Researchers tracked more than 131,000 adults across two long-running Harvard cohorts for up to 43 years and found that those who drank the most caffeinated coffee had roughly 18 percent lower risk of developing dementia. The signal was strongest at two to three cups a day, and caffeinated tea showed a similar pattern. The findings land as global dementia diagnoses continue to rise and millions of people look for straightforward daily habits that might protect brain health.
Why 43 years of dietary data sharpen the coffee–dementia question
Most prior studies linking coffee to cognitive health relied on a single snapshot of what people ate and drank. This analysis drew on repeated diet measures from the Nurses’ Health Study and Health Professionals Follow-up Study, two cohorts that collected food-frequency questionnaires every few years over decades. That design matters because it captures people who changed their coffee habits over time and separates them from those whose intake stayed steady, a methodological strength highlighted in the Harvard Gazette coverage of the work.
The distinction between stable and shifting consumption is central to judging how much weight the 18 percent figure deserves. A person who reported drinking three cups a day in 1980 and still reported three cups in 2005 offers stronger evidence of a real exposure than someone whose intake bounced between zero and five cups across surveys. By using cumulative average intake rather than a one-time measure, the researchers reduced the noise that plagues shorter studies. The result is the longest prospective look at caffeinated beverages and dementia risk published to date from these cohorts.
One hypothesis worth testing against these data is whether the protective association is stronger for ground coffee than for instant coffee, particularly among participants whose records show stable intake over at least two decades. Earlier work using the UK Biobank found heterogeneous dementia risks by coffee type, with ground, instant, and decaffeinated varieties showing different effect sizes. The Nurses’ Health Study and Health Professionals Follow-up Study, however, did not separate coffee preparation methods in their questionnaires the same way the UK Biobank did, which limits direct comparison and leaves open the question of whether brewing method or bean processing meaningfully alters risk.
Long follow-up also helps address another challenge: reverse causation. Subtle cognitive changes can precede a dementia diagnosis by years and might themselves alter a person’s appetite for coffee or tolerance for caffeine. By tracking diet decades before most diagnoses and updating exposure repeatedly, the investigators were better positioned to distinguish long-standing habits from changes that could be early signs of disease.
How 131,000 participants and telephone-based testing produced the 18 percent figure
The study enrolled more than 131,000 nurses and health professionals and examined outcomes including both clinically diagnosed dementia and subjective cognitive decline, a self-reported measure that often precedes formal diagnosis. According to a detailed summary from Harvard Health Publishing, participants who consumed the highest amounts of caffeinated coffee showed approximately 18 percent lower dementia risk compared with the lowest consumers. Decaffeinated coffee showed a weaker link, suggesting caffeine itself may play a role alongside the hundreds of bioactive compounds in coffee beans.
Objective cognitive status was assessed by telephone using the Telephone Interview for Cognitive Status, or TICS, and a modified version known as TICSm. Validation research published in peer-reviewed journals has shown that these tools perform reliably for screening mild cognitive impairment and dementia when compared with in-person neuropsychological testing. Telephone-administered cognitive batteries have been shown to produce results comparable in reliability to face-to-face assessment, which made large-scale follow-up feasible across tens of thousands of geographically dispersed health professionals and nurses.
The researchers also compared caffeinated coffee and tea against decaf, finding that caffeinated tea drinkers showed a pattern similar to coffee drinkers. That parallel result strengthens the case that caffeine, rather than some other compound unique to coffee, contributes to the observed association. Still, coffee and tea contain overlapping but distinct sets of polyphenols and antioxidants, so isolating the active ingredient from observational data alone is not possible, and the investigators have been cautious not to overstate caffeine’s role.
Importantly, the 18 percent figure represents a relative difference between high and low intake groups, not an absolute guarantee of protection. Even among heavy coffee drinkers, some participants developed dementia, and many low-intake participants remained cognitively healthy. The cohorts’ composition-largely health professionals-also means the findings may not translate perfectly to populations with different lifestyles, comorbidities, or access to medical care.
Confounding, causation, and what coffee drinkers still cannot know
The lead author has been clear that these are observational data and that cause and effect cannot be proven from this kind of study. In interviews summarized by reporting in The Guardian, the research team emphasized that people who drink two to three cups of coffee a day may also exercise more, sleep differently, or carry genetic profiles that independently lower dementia risk. The researchers adjusted for many known confounders, including age, sex, education, smoking, and certain cardiovascular risk factors, but residual confounding from unmeasured variables like detailed sleep quality and long-term physical activity patterns has not been fully addressed in publicly available summaries.
Genetic risk stratification adds another layer of complexity. The UK Biobank study examined how coffee interacted with genetic predisposition to dementia, but the new Harvard analysis has not yet detailed how genetic-risk strata were defined or whether the 18 percent reduction held equally across high-risk and low-risk groups. Without that information, it remains unclear whether coffee consumption might be especially relevant for people with strong family histories of dementia or whether the association is broadly similar across genetic backgrounds.
There is also the question of dose. While the lowest dementia risk clustered around two to three cups per day, higher intakes did not appear to confer proportionally greater benefit and may carry downsides such as insomnia, palpitations, or anxiety in susceptible individuals. The available summaries do not suggest a sharp harm threshold at higher doses, but they do not rule out subtle risks either, particularly for people with cardiovascular disease, arrhythmias, or pregnancy-related concerns.
For now, the most defensible interpretation is that moderate caffeinated coffee or tea intake is compatible with, and may modestly support, healthy brain aging in many adults, especially when embedded in an overall pattern that includes regular exercise, a balanced diet, blood pressure control, and cognitive engagement. Coffee alone is unlikely to offset powerful drivers of dementia such as uncontrolled diabetes, smoking, or severe hypertension.
What this means for everyday coffee and tea drinkers
For people who already enjoy coffee or tea, the findings offer some reassurance. Maintaining one to three cups of caffeinated coffee or tea per day appears reasonable for most healthy adults and may be one component of a brain-friendly lifestyle. Those who are sensitive to caffeine, experience sleep disruption, or have been advised to limit stimulants for medical reasons should not feel compelled to increase intake based on these data alone.
Non–coffee drinkers face a different calculus. Starting a daily coffee habit purely for potential dementia protection is not clearly warranted, especially in the absence of randomized trials. Individuals who dislike coffee, have gastrointestinal issues, or are prone to anxiety might reasonably focus instead on better-established interventions such as physical activity, social connection, and management of cardiovascular risk factors.
Future research will need to clarify whether specific coffee types, preparation methods, or genetic profiles modify the observed association and to explore biological mechanisms linking moderate caffeine exposure to brain resilience. Until then, the Harvard cohorts’ four decades of data suggest that, for many adults, a morning cup or two can be enjoyed not only for its taste and alertness boost, but also as a potentially helpful-if not definitive-ally in long-term cognitive health.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
