- Email: [email protected]
Fasting breath hydrogen in celiac disease
GASTROENTEROLOGY 1987;93:53-8
Fasting Breath Hydrogen in Celiac Disease GINO ROBERTO CORAZZA, GIOVANNI GASBARRINI
ALESSANDRA
STROCCHI,
and
I Department of Medical Pathology, S. Orsola University Hospital, University of Bologna, Bologna, Italy
We studied the possible clinical significance of high basal levels of H2 by analyzing the breath excreted by the following fasting subjects: 50 healthy volunteers, 149 subjects with functional bowel disorders, 16 patients with small bowel bacterial overgrowth proven by bacteriology, 34 patients with untreated celiac disease, 40 patients with celiac disease on a gluten-free diet, and 40 patients with disorders of the small intestine other than celiac disease (disease controls). The fasting levels of Hz in untreated celiac patients (mean 22.5 k 19.3 ppm) were significantly higher than those in healthy volunteers (5.8 k 3.1 ppm), patients with functional bowel disorders (6.6 k 4.4 ppm), treated celiac patients (9.9 +- 8.1 ppm), and disease controls (7.0 k 6.7 ppm). No significant difierence was found between patients with untreated celiac disease and bacterial overgrowth (mean 14.7 + 14.0 ppm). The percentage of patients with elevated Hz fasting levels in untreated celiac disease [58.8%) was significantly higher than that in the other groups, except for the patients with bacterial overgrowth (43.7%). In 14 celiac patients, studied before and after a gluten-free diet, fasting H2 levels decreased from 26.6 k 18 to 11.6 2 10 ppm, becoming normal only in those patients with healing of intestinal lesions. Our results show that high fasting Hz levels are a frequent feature of untreated celiac disease and that the return to normal of these levels is predictive of villous regrowth. Measurement of respiratory hydrogen (Hz) excretion provides an accurate means of estimating the rate of Hz production in the intestine (1). As H, production Received July 30, 1986. Accepted January 12, 1987. Address requests for reprints to: G. R. Corazza, M.D., I Patologia Medica, Nuove Patologie, Policlinico S. Orsola, via Massarenti 9, 40138 Bologna, Italy. This work was partly supported by CNR grant 85.00820. The authors thank Dr. M. G. Menozzi (Department of Microbiology, University of Parma) for the bacteriologic 0 1987 by the American Gastroenterological 0016-5085/87/$3.50
studies. Association
greatly increases when a small amount of carbohydrate is supplied to colonic bacteria, the measurement of respiratory H2 excretion has been proposed as an indicator of carbohydrate malabsorption (2). Although breath Hz analysis is performed after an overnight fast, individual variations in fasting breath hydrogen (FBH2) concentration have been noted (3,4). Perman et al. (5) evaluated the possible diagnostic significance of the FBH;! concentration, and a review of their patients with values above the normal range revealed that an elevated FBHz concentration was associated with small-bowel bacterial overgrowth. On this basis, we evaluated FBHz levels in patients with functional bowel disorders and with organic diseases of the small intestine. Our results indicate that untreated celiac disease is the condition most frequently associated with high concentrations of FBH2.
Materials and Methods Subjects Fasting breath samples were obtained from a number of adult subjects who had undergone the H2 breath test in our laboratory. Subjects who had received antibiotics (6) or other drugs known to affect gastrointestinal function, in the z wk before the study, were excluded from this analysis. Cigarette smoking (7) and physical exercise sufficient to produce hyperventilation (8,q) were not allowed in the z h preceding the test. Breath samples were obtained at 9 AM, and all subjects were asked to fast from 9 PM on the day before the test and to avoid eating bread, pasta, and fibers the previous evening, which may cause prolonged excretion of H2 (10,ll). Before obtaining the fasting sample, all subjects were questioned about their compliance with these instructions. The following subjects took part in the study: healthy Abbreviations used fasting
breath
hydrogen.
in this
paper:
CD, celiac
disease;
FBH,,
GASTROENTEROLOGYVol. 93, No. 1
54 CORAZZA ET AL.
Table
1. Final Diagnoses
of 45 Disease Controls
Crohn’s disease of the small bowel Intestinal resection Postenteritis svndrome Dermatitis herpetiformis Intestinal lymphangiectasia Giardiasis Small bowel lymphoma Intestinal tuberculosis Intestinal amyloidosis Gastrointestinal mastocytosis
14 11
5 3
3 3 2 2
HZ Testing
End-expiratory fasting breath samples were collected by a device consisting of a mouthpiece, a T valve, and two collapsible bags (Quintron Instrument Co., Milwaukee, Wis.). The first portion of the expired air (-500 ml) represents dead-space air and is directed into the “discard” bag. As the expiratory effort continues, an alveolar sample is collected in the “collection” bag. The breath sample is then transferred to the sample loop of the chromatograph by a 60-ml syringe and Luer taper connectors. Hydrogen concentration, expressed in parts per million (ppm), was determined on a MicroLyzer model 12 gas chromatograph [Quintron Instruments Co.) (12~3). Bacteriologic
Studies
Samples of jejunal fluid were collected using the capsule described by Shiner et al. (14). Aerobic and anaerobic quantitative bacteriologic investigations were performed on the collected fluid (15,16). The 19 patients with bacterial overgrowth included in this study were diagnosed on the basis of bacterial concentrations >107 colony-forming units per milliliter. Rice-Breath
Statistics unpaired
Results were analyzed using Student’s and paired data, and linear regression.
t-test for
1 1
volunteers (n = 53), patients with functional bowel complaints such as recurrent abdominal pain, chronic diarrhea, or constipation (n = 17l), patients with small intestine bacterial overgrowth proven by bacteriology (n = 191, patients with untreated celiac disease [CD] proven by jejunal biopsy (n = x), patients with CD on a gluten-free diet (n = 43), and patients with other organic diseases of the small intestine, often associated with malabsorption, listed in Table 1 [disease controls) (n = 45). Fifty subjects who took part in the study were sampled twice within a lo-day period. Thirty-seven patients of 366 (10.1%)were found to be Hz nonproducers and were therefore excluded from any evaluation. All groups were matched for sex and age. Informed written consent was obtained from each subject taking part in the study and the investigation was approved by the local ethical committee on human experimentation. Breath
excretion was estimated by calculating the area under the curve of HZ concentration against time using the equation for the sum of the areas of consecutive trapezoids as proposed by Kotler et al. (4).
H2 Test
This test was performed according to Kerlin et al. (17), giving 15 untreated celiac patients a test meal of 100 g of carbohydrate in the form of rice flour. Breath samples were collected at 30-min intervals over 8 h. Cumulative HZ
Results Fasting breath hydrogen concentrations in healthy volunteers and in the five groups of patients studied are shown in Figure 1. In healthy volunteers FBHP ranged from 1 to 14 ppm (mean 5.8 + 3.1); 12.0 ppm (mean + 2 SD) was considered the highest value of the control range. No significant difference was found between healthy volunteers, subjects with functional bowel disorders (mean 6.6 ? 4.41, and disease controls (7.0 t 6.7). In bacterial overgrowth (mean 14.7 * 14.0), untreated CD (22.5 2 19.3), and treated CD (9.9 ? 8.1), FBHz values were significantly higher than in healthy volunteers (p < 0.001 for each of the three comparisons). No significant difference was found between patients with bacterial overgrowth and untreated CD or between patients with bacterial overgrowth and treated CD. In untreated CD, FBHz levels were significantly higher than in treated CD (p < 0.001) and in disease controls (p < 0.001). Figure 2 shows that in 50 subjects (healthy volunteers n = 17, patients with bacterial overgrowth n = 13, patients with untreated CD n = 10, disease controls n = lo), there was a highly significant correlation between the levels of FBHz obtained on two different occasions within a lo-day period (r = 0.89, pcz) < 0.001). In only 4 of the 50 subjects did the FBHz values obtained at the second sampling disagree in terms of normality and abnormality with those obtained at the first sampling. The frequency of subjects with FBH2 values >l2.0 ppm (the upper limit of the control range) is shown in Figure 3. It was found that 8.0% of healthy volunteers, 12.1% of subjects with functional bowel disorders, 43.7% of patients with bacterial overgrowth, 58.8% of patients with untreated CD, 32.5% of patients with treated CD, and 15.0% of the disease controls had FBH2 concentrations higher than the upper limit of the control range. Among the patients with bacterial overgrowth, the 7 patients with high FBHz levels did not differ in clinical symptoms, in isolated bacterial species, or in colony-forming units per milliliter from the 9 patients with normal levels of FBHz. With regard to the percentage of patients with elevated FBHz concentrations, no significant difference was found between healthy volunteers, subjects
FASTING BREATH HYDROGEN IN CELIAC DISEASE
July 1987
HEALTHY VOLUNTEERS
FUNCTIONAL BOUEL DISORDERS
BACTERIAL OVERGROUTH
UNTREATED cELIAc DISEASE
TREATED CELIAC DISEASE
55
DISEASE CONTROLS
- (93)
60-
.
. .
. .
. . ..
-.-
&g&y ............. ..s...
------a-----* . ...: __,_ :...
Figure
i
.. . .. : ... .
:. . .-.
. -1.
_____+____ ___ _______ _______2__ .... ....
.
_________ -
: . .
r----______l__
i_l-
:
--e---..:. . ..
i
. . . .-. . . .
‘I’
. . . .
,, . .::. . ‘-. . . . ..I... ::
1. Fasting breath hydrogen concentrations in the six groups of subjects studied. The dashed line represents control range (mean value of the healthy volunteers + 2 SD).
with functional bowel disorders, and disease controls. However, the percentages of patients with elevated FBHz concentrations in bacterial overgrowth, untreated CD, and treated CD were signifi-
the upper limit of the
cantly higher than those in healthy volunteers (p < 0.005, p < 0.001, p < 0.01, respectively). No significant difference was found between patients with bacterial overgrowth and patients with untreated CD (93;1241
I
1/
60
0
I I
0
I
0 .
Figure
2. Fasting breath hydrogen concentrations in 50 subjects (A, healthy volunteers; 0, untreated CD; A, disease controls; 0, bacterial overgrowth) studied on two separate occasions within a lo-day period.
0
.
i I I I
.
.
lo . I
,
.
FBH2
(PMI)
Initial
I
56
GASTROENTEROLOGY Vol. 93, No.
CORAZZA ET AL.
I
1
??
50
. .
401
.
3
-m30 E
. .
20
.
.
I
. .
IO
.
.
r= 0.21 P
.
NS
.
Figure 3. Percentage of patients with elevated levels of FBH2 in the six groups studied.
or between patients with treated CD and disease controls. In untreated CD, the percentage of patients with an elevated FB.Hz concentration was significantly higher than in treated CD (p < 0.05) and in the disease control group (p < 0.001). Fasting breath hydrogen concentrations in 14 celiac patients before and after a gluten-free diet [mean duration of the diet: 18 mo + 30 mo) are shown in Figure 4. A significant reduction in Hz levels was
BEFORE DIET
AFTER DIET
5,000
10,000
15,000
CUMULATIVE Hz
20,000
25,000
(eom/min)
Figure 5. Fasting breath hydrogen concentrations plotted against H, cumulative excretion over 8 h after ingestion of rice flour in 15 untreated celiac patients.
evident after treatment. In 9 of 14 untreated celiac patients, FBHz levels were higher than the upper limit of the control range. Of the 6 patients with levels that were still abnormal after treatment, 5 had the second jejunal biopsy still compatible with subtotal or severe partial villous atrophy. Figure 5 shows that in 15 untreated celiac patients no correlation was found between the FBHz concentrations and carbohydrate malabsorption assessed by cumulative Hz excretion over 8 h after ingestion of rice flour.
Discussion
-
Figure 4. Fasting patients triangle presented subtotal
p
-
breath hydrogen concentrations in 14 celiac before and after gluten-free diet. An open indicates 5 treated celiac patients who still an intestinal biopsy specimen indicative of or severe partial villous atrophy.
Hydrogen breath testing has been proposed as a method of detecting carbohydrate malabsorption (2), bacterial overgrowth (18), and small bowel transit time (19). Routine use of the test has revealed some pitfalls, one of the most common being the detection in certain subjects of a high FBH2 concentration after an overnight fast. The ingestion of bread and pasta can lead to prolonged excretion of Hz (lo), whereas the ingestion of a carbohydrate-free dinner (4) or of carbohydrate in the form of rice flour only (5) on the evening preceding the test reliably reduces FBHz levels in healthy individuals. All the subjects who took part in our study were asked to fast from 9 PM the evening before the study and to avoid eating bread, pasta, and fibers. Our results confirm that in subjects with normal or high values of FBHz, such values show a considerable stability if tested under the same experimental conditions on two separate occasions (5). Previous studies have shown high levels of FBHz in pneumatosis cystoides intestinalis [zo), cystic
FASTING BREATH HYDROGEN IN CELIAC DISEASE
July 1987
fibrosis (~1)~ and in patients with functional bowel disorders, the majority of whom showed evidence of bacterial overgrowth (5). We demonstrated that not only in bacterial overgrowth but also in CD, FBH2 concentrations were significantly higher than in healthy volunteers. The mean FBHz level in our 50 healthy volunteers was similar to that obtained by other authors in analogous conditions (4,5,17,22). Only 4 of 50 healthy volunteers (8.0%)had FBH2 levels above the mean values of their group + 2 SD. However, the proportion of patients with high concentrations of FBHz was 43.7% in bacterial overgrowth and 58.8% in untreated CD, and we are therefore in agreement with Perman et al. (5) that elevated FBHz values may have clinical implications. As it has been shown that 20% of children with CD may have small bowel bacterial overgrowth (23), it may be argued that the high levels of FBHz that we found in untreated CD are due to a concomitant bacterial overgrowth. However, the fact that the proportion of untreated celiac patients with high FBHz levels (58.8%)is well above that of patients with bacterial overgrowth (43.7%) disproves, in our opinion, this hypothesis. Twelve percent of patients with functional bowel disorders and 15% of patients with organic diseases of the small intestine other than CD (disease controls) had elevated levels of FBH2, but there was no difference between these groups and the healthy volunteers. It has been shown that, unlike what occurs in healthy subjects, in patients with bacterial overgrowth the consumption of a standard dinner meal without bread and pasta does not reduce FBHz levels (5). Moreover, as already stated, all those taking part in our study avoided these substances in the evening meal on the day before the study. Nevertheless, we cannot exclude the possibility that the greater basal Hz excretion observed in our patients with bacterial overgrowth and with CD is due to fermentation in the colon of exogenous carbohydrates that are easily absorbed by the small intestine of normal subjects but malabsorbed in these conditions. Moreover, in untreated CD the colonic fermentation of exogenous sugars may be delayed, based on indicating a preliminary report that in celiac patients with steatorrhea the mouth-cecum transit time is delayed and that this is accompanied by elevated FBH2 levels (24).However, the presence in disease controls of FBHz levels not significantly different from those of normal subjects and the lack of correlation in untreated CD between FBHz levels and rice flour malabsorption suggest that the prolonged fermentation of malabsorbed exogenous sugars may represent an important factor, but probably not the only one, in determining the high levels of FBHz in CD. By means of an in vitro fecal incubation system, HZ
57
production by glycoproteins of both dietary and endogenous origin was also demonstrated (25). Thus the availability of endogenous substances to intestinal bacteria may enhance the basal excretion of Hz. The importance of endogenous glycoproteins as potential sources of Hz has already been considered for bacterial overgrowth (5) and may play an even greater role in untreated CD. In this condition, in fact, glycoproteins of cellular origin, due to the very high enterocyte shedding (26,27), and serum glycoproteins such as albumin (28)and cll-antitrypsin (29), due to their exudation through the damaged mucosa, become available for fermentation by the luminal bacteria. Perman et al. (30) showed that Hz production after exposure of carbohydrate to colonic flora is maximal at neutral pH and is strongly inhibited at acid pH. In untreated CD the intraluminal pH in the upper jejunum is more alkaline than in normal subjects (31), and this may contribute to a greater production of Hz by a fermentable substrate (of dietary or endogenous origin) even in the presence of a quantitatively normal jejunal bacterial flora. In untreated CD the levels of FBHz are significantly reduced after a gluten-free diet. In 6 of the 14 patients treated, however, the FBH2 levels remained within the abnormal range. In 5 of these patients, intestinal lesions varying from severe partial villous atrophy to subtotal villous atrophy were still present, due either to the too short duration of the gluten-free diet or to poor compliance with the diet. In conclusion, our results demonstrate for the first time that elevated levels of FBHz are a frequent feature of untreated CD. In celiac patients with initially high levels of FBH2, the normalizing of such values may be indicative of a normalization of the intestinal lesions and contribute, therefore, to determining the most suitable time to perform a second intestinal biopsy, which is necessary for the diagnosis of this condition. References 1. Levitt MD. Production
2.
3.
4.
5.
and excretion of hydrogen gas in man. N Engl J Med 1969;281:122-7. Levitt MD, Donaldson RM. Use of respiratory hydrogen (H,) excretion to detect carbohydrate malabsorption. J Lab Clin Med 1970;75:937-45. Bond JH, Levitt MD. Use of pulmonary hydrogen (H,) measurements to quantitate carbohydrate malabsorption. Study of partially gastrectomized patients. J Clin Invest 1972;51: 1219-25. Kotler DP, Holt PR, Rosensweig NS. Modification of the breath hydrogen test: increased sensitivity for the detection of carbohydrate malabsorption. J Lab Clin Med 1982;lOO: 798-805. Perman JA, Modler S, Barr RG, Rosenthal P. Fasting breath hydrogen concentration: normal values and clinical application. Gastroenterology 1984;87:1358-63.
58 CORAZZA ET AL.
6. Gilat T, Ben Hur H, Gelman-Malachi E, Terdiman R, Peled Y. Alterations of the colonic flora and their effect on the hydrogen breath test. Gut 1978;19:602-5. 7. Thompson DG, Binfield P, De Belder A, O’Brien J, Warren S, Wilson M. Extra intestinal influences on exhaled breath hydrogen measurements during the investigation of gastrointestinal disease. Gut 1985;26:1349-52. 8. Payne DL, Welsh JD, Claypool PL. Breath hydrogen (H,) response to carbohydrate malabsorption after exercise. J Lab Clin Med 1983;102:147-50. 9. Perman JA, Modler S, Engel RR, Heldt G. Effect of ventilation on breath hydrogen measurements. J Lab Clin Med 1985; 105:436-g. 10. Anderson IH, Levine AS, Levitt MD. Incomplete absorption of the carbohydrates in all-purpose wheat flour. N Engl J Med 1981;304:891-2. 11. Brummer RJM, Armbrecht U, Bosaeus I, Dotevall G, Stockbruegger RW. The hydrogen (H,) breath test. Sampling methods and the influence of dietary fibre on fasting level. Stand J Gastroenterol 1985;20:1007-13. 12. Christman NT, Hamilton LH. A new chromatographic instrument for measuring trace concentrations of breath-hydrogen. J Chromatogr 1982;229:259-65. 13. Solomons NW, Hamilton LH, Christman NT, Rothman D. Evaluation of a rapid breath hydrogen analyzer for clinical studies of carbohydrate absorption. Dig Dis Sci 1983;28: 397-404. 14. Shiner M, Waters TE, Allan Gray JD. Culture studies of the gastrointestinal tract with a newly devised capsule. Results of tests in vitro and in vivo. Gastroenterology 1963;45:625-32. 15. Gorbach SL. Intestinal microflora. Gastroenterology 1971; 60:1110-29. 16. Holdeman LV, Cato EP, Moore WEC. Anaerobe laboratory manual. 4th ed. Blacksburg, Va.: Virginia Polytechnic Institute and State University, 1977. 17. Kerlin P, Wong L, Harris B, Capra S. Rice flour, breath hydrogen, and malabsorption. Gastroenterology 1984;87: 578-85. 18. Metz G, Gassull MA, Drasar BS, Jenkins DJA, Blendis LM.
GASTROENTEROLOGY Vol. 93, No. 1
19.
20.
21. 22.
23.
24. 25.
26. 27.
28.
29.
30.
31.
Breath hydrogen test for small-intestine bacterial colonisation. Lancet 1976;i:668-9. Bond JH, Levitt MD. Investigation of small bowel transit time in man utilizing pulmonary hydrogen (H,) measurements. J Lab Clin Med 1975;85:546-55. Gillon J, Tadesse K, Logan RFA, Holt S, Sircus W. Breath hydrogen in pneumatosis cystoides intestinalis. Gut 1979; 20:1008-11. Bali A, Stableforth DE, Asquith P. Prolonged small-intestinal transit time in cystic fibrosis. Br Med J 1983;287:1011-3, Saltzberg DM, Levine GM, Lubar C. Common pitfalls in interpreting hydrogen breath tests (abstr). Gastroenterology 1985;88:1566. Anderson CM, Langford RF. Bacterial content of small intestine of children in health, in coeliac disease, and in fibrocystic disease of pancreas. Br Med J 1958;1:803-6. Spiller RC, Lee YC, Edge C, et al. Functional significance of the ileal brake in coeliac disease. Gut 1984;25:A545. Perman JA, Modler S. Glycoproteins as substrates for production of hydrogen and methane by colonic bacterial flora. Gastroenterology 1982;83:388-93. Croft DN, Loehry CA, Creamer B. Small-bowel cell-loss and weight-loss in coeliac syndrome. Lancet 1968;ii:68-70. Trier JS, Browning TH. Epithelial-cell renewal in cultured duodenal biopsies in celiac sprue. N Engl J Med 1970; 283:1245-50. Cluysenaer OJJ, Corstens FHM, Hafkenscheid JCM, Yap SH, Van Tongeren JHM. Mechanisms of hypo-albuminaemia in coeliac sprue. In: Hekkens WTJM, Pena AS, eds. Coeliac disease. Leiden: Stenfert Kroese, 1974:386-96. Thomas DW, Sinatra FR, Merrit RJ. Random fecal alpha-lantitrypsin concentration in children with gastrointestinal disease. Gastroenterology 1981;80:776-82. Perman JA, Modler S, Olson AC. Role of pH in production of hydrogen from carbohydrates by colonic bacterial flora. Studies in vivo and in vitro. J Clin Invest 1981;67:643-50. Prizont R, Hersh T, Floch MH. Jejunal bacterial flora in chronic small-bowel disease. I. Coeliac disease. II. Regional enteritis. Am J Clin Nutr 1970;23:1602-7.
Fasting Breath Hydrogen in Celiac Disease GINO ROBERTO CORAZZA, GIOVANNI GASBARRINI
ALESSANDRA
STROCCHI,
and
I Department of Medical Pathology, S. Orsola University Hospital, University of Bologna, Bologna, Italy
We studied the possible clinical significance of high basal levels of H2 by analyzing the breath excreted by the following fasting subjects: 50 healthy volunteers, 149 subjects with functional bowel disorders, 16 patients with small bowel bacterial overgrowth proven by bacteriology, 34 patients with untreated celiac disease, 40 patients with celiac disease on a gluten-free diet, and 40 patients with disorders of the small intestine other than celiac disease (disease controls). The fasting levels of Hz in untreated celiac patients (mean 22.5 k 19.3 ppm) were significantly higher than those in healthy volunteers (5.8 k 3.1 ppm), patients with functional bowel disorders (6.6 k 4.4 ppm), treated celiac patients (9.9 +- 8.1 ppm), and disease controls (7.0 k 6.7 ppm). No significant difierence was found between patients with untreated celiac disease and bacterial overgrowth (mean 14.7 + 14.0 ppm). The percentage of patients with elevated Hz fasting levels in untreated celiac disease [58.8%) was significantly higher than that in the other groups, except for the patients with bacterial overgrowth (43.7%). In 14 celiac patients, studied before and after a gluten-free diet, fasting H2 levels decreased from 26.6 k 18 to 11.6 2 10 ppm, becoming normal only in those patients with healing of intestinal lesions. Our results show that high fasting Hz levels are a frequent feature of untreated celiac disease and that the return to normal of these levels is predictive of villous regrowth. Measurement of respiratory hydrogen (Hz) excretion provides an accurate means of estimating the rate of Hz production in the intestine (1). As H, production Received July 30, 1986. Accepted January 12, 1987. Address requests for reprints to: G. R. Corazza, M.D., I Patologia Medica, Nuove Patologie, Policlinico S. Orsola, via Massarenti 9, 40138 Bologna, Italy. This work was partly supported by CNR grant 85.00820. The authors thank Dr. M. G. Menozzi (Department of Microbiology, University of Parma) for the bacteriologic 0 1987 by the American Gastroenterological 0016-5085/87/$3.50
studies. Association
greatly increases when a small amount of carbohydrate is supplied to colonic bacteria, the measurement of respiratory H2 excretion has been proposed as an indicator of carbohydrate malabsorption (2). Although breath Hz analysis is performed after an overnight fast, individual variations in fasting breath hydrogen (FBH2) concentration have been noted (3,4). Perman et al. (5) evaluated the possible diagnostic significance of the FBH;! concentration, and a review of their patients with values above the normal range revealed that an elevated FBHz concentration was associated with small-bowel bacterial overgrowth. On this basis, we evaluated FBHz levels in patients with functional bowel disorders and with organic diseases of the small intestine. Our results indicate that untreated celiac disease is the condition most frequently associated with high concentrations of FBH2.
Materials and Methods Subjects Fasting breath samples were obtained from a number of adult subjects who had undergone the H2 breath test in our laboratory. Subjects who had received antibiotics (6) or other drugs known to affect gastrointestinal function, in the z wk before the study, were excluded from this analysis. Cigarette smoking (7) and physical exercise sufficient to produce hyperventilation (8,q) were not allowed in the z h preceding the test. Breath samples were obtained at 9 AM, and all subjects were asked to fast from 9 PM on the day before the test and to avoid eating bread, pasta, and fibers the previous evening, which may cause prolonged excretion of H2 (10,ll). Before obtaining the fasting sample, all subjects were questioned about their compliance with these instructions. The following subjects took part in the study: healthy Abbreviations used fasting
breath
hydrogen.
in this
paper:
CD, celiac
disease;
FBH,,
GASTROENTEROLOGYVol. 93, No. 1
54 CORAZZA ET AL.
Table
1. Final Diagnoses
of 45 Disease Controls
Crohn’s disease of the small bowel Intestinal resection Postenteritis svndrome Dermatitis herpetiformis Intestinal lymphangiectasia Giardiasis Small bowel lymphoma Intestinal tuberculosis Intestinal amyloidosis Gastrointestinal mastocytosis
14 11
5 3
3 3 2 2
HZ Testing
End-expiratory fasting breath samples were collected by a device consisting of a mouthpiece, a T valve, and two collapsible bags (Quintron Instrument Co., Milwaukee, Wis.). The first portion of the expired air (-500 ml) represents dead-space air and is directed into the “discard” bag. As the expiratory effort continues, an alveolar sample is collected in the “collection” bag. The breath sample is then transferred to the sample loop of the chromatograph by a 60-ml syringe and Luer taper connectors. Hydrogen concentration, expressed in parts per million (ppm), was determined on a MicroLyzer model 12 gas chromatograph [Quintron Instruments Co.) (12~3). Bacteriologic
Studies
Samples of jejunal fluid were collected using the capsule described by Shiner et al. (14). Aerobic and anaerobic quantitative bacteriologic investigations were performed on the collected fluid (15,16). The 19 patients with bacterial overgrowth included in this study were diagnosed on the basis of bacterial concentrations >107 colony-forming units per milliliter. Rice-Breath
Statistics unpaired
Results were analyzed using Student’s and paired data, and linear regression.
t-test for
1 1
volunteers (n = 53), patients with functional bowel complaints such as recurrent abdominal pain, chronic diarrhea, or constipation (n = 17l), patients with small intestine bacterial overgrowth proven by bacteriology (n = 191, patients with untreated celiac disease [CD] proven by jejunal biopsy (n = x), patients with CD on a gluten-free diet (n = 43), and patients with other organic diseases of the small intestine, often associated with malabsorption, listed in Table 1 [disease controls) (n = 45). Fifty subjects who took part in the study were sampled twice within a lo-day period. Thirty-seven patients of 366 (10.1%)were found to be Hz nonproducers and were therefore excluded from any evaluation. All groups were matched for sex and age. Informed written consent was obtained from each subject taking part in the study and the investigation was approved by the local ethical committee on human experimentation. Breath
excretion was estimated by calculating the area under the curve of HZ concentration against time using the equation for the sum of the areas of consecutive trapezoids as proposed by Kotler et al. (4).
H2 Test
This test was performed according to Kerlin et al. (17), giving 15 untreated celiac patients a test meal of 100 g of carbohydrate in the form of rice flour. Breath samples were collected at 30-min intervals over 8 h. Cumulative HZ
Results Fasting breath hydrogen concentrations in healthy volunteers and in the five groups of patients studied are shown in Figure 1. In healthy volunteers FBHP ranged from 1 to 14 ppm (mean 5.8 + 3.1); 12.0 ppm (mean + 2 SD) was considered the highest value of the control range. No significant difference was found between healthy volunteers, subjects with functional bowel disorders (mean 6.6 ? 4.41, and disease controls (7.0 t 6.7). In bacterial overgrowth (mean 14.7 * 14.0), untreated CD (22.5 2 19.3), and treated CD (9.9 ? 8.1), FBHz values were significantly higher than in healthy volunteers (p < 0.001 for each of the three comparisons). No significant difference was found between patients with bacterial overgrowth and untreated CD or between patients with bacterial overgrowth and treated CD. In untreated CD, FBHz levels were significantly higher than in treated CD (p < 0.001) and in disease controls (p < 0.001). Figure 2 shows that in 50 subjects (healthy volunteers n = 17, patients with bacterial overgrowth n = 13, patients with untreated CD n = 10, disease controls n = lo), there was a highly significant correlation between the levels of FBHz obtained on two different occasions within a lo-day period (r = 0.89, pcz) < 0.001). In only 4 of the 50 subjects did the FBHz values obtained at the second sampling disagree in terms of normality and abnormality with those obtained at the first sampling. The frequency of subjects with FBH2 values >l2.0 ppm (the upper limit of the control range) is shown in Figure 3. It was found that 8.0% of healthy volunteers, 12.1% of subjects with functional bowel disorders, 43.7% of patients with bacterial overgrowth, 58.8% of patients with untreated CD, 32.5% of patients with treated CD, and 15.0% of the disease controls had FBH2 concentrations higher than the upper limit of the control range. Among the patients with bacterial overgrowth, the 7 patients with high FBHz levels did not differ in clinical symptoms, in isolated bacterial species, or in colony-forming units per milliliter from the 9 patients with normal levels of FBHz. With regard to the percentage of patients with elevated FBHz concentrations, no significant difference was found between healthy volunteers, subjects
FASTING BREATH HYDROGEN IN CELIAC DISEASE
July 1987
HEALTHY VOLUNTEERS
FUNCTIONAL BOUEL DISORDERS
BACTERIAL OVERGROUTH
UNTREATED cELIAc DISEASE
TREATED CELIAC DISEASE
55
DISEASE CONTROLS
- (93)
60-
.
. .
. .
. . ..
-.-
&g&y ............. ..s...
------a-----* . ...: __,_ :...
Figure
i
.. . .. : ... .
:. . .-.
. -1.
_____+____ ___ _______ _______2__ .... ....
.
_________ -
: . .
r----______l__
i_l-
:
--e---..:. . ..
i
. . . .-. . . .
‘I’
. . . .
,, . .::. . ‘-. . . . ..I... ::
1. Fasting breath hydrogen concentrations in the six groups of subjects studied. The dashed line represents control range (mean value of the healthy volunteers + 2 SD).
with functional bowel disorders, and disease controls. However, the percentages of patients with elevated FBHz concentrations in bacterial overgrowth, untreated CD, and treated CD were signifi-
the upper limit of the
cantly higher than those in healthy volunteers (p < 0.005, p < 0.001, p < 0.01, respectively). No significant difference was found between patients with bacterial overgrowth and patients with untreated CD (93;1241
I
1/
60
0
I I
0
I
0 .
Figure
2. Fasting breath hydrogen concentrations in 50 subjects (A, healthy volunteers; 0, untreated CD; A, disease controls; 0, bacterial overgrowth) studied on two separate occasions within a lo-day period.
0
.
i I I I
.
.
lo . I
,
.
FBH2
(PMI)
Initial
I
56
GASTROENTEROLOGY Vol. 93, No.
CORAZZA ET AL.
I
1
??
50
. .
401
.
3
-m30 E
. .
20
.
.
I
. .
IO
.
.
r= 0.21 P
.
NS
.
Figure 3. Percentage of patients with elevated levels of FBH2 in the six groups studied.
or between patients with treated CD and disease controls. In untreated CD, the percentage of patients with an elevated FB.Hz concentration was significantly higher than in treated CD (p < 0.05) and in the disease control group (p < 0.001). Fasting breath hydrogen concentrations in 14 celiac patients before and after a gluten-free diet [mean duration of the diet: 18 mo + 30 mo) are shown in Figure 4. A significant reduction in Hz levels was
BEFORE DIET
AFTER DIET
5,000
10,000
15,000
CUMULATIVE Hz
20,000
25,000
(eom/min)
Figure 5. Fasting breath hydrogen concentrations plotted against H, cumulative excretion over 8 h after ingestion of rice flour in 15 untreated celiac patients.
evident after treatment. In 9 of 14 untreated celiac patients, FBHz levels were higher than the upper limit of the control range. Of the 6 patients with levels that were still abnormal after treatment, 5 had the second jejunal biopsy still compatible with subtotal or severe partial villous atrophy. Figure 5 shows that in 15 untreated celiac patients no correlation was found between the FBHz concentrations and carbohydrate malabsorption assessed by cumulative Hz excretion over 8 h after ingestion of rice flour.
Discussion
-
Figure 4. Fasting patients triangle presented subtotal
p
-
breath hydrogen concentrations in 14 celiac before and after gluten-free diet. An open indicates 5 treated celiac patients who still an intestinal biopsy specimen indicative of or severe partial villous atrophy.
Hydrogen breath testing has been proposed as a method of detecting carbohydrate malabsorption (2), bacterial overgrowth (18), and small bowel transit time (19). Routine use of the test has revealed some pitfalls, one of the most common being the detection in certain subjects of a high FBH2 concentration after an overnight fast. The ingestion of bread and pasta can lead to prolonged excretion of Hz (lo), whereas the ingestion of a carbohydrate-free dinner (4) or of carbohydrate in the form of rice flour only (5) on the evening preceding the test reliably reduces FBHz levels in healthy individuals. All the subjects who took part in our study were asked to fast from 9 PM the evening before the study and to avoid eating bread, pasta, and fibers. Our results confirm that in subjects with normal or high values of FBHz, such values show a considerable stability if tested under the same experimental conditions on two separate occasions (5). Previous studies have shown high levels of FBHz in pneumatosis cystoides intestinalis [zo), cystic
FASTING BREATH HYDROGEN IN CELIAC DISEASE
July 1987
fibrosis (~1)~ and in patients with functional bowel disorders, the majority of whom showed evidence of bacterial overgrowth (5). We demonstrated that not only in bacterial overgrowth but also in CD, FBH2 concentrations were significantly higher than in healthy volunteers. The mean FBHz level in our 50 healthy volunteers was similar to that obtained by other authors in analogous conditions (4,5,17,22). Only 4 of 50 healthy volunteers (8.0%)had FBH2 levels above the mean values of their group + 2 SD. However, the proportion of patients with high concentrations of FBHz was 43.7% in bacterial overgrowth and 58.8% in untreated CD, and we are therefore in agreement with Perman et al. (5) that elevated FBHz values may have clinical implications. As it has been shown that 20% of children with CD may have small bowel bacterial overgrowth (23), it may be argued that the high levels of FBHz that we found in untreated CD are due to a concomitant bacterial overgrowth. However, the fact that the proportion of untreated celiac patients with high FBHz levels (58.8%)is well above that of patients with bacterial overgrowth (43.7%) disproves, in our opinion, this hypothesis. Twelve percent of patients with functional bowel disorders and 15% of patients with organic diseases of the small intestine other than CD (disease controls) had elevated levels of FBH2, but there was no difference between these groups and the healthy volunteers. It has been shown that, unlike what occurs in healthy subjects, in patients with bacterial overgrowth the consumption of a standard dinner meal without bread and pasta does not reduce FBHz levels (5). Moreover, as already stated, all those taking part in our study avoided these substances in the evening meal on the day before the study. Nevertheless, we cannot exclude the possibility that the greater basal Hz excretion observed in our patients with bacterial overgrowth and with CD is due to fermentation in the colon of exogenous carbohydrates that are easily absorbed by the small intestine of normal subjects but malabsorbed in these conditions. Moreover, in untreated CD the colonic fermentation of exogenous sugars may be delayed, based on indicating a preliminary report that in celiac patients with steatorrhea the mouth-cecum transit time is delayed and that this is accompanied by elevated FBH2 levels (24).However, the presence in disease controls of FBHz levels not significantly different from those of normal subjects and the lack of correlation in untreated CD between FBHz levels and rice flour malabsorption suggest that the prolonged fermentation of malabsorbed exogenous sugars may represent an important factor, but probably not the only one, in determining the high levels of FBHz in CD. By means of an in vitro fecal incubation system, HZ
57
production by glycoproteins of both dietary and endogenous origin was also demonstrated (25). Thus the availability of endogenous substances to intestinal bacteria may enhance the basal excretion of Hz. The importance of endogenous glycoproteins as potential sources of Hz has already been considered for bacterial overgrowth (5) and may play an even greater role in untreated CD. In this condition, in fact, glycoproteins of cellular origin, due to the very high enterocyte shedding (26,27), and serum glycoproteins such as albumin (28)and cll-antitrypsin (29), due to their exudation through the damaged mucosa, become available for fermentation by the luminal bacteria. Perman et al. (30) showed that Hz production after exposure of carbohydrate to colonic flora is maximal at neutral pH and is strongly inhibited at acid pH. In untreated CD the intraluminal pH in the upper jejunum is more alkaline than in normal subjects (31), and this may contribute to a greater production of Hz by a fermentable substrate (of dietary or endogenous origin) even in the presence of a quantitatively normal jejunal bacterial flora. In untreated CD the levels of FBHz are significantly reduced after a gluten-free diet. In 6 of the 14 patients treated, however, the FBH2 levels remained within the abnormal range. In 5 of these patients, intestinal lesions varying from severe partial villous atrophy to subtotal villous atrophy were still present, due either to the too short duration of the gluten-free diet or to poor compliance with the diet. In conclusion, our results demonstrate for the first time that elevated levels of FBHz are a frequent feature of untreated CD. In celiac patients with initially high levels of FBH2, the normalizing of such values may be indicative of a normalization of the intestinal lesions and contribute, therefore, to determining the most suitable time to perform a second intestinal biopsy, which is necessary for the diagnosis of this condition. References 1. Levitt MD. Production
2.
3.
4.
5.
and excretion of hydrogen gas in man. N Engl J Med 1969;281:122-7. Levitt MD, Donaldson RM. Use of respiratory hydrogen (H,) excretion to detect carbohydrate malabsorption. J Lab Clin Med 1970;75:937-45. Bond JH, Levitt MD. Use of pulmonary hydrogen (H,) measurements to quantitate carbohydrate malabsorption. Study of partially gastrectomized patients. J Clin Invest 1972;51: 1219-25. Kotler DP, Holt PR, Rosensweig NS. Modification of the breath hydrogen test: increased sensitivity for the detection of carbohydrate malabsorption. J Lab Clin Med 1982;lOO: 798-805. Perman JA, Modler S, Barr RG, Rosenthal P. Fasting breath hydrogen concentration: normal values and clinical application. Gastroenterology 1984;87:1358-63.
58 CORAZZA ET AL.
6. Gilat T, Ben Hur H, Gelman-Malachi E, Terdiman R, Peled Y. Alterations of the colonic flora and their effect on the hydrogen breath test. Gut 1978;19:602-5. 7. Thompson DG, Binfield P, De Belder A, O’Brien J, Warren S, Wilson M. Extra intestinal influences on exhaled breath hydrogen measurements during the investigation of gastrointestinal disease. Gut 1985;26:1349-52. 8. Payne DL, Welsh JD, Claypool PL. Breath hydrogen (H,) response to carbohydrate malabsorption after exercise. J Lab Clin Med 1983;102:147-50. 9. Perman JA, Modler S, Engel RR, Heldt G. Effect of ventilation on breath hydrogen measurements. J Lab Clin Med 1985; 105:436-g. 10. Anderson IH, Levine AS, Levitt MD. Incomplete absorption of the carbohydrates in all-purpose wheat flour. N Engl J Med 1981;304:891-2. 11. Brummer RJM, Armbrecht U, Bosaeus I, Dotevall G, Stockbruegger RW. The hydrogen (H,) breath test. Sampling methods and the influence of dietary fibre on fasting level. Stand J Gastroenterol 1985;20:1007-13. 12. Christman NT, Hamilton LH. A new chromatographic instrument for measuring trace concentrations of breath-hydrogen. J Chromatogr 1982;229:259-65. 13. Solomons NW, Hamilton LH, Christman NT, Rothman D. Evaluation of a rapid breath hydrogen analyzer for clinical studies of carbohydrate absorption. Dig Dis Sci 1983;28: 397-404. 14. Shiner M, Waters TE, Allan Gray JD. Culture studies of the gastrointestinal tract with a newly devised capsule. Results of tests in vitro and in vivo. Gastroenterology 1963;45:625-32. 15. Gorbach SL. Intestinal microflora. Gastroenterology 1971; 60:1110-29. 16. Holdeman LV, Cato EP, Moore WEC. Anaerobe laboratory manual. 4th ed. Blacksburg, Va.: Virginia Polytechnic Institute and State University, 1977. 17. Kerlin P, Wong L, Harris B, Capra S. Rice flour, breath hydrogen, and malabsorption. Gastroenterology 1984;87: 578-85. 18. Metz G, Gassull MA, Drasar BS, Jenkins DJA, Blendis LM.
GASTROENTEROLOGY Vol. 93, No. 1
19.
20.
21. 22.
23.
24. 25.
26. 27.
28.
29.
30.
31.
Breath hydrogen test for small-intestine bacterial colonisation. Lancet 1976;i:668-9. Bond JH, Levitt MD. Investigation of small bowel transit time in man utilizing pulmonary hydrogen (H,) measurements. J Lab Clin Med 1975;85:546-55. Gillon J, Tadesse K, Logan RFA, Holt S, Sircus W. Breath hydrogen in pneumatosis cystoides intestinalis. Gut 1979; 20:1008-11. Bali A, Stableforth DE, Asquith P. Prolonged small-intestinal transit time in cystic fibrosis. Br Med J 1983;287:1011-3, Saltzberg DM, Levine GM, Lubar C. Common pitfalls in interpreting hydrogen breath tests (abstr). Gastroenterology 1985;88:1566. Anderson CM, Langford RF. Bacterial content of small intestine of children in health, in coeliac disease, and in fibrocystic disease of pancreas. Br Med J 1958;1:803-6. Spiller RC, Lee YC, Edge C, et al. Functional significance of the ileal brake in coeliac disease. Gut 1984;25:A545. Perman JA, Modler S. Glycoproteins as substrates for production of hydrogen and methane by colonic bacterial flora. Gastroenterology 1982;83:388-93. Croft DN, Loehry CA, Creamer B. Small-bowel cell-loss and weight-loss in coeliac syndrome. Lancet 1968;ii:68-70. Trier JS, Browning TH. Epithelial-cell renewal in cultured duodenal biopsies in celiac sprue. N Engl J Med 1970; 283:1245-50. Cluysenaer OJJ, Corstens FHM, Hafkenscheid JCM, Yap SH, Van Tongeren JHM. Mechanisms of hypo-albuminaemia in coeliac sprue. In: Hekkens WTJM, Pena AS, eds. Coeliac disease. Leiden: Stenfert Kroese, 1974:386-96. Thomas DW, Sinatra FR, Merrit RJ. Random fecal alpha-lantitrypsin concentration in children with gastrointestinal disease. Gastroenterology 1981;80:776-82. Perman JA, Modler S, Olson AC. Role of pH in production of hydrogen from carbohydrates by colonic bacterial flora. Studies in vivo and in vitro. J Clin Invest 1981;67:643-50. Prizont R, Hersh T, Floch MH. Jejunal bacterial flora in chronic small-bowel disease. I. Coeliac disease. II. Regional enteritis. Am J Clin Nutr 1970;23:1602-7.