PNAS Publication Of AIDS Article Spurs Debate Over Peer Review

Some members of the National Academy of Sciences no doubt were shocked this past February when the latest edition of their thick, pale gray journal—Proceedings of the NAS— arrived in their mailboxes. Here was the National Academy—the most respected, and surely the most cautions, scientific body in the United States—publishing in its very own “house organ” the work of Peter Duesberg, the respected but controversial University of California, Berkeley, retrov

Apr 3, 1989
Anthony Liversidege

Some members of the National Academy of Sciences no doubt were shocked this past February when the latest edition of their thick, pale gray journal—Proceedings of the NAS— arrived in their mailboxes. Here was the National Academy—the most respected, and surely the most cautions, scientific body in the United States—publishing in its very own “house organ” the work of Peter Duesberg, the respected but controversial University of California, Berkeley, retrovirologist who has been arguing for two years now that HIV is in no way the cause of AIDS. This opinion has earned Duesberg little more than opprobrium throughout the AIDS research community, since his ideas are totally at odds with the consensus view in the biomedical field. And yet, there he was, in PNAS [volume 86(3), pages 755-764], not only raising the same heretical notions all over again, but doing so at greater length than ever and with 196 references.

However, after several rounds of published—and public—debates over Duesbeg’s views, it is not merely the findings he presents in the recent PNAS that are of concern; it is also the remarkable set of events that led up to the publication of this latest paper on the etiology of AIDS. These events may well lead scientists to question just what constitutes fair play in the science publishing arena. For as it ‘turned out, PNAS forced Duesberg’s paper to clear the hurdles ohio fewer than six separate peer reviews, even though the journal typically requires no formal screening of papers written by academy members—and Duesberg, one should not forget, is an academy member. The challenging procedure left a paper trail of no less than 60 pages of combative correspondence between academician Dues- berg and the journal’s editor, Igor Dawid.

Further, the 10,000 word-manuscript appeared in print despite the fact that the three main reviews ordered by PNAS were unfavorable. This raises other intriguing questions, such as: Had the peer review process been a sham? Or had the negative reviews failed to shoot down arguments that have been shunned and scorned by every key player in AIDS research? And if that were the case, what is being said about peer review? Is the system vindicated by having permitted a controversial notion to be aired in public? Or is it deficient for having allowed further dissemination of what most scientists consider a wrong-headed notion and what some feel is actually a public danger?

In Peter Duesberg’s case, the answers to these important questions may well be embedded in his running battle with the AIDS research establishment—a battle that began long before his submission of the lengthy (two pages over the usual limit) paper now published in the academy’s journal. This most recent effort, “Human immunodeficiency virus and acquired immunodeficiency syndrome: Correlation but not causation,” expands on point a Duesberg first raised in the March 1987 issue of Cancer Research [pages 1199-1220]. Tempered in public debate and tested since then, the Duesberg arguments reappeared most notably in the July 29, 1988, issue of Science: There, Duesberg’ was pitted against a seemingly formidable trio: Robert Gallo, chief of the Laboratory of Tumor Cell Biology at the-National Cancer Institute; his colleague Bill Blattner; and Nobelist Howard Temin, a retrovirologist at the University of Wisconsin’s McArdle Laboratory. The result was a four-page confrontation labeled “HIV Is/Is Not The Cause Of Aids.”

In each case, Duesberg boldly— some might say brashly—pronounced his conviction that “the AIDS virus is just the most common among the occupational viral infections of AIDS...rather than the cause of AIDS.” He points out that the virus is hardly detectable after the initial period of infection, that it is not biochemically active when it supposedly much later causes the disease, that the body can easily regenerate any T cells killed by the virus, and that the lab evidence for the killing of T cells by the virus is no model for its claimed lethality in the body. An authority on the genetic structure of retroviruses, Duesberg also asserts that the virus is genetically inadequate to the challenge of producing a new effect years after infection.

Duesberg suggests that AIDS is almost certainly not caused by any one virus or microbe but “may be caused by new combinations of conventional pathogens, including acute viral or microbial infections and chronic drug use and malnutrition.” He says studies are needed to unbundle AIDS diseases and see “how the nature, frequency, and duration of AIDS risks generate risk-specific diseases.”

Gallo and company roundly reject such arguments-, insisting—at least until recently—that HIV is both a necessary and sufficient cause of the fatal immune deficiency, which they say kills many, perhaps all AIDS patients. Most recently, however, in a magazine interview [Spin, March 1989, pages 54-81], Gallo, a National Cancer Institute virologist, seemed to have stepped back somewhat from his total opposition to Duesberg ‘s view, and to concede that the virus in some- variants may be necessary, but in itself is not sufficient to trigger the deadly syndrome. In the interview, Gallo said: “You and I could probably live with the Virus for 30 years and die of old age.”

Still, Gallo and his fellow researchers such as Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, do their level best to dismiss Duesberg as way off base. Indeed, only a handful of prominent scientists will defend Duesberg’s efforts. One of these is microbiologist and Nobel Prize winner Walter Gilbert of Harvard University. Gilbert says, “At one level some of what Peter does is exaggerated, just for the forms of argument. At another level-, he has made a very serious argument on the role of HIV in AIDS. It is good to have it questioned, and argued. I absolutely do consider it a valid debate, and I am glad his article has appeared in the [NAS] Proceedings. The editors made it too much of a rough going.”

Rough going it was. Duesberg first mailed his paper to the academy last June. What he expected were the - usual delays of a fortnightly scholarly journal; what he didn’t expect was four months of conflict—first, an outright rejection by the editor and then, later, under the regime of a new editor, months of battling over reviewers’ comments. Normally, papers contributed to the PNAS by academy members—including more than 20 by Duesberg himself—are printed without formal peer review, although they are run past one informed colleague before submis- sion. Duesberg had his paper read by fellow academy member Harry Rubin and also Steven Martin, both retrovirologists at Berkeley. This, he expected, would be sufficient for publication.

However, Maxine Singer saw it differently. As editor of PNAS at the time the paper was first submitted in mid-June of last year, she saw the essay as too similar in its conclusions to Duesberg’s earlier paper in Cancer Research. “I felt on the basis of lack of originality that it just didn’t belong in the Proceedings,” she explains. “I felt he had published his view in full in that prior article and he didn’t add very much.”

After consulting with referees, Singer decided to reject the paper. She returned Duesberg’s manuscript to him the same week she stepped down from the part-time post she held as PNAS editor at the end of June.

This didn’t stop Duesberg, who. resubmitted his paper to Singer’s successor, Igor Dawid. The new PNAS editor took a different tack: He sent the paper out to three referees in what was now to be a formal review process. The three official reviews were all unfavorable, liberally attacking alleged “non sequiturs,” “misleading arguments,” “completely incorrect statements,” “conceptual errors,” and “partial truths.” For all that, in 14 single-spaced pages, they provided only two citations from the published literature to back their criticisms.

Duesberg’s response was of two kinds: On the one hand, he irritably challenged what he called “worthless, undocumented opinions.” (David requested more references but elicited only two, both dating from before AIDS was identified.) And on the other hand, Duesberg made changes and clarifications in his paper that sufficed to accommodate each reviewer objection.

But none of this came easily. There were months of sometimes tense exchanges between Duesberg and David over what was expected and what was reasonable. The 60 pages of correspondence that grew up between David and Duesberg make lively reading.

In the end, though, Duesberg acknowledges Dawid’s “substantial efforts” in developing a paper he is “proud of,” writing that “despite the good battle, there are no hard feelings.” And Dawid, for his part, authorized the paper for publication in the February issue, along with a caveat of sorts—a promise to present a divergent view in a future issue.

Two weeks after the public dissemination of the paper, the major HIV proponents either wouldn’t comment on its contents or were claiming not to have read it properly. Anthony Fauci, director of the NIH Office of AIDS Research, said he preferred not to make comment. Fauci, by the way, may have been one of David’s reviewers, judging from the fact that some of the most rhetorical reviewer comments match closely statements Fauci has made in public. To cite one such comment: “What kind of aberrant lifestyle abnormalities does a monogamous heterosexual 55-year-old wife of a hemophiliac have that causes her to develop AIDS?” the reviewer asks, in what is intended to be a swipe at Duesberg’s claim that aberrant lifestyles, and not the virus itself, are key to the condition.

Asked her view of the final piece, Maxine Singer says, “I’m still reading the November issue of the Proceedings. Keeping up is a burden.” She then notes that, after all, she had read the initial version very thoroughly. And then there’s Robert Gallo, who has been asked to write a counter review for PNAS. Gallo told The Scientist that there is a copy on his desk but “No, I haven’t read it. I have to work for a living.” Has he changed his mind about the potential impact of the article? “I haven’t heard a single scientist discuss it for one second,” he says. However, Gallo adds, “If it concludes that HIV is not the cause of AIDS, it is wrong, very wrong.

It does, of course, conclude that HIV is not the cause of AIDS, and that raises an intriguing question for Gallo: Was David forced to ignore the universally negative reviews because the peer reviewers had failed to disprove Duesberg’s contentions—indeed, because they somehow did a bad job? Not according to Gallo. “As a reviewer—and I wouldn’t dare review it, it is too emotional and too strange—I might have accepted it for publication too, even though I didn’t agree with it. The Proceedings is a great journal, but you can’t stop a member from publishing unless it is totally off the wall.”

To support his point, Gallo asserts that Nobelist Linus Pauling published several papers in PNAS, “which were disagreed with by the scientific community,” even though Pauling’s original paper on Vitamin C and cancer was turned down by PNAS in 1972. Gallo affects a nonchalance about Duesberg: “It mystifies me,” he laughs, “that Peter’s [paper] was able to be stopped for such a long time.”

But shouldn’t competent peer reviewers have been able to refute a scientific assertion that, to Gallo, is clearly way off base? Gallo replies: "Maybe they didn’t know every arguement, maybe they didn’t check every last detail. You would have to ask Dawid. But I can tell you one reviewer told me he found it abominable.”

Dawid’s response is guarded, but he cdnfirms that Duesberg was not forced to go along with every last demand of the reviewers.(Duesberg insists that, on the contrary, all critical points raised by the reviewers were answered to Dawid’s satisfaction.) Peer review in this case does not guarantee that the paper is valid, he insists. Says Dawid: “Review means that people have looked at the paper, and commented, and after this the editorial board has decided that the paper could be published. This does not mean that the editorial board of the journal vouches for the accuracy or validity of the published material.”

Despite his own intense familiarity with Duesberg’s manuscript, Dawid, a developmental biologist and chief of the laboratory of molecular genetics at the NIH Institute of Child Health and Development, won’t say what he thinks of the paper or even whether he believes the months of review improved the original manuscript. “I certainly have an opinion,” he says, “but I don’t count myself an expert.” Likening himself to a family physician who leaves the deciphering of an electrocardiogram to a heart specialist, Dawid defers to “the people in the field.”

But then the question becomes: Just how far out does a paper have to be before it will, be rejected? For example, would Dawid have done what Nature editor John Maddox did: that is, would he have published the claims, later debunked by Nature’s own three-man investigative team, of French chemist and pharmacologist Jacques Benveniste, director of INSERM, a medical research laboratory at the University of Paris-South, who asserted that he had proved that water has memory? As Harvard’s Gilbert puts it: “There is Nature happily publishing Benveniste, and everyone objects, saying they are giving some credence to homeopathy. But the editors quite rightly say that appearance an their journal is not a guarantee of truth.”

“The Beneviste paper,” Gilbert concludes, “is a perfect example of a paper that is wildly wrong and contradicts the current view, and yet is published anyway.” In that example, a disclaimer by the editors was published alongside the Beneviste paper. In Doesberg's case, Dawid published this footnote: ‘This paper, which reflects the author’s views on the causes of AIDS, will be followed in a future issue by a paper presenting a different view of the subject.” Gallo has accepted the assignment.

But other academy members who believe the HIV question is settled resist the notion that PNAS would publish “wildly wrong” papers on principle. Instead, they cling to the thesis that Duesberg was able to slip the leash purely because he is an academy member. Virologist Harold Ginsberg of Columbia University and retrovirologist Howard Temin not only join Gallo in saying that in any other journal Duesberg would have had to abide by the views of the negative reviewers called in by Dawid, but they claim that, despite the intense reviewing, the paper still has errors. Publishing the paper after considerable review does not argue for its merit, says Temin. “It argues only for a belief that any member has a right to publish in PNAS.” Says Ginsberg, “If it had been left up to the peer review system, [the paper] would not have been published [since] everybody who read it said it should'nt be." maxine singer however does not share this view. She argues that the paper could have been printed by any journal despite the objections of the reviewers because peer review stops short of assessing interpretation. It is intended to evaluate “the design of experiments and validity of data, but is almost irrelevant at the point of interpretation,” she says, adding that, at the point where data is insufficient to prove a point one way or the other, judgment enters. Says Singer “You are always left with things the data may suggest but not prove. Then it is a matter of judgment, and this is an activity of human beings, where peer review will rarely say what is wrong or right. A reviewer might say ‘I don’t like this interpretation,’ but that would rarely be a reason to decline a paper.” Daniel Koshland, editor of Science, agrees: “Peer review is designed to find out if a paper is basically valid in terms of the solid data presented. The leeway provided for discussion and speculation that goes beyond that is a matter of editorial policy.”

For all the distress the Duesberg paper may have caused its peer reviewers, the broader scientific community has yet to react strongly, at least in public. Before it appeared, Gallo was asked what significance he thought the paper would have. He replied: “Believe me, it is going to have zero impact.” His forecast so far seems to be accurate. Science ran a news item the week of the paper’s release under the heading, “AIDS Paper Raises Red Flag at PNAS.” To Frances Zwanzig, managing editor of PNAS, that headline meant the flap might “end up on the front page of the Washington Post.” It didn’t. Indeed, coverage by the major media has been nonexistent.

Since PNAS does not publish let- ters to the editor, the mail goes to Duesberg, who reports several supportive letters. One is from a former PNAS managing editor, now at the Mayo Clinic. Bernard Forsher calls the article a superb example of sci entific scholarship. We have a lot of laboratory mechanics but few thinkers, a dreadful imbalance.”

How does Duesberg feel about the wringing he went through? “Dawid did a very thorough job,” he. says. “I give him credit for that. He could have behaved more as an impartial arbiter. But all reviews are beneficial if you can survive them.”

To Duesberg, part of his problem is political. “It is harder to find an objective reviewer on this question than an AIDS virus in an AIDS patient. Rubin and Martin (the two virologists at UC-Berkeley] are the only two [objective reviewers] I know in the nation. All the others have companies or contracts or both.”

Final Answer

To many, this statement will be nearly as contentious as Duesberg’s theory on the triggering mechanism of AIDS. But even if he is correct in his perception of the dangers of peer review in a politicized environment, the question remains: Did the system work in his case? The final answer and any possible vindication of peer review must await ajudgment on the correctness of Duesberg’s theory.

Should Duesberg be proved wrong, then peer review will be faulted for failing to prevent bad science from reaching the public. But if he turns out to be right—or even right in part—then the peer review system will have been vindicated, since in this case it will have tested, but not censored, some version of the truth, however unwillingly.

Maxine Singer, who didn’t want to publish Duesberg in the first place, might seem the least likely person to accept this notion. Yet she views the possibility of Duesberg’s turning out to be right in the end with equanimity. Says Singer: ‘The his- tory of science is one of overturning accepted dogma, and I think we are all prepared for that. If we aren,t, I don’t think we are very good scientists.”

Anthony Liversidge is a science writer living in New York City.