Small-bowel bacterial overgrowth (SBBO) is a common complication in a number of pediatric conditions, including short-bowel syndrome, idiopathic pseudoobstruction syndrome, bowel strictures, malnutrition, and achlorhydria. Significant complications associated with SBBO include malabsorption of fat, vitamin deficiencies, malnutrition, chronic diarrhea, abdominal pain, and anemia.
The diagnosis of SBBO is fraught with difficulties. The “gold standard,” aspiration and culture of small-bowel secretions, is invasive and technically difficult. Therefore, indirect measures of the presence of bacteria within the intestinal lumen are frequently employed. Many of these measures, including serum folate and urinary indican levels, have inadequate sensitivity and specificity (1,2). Other methods, including measurement of unconjugated bile acids in duodenal fluid or serum, require invasive techniques to obtain samples (3) or sophisticated laboratory procedures for analysis (4), limiting their clinical utility.
Breath hydrogen tests (in fasted state and/or after ingestion of a fermentable substrate) are convenient, widely available, and safe; however, sensitivity is compromised by the presence of bacteria that do not produce hydrogen (5-9). 14C-bile acid breath tests are not specific for SBBO in the presence of ileal damage or resection, secondary to bacterial deconjugation of the unabsorbed 14C-bile salt in the colon (10). 14C-xylose was found to be superior to 14C-labeled bile acids (10) and fermentable sugars (6) as a substrate in breath tests for the diagnosis of SBBO in adults. Despite disagreement about the sensitivity and specificity of the 14C-xylose breath test in various groups of patients (6,10-14) it is the most promising diagnostic method to date. However, it has not been studied in children because of the use of a radioactive label. A breath test, in which xylose is labeled with 13C, a safe, nonradioactive isotope, has recently been developed (Martek, Columbia, MD). The 13C-xylose breath test was found to be 100% specific and 100% sensitive for the diagnosis of SBBO in a small group of adults with bacterial overgrowth when compared with its specificity and sensitivity in control subjects (15).
The purpose of this study was to evaluate the utility of the 13C-xylose breath test for the diagnosis of SBBO in children. We hypothesized that children with bacterial culture-documented SBBO would have an increased 13CO2 in expired air after ingestion of 13C-xylose compared with that in normal children, and effective antibiotic treatment of bacterial overgrowth would return 13CO2 excretion in breath to normal levels after adminstration of 13C-xylose.
MATERIALS AND METHODS
For the first part of the study, the determination of the optimal dose of 13C-xylose in healthy children, thirty children between 2 and 12 years old were recruited. Exclusion criteria included pulmonary, hepatobiliary, or gastrointestinal disease; malnutrition; and antibiotic use within the past month.
Each subject reported to the General Clinical Research Center at Cincinnati, Ohio, after an overnight fast of at least 8 hours. Each subject was instructed to fast except for water and to engage in quiet play (no strenuous activity) during the testing period, to minimize the variation of 13C background in respiratory CO2 (16). Weight and height were measured with a digital scale and stadiometer. A baseline breath sample was obtained, and the subject was given a capsule containing 13C-xylose with 30 ml of water. Three doses were tested (10, 25, and 50 mg), based on extrapolation from adult data (15). Ten children were randomly assigned to each dose. The 13C-xylose was provided by Martek Biosciences (Columbia, MD) in randomly numbered kits, and the clinical investigators were blinded to the dose until the end of the standardization phase. The analysis was performed by a third party (Metabolic Solutions; Merrimack, NH) as an additional measure against sponsor or investigator bias. Those children who could not swallow the capsule were given the contents mixed in 5 ml of water. Serial samples of expired air were collected every 30 minutes for 4 hours, using the Quintron EasySampler (Quintron; Menomonee Falls, WI) and vacuum-sealed tubes.
For the second part of the study, determination of diagnostic utility of 13C-xylose breath test in children with suspected SBBO, six patients between 2 and 12 years old were enrolled in the study. Each patient had presumed SBBO based on symptoms (diarrhea, abdominal pain, and bloating) or clinical improvement with cycled antibiotic treatment. None of the patients (including those on cycled antibiotics) had received antibiotics for at least 3 weeks before the test, except for one, who was receiving trimethoprim-sulfamethoxazole as prophylaxis against urinary tract infections. The patients were drawn from a population of 40 infants and children followed in the Comprehensive Nutritional Care Clinic of Children's Hospital Medical Center for conditions predisposing to SBBO, including immunodeficiency syndrome, short-bowel syndrome, and intestinal pseudoobstruction.
After an overnight fast, the patients reported to the research center for the study. Two of the children had jejunostomy tubes from which 2 ml of fluid was withdrawn into a sterile tube and sent within 30 minutes of collection, stored on ice, to the laboratory for quantitative aerobic and anaerobic bacterial culture. The four remaining children had drainage obtained from a nasoduodenal tube passed under fluoroscopic guidance. Cultures were considered positive when greater than 105 bacteria were identified. The breath test was performed as described in Methods part I, using a 50-mg dose of 13C-xylose (determined from the results of part I). Children with SBBO documented by a positive findings in a duodenal aspirate culture and in a 13C-xylose breath test were treated for 2 weeks with an appropriate antibiotic, chosen based on the results of bacterial sensitivity testing. Follow-up breath tests were performed when possible.
Written, informed consent was obtained from the parents of all subjects. The study was approved by the Institutional Review Board of Children's Hospital Medical Center.
Breath 13CO2 Analysis
The breath samples were analyzed at Metabolic Solutions (Merrimack, NH). Atom percentage of 13CO2[13C/(13C + 12C) × 100] was determined by gas chromatography with mass spectrometry using the Europa Scientific Automated Breath Carbon Analyzer.
For each sample, atom percentage of excess (APE), the percentage of 13C above naturally occurring levels, was calculated by subtracting the baseline percentage of 13C (at time 0) from the percentage of 13C of the sample. Results were expressed as percentage of dose of 13C recovered in the breath per hour (% dose/hr), after correction for variability of endogenous CO2 production caused by differences in body surface area (16).
Aerobic and anaerobic cultures of duodenal aspirates were performed in the microbiology laboratory of the Children's Hospital Medical Center. In brief, at least one milliliter of duodenal drainage was collected and stored on ice. Varying dilutions of aliquots were placed in Mueller-Hinton broth. Thereafter, aliquots of the broth dilutions were plated on TSA II 5% SB Blood Agar plates and incubated either in CO2 or in anaerobic conditions at 35°C for 3 days. Colony counts were assessed at 24 hours and reported as colony-forming units (CFU) per milliliter. Identification of organisms was performed utilizing standard techniques in the microbiology laboratory.
Demographics of the three control groups (and the patients) were compared using Kruskal-Wallis one-way analysis of variance (ANOVA; nonparametric). For each dose (10, 25, and 50 mg), group distribution of percentage of 13C dose/hr at each time point were calculated. Nonnormally distributed data were reported as median (range). Data were plotted as percentage of dose/hr versus time. The dose chosen for part II was that which produced the highest median peak percentage of 13C dose/hr with the least variance.
Analysis of data from part II
Results of the breath test were reported as described in the methods for Part I. Breath tests were considered abnormal if one or more of the values exceeded the corresponding maximum values of the control subjects. The test was subjected to sensitivity and specificity analysis.
The characteristics of the control group are summarized in Table 1. One of the children in group 2 was unable to complete the breath test (for lack of cooperation), and therefore was excluded from the analysis. There were no significant differences among the groups.
The results of the breath test studies are shown in Fig. 1. The 50-mg dose was determined to be the optimal dose because it had the smallest variation for each period, and it produced the highest median peak percentage dose/hr (4.01%; range 2.45-4.53%) at 2.5 hours.
The diagnoses and clinical features of the six patients with presumed SBBO are presented in Table 2. Two of the patients (nos. 4 and 6) were tested twice, with a period of at least 2 months between the two tests. Only one (no. 5) was available for a repeat breath test after treatment for SBBO.
The results of the breath tests are shown in Fig. 2. Two of the patients with a history of small bowel resection (nos. 2 and 5a) had positive cultures and peak excretion in breath above the normal range at 2.5 hours after ingestion of 13C-xylose. After treatment of SBBO, the one patient who was available for repeat testing (no. 5b) had a negative culture and a negative breath test.
The three patients with motility disturbances secondary to idiopathic pseudoobstruction or gastroschisis had concurrently positive cultures and abnormal breath tests, but the timing of peak excretion in breath varied. When the 13C-xylose was given through the gastrostomy tube (in nos. 3 and 4a), the percentage of 13C dose/hr in breath did not rise above the normal range until 3 and 4 hours after ingestion of the xylose, respectively. When the xylose was given through a feeding jejunostomy (nos. 1 and 4b), the percentage of 13C dose/hr in breath of one patients (no. 1) peaked above the normal range at 1.5 hours, and returned to normal by 2 hours after ingestion of the isotope. In the other patient (no. 4b), percentage of 13C dose/hr in breath reached a peak at 2.5 hrs after administration.
The patient with immunodeficiency and diarrhea (no. 6a) was tested on two different occasions. The first time, she had a negative bacterial culture finding and a percentage of 13C dose/hr in breath curve within the normal range. The second time (no. 6b), she again had a negative bacterial culture finding; however, her breath test result was abnormal.
The current study provides the basis for future work defining the value of the 13C-xylose breath test for the diagnosis of bacterial overgrowth in susceptible children. We have demonstrated that the optimal dose of 13C-xylose for testing children ages 3 to 12 years is 50 mg. Using the 50-mg dose, we then studied the breath test results in a limited number of patients with short bowel syndrome, immunodeficiency-associated diarrhea, and motility disorders. We found that all patients with bacterial overgrowth, defined by standard culture criteria, had positive breath test results (100% sensitivity). Of the three patients with negative cultures, two had negative results and one had a positive result (67% specificity). In addition, we demonstrated normalization of both duodenal culture and breath test findings after antibiotic treatment of SBBO in one patient.
Our study results confirm some of the potential problems observed with the 14C-xylose breath test in adults (12-14). Intersubject variability in the rate of xylose delivery to fermentative bacteria was shown in our patients with motility disorders. Two of our patients with delayed gastric emptying and slow transit had delayed peak 13C excretion in breath (3 hours or more) when compared with control subjects. These results are consistent with those of studies of the 14C-xylose breath test in adults with dysmotility (13,14). This problem was avoided in subsequent patients with motility disorders by administering the xylose through a feeding jejunostomy instead of into the stomach. Despite the potential for false negative breath test results caused by delayed peak 13CO2 excretion in breath, our patients with motility disorders had excessive 13CO2 excretion within the 4-hour testing period.
We also had one patient with a negative culture findings but a positive breath test result. It is possible that we may have missed a localized region of bacterial overgrowth (a dilated loop of bowel, for example) with our method of obtaining small-bowel fluid for culture. Although checking for the presence of products of bacterial metabolism (including bile acids, fatty acids, and other products) would add additional evidence of the presence or absence of bacterial overgrowth, the potential for false negative and false positive findings exists for the methods as well.
There are other potential problems with the 13C-xylose breath test that we did not encounter. Metabolism after absorption of 14C-xylose contributes to a rise in breath 14CO2 in control subjects; therefore, it has been suggested that patients with malabsorption may have falsely negative results of breath tests because of decreased absorption of the 14C-xylose (14). This has not been confirmed by other investigators. In patients with rapid transit, secondary to gastric or intestinal resection, an early peak of 13CO2 excretion in breath may result from early delivery of xylose to the colon where fermentation normally takes place, thereby creating a false positive breath test result (12,13).
Unfortunately, the sample size of this study is small, and the study had to be terminated because of the unavailability of isotope. Larger numbers of subjects should be studied using a dose of 50 mg 13C-xylose to provide more definitive data on which to base conclusions regarding the utility of this test. Despite these limitations, our preliminary results suggest that the 13C-xylose breath test appears to be a reliable, noninvasive method for diagnosing bacterial overgrowth.
Acknowledgments: Supported in part by U.S. Public Health Service Grant #M01 RR-08084 from the General Clinical Research Centers Program, National Center for Research Resources, National Institutes of Health.
The authors thank the parents and study children for their participation and the Children's Hospital Medical Center, General Clinical Research Center nurses, Kim Klotz, RN and Gerry Hennies, RN for their assistance in subject recruitment and in data collection.
1. Hoffbrand AV, Tabaquchali S, Mollin DL. High serum folate levels in intestinal blind-loop syndrome. Lancet 1996;1:1339-42.
2. Aarbakke J, Schjonby H. Value of urinary simple phenol and indican determinations in the diagnosis of the stagnant loop syndrome. Scan J Gastroenterology 1976;11:409-14.
3. Kocoshis SA, Schletewitz K, Lovelace G, and Laine RA. Duodenal bile acids among children: Keto derivatives and aerobic small-bowel bacterial overgrowth. J Pediatr Gastroenterol Nutr 1987;6:686-96.
4. Bolt MJG, Stellaard F, Sitrin MD, Paumgartner G. Serum unconjugated bile acids in patients with small-bowel bacterial overgrowth. Clin Chim Acta 1989;181:87-102.
5. Perman JA, Modler S, Barr RG, Rosenthal P. Fasting breath hydrogen concentration: Normal values and clinical application. Gastroenterology 1984;87:1358-63.
6. King CE, Toskes PP. Comparison of the 1-gram [14C]xylose, 10-gram lactose-H2, and 80-gram glucose-H2 breath test in patients with small intestinal bacterial overgrowth. Gastroenterology 1986;91:1447-51.
7. Kerlin P, Wong L. Breath hydrogen testing in bacterial overgrowth of the small intestine. Gastroenterology 1988;95;982-8.
8. Corazza GR, Menozzi MG, Strocchi A, et al. The diagnosis of small-bowel bacterial overgrowth. Reliability of jejunal culture and inadequacy of breath hydrogen testing. Gastroenterology 1990;98:302-9.
9. Khin-Maung-U, Tin-Aye, Ku-Tin-Myint, et al. In vitro hydrogen production by enteric bacteria cultured from children with small-bowel bacterial overgrowth. J Pediatr Gastroenterol Nutr 1992;14:192-7.
10. King CE, Toskes PP, Guilarte TR, et al. Comparison of the one-gram d-[14C]xylose breath test to the [14C] bile acid breath test in patients with small-intestinal bacterial overgrowth. Dig Dis Sci 1980;25:53-8.
11. Schneider A, Novis B, Chen V, et al. Value of the 14C-D-xylose breath test in patients with intestinal bacterial overgrowth. Digestion 1985;32:86-91.
12. Rumessen JJ, Gudmand-Hoyer E, Bachmann E, Justeesen T. Diagnosis of bacterial overgrowth of the small intestine. Scand J Gastroenterol 1985;20:1267-75.
13. Valdovinos MA, Camilleri M, Thomforde GM. Reduced accuracy of 14C-D-xylose breath test for detecting bacterial overgrowth in gastrointestinal motility disorders. Scan J Gastroenterol 1993;28:963-8.
14. Riordan SM, McIver CJ, Duncombe VM, Bolin TD, Thomas MC. Factors influencing the 1-g 14C-D-xylose breath test for bacterial overgrowth. Am J Gastro 1995;1455-60.
15. Lim H, Wagner DA, Toskes PP. A 13C-xylose breath test for bacterial overgrowth. Gastroenterology 1993;104:A259.
16. Boutton TW. Tracer studies with 13C-enriched substrates: Humans and large animals. In: Coleman DC, Fry B, eds. Carbon Isotope Techniques. San Diego: Academic Press, 1991;219-42