Hydrogen Breath Test (SIBO, Lactose, Fructose): How to Prepare and How to Read the Result

Human cells do not produce hydrogen. It arises almost exclusively from the fermentation of carbohydrates by gut bacteria. If the sugar given in the test is absorbed in the small intestine, the bacteria have nothing to ferment and the hydrogen concentration in exhaled air does not change. If it reaches the large intestine (because it is not absorbed) or is fermented earlier by bacteria in the small intestine, the hydrogen concentration in the breath rises, because part of the hydrogen is absorbed into the blood and exhaled by the lungs. It is measured every 15-30 minutes with a special analyzer, often together with methane, which is produced by other microorganisms (archaea).

Several practical consequences follow from this design. First, the test measures bacterial activity rather than absorption directly, so everything that changes bacterial activity or numbers (antibiotics, bowel preparation for colonoscopy, the previous day's fibre, diet) changes the result. Second, in some people bacteria produce little hydrogen or consume it to make methane or hydrogen sulfide, so the test can be falsely negative. According to the European consensus this low hydrogen production affects up to about 20% of tests, and in children hydrogen non-excretion is found in 10-15%. Third, samples for analysis must come from end-expiratory air, and stored samples are stable for about 6 hours at room temperature.

Where does each of these recommendations come from? Antibiotics and bowel cleansing change the composition and metabolism of bacteria and therefore hydrogen production. The guideline notes there are no data on how long this effect lasts, so a safe interval of 1 to 4 weeks was adopted (finally 4 weeks after an antibiotic). Fibre and fermentable carbohydrates from the previous evening can raise baseline hydrogen, which makes it harder to detect the rise after the test sugar; hence the advice of a low-fibre diet before the test, and in constipated people even up to 48 hours of avoiding gas-forming foods (although data are lacking). Smoking and physical exertion (hyperventilation) change gas concentrations in exhaled breath. Rinsing with chlorhexidine prevents abrupt rises in hydrogen coming from oral bacteria.

The guideline does not give a separate washout period for proton pump inhibitors (PPIs) before the test, but requires that the person performing the test knows all gastrointestinal medicines taken in the last 24 hours, and notes that the decision on stopping symptomatic drugs such as laxatives, antidiarrheals and antispasmodics depends on the clinical situation and is for the doctor. It does list hypochlorhydria, including that caused by PPIs, among the risk factors for SIBO. The practical conclusion: do not stop chronic medicines on your own, discuss them with the ordering doctor and report them at the laboratory. Laxatives containing fermentable carbohydrates (lactulose) or fermentable fibre must be treated separately: according to the guideline they need to be stopped and, if necessary, replaced by non-carbohydrate agents such as polyethylene glycol.

The type of sugar given depends on the question being asked. The lactose test checks whether lactose is digested (lactase deficiency), the fructose test whether fructose is absorbed, and tests with glucose or lactulose are used to assess SIBO. Doses and duration differ between documents. The North American Consensus (Rezaie et al., 2017) recommended doses of 25 g for lactose and fructose, 75 g for glucose and 10 g for lactulose. The European guideline (Hammer et al., 2022) gives for adults 25-50 g of lactose for diagnosing malabsorption (25 g for diagnosing intolerance), 20-25 g of fructose, 50 g of glucose in 250 ml of water and 10-20 g of lactulose in 250 ml of water.

The measurement time also differs. For carbohydrate malabsorption tests the recommended duration is 3-5 hours with samples every 30 minutes (intervals up to 60 minutes suffice to assess malabsorption alone, while shorter 10-15-minute intervals are needed if you want to show a temporal link between the rise in hydrogen and symptoms). For SIBO tests the standard according to the European guideline is 120 minutes with samples every 15 minutes. For lactose and fructose the positive threshold is a hydrogen rise of at least 20 ppm over baseline at a single time point, which agrees with the North American Consensus. For SIBO the North American Consensus regarded a rise of at least 20 ppm by 90 minutes of a glucose or lactulose test as positive, whereas according to the European guideline the most commonly used threshold for glucose is 10-12 ppm, and the criteria for diagnosing SIBO from the test are not uniformly confirmed.

Lactose, fructose and sorbitol should not be used as substrates in SIBO diagnosis (Saad and Chey, 2014): they are absorbed or digested in the small intestine and are not suited to detecting bacteria in its lumen. For glucose and lactulose the situation is more complex. Glucose is rapidly absorbed in the duodenum and jejunum, which may limit the test's sensitivity but reduces the rate of false positives; lactulose is not absorbed and reaches the large intestine, so the test is more sensitive but more prone to a false result when transit is fast. The European guideline states that in non-surgical patients, without a parallel transit study, glucose should be preferred because the rate of false positives is lower than with lactulose.

The European guideline stresses that a rise in hydrogen after a sugar alone does not confirm that the patient's symptoms result from malabsorption of that sugar. Combining the breath test with symptom assessment distinguishes four situations: malabsorption with symptoms, malabsorption without symptoms, symptoms without malabsorption, and neither. Only the first corresponds to intolerance in the clinical sense. That is why a validated tool for symptom assessment is recommended, and a good, lasting response to diet may settle the diagnosis of intolerance. Symptoms can also appear without evidence of malabsorption on the test, for example because of visceral hypersensitivity, a nocebo effect or a false-negative result.

False-negative results for carbohydrate malabsorption can occur in people with low hydrogen production, with slow transit from mouth to cecum (when fermentation starts after the test has ended) and with an elevated baseline hydrogen. False positives can result from SIBO or rapid transit. That is why SIBO should be excluded before a carbohydrate malabsorption test, and measuring methane (a rise of at least 10 ppm) is sometimes helpful in people with low hydrogen, although its contribution to sensitivity may be offset by a fall in specificity.

For years the gold standard for diagnosing SIBO was culture of a small-bowel aspirate, but it is invasive, difficult to perform and unstandardized, so in practice breath tests are used (Saad and Chey, 2014). The European guideline allows them until a true gold standard is established and recommends considering them with bloating, discomfort and gas or signs of malabsorption without another diagnosis on endoscopy or imaging, especially with risk factors for SIBO (moderate to high pre-test probability). At the same time it states that the criteria for diagnosing SIBO from the test are not standardized and that the clinical value of a positive result must be judged in light of the pre-test probability.

The main problem is the variability of transit time from mouth to cecum. When transit is rapid (common in IBS), hydrogen appears early because the substrate reached the large intestine, not because bacteria populate the small intestine. The scale of discrepancy can be large: in the meta-analysis by Poon et al. (2022), in patients meeting IBS criteria the frequency of SIBO assessed by breath test was 49% with lactulose but 19% with glucose. The review by Bushyhead and Quigley (2022) writes of problems with the specificity of breath tests that have clouded interpretation and aroused skepticism about the role of SIBO in the broad spectrum of symptoms, including IBS, and the paper by Suárez Terán and Guarner Aguilar (2024) asks outright whether SIBO is clinically over-diagnosed.

Figures from one of the newer studies illustrate the limitations well. In the study by Dahlgren et al. (2025) using a lactulose test and a 20 ppm threshold (with an 80-minute transit criterion), sensitivity for SIBO in patients with intestinal lesions was 77% with a specificity of 88%, but in IBS patients sensitivity fell to 39% with a specificity of 78%. This is one study, so we treat it as an illustration, not final values. The practical conclusion is that the breath test is a useful tool in risk groups with a high probability of SIBO (after bowel surgery, with motility disorders, hypochlorhydria), and much less reliable as a screening test for everyone with bloating.

The opposite belief is equally mistaken: that since the test is imperfect it is worthless. A lactose test with symptom assessment has, according to the European consensus, acceptable sensitivity and specificity and can direct effective therapy, and SIBO tests can help in risk groups. The point is to know what the test cannot tell you.

The European guideline recommends that patients with alarm features (weight loss, blood in the stool, anemia, nocturnal diarrhea, symptoms after age 45-50, a family history of bowel cancer, celiac disease or inflammatory bowel disease) are investigated biochemically, endoscopically and by imaging before a breath test. The breath test is for assessing carbohydrate intolerance in people without organic disease, not a substitute for investigating serious disease. For bloating and chronic diarrhea it is best to start with the steps described in our guides on bloating after meals and on chronic diarrhea.

The hydrogen breath test is a good tool in the hands of someone who knows what to expect from it. Preparation is a dull but crucial stage: 4 weeks after an antibiotic, a day of low-fibre diet, 8 hours of fasting, no cigarette and no exertion. Interpretation must always be combined with symptoms, the test dose and the pre-test probability. The most harm comes from believing that the number in ppm is a verdict: a positive result for SIBO or fructose is sometimes a physiological feature of the gut, and a negative one does not always exclude a problem.