Thirty-one habitual coffee drinkers in Ireland gave up coffee for two weeks so scientists could watch what happened to their gut bacteria. Then they started again, half on caffeinated coffee and half on decaf, without knowing which.
The bacteria moved both times, and the caffeine turned out not to be the main driver.
The study, published in Nature Communications by researchers at APC Microbiome Ireland and University College Cork with John Cryan as corresponding author, is one of the few in this field to follow the same people through removal and reintroduction rather than simply comparing drinkers to non-drinkers.
One disclosure belongs up front: the work was sponsored by the Institute for Scientific Information on Coffee, an organization funded by the coffee industry. That does not invalidate the data, which is registered, blinded and openly published, but it is context readers deserve when the conclusions lean favorably.
The Design That Makes It Unusual
Sixty-two healthy adults aged 30 to 50 took part, recruited in Ireland between September 2021 and January 2023 and split evenly between people who never drink coffee and people who habitually drink three to five cups a day. Both groups gave stool, blood, urine and saliva samples and completed cognitive testing at baseline.
Then only the coffee drinkers continued. They abstained completely for 14 days, after which they were block-randomized in a double-blind, parallel design to three weeks of either caffeinated or decaffeinated instant coffee, four sachets daily at 1.8 grams each. Food intake was tracked with weighed seven-day diaries at each stage, and no meaningful dietary differences emerged between groups. The trials were registered as NCT05927038 and NCT05927103.
Specific Species, Not Overall Diversity
Coffee drinkers carried higher levels of Cryptobacterium curtum, two Eggerthella strains and a Firmicutes strain. Non-drinkers carried more Veillonella parvula, a related Veillonella strain and Haemophilus parainfluenzae.
During the two-week abstinence, Cryptobacterium curtum dropped across three separate timepoints. Two other coffee-associated strains returned to levels resembling those of people who never drink coffee.
The pattern is narrower than the usual microbiome headline. Alpha diversity differed between drinkers and non-drinkers but did not change within the coffee drinkers across abstinence or reintroduction. Coffee appears to act on particular species rather than on overall richness.
An odd detail: several of the differentially abundant species normally live in the mouth. The authors flag the oral cavity as a possible reservoir feeding the gut and suggest it deserves separate study.
Fecal chemistry shifted too. Coffee drinkers had lower levels of indole-3-propionic acid, indole-3-carboxaldehyde and the neurotransmitter GABA, compounds with plausible relevance to gut barrier function and mood.
Decaf Did Much of What Caffeine Did
When coffee came back, every one of the seven strains that had differed between groups shifted again, and it happened regardless of caffeine content.
That aligns with the larger picture from a 2024 Nature Microbiology analysis of 22,867 participants integrated with public data from 211 cohorts totalling 54,198 samples, which found the coffee signal highly reproducible and driven largely by the species Lawsonibacter asaccharolyticus. Decaffeinated drinkers showed the same enrichment, and the authors pointed toward quinic acid and its derivatives rather than caffeine. Coffee's effects on other metabolic markers, including blood sugar control, have also been linked to microbial pathways.
The behavioral results were mixed and should not be oversold. Coffee drinkers scored higher on impulsivity and emotional reactivity than non-drinkers, and both measures fell during abstinence. Non-drinkers performed better on memory. Caffeinated coffee reduced anxiety and psychological distress; decaffeinated coffee was associated with gains in memory, sleep quality and physical activity. Some of that may reflect practice effects on repeated cognitive tests, which the authors acknowledge.
Cryan has framed the takeaway broadly, noting that "coffee is more than just caffeine". Notably, the paper also states that non-drinkers appeared to benefit from lower impulsivity, better cognitive stability and more stable blood pressure, a conclusion that sits awkwardly alongside the coffee-positive headlines.
Sixty-Two People, and No Dietary Advice Follows
The caveats matter more than usual here.
The sample was small, recruited entirely in Ireland, and too homogeneous to analyze by ethnicity. The study did not directly measure stool transit time, which coffee is known to affect and which independently shapes microbiome readings. The non-drinkers had given up small amounts of caffeine before baseline. And the authors state the study may not have been sufficiently powered to detect small to medium effects beyond its primary microbiome hypothesis.
More fundamentally, the functional meaning of these bacterial shifts is unknown. Cryptobacterium curtum is mostly studied in the context of oral health. The role of the Eggerthella strains is poorly characterized. Finding that coffee moves them is not the same as knowing whether it matters.
Nothing here supports starting or stopping coffee for gut health. What it does support is a more precise question than the one usually asked: not whether coffee is good or bad, but which of its hundreds of compounds the microbes are actually responding to.
Key Questions Answered
What did the study do differently?
It followed the same people through coffee drinking, two weeks of complete abstinence, and three weeks of blinded reintroduction with either caffeinated or decaffeinated coffee.
Which bacteria changed?
Coffee drinkers had more Cryptobacterium curtum, Eggerthella, and Firmicutes strains; non-drinkers had more Veillonella and Haemophilus parainfluenzae. Levels shifted during abstinence and again on reintroduction.
Is caffeine responsible?
Largely not. Microbial changes on reintroduction occurred with both caffeinated and decaffeinated coffee, pointing toward polyphenols and related compounds.
Did overall gut diversity change?
No. Diversity differed between drinkers and non-drinkers but did not shift within individuals across abstinence and reintroduction, suggesting coffee acts on specific species.
Who funded the research?
The study was sponsored by the Institute for Scientific Information on Coffee, a coffee-industry-funded organization, a relationship readers should weigh alongside the findings.
Does this mean coffee is good for the gut?
The study cannot answer that. The health significance of the specific bacteria involved is not established.
