Secondary Logo

Altered Intestinal Permeability to Mannitol in Diabetes Mellitus Type I

Carratù, Romano; Secondulfo, Mario; de Magistris, Laura; Iafusco, Dario*; Urio, Anna*; Carbone, Maria Grazia*; Pontoni, Gabriele; Cartenì, Marilena; Prisco, Francesco*

Journal of Pediatric Gastroenterology & Nutrition: March 1999 - Volume 28 - Issue 3 - p 264-269
Original Articles

Background: Intestinal permeability has seldom been investigated in diabetes mellitus, even though patients frequently report gastrointestinal symptoms, and it has recently been shown that the prevalence of celiac disease associated with diabetes mellitus is higher than expected.

Methods: Intestinal permeability to cellobiose and mannitol was investigated in 31 patients affected by type I uncomplicated diabetes mellitus. Values were compared with those obtained in 32 normal subjects.

Results: The percentage of mannitol recovery was far higher than normal in two thirds of the investigated patients and correlated with the length of disease, even though the probes' ratio (cellobiose/mannitol) was in the normal range.

Conclusions: A not previously reported increase of intestinal permeability to mannitol, clear-cut and not associated with that of the larger probe, is found in type I uncomplicated diabetes mellitus. These results may describe a primary feature of type I diabetes mellitus and the initial steps of evolution of celiac disease.

Gastroenterology Unit, *Department of Pediatrics, and †Macromolecular Biochemical Institute, Seconda Università di Napoli, Napoli, Italy

Received October 16, 1997; revised July 13, 1998; accepted September 30, 1998.

Address correspondence and reprint requests to Romano Carratù, MD, Professor of Gastroenterology, Gastroenterology Unit, II Università degli Studi di Napoli, Piazza Miraglia 1, 80138 Napoli, Italy.

Gastrointestinal symptoms in diabetes mellitus are generally ascribed to altered intestinal motility (1) subsequent to autonomic neuropathy, which is well known to be found in 15% of patients with diabetes (2). Anatomic changes of gastrointestinal mucosa have been observed only in cases of diabetes mellitus with celiac disease (approximately 4-5% of patients with diabetes mellitus) (3), although a direct effect of reduced blood flow caused by diabetic microangiopathy or macroangiopathy would be expected (4).

In previous studies of absorptive function of the small intestine in diabetes, either the absorption of glucose and related molecules or the effects of disaccharides such as lactulose on metabolism of carbohydrates were investigated (5-8). Studies in diabetic patients of mucosal permeability to sugars, in which the differential passive permeability of monosaccharides and disaccharides through the intestinal mucosa was examined, have been rare (9-10). Most of these studies concern adult patients affected by non-insulin-dependent diabetes mellitus (diabetes mellitus type II).

Insulin-dependent diabetes mellitus (IDDM; type I) is an autoimmune condition, sometimes associated with diseases that are characterized by marked immunologic features, such as celiac disease and thyroiditis (3,11). In patients with IDDM, prevalence of subclinical forms of celiac disease, unexpectedly ranging from 1.0% to 7.8%, has been reported (12) in contrast with its prevalence in type II diabetes mellitus. It has therefore been suggested (12) that all patients with IDDM should be screened for celiac disease by means of immunoglobulin (Ig) A endomysial antibodies.

The purpose of the present investigation was to search in children, adolescents, and young adults with uncomplicated IDDM for alterations of intestinal permeability to two different sized sugar probes, cellobiose and mannitol, that have different permeability pathways and are easily and reliably measured in urine (13-24). Unless stated differently, values are mean ± standard error (SE).

Back to Top | Article Outline



Thirty-one consecutive children, adolescents, and young adults with a known history of IDDM entered the study. Sex and age allocation were: 15 girls, 16 boys; mean age, 13.6 ± 3.8 (Table 1). The mean duration of IDDM was 6.1 ± 3.2 years (range, 1-15 years). All patients were maintained with subcutaneous insulin therapy with a mean dosage of 0.96 ± 0.11 U/kg per day. All patients were studied when the metabolic disease was uncomplicated, although 11 of them showed glycosuria. During the year preceding the admission, mean glycosylated hemoglobin value were 8.7 ± 1.7% (normal, 4.4-6.2%). Microalbuminuria, measured in night urine by means of an immunoturbidimetric method (DCA 2000, Bayer Diagnostics, Leverkusen, Germany), was in the normal range (6.38 ± 10.38 µg/m2 per minute; normal, <20 µg/m2 per minute); blood urea and creatinine levels were also normal. No signs of retinopathy were detected. A few associated pathologic conditions were found at admission: secondary amenorrhea in two patients (Table 1, LF and SaF), thyroiditis in three (Table 1, SoF, DGA, and IR), rheumatoid arthritis in one (Table 1, BC), and multiple sclerosis in another (Table 1, DRC).



On admission to the study, all patients were screened for the presence of celiac disease through serologic analysis of anti-gliadin antibodies (by enzyme-linked immunosorbent assay; Eurospital, Naples Italy) and anti-endomysial antibodies (by immunofluorescence, Eurospital), and results in all were negative. During the 1-year follow-up in two patients (Table 1, GF and IC) a subclinical form of celiac disease was diagnosed based on the presence of anti-gliadin and anti-endomysial antibodies, confirmed by mucosal atrophy at biopsy.

All the patients had normal growth; none had gastrointestinal symptoms or signs of autonomic neuropathy when assessed by postural blood pressure, alteration of the normal heart rate beat in response to breathing or the Valsalva maneuver, presence of impotence, or atonic bladder. The patients had not had nonsteroidal anti-inflammatory or antibiotic therapy in the previous months, and none, even the oldest ones, used alcohol or tobacco.

Back to Top | Article Outline

Control Groups

To define normal values of the cellobiose-mannitol (CE/MA) permeability test (13-24) in our laboratory, 12 healthy children were compared with the patients. Sex and age allocation: 7 girls, 5 boys; mean age, 13.08 ± 2.88 years. Twenty healthy adult volunteers were also assessed: 11 women, 9 men; mean age, 40.0 ± 6.8.

Back to Top | Article Outline

Cellobiose-Mannitol Permeability Test

After an overnight fast (6-8 hours), subjects ingested the test sugar solution composed of 5 g cellobiose with 2 g mannitol (C7252 and M4125; Sigma, St. Louis, MO, U.S.A.) in 100 ml of water. The solution had an osmolality of approximately 270 mOsm. After 1 hour, subjects were encouraged to drink water to promote diuresis. All urine passes within 5 hours was collected, the volume measured, and two aliquots immediately stored at -25°C. The subsequent analysis was performed during the next 10 days.

Mannitol in urine was determined according to the method of Corcoran and Page (25). Briefly, mannitol was oxidized to formaldehyde by periodic acid, and the excess was removed by stannous chloride. The formaldehyde was reacted with chromotropic acid to form a purple complex, the absorbance of which was measured at 570 nm. Cellobiose in urine was determined in accordance with Strobel et al. (13) by a double-step enzymatic reaction: first, by means of β-glucosidase (G0395; Sigma), which determines hydrolysis of cellobiose to glucose, and then by means of an hexokinase-coupled reaction (glucose HK/16-50; Sigma), with final detection of reduced nicotinamide adenine dinucleotide phosphate (NADPH) at 340 nm. The percentage of urinary recovery in 5 hours was calculated for cellobiose and mannitol as was the ratio between the percentage recovery of cellobiose and the percentage recovery of mannitol (CE/MA).

Urine specimens collected for cellobiose and mannitol analysis were also analyzed to exclude the presence of glucose, aldehydes and ketone bodies by means of the strip test: Ketodiaburtest (Boehringer, Mannheim, Germany).

Back to Top | Article Outline

Nuclear Magnetic Resonance

The method for mannitol detected (25) is based on a rather unspecific reaction. Possible shortcomings of the method were checked by submitting urine samples to nuclear magnetic resonance (NMR) analysis. This is a powerful technique because of its potential for detecting essentially all hydrogen-containing metabolites in (and often below) the 100-µM range in biologic fluids (26). Several inherited diseases have already been identified by NMR by using for diagnosis the typical fingerprint patterns of the proton NMR spectra of both urine and serum (27).

Urine specimens (300 µl) were brought to pH 2.5 ± 0.03 (to minimize possible slight pH-dependent changes in chemical shift) by means of an automatic titration unit (Radiometer, Copenhagen, Denmark) equipped with a 2.5-ml automatically filled buret with 3 M HCl. The sample was then mixed with 100 µl of deuterated water, to which a known amount of perdeuterated sodium trimethylsylilpropionate (TSP) was added. The obtained solution was then inserted into a 5-mm NMR tube and analyzed in a 4.7-Tesla NMR spectrometer (model 200 AC-E; Bruker, Billerica, MA, U.S.A.) with 200 mHz proton resonance. Th TSP was used as a qualitative standard for the chemical shift scale and as a quantitative external standard for peak are calculations, according to the method of Tofts and Wray (28); 128 3.15-second scans were acquired. All other NMR experimental details were according to Lennert and Hunkler (27), with minor modifications. Five urine samples with the highest mannitol concentrations were analyzed by NMR, compared with the administered mannitol (Sigma M4125), as an internal standard (Figs 1 and 2).

FIG. 1

FIG. 1

FIG. 2

FIG. 2

Back to Top | Article Outline

Statistical Analysis

Changes, if any, in intestinal permeability to cellobiose and mannitol and in the CE/MA ratio are presented as mean ± SE and/or median (range) of the percentage recoveries of each probe molecule in the 5-hour urine collection. The normal ranges of the two control groups were calculated in the same way, and the two values were considered separately in the calculations. To define the normal upper limit of CE/MA, the mean + 2 standard deviations (SD) of cumulative control groups was calculated. Differences between mean values of control (pediatric and adult) and IDDM patient groups were assessed using the nonparametric Mann-Whitney test.

Correlation of NMR in contrast with the Corcoran and Page (25) methods for mannitol analysis was achieved by means of the linear regression analysis (r correlation factor). The r correlation factor was also applied in the comparison of percentage of CE, percentage of MA, and CE/MA with duration of disease, glycosylated hemoglobin, and amount of insulin therapy.

Back to Top | Article Outline


Informed consent was obtained from the patients or children's parents before submitting to the CE/MA test. The test is noninvasive and was well tolerated, as expected, by all patients and control subjects.

Back to Top | Article Outline


All values are express as mean ± standard error. The orally administered sugars were well tolerated by patients and control subjects; there were no symptoms suggestive of lactose intolerance. All urine specimens were free of aldehydes and ketone bodies.

Back to Top | Article Outline

Cellobiose/Mannitol Test

All the enrolled subjects completed the test, and the individual percentages of recovered molecular probes are shown in Table 1. In 12 healthy children and 20 adult healthy volunteers the CE/MA ratio was 0.016 ± 0.010 and 0.009 ± 0.002, respectively. To define an upper limit of normal range for the data, the mean ± SE for our 32 controls was calculated (0.008 ± 0.010), and values higher than 0.028 (i.e., mean + 2 SD) were considered significantly changed.

In 11 patients with IDDM, glycosuria was present in a range of 0.25-0.50 g/l; therefore, urinary cellobiose was measured by the enzymatic assay in the remaining 20 patients. The median (range) percentages of cellobiose recovered in these 20 patients was 0.0% (range, 0-3.6%) and did not differ significantly from the levels in either group of control subjects.

The mean percentage of mannitol recovered in all 31 children with IDDM was far higher (50.15 ± 4.4%) than in control groups (children, 17.5 ± 2.2%; adults, 31.9 ± 2.5%). The difference in results between the study and the control children's groups was statistically significant (p = 0.003) and also between the study group and the adult control group (p = 0.03). There were higher concentrations than normal upper limit in 22 of 31 patients.

In 16 of these 22 patients and in 4 out of 9 patients with a normal percentage of mannitol excretion, cellobiose was measured. Individual distribution of the percentage of mannitol recovery is shown in Figure 3. A high percentage of mannitol recovery in urine was confirmed by NMR (Figs. 1 and 2; see also the NMR results section).

FIG. 3

FIG. 3

The median (range) value of the CE/MA ratio in 20 patients with IDDM was 0.0% (range, 0-0.06%), which was not statistically significant when compared with ratios in both control groups. Two patients (Table 1, CA and ML) had values beyond the upper normal limit (mean + 2 SD). In both patients no clinical or laboratory signs of inflammatory or malabsorptive intestinal disease were found, even after 1 year of follow-up. None of the investigated parameters correlated, except percentage of mannitol versus length of disease (r = 0.914).

Back to Top | Article Outline

Nuclear Magnetic Resonance

In Figure 1 in spectrum for urine of patients with IDDM after mannitol administration is compared with that of standard mannitol in water, under the same pH conditions, as described in Materials and Methods. The shape of the spectra in the illustrated region unambiguously demonstrates the presence of mannitol in high concentration in the urine sample. The NMR values for five points strictly correlated with the values obtained according to the Corcoran and Page method (r = 0.993; Fig. 2).

Back to Top | Article Outline


Intestinal permeability in patients with diabetes mellitus has seldom been reported (9-10). In one of these studies (9) lactulose and rhamnose were administered as a hyperosmolar solution to 48 men with diabetes mellitus, of whom 8 had IDDM. The overall data showed a significantly augmented permeability to lactulose and a similar trend in rhamnose, with a ratio of the two probes that was not statistically different from ratios in 13 healthy subjects and 40 patients with type II diabetes mellitus. The increase in permeability was the highest in the 8 patients with IDDM. Other specific data of interest were not presented in the article, such as the differential effect of insulin treatment in the two groups of patients with diabetes mellitus, which on the whole caused an increase in permeability to both sugars, and the effect, if any, of neuropathy or angiopathy, which were reported by most patients.

In the other study reported by Cooper et al. (10), intestinal permeability in 12 patients with diabetes mellitus who had diarrhea, neuropathy, and retinopathy, and in 10 patients with uncomplicated diabetes and 25 healthy control subjects, was investigated by a lactulose-mannitol test. We do not know how many if any patients with IDDM were enrolled in the study. The investigators found that the diarrhea was associated with an augmented excretion ratio of both probes, significant because of a lowered permeability to mannitol, in comparison with the same ratio in the patients with uncomplicated diabetes mellitus.

In both studies the ratios between the probes used in patients with diabetes mellitus and control subjects largely overlapped. No correlations with metabolic parameters (blood and urinary glucose, glycosylated hemoglobin) or with renal function parameters (creatinemia, urinary proteins) were found (9,10). Also, in the patients in our study, none of these parameters, or any associated pathologic conditions, such as thyroiditis, rheumatoid arthritis, multiple sclerosis, or secondary amenorrhea, was correlated to intestinal permeability values. The data reported from the literature (9,10) show an increase of permeability to both probes, with only the renal excretion of the larger one significantly higher in certain patients with diabetes mellitus than in control subjects. Although the ratios were in the normal range, the results seem to be not specific and similar to those reported in bacterial overgrowth of the small intestine (29).

We studied a homogeneous series of young patients with uncomplicated IDDM, and our data differ from those in the reported literature in that a significantly higher renal excretion of only the small probe was shown in most cases.

The CE/MA test procedure has been proposed by Strobel et al. (13) as a relatively convenient tool in the study of intestinal permeability and has since been used by other investigators (14-18). In our laboratory it was reliable with no substantial flaws in healthy children and adults; in diseases such as scleroderma (21), radiation enteritis (19-20), cancer chemotherapy (23), postcolectomy pouch (24), and in type II diabetes mellitus (30). The increase of the ratio, which has been shown in different pathologic conditions such as celiac disease (14,16,17), malnutrition (31), iron deficiency (32), idiopathic and infectious diarrhea (33), food allergy (18), cystic fibrosis (15), and Crohn's disease (34-35), was not reproduced in our patients with IDDM. Although we measured percentages of recovered CE in 20 of 31 children, our results may be limited by the 11 (35%) undetectable cases with no data on cellobiose urinary excretion. We cannot exclude that all these cases had a higher than normal CE/MA ratio. The increased mannitol excretion in 6 of 11 such as cases makes that possibility unlikely.

We cannot explain the increased intestinal permeability to mannitol. The endogenous production of mannitol can reasonably be excluded because we administered 2 g sugar, an amount far above the eventual endogenously produced quantity (36). An increase of the intestinal flow rate, although within the normal range, was shown to decrease the estimated pore size of the jejunal diffusion pathway from 13 to 8 angstrom (37) in the normal subject. That could imply a preferential route for mannitol compared with that of cellobiose. In symptom-free patients with diabetes, however, jejunal cecal transmit time seems to be longer than that in healthy subjects (38). Another possibility is that the selective increase in permeability to mannitol is secondary to an increased rate of the mechanism experimentally investigated and proposed by Bijlsma et al. (39) but not confirmed however by others (40) (i.e., a paracellular solvent drag caused by the intravillous hyperosmolarity produced by local countercurrent movements).

The prevalence of celiac disease associated with IDDM has been shown recently to be higher than expected. Patients with celiac disease determined by screening with anti-gliadin (3) or anti-endomysial antibody tests (11) often have few or no symptoms and an earlier age of onset of IDDM (41) and are predominantly female. Some with IDDM who are antibody negative, may later test positive (3), even in the presence of persistently negative results in analyses of small bowel biopsy specimens. From our series we report on two patients (6.2%) who at the 1-year follow-up, showed seroconversion to positive for anti-endomysial antibodies and a typical histologic pattern in a small bowel specimen.

In conclusion, our study showed an increase in intestinal permeability to mannitol not previously reported, clear-cut and unassociated with that of the larger probe, in at least two thirds of 31 young patients, all with uncomplicated IDDM and free from clinical and serologic signs of celiac disease at the time of their recruitment. The high intestinal permeability to mannitol represents a functional alteration apparently opposite to those reported in overt celiac disease, in which the reduced mannitol urinary recovery is caused by the loss of absorptive villous surface which produced an increase in CE/MA ratio. The intestinal permeability to mannitol was normal in our healthy children, young adult controls, and adult patients with type II diabetes mellitus (30). It is difficult to define the impact of this finding, whether it is a primary feature in patients with IDDM or an initial alteration of the villi's function, preceding their atrophy in patients with IDDM, in whom celiac disease could develop later.

Back to Top | Article Outline


1. Feldman M, Schiller LR. Disorders of gastrointestinal motility associated with diabetes mellitus. Ann Intern Med 1983;98:378-80.
2. Ellenberg M. Diabetic enteropathy. Am J Gastroenterol 1963;40:269-75.
3. Maki M, Huupponen T, Holm K, Hallstrom O. Seroconversion of reticulin autoantibodies predicts coeliac disease in insulin dependent diabetes mellitus. Gut 1995;36:239-42.
4. Lanza G. Pancreas. In: Lanza G. ed. Anatomia Patologica Sistematica. Piccin: Padova, 1986:781-825.
5. Kirkman MS, Zimmerman DR, Filippini SA. Marked deterioration in glycemic control with change in brand of lactulose syrup. South Med J 1995;88:492-3.
6. Olsen WA, Rosenberg IH. Intestinal transport of sugars and aminoacids in diabetic rat. J Clin Invest 1970;49:96-105.
7. Jenkins DJA, Taylor RH, Wolever TMS. The diabetic diet, dietary carbohydrate, and difference in digestibility. Diabetologia 1982;23:474-84.
8. Csaky T, Fisher E. Intestinal sugar transport in experimental diabetes. Diabetes 1981;30:568-74.
9. Mooradian AD, Morley JE, Levine AS, Prigge WF, Gebhard RL. Abnormal intestinal permeability to sugars in diabetes mellitus. Diabetologia 1986;29:221-4.
10. Cooper BT, Ukabam SO, O'Brein AD, Hare JPO, Corrall RJM. Intestinal permeability in diabetic diarrhoea. Diabetes Med 1987;4:49-52.
11. Collin P, Salmi J, Hallstrom O. High frequency of coeliac disease in adult patients with type I diabetes. Scand J Gastroenterol 1989;24:81-4.
12. Cronin CC, Shanahan F. Insulin-dependent diabetes mellitus and coeliac disease. Lancet 1997;349:1096-7.
13. Strobel S, Brydon WG, Fergusson A, Cellobiose-mannitol sugar permeability test complements biopsy histopathology in clinical investigation of the jejunum. Gut 1984;25:1241-46.
14. Juby LD, Rothwell J, Axon ATR. Cellobiose/mannitol sugar test. A sensitive tubeless test for coeliac disease: Results on 1010 unselected patients. Gut 1989;30:476-80.
15. Dalzell AM, Freestone NS, Billington D, Heaf DP. Small intestinal permeability and orocaecal transit time in cystic fibrosis. Arch Dis Child 1990;65:585-8.
16. Troncone R, Starita A, Coletta S, Mayer M, Greco L. Antigliadin antibody, D-xylose, and cellobiose/mannitol permeability test as indicators of mucosal damage in children with coeliac disease. Scand J Gastroenterol 1992;27:703-6.
17. Troncone R, Caputo N, Micillo M, Maiuri L, Poggi V. Immunological and intestinal permeability tests as predictors of relapse during gluten challenge in childhood coeliac disease. Scand J Gastroenterol 1994;29:144-7.
18. Troncone R, Caputo N, Florio G, Finelli E. Increased intestinal sugar permeability after challenge in children with cow's milk allergy or intolerance. Allergy 1993;49:142-6.
19. Carratù R, Secondulfo M, deMagistris L, et al. Assessment of damage to small intensive in women treated with pelvic radiotherapy. Oncol Rep 1998;5:635-9.
20. Carratù R, Secondulfo M, de Magistris L, Lubrano C, Sasso FS. Intestinal permeability and radiation enteritis (abstract). Ital Gastroenterol 1995;27(Suppl 1):56.
21. Secondulfo M, deMagistris L, De Luca A, Tirri E, Valentini C, Carratù R. Intestinal permeability in systemic sclerosis (abstract). Gut 1996;39(Suppl 3):A80.
22. deMagistris L, Secondulfo M, Iafusco D, et al. Altered mannitol absorption in diabetic children (letter). Ital J Gastroenterol 1996;28:367.
23. Daniele B, Secondulfo M, Pignata S, et al. Effect of chemotherapy with 5-fluorouracil on intestinal permeability of patients with advanced colorectal cancer (abstract). Eur J Cancer 1997;33(suppl 8):S169.
24. Riegler G, de Magistris L, Secondulfo M, Tartaglione M, Selvaggi F, Carratù R. Assessment of intestinal permeability in patients with ileal pouch (IPAA) (abstract). Gastroenterology 1997;112(suppl):A397.
25. Corcoran AC, Page IH. A method for the determination of mannitol in plasma and urine. J Biol Chem 1947;170:165-71.
26. Pontoni G, Rotondo F, Dardo G, Cartenì-Farina M, Zappia V. Proton magnetic resonance spectroscopy of biological fluids: A powerful tool in the diagnosis of inherited metabolic diseases (abstract). Eur J Clin Chem Clin Biochem 1994;32:A7.
27. Lehnert W, Hunkler D. Possibilities of selective screening for inborn errors of metabolism using high resolution 1H-FT-NMR spectrometry. Eur J Pediatr 1986;145:260-6.
28. Tofts PS, Wray S. A critical assessment of methods of measuring metabolite concentration by NMR spectroscopy. NMR Biomed 1988;1:1-10.
29. Pignata C, Budillon G, Monaco G, et al. Jejunal bacterial overgrowth and intestinal permeability in children with immunodeficiency syndromes. Gut 1990;31:879-82.
30. deMagistris L, Secondulfo M, Iafusco D, et al. Assessment of intestinal permeability diabetes mellitus type I (IDDM) and type II (NIDDM) (abstract). Gastroenterology 1997;112(suppl): A358.
31. Penny DI, Ingall CB, Boulton NP, Walker-Smith TA, Basheer S. Intestinal malabsorption in cystic fibrosis. Arch Dis Child 19861:1127-8.
32. Bernant M, Khourie M, Menzies IS. Effects of iron deficiency on small intestinal permeability in infants and young children. J Pediatr Gastroenterol Nutr 1992;14:17-20.
33. Keating J, Bjarnason I, Somesundaram S, et al. Intestinal absorptive capacity, intestinal permeability and jejunal histology in HIV and their relation to diarrhoea. Gut 1995;37:623-9.
34. Hollander D. Intestinal permeability barrier. Scand J Gastroenterol 1992;27:721-6.
35. Olaison G, Tagesson C. Abnormal intestinal permeability in Crohn's disease. A possible pathogenetic factor. Scand J Gastroenterol 1990;25:321-8.
36. Bjarnason I, MacPherson A, Hollander D. Intestinal permeability: an overview. Gastroenterology 1995;108:1566-81.
37. Fine KD, Carol A, Santana E, Porter JI, Fordtran JS. Effect of changing intestinal flow rate on a measurement of intestinal permeability. Gastroenterology 1995;108:983-9.
38. Spengler U, Stellard F, Ruckdeschelg K, Scheurlen C, Kruis W. Small intestinal transit, bacterial growth, and bowel habits in diabetes mellitus. Pancreas 1989;4:65-70.
39. Bijlsma PB, Peeters RA, Proot JA, Dekker PR, Taminiau JAJM, Van der Reer R. Differential in vivo and in vitro intestinal permeability to lactulose and mannitol in animals and humans: A hypothesis. Gastroenterology 1995;108:676-96.
40. Gabe SM, Crane R, Menzies IS, Silk DBA, Bjarnason I. Intestinal permeability assessment evidence against the solvent drug hypothesis (abstract). Gut 1996;39(suppl 1):A6.
41. Cacciari E, Salardi S, Volta U. Prevalence and characteristics of coeliac disease in type I diabetes mellitus. Acta Pediatr Scand 1987;76:671-2.

Cellobiose; Diabetes mellitus type I; Intestinal permeability; Mannitol

© 1999 Lippincott Williams & Wilkins, Inc.