The director of the nation’s top health research agency pulled the plug on a study of alcohol’s health effects without hesitation on Friday, saying a Harvard scientist and some of his agency’s own staff had crossed “so many lines” in pursuit of alcohol industry funding that “people were frankly shocked.”
A 165-page internal investigation prepared for Dr. Francis Collins, director of the National Institutes of Health, concluded that Kenneth J. Mukamal, the lead investigator of the trial, was in close, frequent contact with beer and liquor executives while designing the study.
Buried in that document are disturbing examples of the coziness between the scientists and their industry patrons. Dr. Mukamal was eager to allay their concerns, respond to their questions and suggestions, and secure the industry’s buy-in.
Dr. Mukamal has repeatedly denied communicating with the alcohol industry while planning the trial, telling The Times last year that he had “literally no contact with the alcohol industry.”
In a statement issued on Friday, Dr. Mukamal said he and his colleagues “stand fully and forcefully behind the scientific integrity” of the trial.
But the report documented conference calls he held with alcoholic beverage companies and lengthy memos written in response to their concerns, long before the N.I.H. even announced it would sponsor the trial.
Beer and liquor companies offered their own suggestions for carrying out the trial. Carlsberg, the Danish beer company, at one point suggested that clinical trial centers be established in Russia, China and Denmark. (A trial site was located in Copenhagen, but not in Russia or China.)
The strategy of engagement with industry was effective. Five large beer and liquor companies eventually agreed to pick up most of the $100 million tab for the 10-year-long randomized trial.
The study was intended to test the hypothesis that one drink a day is better for one’s heart than none, among other benefits of moderate drinking. But its design was such that it would not pick up harms, such as an increase in cancers or heart failure associated with alcohol, the investigation found.
Scientists who designed the trial were aware it was not large enough to detect a rise in breast cancer, and acknowledged to grant reviewers in 2016 that the study was focused on benefits and “not powered to identify negative health effects.”
“Clearly, there was a sense that this trial was being set up in a way that would maximize the chances of showing a positive effect of alcohol,” Dr. Collins said last week as he accepted his advisers’ recommendation to terminate the trial.
“Understandably, the alcoholic beverage industry would like to see that.”
But even as Dr. Mukamal and staff from the N.I.H.’s National Institute on Alcohol Abuse and Alcoholism shared details of the trial with liquor and beer manufacturers, institute staff were engaged in subterfuge within the federal agency.
They denied that discussions with industry were taking place, the report found, and withheld pertinent information from “Building 1,” the location of Dr. Collins’s office.
When setting up a meeting about the study, a staff member sent an email on June 30, 2013, warning that one senior official who would be present “knows NOTHING about the possible funding source, and we should probably keep it that way for now.”
All along, senior staff at the alcohol institute were well aware the industry was likely to use the study’s findings for marketing and promotion purposes.
In June 2013, institute staff drew up a business plan making the case for the industry’s financial support of the study. “Once the data are released into the public domain via publication,” it said, “the industry can use that information to make or bolster whatever arguments and claims they choose.”
It continued, “At that point, the N.I.A.A.A. and the N.I.H. are out of the process.”
If the study failed to find health benefits in moderate drinking but provided no evidence of harm, the results still would be a boon for the beverage makers. The findings would counter a 2014 World Health Organization edict that no level of alcohol consumption is safe because it raises the risk of cancer.
Indeed, on Feb. 26, 2015, Dr. Mukamal and alcohol institute staff weighed in on an email to an industry group, editing it to say that one of the important findings of the study “will be showing that moderate drinking is safe.”
“As we discussed, this will be the first R.C.T., i.e. ‘gold standard’ evidence of this,” they added, “and it is important to answer statements made by W.H.O. and others that ‘no level of alcohol is safe’ with certainty.” (R.C.T. refers to a randomized controlled trial.)
Alcohol, which is classified as a carcinogen, is linked to a slight increase in breast cancer risk at even one drink a day. One of the main criticisms of the alcohol trial from the start was that it was not large enough, and would not last long enough, to detect a rise in cancers, which are slow-growing, among drinkers.
Barry S. Kramer, director of cancer prevention at the National Cancer Institute, who reviewed the trial design as part of the advisory committee’s report, agreed with this assessment. “The trial is set to show the benefit while missing the harm” of alcohol consumption, he wrote.
Indeed, notes from Dr. Mukamal’s communications with alcohol companies suggests he wanted to allay their concerns that the trial might pick up a rise in breast cancer.
The trial protocol called for enrolling adults aged 50 and older who are at high risk for heart disease. But breast cancer rates also rise at this age.
“Will breast cancer be one of the safety outcomes measured?” was one question from an alcohol industry-supported group.
“Is the intention to publish results even if they are less desirable, e.g. negative or mixed?” asked a representative of Suntory, a company that sells wine, beer and whisky, during a Dec. 8, 2014, conference call with Dr. Mukamal.
“Yes,” Dr. Mukamal responded, according to notes from the call published in the advisory committee report. But, he added, “We are not enrolling people of high risk for breast cancer.”
Earlier in the same call, an alcohol industry representative asked whether it was possible that another health agency, like the federal Centers for Disease Control and Prevention, might “pick up the idea” of an alcohol trial “and go a different way with it?”
“Yes, we have been approached by different groups, and some of these groups have very different motives, e.g. investigating the relationships between alcohol and breast cancer,” Dr. Mukamal responded.
Perhaps the earliest signs of bias could be seen in emails between N.I.A.A.A. staff as they debated what they should call the trial. Scientists often come up with acronyms that serve as nicknames for studies and are shorthand for long, complicated scientific titles.
One staffer informed another on June 13, 2013, that the name of this clinical trial would be “Cheers” — short for “Cardiovascular Health Effects of Ethanol Research Study.”
“And it will be a new drinking game,” the official added. “Every time you hear it, you must assume its [sic] a toast, and so have a drink.”
That idea was later abandoned in favor of the more sober acronym, M.A.C.H., which stood for “Moderate Alcohol and Cardiovascular Health.”
Correction:
Because of an editing error, an earlier version of this article misstated the meaning of the acronym R.C.T. It means randomized controlled trial, not randomized clinical trial.
