The IARC Announces Water Causes Cancer

Well, drinking water very hot, or other scalding beverages, probably does cause cancer. Earlier this week, the International Agency for Research on Cancer (IARC) issued a press release that one of its working groups had reviewed the data on the carcinogencity of coffee, maté, and very hot beverages, and concluded that maté, which is often served very hot, “probably” causes esophageal cancer. IARC Press Release N° 244, “IARC Monographs evaluate drinking coffee, maté, and very hot beverages” (June 15, 2016). Very hot beverages were rated 2A, for their probably causing human esophageal cancer.

The good news is that “probably” does not mean “more likely than not” in IARC-speak, and the working group was evaluating hazard not risk.[1] IARC classifications do not attempt to quantify the magnitude of risk that may result from exposure to a classified “hazard.” Id. at Note to the Editor. Because all empirical propositions have a probability of being true, somewhere between 0 and 100%, (with P ≠ 0; P ≠ 100%), the IARC classifications of “probably” causing cancer are probably not particularly meaningful.  Everything “probably” causes cancer in this sense. See Ed Yong, “Beefing With the World Health Organization’s Cancer Warnings,” The Atlantic (Oct 26, 2015).

The IARC group’s evaluation of “very hot drinks” accords with the World Health Organization’s Technical Report Series 916 on Diet, Nutrition and the Prevention of Chronic Diseases, which recommends against consumption of scalding hot temperatures. See Anahad O’Connor, “Coffee May Protect Against Cancer, W.H.O. Concludes,” N.Y. Times (June 15, 2016)[O’Connor]. As though people, other than McDonald’s coffee drinkers, needed such a recommendation. The IARC group found no conclusive evidence to implicate drinking cold maté, or maté at temperatures below scalding levels.

An IARC Decision We Can Like a Latte

The Working Group found no conclusive evidence for a carcinogenic effect of drinking coffee, and placed coffee in its category 3, “not classifiable” with respect to carcinogenicity.[2] The working group’s evaluation included over 1,000 observational and experimental studies, including randomized trials, and found no evidence to support the claims that coffee causes human cancer. The IARC also found a good deal of evidence supporting the claim that drinking coffee reduces the risk of various human cancers.

There is a Group 4, for exposures probably not carcinogenic in humans, but in its 45 years of evaluations, the IARC has found only one substance on Planet Earth, which does not cause cancer:  caprolactam.  Perhaps after another 1,000 studies, coffee will reach this exalted category. For now, coffee is unclassifiable with “inadequate” evidence of human carcinogenicity in the IARC’s view.

The New York Times, not particularly expertly, and without supporting citations, declared that the evidence for coffee’s health benefits could not establish actual causation of benefit because the data came from epidemiologic studies.  See O’Conner. This would not be the first time that the New York Times made up things.

In 1991, the IARC evaluated coffee drinking as a “possible” human carcinogen (Group 2B), based upon limited evidence of an association with urinary bladder cancer in case-control studies, and some evidence in experimental animals.[3] This year’s evaluation of coffee as Group 3 thus represents a rare reversal of opinion, in the face of additional evidence, from the IARC.


[1] The IARC Preamble definition of probable reveals that “probable” does not mean greater than 50%. See alsoThe IARC Process is Broken” (May 4, 2016).

[2] See Dana Loomis, Kathryn Guyton, Yann Grosse, Béatrice Lauby-Secretan, Fatiha El Ghissassi, Véronique Bouvard, Lamia Benbrahim-Tallaa, Neela Guha, Heidi Mattock, Kurt Straifon behalf of the IARC Monograph Working Group, “Carcinogenicity of drinking coffee, mate, and very hot beverages,” Lancet Oncology (2016 in press).

[3] IARC, “Coffee, tea, mate, methylxanthines and methylglyoxal,” 51 IARC Monogr Eval Carcinog Risks Humans 1 (1991).

Print Friendly, PDF & Email

Comments are closed.