ORIGINAL ARTICLE

Fructose Intolerance in IBS and Utility

of Fructose-Restricted Diet
Young K. Choi, MD,* Nancy Kraft, RD, LD,w Bridget Zimmerman, PhD,z Michelle Jackson,w
and Satish S. C. Rao, MD, PhD, FRCPw

Introduction: Whether dietary fructose intolerance causes

symptoms of irritable bowel syndrome (IBS) is unclear. We
F ructose is naturally present in fruits and vegetables,
but today it is avidly consumed in beverages and
canned products as high fructose corn syrup.1 In the
examined the prevalence of fructose intolerance in IBS and long- United States, the consumption of fructose has increased
term outcome of fructose-restricted diet. several fold during the last 2 decades. A United States
Methods: Two hundred and nine patients with suspected IBS Department of Agriculture study estimated that the
were retrospectively evaluated for organic illnesses. Patients annual consumption of fructose has risen from less than
with IBS (Rome II) and positive fructose breath test received a ton in 1966 to 8.8 million tons in 2003.2 Unlike other
instructions regarding fructose-restricted diet. One year later, sugars such as sucrose or lactose which are digested by
their symptoms, compliance with, and effects of dietary sucrase or lactase enzymes produced by the intestinal
modification on lifestyle were assessed using a structured brush border, the gut does not appear to have a specific
interview. enzyme for digesting or transporting fructose.3,4 Recent
studies have shown that GLUT-5 and GLUT-2 which
Results: Eighty patients (m/f = 26/54) fulfilled Rome II criteria. actively transport glucose across the mucosa may also
Of 80 patients, 31 (38%) had positive breath test. Of 31 patients, play a role in facilitating absorption of fructose.4
26 (84%) participated in follow-up (mean = 13 mo) evaluation. Consequently, if fructose is ingested in large quantities,
Of 26 patients, 14 (53%) were compliant with diet; mean the capacity of the gut to absorb fructose can be easily
compliance = 71%. In this group, pain, belching, bloating, overwhelmed leading to fructose malabsorption and
fullness, indigestion, and diarrhea improved (P<0.02). Of 26 symptoms.5–9
patients, 12 (46%) were noncompliant, and their symptoms were Dietary fructose intolerance is associated with many
unchanged, except belching. The mean impact on lifestyle, common symptoms such as abdominal bloating, pain,
compliant versus noncompliant groups was 2.93 versus 2.57 flatulence, and diarrhea.5,8 These symptoms are also
(P>0.05). similar to those described by patients with lactose10 or
Conclusions: About one-third of patients with suspected IBS had sorbitol intolerance6,7,11 or bacterial overgrowth.12 Pre-
fructose intolerance. When compliant, symptoms improved on viously, we8 and others5 have reported a higher pre-
fructose-restricted diet despite moderate impact on lifestyle; valence of fructose intolerance in patients with
noncompliance was associated with persistent symptoms. unexplained gastrointestinal symptoms and dyspepsia.
Fructose intolerance is another jigsaw piece of the IBS puzzle In contrast, a controlled study showed that the frequency
that may respond to dietary modification. of gastrointestinal symptoms and the occurrence
of fructose malabsorption were similar between
Key Words: IBS, fructose intolerance, diet, breath test controls and patients with functional bowel disorder.13
(J Clin Gastroenterol 2008;42:233–238) Furthermore, the prevalence of fructose intolerance in
patients with irritable bowel syndrome (IBS) is not clearly
known.
IBS affects 10% to 15% of the population in the
Received for publication August 23, 2006; accepted October 20, 2006. United States.14 The prevalence of fructose intolerance
From the *Immanuel St Joseph’s, Mayo Health System, Mankato, MN; in patients with suspected IBS has not been systematically
wDepartment of Internal Medicine; and zClinical Research Center, examined. Furthermore, whether IBS patients with
University of Iowa Carver College of Medicine, Iowa City, IA. fructose intolerance benefit from a fructose-restricted diet
The authors declare no conflict of interest.
Supported in part by NIH RO1 grant DK 57100-05 and grant RR00059 is not known.
from the General Clinical Research Centers program, National The aims of our study were to examine: (1) the
Center for Research Resources. prevalence of fructose intolerance in patients with
Reprints: Satish S.C. Rao, MD, PhD, FRCP, Department of Internal suspected IBS and to assess their symptom profiles and
Medicine, University of Iowa Hospitals and Clinics, 200 Hawkins
Drive/4612 JCP, Iowa City, IA 52242-1009 (e-mail: satish-rao@ (2) the effects of fructose-restricted diet on symptom
uiowa.edu). patterns and lifestyle in patients with IBS and fructose
Copyright r 2008 by Lippincott Williams & Wilkins intolerance.

J Clin Gastroenterol  Volume 42, Number 3, March 2008 233

Choi et al J Clin Gastroenterol  Volume 42, Number 3, March 2008

METHODS From these data, a mean score was calculated for each
Patients referred to our tertiary care center with symptom.
suspected IBS or with persistent, unexplained, nonspecific Patients with a positive fructose breath test received
gastrointestinal complaints between January 2001 and both written and verbal dietary instructions by a dietician
January 2002 were evaluated with appropriate diagnostic regarding a fructose exclusion or restricted diet. The
tests to identify organic illnesses including fructose breath written instructions comprised of a fructose-restriction
test. Subsequently, case records were reviewed to identify diet manual that was developed by our dietician. This
those patients who fulfilled the Rome II criteria for IBS patient-directed, self-guide informs the patient about the
and/or functional abdominal bloating,15 together with an fructose content of common foods and how to avoid food
absence of alarm symptoms. In particular, we excluded items such as fruit juice, cola products, carbonated
patients with abnormal findings on a barium study, beverages, corn products, chocolates, and foods contain-
computed tomography/ultrasound scan of abdomen, ing high fructose corn syrup. One year later, fructose
upper or lower gastrointestinal endoscopy, hematologic intolerant patients were invited to participate in a follow-
or biochemical studies or stool tests. Patients were also up telephone survey. Through a structured interview
excluded if they had any coexisting active or inactive conducted by one of the investigators (Y.C.), we enquired
gastrointestinal problems such as previous abdominal about their current symptoms, their compliance with the
surgery (except appendectomy or hysterectomy), lactose fructose-restricted diet, and the effects of dietary restric-
intolerance (negative lactose breath test), bacterial over- tion on their lifestyle. The patient’s symptoms at 1 year
growth (negative glucose breath test), peptic ulcer disease, were compared with those obtained at baseline. The
gastroparesis, dumping syndrome, pancreatic disorders, compliance with a fructose-restricted diet was designated
biliary disease, liver disease, any malignancy, celiac as compliant, if the patient reported that they had
disease, inflammatory bowel disease, gastrointestinal modified their diet substantially to reduce consumption
reflux disease, eosinophilic gastritis, or lymphocytic of fructose products by Z50% of the amount consumed
colitis. before testing and as noncompliant if it was <50%. The
The following protocol was used to test dietary effect of dietary restriction on lifestyle was scored on a
fructose intolerance. Patients were asked to refrain from modified Likert-like scale as follows: 0 (none), 1 (mini-
taking high fat, lactose, or fructose containing foods for 1 mal), 2 (mild), 3 (moderate), 4 (significant), and 5
day before the test. Patients were also asked to fast from (extreme).
midnight and refrain from smoking. After arrival in the The extra amount of time spent per week to comply
motility laboratory, the patient was instructed to blow with the diet was assessed during the follow-up interview
into a modified Haldane-Priestley bag (QuinTron, Mil- using the following scale: 1 = no extra time spent per
waulkee, WI), and an end-expiratory breath sample was week; 2 = spending 1 extra hour; 3 = spending 3 extra
collected. A 50 cm3 sample of air was taken from the bag hours; 4 = spending 4 extra hours; and 5 = spending
and injected into a gas chromatography analyzer (Quin- more than 5 extra hours.
tron Microlyzer Self Correcting Model SC, QuinTron, To assess the test-retest reliability of the bowel
Milwaukee, WI), and baseline values for hydrogen (H2) symptom questionnaire, we administered the symptom
and methane (CH4) were measured. These values were questionnaire twice, at 1-week interval to 25 additional
corrected for CO2. Next, the patients were asked to drink participants that included 8 healthy volunteers and 17
a solution containing 25 g of fructose dissolved in 250 mL patients who were undergoing breath tests for either
of water (10% solution). This dose was chosen on dose research or clinical purposes. During this period of
response studies of fructose absorption in healthy hu- assessment, these participants were advised to continue
mans.16 Thereafter, breath samples were collected at 30- with their usual diet and were unaware of the results of
minute intervals for 5 hours, and analyzed for H2 and the breath test.
CH4. During the test, any symptoms experienced by the The study was approved by the University of Iowa
patient were recorded. An incremental rise in breath H2 Institutional Review Board.
and/or CH4 of Z5 ppm in 3 or more consecutive breath
samples, over and above the baseline value or a value Statistics
Z20 ppm above the baseline value in 2 consecutive The differences in symptom profiles in IBS patients
samples was interpreted as a positive breath test.8,16,17 We who tested positive or negative to the fructose breath test
used a stricter definition of incremental rise over 2 or was compared using the Wilcoxon signed-rank test. The
more samples to minimize false positive test.16,17 symptom scores reported by our patients during the
Before the test, patients were asked to fill out a fructose breath test were compared using the student t
symptom questionnaire that assessed the presence of 9 test. After the fructose-restricted diet, the difference in
common bowel symptoms8; abdominal discomfort or symptom profiles between the compliant and the non-
pain, belching, bloating, postprandial fullness, indiges- compliant groups was compared, using the Wilcoxon
tion, nausea, diarrhea, vomiting, and flatulence. The signed-rank test.
severity of each symptom was rated in terms of frequency, The bowel symptom questionnaire was validated
duration, and intensity on a scale of 0-3, and the using the weighted k statistic. For each symptom, the
maximum possible total core for any 1 symptom was 9. individual scores for frequency, intensity, and duration

234 r 2008 Lippincott Williams & Wilkins

J Clin Gastroenterol  Volume 42, Number 3, March 2008 IBS, Diet, and Fructose Intolerance

TABLE 1. The Distribution of Symptom Scores, and the Proportion of Patients Who did not Report a Particular Symptom During
Their First and Second Symptom Assessment
Observed Score Range = Min-Max
Freq (%) With no Symptom Distribution of Score Difference
Symptom First Study Second Study Equal Differ by 1 Differ by 2 or Greater Weighted j (95% CI)
Abdominal pain 0-9 0-9 11 6 6 0.68
7 (30.4%) 5 (26.1%) (Max diff = 6) (0.49, 0.87)
Belching 0-9 0-9 11 7 5 0.68
8 (34.8%) 6 (26.1%) (Max diff = 5) (0.49, 0.86)
Bloating 0-9 0-9 13 5 2 0.78
6 (26.1%) 7 (30.4%) (Max diff = 6) (0.63, 0.93)
Diarrhea 0-9 0-8 13 2 8 0.51
15 (65.2%) 13 (56.5%) (Max diff = 7) (0.24, 0.78)
Flatulence 0-9 0-9 10 3 10 0.49
5 (21.7%) 7 (30.4%) (Max diff = 7) (0.24, 0.74)
Fullness 0-9 0-9 6 8 9 0.40
7 (30.4%) 7 (30.4%) (Max diff = 7) (0.13, 0.68)
Indigestion 0-9 0-9 13 2 8 0.61
9 (39.1%) 11 (47.8%) (Max diff = 6) (0.39, 0.84)
Nausea 0-9 0-9 16 2 5 0.81
13 (56.5%) 13 (56.5%) (Max diff = 3) (0.71, 0.91)
Vomiting 0-4 0-4 22 0 1 0.82
21 (91.3%) 21 (91.3%) (Max diff = 2) (0.50, 1.0)

were summated to develop a total symptom score and Reproducibility of Symptoms During
this was compared. We assessed the agreement for the the Breath Test
patient’s total symptom score and for each of the 9 Among those who tested positive (fructose intoler-
symptoms. ant), 28/31 (91%) patients reported that the breath test
reproduced their typical symptom(s) such as bloating,
diarrhea, gas, or abdominal pain (Fig. 1). Similar
RESULTS
symptoms were also reproduced in 40% of the patients
Demographics who had a negative breath test (fructose tolerant).
Two hundred and nine patients with suspected IBS Fructose intolerant patients were more likely
or unexplained gastrointestinal symptoms were evaluated. (P = 0.006) to experience symptoms during the breath
Of these, 129 patients were excluded, because of organic test when compared with those who were fructose
gastrointestinal illness or coexisting disorder(s), and 80 tolerant (Fig. 1).
patients who fulfilled the Rome II criteria for IBS and
functional abdominal bloating were included. Data from Symptom Profiles
these 80 patients [m/f = 26/54; mean age of 42 y All 80 suspected IBS patients had reported more
(range = 20 to 76)] were analyzed. than 1 gastrointestinal symptom. At baseline, irrespective
of whether they tested positive or negative with the
Validation of Bowel Symptom Questionnaire fructose breath test, the prevalence of symptoms was not
Twenty-five patients (m/f = 6/19, mean age = 39 y) statistically different, except for diarrhea. For example,
completed the bowel symptom questionnaire on 2 the mean total symptom score for abdominal pain was 7.0
separate occasions to assess the test-retest reliability.
Data obtained for each of the 9 individual symptoms are
shown in Table 1. The weighed k statistic with 95% CI
varied from 0.4 (0.13-0.68) to 0.82 (0.5-1.0). Overall
reproducibility of the questionnaire was fairly good
(Table 1).

Fructose Breath Test

Among the 80 patients with suspected IBS, 31
(33%) patients had a positive breath test (fructose
intolerant), and 49 (67%) had a negative breath test
(fructose tolerant). Among the 31 fructose intolerant
patients, 28 (90%) had elevated breath H2, 2 (7%) had
elevated H2 and CH4, and 1 other patient (3%) had FIGURE 1. Reproducibility of symptoms during the fructose
elevated CH4 only. breath test in patients with and without fructose intolerance.

r 2008 Lippincott Williams & Wilkins 235

Choi et al J Clin Gastroenterol  Volume 42, Number 3, March 2008

versus 6.42, bloating was 7.0 versus 6.33, and flatulence compliant group reported a temporal correlation between
was 6.07 versus 6.83 in the fructose intolerant patients a trial of consuming fructose containing foods and relapse
versus fructose tolerant patients (P>0.05). However, of symptoms. None of the noncompliant patients
loose stools/diarrhea was reported by all patients who reported such a correlation.
were fructose intolerant when compared with 35/49
(71%) patients who were fructose tolerant (P = 0.007). Effects on Lifestyle
The effect of fructose-restricted diet on lifestyle in
Effect of Fructose-restricted Diet on the dietary compliant group ranged from 1 to 3.5, with a
Bowel Symptoms mean value of 2.93. This suggests a mild to moderate
After 1 year, 26/31 fructose intolerant patients were effect on their lifestyle. In this group, the extra time spent
available and agreed to participate in a follow-up by the patients per week to comply with the fructose-
evaluation. Among these, 14/26 patients were designated restricted diet ranged from 1 to 3, with a mean value of
as complaint, with a mean self-estimate of 71% (range 2.57. Despite these effects on lifestyle, all of the compliant
50% to 90%) compliance with a fructose-restricted diet. patients reported that they were planning to continue
The mean duration of dietary restriction was 12.6 months with their dietary modification. The effect of fructose-
(range 6 to 18 mo). restricted diet among the noncompliant group was
In this group, there was significant improvement of reported as 1, suggesting that it did not alter their
symptoms, notably abdominal pain, belching, bloating, lifestyle.
fullness, indigestion, and diarrhea (Fig. 2A). In contrast,
12/26 patients were designated as noncompliant with a DISCUSSION
mean self-estimate of 23.4% (range 0% to 30%) In this series of patients with suspected IBS, we
compliance with the diet. In this group, there was no found that approximately one-third had a positive
significant change in bowel symptoms, except for belching fructose breath test. In over 90% of these patients, the
and nausea (Fig. 2B). Additionally, 12/16 patients in the breath test also reproduced their symptoms suggesting

FIGURE 2. A, Changes in symptom patterns in patients with fructose intolerance who complied with a fructose-restricted diet.
B, Changes in symptom patterns in patients with fructose intolerance who did not comply with a fructose-restricted diet.

236 r 2008 Lippincott Williams & Wilkins

J Clin Gastroenterol  Volume 42, Number 3, March 2008 IBS, Diet, and Fructose Intolerance

dietary fructose intolerance. Thus, it seems that in this ratio, most individuals who consume these products may
selected group of patients presenting to a tertiary care not be symptomatic.9 However, many food products have
center with symptoms suspicious of IBS, a significant excess fructose,20,21 and when these are ingested in
proportion had fructose intolerance. significant amounts, fructose intolerance is likely to
At baseline, patients with and without fructose ensue. In such individuals, our data show that dietary
intolerance appeared to have similar symptom profiles. fructose restriction could be effective in ameliorating
Thus, symptoms alone were not sufficiently reliable to symptoms.
identify patients with fructose intolerance. However, one However, the adherence to dietary modification was
interesting observation was that all patients who reported not easy and had some impact on the lifestyle of these
loose stools/diarrhea at baseline had a positive fructose patients. We found that patients who were complaint with
breath test. Also, the pretest incidence of loose stools/ the diet reported mild to moderate effects on their
diarrhea was significantly higher in those with a positive lifestyle. Furthermore, they had to spend extra time, on
breath test compared to those with a negative test. This average 2 to 3 hours per week, to maintain their dietary
suggests that patients presenting with diarrhea-predomi- restriction. Despite these effects, when inquired, all of our
nant IBS are more likely than others to have dietary compliant patients reported that they were willing to
fructose intolerance. Many patients with a negative continue with dietary restriction, because of the overall
fructose breath test also reported IBS symptoms such as benefit.
bloating and abdominal pain. These patients may have We designated all individuals with a positive
visceral hypersensitivity to physiologic stimuli.18,19 fructose breath test as having fructose intolerance and
To our knowledge, there is no standard question- provided them with dietary advice. Only 9% of our
naire for assessing symptoms in patients with suspected subjects did not develop symptoms during the breath test.
carbohydrate intolerance or bacterial overgrowth. We Thus, a possible overestimation of fructose intolerance
designed a simple questionnaire to assess the presence of in our patients may only apply to a small group of
bowel symptoms related to carbohydrate malabsorption individuals.
and its intensity and duration. Because this questionnaire Our study has several limitations that include the
formed an integral part of our analysis, we also examined small sample size, the referral bias, the lack of blinding, a
the test-retest reliability in a separate group of patients. possible recall bias for dietary restriction, the study of
We found that the results were quite reproducible; hence, patients from a tertiary care center, and the retrospective
this questionnaire may be useful for the routine assess- analysis of our data. Also, whether the noncompliant
ment of symptoms at baseline and after therapeutic patients did not try or the diet was ineffective cannot be
interventions. It also provides qualitative and semiquan- established. Nevertheless, these findings reveal an asso-
titative information regarding the occurrence of symp- ciation between dietary fructose intolerance and IBS
toms during the breath test. symptoms and that dietary intervention may be helpful in
One potential explanation for the elevations in these patients. These observations merit further confirma-
breath hydrogen and/or methane after ingestion of tion in a larger, prospective study.
fructose could be the presence of small bowel bacterial In conclusion, our study shows that about one-third
overgrowth. However, all patients included in this study of patients with suspected IBS may have fructose
had a negative lactose and a negative glucose breath test intolerance. Fructose breath test may identify this subset
making it less likely that they had proximal small bowel of treatable IBS patients. Without investigating this
bacterial overgrowth. Whether some of our patients had possibility, many such patients, particularly those with
distal small bowel bacterial overgrowth12 cannot be diarrhea-predominant illness may be mistakenly labeled
excluded from this study. as having IBS. Fructose-restricted diet imposes some
Although, we and others5–8 have reported dietary changes in lifestyle, but when adhered to may confer
fructose intolerance as a potential cause of unexplained significant relief of symptoms.
gastrointestinal symptoms, the efficacy of fructose-re-
stricted diet in the management of these patients is not ACKNOWLEDGMENT
known. In this study, we found that 54% of suspected The authors thank the secretarial assistance of
IBS patients were mostly compliant with a fructose- Mrs Heidi Vekemans.
restricted diet and showed significant improvement in
their bowel symptoms. In contrast, 46% were not
compliant with the diet and showed very little improve- REFERENCES
1. Position of American Dietetic Association: use of nutritive and non-
ment. These observations establish a direct correlation nutritive sweetners. J Am Diet Assoc. 2004;104:255–275.
between the consumption of fructose-containing products 2. Economic Research Service, USDA. Table 49–US total estimated
and the occurrence of symptoms that mimic IBS. If this deliveries of caloric sweeteners for domestic food and beverage use,
association is not recognized and treated, it is likely that by calendar year. 2003.
many of these patients may be misdiagnosed. 3. Riby JE, Fujisawa T, Kretchmer N. Fructose absorption. Am J Clin
Nutr. 1993;58:748–753.
Because fructose is often consumed with glucose 4. Wasserman D, Hoekstra JH, Tolia V, et al. Molecular analysis of
either in the form of fruits or certain forms of high the fructose transporter gene (GLUT 5) in isolated fructose
fructose corn syrup that has an excess glucose/fructose malabsorption. J Clin Invest. 1996;98:2398–2402.

r 2008 Lippincott Williams & Wilkins 237

Choi et al J Clin Gastroenterol  Volume 42, Number 3, March 2008

5. Rumessen JJ, Gudmand-Hoyer E. Functional bowel disease: 13. Fernandez-Banares F, Esteve-Pardo M, de Leon R, et al. Sugar
malabsorption and abdominal distress after ingestion of fructose, malabsorption in functional bowel disease: clinical implications.
sorbitol, and fructose-sorbitol mixtures. Gastroenterology. 1988;95: Am J Gastroenterol. 1993;88:2044–2050.
694–700. 14. Drossman DA, Whitehead WE, Camilleri M. Irritable bowel
6. Mishkin D, Sablauskas L, Yalovsky M, et al. Fructose and sorbitol syndrome: a technical review for practice guideline development.
malabsorption in ambulatory patients with functional dyspepsia. Gastroenterology. 1997;112:2120–2137.
Comparison with lactose maldigestion/malabsorption. Dig Dis Sci. 15. Thompson WG. Functional bowel disorders and functional
1997;42:2591–2598. abdominal pain. Gut. 1999;45:II43–II47.
7. Rumessen JJ. Fructose and related food carbohydrates. Sources, 16. Rao SSC, Attaluri A, Anderson L, et al. Ability of the normal
intake, absorption, and clinical implications. Scand J Gastroenterol. human small intestine to absorb fructose: evaluation of breath
1992;27:819–828. testing. Clin Gastroenterol Hepatol. 2007. In press.
8. Choi YK, Johlin FC, Summers RW, et al. Fructose intolerance: an 17. Read NW, Al-Janabi MN, Bates TE, et al. Interpretation of
under-recognized problem. Am J Gastroenterol. 2003;98:1348–1353. the breath hydrogen profile obtained after ingesting a solid
9. Skoog SM, Bharucha AE. Dietary fructose and gastrointestinal meal containing unabsorbable carbohydrate. Gut. 1985;26:834–842.
symptoms: a review. Am J Gastroenterol. 2004;99:2046–2050. 18. Serra J, Azpiroz F, Malagelada JR. Impaired transit and tolerance
10. Suarez FL, Savaiano DA, Levitt L. Comparison of symptoms after of intestinal gas in the irritable bowel syndrome. Gut. 2001;48:14–19.
the consumption of milk or lactose-hydrolyzed milk by people with 19. Bouin M. Intolerance to visceral distension in functional dyspepsia
self-reported severe lactose intolerance. N Engl J Med. 1995;333:1–4. or irritable bowel syndrome: an organ specific defect or a pan
11. Hyams JS. Sorbitol intolerance: an unappreciated cause of functional intestinal dysregulation? Neurogastroenterol Motil. 2004;16:311–314.
gastrointestinal complaints. Gastroenterology. 1983;84:30–33. 20. Hanover LM, White JS. Manufacturing, composition, and applica-
12. Pimentel M, Chow EJ, Lin HC. Eradication of small intestinal tions of fructose. Am J Clin Nutr. 1993;58:724S–732S.
bacterial overgrowth reduces symptoms of irritable bowel syndrome. 21. Park YK, Yetley EA. Intakes and food sources of fructose in the
Am J Gastroenterol. 2000;95:3503–3506. United States. Am J Clin Nutr. 1993;58:737S–747S.

238 r 2008 Lippincott Williams & Wilkins