Skip to content
Years & Yield

Coffee: what the research shows about 3 to 4 cups a day

Coffee was long seen as a vice. Hundreds of studies now point the other way. That is not a free pass, though: pregnancy and sleep come with clear limits.

Moderate evidenceHow we rate, 4 min read

In short

  • A review of more than 200 meta-analyses found a 17% lower risk of death at 3 to 4 cups of coffee a day, compared with people who drink none.
  • Almost all of the data are observational. Whether coffee itself is the cause has not been proven.
  • For healthy adults, up to 400 mg of caffeine a day is considered safe, a little over 4 cups of filter coffee. In pregnancy the limit is 200 mg.
  • A large dose of caffeine (400 mg, a little over 4 cups of filter coffee) cut sleep by more than an hour even when taken 6 hours before bedtime.
A steaming cup of coffee next to a hand grinder and a few coffee beans.
AI illustration

Coffee is one of the most thoroughly studied foods there is. A 2017 review in the BMJ gathered every meta-analysis available: 201 analyses of observational studies covering 67 health outcomes, plus 17 analyses of intervention studies. The result: coffee was linked to benefit more often than to harm.1

What the studies show

The link was strongest at 3 to 4 cups a day, compared with people who drink no coffee:1

Health outcomeRisk at 3 to 4 cups a day
Death from any cause17% lower
Death from cardiovascular disease19% lower
Cardiovascular disease15% lower

Heavy coffee drinkers were also 18% less likely to develop cancer than light drinkers. The researchers also found lower risks for several specific cancers and for liver, metabolic and neurological conditions.1

The relationship is not linear. More coffee does not always mean more benefit; the largest risk reduction came at 3 to 4 cups.1

Why this is not a free pass

Almost all of these figures come from observational studies. Smoking shows how much other factors matter: most harmful associations disappeared once studies properly accounted for smoking. The authors therefore call for randomized trials to establish whether coffee itself is the cause.1

Two exceptions remained. In pregnancy, high coffee intake was linked to lower birth weight, more preterm births and more pregnancy loss. And women who drank a lot of coffee had more fractures; men did not.1

How much caffeine is safe

In 2015, the European Food Safety Authority assessed which amounts are safe for healthy people:2

  • Adults: up to 400 mg of caffeine a day, spread over the day, and up to 200 mg at once.
  • Pregnant and breastfeeding women: up to 200 mg a day.
  • Children and adolescents: up to 3 mg per kilogram of body weight a day.

By EFSA’s figures, a cup of filter coffee (200 ml) contains about 90 mg of caffeine and an espresso (60 ml) about 80 mg.5 So 400 mg is a little over 4 cups of filter coffee or 5 espressos. Caffeine from tea, cola, chocolate and energy drinks counts too.

Coffee and sleep

In a small study, 12 healthy sleepers took 400 mg of caffeine, by EFSA’s figures a little over 4 cups of filter coffee at once. They took it at bedtime, 3 hours before or 6 hours before. In all three cases they slept measurably worse than with a placebo; even 6 hours before, their sleep was cut by more than an hour. The researchers therefore advise avoiding substantial amounts of caffeine for at least 6 hours before bed.3 If you go to sleep at 10:30 pm, larger amounts of caffeine should come by 4:30 pm, and earlier is better.

Caffeine calculator

On a normal day
Filter coffeecup, 200 ml, 90 mg2
Espresso60 ml, 80 mg1
Black teacup, 220 ml, 50 mg0
Colacan, 355 ml, 40 mg0
Energy drinkcan, 250 ml, 80 mg0
Plain chocolate50 g, 25 mg0
Milk chocolate50 g, 10 mg0
For whom
Caffeine per day260 mgabout 3 cups of filter coffee
Limit: 400 mg65%

Below the 400 mg a day that EFSA considers safe for healthy adults.

  • At most 200 mg at once, about 2 cups of filter coffee.
  • Larger amounts no later than 16:30, 6 hours before bed. Even 100 mg close to bedtime can disturb sleep.

EFSA reference values per portion. The actual content varies widely with the bean and the brewing method.

Your entries stay in this browser only.Open as its own page

Filtered or unfiltered

A meta-analysis of 12 randomized trials with 1,017 participants found that drinking coffee for about 45 days raised total cholesterol by 8.1 mg/dl (about 0.21 mmol/l) and LDL cholesterol by 5.4 mg/dl (about 0.14 mmol/l) on average. The rise was larger with unfiltered coffee, and people with high blood lipids reacted more strongly.4 If you need to watch your cholesterol, filter coffee is the better choice.

Put it into practice

0 of 5 done

What we don’t know

  • Coffee drinkers differ from non-drinkers in other ways too. Whether the lower risk comes from the coffee would need large randomized trials.
  • How much caffeine a cup contains varies widely with the coffee and how it is made. Figures per cup are rough guides.
  • Why women who drank a lot of coffee had more fractures in some studies, but men did not, is unclear.

Sources

  1. 1Coffee consumption and health: umbrella review of meta-analyses of multiple health outcomesPoole R, Kennedy OJ, Roderick P, Fallowfield JA, Hayes PC, Parkes J, BMJ 359, j5024, 2017
  2. 2Scientific Opinion on the safety of caffeineEuropean Food Safety Authority (EFSA), EFSA Journal 13 (5), 4102, 2015
  3. 3Caffeine effects on sleep taken 0, 3, or 6 hours before going to bedDrake C, Roehrs T, Shambroom J, Roth T, Journal of Clinical Sleep Medicine 9 (11), pp. 1195–1200, 2013
  4. 4The effect of coffee consumption on serum lipids: a meta-analysis of randomized controlled trialsCai L, Ma D, Zhang Y, Liu Z, Wang P, European Journal of Clinical Nutrition 66 (8), pp. 872–877, 2012
  5. 5CaffeineEuropean Food Safety Authority (EFSA), topic page, 2015
Written by
Roman Henggeler, Publisher
Published

Read this article in German

Two articles a week

Health and money, checked and read in five minutes.

Two articles every Friday. Unsubscribe with one click. We measure whether the newsletter is opened and which links are clicked. Privacy