Is hot tea or coffee carcinogenic? Why IARC points to the temperature, not the drink

1991: coffee in Group 2B

IARC first evaluated coffee in 1991 and classified it as possibly carcinogenic to humans (Group 2B). The decision rested on limited evidence for one cancer site: the urinary bladder.

Twenty-five years later IARC explained what had happened to that evidence. Many of the early studies that found a link with bladder cancer did not adequately account for tobacco smoking, which is itself a strong cause of bladder cancer and was common among heavy coffee drinkers. Most of the later, better studies showed no consistent association. The bladder evidence was therefore judged too weak to say whether coffee causes the disease.

This is an ordinary part of how the Monographs work. A classification reflects the evidence available at the time of the meeting, and it can move in either direction when new studies appear. What “limited evidence” means in IARC’s vocabulary is explained on Groups 1, 2A, 2B and 3.

2016: Group 3 for coffee

The re-evaluation took place in Lyon in June 2016. A Working Group of 23 scientists reviewed more than 1000 studies in humans and in experimental animals; IARC’s Q&A counts nearly 500 relevant epidemiological studies covering more than 20 types of cancer. The conclusions were announced on 15 June 2016, a summary appeared the same day in The Lancet Oncology, and the full assessment was published as Volume 116 of the Monographs.

The overall evidence that drinking coffee causes cancer was judged inadequate, and coffee was placed in Group 3. For some cancers the studies were quite clear: they showed no carcinogenic effect for cancers of the pancreas, the female breast and the prostate, and reduced risks for cancers of the liver and of the lining of the womb (the endometrium). For more than 20 other cancers the evidence was inconclusive.

IARC was careful about what Group 3 means. Its Q&A states that a Group 3 evaluation “does not mean that a substance has been proven to be safe”; it means that the data do not allow a conclusion about whether it causes cancer. The same rule applies to every agent in that group, as described on Groups 1, 2A, 2B and 3.

Very hot drinks: Group 2A

The more striking result of the same meeting concerned not a substance but a habit. Drinking very hot beverages was classified as probably carcinogenic to humans (Group 2A), mainly on the basis of evidence for cancer of the oesophagus, the tube that carries food from the mouth to the stomach.

IARC drew the line at 65 °C. The press release defines “very hot” as any beverage consumed above 65 °C; the Q&A speaks of temperatures of 65 °C or above. The threshold comes mainly from animal experiments, which suggested that carcinogenic effects probably occur from that point upwards, and it is supported by measurements of real drinking temperatures in regions where the cancer is common.

IARC also stated that the typical drinking temperature of tea and coffee in most parts of the world is below 65 °C. The classification therefore describes how hot a drink is when it is swallowed, not which drink it is.

The evidence from maté

Much of the evidence came from maté, an infusion drunk mostly in South America and traditionally taken very hot – at about 70 °C – through a metal straw. It can also be drunk warm or cold, and that variety made it possible to separate the effect of the drink from the effect of its temperature.

A combined analysis of several epidemiological studies, with about 1400 cases of oesophageal cancer, looked at both the temperature and the amount of maté consumed. The risk of cancer rose with temperature, independently of the amount drunk; the increase was significant for very hot maté but not for warm maté. Studies in places such as China, the Islamic Republic of Iran and Turkey, as well as South America, pointed the same way: the hotter the drink, the higher the risk of oesophageal cancer.

Animal experiments matched this picture. In IARC’s summary, very hot water at 65 °C promoted the development of oesophageal tumours, whereas maté itself did not. Maté drunk warm or cold was placed in Group 3. As the IARC Director of the time, Christopher Wild, put it, it appeared to be the temperature, rather than the drinks themselves, that was responsible.

Why it matters, and what the label does not say

Oesophageal cancer is not rare. In 2016 IARC described it as the eighth most common cancer worldwide, with about 400 000 deaths recorded in 2012, or 5% of all cancer deaths (figures as quoted by IARC in 2016).

Group 2A, like every Monographs group, describes the strength of the evidence that something can cause cancer, not how large the risk is for a given person. IARC does not say how many cancers very hot drinks cause, and the Monographs do not recommend what anyone should do; the results are for governments and health agencies to weigh. The difference between those two questions is the subject of Can it, or will it?.

Not the final word

Classifications are revisited when evidence accumulates. In March 2024 an IARC Advisory Group of 28 scientists from 22 countries met in Lyon to recommend priorities for 2025–2029. It listed “very hot beverages and food” among the high-priority agents for re-evaluation, citing new evidence from studies in people, and noted that very hot food had never been evaluated (status as of 3 November 2026).

The coffee story is a useful reminder when reading old headlines. A classification quoted in an article may already have been replaced, and IARC’s own list shows the current group for each agent. Two other cases in which a label is often misread are described in Do mobile phones cause cancer? and Smog and cancer.

This article explains how evaluations work; it is not medical advice.