Did the military secretly doctor its anthrax vaccine?
In Vaccine A: The Covert Government Experiment That's Killing Our Soldiers—and Why GI's Are Only the First Victims, Gary Matsumoto struggles for nearly 300 pages to justify his book's subtitle. Matsumoto's contention, as he first reported in a 1999 Vanity Fair article that triggered a congressional hearing, is that the U.S. military's anthrax vaccine causes a constellation of autoimmune diseases and symptoms such as lupus, rashes, hair loss, and aching joints that fall under the rubric known as Gulf War Syndrome. He makes the case that the military secretly spiked the vaccine with squalene, a fat naturally found in the body that, when injected as part of a vaccine, supercharges the immune system's ability to make antibodies. He accuses the Department of Defense of slipping this experimental vaccine past the Food and Drug Administration and involving unwitting troops in a human test that Matsumoto repeatedly equates with Nazi experimentation and the infamous Tuskegee syphilis study. The vaccine harmed "possibly tens of thousands" of troops, he warns, and millions of doses now are being prepared for public use.
To be sure, a myriad of real problems have stained the reputation of the military's anthrax vaccine, and Matsumoto thoroughly documents them. (To read more about the anthrax vaccine, click
Developed in the 1950s and approved by the FDA in 1970, the anthrax vaccine requires an extraordinary six doses—most vaccines work with three or fewer—and contains what's called "protective antigen," a toxin made by the anthrax bacterium. In theory, injections of the protective antigen teach people's immune systems to make antibodies that later will protect them should they become exposed to anthrax. To help rev up the antibody response against the protective antigen, the anthrax vaccine also contains a chemical called an adjuvant, which essentially serves as an irritant to command the immune system's full attention. The adjuvant used in the FDA-licensed anthrax vaccine is aluminum hydroxide, or alum, which has a long safety record in other vaccines.
Adjuvants are the neglected stepchild of the vaccine business: There's something almost alchemical about the nonspecific way in which they work, and scant effort has gone into making a more powerful adjuvant than alum. But for more than 50 years, a small group of researchers has tinkered with adjuvants made from oils, and squalene is one of the modern concoctions. Although it is not currently in any FDA-licensed vaccines and has caused autoimmunity in some animal experiments, it is, as we learn two-thirds of the way into the narrative, in a flu vaccine that's licensed in Europe—and, according to Chiron Corporation, the vaccine's maker, has been safely injected into more than ten million people. *
In the late 1980s, military researchers demonstrated that a single dose of a vaccine made from protective antigens and squalene prevented a particularly nasty anthrax strain from harming guinea pigs and mice. On the eve of the Persian Gulf War in October 1990, the Department of Defense did not have enough anthrax vaccine for its troops, and the DoD's pressing deployment plans created a time crunch—ideally, the vaccine requires six shots over 18 months to develop a full antibody response. A military task force with the eerie code name of Project Badger set out to find companies that could quickly manufacture more doses, and, according to Matsumoto, this provided a perfect opportunity for the DoD to throw in some squalene and see whether the animal results translated to humans. "More immunity in less time with one shot—just what the Army needed," Matsumoto writes.
According to Matsumoto, the urgent situation was the military's motive for using squalene. Matsumoto attempts to bolster his theory by stressing that most Gulf War troops received only two shots at most, which, without the added squalene, may not have provided protection. The anthrax vaccination program was also shrouded in secrecy—most shots were not marked on their recipients' immunization cards. When they were, the shot was referred to as "Vaccine A" or by some other coded name.
But there are simple explanations for Matsumoto's allegations—which not a single person with inside information ever suggests actually took place—and in the end Vaccine A smells like a conspiracy theory. First, the military had no idea that the war would end in a mere 42 days, so it seems prudent to at least start a vaccination program—wouldn't some immunity be better than none? As for the military's obsession with keeping the vaccination hush-hush, there is an obvious tactical advantage to not revealing that you've vaccinated your troops against anthrax.
So how did the vaccine's supposed link to Gulf War Syndrome come about? In 1994, Pam Asa, an immunologist in Memphis who had only two obscure scientific papers to her name, noticed that several male Gulf War veterans seen by her physician husband had lupus, an autoimmune disease that rarely occurs in men. Asa soon teamed up with virologist Robert Garry of Tulane University; the pair developed a test to measure the amount of squalene antibodies in blood, which they claimed to have found in patients with Gulf War Syndrome. Years after circulating their data to Congress and the DoD, they finally published their evidence in February 2000; military scientists quickly blasted their findings for relying on questionable methodology and "circular logic." Scientists from a private company hired by the DoD also tested the vaccine and did not found any squalene.
However, a startling finding then surfaced that seemed to give the theory credence: In response to the 1999 Vanity Fair article, FDA scientists tested specific lots of the anthrax vaccine that Matsumoto cited and reported that they did find trace amounts of squalene.
Yet the Aha! factor here, which is scientifically valid and has the illusion of being a smoking gun, loses its impact when you look closely at the findings. The FDA scientists found levels of squalene in the vaccine that measured between 10 parts and 83 parts per billion. Squalene found naturally in human blood, the DoD responded, measures 250 parts per billion; the trace amounts, they suggested, could have come from a technician's fingertip. The licensed European flu vaccine, moreover, has 20 million parts per billion. As the DoD archly notes on a Web site it runs that is dedicated to the Anthrax Vaccine Immunization Program, the difference between the trace amounts of squalene found in the vaccine and the concentration needed to make squalene work properly as an adjuvant "is like the difference between a teaspoon of oil and 2,000 pounds of mayonnaise."
Matsumoto also notes the FDA's claim that it found similar trace levels of squalene in a tetanus vaccine made by Wyeth and a diphtheria vaccine made by Connaught. To the FDA, this simply showed that tiny amounts of squalene can contaminate many vaccines. To Matsumoto, this was a hot lead: Neither of these products is on the FDA's list of licensed vaccines, raising the prospect that they were actually experimental, unlicensed vaccines that also had squalene intentionally added to them. How could it be that the FDA researcher who presented these findings—"in one of those moments of almost cosmic irony, testifying to the House Committee for Government Reform," Matsumoto writes—omitted these details? Here's how: Matsumoto mistakenly wrote that the FDA tested Wyeth's tetanus vaccine and Connaught's diphtheria vaccine (it's actually Connaught's tetanus vaccine and Wyeth's diphtheria vaccine), and then did not note that Connaught is part of what's now called Aventis Pasteur, which has an FDA-licensed tetanus vaccine. Cosmic irony, indeed. *
But in Matsumoto's mind, the deception goes deeper: The DoD, the National Institutes of Health, and the FDA conspired to quash concerns about squalene because it was the magical ingredient needed to make future vaccines—not just against bioweapons, but against AIDS and cancer. He writes that "by questioning the safety of squalene, Asa imperiled more than 80 percent of the existing NIH-sponsored clinical trials for vaccines to prevent HIV." This is fiction. Here is a 2001 list of AIDS vaccines in clinical trials and in the pipeline. Only one product, not yet tested in humans, uses squalene, and many don't use any adjuvants at all. Chiron Corp. did use squalene in earlier human tests of an experimental AIDS vaccine, but that project crashed and burned because of the unimpressive results with the HIV ingredients in the vaccine, not the adjuvant.
The shaky premise of Vaccine A falls apart completely when Asa and Garry—and, separately, military researchers—compare squalene antibody levels in people who received the anthrax vaccine and controls who did not. Logically, if the vaccine contained significant amounts of squalene, vaccinated people should have higher levels of squalene antibodies than the unvaccinated. Asa and Garry found the antibodies in eight of 25 vaccinated people (32 percent) and three of 19 controls (15.8 percent). "This difference is not statistically significant in this size sample," they reported in a 2002 paper. A military study tested for squalene antibodies in more than 700 people from three separate groups, one of which included 34 people who had received the anthrax vaccine. Each group had the antibodies, and there again was no significant difference between them. Matsumoto challenges the quality of the military data near the book's end, but at that point it's like listening to a Yankee fan at a bar who still can't accept the ump's interference call on A-Rod.
Vaccine accidents do happen, they do get covered up, and they deserve close scrutiny. One of the most famous vaccine accidents is the contamination of the first polio vaccines with a monkey virus, and a new book on the accident by Debbie Bookchin and Jim Schumacher, The Virus and the Vaccine, has many parallels with Vaccine A. But the research in this book is meticulous, and Bookchin and Schumacher's tone is so measured and thoughtful that my mind opens to their controversial thesis (that the monkey virus accidentally injected into people through polio vaccines causes human cancers). I'm not convinced that the monkey virus does, in fact, cause harm, but I am convinced that the question merits serious attention. Matsumoto, in contrast, with his sloppy errors, remote possibilities trumped up as facts, and outright dismissal of evidence that doesn't support his thesis, leaves me groaning.
Science is provisional. Data can surface tomorrow that overturns everything we think we know. But given the evidence to date, did the military spike its current anthrax vaccine with squalene and unintentionally cause Gulf War Syndrome? No. And part of my conviction comes from the fact that Matsumoto tried so hard to make this case and failed.
Corrections, Nov. 24, 2004: Jon Cohen did not intend to imply that the book Vaccine A stated that a flu vaccine licensed in Europe "has been safely injected into tens of millions of people." In fact, the book questions the safety of that vaccine, but does not explicitly state how many people it has been injected into. Cohen stated that the vaccine "has been safely injected into tens of millions of people." This figure was taken from the Department of Defense Web site; according to Chiron Corporation, the vaccine's maker, the vaccine has been safely injected into more than ten million people. (Return to the corrected sentence.)
Jon Cohen incorrectly wrote that Wyeth made an FDA-licensed tetanus vaccine, and Connaught made an FDA-licensed diphtheria vaccine. The error, however, was originally made by Gary Matsumoto in Vaccine A, and Cohen did not notice it. In fact, Wyeth made an FDA-licensed diphtheria vaccine and Connaught, which is part of what's now called Aventis Pasteur, made an FDA-licensed tetanus vaccine. (Return to the corrected sentence.)
Jon Cohen writes for Science magazine. You can reach him at firstname.lastname@example.org.
Photograph of the syringe on Slate's home page from Corbis.