Key points are not available for this paper at this time.
The efficacy of acetylcholinesterase inhibitors (AChEI) and memantine (ME) in severe dementia remains uncertain, and real world data are limited. We evaluated the progression of cognitive decline and mortality rates in a cohort of outpatients with severe dementia treated with AChEI and/or ME or not treated, from the National Alzheimer's Coordinating Center Uniform Data Set (NACC-UDS). Overall, 703 patients (mean age:78.3 ± 7.2 years; females:56.1%) with severe dementia (Mini-Mental State Examination-MMSE ≤ 10/30) were included 407 late-onset Alzheimer's disease (LOAD), 126 Lewy body dementia (LBD), 170 vascular dementia (VD). Patients were grouped by treatment status: untreated (AChEI - /ME - ; n.70), AChEI only (AChEI + ; n.294), ME only (ME + ; n.215), combined therapy (AChEI + /ME + ; n.124). During follow-up (mean:1.7 ± 0.3 years; range:0.2-2.9 years) the mean MMSE scores declined by - 1.7 points (Standard Deviation: -1.55 to -1.9) in AChEI - /ME - (p for trend:0.001), - 1.5 points (-1.55 to -1.9) in AChEI + (p:0.001), - 0.7 points (-0.5 to -0.9) in ME + (p:0.09), while it increased by + 0.7 points (+ 0.4 to + 1.0) in AChEI + /ME + (p:0.001). Survival rates were 8% in AChEI - /ME - , 10% in AChEI + , 28% in ME + , and 49% in AChEI + /ME + . In Cox proportional hazards analysis (age and sex adjusted), AChEI + /ME + exhibited a 39% reduction in total mortality compared with no treatment (HR:0.61; 95%CI:0.49-0.77; p < 0.001), ME + displayed a 22% reduction (HR:0.78; 95%CI:0.65-0.94; p:0.007), while AChEI + showed no association (HR:0.97; 95%CI:0.80-1.18; p:0.75). When the Cox proportional hazards models were stratified by dementia subtype, consistent significant patterns were observed in LOAD and LBD. In this real-world cohort of older adults with severe dementia, the combination therapy (ME + AchEI) seems to be associated with the best outcomes, including a stabilization of cognitive function and reduced total mortality.
Zuliani et al. (Mon,) studied this question.