The study that measured it properly

In 2014, Killer, Blannin and Jeukendrup published a trial designed specifically to answer this question, and the design is worth walking through because the details are what make it convincing.

Fifty-two men were recruited and fifty completed. All were habitual coffee drinkers, at three to six cups a day. They completed two three-day trials in counterbalanced order, separated by a ten-day washout. In one trial they drank four 200 ml cups of black coffee a day, dosed at 4 mg of caffeine per kilogram of body mass, which came to a mean of 308 mg of caffeine daily. In the other they drank four 200 ml cups of water. Total test beverage was 800 ml a day in both arms. Their entire diet was provided and replicated across both trials.

Then they measured hydration properly. Total body water by deuterium oxide dilution is the reference method, not an estimate. They also measured 24-hour urine volume, urine specific gravity and osmolality, sodium and potassium excretion, haematocrit, serum osmolality, total plasma protein, creatinine, blood urea nitrogen, and body mass. Blood caffeine was measured to confirm people drank what they were given.

The results:

• Total body water: 51.5 kg on coffee, 51.4 kg on water. No significant difference (p = 0.90).
• 24-hour urine volume: 2,409 ml on coffee, 2,428 ml on water. No significant difference.
• No differences in any blood marker of hydration.

The authors' conclusion: coffee, consumed in moderation by caffeine-habituated males, provides similar hydrating qualities to water.

Black tea has been tested the same way, in a randomised counterbalanced trial where 21 men were recruited and 19 completed all four conditions. They drank either four or six 240 ml cups of black tea (168 or 252 mg of caffeine) or identical volumes of boiled water, with all food controlled and blood sampled over 24 hours. The finding was no significant differences between tea and water for any of the mean blood or urine measurements.

Why the myth is not completely baseless

Caffeine can be a diuretic. The question is at what dose, and in whom.

A review of the literature from 1966 to 2002 lays it out precisely, and the qualifications are the whole story:

Acute ingestion of caffeine in large doses, at least 250 to 300 mg, equivalent to the amount found in 2 to 3 cups of coffee or 5 to 8 cups of tea, results in a short-term stimulation of urine output in individuals who have been deprived of caffeine for a period of days or weeks.

And then:

A profound tolerance to the diuretic and other effects of caffeine develops, and the actions are much diminished in individuals who regularly consume tea or coffee.

And the conclusion:

Doses of caffeine equivalent to the amount normally found in standard servings of tea, coffee and carbonated soft drinks appear to have no diuretic action.

So the original observation was real, in people who had gone without caffeine for days or weeks and then took a substantial dose. That is not most coffee drinkers, most days. Tolerance develops, and it develops fast: one group noted that caffeine withdrawal of as little as four days is enough for that tolerance to be lost.

A 2015 meta-analysis of 16 studies and 28 effect sizes, at a median dose of 300 mg, puts a number on the residual effect at rest: about 109 ml of extra urine on average, with a spread of ±195 ml (roughly 16%, ±19%). The variation between studies is larger than the effect itself, which is its own kind of answer. The authors' own summary: "Caffeine exerted a minor diuretic effect which was negated by exercise. Concerns regarding unwanted fluid loss associated with caffeine consumption are unwarranted."

The idea's origin, and the gap in this article

The "coffee dehydrates you" belief is usually traced back to a study from 1928, by researchers named Eddy and Downs, who reported acute caffeine diuresis, and who, even then, suggested that regular consumption might lead to tolerance developing against that effect.

We could not retrieve the 1928 paper itself. That attribution comes from a later paper describing it, so treat it as secondhand.

There is also a real physiological mechanism behind the effect. Caffeine acts as an adenosine receptor antagonist, reducing how much sodium the kidney reabsorbs. One source puts the dose at which this produces a diuretic effect at 500 mg or more, which is higher than the 250 to 300 mg figure above. We are giving you both rather than picking one, because two credible sources give two different thresholds and pretending to a single crisp number would be false precision.

An important limit on all of this

Both of the headline trials above, the coffee one and the tea one, enrolled men only. The coffee study excluded women explicitly, on the grounds of possible menstrual-cycle effects on fluid balance.

That matters more than it might seem, because the 2015 meta-analysis found a sex difference in caffeine's diuretic effect: an effect size of 0.75 in women versus 0.13 in men, with the difference between them statistically significant.

In habitual male coffee and tea drinkers, careful trials found hydration equivalent to water. In women, the pooled data suggest a larger diuretic response, and the equivalent controlled hydration trials have not been done. That is a real gap, not a quibble, and anyone stating "coffee hydrates you just like water" as a universal fact is overreaching.

A few other limits from the coffee trial, stated by its own authors: it was free-living rather than a metabolic ward study; there was no decaffeinated coffee arm; three days is short; and body mass fell slightly in both arms, which the authors could not fully explain and suggested might mean participants were not given quite enough fluid overall. Also, while total fluid balance did not differ, urinary sodium excretion was significantly higher on coffee, so "no difference at all" would overstate it.

One more disclosure: we could not retrieve the funding statements for either the coffee trial or the tea trial. Beverage-industry funding is common in hydration research, and the lead author of the tea study is affiliated with a nutrition consultancy rather than a university department. We are not alleging anything. We simply could not check, and you should know that.

Do they count toward your fluid intake?

Yes, according to the National Academies panel that set the reference intakes for water, which concluded that beverages containing caffeine contribute to total water intake, with any diuretic effect being transient.

This fits the rest of the picture. The Adequate Intake those panels set is for total water (every beverage plus the water in food), and fluids supply about 81% of it. Coffee and tea are mostly water, and they behave like mostly water.

How much caffeine is in what

FDA's consumer figures, all per 12 fluid ounces:

The ranges in that table are wide, particularly for coffee, where brewing method, bean and strength all move it substantially. The figures are also per 12 ounces, not per cup; a large takeaway coffee is often 16 or 20.

We could not retrieve a figure for espresso, so we are not giving you one.

FDA describes 400 mg a day for healthy adults as "an amount not generally associated with negative effects," which it puts at about two to three 12-ounce cups of coffee. The wording is careful. It is neither an approval nor a recommendation, just a description.

FDA also emphasises the "wide variation in both how sensitive people are to the effects of caffeine and how fast they eliminate it from the body." If 200 mg keeps you awake all night, that is a real physiological difference, not a failure of willpower.

At the far end, FDA notes that toxic effects can be observed with rapid consumption of around 1,200 mg. The warning is aimed at pure and highly concentrated caffeine powders, where that amount is less than half a teaspoon. It is not a statement about coffee.

On pregnancy: FDA's page advises anyone pregnant, trying to become pregnant, or breastfeeding to consult their healthcare provider about limiting caffeine, and does not state a milligram figure. You will see a specific daily limit quoted in a lot of places. We could not verify where it comes from, so we are not repeating it. Ask your obstetric provider; this is exactly the kind of question they answer routinely.

The tea-and-iron question

This one is real, and also routinely overstated.

Polyphenols in tea and coffee do reduce the absorption of non-heme iron, the form found in plant foods, eggs, dairy and iron supplements. That is well established.

But the NIH fact sheet on iron includes a sentence that changes the practical picture: "the effects of enhancers and inhibitors of iron absorption are attenuated by a typical western diet, so they have little effect on most people's iron status."

That fact sheet does not quantify the polyphenol effect and gives no timing advice for tea or coffee. Its timing guidance concerns calcium supplements. So the very common instruction to "wait an hour after meals before drinking tea" is not coming from that source, and we are not going to invent one.

Where it matters more is in the groups at real risk of low iron: people who are pregnant, infants and young children, people who menstruate, blood donors, and people with certain digestive conditions or chronic illnesses. If you are in one of those groups, or have been told your iron is low, the timing of tea relative to iron-rich meals or iron supplements is a reasonable thing to ask your doctor about. For most people, it is not a reason to change anything.

And a small point of vocabulary: herbal "teas" are not tea. Tea means Camellia sinensis: black, green, white, oolong. Peppermint and chamomile are infusions of other plants, contain no caffeine, and do not carry the same polyphenol profile.

What about all those headlines saying coffee is good for you?

They are real studies, and they are observational, which is an important distinction.

A meta-analysis of 20 prospective cohort studies covering 973,904 participants and 129,538 deaths found that high coffee consumption versus low was associated with lower total mortality, with a relative risk of 0.86. A separate meta-analysis of 30 prospective studies covering more than 1.1 million people found coffee consumption associated with lower risk of type 2 diabetes.

Three reasons to hold these lightly.

They cannot establish cause. These are cohort studies. People who drink more coffee differ from people who drink less in many ways, and statistical adjustment only goes so far. There is no randomised trial assigning people to decades of coffee.

The US estimate is weaker than the headline. In the mortality analysis, the pooled result for Europe was 0.78 and for Japan 0.82, but for the USA it was 0.92, with a confidence interval reaching 1.00. For a US reader, that is the relevant number and it barely clears the null.

Decaf shows the same association for diabetes. The relative risk is 0.94, essentially identical to caffeinated, which rather undercuts any story in which caffeine is the active ingredient and points toward either non-caffeine compounds or confounding.

So coffee drinkers do not appear to be doing themselves harm at ordinary intakes, which is a reasonable and modest thing to take from a large body of observational data. That is different from coffee being good for you, and we are not going to make that step.