SIBO Breath Testing: Choosing and Reading the Right Test

3 hours ago

6 min read

A negative SIBO breath test in a symptomatic patient is more often the wrong substrate, an incomplete gas panel, or a rise read outside the diagnostic window than a true negative. Here is how to choose the test and read what comes back.

Substrate: lactulose or glucose

The two substrates don't test the same bowel, so the one you order decides what you are able to find.

Glucose is absorbed in the proximal small intestine — roughly the first two to three feet. Bacteria living past that point never meet the substrate, so a glucose test interrogates the proximal bowel and nothing beyond it. A positive is therefore highly credible. A negative says very little about distal overgrowth.

Lactulose is not absorbed by humans at all. It travels the length of the small intestine and on into the colon, so it can reach overgrowth that glucose never sees. The price is the colon itself: once lactulose arrives at the cecum, colonic bacteria ferment it and produce a large rise. In a patient with fast transit, that colonic rise lands inside the test window and reads as a positive.

	Lactulose	Glucose
Reaches	Whole small intestine, then colon	Proximal small intestine only
Sensitivity	Higher — finds distal overgrowth	Lower — misses distal overgrowth
Specificity	Lower — colonic fermentation reads as positive	Higher — a positive is usually real
Fails when	Transit is fast	The overgrowth is distal

Most practitioners use lactulose as the default, which is reasonable. Glucose earns its place in two situations: confirming a borderline lactulose positive, and testing someone whose transit you expect to be fast — prior bowel resection, altered post-surgical anatomy, a firmly diarrhea-predominant picture.

Hydrogen, methane and hydrogen sulfide: reading all three

Methane does not behave like hydrogen, and a two-gas panel cannot see the third gas at all. Both facts produce negatives in patients who are not negative.

Gas	Produced by	Read as positive when	Usually presents as
Hydrogen	Bacteria	A rise of 20 ppm or more over baseline within 90 minutes	Diarrhea or mixed
Methane	Archaea, chiefly Methanobrevibacter smithii	10 ppm or more at any single timepoint, baseline included	Constipation
Hydrogen sulfide	Sulfate-reducing bacteria	Only measurable on a three-gas panel	Diarrhea, sulfur intolerance

The hydrogen and methane thresholds come from the 2017 North American Consensus, and the difference between them is the part that gets missed.

Methane is not a rise, it is a level. A patient whose methane sits at 18 ppm from baseline to the end of the test, perfectly flat, is positive. Nothing rose. If you were trained to look for a climb, that result reads as normal, and it is the single most common misread on a breath test. It is also why the preferred term is now intestinal methanogen overgrowth, or IMO — archaea are not bacteria, and the overgrowth is not confined to the small bowel, so calling it methane SIBO describes neither the organism nor the location.

A two-gas panel is blind to hydrogen sulfide. If your lab reports hydrogen and methane only, a patient whose overgrowth is hydrogen-sulfide predominant will come back negative every time, no matter how many times you repeat it.

The 90-minute window: when a rise is the colon

A hydrogen rise after 90 minutes is usually the substrate arriving in the colon, not bacteria in the small bowel.

That is the whole reason the window exists. Lactulose is designed to reach the colon, and when it does, colonic bacteria ferment it and hydrogen climbs steeply. That climb is normal physiology. Counting it as a positive turns a healthy colon into a diagnosis.

Plenty of labs still run 120 or 180 minutes and flag any rise in that whole stretch. If you read the report's own verdict rather than the curve, you will treat patients who do not have small bowel overgrowth — and then wonder why the protocol did nothing.

The honest exception is transit. The 90-minute rule assumes it is normal. In a markedly constipated patient, particularly one with elevated methane, the substrate may genuinely still be in the small intestine past 90 minutes, and a rise at 100 or 110 minutes can be real. So the window is a strong prior, not a law. What it should do is stop you accepting a late rise uncritically, and send you back to the patient's bowel habit before you decide.

The double-peak pattern is worth knowing about and not worth relying on. An early small-bowel peak followed by a later colonic one is taught widely, but it appears inconsistently and its absence tells you nothing.

A flat SIBO breath test in a symptomatic patient

A hydrogen curve that never moves is a finding, not an absence of one.

Hydrogen is the shared currency of the gut's microbial economy. Other organisms consume it. Methanogens use four parts hydrogen and one of carbon dioxide to make methane, which is why high methane sits alongside low or flat hydrogen — those two results are not in conflict, one explains the other. Sulfate-reducing bacteria do the same thing to produce hydrogen sulfide.

So in a symptomatic patient whose hydrogen is flat and whose methane is normal, something may be eating the hydrogen before it reaches the breath. On a two-gas panel that patient is negative, and will be negative again on the repeat, and again after that. The test cannot see the organism responsible. A three-gas panel can.

Before you land there, rule out the mundane explanations:

Antimicrobials. Antibiotics, herbal antimicrobials or a recent course of anything suppressive within the washout period.
Over-restricted prep. A prep diet taken too far starves the organisms you are trying to provoke.
The test was not followed. Smoking, exercise, sleeping during the test, eating something during it.
It is genuinely negative. Exocrine pancreatic insufficiency, bile acid malabsorption, post-infectious IBS and dysbiosis without overgrowth all produce this symptom picture and none of them show up on a breath test.

The mistake to avoid is treating a flat test as proof that nothing is wrong, and the second mistake is repeating the identical two-gas test and expecting a different answer.

Prep failures that invalidate the result

More breath tests are ruined before the patient breathes than are misread afterwards.

Exact windows differ between labs and the lab's own protocol wins. These are the ones patients get wrong, and they are worth confirming out loud before you interpret anything:

The prep diet in both directions. Too lax and residual fermentable substrate inflates the baseline. Too strict — patients who decide a day of plain chicken is good, so three days must be better — and you starve the organisms you were trying to provoke.
The washout. Antibiotics and herbal antimicrobials need the full window, usually four weeks. A course finished a fortnight ago makes the result uninterpretable, and patients rarely volunteer the herbal ones because they do not think of them as medication.
Bowel prep. A recent colonoscopy clears the field. Testing too soon after one tests the prep, not the patient.
Behaviour during the test. No smoking, no exercise, no sleeping through it. Sleep changes breathing and flattens the curve.

One question answers most of this: ask the patient to walk you through what they actually did, rather than asking whether they followed the instructions. Everyone says yes to the second question.

What the SIBO subtype changes about treatment

The gas does not just label the case, it changes where you start and how long you expect it to take.

Hydrogen-predominant overgrowth is the one most protocols are written for, and it generally behaves. Methanogen overgrowth does not. Archaea are not bacteria, they do not respond to the same agents, and a protocol that clears hydrogen can leave methane untouched — which is why the constipated patient so often reports that treatment did nothing. It usually needs a different combination, more time, and real attention to motility rather than antimicrobials alone. Hydrogen sulfide is different again, and the substrate load matters in a way it does not for the other two.

The part that decides whether any of it holds is the question underneath: why did this patient get overgrowth in the first place? Impaired migrating motor complex, post-infectious IBS, a structural problem, adhesions, an ileocecal valve that is not doing its job, long-term acid suppression. Clear the organisms without answering that, and you are booking the same patient in again in four months. The relapse was not a failure of the antimicrobial.

That is the part breath testing cannot tell you, and it is where the clinical work actually is.

The GHT Academy Gut Guide carries the rest of this — protocols by SIBO subtype with specific supplements, dosing and timelines, root cause analysis for each pattern, GI Map and IBS-Smart interpretation marker by marker, plus video walkthroughs of live interpretation. It is the reference I built for my own team.

Common questions

What methane level counts as positive?

10 ppm or more at any single timepoint, baseline included. It does not have to rise.

How long should a SIBO breath test run?

Collection often continues to 120 or 180 minutes, but the diagnostic window for hydrogen is the first 90. A rise after that is usually the substrate reaching the colon.

Lactulose or glucose?

Lactulose travels the whole small intestine and finds distal overgrowth. Glucose is absorbed proximally and misses it. Lactulose is more sensitive, glucose more specific.

Can a patient have SIBO with a negative breath test?

Yes. A two-gas panel cannot detect hydrogen sulfide at all, and hydrogen-consuming organisms can flatten the hydrogen curve.

What is IMO, and how is it different from SIBO?

Intestinal methanogen overgrowth. The organisms are archaea rather than bacteria, they are not confined to the small bowel, and they do not respond to the same antimicrobials.