Complementary analyses of the AMBAR trial: Individual items of CDR and ADAS‐Cog12 in Alzheimer’s disease patients treated with plasma exchange with albumin replacement

Abstract

AbstractBackgroundAMBAR (Alzheimer’s Management By Albumin Replacement) study results demonstrated that plasma exchange and albumin replacement (PE‐A) with or without intravenous immunoglobulin slows the progression of Alzheimer’s disease (AD) as measured by cognitive (ADAS‐Cog12), functional (ADCS‐ADL), and global assessment (CDR) (Boada et al. Alzheimers Dement 2020;16:1412). In this complementary analysis, we explored the association of ADAS‐Cog12 and CDR subdomains with treatment effects overall as well as stratified by AD severity.MethodAD patients (N=347) were randomized (1:1:1:1) to one of the three different PE‐A treatments or placebo (sham PE‐A). We analyzed the performance of active (all 3 PE‐A‐treated groups) relative to placebo overall as well as stratified by AD severity for each of the 6 domains of CDR scale and 12 items of ADAS‐Cog12 grouped in three categories of “Memory”, “Language” and “Praxis” (Verna et al. Alzheimer Res Ther 2015;7:64). Language was also assessed through Phonetic and Semantic Verbal Fluency (PVF, SVF) tests and Neuropsychological Assessment Battery (NAB) naming test. Changes from baseline were analyzed using a mixed model for repeated measures (MMRM).ResultPE‐A‐treated patients showed improvement by month 14 (p<0.05) in 5 of 6 CDR domains (“Memory”; “Orientation”; “Judgement & Problem solving”; “Home & Hobbies”; “Personal care”), with effect sizes from 71% (less decline) to >100% (improvement over baseline). The “Community affairs” domain had borderline significance (p=0.08). PE‐A‐treated moderate AD patients (MMSE: 18‐21) exhibited improvement over control whereas PE‐A‐treated mild AD patients (MMSE: 22‐26) were not significantly different. Among ADAS‐Cog12 items, PE‐A‐treated patients showed improvement by month 12 or 14 (p<0.05) in at least one of the categories “Language”, “Memory” or “Praxis” with effect sizes from 76 to >100%. PVF and SVF also demonstrated improvements by month 14 (all effect sizes >100%; p‐values: 0.003‐0.04). For the stratified analysis, this improvement, favoring PE‐A‐treated individuals, was only observed in the mild AD group (p= 0.0040 [PVF] and 0.02 [SVF]). Finally, NAB demonstrated improvement by 12 months only in the moderate group (p=0.04).ConclusionPE‐A demonstrated beneficial effects across all CDR domains and language tests, while the ADAS‐Cog12 benefit may be more prominent in specific subdomains of memory, language and praxis.