Immune Response (Carlsbad, CA) has started arbitration proceedings againstmedical researchers at the University of California (San Francisco) and HarvardUniversity (Cambridge, MA) over publication of negative data from a phaseIII trial of the firm's anti-HIV drug Remune. Opinion is mixed as to whetheror not the company has tried to manipulate unfavorable clinical trial dataabout a potential blockbuster drug, or whether the university researchershave violated a research agreement. The dispute raises questions about inherentconflicts of industry-supported medical research, and highlights the lackof regulatory oversight of research contracts between academia and industry.

Remune, a disabled form of HIV (HIV-1 immunogen), was supposed to stimulatethe body to produce more antibodies to fight the active virus. But analysisof a three-year study of 2,527 patients has found that Remune is no more effectivethan placebo at reducing disease progression. Designed as a “real-world”test, the study looked at patients who may have been switching to, or alreadytaking, other drugs with a different mechanism of action, such as proteaseinhibitors, according to the study's lead researcher, Joseph Kahn, associateprofessor of medicine at UCSF. Patients receiving Remune showed no differencein viral load and a clinically insignificant change in CD4 cell counts comparedwith the control group. Consequently, an independent review board halted thestudy in May 1999 before it was completed, and Kahn and Stephen Lagakos, professorof biostatistics at the Harvard School of Public Health, began preparing tosubmit the results to the Journal of the American Medical Association.

However, Immune Response, which has spent $191 million since 1986 on developingand testing Remune (its only drug in phase III trials), claims its own statisticalanalysis of a random subset of 252 patients who had blood samples taken moreoften (3 months) than in the main study (6 months) show some benefit fromtaking Remune. The company says viral load was “significantly”lower in Remune patients compared with the control group.

Kahn says he included these data in his manuscript, but refused to includeImmune Response's analysis of the data because it was not part of the originalprotocol. Kahn also says that the researchers own analysis of that data didnot show a benefit of Remune in this subset. “This is not data thatcan be reanalyzed,” says Kahn. “Their analysis was post-hoc andunspecified. We felt we had to stay firm about using the statistical methodologythat we specified up front.”

Immune Response tried to assert right of ownership over the entire dataset,and informed Kahn and Lagakos in January 2000 that it would not provide themwith the final dataset for the study. Kahn and Lagakos, however, used interimdata from the independent Data Safety Monitoring Board (accounting for 95%of the results) and submitted the paper. Immune Response began legal proceedingsagainst the pair on September 1, but the paper was published in JAMAon November 1, prompting a 25% fall in Immune Response share price to $4.63.A week later, shares were still trading at $4.78, down 73% from the 52-weekhigh of $18.31 in March and knocking $39 million off the firm's market capitalization.

Immune Response is claiming damages of $7 to 10 million as a result ofthe JAMA publication. “We felt that involvement of a third partythrough arbitration was the only way to ensure that all parties, includingthe clinical trial investigators, would have input on what would be published,“vice president Ronald Moss said in a statement. The company's claim will beheard by the American Arbitration Association, a third-party alternative tocivil court in California.

While officials at Immune Response would not speak to Nature Biotechnology, Agouron Pharmaceuticals (La Jolla, CA), co-developer of Remune and subsidiaryof drug giant Pfizer, stands by Immune Response, and analysts Alan Auerbachof First Security Van Kasper (Los Angeles, CA) trusts their judgement. “Ifthere isn't [any effect on patients], why is Pfizer putting so much moneyinto this?” asks Auerbach. “Are you telling me that Jim Kahn issmarter than Pfizer? I have a problem believing that.”

But Charles Engleberg, a biotech analyst at Americal Securities (San Francisco,CA), says he's had a sell recommendation on Immune Response for three yearsbecause of his underlying skepticism about how Remune is supposed to work.“The patients have huge viral loads and giving antibodies isn't goingto make a difference,” says Engleberg. “I've been following thissince 1993 and the company has been guilty of massaging data all along andall our clients know it.”

A similar incident occurred in August 1997, when Knoll Pharmaceuticals(Mt. Olive, NJ), a subsidiary of BASF, agreed to pay up to $135 million tosettle allegations that it tried to suppress research showing its prescriptionthyroid drug Synthroid is no better than cheaper alternatives—a studythat was published in JAMA in April 1997. Knoll had threatened a lawsuitto stop its publication, and Betty Dong, a UCSF researcher who conducted thestudy, told the journal that Knoll had suppressed her findings for more thansix years. Biomedical firms cover up unfavorable clinical data “moreoften than not,” claims Kahn, “We just don't hear about it.”

Richard Jennings, director of the Wolfson Industrial Liaison office atCambridge University, points out the inherent conflict between profit-seekingcompanies and truth-seeking researchers who rely on them for funding. “It'sa difficult ethical issue,” says Jennings. “The company is tryingto make the best of results with analysts breathing down their necks, whilescientific researchers also feel the heat. There is major pressure for thingsto get perverted.”

Nevertheless, there is currently no governmental oversight of researchagreements between companies and universities. It is generally understoodthat if a company has paid for the research, it owns the data, but it can'trestrict researchers from publishing. For instance, Cambridge University (Cambridge,UK), which receives about £18 million ($27 million) research fundingyearly from private industry, has its own guidelines that allow publicationof all data as long as sponsors have advance notice of the findings. So far,Cambridge has avoided the kind of conflict that has occurred at UCSF but,says Jennings, “No system is perfect.”

Moreover, a Stanford University study, published in the same issue of JAMA, has found that 60% of the US's top research universities lack specificguidelines governing relationships between private companies that pay forresearch conducted on campus (284, 17, 2000). This lack of clear guidancecan cause confusion among industrial partners, competition among universitiesfor corporate money, or erode academic standards, according to study authorMildred Cho, a research fellow at Stanford's Center for Biomedical Ethics.Cho recommends extending some kind of federal oversight to industry-academicresearch agreements, similar to the kind given to patient consent for humanexperimentation. She says the disputes involving Dong, and now Kahn, havecome out into the open “only because these researchers and institutionswere willing to risk multimillion settlement to get information out there—mostpeople are not willing to do that.”