The phase 3 study
The approval of GV-971 was based on the subsequent 36-week phase 3, placebo-controlled study of 818 patients with mild to moderate AD, which was reported last year. Patients were randomized to placebo or to GV-971 450 mg twice daily. Amyloid PET imaging was not required at entrance to the study, so there was no measure of baseline amyloid load. However, all patients showed MRI evidence of cortical atrophy. The Mini Mental State Exam (MMSE) ranged from 11 to 26, indicating mild to moderate AD.
Secondary endpoints included change on the Alzheimer’s Disease Cooperative Study Activities of Daily Living, the Clinician’s Interview-Based Impression of Change–Plus, and the Neuropsychiatric Index.
Patients taking GV-971 experienced a statistically significant 2.54-point difference on the ADAS-cog12, compared with placebo. The difference was apparent by week 4 and was seen at every clinical visit. When patients were grouped according to baseline MMSE (11-14, 15-19, and 20-26), the drug performed similarly.
Michele G Sullivan/MDedge News
Dr. David Knopman
But none of the secondary endpoints significantly favored of GV-971. And the placebo group behaved in an unexpected way, which could throw the data interpretation off-kilter somewhat, David Knopman, MD, said in an interview.
“It was a very weird and unusual-looking trajectory of the placebo and treated groups,” said Dr. Knopman of the Mayo Clinic, Rochester, Minn., with those taking placebo staying relatively stable for some time before the cognitive scores dropped precipitously. “The data were unconvincing to me.”
He also pointed out that although the extensive preclinical data appeared in the recent peer-reviewed Cell Research paper, the phase 3 data has not appeared in any peer-reviewed forum, “even though the trial has been completed for some time now. Furthermore, the duration of the study was inadequate.”
Finally, none of the study subjects were taking the standard-of-care cholinesterase inhibitors, which virtually every AD patient in the United States does take.
“This makes it almost completely inapplicable to the U.S.,” he said. “It’s not bad news. I’m just not convinced.”
Dr. Holtzman is a cofounder of C2N Diagnostics. He is on the scientific advisory board of Genentech, Denali, and C2N Diagnostics. He consults for Idorsia. Dr. Knopman is a consultant for the Bluefield Project to Cure Frontotemporal Dementia and for Lundbeck.