Donepezil (Aricept)
What studies of Donepezil (Aricept) found
At a glance, by condition
Long COVID: D Little or no difference
Based on the strongest evidence, the placebo trial and the randomised trial (very low certainty): 6 findings, all D.
6 outcomes rated · see each below
How we got this
Each condition’s headline comes from its graded findings (one for each outcome and kind of study) rated with the highest certainty: the grade most of them share; a tie goes to the finding with the most people, then the most studies, then the outcome’s name. It sums up the grades below, and conditions are never combined.
What this means: Studies found little or no difference with Donepezil (Aricept) for long COVID. If you’re thinking about it, talk it through with a doctor who knows ME/CFS or long COVID.
Ask about this treatment
The grades, in plain words
Each outcome gets a direction, a certainty and a letter read from the two. Certainty is judged with GRADE, the method Cochrane and WHO use.
- AHelpsImprovement, high certaintyProbably helpsImprovement, moderate certainty
- BMay helpImprovement, low certainty
- CMight help, very uncertainImprovement, very low certaintyMixed resultsStudies disagree
- Donepezil (Aricept) · Long COVID is hereDLittle or no differenceStudies found little or no difference
- FHarmsIt made things worse
By symptom and daily life
Long COVID
Symptoms
Depression1 study · 55 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: high risk of bias (RoB 2) in the results that carry most of the weight; one trial of 110 people found no clear difference; a small benefit can't be ruled out.
The placebo trial120 randomised · 119 started · 110 analysed · 3-week period · 3 mg/d for the first week and then 5 mg/d for 2 weeks
No clear difference in depression.
The evidence is very uncertain about the effect of donepezil (Aricept) on depression in long COVID. Little or no difference · very low certainty D
- High risk of bias (RoB 2) in the results that carry most of the weight
- One trial of 110 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 120 randomised · 119 started · 110 analysed
- Dose
- 3 mg/d for the first week and then 5 mg/d for 2 weeks
- Outcome
- PHQ-9 at 3 wk Patient Health Questionnaire
- Effect
- no clear difference
- Risk of bias
RoB 2: High
- Randomisation process
- Some concerns
- Deviations from intended interventions
- High
- Missing outcome data
- Low
- Measurement of the outcome
- Low
- Selection of the reported result
- Low
Fatigue1 study · 55 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: high risk of bias (RoB 2) in the results that carry most of the weight; one trial of 110 people found no clear difference; a small benefit can't be ruled out.
The placebo trial120 randomised · 119 started · 110 analysed · 3-week period · 3 mg/d for the first week and then 5 mg/d for 2 weeks
No clear difference in fatigue.
The evidence is very uncertain about the effect of donepezil (Aricept) on fatigue in long COVID. Little or no difference · very low certainty D
- High risk of bias (RoB 2) in the results that carry most of the weight
- One trial of 110 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 120 randomised · 119 started · 110 analysed
- Dose
- 3 mg/d for the first week and then 5 mg/d for 2 weeks
- Outcome
- CFS at 3 weeks (absolute value) Chalder Fatigue Scale
- Effect
- no clear difference
- Risk of bias
RoB 2: High
- Randomisation process
- Some concerns
- Deviations from intended interventions
- Low
- Missing outcome data
- Low
- Measurement of the outcome
- Low
- Selection of the reported result
- High
Low mood or anxiety1 study · 55 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: some concerns about bias (RoB 2); one trial of 110 people found no clear difference; a small benefit can't be ruled out.
The placebo trial120 randomised · 119 started · 110 analysed · 3-week period · 3 mg/d for the first week and then 5 mg/d for 2 weeks
No clear difference in low mood or anxiety.
The evidence is very uncertain about the effect of donepezil (Aricept) on low mood or anxiety in long COVID. Little or no difference · very low certainty D
- Some concerns about bias (RoB 2)
- One trial of 110 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 120 randomised · 119 started · 110 analysed
- Dose
- 3 mg/d for the first week and then 5 mg/d for 2 weeks
- Outcome
- HADS at 3 wk Hospital Anxiety and Depression Scale
- Effect
- no clear difference
- Risk of bias
RoB 2: Some concerns
- Randomisation process
- Some concerns
- Deviations from intended interventions
- Low
- Missing outcome data
- Low
- Measurement of the outcome
- Low
- Selection of the reported result
- Low
Cognition (brain fog)1 study · 10 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: high risk of bias (RoB 2) in the results that carry most of the weight; one trial of 25 people found no clear difference; a small benefit can't be ruled out.
The randomised trial25 randomised · at baseline, one month, and three months later · 5 mg
No clear difference in cognition (brain fog).
The evidence is very uncertain about the effect of donepezil (Aricept) on cognition (brain fog) in long COVID. Little or no difference · very low certainty D
- High risk of bias (RoB 2) in the results that carry most of the weight
- One trial of 25 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 25 randomised
- Dose
- 5 mg
- Outcome
- memory scales
- Effect
- no clear difference
- Risk of bias
RoB 2: High
- Randomisation process
- Some concerns
- Deviations from intended interventions
- High
- Missing outcome data
- High
- Measurement of the outcome
- High
- Selection of the reported result
- Some concerns
Daily life and health overall
Overall symptoms1 study · 55 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: high risk of bias (RoB 2) in the results that carry most of the weight; one trial of 110 people found no clear difference; a small benefit can't be ruled out.
The placebo trial120 randomised · 119 started · 110 analysed · 3-week period · 3 mg/d for the first week and then 5 mg/d for 2 weeks
No clear difference in overall symptoms.
The evidence is very uncertain about the effect of donepezil (Aricept) on overall symptoms in long COVID. Little or no difference · very low certainty D
- High risk of bias (RoB 2) in the results that carry most of the weight
- One trial of 110 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 120 randomised · 119 started · 110 analysed
- Dose
- 3 mg/d for the first week and then 5 mg/d for 2 weeks
- Outcome
- Daily Health Status at 3 wk Daily Health Status
- Effect
- no clear difference
- Risk of bias
RoB 2: High
- Randomisation process
- Some concerns
- Deviations from intended interventions
- High
- Missing outcome data
- Low
- Measurement of the outcome
- Low
- Selection of the reported result
- Low
Quality of life1 study · 55 people took Donepezil (Aricept) Little or no difference D
Why this grade: Little or no difference, very low certainty, from randomised trials: high risk of bias (RoB 2) in the results that carry most of the weight; one trial of 110 people found no clear difference; a small benefit can't be ruled out.
The placebo trial120 randomised · 119 started · 110 analysed · 3-week period · 3 mg/d for the first week and then 5 mg/d for 2 weeks
No clear difference in quality of life.
The evidence is very uncertain about the effect of donepezil (Aricept) on quality of life in long COVID. Little or no difference · very low certainty D
- High risk of bias (RoB 2) in the results that carry most of the weight
- One trial of 110 people found no clear difference; a small benefit can't be ruled out
See the 1 study
-
- People in the study
- 120 randomised · 119 started · 110 analysed
- Dose
- 3 mg/d for the first week and then 5 mg/d for 2 weeks
- Outcome
- EQ-5D-5L at 3 wk EuroQol 5-Dimension 5-Level Version
- Effect
- no clear difference
- Risk of bias
RoB 2: High
- Randomisation process
- Some concerns
- Deviations from intended interventions
- High
- Missing outcome data
- Low
- Measurement of the outcome
- Low
- Selection of the reported result
- Low
Side effects
Side effects studies reported6 kinds
| Side effect | Studies that reported it |
|---|---|
| Fatigue | 1 study |
| Headache | 1 study |
| Nausea or stomach upset | 1 study |
| Trouble sleeping | 1 study |
| Vivid dreams or nightmares | 1 study |
| Worse symptoms or a flare | 1 study |
Doses studied
What researchers used, as the papers print it; not a recommendation.
- 1Start
3 to 5 mg at the start
- 2Top dose
Most often 3 mg (1 of 2 studies)
Doses across 2 studies
3 to 5 mg.
Each study’s dose, as printed
- 3 mg/d for the first week and then 5 mg/d for 2 weeks Donepezil for Fatigue and Psychological Symptoms in Post-COVID-19 Condition: A Randomized Clinical Trial. (2025)
- 5 mg The effect of donepezil hydrochloride on post-COVID memory impairment: A randomized controlled trial. (2023)
All papers
- Donepezil ameliorates fatigue and depression in PASC patients with HHV-6B SITH-1-induced acetylcholine deficiency.2026 · placebo trial · 73 analysed
Summary
This retrospective subgroup analysis used stored serum samples from a previously reported multicenter, randomized, placebo-controlled, double-blind trial of donepezil in Japan. The subgroup included 73 of 110 trial participants with post-acute sequelae of SARS-CoV-2 infection (PASC); 38 were analysed in the donepezil group and 35 in the placebo group. Participants received donepezil or placebo for 3 weeks, and assessments were conducted at 3 and 8 weeks after treatment initiation. Outcomes included the Chalder Fatigue Scale (CFS), Hospital Anxiety and Depression Scale (HADS), and Patient Health Questionnaire-9 (PHQ-9). Among the 73 patients, 52 were seropositive for anti-SITH-1 antibodies. In anti-SITH-1 antibody positive PASC patients, donepezil significantly reduced HADS depression scores at week 3 compared with placebo and CFS total scores at week 8; the seropositive comparison included 26 placebo and 26 donepezil patients. No significant improvement was observed in PHQ-9 scores. In seronegative individuals, donepezil tended to worsen symptoms. In the original trial, ten individuals withdrew or were lost to follow-up. The authors state that the subgroup analysis was retrospective, exploratory, and not pre-specified.
- Donepezil for Fatigue and Psychological Symptoms in Post-COVID-19 Condition: A Randomized Clinical Trial.2025 · placebo trial · 120 randomised · 119 started · 110 analysed
Summary
This multicenter, placebo-controlled, double-blind trial in Japan examined donepezil for fatigue and psychological symptoms in post–COVID-19 condition. A total of 120 eligible patients were enrolled; 119 were assigned to the safety analysis after 1 patient withdrew consent, and 110 (55 in each group) were included in the efficacy analysis. Donepezil was given at 3 mg/d for the first week and then 5 mg/d for 2 weeks during a 3-week intervention, compared with lactose placebo. The primary outcome was Chalder Fatigue Scale score at 3 weeks. Mean (SD) CFS at 3 weeks was 14.6 (6.9) with donepezil and 14.1 (6.7) with placebo (P = .31); the baseline-adjusted estimating treatment effect was 0.34 (95% CI, −2.23 to 2.91; P = .79). Secondary outcomes at 3 and 8 weeks, including HADS, IES-R, EQ-5D-5L, PHQ-9, and Daily Health Status, showed no significant differences. Twenty-one patients (36%) in the donepezil group and 25 (42%) in the control group had 1 or more adverse events; 12 (20%) and 11 (18%), respectively, were suspected to be causally related. No serious adverse events occurred, and severe adverse events were 0 in both groups. Ten patients withdrew or were lost to follow-up; 4 in each group dropped out before the 3-week outcome assessment.
- The effect of donepezil hydrochloride on post-COVID memory impairment: A randomized controlled trial.2023 · randomised trial · 25 randomised
Summary
This randomized, controlled trial studied donepezil hydrochloride for memory impairment after COVID-19. The authors state that post-COVID syndrome is associated with complications persisting after recovery, and that cognitive dysfunction including memory impairment has been relatively common, affecting executive function and quality of life, with no approved treatment for this specific complication. The trial enrolled 25 patients with post-COVID memory impairment; participants with a history of hospitalization were randomly assigned to a drug group (n = 10) or a control group (n = 15). The drug group received donepezil 5 mg. Memory indices were assessed at baseline, one month, and three months later using the Wechsler Memory Scale-Revised test. The statistical analysis revealed no significant difference in WMS-R subtest and index scores between the drug and control groups at the 4-week and 12-week follow-up periods. However, within the drug group, there was a notable increase in the visual reproduction I and verbal paired associates II subtests during the specified time intervals. The authors write that donepezil 5 mg did not exhibit a significant overall increase in memory scales compared to the control group over time, while suggesting the medication may exert a positive effect on specific memory subtests. No information on side effects or withdrawals is given in the abstract, and the authors state that further research and exploration are warranted.
- Systematic Review of Randomized Clinical Trials for the Treatment of Long COVID Syndrome.2026 · systematic review
Summary
This systematic review asked whether treatments for long COVID improve outcomes. It searched MEDLINE through PubMed from 2020 to August 2024 and included 43 randomized clinical trials of people with long COVID. The trials compared interventions with control groups or other treatments and reported a primary outcome. Treatments included nutraceuticals, respiratory muscle training, supervised exercise or pacing, brain stimulation, and medications. Outcomes included symptoms, exercise capacity, Vo2 max, fatigue, cognitive functions, inspiratory pressure, pain, and olfactory and gustatory function. The review reports that all studies testing supervised exercises improved the primary outcome except for one; all but one brain stimulation study improved the primary outcome; and only 2 studies using medications improved the primary outcome. Nirmatrelvir/ritonavir did not improve symptoms; donepezil and lithium did not improve their primary outcomes; thermal water inhalation and a metabolic modulator did not improve the primary outcome. Creatine improved fatigue; vitamin C improved exercise capacity; inspiratory muscle training improved symptoms, exercise capacity and inspiratory pressure. Reported harms included 33% with adverse events (dry mouth) with amantadine, 16% nausea with creatine, 17%–37% adverse events with TMS, two to four percent TMS discontinuation, and 8% inspiratory muscle training discontinuation. The authors state the evidence was mostly low certainty.
- Research evidence on the management of the cognitive impairment component of the post-COVID condition: a qualitative systematic review.2024 · systematic review
Summary
This qualitative systematic review asked which treatments have been studied for cognitive impairment after COVID-19. The search covered PubMed, Scopus, and ProQuest up to December 31, 2023, and included 29 studies of post-COVID cognitive impairment; samples ranged from single patients to larger groups. Treatments included cognitive remediation, multidisciplinary rehabilitation, noninvasive brain stimulation, hyperbaric oxygen, pharmacological and nutraceutical agents, neurofeedback, photobiomodulation, and meditation. Comparisons, where reported, included control, sham, placebo, or usual-care groups. Cognitive functioning was assessed with standardized tests, test batteries, or structured interviews. Findings varied. A randomized sham-controlled HBOT trial in 37 subjects reported significant improvements in global cognitive functioning, attention, and executive functions versus control. A famotidine RCT reported significantly improved global cognition in the treatment group (N = 25). A donepezil RCT found no differences between N = 15 and N = 10 in memory after 4 and 12 weeks. A tDCS randomized sham-controlled trial in 23 subjects found no significant effects on executive functioning and processing speed after eight tDCS sessions. A remote multidisciplinary RCT found no significant differences between N = 52 and N = 48. A multidisciplinary study in 80 subjects found no significant improvements in memory and attention at discharge. Other studies reported improvements in smaller or uncontrolled samples. No worsening, stopping, or side effects were reported in the recorded data. The authors state limitations including heterogeneity, lack of long-term follow-up, and under-characterization of samples.