Harvard-led data suggest 2–3 cups of real coffee may protect cognition while decaf falls flat
Two to three cups of caffeinated coffee a day may cut dementia risk by more than half in older adults, with similar benefits for tea drinkers, according to recent research.
The study, conducted by the Journal of the American Medical Association, tracked intake of caffeinated coffee, decaffeinated coffee and tea, and collected dietary data every two to four years, then looked at both objective measures of dementia and subjective reports of cognitive decline.
Forbes reported that daily caffeinated coffee consumption correlated with better cognitive function.
The development of dementia decreased by more than half, while subjective cognitive decline dropped from 9.5 percent to 7.8 percent.
Tea consumers reported similar benefits to caffeinated coffee drinkers.
Those who consumed decaffeinated coffee did not see reductions in dementia risk or subjective cognitive decline.
As per Forbes, the benefits of caffeinated coffee peaked at about two to three cups per day and then plateaued with higher daily intake.
The study populations included health professionals only, which the authors noted could affect responses and outcomes data.
They also stressed that “association as opposed to causation remain hard to differentiate,” even with large population studies over several decades.
Cleveland Clinic reported on a related analysis involving more than 130,000 people.
Dylan Wint, MD, a neurologist at Cleveland Clinic, said the findings were consistent with earlier research showing that coffee and tea drinkers have a lower risk of dementia and cognitive decline.
He said the study “was also careful to include” people who drank decaffeinated coffee so researchers could compare them with those who drank caffeinated coffee.
He noted that those who drank two to three cups of caffeinated coffee per day saw a lower risk for developing dementia, with similar results for those who drank one to two cups of caffeinated tea.
According to Cleveland Clinic, Wint said the findings suggest caffeine, rather than coffee or tea themselves, may be responsible for the effect.
“Caffeine might reduce the accumulation of amyloid-beta, which is a toxic protein that can be found in the plaques that are associated with Alzheimer’s disease. There are also chemical effects of caffeine that may be neuroprotective,” he said.
Forbes noted that Alzheimer’s Disease is the most common cause of dementia in the US, usually beginning with subjective cognitive decline and progressing to deficits in memory and activities of daily living.
More than 6 million people in the country are living with Alzheimer’s, and this is likely to rise to 13 million over the next two decades.
Forbes also pointed out that engaging in learning activities, exercise, social connections, getting adequate sleep and minimising alcohol intake and use of sedatives have all been linked to reduced likelihood of developing dementia later in life.
To move beyond association, Forbes highlighted an ongoing placebo-controlled multi-centre randomized double-blind trial at the University of Lille in France.
Patients with very early signs of declining cognition and early Alzheimer’s Disease receive either a placebo capsule or a capsule containing 200mg of caffeine (equivalent to two cups of coffee) twice per day, for a total of 400mg per day, over 30 weeks, followed by cognitive testing to assess any caffeine-specific impacts.
Cleveland Clinic reported that, while the results look positive for coffee and tea drinkers, Wint said more research is needed and that people should still focus on a healthy diet, regular exercise and maintaining high levels of cognitive engagement.
Forbes added that, for now, the data point towards some benefits, and likely no risks, to drinking real coffee or tea, while advising that people consume any caffeinated substance in the early part of the day to protect sleep quality.
According to Forbes, the study is in the February 2026 issue of the Journal of the American Medical Association reviewed data from over 100,000 adults followed for up to four decades.
