A Cork RCT published in Nature Communications on April 21 reports that caffeinated and decaffeinated coffee shift the gut microbiome in similar ways — the RCT that hardens the paper's observational write-up yesterday. [1] The trial design is the headline finding for a field that usually relies on observational cohorts. Sixty-two adults — thirty-one regular coffee drinkers and thirty-one non-drinkers — were registered through ClinicalTrials.gov (NCT05927038 and NCT05927103). The drinkers stopped all coffee for fourteen days. Then a blinded randomization assigned them either caffeinated (sixteen subjects) or decaffeinated (fifteen subjects) for twenty-one days at four standardized sachets daily. [2]

The shifts in gut composition are similar across the two arms. Coffee drinkers showed higher relative abundance of Cryptobacterium curtum (associated with gastric acid secretion), Eggerthella species (bile-acid synthesis), Lawsonibacter, and several Firmicutes. [1] The abstinence period drove these populations back toward non-drinker baseline; reintroduction restored them, in both arms. The likely active compounds are chlorogenic acids — polyphenols that survive decaffeination — and the nine metabolites the integrated model identified include theophylline, caffeine, and several phenolic acids that survive decaf. [1]

The behavioral measures matter only with caveats. Coffee drinkers scored higher on impulsivity and emotional reactivity at baseline than non-drinkers; both scores dropped during abstinence and partially returned during reintroduction. Caffeinated coffee specifically tracked with reduced anxiety and improved vigilance; decaf specifically tracked with improvements in learning, memory, sleep quality, and physical activity. [3] The authors caution that some cognitive gains may reflect task repetition.

What separates this paper from the wellness-press wave around it is the registered RCT design. Most microbiome-mood research lives in observational cohorts that cannot establish causation. The UCC team's pre-registered abstinence-and-reintroduction protocol — with stool, urine, blood, and psychometric measures across three phases — is closer to a CONSORT-grade pharmaceutical trial than to a typical food-microbiome study. [3] The clinical-trial register IDs are the receipt.

The practical takeaway is restrained. Coffee is not a treatment. The trial is small, short, and concentrated in one Irish cohort. But the methodological gap the paper closes is real: caffeine is not the only mechanism by which coffee touches the gut-brain axis, and decaf is doing more chemical work than most clinicians have been telling patients.

-- NORA WHITFIELD, Chicago