Does Drinking Tea or Coffee Above 65°C (149°F) Really Triple Esophageal Cancer Risk?

No—not in the precise way that claim implies. A new study of 977,282 UK adults found that people who described their preferred drinks as “very hot” had about 3.17 times the relative risk of esophageal squamous cell carcinoma (SCC) compared with people who preferred them warm. But researchers did not measure the temperature of participants’ drinks in degrees Celsius or Fahrenheit.

The widely repeated 65°C (149°F) number comes instead from an older International Agency for Research on Cancer (IARC) classification defining beverages above 65°C as “very hot.” Combining that older threshold with the new study’s 3.17 risk estimate produces the viral statement that “drinks above 65°C triple cancer risk.” The new study did not establish that.

That correction does not mean the underlying concern is bogus. The broader evidence is increasingly persuasive that repeatedly drinking beverages at very high temperatures increases the risk of SCC, probably because of thermal injury to the esophageal lining. And one major prospective study that actually measured tea temperature found higher SCC risk at 60°C (140°F) and above.

The best practical interpretation is therefore not “65°C triples your risk” or “the study proves nothing.” It is:

Regularly drinking beverages very hot appears to be a genuine and avoidable SCC risk, but science has not identified a magic temperature at which risk suddenly switches on.

Where the “65°C triples your risk” claim came from

Two separate findings are being fused together.

The distinction matters because there was no thermometer attached to participants’ cups. Cancer Research UK, which funded the research, explicitly says the investigators were unable to measure actual beverage temperatures. Participants instead classified their preferred drinks as warm, hot or very hot.

IARC’s 65°C number predates the new research by a decade. Its 2016 working group considered beverages from roughly 50°C to 65°C (122°F to 149°F) “hot” and those above 65°C (149°F) “very hot.” The agency classified consumption of very-hot beverages as Group 2A, probably carcinogenic to humans.

IARC also stresses that its hazard classifications do not quantify how much risk an individual exposure creates. Group 2A tells us something about evidence that an exposure can cause cancer; it does not mean every exposure has the same risk or establish the size of that risk at 65°C.

What the new million-person study actually found

Researchers led by Keren Papier at the University of Oxford combined two large prospective UK cohorts: the Million Women Study and UK Biobank.

There were 977,282 participants in the analysis. Average follow-up was approximately 14.2 years in the Million Women Study and 11.1 years in UK Biobank. During follow-up researchers identified:

• 2,348 total esophageal cancers
• 1,045 squamous cell carcinomas
• 1,303 adenocarcinomas

The paper was published online in the International Journal of Cancer on September 8, 2026.

Temperature was the striking result.

Compared with participants who preferred their beverages warm, people reporting “very hot” drinks had an SCC relative risk of:

RR 3.17

That means roughly 217% higher relative risk, or just over three times the reference group’s risk—not a 317% increase.

People describing their drinks merely as “hot” also had higher risk, at roughly 1.7 times the risk among warm-drinkers.

For esophageal adenocarcinoma, however, the very-hot temperature result was essentially null:

RR 0.92, 95% CI 0.76–1.12.

That histologic difference is important. The result was not simply “hot drinks caused more of every type of esophageal cancer.”

The study measured temperature preference—not 65°C (149°F)

This is the central limitation behind the viral headline.

A person selecting “very hot” on a questionnaire might routinely drink at 62°C (144°F), 66°C (151°F), 72°C (162°F) or another temperature. Another person’s idea of “very hot” may differ substantially.

That makes the categories useful epidemiologically—they probably capture meaningful differences in drinking behavior—but they cannot establish:

64°C (147°F) = one risk 65°C (149°F) = three times that risk

The authors were studying whether people’s usual temperature preference within a Western population predicted cancer. Previous studies had produced substantial evidence in regions where tea or maté may be consumed around 70°C (158°F), while Western consumption is generally cooler.

Cancer Research UK consequently says more research is needed to determine the actual temperatures associated with higher risk.

So is 65°C (149°F) actually a cancer threshold?

No evidence shows that 65°C (149°F) is a biological on/off switch.

IARC’s >65°C definition is a useful classification based on the available epidemiological and experimental evidence. It should not be interpreted to mean that 64°C (147°F) has been proven safe while 65°C suddenly becomes dangerous.

One of the best reasons comes from an earlier study that did what the 2026 UK research could not: researchers actually measured the temperature of participants’ tea.

The Golestan Cohort Study followed 50,045 adults in northeastern Iran for a median 10.1 years. Researchers used a validated procedure to objectively measure tea-drinking temperature at baseline. During follow-up, 317 participants developed esophageal squamous cell carcinoma.

Here were the adjusted estimates:

The confidence intervals around the two narrower temperature groups overlap, so those numbers should not be read as evidence that 60–64°C is somehow worse than ≥65°C.

The important point is the opposite:

The measured human data do not reveal a sharp jump at 65°C (149°F).

Risk in the 60–64°C (140–147°F) group was already elevated, and its estimate was very similar to the ≥65°C (≥149°F) group.

That supports thermal exposure as a concern while undermining the idea of a precise 65°C cliff.

What temperature should you let tea or coffee cool to?

If someone wants a numeric precautionary target, below about 60°C (140°F) is more defensible than treating 65°C (149°F) as the dividing line between safe and dangerous.

The Golestan researchers concluded that advising people to let beverages cool to below 60°C (140°F) before drinking would be a reasonable public-health measure.

That does not establish that 59°C (138°F) carries zero risk.

The available evidence is more consistent with thermal dose than with a binary threshold. Actual exposure probably depends on several things at once:

temperature × amount swallowed × contact time × frequency of drinking

Sip size and repeated exposure may therefore matter alongside the temperature measured at a single moment. IARC’s more recent research-priority work likewise notes that reliance on self-reported temperature alone is an incomplete way to characterize thermal exposure.

A practical interpretation is:

Above 65°C (149°F): firmly within IARC’s “very hot” category and sensible to avoid as a routine habit.

60–65°C (140–149°F): not proven safe; prospective temperature-measured evidence found elevated SCC risk in this range.

Below 60°C (140°F): a reasonable precautionary target supported by the strongest prospectively measured temperature study, but not a scientifically proven zero-risk boundary.

Does drinking six cups a day increase cancer risk too?

The UK researchers also examined how many hot drinks people consumed, separately from their stated temperature preference.

After statistically adjusting for temperature, people consuming six or more hot drinks per day had an SCC relative risk of 1.71 compared with people consuming fewer than six.

For adenocarcinoma the corresponding estimate was much smaller: 1.19.

But “six cups” should not become another magic threshold.

The analysis does not establish that:

• cup number five is safe but cup number six causes cancer;
• the six drinks were all above 65°C (149°F);
• drinking six ordinary-temperature cups raises risk by exactly 71%;
• quantity matters equally for everyone.

It is an observational comparison between consumption groups after statistical adjustment.

A plausible explanation is that more cups mean more repeated thermal exposures. The objectively measured Golestan study lends some support to this cumulative-exposure idea: people consuming at least 700 mL of tea per day at ≥60°C (140°F) had about 90% higher SCC risk than those consuming less than 700 mL at below 60°C.

Still, the temperature finding is easier to interpret than “six cups.” The evidence does not identify six as a biological cutoff.

Why the earlier temperature-measured study matters so much

The Golestan study also exposes another problem with translating words like “very hot” directly into Celsius or Fahrenheit.

In that same cohort:

• objectively measured ≥60°C (≥140°F) tea was associated with an adjusted HR of 1.41;
• people who merely reported preferring “very hot” tea had an HR of 2.41 compared with those describing it as cold or lukewarm.

Those are different comparisons and should not be treated as interchangeable estimates. But the disparity demonstrates why subjective temperature labels and thermometer readings are not the same exposure variable.

A “very hot” preference may capture more than one temperature reading. It could reflect consistently drinking quickly after pouring, larger volumes, repeated exposures or other behaviors correlated with temperature.

That makes subjective preference potentially useful. It does not make “very hot” synonymous with “65°C (149°F) or higher.”

This also wasn’t the first UK study to see the pattern

Another useful piece of context is being lost in the rush around the new million-person result.

A 2025 British Journal of Cancer prospective analysis already examined 454,796 UK Biobank participants. During an average 11.6 years of follow-up, researchers identified 242 SCC cases and 710 adenocarcinomas.

SCC risk increased with hotter reported beverages and greater intake. For example, among participants preferring very-hot beverages, the SCC hazard ratios increased across consumption groups, reaching 5.64 among those consuming more than eight cups per day in that study’s analysis. No comparable consistent pattern appeared for adenocarcinoma.

The 2026 research therefore did not discover the hot-drink hypothesis from scratch.

Its importance is that it combines UK Biobank with the much larger Million Women Study, producing far more SCC cases and substantially strengthening the evidence that the relationship exists in a Western population where tea and coffee are generally consumed cooler than the roughly 70°C (158°F) beverages studied in some earlier high-risk regions.

Does this prove that very hot drinks cause cancer?

Not by itself.

The study is observational. Participants were not randomly assigned to drink warm or scalding beverages for a decade—which would obviously be an unrealistic and unethical cancer trial.

Observational studies can be distorted by confounding: people who drink beverages very hot may differ from warm-drinkers in other ways that influence cancer risk.

Researchers can statistically adjust for known differences, but no model can guarantee that every confounder has been identified and perfectly measured.

For that reason, the most defensible description of the new result is a strong association rather than experimental proof of causality.

But stopping at “correlation doesn’t equal causation” would also understate the evidence.

The thermal explanation is supported by several independent strands:

• prospective studies show exposure occurring before cancer;
• hotter reported beverages generally correspond to higher SCC risk;
• an Iranian prospective cohort found an association using objectively measured temperature;
• the association repeatedly appears more strongly for squamous cell carcinoma than adenocarcinoma;
• IARC found supporting animal evidence involving very hot water;
• studies involving different beverages point toward temperature rather than coffee, tea or maté chemistry alone.

Taken together, it is reasonable to infer that repeated thermal injury probably contributes causally to SCC risk, while acknowledging that the exact human temperature-response curve remains uncertain.

That is an inference from the total evidence—not something the 2026 observational study alone proves.

Why would heat affect SCC but not adenocarcinoma?

The distinction fits what is already known about the two major forms of esophageal cancer.

Squamous cell carcinoma arises from squamous cells lining much of the esophagus. Adenocarcinoma develops from glandular-type tissue and has a different disease pathway, strongly associated with conditions including gastroesophageal reflux and Barrett’s esophagus.

Earlier observational research has similarly found hot beverage and food consumption associated more clearly with SCC, while pooled associations for adenocarcinoma have often been absent. A meta-analysis of 39 observational studies reported an odds ratio of 1.60 for SCC but 0.79 for adenocarcinoma, the latter not statistically significant.

The likely mechanism is that repeated high-temperature exposure injures the esophageal squamous lining, potentially creating inflammation or conditions that make malignant changes more likely.

That mechanism is biologically plausible and consistent with the pattern of evidence, but the precise sequence by which thermal injury contributes to human cancer has not been definitively mapped.

Is it coffee and tea that are dangerous, or the heat?

The evidence points much more strongly toward temperature.

When IARC reassessed coffee in 2016, its working group found inadequate evidence that coffee drinking itself was carcinogenic to humans. At the same time, it classified drinking very-hot beverages—generally above 65°C (149°F) under its framework—as probably carcinogenic and said the available evidence suggested temperature rather than the drinks themselves was responsible for the esophageal-cancer association.

A 2026 systematic review and meta-analysis of coffee consumption likewise found no statistically significant overall association between coffee intake and esophageal cancer, although the authors rated the evidence as low certainty and noted persistent problems measuring beverage temperature consistently.

So this research should not be interpreted as evidence that a cup of comfortably warm coffee or tea is itself carcinogenic.

The concern is chronic exposure to excessive heat.

What does “three times the risk” mean in absolute terms?

This is where relative-risk headlines can become misleading even when the number itself is correct.

Cancer Research UK estimates that people in the UK have approximately a 1% lifetime risk of developing esophageal SCC.

It would be tempting to multiply:

1% × 3.17 = 3.17%

and claim that drinking very-hot beverages raises lifetime risk from 1% to 3.17%.

Do not make that calculation.

The figures describe different things.

The 1% number is a broad population lifetime-risk estimate. The 3.17 number is a relative risk between exposure categories in the study cohorts during their follow-up.

The warm-drinking reference group’s lifetime risk is not established as exactly 1%. Age, sex, smoking, alcohol consumption and other individual factors also substantially affect baseline risk.

The responsible interpretation is therefore:

A threefold relative association is substantial, but SCC remains uncommon in the UK, and this study does not provide enough information to tell an individual that their lifetime risk changes from 1% to a specific new percentage.

What about the estimate that 14% of SCC cases could be prevented?

The authors calculated that approximately 14% of SCC cases in UK adults could potentially be avoided if people reporting “very hot” drinking instead consumed beverages at temperatures representative of the study’s “hot” category.

That is a population-attributable estimate, not the result of an intervention.

It depends on the assumption that the observed relationship is genuinely causal and that lowering exposure would produce the expected reduction.

The accurate phrasing is therefore:

The researchers estimated that about 14% of SCC cases might be avoidable if the relationship is causal and very-hot drinkers lowered their drinking temperature.

It is not accurate to say a trial demonstrated that cooling drinks prevented 14% of cancers.

How long should you let coffee or tea cool?

There is no universal evidence-based number of minutes.

Cooling depends on:

• starting temperature;
• volume;
• cup size and material;
• whether the container is insulated;
• whether there is a lid;
• room temperature;
• stirring;
• whether cold milk or another liquid is added.

An exploratory peer-reviewed study measured 356 coffees in food-service settings and 110 from household machines. The average measured serving or dispensing temperature was about 75°C (167°F).

Under the conditions studied, waiting more than about 10 minutes or adding more than 20 mL of cold milk generally brought coffee below 65°C (149°F).

That finding is useful for demonstrating how cooling works. It is not a universal ten-minute medical rule.

A small espresso, a large ceramic mug and a sealed insulated travel cup will not cool at the same rate.

If knowing the actual temperature matters to you, an ordinary food thermometer is more reliable than guessing from elapsed time.

Does adding milk make tea or coffee safer?

It can reduce thermal exposure if it actually lowers the final drinking temperature.

Adding refrigerated milk mixes a cooler liquid into the beverage and therefore lowers its temperature. The effect depends on the starting temperatures and how much milk is added. Experimental coffee-temperature research has found that cold milk can materially accelerate cooling.

That is different from saying that milk itself protects against esophageal cancer.

The relevant mechanism is temperature. A beverage that remains extremely hot after a small splash of milk is still extremely hot.

So the practical answer is:

Adding cold milk may help because it cools the drink—not because milk has been shown to neutralize the cancer risk.

Is “comfortable to drink” a good enough rule?

For most people, avoiding beverages that feel scalding, burning or uncomfortably hot is a sensible low-burden precaution, and Cancer Research UK recommends allowing tea or coffee to cool before drinking.

But personal comfort cannot define an exact Celsius or Fahrenheit threshold.

Sensory research has found meaningful variation in people’s temperature preferences and pain thresholds; one small study found an average preferred coffee temperature around 63°C (145°F) and an average pain threshold around 67°C (153°F).

That means some people can comfortably tolerate temperatures around or above the ranges implicated in epidemiological research.

You do not need to turn drinking coffee into a laboratory procedure. But if someone specifically wants to keep beverages under a numerical target such as 60°C (140°F), measuring is more reliable than relying on sensation alone.

What about hot soup and other foods?

The new UK study does not answer that question. It analyzed hot beverages, principally tea and coffee.

The broader thermal-injury hypothesis would make sufficiently hot foods biologically relevant, and older observational research has reported associations involving very hot foods as well as beverages.

A meta-analysis involving 39 studies found overall hot-beverage and hot-food consumption associated with higher esophageal-cancer risk, particularly SCC. However, the studies were heterogeneous, many were case-control studies, and the evidence is not as clean as the large prospective beverage cohorts.

IARC’s established Group 2A classification specifically concerns very hot beverages, particularly those above 65°C (149°F) under its definition, although the agency has identified very-hot foods and beverages as an area deserving further evaluation.

So it is reasonable to avoid repeatedly swallowing scalding-hot soup or food, but this particular 2026 study should not be cited as proving a soup-cancer relationship.

What is verified, what is inference, and what is unsupported?

The bottom line

The viral sentence “drinking tea or coffee above 65°C (149°F) triples your risk of esophageal cancer” contains a real scientific concern wrapped in false precision.

The September 2026 Oxford study found roughly three times the relative risk of esophageal squamous cell carcinoma among people who said they preferred “very hot” drinks. It did not measure those beverages at 65°C (149°F).

The 65°C number comes from IARC’s older definition of a very-hot beverage.

But correcting that claim should not obscure the larger evidence. A prospective cohort that actually measured drinking temperature found increased SCC risk at 60°C (140°F) and above, and its estimates for 60–64°C (140–147°F) and ≥65°C (≥149°F) were remarkably similar. That is much more consistent with heat-related risk along a continuum than with an abrupt threshold at exactly 65°C.

For someone looking for a practical number, letting beverages cool below roughly 60°C (140°F) is a reasonable precaution supported by directly measured prospective evidence. It should not be presented as a proven line between danger and safety.

And in perspective, temperature is only one modifiable factor. For SCC, avoiding tobacco and limiting alcohol remain particularly important established ways of reducing risk.

References and Further Reading

Primary studies

• Papier et al., “Hot Drinks and Oesophageal Cancer: Prospective Analyses in Around 1 Million UK Adults” — International Journal of Cancer (2026) — The primary September 2026 study combining the Million Women Study and UK Biobank; source of the RR 3.17 result and the 977,282-participant analysis.

• Islami et al., “A Prospective Study of Tea Drinking Temperature and Risk of Esophageal Squamous Cell Carcinoma” — International Journal of Cancer — Golestan Cohort Study of 50,045 participants; particularly important because tea temperature was objectively measured.

• “Hot Beverage Intake and Oesophageal Cancer in the UK Biobank: Prospective Cohort Study” — British Journal of Cancer (2025) — Earlier UK prospective study showing temperature- and intake-related SCC associations in UK Biobank.

Cancer and public-health authorities

• Cancer Research UK — “New UK Study Links Hot Drink Consumption to One Type of Oesophageal Cancer” — Funder’s explanation of the 2026 study; explicitly states that actual beverage temperatures were not measured and provides absolute-risk context.

• International Agency for Research on Cancer — Drinking Very Hot Beverages — IARC explanation of its >65°C (>149°F) definition and Group 2A classification.

• IARC Monographs Volume 116 — Exposure Data for Coffee, Maté and Very Hot Beverages — Detailed basis for distinguishing “hot” beverages from those classified as “very hot.”

Supporting evidence and practical temperature research

• Verst, Winkler and Lachenmeier, “Dispensing and Serving Temperatures of Coffee-Based Hot Beverages” — Peer-reviewed exploratory measurements of real-world coffee temperatures, cooling time and cold-milk effects.

• “Consumption of Hot Beverages and Foods and the Risk of Esophageal Cancer: A Meta-Analysis of Observational Studies” — BMC Cancer — Earlier pooled evidence showing a stronger association with SCC than adenocarcinoma.

• Science Media Centre — Expert Reaction to the 2026 Hot-Drinks Study — Independent methodological context emphasizing that the findings are strong associations rather than randomized causal proof.

Editorial currency note: This article reflects the evidence available through September 9, 2026. The exact temperature-response relationship remains an active research question, and public-health classifications or recommendations may be updated as new measured-temperature studies become available.