Donepezil

Limits on what this shows

• Key limitation: The human study is observational, not randomized; treatment assignment was not controlled, so confounding by indication (healthier or more treatment-adherent patients being the ones on donepezil) cannot be excluded per the standard applied throughout this series.

Mechanism

Named, drug-specific, Tier B: in transgenic amyloid-precursor-protein mice and in vitro systems, donepezil reduced mitochondrial accumulation of amyloid-beta, modulated the mitochondrial permeability transition pore, preserved glutathione, and restored mitochondrial energy production capacity, independent of acetylcholinesterase inhibition.

Population evidence

A human study comparing 24 cognitively healthy older adults and 26 adults with mild cognitive impairment (some donepezil-treated, some not) found untreated MCI participants had reduced skeletal-muscle mitochondrial respiratory capacity relative to both healthy controls and donepezil-treated MCI participants; donepezil treatment was also associated with altered expression of over 600 skeletal-muscle genes.

Disease-vs-drug note

Mitochondrial dysfunction is a well-established, disease-intrinsic feature of Alzheimer's disease and MCI. The finding here is specifically that donepezil-treated patients look less impaired than untreated patients with the same disease, which supports a drug effect but does not establish it with the certainty a randomized trial would.

Limits on what this shows

• Key limitation: Single-site, comparative (not randomized or placebo-controlled), moderate sample size (n=52 patients). Confounding by indication cannot be excluded: patients selected for or tolerating donepezil treatment may differ systematically from untreated patients in disease severity, comorbidity, or frailty.

Population evidence

A comparative human study of 52 Alzheimer's patients (23 on donepezil, remainder untreated) and 32 healthy controls measured protein carbonyl (PCO), advanced protein oxidation products, ischemia-modified albumin, ferric reducing antioxidant power (FRAP), total and protein thiol, and kynurenine metabolites. Untreated Alzheimer's patients had significantly higher PCO than both the donepezil-treated group and healthy controls; the treated group showed higher FRAP, total thiol, and protein thiol, and lower kynurenine and N-formyl kynurenine, than the untreated group.

Mechanism

Not a direct antioxidant chemistry of donepezil itself; the most plausible pathway, supported by the mitochondrial entries below, is that reducing amyloid-beta-driven mitochondrial and oxidative injury secondarily improves systemic redox markers.

Disease-vs-drug note

Alzheimer's disease itself is independently associated with substantial oxidative stress in the broader literature; this entry reflects a comparison between treated and untreated disease, not a comparison to nutrient-replete healthy status, and is graded as a bounded, single-study finding for that reason.

Counter-evidence held on record

This is not evidence that donepezil requires antioxidant supplementation, nor that withholding donepezil protects redox status. Direction runs opposite to a depletion story: less treatment is associated with worse markers in this dataset, not better.

Population evidence

A meta-analysis of 44 case-control studies found serum copper significantly higher in Alzheimer's disease patients than controls (mean difference 9.27, 95% CI 5.02-13.52, p<0.0001). This meta-analysis did not evaluate treatment status or any specific drug.

Drug-specific evidence

None located. A rat study using copper sulfate to induce Alzheimer's-like neurodegeneration tested donepezil as a treatment for the resulting cognitive and inflammatory endpoints (acetylcholine, BACE1, phosphorylated tau, apoptotic markers), but did not report donepezil's effect on copper levels themselves, so it does not support a copper-specific claim for this drug.

Mechanism

None identified for donepezil directly. No chelation, transporter, or renal-handling mechanism connects donepezil to copper in the pharmacokinetic or mechanism screen.

Disease-vs-drug note

Per the confounding-by-indication standard applied throughout this series, an elevated copper finding present in Alzheimer's disease broadly cannot be attributed to donepezil without a study comparing treated to untreated or to pre-treatment baseline, which does not exist in the literature reviewed here.

Population evidence

The same 44-study case-control meta-analysis found serum iron lower in Alzheimer's disease patients than controls after sensitivity exclusion of two outlying studies (mean difference -13.01, 95% CI -20.75 to -5.27, p=0.001). Serum zinc was also lower (mean difference -6.12, 95% CI -9.55 to -2.69, p=0.0005). Neither analysis evaluated treatment status.

Drug-specific evidence

A rat study of nanocurcumin and donepezil in an aluminum chloride/D-galactose Alzheimer's model tested seven groups, including a donepezil-alone arm, and reported that nanocurcumin alone or combined with donepezil improved iron profile and ferroptosis markers; the abstract and available data do not isolate an iron-specific effect for the donepezil-alone arm apart from the nanocurcumin-containing arms, so this cannot support a donepezil-specific iron claim.

Mechanism

None identified for donepezil directly in the pharmacokinetic or mechanism screen.

Disease-vs-drug note

As with copper, the lower-iron finding is a disease association, not a treatment effect, per the same confounding-by-indication standard.

Limits on what this shows

• Key limitation: All three sources are rodent studies; no human gut-microbiome data for donepezil were located in this review.

Mechanism

Not fully characterized; proposed pathway is the cholinergic anti-inflammatory reflex (vagal cholinergic signaling modulating gut inflammation and barrier function) rather than direct antimicrobial action.

Population evidence

In female APP/PS1 transgenic mice given donepezil 3 mg/kg/day orally for 24 weeks, treatment increased Akkermansia (about 54% higher), Clostridium, Odoribacter, and Ruminiclostridium, decreased Bacteroides, Helicobacter, Mucispirillum, Lactobacillus, and Prevotellaceae, restored intestinal barrier integrity, and reduced colonic inflammation relative to untreated APP/PS1 mice. A second, independent amyloid-beta-induced mouse pilot model also reported donepezil-associated shifts in gut microbiota and gut metabolites alongside brain metabolite changes. A third study found donepezil ameliorated gut barrier disruption in a doxorubicin-treated rat model unrelated to Alzheimer's disease, supporting a general barrier-protective mechanism rather than one specific to amyloid pathology.

Counter-evidence held on record

The available animal direction is toward barrier restoration and reduced inflammation, not toward a harmful microbiome disruption. This entry should not be read as a caution against donepezil on gut grounds; it is graded yellow because the evidence tier (animal-only) caps it there regardless of direction.

Limits on what this shows

• Key null finding: A donepezil-specific systematic review and meta-analysis of 60 randomized trials (over 12,000 participants) found no association between donepezil and fatal, QT-related arrhythmia or associated death. A separate pooled meta-analysis across donepezil, rivastigmine, and galantamine (25 studies, n=258,388) found elevated relative risk for both bradycardia (RR 1.23, 95% CI 1.14-1.33) and syncope (RR 1.40, 95% CI 1.20-1.64), but did not stratify results by individual drug.
• At-risk subgroups: Patients with pre-existing sinus node disease, AV conduction disease, or on other rate-slowing agents; risk is concentrated at initiation and dose escalation (5 mg to 10 mg to 23 mg).

Regulatory status

The FDA label states directly that cholinesterase inhibitors "may have vagotonic effects on the sinoatrial and atrioventricular nodes," manifesting as bradycardia or heart block, with or without pre-existing conduction abnormality.

Population evidence

A case-time-control study of Ontario residents aged 67 and older (161 cases hospitalized for bradycardia, 466 matched controls; donepezil was the dominant drug, 73% of cases and 63% of controls) found recent initiation of a cholinesterase inhibitor associated with more than double the risk of hospitalization for bradycardia (adjusted OR 2.13, 95% CI 1.29-3.51). Pacemaker insertion occurred in 11% of hospitalized cases.

Mechanism

Named: peripheral acetylcholinesterase inhibition increases vagal tone at the sinoatrial and atrioventricular nodes, slowing conduction. This is an off-target systemic extension of the drug's cholinergic action, not its intended CNS therapeutic mechanism.

Counter-evidence held on record

This is a cardiac conduction adverse effect, not a nutrient or mineral deficiency, and nothing about it is corrected by any supplement. There is no FDA label language calling for correcting potassium or magnesium before use, and none should be inferred. The action is clinical monitoring of heart rate and syncope symptoms, particularly at dose changes, not a nutritional intervention. Do not read this entry as evidence that donepezil depletes a mineral involved in cardiac conduction. No such mechanism is documented.

Limits on what this shows

• Key limitation: All supporting evidence is cell-based, ex vivo, or animal; no human study in this review measured ATP production or mitochondrial biogenesis markers directly in donepezil-treated patients.

Mechanism

Named, drug-specific, Tier B: in neuroblastoma cells, primary hippocampal neurons, and organotypic hippocampal tissue from adult mice, donepezil dose-dependently increased mitochondrial biogenesis and ATP levels along with PGC-1α and NRF-1 expression; blocking AMPK eliminated the effect, identifying AMPK as the mediator. A separate transgenic-mouse and in vitro study reported donepezil restored mitochondrial energy production capacity reduced by amyloid-beta toxicity.