Caffeinated or Decaffeinated: Which Coffee Is Linked to a Lower Risk of Dementia?

Masahiro Kami, M.D., Ph.D.

President, Medical Governance Research Institute, Tokyo, Japan

The health effects of coffee and tea are once again attracting attention. In my previous article, I discussed the possibility that coffee may help suppress arrhythmias. So, are people who regularly drink coffee or tea less likely to develop dementia later in life?

A recently published study offers one clue. A research team from Harvard University and affiliated institutions, including Brigham and Women’s Hospital, published the study in the February 9, 2026 issue of the Journal of the American Medical Association (JAMA). The findings suggest that regularly consuming moderate amounts of caffeinated coffee or tea is associated with a lower risk of developing dementia. Here, I would like to offer my interpretation of the study.

What Did the Study of Healthcare Professionals Find?

First, let us look at the study itself. The researchers used data from two well-known long-term cohort studies in the United States.

One was the Nurses’ Health Study (NHS), which follows female nurses. It began in 1976 and enrolled women between the ages of 30 and 55. The other was the Health Professionals Follow-up Study (HPFS), which follows male health professionals. It began in 1986 and enrolled men between the ages of 40 and 75.

If we were to compare these participants with generations in Japan, most would belong to the generations born before or during World War II. Not only the level of medical care but also diets and attitudes toward health were entirely different from those of today. All of these changed dramatically during the study period. Recognizing these differences across eras is important when interpreting the findings.

For this analysis, the researchers used data from the point when detailed dietary surveys began in each cohort—1980 for the NHS and 1986 for the HPFS—and followed approximately 130,000 participants for more than 40 years.

Every two to four years, participants provided detailed information about their diets. The researchers repeatedly assessed their intake of caffeinated coffee, decaffeinated coffee, tea, and total caffeine, and examined how these were associated with the subsequent development of dementia and changes in cognitive functions such as memory and attention.

The results were fairly clear. People who consumed more caffeinated coffee tended to develop dementia less often, and a similar association was observed for tea.

How Many Cups Are Beneficial?

By contrast, decaffeinated coffee—coffee from which most of the caffeine has been removed—showed no clear association with a reduced risk of dementia. On cognitive tests, people who drank caffeinated coffee or tea also performed slightly better on average.

[Photo caption: Coffee can help wake up the brain. Photo by Toshiki Miyama, Chiyoda Ward, Tokyo, November 28, 2023.]

When it came to the amount consumed, more was not necessarily better. Moderate intake appeared to be most favorable. The strongest association with a lower risk of dementia was observed at approximately two to three cups of coffee per day, one to two cups of tea per day, and a total caffeine intake of around 300 mg.

Why, then, might caffeine have beneficial effects on brain health?

Without going into too much detail, caffeine acts on mechanisms in the brain involved in arousal and attention, and studies suggest that it may also reduce inflammation and oxidative stress. Basic research has pointed to possible effects on the accumulation of abnormal proteins associated with Alzheimer’s disease, such as amyloid beta and tau, as well as on inflammatory responses in the brain. Caffeine may also improve the brain’s blood flow and metabolic environment through effects on vascular function and glucose metabolism.

The findings of this study are consistent with such basic research. The fact that no clear association was observed with decaffeinated coffee suggests that caffeine itself may play an important role.

An “Unhealthy Indulgence”?

There is, however, an important point to keep in mind when interpreting these results. The cohorts in this study were followed from the 1980s through the 2020s, a period during which public perceptions of coffee’s health effects changed dramatically.

A typical example is a study published in the New England Journal of Medicine in October 1986 by a research team centered at Johns Hopkins University. The cohort study examined coffee consumption and the risk of coronary heart disease, including myocardial infarction, angina, and sudden death. It followed 1,130 male medical students in the United States for 19 to 35 years.

In that study, people who drank five or more cups of coffee per day had approximately 2.5 times the risk of coronary heart disease compared with those who did not drink coffee at all.

Many similar studies were published at the time. There was a broad international consensus that coffee was an “unhealthy indulgence.”

Lifestyle Changes Matter, Too

Beginning in the 2000s, perceptions of coffee’s health effects changed dramatically. Researchers increasingly suggested that moderate coffee consumption might actually have health benefits.

In other words, the public perception of coffee—and people’s behavior toward it—was quite different in the first and second halves of the follow-up period in the study discussed here. This is an important consideration when interpreting the results.

Even more importantly, management of lifestyle-related diseases, which are also risk factors for dementia, improved substantially during the same period.

After World War II, approaches to lifestyle-related diseases changed greatly. In the case of hypertension, until the 1950s and 1960s there was little recognition that it was a condition requiring treatment. Some even viewed high blood pressure positively, as a sign of “strength.”

This began to change during the 1950s and 1960s. Studies such as the Framingham Heart Study in the United States showed that hypertension was a risk factor for stroke and myocardial infarction. Early antihypertensive drugs such as thiazide diuretics were developed around this time, followed in the 1960s and 1970s by various medications including beta-blockers, calcium channel blockers, and angiotensin-converting enzyme (ACE) inhibitors.

These medications became widely used during the 1980s and 1990s, corresponding to the first half of the observation period in the present study.

In the latter half of the study, as the 21st century began, researchers increasingly discussed the relationship between hypertension and dementia. Numerous cohort studies suggested that hypertension in middle age may increase the risk of dementia later in life.

Through the accumulation of such clinical research, today’s approaches to managing lifestyle-related diseases were gradually established. In historical terms, however, only a relatively short time has passed since these approaches became standard.

The participants in the present study were nurses and other health professionals, a population likely to be more health-conscious than the general public. After the 2000s, as evidence about the safety and potential benefits of coffee became more widely known, these individuals may have felt more comfortable increasing their coffee consumption.

At the same time, many participants were entering middle age, when lifestyle-related diseases become a greater concern, and they were likely managing their blood pressure, cholesterol, and blood glucose more carefully.

Given these circumstances, it is difficult to completely distinguish between two possibilities: did drinking coffee reduce the risk of dementia, or were people who managed their health well also more likely to drink coffee?

The possibility of confounding therefore needs to be considered.

When Did Participants Start Drinking Coffee?

Another issue is timing—when in their lives the participants were drinking coffee.

If coffee or caffeine affects the early stages of neurodegeneration, including the accumulation of amyloid and tau, exposure decades before symptoms appear—that is, before middle age—may be particularly important.

In this study, long-term intake was assessed using a “cumulative average.” As a result, differences in when during a person’s life they consumed larger amounts were averaged out. The study may therefore not have fully captured exposure during the “early” stages when neurodegeneration begins.

This “early” period corresponds to the first half of the study, particularly around 1986 to 2000.

By then, the importance of managing lifestyle-related diseases such as hypertension and hyperlipidemia was becoming increasingly recognized. At the same time, coffee and caffeine were still widely regarded as unhealthy indulgences.

If the analysis were restricted to this period, the kind of confounding that becomes problematic in the latter half of the study would not arise in the same way. Unfortunately, the research team did not present an analysis limited to this period in the published paper.

JAMA encourages readers who have questions about published papers to submit letters to the journal. Letters that the editors consider useful to a broader readership may be published both in print and online.

Together with Dr. Makoto Kosaka of Navitas Clinic, I wrote a letter and submitted it to the journal. I look forward to seeing how the editors respond.

There are several other limitations to the study.

Because it is an observational study, it does not directly establish a causal relationship. Dietary information was self-reported and is therefore subject to measurement error. Dementia diagnoses were based on physicians’ records and death records, so mild cases may have been missed and some variation in classification is unavoidable. In addition, because the participants were predominantly health professionals, caution is needed when applying the findings to the general population.

A Doctor’s Recommended Way to Drink Coffee

Even after taking these limitations into account, the capacity of U.S. medical research to conduct a cohort study on this scale is truly remarkable.

I believe this study reflects the “will” of the United States as a nation.

The research was funded by the U.S. National Institutes of Health (NIH) and conducted by a team at Harvard University, one of the country’s leading research institutions. In an aging United States, they are likely thinking seriously about dementia as a major national challenge and about how people can improve their lifestyles.

It is also worth noting that the study received no funding from companies in the coffee industry or other commercial sources. In this respect, its independence and neutrality are also supported.

So, what should we take away from this study?

As a practical lesson for everyday life, a reasonable approach is simply to enjoy coffee or tea in moderate amounts without overdoing it.

For people who already have a habit of drinking coffee or tea, a few cups a day may have a positive effect on long-term brain health. On the other hand, people who have difficulty sleeping or are prone to palpitations should not force themselves to drink it and should choose an amount that suits their individual tolerance.

Coffee and tea are not “medicine.” But when combined with basic healthy habits such as exercise, smoking cessation, and the management of blood pressure, blood glucose, and lipid levels, they may be one small part of an overall effort to reduce the future risk of dementia.

Reference:“Coffee and Tea Intake, Dementia Risk

This article was originally published in Mainichi Medical on March 10, 2026.