After languishing in limbo for more than a year, the report of acrucial clinical study appeared in today's Journal of theAmerican Medical Association (JAMA), despite attempts tosuppress it.
Titled "A Controlled Trial of HA-1A in a Canine Model of Gram-negative Septic Shock," the seven-page paper tells in detail howresearchers at the NIH Clinical Center's Department of CriticalCare Medicine tested Centocor Inc.'s sepsis-fighting antibody,HA-1A, in 28 dogs, and found the drug more lethal than theplacebo.
Begun in mid-1991, the double-blinded, controlled study hadlined up 40 specially bred beagles to be inoculated, four at atime, a week apart, with virulent endotoxin-containing E. coli.
Two of each quartet would receive therapeutic doses of themonoclonal antibody, directed at lipid A epitopes on the toxin;the other pair, placebos. All got the full panoply of antibioticsand other therapeutic measures accorded to human patients.
As William Hoffman, senior author of the JAMA paper, toldBioWorld: "When we were doing the study, we thought thedrug was working, because we were blinded as to which groupgot which. Always, half the animals looked worse, with septicshock, than the other half. We thought, 'Gee, this stuff must beworking like gangbusters. Always two real sick ones; two not sosick.' "
One non-member of the team, an independent statistician, wasnot blinded. His daily data revealed that the dogs that diedwere preponderantly those on the Centocor antibody. Afterseven weeks and 28 beagles, he called a halt to the proceedingsbecause additional canine subjects would not alter thesurprising statistical results.
To the investigators' chagrin, only two (15 percent) of the 13testable animals in the antibody group survived day 28. Eight(57 percent) of the 14 placebo-control dogs fought off theirseptic shock.
Right after Thanksgiving in 1991, Hoffman and his colleaguesalerted the FDA to the antibody's apparent lethality and failureto protect against sepsis. In September 1991, an FDA advisorycommittee had recommended approval of HA-1A, which theagency was on the verge of licensing.
"It was the worst possible moment for Centocor," reflectedHoffman, "but the best possible moment if you happened to bea sepsis patient."
As BioWorld Financial Watch reported Jan. 25, "The dog datawere discussed at a meeting between the NIH, FDA andCentocor in mid-December 1991. Sources described the meetingas "tempestuous" and "stormy." Company officials subsequentlywrote several strongly worded letters to National Institutes ofHealth, pressing the agency to suppress the study, as resultscould be used unfairly by Pfizer Inc., the marketing partner ofXoma Corp., which has a rival sepsis antibody, E5, targeting thesame endotoxin.
An NIH source subsequently explained to BioWorld that thecanine trial had been conducted with taxpayers' money, andthat federal employees had no right not to publish it. Instead,data from the dog trial were first released as abstracts in theApril 1992 issue of Clinical Research. More findings werepresented at the 32nd Interscience Conference on Anti-microbial Agents and Chemotherapy (ICAAC) last October inAnaheim, Calif.
Meanwhile, Hoffman and his co-authors were shopping the fullreport to one scientific and clinical journal after another. JAMAeventually accepted it for today's issue.
In mid-January, Centocor (NASDAQ:CNTO) of Malvern, Pa.,suspended HA-1A trials for septic shock, owing to "excessivedeaths," but is continuing a smaller study of fulminantmeningococcemia in children.
Xoma Corp. is about to launch a redesigned Phase III clinicaltrial of its E5 antibody. "A lot of people feel sepsis products areout of vogue," said Patrick Scannon, the Berkeley, Calif.,company's chief scientific and medical officer. "Unfortunately,they think so at a time when we've learned more about sepsisthan ever in medical history."
Scannon plans to put this learning curve at the service of thenew study, to begin in "the next few weeks," he told BioWorld.
"What Centocor and Synergen showed is that these productsappear to work in patients with severe sepsis," Scannon said."Our new trial design takes this into account by identifying justthose patient groups most likely to benefit from E5."
At one extreme, he explained, are the mildly infected, who willrecover on their own, but who if treated can't be told apartfrom placebo. At the other end of the spectrum are patientswho are virtually dead, with irreversible disease, hence notcandidates for E5 treatment.
In between are the seriously ill, with developing complications,such as renal failure and other organ damage. Focusing onthese, Scannon explained, not only "maximizes our chances ofsuccess" but introduces an essential cost-effectiveness factor.
Xoma already has two E5 Phase III trials under its belt, and itsmarketing partner, Pfizer "has given the drug to more than 500compassionate-IND patients, with continued enthusiasm."
Financial analyst Stuart Weisbrod's enthusiasm is somewhattempered. He follows biotechnology for the equity researchgroup of Merrill Lynch & Co. Anent today's report of the caninefiasco with HA-1A in JAMA he told BioWorld, "I don't think anymore bad news about the Centocor anti-sepsis drug is going tohurt the biotechnology stock market."050593HA-1A
-- David N. Leff Science Editor
(c) 1997 American Health Consultants. All rights reserved.