ORIGINAL CONTRIBUTIONS
nature publishing group
1
Breath Methane Excretion Is not An Accurate Marker of Colonic Methane Production in Irritable Bowel Syndrome Michele Di Stefano, MD1, Caterina Mengoli, MD1, Manuela Bergonzi, MD1, Catherine Klersy, MD2, Elisabetta Pagani, MD1, Emanuela Miceli, MD1 and Gino Roberto Corazza, MD1 OBJECTIVES:
The role of colonic methane production in functional bowel disorders is still uncertain. In small samples of irritable bowel syndrome (IBS) patients, it was shown that methane breath excretion correlates with clinical presentation and delayed gastrointestinal transit time. The aim of this study was to evaluate the relationship between intestinal production and breath excretion of CH4 and to correlate CH4 production with the presence and the severity of symptoms, in a large cohort of IBS patients and in a group of healthy volunteers.
METHODS:
A group of 103 IBS patients and a group of 28 healthy volunteers were enrolled. The presence and severity of symptoms and gastrointestinal transit were evaluated in all subjects, who underwent breath H2/CH4 measurement for 7 h after lactulose to identify breath excretors of these gases; H2 and CH4 were also measured in rectal samples to identify colonic producers. Cumulative H2 and CH4 excretion and production were evaluated by the area under the time–concentration curve calculation (AUC).
RESULTS:
In IBS patients, CH4 was detected in rectal samples in 48 patients (47%), but only 27 of them (26% of the 103 enrolled patients) excreted this gas with breath. In CH4 producers, the prevalence and severity of symptoms and gastrointestinal transit time were not significantly different with respect to non-producers. IBS subtypes were homogeneously represented in CH4 producers and in non-producers. Healthy volunteers, compared with IBS patients, showed a significantly lower prevalence of CH4 excretion, whereas no difference was found in the prevalence of colonic CH4 production; moreover, in healthy volunteers compared with IBS, CH4 breath excretion and CH4 production were not different in quantitative terms.
CONCLUSION:
Our data show that colonic CH4 production is not associated with clinical presentation in IBS patients and does not correlate with symptom severity or with gastrointestinal transit time. Clinical inferences based on breath CH4 excretion should undergo an in-depth revision, as this method is not a good marker of CH4 colonic production.
SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg
Am J Gastroenterol advance online publication, 24 March 2015; doi:10.1038/ajg.2015.47
INTRODUCTION Up to 20% of ingested carbohydrates escape digestion in the human small intestine (1) and are available for fermentation by colonic bacteria. Besides water and short-chain fatty acids (SCFAs), several gases are formed, mainly carbon dioxide (CO2),
nitrogen (N2), hydrogen (H2), and methane (CH4). Some papers have previously shown a regulatory role of products of fermentation on the sensorimotor activity of the gastrointestinal tract: the exposure of the proximal colon to short-chain fatty acids induces the relaxation of the proximal stomach (2) and modifies
1
1st Department of Medicine, University of Pavia Foundation IRCCS “S. Matteo” Hospital, Pavia, Italy; 2Biometry and Clinical Epidemiology, Foundation IRCCS “S. Matteo” Hospital, Pavia, Italy. Correspondence: Gino Roberto Corazza, MD, 1st Department of Medicine, University of Pavia, Foundation IRCCS “S. Matteo” Hospital, Pavia 27100, Italy. E-mail: [email protected] Received 25 June 2014; accepted 29 January 2015
© 2015 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
FUNCTIONAL GI DISORDERS
see related editorial on page x
FUNCTIONAL GI DISORDERS
2
Di Stefano et al.
lower esophageal sphincter function, triggering transient relaxations (3,4); an increased short-chain fatty acid concentration in the ileum stimulates peristaltic contractions and increases tonic activity, as an emptying response (5); conflicting results are available on the effect on colonic motility, as a stimulation of colonic transit via intraluminal release of 5HT in animals (6), but negative results are reported in humans (7). We have recently shown that colonic fermentation of orally administered lactulose induces a modification of the discomfort threshold in a subgroup of IBS patients (8), suggesting a putative role of fermentation products in intestinal sensitivity. Moreover, a regulatory role of intraluminal CH4 in intestinal motility was recently suggested (9). In particular, it was shown that intraluminal CH4 slows intestinal transit, and an increase of small bowel contractile activity, shown in vitro, was proposed as a putative mechanism responsible for this alteration (9). On clinical grounds, the presence of CH4 in breath excretion was suggested as an indirect indicator of the severity of constipation in patients with IBS (10), and a further study on a very small cohort of CH4 producer IBS patients (only six patients) suggested that the presence of CH4 in breath excretion predicts the improvement of constipation after antibiotic administration (11). These results were obtained in very small cohorts of patients and many doubts could be raised on their reproducibility in a larger population group. Some years ago, it was shown that colonic CH4 production does not always result in breath CH4 excretion (12). Consequently, monitoring breath CH4 excretion may not be the best way to select CH4 producers. We therefore conducted a prospective study to evaluate, first of all, the association between breath CH4 excretion and the presence and severity of functional bowel symptoms in a large cohort of IBS patients. Moreover, to perform a more in-depth analysis of the relationship between CH4 and IBS, we also measured the colonic intraluminal CH4 level to verify its direct role in the pathogenesis of functional symptoms.
METHODS Subjects
A group of 103 consecutive breath H2 excretor IBS patients (76 females, mean age 41±14 years, range 18–89 years), diagnosed according to the Rome III criteria (13), was enrolled. The IBS subtype was constipation predominant in 44 (35 females, mean age 44.7±14.1 years), diarrhea predominant in 53 (39 females, mean age 41.5±15.3 years), and mixed IBS in 6 patients (2 females, mean age 50±27 years). Twenty-eight H2 excretor healthy volunteers (HVs) (12 females, mean age 44.1±17.5 years), members of the medical and paramedical staff, were enrolled on a voluntary basis as a control group. None of them suffered from intestinal, liver, renal, chest, cardiac, metabolic, or neurologic diseases; in particular, no one suffered from functional bowel disorders. Both patient and healthy volunteer groups consisted in already known H2 excretors, as this condition was tested before the subjects entered the study. The American Journal of GASTROENTEROLOGY
Study design
At entry, all subjects underwent symptom evaluation and gastrointestinal transit time study. They then underwent H2 and CH4 measurement in breath and simultaneously in rectal gas samples after lactulose administration, and were divided into two subgroups according to the presence of CH4 detection in rectal samples. The term “excretor” refers to breath sample results and the term “producer” refers to rectal gas samples. It was then analyzed whether symptom presence and severity were different among the two subgroups and HVs and whether the prevalence of CH4 producers was different among IBS subtypes. Finally, the relationship among intraluminal CH4 and H2 production and the subjects’ breath excretion was analyzed. The protocol was approved by the Local Ethics Committee (IRCCS Policlinico S. Matteo Foundation). Symptom evaluation
The presence and severity of abdominal pain, bloating, flatulence, and borborygmi were evaluated according to Tack et al. (14,15). The Bristol Stool Form Scale was used to evaluate stool consistency (16). For each patient, a mean score was calculated averaging the scores of the last six bowel movements before enrolment. Breath testing
All patients underwent breath H2 and CH4 excretion monitoring after oral lactulose administration, in accordance with the Rome Consensus Conference (17). To avoid prolonged intestinal gas production, due to the presence of non-absorbable or slowly fermentable material in the colonic lumen, subjects consumed a meal consisting of only rice, meat, and olive oil the evening before the test day (18). This meal was followed by a 12-h fasting period. Breath testing started between 0830 and 0930 hours, after thorough mouthwash with 40 ml of 1% chlorhexidine solution (19). Smoking (20) and physical exercise (21) were not allowed for 1 h before and throughout the test. Sampling of alveolar air was performed by means of a commercial device (GasamplerQuintron, Milwaukee, WI), which allows the first 500 ml of dead space air to be separated and discarded while the remaining 700 ml of end-alveolar air are collected in a gas-tight bag. Subjects were instructed to avoid deep inspiration and not to hyperventilate before exhalation. A gas chromatograph dedicated to the detection of H2 and CH4 in air samples was used for breath sample analysis (Model DP12; Quintron Instrument, Milwaukee, WI). The accuracy of the detector was ±2 p.p.m. with a linear response range between 2 and 150 p.p.m. of H2 and between 2 and 50 p.p.m. of CH4. Twelve grams of lactulose in a 250 ml water solution were administered per os, and air samples were collected in fasting condition and every 15 min. Breath H2 and CH4 excretion were monitored for a 7-h period. The presence of an average breath CH4 concentration >5 p.p.m. above that of room air during breath testing was considered indicative of CH4 production (22). The area under the time–concentration curve (AUC) and the mean gas concentration were calculated www.amjgastro.com
Methane Production and Excretion in IBS
Colonic gas measurement
After an overnight fast, a preliminary gentle, non-chemical, up to 200 ml tap-water enema was performed to avoid surface damage while ensuring rectal emptying (24). A double lumen polyvinyl tube with an adherent, infinitely compliant cylindrical plastic bag (200 ml capacity, 10 cm length, 5 cm maximal diameter) was inserted through the anus, positioned at the rectosigmoid junction, and secured with adhesive tape. The proximal border of the bag was 2 cm from the distal tip of the tube. The distal border of the bag was 14±3 cm (range 12–16 cm) from the anal verge. A fluoroscopic control with about 50 ml of air in the balloon ensured the correct positioning. The bag was inflated through channel 2 to obtain a slight dilation of the viscus, avoiding adherence between the colonic mucosa and the intraluminal tube during sampling of intestinal gas. At the same time as breath sample collection, in fasting condition and every 15 min after the lactulose intake, a volume of 20 ml of rectal gas was obtained through channel 1 of the two lumens of the polyvinyl tube. The first sample was obtained after a 20 min resting period following insertion of the balloon. At the end of each test, a second fluoroscopic control was performed to exclude dislocation of the balloon. The reproducibility of the method for rectal gas measurement was tested on a subgroup of 10 patients. In this subgroup, on two separate days at least a month apart, the colonic gas measurement was repeated. In all patients, CH4 production or non-production at the first test was confirmed at the second test, resulting in perfect reproducibility of the producer/non-producer status. Moreover, in terms of absolute values, the difference of gas levels between the first and the second test was very low. To assess the reproducibility of measures, we computed the coefficient of variation and 95% confidence interval (95% CI), the Lin’s concordance correlation coefficient (95% CI), and the Bland–Altman limits of agreement. For H2, the coefficient of variation was 5.2% (95% CI: 2.2–7.0), the Lin’s concordance correlation coefficient was 0.999 (95% CI: 0.997–0.999), and the Bland–Altman limits were between 21,531 and 15,447; for CH4, the coefficient of variation was 5.5% (95% CI: 4.0–6.6), the Lin’s concordance correlation coefficient was 0.994 (95% CI: 0.977–0.999), and the Bland–Altman limits were between 5,166 and 5,590. As for gas breath excretion, the AUC (23) and the mean gas concentration were calculated to estimate intestinal gas production both for H2 and CH4. © 2015 by the American College of Gastroenterology
Gastrointestinal transit measurement
In all the subjects, an evaluation of total gastrointestinal transit time with radio-opaque markers was performed, according to Metcalf et al. (25). Briefly, on three consecutive days, each day in the morning during breakfast, a set of 20 radiopaque markers, a different shape each day, was administered orally, and an X-ray plan of the abdomen was taken on the fourth day. The total gastrointestinal transit was calculated in hours, multiplying the number of persistent markers in the colon by a factor of 1.2. Data analysis
Considering an α value of 1% and a Pearson’s correlation coefficient (r) value between methane production and any symptom scores of 0.65 as biologically meaningful, 100 subjects were needed to give the study a power of 81%. The α value of 1% will allow the statistical analysis of subgroups of patients. Previous results on the same field yielded an r value of 0.60–0.70 (10). All variables, except cumulative gas production and excretion (AUC), are expressed as mean±s.d. Gas AUCs, due to their wide intersubject variability, are expressed as median (25th–75th percentile). Distribution of data was tested with the Kolmogorov– Smirnov normality test. All variables except gas AUC followed a parametric distribution: paired and unpaired t-tests were computed for the comparison between the two groups. Pearson’s correlation coefficient was computed for the parametric estimates of the level of association between two variables. For nonparametric distributions (H2 and CH4 AUC and mean concentrations), the Wilcoxon’s signed-rank test was computed to evaluate the difference within groups and the Mann–Whitney U-test was computed to evaluate the difference between groups. Spearman’s coefficient of correlation was computed to evaluate the level of association between two variables. A P value
nature publishing group
1
Breath Methane Excretion Is not An Accurate Marker of Colonic Methane Production in Irritable Bowel Syndrome Michele Di Stefano, MD1, Caterina Mengoli, MD1, Manuela Bergonzi, MD1, Catherine Klersy, MD2, Elisabetta Pagani, MD1, Emanuela Miceli, MD1 and Gino Roberto Corazza, MD1 OBJECTIVES:
The role of colonic methane production in functional bowel disorders is still uncertain. In small samples of irritable bowel syndrome (IBS) patients, it was shown that methane breath excretion correlates with clinical presentation and delayed gastrointestinal transit time. The aim of this study was to evaluate the relationship between intestinal production and breath excretion of CH4 and to correlate CH4 production with the presence and the severity of symptoms, in a large cohort of IBS patients and in a group of healthy volunteers.
METHODS:
A group of 103 IBS patients and a group of 28 healthy volunteers were enrolled. The presence and severity of symptoms and gastrointestinal transit were evaluated in all subjects, who underwent breath H2/CH4 measurement for 7 h after lactulose to identify breath excretors of these gases; H2 and CH4 were also measured in rectal samples to identify colonic producers. Cumulative H2 and CH4 excretion and production were evaluated by the area under the time–concentration curve calculation (AUC).
RESULTS:
In IBS patients, CH4 was detected in rectal samples in 48 patients (47%), but only 27 of them (26% of the 103 enrolled patients) excreted this gas with breath. In CH4 producers, the prevalence and severity of symptoms and gastrointestinal transit time were not significantly different with respect to non-producers. IBS subtypes were homogeneously represented in CH4 producers and in non-producers. Healthy volunteers, compared with IBS patients, showed a significantly lower prevalence of CH4 excretion, whereas no difference was found in the prevalence of colonic CH4 production; moreover, in healthy volunteers compared with IBS, CH4 breath excretion and CH4 production were not different in quantitative terms.
CONCLUSION:
Our data show that colonic CH4 production is not associated with clinical presentation in IBS patients and does not correlate with symptom severity or with gastrointestinal transit time. Clinical inferences based on breath CH4 excretion should undergo an in-depth revision, as this method is not a good marker of CH4 colonic production.
SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg
Am J Gastroenterol advance online publication, 24 March 2015; doi:10.1038/ajg.2015.47
INTRODUCTION Up to 20% of ingested carbohydrates escape digestion in the human small intestine (1) and are available for fermentation by colonic bacteria. Besides water and short-chain fatty acids (SCFAs), several gases are formed, mainly carbon dioxide (CO2),
nitrogen (N2), hydrogen (H2), and methane (CH4). Some papers have previously shown a regulatory role of products of fermentation on the sensorimotor activity of the gastrointestinal tract: the exposure of the proximal colon to short-chain fatty acids induces the relaxation of the proximal stomach (2) and modifies
1
1st Department of Medicine, University of Pavia Foundation IRCCS “S. Matteo” Hospital, Pavia, Italy; 2Biometry and Clinical Epidemiology, Foundation IRCCS “S. Matteo” Hospital, Pavia, Italy. Correspondence: Gino Roberto Corazza, MD, 1st Department of Medicine, University of Pavia, Foundation IRCCS “S. Matteo” Hospital, Pavia 27100, Italy. E-mail: [email protected] Received 25 June 2014; accepted 29 January 2015
© 2015 by the American College of Gastroenterology
The American Journal of GASTROENTEROLOGY
FUNCTIONAL GI DISORDERS
see related editorial on page x
FUNCTIONAL GI DISORDERS
2
Di Stefano et al.
lower esophageal sphincter function, triggering transient relaxations (3,4); an increased short-chain fatty acid concentration in the ileum stimulates peristaltic contractions and increases tonic activity, as an emptying response (5); conflicting results are available on the effect on colonic motility, as a stimulation of colonic transit via intraluminal release of 5HT in animals (6), but negative results are reported in humans (7). We have recently shown that colonic fermentation of orally administered lactulose induces a modification of the discomfort threshold in a subgroup of IBS patients (8), suggesting a putative role of fermentation products in intestinal sensitivity. Moreover, a regulatory role of intraluminal CH4 in intestinal motility was recently suggested (9). In particular, it was shown that intraluminal CH4 slows intestinal transit, and an increase of small bowel contractile activity, shown in vitro, was proposed as a putative mechanism responsible for this alteration (9). On clinical grounds, the presence of CH4 in breath excretion was suggested as an indirect indicator of the severity of constipation in patients with IBS (10), and a further study on a very small cohort of CH4 producer IBS patients (only six patients) suggested that the presence of CH4 in breath excretion predicts the improvement of constipation after antibiotic administration (11). These results were obtained in very small cohorts of patients and many doubts could be raised on their reproducibility in a larger population group. Some years ago, it was shown that colonic CH4 production does not always result in breath CH4 excretion (12). Consequently, monitoring breath CH4 excretion may not be the best way to select CH4 producers. We therefore conducted a prospective study to evaluate, first of all, the association between breath CH4 excretion and the presence and severity of functional bowel symptoms in a large cohort of IBS patients. Moreover, to perform a more in-depth analysis of the relationship between CH4 and IBS, we also measured the colonic intraluminal CH4 level to verify its direct role in the pathogenesis of functional symptoms.
METHODS Subjects
A group of 103 consecutive breath H2 excretor IBS patients (76 females, mean age 41±14 years, range 18–89 years), diagnosed according to the Rome III criteria (13), was enrolled. The IBS subtype was constipation predominant in 44 (35 females, mean age 44.7±14.1 years), diarrhea predominant in 53 (39 females, mean age 41.5±15.3 years), and mixed IBS in 6 patients (2 females, mean age 50±27 years). Twenty-eight H2 excretor healthy volunteers (HVs) (12 females, mean age 44.1±17.5 years), members of the medical and paramedical staff, were enrolled on a voluntary basis as a control group. None of them suffered from intestinal, liver, renal, chest, cardiac, metabolic, or neurologic diseases; in particular, no one suffered from functional bowel disorders. Both patient and healthy volunteer groups consisted in already known H2 excretors, as this condition was tested before the subjects entered the study. The American Journal of GASTROENTEROLOGY
Study design
At entry, all subjects underwent symptom evaluation and gastrointestinal transit time study. They then underwent H2 and CH4 measurement in breath and simultaneously in rectal gas samples after lactulose administration, and were divided into two subgroups according to the presence of CH4 detection in rectal samples. The term “excretor” refers to breath sample results and the term “producer” refers to rectal gas samples. It was then analyzed whether symptom presence and severity were different among the two subgroups and HVs and whether the prevalence of CH4 producers was different among IBS subtypes. Finally, the relationship among intraluminal CH4 and H2 production and the subjects’ breath excretion was analyzed. The protocol was approved by the Local Ethics Committee (IRCCS Policlinico S. Matteo Foundation). Symptom evaluation
The presence and severity of abdominal pain, bloating, flatulence, and borborygmi were evaluated according to Tack et al. (14,15). The Bristol Stool Form Scale was used to evaluate stool consistency (16). For each patient, a mean score was calculated averaging the scores of the last six bowel movements before enrolment. Breath testing
All patients underwent breath H2 and CH4 excretion monitoring after oral lactulose administration, in accordance with the Rome Consensus Conference (17). To avoid prolonged intestinal gas production, due to the presence of non-absorbable or slowly fermentable material in the colonic lumen, subjects consumed a meal consisting of only rice, meat, and olive oil the evening before the test day (18). This meal was followed by a 12-h fasting period. Breath testing started between 0830 and 0930 hours, after thorough mouthwash with 40 ml of 1% chlorhexidine solution (19). Smoking (20) and physical exercise (21) were not allowed for 1 h before and throughout the test. Sampling of alveolar air was performed by means of a commercial device (GasamplerQuintron, Milwaukee, WI), which allows the first 500 ml of dead space air to be separated and discarded while the remaining 700 ml of end-alveolar air are collected in a gas-tight bag. Subjects were instructed to avoid deep inspiration and not to hyperventilate before exhalation. A gas chromatograph dedicated to the detection of H2 and CH4 in air samples was used for breath sample analysis (Model DP12; Quintron Instrument, Milwaukee, WI). The accuracy of the detector was ±2 p.p.m. with a linear response range between 2 and 150 p.p.m. of H2 and between 2 and 50 p.p.m. of CH4. Twelve grams of lactulose in a 250 ml water solution were administered per os, and air samples were collected in fasting condition and every 15 min. Breath H2 and CH4 excretion were monitored for a 7-h period. The presence of an average breath CH4 concentration >5 p.p.m. above that of room air during breath testing was considered indicative of CH4 production (22). The area under the time–concentration curve (AUC) and the mean gas concentration were calculated www.amjgastro.com
Methane Production and Excretion in IBS
Colonic gas measurement
After an overnight fast, a preliminary gentle, non-chemical, up to 200 ml tap-water enema was performed to avoid surface damage while ensuring rectal emptying (24). A double lumen polyvinyl tube with an adherent, infinitely compliant cylindrical plastic bag (200 ml capacity, 10 cm length, 5 cm maximal diameter) was inserted through the anus, positioned at the rectosigmoid junction, and secured with adhesive tape. The proximal border of the bag was 2 cm from the distal tip of the tube. The distal border of the bag was 14±3 cm (range 12–16 cm) from the anal verge. A fluoroscopic control with about 50 ml of air in the balloon ensured the correct positioning. The bag was inflated through channel 2 to obtain a slight dilation of the viscus, avoiding adherence between the colonic mucosa and the intraluminal tube during sampling of intestinal gas. At the same time as breath sample collection, in fasting condition and every 15 min after the lactulose intake, a volume of 20 ml of rectal gas was obtained through channel 1 of the two lumens of the polyvinyl tube. The first sample was obtained after a 20 min resting period following insertion of the balloon. At the end of each test, a second fluoroscopic control was performed to exclude dislocation of the balloon. The reproducibility of the method for rectal gas measurement was tested on a subgroup of 10 patients. In this subgroup, on two separate days at least a month apart, the colonic gas measurement was repeated. In all patients, CH4 production or non-production at the first test was confirmed at the second test, resulting in perfect reproducibility of the producer/non-producer status. Moreover, in terms of absolute values, the difference of gas levels between the first and the second test was very low. To assess the reproducibility of measures, we computed the coefficient of variation and 95% confidence interval (95% CI), the Lin’s concordance correlation coefficient (95% CI), and the Bland–Altman limits of agreement. For H2, the coefficient of variation was 5.2% (95% CI: 2.2–7.0), the Lin’s concordance correlation coefficient was 0.999 (95% CI: 0.997–0.999), and the Bland–Altman limits were between 21,531 and 15,447; for CH4, the coefficient of variation was 5.5% (95% CI: 4.0–6.6), the Lin’s concordance correlation coefficient was 0.994 (95% CI: 0.977–0.999), and the Bland–Altman limits were between 5,166 and 5,590. As for gas breath excretion, the AUC (23) and the mean gas concentration were calculated to estimate intestinal gas production both for H2 and CH4. © 2015 by the American College of Gastroenterology
Gastrointestinal transit measurement
In all the subjects, an evaluation of total gastrointestinal transit time with radio-opaque markers was performed, according to Metcalf et al. (25). Briefly, on three consecutive days, each day in the morning during breakfast, a set of 20 radiopaque markers, a different shape each day, was administered orally, and an X-ray plan of the abdomen was taken on the fourth day. The total gastrointestinal transit was calculated in hours, multiplying the number of persistent markers in the colon by a factor of 1.2. Data analysis
Considering an α value of 1% and a Pearson’s correlation coefficient (r) value between methane production and any symptom scores of 0.65 as biologically meaningful, 100 subjects were needed to give the study a power of 81%. The α value of 1% will allow the statistical analysis of subgroups of patients. Previous results on the same field yielded an r value of 0.60–0.70 (10). All variables, except cumulative gas production and excretion (AUC), are expressed as mean±s.d. Gas AUCs, due to their wide intersubject variability, are expressed as median (25th–75th percentile). Distribution of data was tested with the Kolmogorov– Smirnov normality test. All variables except gas AUC followed a parametric distribution: paired and unpaired t-tests were computed for the comparison between the two groups. Pearson’s correlation coefficient was computed for the parametric estimates of the level of association between two variables. For nonparametric distributions (H2 and CH4 AUC and mean concentrations), the Wilcoxon’s signed-rank test was computed to evaluate the difference within groups and the Mann–Whitney U-test was computed to evaluate the difference between groups. Spearman’s coefficient of correlation was computed to evaluate the level of association between two variables. A P value
