DEMENTIA CLINICAL GLOBAL ASSESSMENT SCALES REQUIRE MORE VALIDATION; CAREGIVER ASSESSMENTS ARE VALID MEASURES, RESEARCHERS TELL ADVISORY CMTE.

Clinical global assessment scales currently used to evaluate the clinical relevance of antidementia drug trial results have not been validated sufficiently to show that they are reliable, audience members from industry and academia told FDA's Peripheral & Central Nervous System Drugs Advisory Committee Feb. 13. The committee, joined by a panel of invited experts, met for two days, on Feb. 13-14, to discuss the agency's proposed guidelines for the clinical evaluation of antidementia drugs. FDA began formulating the guidelines in 1988 in response to the debate over the efficacy of tacrine that followed the controversy over the Summers tacrine and lecithin trial. The draft has been circulating since Nov. 8, 1990 when Division of Neuropharmacological Drug Products Director Paul Leber, MD, sent them to selected companies, dementia researchers in academia, and trade groups, including the Pharmaceutical Manufacturers Association, for comments. FDA said that no timeframe has been set for publishing of the guidelines for formal comment or for finalization. "The issue of clinical meaningfulness can't be solved simply by mandating the use of a clinician's global assessment scale as a primary measure of that meaningfulness," Lon Schneider, MD, University of Southern California, commented in prepared remarks to the committee. Schneider was reacting to a statement in the draft guidelines that "the clinician's global assessment serves as the primary measure of the clinical utility of a product's antidementia effect." Schneider said the guidelines' endorsement of clinical global assessments "is premature if for no other reason than that previous research has demonstrated the very poor reliability of these global assessments." The battery of clinician global assessments involves a standard series of objective questions asked by a physician of a dementia patient to test the patient's cognitive abilities (retention of names, faces, etc. and reasoning) and repeated over time in order to build a database on the decline in cognitive function. Warner-Lambert suggested that FDA allow companies more flexibility in choosing the assessment scales used to measure the drug's clinical relevance to Alzheimer's patients. Warner-Lambert appeared before the committee twice in 1991 to present clinical data on its drug for Alzheimer's disease Cognex (tacrine). The company was able to show a statistically significant improvement in the Alzheimer's Disease Assessment Scale (ADAS-COG) among patients receiving Cognex but not on a second clinical global assessment scale, the Clinical Global Impression of Change, or CGIC ("The Pink Sheet" March 25, p. 4). Cognex, which failed to get an advisory committee recommendation for approval, is currently undergoing further, longer-term clinical studies and is also being made available under a Treatment IND ("The Pink Sheet" Jan. 13, T&G-3). Warner-Lambert Director of Clinical Development for CNS Compounds Stephen Gracon told the committee that the major global assessment scales now used, including CGIC, the Clinical Interview-Based Impression of Change and the Consensus Clinical Assessment, "require further validation in patients with Alzheimer's disease." He added that "given the current level of different opinions, it seems to me that the agency and the committee would want to defer final guidelines or provide for greater flexibility in the guidelines until more data is available on these scales." The Clinical Interview-Based Impression of Change, or CIBIC, came under some strong criticism by a Sandoz clinical researcher. Rene Spiegel, PhD, commented that "the CIBIC has not been shown to be reliable, valid or sensitive to change." Spiegel noted that FDA endorsed the validity of the CIBIC in a Nov. 8, 1991 letter from FDA's Leber to companies involved in antidementia research. The letter states that the CIBIC, for purposes of regulatory assessment of antidementia drugs, "will be assigned the same amount of weight as would have been assigned to the clinical global change measure described in the draft antidementia guidelines," Spiegel said. "In other words," he commented, "the CIBIC will carry the same weight for regulatory purposes as performance-based assessments of cognitive function like the ADAS Cognitive and similar instruments." In the relatively new CIBIC test, clinicians interview dementia patients one-on-one without benefit of other information about the patient. Spiegel described CIBIC clinician raters as operating in "total isolation from any source of information other than their own observations during patient interviews," which he said are "performed in an almost laboratory-like atmosphere." He characterized the interviews as "artificial." Cambridge Neuroscience President and Chief Operating Officer Elkan Gamzu, PhD, agreed with Spiegel, saying: "the CIBIC is not very practical or pragmatic, and I would just like to say there are a number of people in the audience who have also told me that." The guidelines' specific endorsement of global assessments performed by physicians was interpreted by many of the public commentators at the meeting as disparaging the validity of global assessments made by non-clinician caregivers. USC's Schneider told the advisory committee that one "problem with emphasizing the [clinical] global is that, by implication, the emphasis of a functional assessment devalues the fact that favorable changes in behavior and functional activities in daily living are clinically meaningful in and of themselves." Warner-Lambert's Gracon concurred, saying that caregiver-based assessments "should also be regarded as representing clinically important endpoints." The weight of caregiver assessments in measuring the clinical relevance of an antidementia drug was a major subject at the first Cognex advisory committee, since some patients receiving Cognex showed signs of cognitive improvement based on caregiver assessments even though clinician assessments showed no statistically significant change. Cambridge Neuroscience exec Gamzu also suggested that quality of life measures such as the Activity of Daily Living, the Instrumental Activities of Daily Living Assessment, the Progressive Deterioration Scale and the Personal Self-Maintenance Scale, "are actually very valuable outcome measures and should be considered as an alternative way of looking at clinical meaningfulness." Gamzu was the principal data presenter for W-L at the first Cognex advisory committee meeting. The FDA draft guidelines are based largely on the conclusions of a June 15-16, 1989 special meeting of the PCNS advisory committee, which concluded that antidementia claims should be supported by both an improvement in cognition demonstrated by objective tests and an overall clinical response as assessed by a physician ("The Pink Sheet" June 26, 1989, p. 3). The Feb. 13-14 meeting was convened to discuss the comments received as recently as October 1991 and, according to FDA's Leber, to "produce a document that will provide the regulated industry with at least one path to developing antidementia drugs." Several comments expressed concern about the draft guidelines' statement that "a treatment cannot be considered to exert an antidementia action unless it beneficially affects a demented patient's ability to learn new and retrieve old, previously learned information." Leonard Berg, MD, Washington University, who spoke on behalf of the Alzheimer's Association, said: "There is clearly a compelling need for studies that will evaluate the effectiveness of drugs to treat the non-cognitive manifestations of Alzheimer's disease" such as "agitation, screaming, hostile and aggressive actions, delusions, depression and sleep disorders." USC's Schneider said because the guidelines do not address treatments for non-cognitive Alzheimer's symptoms, they "apply only to mildly impaired Alzheimer's patients and...these patients are a minority of the dementia population." FDA's Leber responded that products to treat non-cognitive Alzheimer's symptoms "are surely valuable," but that the proposed guidelines would have to be "scrapped" to include behavioral treatments. He emphasized that the guidelines' focus on cognitive symptoms "in no way precludes sponsors from developing drugs for behavior in dementia" and suggested that "other guidelines could be developed" for the behavioral symptoms of Alzheimer's. Berg also said the Alzheimer's Association supports expanding the population treated in antidementia clinical trials, to comprise "trials that include those with more severe dementia" and trials involving patients "with significant comorbidity and cotherapy if the other conditions are stable and the investigational drug is unlikely to conflict." The committee also discussed whether the guidelines should include only Alzheimer's-type dementia or dementia associated with other medical conditions, such as multi-infarct dementia. The draft states: "This guideline is intended primarily to provide advice about developing treatments for patients who would be deemed to suffer from Alzheimer's Dementia. However, many of the principles enumerated in the guideline apply equally well to other types of chronic dementing illness." Leber observed that, "as a practical consequence" the distinction is not important because the "two major causes of dementing illness" -- Alzheimer's disease and multi-infarct dementia are frequently confused in diagnosis.