Does alcohol cause cancer? It depends upon whom you ask.
The International Agency for Research on Cancer (IARC), a branch of the United Nations' World Health Organization (WHO), issued a statement in 1987 that said, "the consumption of alcoholic beverages is causally related [emphasis in original document] to the occurrence of cancers of the oral cavity [mouth], pharynx [throat], larynx [voice box], esophagus and liver which tend to be rare [emphasis added]."
The IARC study found "no indication that alcohol increases the risk for cancers of the stomach, lung, urinary bladder, kidney, ovary, prostate, lymph system or blood and blood marrow." It said it found "suggestive but inconclusive" links with cancers of the rectum.
A 1981 article by Doll and Petro in the Journal of the National Cancer Institute estimated that perhaps 3 percent of cancer deaths in the U.S. were related to alcohol. That study also estimated that another 3 percent of cancers were caused by "geophysical" factors -- such as living in Denver as opposed to Memphis - (Denver's higher altitude means that the background radiation from cosmic rays is higher than Memphis.)
But moderate consumers of alcohol don't seem to differ much from the population as a whole, except that their overall risk of dying from all causes is 10 percent lower than abstainers or heavy drinkers.
According to Dr. Emanuel Rubin, M.D., who is a professor and chairman of the Department of Pathology and Cell Biology at the Jefferson Medical College of Thomas Jefferson University, Philadelphia, and also Adjunct Professor of Biochemisty and Biophysics at the University of Pennsylvania Medical School:
"A review of the evidence relied upon by the IARC working group reveals that,
-- "the putative causal relationship between alcohol and cancer is based upon inconsistent and limited epidemiological evidence,
-- "that the animal studies fail to support a causal link and
-- "that no satisfactory explanation for a possible causal relationship has been elucidated."
Dr. Rubin is not alone in his conclusions. He said that one of the biggest flaws in the IARC's conclusion is that studies of laboratory animals exposed to very high levels of alcohol consistently fail to develop cancer. Even the few that have shown a relationship have done so only at the very highest levels.
"The animal data support the hypothesis that the observed statistical associations [between alcohol and cancer] are due to the confounding factors [unconnected with alcohol] rather than the alcohol," Dr. Rubin said. "Test animals have been exposed to considerably greater amounts of alcohol than have ever been reported in epidemiological studies (the equivalent of four 25-ounce bottles of 86-proof whiskey a day for life for a 7O kg [155-pound] human) ... The animal studies have failed to support a conclusion that alcohol is carcinogenic at any level."
As you read in this book's chapters on breast cancer and interpreting study results, epidemiological studies involve the assembly of vast amounts of data -- much of it self-reported (therefore somewhat inaccurate) -- which is analyzed using a variety of mathematical and statistical techniques. Before a scientific or public health conclusion can be based on epidemiological data, scientists use a number of other key tests to assess the credibility of the epidemiological data. Those key factors include:
-- Is the conclusion verified by laboratory animal studies?
-- Are the results consistent from one study to another?
-- Is the conclusion supported by the total body of research?
-- Is the effect dose-related?
-- Has the epidemiologist corrected for biases such as picking a representative sample and adjusted for confounding factors such as smoking and other related diseases?
-- Is there a logical biological mechanism to explain the association?
Dr. Rubin pointed out that animal tests repeatedly have not supported the alcohol-cancer association and that "estimates" of cancer risk in this case have not been verified.
"At least one commentator who posits a causal link between alcohol and cancer has estimated [emphasis in original document] that 9 grams per kilogram of body weight per day of ethanol administered to rats would cause hepatocellular carcinomas [liver cancer] in 50 percent of the animals. In fact, rats fed alcohol at higher levels for their entire lifetimes, and subhuman primates intoxicated for more than six years have failed to show increased liver cancers," said Dr. Rubin.
It's significant to note that 9 grams is approximately one "drink" and that a 70-kg person (155 pounds) would have to drink 70 drinks per day to reach this level -- acute alcohol poisoning and death would result before this level was reached.
The IARC studies also fail the tests of consistent results and support by the total body of research. Dr. Rubin said that the IARC was guilty of stacking the deck toward their conclusion by ignoring data that did not fit.
Dr. Rubin also points out that, although having cirrhosis of the liver puts a person at increased risk of liver cancer, "99 percent or more [of abusers] do not [emphasis added] get liver cancer.
"It is impermissible to select from these studies only those data which support an association and ignore those that do not. For example, the four prospective studies [used by the IARC] which report associations between the ingestion of alcoholic beverages and cancers of the oral cavity, larynx, esophagus and liver come to different conclusions."
One of these studies found an association with lung and rectal cancer and not with esophagus, stomach, colon, prostate or liver cancer. In a similar vein, other studies found an association with one type of cancer not found in other studies while finding no associations which were found in those other studies.
This lack of consistency and the failure of the total body of research to support the alcohol-cancer link may point to biases in the study, particularly the failure to control for known cancer risks.
For example, Dr. Rubin said that the primary external cause of liver cancer is the hepatitis 13 Virus; but "none [emphasis added] of the studies purporting to relate the incidence of liver cancer to alcohol consumption have properly controlled for hepatitis B. Such an omission is particularly significant because there is reason to believe that the chronic alcohol abuser population has an increased prevalence of hepatitis B."
Dr. Rubin also pointed out that to heavy drinkers were almost always tobacco smokers as well and that it was impossible to get a statistically significant sample of heavy drinkers who did not also smoke. Thus, studies that concluded that consumption was related to lung, mouth, throat and larynx cancers are probably flawed by this inability to correct.
All of the other types of cancer which the IARC concludes are linked to alcohol consumption are also marred by the failure to correct for known causal links. These uncorrected-for risks include:
-- mouth and oral cavity cancer risk: tobacco use, poor dental hygiene and poorly fitting dentures;
-- esophageal cancer risk: nutritional deficiencies, consumption of hot (non-alcoholic) drinks, socioeconomic factors;
-- colon/rectal cancer risk: diet, lack of fiber;
-- breast cancer risk: causes unknown (but not linked to alcohol by the IARC);
-- stomach cancer risk: diet, ulcers.
Finally, the studies relied upon by the IARC do not demonstrate any consistent dose relationship (except at the very highest levels of abuse) and the few mechanisms proposed as possible biological mechanisms to explain alcohol's effects have not stood up under scientific scrutiny.
Among the three most common proposed mechanisms is that alcohol is a promoter, not a prime cause, of cancer. "If it were a promoter," Dr. Rubin said, "alcohol would be expected, under these theories, to cause cell replication. In fact, alcohol does not induce cell proliferation, and, under some circumstances actually inhibits it."
A second theory proposes that ingestion of large amounts of alcohol (far above that of moderate consumption) causes certain enzymes, known as P-450, to act on certain non-carcinogenic molecules turning them into carcinogens. "But this theory of mechanism fails to take account of the evidence that the presence of alcohol itself inhibits the activities of the P-450 class of enzymes," said Dr. Rubin.
Finally, a third theory involves compounds produced by the metabolism of alcohol. The chain of chemical reactions in alcohol metabolism first produces acetaldehyde, then acetate and finally water and carbon dioxide. Acetate, carbon dioxide and water are all normal metabolites produced when the body uses most food sources. Acetaldehyde is a short-lived intermediate compound produced when alcohol is first metabolized and has an established role in metabolism. Acetaldehyde has been extensively studied and acts as a carcinogen only when inhaled in high concentrations.
Among other medical experts who agree with Dr. Rubin are:
-- Dr. Parviz Pour, M.D., Professor at the University of Nebraska's Epply Institute for Research in Cancer, "... after careful analysis of the reported data in epidemiologic and laboratory studies on the possible association between alcohol and esophageal cancer, I have concluded that alcohol cannot be considered a carcinogen or co-carcinogen of the esophagus nor can it be regarded as a promoter of esophageal carcinogenesis."
-- Dr. Michael Anderson, M.D., Senior Research Fellow, Cancer Research Campaign (the British equivalent of the American Cancer Society), who said that many of the IARC studies were "poorly designed or reported ... and that any findings in the better-designed and reported studies are likely to reflect the fact that the population of chronic heavy alcohol users exposed themselves to known or unknown risk factors (poor diet, smoking, infection) not adequately accounted for in the studies."
-- Marvin Goldman, Ph.D., Professor of Toxicology at the University of California, Davis, who said, "the overwhelming and consistent trend in the laboratory experiments is that alcohol is not a carcinogen."
-- Paul Levy, Sc.D., Professor and Director of the Epidemiology and Biometery programs at the University of Illinois, School of Public Health, who found the IARC's conclusions were based on flawed methodology.
-- Dr. John Doull, M.D., Ph.D., Professor and Director of the Center of Environmental and Occupational Health at the University of Kansas Medical Center who said that the "experimental studies do not provide a basis for any conclusion other than there is, at this time, no demonstration of the carcinogenicity of alcohol."
The answer on the link's between cancer and alcohol seems to be a contest of dueling scientists. Conclusions in this matter, like those in other areas, must be individual ones based on the best available information. That best information may indict heavy drinking, but does not indict moderate consumption, particularly when alcohol is consumed in the absence of know cancer-causing factors such as tobacco smoking.
Making a decision is often complicated, as it is here, by scientific investigators who may have a point of view they wish to support. In this case, the World Health Organization has a publicly stated policy goal to reduce -- and if possible, eliminate -- alcohol consumption of all sorts. To link alcohol consumption with cancer would advance that organizational policy.
The IARC and other anti-alcohol activities of the WHO are also heavily funded by Islamic countries who oppose alcohol on religious grounds. On the other hand, Dr. Rubin and his colleagues have, on occasion, conducted studies that have been funded by portions of the alcoholic beverage industry.
In assessing the possible (but by no means causal) associations and the possible points of view behind them, it is very important to realize that the research conducted by Dr. Rubin and his associates has been reviewed by their scientific peers (unassociated with any industry-funded research) and published in respected scientific journals. By contrast, the IARC's conclusion was not peer-reviewed, has not been published in a respected scientific journal and must be viewed as a government statement which reflects both scientific and policy considerations.