The largest studies associate regular coffee drinking with lower all-cause mortality, strongest at around three to four cups a day, where the relative risk sits near 0.83. The association is clearest for filtered coffee and appears to vanish when sugar is added. These are observational findings. They show association, not causation, and they describe plain coffee, not what is often poured on top of it.
Here is what the strongest evidence says, read slowly and without enthusiasm.
What the mortality studies show
Large cohort studies, pooled across hundreds of thousands of people, consistently find that coffee drinkers die at lower rates over the study period than people who drink none. The association is strongest at around three to four cups a day, where the relative risk for all-cause mortality sits near 0.83. In plain words, moderate coffee drinkers in these datasets were about 17 per cent less likely to die over the follow-up period than non-drinkers. Higher intakes do not show a stronger association.
Say it carefully. Coffee drinking is associated with lower mortality. That sentence, and no more. The studies observe people who already drink coffee. They do not hand coffee to half a population and wait.
Filtered coffee reads better than unfiltered
The clearest single dataset is a Norwegian cohort of roughly 508,000 people. Filtered coffee was associated with lower mortality than drinking no coffee at all. Unfiltered coffee, boiled or pressed, showed a weaker picture. The researchers point to compounds in unfiltered coffee's oils, which a paper filter holds back, as a likely explanation.
If you want your habit to match the evidence, the evidence drank filter coffee.
Sugar appears to erase the association
Analyses that separate sweetened from unsweetened coffee find the favourable association concentrated in the unsweetened cups. Add sugar, and the association fades towards nothing. The studies cannot say exactly why. What they can say is that the pattern belongs to plain coffee. Sweetened coffee drinks are a different food, and the evidence treats them that way.
Association is not causation
This deserves its own section, stated as plainly as we can manage.
Everything above is observational. People who drink moderate amounts of coffee differ from people who drink none, in income, in habits, in health. Some people give up coffee because they are already unwell, which alone can make abstainers look worse in the data. And studies that try to strip these biases out using genetic methods have not confirmed a causal effect.
So the honest position is narrow. The association is real, consistent, and large enough to notice. Whether coffee causes any of it is unproven. We sell coffee, which is a reason to read our reading with suspicion. It is also why we keep the claim small.
Who should be careful
Averages are not advice.
In pregnancy, caffeine guidance is stricter, and the sensible move is to follow your clinician's limit rather than a cohort study. Anyone whose doctor has said to limit caffeine should take the doctor over the dataset. And caffeine holds on for hours, so a late cup quietly taxes sleep. Poor sleep is not a fair exchange for any association in this article.
A quiet note on dose. Liberica carries roughly half the caffeine of arabica, which makes it a gentler afternoon cup than its heavy body suggests.
How we read all this
The evidence describes a modest habit: filtered coffee, unsweetened, a few cups, most days. Nothing in it licenses health promises, and we make none. What it quietly supports is the idea that coffee brewed simply and drunk black needs no defending.
What we can control is what goes into the bag. We screen green beans at independent laboratories accredited to ISO/IEC 17025 before roasting, and publish results as each lot clears, as set out on our standard. We apply the same slow reading to organic claims in is organic coffee worth it and to a persistent pesticide myth in is coffee one of the most sprayed crops. The coffees themselves are at the coffees.
The short answers
Is coffee good for you?
The largest observational studies associate regular coffee drinking with lower all-cause mortality, strongest around three to four cups a day. Association is not causation, and no study proves coffee causes better health.
How many cups a day show the strongest association?
Around three to four cups daily, where relative risk for all-cause mortality sits near 0.83 in pooled cohort data. Higher intakes do not show a stronger association.
Is filtered coffee better than unfiltered?
In a Norwegian cohort of about 508,000 people, filtered coffee showed the clearest association with lower mortality. Unfiltered coffee showed a weaker picture.
Does adding sugar change the picture?
It appears to. The favourable association concentrates in unsweetened coffee and fades towards nothing when sugar is added.
Who should be careful with coffee?
Anyone pregnant, anyone advised by a doctor to limit caffeine, and anyone whose sleep suffers. Clinical guidance outranks cohort studies every time.
