Ignore the EFSA. This is the real science of caffeine

Posted on Updated on

I am currently in a state of outrage at a new “study” by the European Food Safety Authority. The EFSA, a body entrusted with keeping food safe, has just produced scientifically indefensible claims about caffeine. I know they are indefensible because I trust my body more than a bunch of badly trained chemists in Brussels.

These badly trained chemists have worked out that having more than 5 espressos a day will cause damage to your health. Piffle I say. And a real threat to the health of Europeans.

Here is why.

First, not all espressos are alike. A seminal study, carried out by high school students in Byres Road, Glasgow (10 minutes from where I live), showed conclusively that the caffeine content of espresso is all over the place.

Caffeine content of a Starbucks espresso is a well regulated 50 mg per shot, far too low as all thinking persons know. Costa Coffee, however, is 3 times greater.

Patisserie Francoise, an independent Glaswegian joint, takes it up one notch. They deliver 322 mg per shot, six times that of Starbucks.

Naturally, like all good scientists, I believe in replication. I rushed out to test the effectiveness of Patisserie Francoise’s medicine, but was sadly let down.

But there you go. “Don’t drink more than five espressos a day” is meaningless advice.

And that’s ignoring the real science.

I will now give it to you. And I am sure all scientists who have studied coffee each morning throughout their careers will agree with me here.

Performance is always zero when caffeine levels is zero. That is not a controversial issue, and it needs no further study. I will take it as read.

What happens when it goes up? There are three possible responses, and only three.

First, there is the linear response. It goes like this:


The evidence for this response is limited. However, it still has some dogged adherents within the scientific community.

Then there is the saturating response. After 4 or 5 cups of espresso the brain can’t take any more, and it just keeps going as before. There is no, note no, decline in performance; there is never a decline in performance.


The evidence that the above response holds on a Friday is robust (p value = 0.01).

However, on Mondays, Tuesday, and especially Wednesdays, the most credible response is the quadratic response (or as many commentators like to call it, the exponential response).


That dear reader is the real science of caffeine.


Wilson, R.J., 2015. Ten cups is better than one. The response of intellectual performance to caffeine intake. Science. In review.