CRYSTAL CITY, Va. _ The FDA and the scientific communitycontinue to grapple with the use of surrogate markers in clinical trialsof AIDS drugs. The thorny issue was at the top of the agenda here onThursday at the third meeting of the National Task Force on AIDSDrug Development (NTFADD).An FDA official floated two new agency initiatives at the meeting toaddress the complex surrogate marker question. First, the agencyproposed the establishment of a subcommittee of experts devotedsolely to the study of surrogate markers that could advise variousFDA advisory committees and attend product review meetings. Thesubcommittee would provide expertise to advisory committeesreviewing both biologics and drugs.In addition, the FDA wants to establish a working group of topresearchers to launch an "organized, proactive effort" to validatesurrogate markers. Validation would require establishment of a clearlink between changes in a surrogate marker and changes in clinicaloutcome. The working group could establish a data base forretrospective analyses of relevant completed clinical trials, as well astracking the progress of ongoing and planned studies that couldgather relevant information about surrogate markers."The FDA is forced to address these issues, however intractable theymay be, because of products in development," Steve Gitterman fromthe FDA's Center for Drug Evaluation and Research told NTFADDmembers.Surrogate markers are indirect measures of disease, such as tumorshrinkage in cancer patients or critical immune system cell (CD4)counts in AIDS patients, that provide researchers with a substitutefor "hard" clinical endpoints, such as disease progression and death.They are a critical tool for evaluating new therapies in life-threatening diseases where trials based on hard endpoints would takeyears to conduct.To date, the FDA has based its accelerated approval policies forAIDS drugs on the use of surrogate markers, specifically CD4counts. Its approval of the nucleoside analogs Didanosine (ddI) andZalcitabine (ddC) were based in large part on the drugs' effects onCD4 counts, even though those effects were of small magnitude andshort duration. Now, experts are questioning whether changes inCD4 counts translate into clinical benefits for patients.For example, a study of 16 AIDS clinical trials done by University ofWashington statistician Tom Fleming showed that CD4 counts werea weak predictor of key clinical endpoints: time to progression toAIDS and overall survival. Simply put, his analysis showed thatdrugs which exerted a statistically significant effect on CD4 countsdid not always exert a statistically significant effect on clinicalendpoints.In one trial, the drug ddI increased CD4 counts but produced acounter-intuitive negative impact on survival. In the 1,749-patientCONCORDE study of AZT, drug-induced increases in CD4 cellcounts that were maintained for three full years still didn't improvesurvival rates. Fleming's results have cast doubt on the validity ofCD4 counts as a surrogate marker in AIDS.Surrogate Markers Not Right For Pivotal TrialsAt Thursday's NTFADD meeting, Fleming argued that surrogatemarkers might be appropriate as endpoints in early, exploratoryPhase I and II clinical trials, but not in definitive Phase III trials. "Inpivotal trials, it's critical to obtain data on true clinical endpoints,"argued Fleming. "You could use surrogate markers in these trials togather supportive data on a drug's mechanism of action and otherauxiliary information but not as primary endpoints."Fleming also warned that a surrogate marker that works to predictthe effects of an anti-retroviral therapy might not be useful inevaluating an immune-based therapy. That's because the two typesof drugs attack AIDS from different angles.Precipitating the recent crisis over surrogate markers is a new classof drugs called protease inhibitors. The FDA believes that companiesdeveloping protease inhibitors intend to rely heavily on a newsurrogate marker, viral load, in their accelerated approvalapplications. Viral load is a measure of the amount of HIV in apatients' blood.So far, experts say that early, anecdotal reports from proteaseinhibitor studies show the drugs clearly reduce viral burden inpatients while boosting CD4 cell counts. However, resistance to thedrugs does develop over time, according to University of Coloradoresearcher Robert Schooley.Schooley also said that more and larger studies would be needed tovalidate viral load as a surrogate marker in AIDS, even though hiswork has suggested that baseline levels of virus RNA in the bloodwere predictive of disease progression after adjusting for otherfactors. "We still don't know enough about the significance of viralload at baseline or the significance of changes in viral load in termsof clinical outcome," Schooley told NTFADD members.Some experts and AIDS activists believe that relying on unvalidatedsurrogate marker data as a measure of protease inhibitors' potential,as regulators did when they released nucleoside analogs forwidespread use, would be a mistake. They have called for a moreorganized and rigorous approach to evaluating protease inhibitors,potentially raising the hurdle for companies seeking acceleratedapproval by mandating larger trials and validated surrogate markers.Industry representatives have argued against such a move. "We have35 drugs in development in six different disease areas and, for amajority of those drugs, surrogate marker data obtained fromcompleted Phase II trials is the key `go, no-go' decision factor,"explained Stephen Carter, senior vice president of worldwide clinicalresearch and development at Bristol-Myers Squibb and NTFADDmember. He argued that CD4 counts are "one of the most potentprognostic variables for clinical deterioration in AIDS patients.""Surrogate markers are chosen on the basis of biologic plausibilityand practical feasibility," said Carter. "We're getting religious aboutthe analytical evaluation of surrogates versus the pragmatic. Withviral load, the plausibility is there. If we wait for validation fromclinical endpoint trials, we'll have to wait for years."Although experts disagreed on Thursday about the validity ofsurrogate markers, all agreed that an ideal surrogate marker forAIDS is the holy grail of the field at this time. A reliable, accuratesurrogate marker would greatly accelerate AIDS drug development.The NTFADD is attempting to formulate recommendations onsurrogate markers.NTFADD was convened in November 1993 to expedite the searchfor new AIDS therapies by identifying and, when possible, removingobstacles to drug development. To that end, the task force canrecommend actions directly to the Secretary of Health and HumanServices, Donna Shalala, for immediate implementation. n
-- Lisa Piercey Washington Editor
(c) 1997 American Health Consultants. All rights reserved.