Skip to main content

Long-Term Effects of Cholinesterase Inhibitors on Cognitive Decline and Mortality

By Michael T. Lin, MD

Associate Professor of Neurology and Neuroscience, Weill Cornell Medical College

SYNOPSIS: Cholinesterase inhibitors are one of the few drug classes approved by the Food and Drug Administration for the treatment of patients with Alzheimer’s disease. This study shows a long-term benefit in slowing the decline of cognition, as measured by the Mini-Mental State Exam, but it is unclear if there is any benefit in quality of life.

SOURCE: Xu H, Garcia-Ptacek S, Jonsson L, et al. Long-term effects of cholinesterase inhibitors on cognitive decline and mortality. Neurology 2021;96:e2220-e2230.

Cholinesterase inhibitors (CHEIs) are the main Alzheimer’s disease (AD) drug therapies. However, the follow-up in most randomized clinical trials of CHEIs is < 1 year, and there are few studies of their long-term effects. In their recent article, Xu and colleagues found that CHEIs were associated with sustained benefit in cognition and decreased mortality over five years. One specific CHEI, galantamine, also was associated with a lower risk of severe dementia. The authors used the Swedish Dementia Registry, established in 2007 to register all patients with incident dementia in Sweden and follow them annually. From 2007 to 2017, 39,196 patients with AD or mixed AD dementia were registered. Subjects were considiered CHEI nonusers if they were never given a CHEI and CHEI users if they were given a CHEI within three months of baseline Mini-Mental State Exam (MMSE). Subjects given their first CHEI more than three months after baseline were excluded because of faster decline. Without adjustment, CHEI nonusers were older, had lower baseline MMSE scores, had more comorbidities, and took more medications. To balance these potential confounders, a propensity score matching scheme was used. The final cohort consisted of 11,652 CHEI users and 5,826 nonusers, well-matched for demographics, comorbidities, and baseline characteristics. The investigators compared the groups with respect to cognitive trajectory, incidence of severe dementia, and mortality.

Over an average of five years of follow-up, MMSE scores in CHEI users were slightly higher than in CHEI nonusers at every time point, and the benefit increased slightly over time (0.13 points/year). The benefit increased with dose up to a certain point (donepezil 7.5 mg, rivastigmine 9 mg, galantamine 16 mg) and then plateaued. The different CHEIs produced similar degrees of improvement. For all CHEI users as a group, the risk of developing severe dementia (MMSE < 10) was slightly decreased compared to nonusers, but not statistically significant (hazard ratio [HR], 0.84; 95% confidence interval [CI], 0.63-1.13). However, for galantamine users, the decrease in risk of severe dementia reached the edge of statistical significance (HR, 0.69; 95% CI, 0.47-1.00).

The mortality rate was decreased for CHEI users as a group compared to nonusers (HR, 0.73; 95% CI, 0.69-0.77), as well as for each CHEI individually.


The strengths of this study are a large sample size and long follow-up period. However, it is observational and can provide only class III evidence. Also, the main criticism of CHEIs is their very modest benefit, and that remains the case here. Although the cognitive benefit was sustained, and even increased slightly over time, after five years, the benefit of CHEIs was less than 0.5 MMSE points. This difference typically is not noticeable, and a graph of the MMSE scores, as shown in Figure 1, illustrates that the CHEI user and nonuser trajectories are virtually superimposed. A 27% decreased risk of mortality over the follow-up period is of interest, but a more meaningful figure would be the duration of increased survival, presented as a Kaplan-Meier plot. It also is reasonable to ask whether prolongation of severe dementia even is desirable. Perhaps the most important question, still unanswered, is whether CHEIs produce long-term improvement in quality of life. It is important to make a distinction between a statistically significant improvement in a test and a quality-of-life improvement.