual because breath hydrogen tests have good
Previously, studies using breath hydrogen (Wursch reproducibility (Welsh et al., 1981; Barillas-Mury & Solo-
et al., 1989) or recovery in ileostomy effluent (Saunders mons, 1987) and the order of testing was randomised.
& Wiggins, 1981; Beaugerie et al., 1997) have consistently Two reasons for the discordance are possible. First, bacte-
demonstrated at least partial absorption of sorbitol and rial fermentation of mannitol may have different kinetics
mannitol, with estimates suggesting that an average of to those for sorbitol in the IBS group. Sorbitol has slower
26% and 40% of 10-g doses of sorbitol and mannitol, and more complex bacterial fermentation than rapidly
respectively, are absorbed. The present study has similarly fermentable sugars such as lactose and fructose (Mishkin
found evidence for substantial absorption of both polyols et al., 1997; Clausen et al., 1998). In support of this, the
in healthy and IBS subjects. What has not been reported pattern of hydrogen production after mannitol over the 4-
previously is that the degree of absorption of sorbitol and h period was different to that for sorbitol in IBS patients.
mannitol was at least two-fold greater from the small A comparison of the kinetics of fermentation of sorbitol
intestine in patients with IBS than in healthy controls, as and mannitol has not been reported. Another likely expla-
shown by a significantly lower hydrogen production. This nation, however, is that, despite the structural similarities
observation was not a reflection of different functional that exist between the two polyols, the different orienta-
capabilities of the intestinal microbiota in the two groups tion of the hydroxyl position in each polyol may result in
because their hydrogen-producing capacity was similar dissimilarities in polyol–water interactions (Grigera, 1988)
after lactulose and there was no evidence of diversion of and, subsequently, their passage across the intestinal bar-
gas production towards a methane pathway, as we previ- rier. In water, mannitol adopts a planar configuration that
ously observed in when a low FODMAP diet is consumed attracts more water molecules and may enhance its per-
(Ong et al., 2010). The reason for the differences must lie meation via water-filled intracellular pores (Grigera,
in the physiology of the intestine. It might reflect slower 1988). It is also possible that the unstirred water layer,
oro-caecal transit, as defined by the first rise of breath suggested to regulate osmotic gradients for the passive
hydrogen after lactulose (Madsen et al., 2006). For mole- uptake of water-soluble molecules (Barry & Diamond,
cules that are slowly and passively absorbed, this would 1984), is a key player in differentiating permeation rate
give more time for absorption to occur. Alternatively, the between sorbitol and mannitol. This unstirred water layer
differences in absorption patterns between healthy and has also been shown to be altered in disease. Strocchi
IBS subjects might reflect epithelial abnormalities, which et al. (1996) demonstrated that patients with coeliac dis-
have been reported in some subgroups of patients with ease were shown to have increased jejunal unstirred water
IBS (Marshall et al., 2004; Dunlop et al., 2006). Further layer thickness compared to healthy controls, reducing
enquiry is warranted to provide an explanation for these nutrient absorption. Alterations in this layer in patients
observations. with IBS have yet to be examined but may potentially

10 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

explain the discordance in sorbitol and mannitol absorp- polyols can be achieved in patients with IBS. Fruits, par-
tion. ticularly in their dried forms such as prunes and dried
As reported previously (Fernandez-Banares et al., 1993), pears, are concentrated sources of sorbitol that may be
an increase in abdominal symptoms was induced by sorbi- the major problem FODMAP, instead of the much-
tol and mannitol over the duration of the breath testing in maligned fructose in such sources. Existing food compo-
patients with IBS but not in healthy controls. However, the sition data indicating that certain vegetables are relatively
induction of symptoms for both polyols was independent rich sources of mannitol were confirmed. Amounts
of whether the polyol was malabsorbed and fermented. approaching 10 g day 1 mannitol can be derived from
Over this short duration, another mechanism other than large intake of these vegetables, particularly celery, cauli-
fermentation-driven gaseous distension of the intestine flower, mushrooms and snow peas. However, it may not
must underlie symptom exacerbation. Fluid distension of require such a dose when consumed in combination with
the small intestine is proposed because the slowly-absorbed sorbitol or other FODMAPs, as discussed above. In
polyols can exert an osmotic effect around most of the comparison with published data from the USDA, the
small intestine, as elegantly shown by magnetic resonance quantities of sorbitol in these fruits were quite similar to
methodology (Marciani et al., 2010). Hence, the implica- those reported in the present study, whereas there were
tion for dietary therapy is that these polyols can trigger some variations with respect to the mannitol content of
symptoms whether they enter the large bowel to be subse- food. These differences can be explained by several fac-
quently fermented or not, as a result of their slow rate of tors, including seasonal and climate variations, fruit or
absorption and osmotically-mediated distension of the vegetable varieties analysed, as well as differences in the
small intestinal. Breath testing to assess polyol absorption, quantification methodologies used (Makinen & Soderling,
as is frequently performed for fructose and lactose, may not 1980; Wrolstad & Shallenberger, 1981).
then be clinically useful or relevant. Alternatively, mannitol ConclusionsThe present randomised, double-blinded,
and sorbitol absorption might directly affect the function placebo-controlled cross-over study has demonstrated dis-
of the enteric nervous system via local osmolytic effects or tinct differences in the small intestinal handling of sorbi-
the involvement of inflammatory mediators. Perhaps mast tol and mannitol between healthy and IBS individuals. By
cell-histamine release, suggested as an explanation for ana- contrast to healthy subjects who appear to absorb both
phylactoid reactions to mannitol in several case reports polyols to an almost similar extent, patients with IBS not
(Findlay et al., 1983; Eggleston et al., 1987; Hegde & Venk- only have a greater ability to absorb both polyols, but
atesh, 2004), may be involved. also absorb mannitol more readily than sorbitol. The
Such an effect on symptom induction reinforces atten- mechanisms by which such differences occur require clar-
tion to those polyols in the design of dietary therapy to ification. Despite superior absorption, both sorbitol and
alleviate IBS symptoms. Thus, sorbitol and mannitol are mannitol specifically induced abdominal symptoms, sug-
appropriately included in the class of poorly absorbed gesting that mechanism(s) other than (or additional to)
short-chain carbohydrates, termed FODMAPs, for which their fermentation are involved. Such observations suggest
an accumulating evidence-base is now available as major that the dietary restriction of these polyols should be
dietary triggers of functional gut symptoms (Gibson & included in dietary restriction of FODMAPs for patients
Shepherd, 2010). The principles of the low FODMAP diet with IBS, irrespective of their ability to absorb them
indicate that individual FODMAPs have additive effects according to breath hydrogen results. Comprehensive
on symptoms and should not be considered separately food knowledge on polyol content not only eases the
but rather together in dietary interventions. The results of design of such dietary therapy, but also provides prebiotic
the present study suggest not only that both sorbitol and food sources for asymptomatic individuals. Additionally,
mannitol should be restricted in patients with IBS, but the role of xylitol and other commonly-used polyols in
also that this decision should be independent of the dem- the food industry (Zumbe et al., 2001) as dietary triggers
onstration of their malabsorption via breath hydrogen in IBS also warrants further study because evidence from
testing. Whether acute challenges with sorbitol and man- the healthy population cannot be extrapolated to this
nitol serve as a better guide of gastrointestinal tolerance patient population.
for patients who may benefit from such restrictions, as
recently suggested for fructose (Wilder-Smith et al.,
2013), would require further study. Acknowledgments
Food composition data from the present study rein- We would also like to thank all laboratory research staff
force previously published data showing that the intake and breath testing staff at Box Hill Hospital for the help
of up to 10 g day 1 from foods with naturally-occurring and technical support received.

ª 2013 The British Dietetic Association Ltd. 11

 Sorbitol and mannitol absorption in IBS C. K. Yao et al.

Chey, W.Y., Jin, H.O., Lee, M.H., Sun, S.W. & Lee, K.Y.
Conflict of interests, source of funding and
(2001) Colonic motility abnormality in patients with
authorship irritable bowel syndrome exhibiting abdominal pain and
The authors declare that there are no conflicts of diarrhea. Am. J. Gastroenterol. 96, 1499–1506.
interest. Clausen, M.R., Jorgensen, J. & Mortensen, P.B. (1998)
No external funding was received for the present study. Comparison of diarrhea induced by ingestion of
JGM, JSB and PRG designed the study. KL and RR fructooligosaccharide idolax and disaccharide lactulose (role
conducted the food analysis. CKY, HLT and DRVL of osmolarity versus fermentation of malabsorbed
carried out the clinical study. CKY analysed and carbohydrate). Dig. Dis. Sci. 43, 2696–2707.
interpreted study data. CKY, JGM, JSB, DRVL and Dunlop, S., Hebden, J., Naesdal, J., Campbell, E., Olbe, L.,
PRG drafted and revised the manuscript for publication. Perkins, A.C. & Spiller, R.C. (2006) Abnormal intestinal
All authors approved the final manuscript submitted for permeability in subgroups of diarrhea-predominant irritable
publication. bowel syndromes. Am. J. Gastroenterol. 101, 1288–1294.
Eggleston, P.A., Kagey-Sobotka, A. & Lichtenstein, L.M.L.
(1987) A comparison of the osmotic activation of basophils
and human lung mast cells. Am. Rev. Respir. Dis. 135, 1043–
References 1048.
Fernandez-Banares, F., Esteve-Pardo, M., De Leon, R.,
Agrawal, A., Houghton, L.A., Lea, R., Morris, J., Reilly, B. & Humbert, P., Cabre, E., Llovet, J.M. & Gassull, M.A. (1993)
Whorwell, P.J. (2008) Bloating and distension in Irritable Sugar malabsorption in functional bowel disease: clinical
Bowel Syndrome: the role of visceral sensation. implications. Am. J. Gastroenterol. 88, 2044–2050.
Gastroenterology 134, 1882–1889. Findlay, S.I., Kagey-Sobotka, A. & Lichtenstein, L.M. (1983)
Barillas-Mury, C. & Solomons, N.W. (1987) Test-retest Mannitol in a patient with a mannitol-induced
reproducibility of hydrogen breath test for lactose anaphylactoid reaction. J. Aller. Clin. Immunol. 73, 578–583.
maldigestion in preschool children. J. Pediatr. Gastroenterol. Francis, C.Y., Morris, J. & Whorwell, P.J. (1997) The irritable
Nutr. 6, 281–285. bowel severity scoring system: a simple method of
Barrett, J.S. & Gibson, P.R. (2010) Development and monitoring irritable bowel syndrome and its progress.
validation of a comprehensive semi-quantitative food Aliment. Pharmacol. Ther. 11, 395–402.
frequency questionnaire that includes FODMAP intake and Gibson, P.R. & Shepherd, S.J. (2010) Evidence-based dietary
glycemic index. J. Am. Diet. Assoc. 110, 1469–1476. management of functional gastrointestinal symptoms: The
Barrett, J.S., Gearry, R.B., Muir, J.G., Irving, P.M., Rose, R., FODMAP approach. J. Gastroenterol. Hepatol. 25, 252–258.
Rosella, O., Haines, M.L., Shepherd, S.J. & Gibson, P.R. Grigera, J.R. (1988) Conformation of polyols in water.
(2010) Dietary poorly absorbed, short-chain J. Chem. Soc., Faraday Trans. 1, 2603–2608.
carbohydrates increase delivery of water and fermentable Hegde, V.L. & Venkatesh, Y.P. (2004) Anaphylaxis to excipient
substrates to the proximal colon. Aliment. Pharmacol. Ther. mannitol: evidence for an immunoglobulin E-mediated
31, 874–882. mechanism. Clin. Exp. Allergy 34, 1602–1609.
Barry, P.H. & Diamond, J.M. (1984) Effects of unstirred layers Hyams, J.S. (1983) Sorbitol intolerance: an unappreciated
on membrane phenomena. Physiol. Rev. 64, 763–872. cause of functional gastrointestinal complaints.
Bate, J.P., Irving, P.M., Barrett, J.S. & Gibson, P.R. (2010) Gastroenterology 84, 30–33.
Benefits of breath hydrogen testing after lactulose Kotler, D., Holt, P. & Rosensweig, N. (1982) Modification of the
administration in analysing carbohydrate malabsorption. breath hydrogen test: increased sensitivity for the detection of
Eur. J. Gastroenterol. Hepatol. 22, 318–326. carbohydrate malabsorption. J. Lab. Clin. Med. 100, 798–805.
Beaugerie, L., Flourie, B., Marteau, P., Pellier, P., Franchisseur, Krugliak, P., Hollander, D., Schlaepfer, C.C., Nguyen, H. &
C. & Rambaud, J.-C. (1990) Digestion and absorption in the Ma, T.Y. (1994) Mechanisms and sites of mannitol
human intestine of three sugar alcohols. Gastroenterology 99, permeability of small and large intestine in the rat. Dig. Dis.
717–723. Sci. 39, 796–801.
Beaugerie, L., Flourie, B., Pernet, P., Achour, L., Franchisseur, Le, A.S. & Mulderrig, K.B. (2001) Sorbitol and mannitol. In
C. & Rambaud, J.-C. (1997) Glucose does not facilitate the Alternative Sweeteners. eds L.O. Nabors & R.C. Gelardi. New
absorption of sorbitol perfused in situ in the human small York, NY: Marcel Dekker Inc.
intestine. J. Nutr. 127, 341–344. Lewis, D.H. & Smith, D.C. (1967) Sugar alcohols (polyols) in
Biesiekierski, J.R., Rosella, O., Rose, R., Liels, K., Barrett, J.S., fungi and green plants: distribution, physiology and
Shepherd, S.J., Gibson, P.R. & Muir, J.G. (2011) metabolism. New Phytol. 66, 143–184.
Quantification of fructans, galacto-oligosaccharides and Longstreth, G.F., Thompson, W.G., Chey, W.D., Houghton,
other short-chain carbohydrates in processed grains and L.A., Mearin, F. & Spiller, R.C. (2006) Functional bowel
cereals. J. Hum. Nutr. Dietet. 24, 154–176. disorders. Gastroenterology 130, 1480–1491.

12 ª 2013 The British Dietetic Association Ltd.

 C. K. Yao et al. Sorbitol and mannitol absorption in IBS

Madsen, J.L., Linnet, J. & Rumessen, J.J. (2006) Effect of Strocchi, A., Corazza, G., Furne, J., Fine, C., Di Sario, A.,
nonabsorbed amounts of a fructose–sorbitol mixture on Gasbarrini, G. & Levitt, M.D. (1996) Measurements of the
small intestinal transit in healthy volunteers. Dig. Dis. Sci. jejunal unstirred layer in normal subjects and patients with
51, 145–153. celiac disease. Am. J. Physiology 270, G487–G491.
Makinen, K. & Soderling, E. (1980) A quantitative study of Thompson, D.G., Binfield, P., De Belder, A., O’Brien, J.,
mannitol, sorbitol, xylitol and xylose in wild berries and Warren, S. & Wilson, M. (1985) Extra intestinal
commercial fruits. J. Food Sci. 45, 367–374. influences on exhaled breath hydrogen measurements
Marciani, L., Cox, E.F., Hoad, C.L., Pritchard, S., Totman, J.J., during the investigation of gastrointestinal disease. Gut 16,
Foley, S., Mistry, A., Evans, S., Gowland, P.A. & Spiller, R.C. 1349–1352.
(2010) Postprandial changes in small bowel water content in US Department of Agriculture, Agricultural Research Service.
healthy subjects and patients with irritable bowel syndrome. (2011) USDA National Nutrient Database for Standard
Gastroenterology 138, 469–477. Reference, Release 24. Available at: http://www.ars.usda.gov/
Marshall, J., Thabane, M., Garg, A., Clark, W., Meddings, J. & ba/bhnrc/ndl (accessed on 26 October 2011).
Collins, S. (2004) Intestinal permeability in patients with Welsh, J.D., Payne, D.L., Manion, C., Morrison, R.D. &
irritable bowel syndrome after a waterborne outbreak of Nichols, M.A. (1981) Interval sampling of breath hydrogen
acute gastroenteritis in Walkerton, Ontario. Aliment. (H2) as an index of lactose malabsorption in
Pharmacol. Ther. 20, 1317–1322. lactase-deficient subjects. Dig. Dis. Sci. 26, 681–685.
Mishkin, D., Sablauskas, L., Yalovsky, M. & Mishkin, S. (1997) Wilder-Smith, C.H., Materna, A., Wermelinger, C. & Schuler,
Fructose and sorbitol malabsorption in ambulatory patients J. (2013) Fructose and lactose intolerance and
with functional dyspepsia. Dig. Dis. Sci. 42, 2591–2598. malabsorption testing: the relationship with symptoms in
Muir, J.G., Rosemary, R., Rosella, O., Liels, K., Barrett, J.S., functional gastrointestinal disorders. Aliment. Pharmacol.
Shepherd, S.J. & Gibson, P.R. (2009) Measurement of Ther. 37, 1074–1083.
short-chain carbohydrates in common Australian vegetables Wrolstad, R.E. & Shallenberger, R.S. (1981) Free sugars and
and fruits by High-Performance Liquid Chromatography sorbitol in fruits – a compilation from the literature.
(HPLC). J. Agri. Food. Chem. 57, 554–565. J. Assoc. Off. Anal. Chem. 61, 91–103.
Nelis, G., Vermeeren, M. & Jansen, W. (1990) Role of Wursch, P., Koellreutter, B. & Schweizer, T.F. (1989)
fructose-sorbitol malabsorption in the irritable bowel Hydrogen excretion after ingestion of five different sugar
syndrome. Gastroenterology 99, 1016–1020. alcohols and lactulose. Eur. J. Clin. Nutr. 43, 819–825.
Ong, D.K., Mitchell, S.B.M., Barrett, J.S., Shepherd, S.J., Irving, Zumbe, A., Lee, A. & Storey, D. (2001) Polyols in
P.M., Biesiekierski, J.R., Smith, S., Gibson, P.R. & Muir, J.G. confectionery: the route to sugar-free, reduced sugar
(2010) Manipulation of dietary short chain carbohydrates and reduced calorie confectionery. Br. J. Nutr. 85,
alters the pattern of gas production and genesis of S31–S45.
symptoms in irritable bowel syndrome. J. Gastroenterol.
Hepatol. 25, 1366–1373.
Rumessen, J.J., Hamberg, O. & Gudmand-Hoyer, E. (1990) Supporting information
Interval sampling of end-expiratory hydrogen (H2)
concentrations to quantify carbohydrate malabsorption by Additional Supporting Information may be found in the
means of lactulose standards. Gut 31, 37–42. online version of this article:
Saunders, D.R. & Wiggins, H.S. (1981) Conservation of Figure S1. Isomeric configuration of sorbitol and man-
mannitol, lactulose and raffinose by the human colon. Am. nitol. Dotted boxes highlight the orientation of their
J. Physiol. 241, G397–G402. hydroxyl group.

ª 2013 The British Dietetic Association Ltd. 13