Dementia is a debilitating condition characterized by impairments in memory, behavior, and emotional regulation. Globally, more than 50 million people suffer from dementia, with projections indicating a continual rise in prevalence [1]. Alzheimer's disease (AD), comprising 60%–80% of all dementia cases [2], is attributed to mechanisms such as the formation of extracellular amyloid-beta plaques and neurofibrillary tangles, leading to neuronal dysfunction and synaptic loss [3].

Acetylcholinesterase inhibitors (AChEIs) have traditionally been the primary pharmacological approach for AD, primarily aimed at alleviating symptoms. They enhance acetylcholine levels in the brain by inhibiting its degradation [4]. Recently, lecanemab, an antibody targeting amyloid plaques in the brain, was introduced to the market, offering a potential disease-modifying therapy for AD. Furthermore, on July 2, 2024, the U.S. Food and Drug Administration approved another antibody therapy, donanemab. However, these innovative therapies are currently indicated only for patients in the early stages of AD, limiting their broader application. Moreover, the high cost of these drugs presents a significant barrier to their widespread use. Consequently, AChEIs remain the predominant treatment modality at this time.

Despite being the cornerstone of treatment, AChEIs demonstrate significant variability in clinical response among patients [5], highlighting the need to elucidate contributing factors. One such factor of interest is genetic variability, particularly in the cytochrome P450 2D6 (CYP2D6) gene, known to influence the metabolism of donepezil, the most widely prescribed AChEI. CYP2D6 polymorphisms have been linked to variations in donepezil plasma concentrations [6-9], which may in turn affect therapeutic efficacy. However, conflicting findings exist regarding whether these concentration differences translate into treatment response.

Furthermore, understanding the relationship between donepezil concentration and its side effects is crucial for optimizing treatment strategies. Donepezil commonly induces gastrointestinal disturbances and other adverse events, potentially impacting treatment adherence and patient outcomes [10]. Yet, there remains a dearth of comprehensive studies exploring the correlation between donepezil plasma levels and adverse effects.

Therefore, this study aims to investigate the relationship between CYP2D6 genotype and donepezil plasma concentration in Korean AD patients. Furthermore, it seeks to explore how donepezil concentration and CYP2D6 genotype influence treatment effect and adverse events, thereby contributing to personalized treatment approaches in AD management.

Participants & study design

This study was conducted at Konkuk Medical Center, Seoul, Korea, in an outpatient setting. Patients diagnosed with probable AD dementia were included in this study. Dementia, and probable AD were diagnosed based on the Diagnostic and Statistical Manual of Mental Disorders (4th edition), the criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association, respectively. Participants who had been receiving either 5 mg or 10 mg of donepezil continuously were included.

Based on safety considerations outlined in clinical guidelines and literature concerning the use of donepezil, patients with known hypersensitivity to donepezil or its components were excluded from the study. Additionally, individuals with moderate to severe hepatic impairment were excluded due to the potential risk of increased drug accumulation. Furthermore, patients with a resting heart rate below 50 beats per minute or a QT interval prolongation exceeding 500 milliseconds were also excluded. We assessed all available data, such as age, sex, and vascular factors (i.e., history of hypertension, diabetes mellitus, and dyslipidemia) at the initial visit. Body mass index (BMI) is calculated as the body weight divided by the square of the body height. All patients initially commenced treatment with 5 mg of donepezil and underwent a 30-day surveillance period to monitor the onset of side effects. In cases where side effects manifested during this period, administration of the medication was ceased, and the nature of the side effects was documented. Upon subsequent outpatient clinic visits following the 30-day monitoring period, patients who did not experience side effects had their dosage escalated to 10 mg of donepezil. For patients who exhibited side effects, the decision to resume or discontinue treatment was made based on the patient's and guardian's preferences and the patient's clinical condition. All patients underwent a re-administration attempt, with successful maintenance of the 5 mg dosage achieved upon re-initiation.

The protocol was approved by the Institutional Review Board of Konkuk Medical Center, Seoul, Korea (IRB No. 2019-03-030), and conducted in accordance with the principles of the Declaration of Helsinki and applicable regulatory guidelines. Written informed consent was obtained from all participants prior to their enrollment in the study.

CYP2D6 genotyping

Genomic DNA was extracted from blood using the MagNA Pure automated extraction system (MagNA Pure 24 Instrument; Roche Diagnostics Limited) following the manufacturer's instructions. The genotyping assay targeted 9 variants (CYP2D6*2, *5, *10, *14, *41, *65, *69, *114, and XN), including four CYP2D6 single nucleotide polymorphisms (SNPs): 100C>T (rs1065852, g.5119C>T, c.100C>T), 1759G>A (rs5030865, g.6778G>A, c.505G>A), 2851C>T (rs16947, g.7870C>T, c.886C>T), and 2989G>A (rs28371725, g.8008G>A, c.985+39G>A), as well as gene deletion and duplication. The genotyping of the four CYP2D6 SNPs was performed using pyrosequencing on the Pyromark Q96 ID instrument with PyroMark Gold Q96 reagents, following Qiagen's protocols. PCR and pyrosequencing primers were designed using PyroMark Assay Design software 2.0 (Qiagen). For genotyping CYP2D65 and *XN, the long-PCR method was employed with minor modifications as previously described [11,12].

Allelic variants identified included CYP2D6*1 (assigned in the absence of variants; default assignment), *2 (normal function), *10, *14, and *41 (decreased function), and 5 (no function), as well as the presence of duplications. CYP2D6 activity scores (AS) and metabolizer statuses were assigned according to guidelines from the Clinical Pharmacogenetics Implementation Consortium (CPIC) [13,14].

To calculate the AS, values were assigned based on the alleles identified in the study cohort: no function allele (*5) = 0; decreased function allele *10 = 0.25; other decreased function alleles (*14 and *41) = 0.5; and normal function alleles (*1 and *2) = 1. The AS for each diplotype was determined by summing the assigned values for each allele. Subjects were categorized as poor metabolizers (PMs) with an AS of 0, intermediate metabolizers (IMs) with an AS between 0 and 1.25, normal metabolizers (NMs) with an AS between 1.25 and 2.25, and ultrarapid metabolizers (UMs) with an AS greater than 2.25, in accordance with the CPIC guidelines [13,14].

Bioanalysis of plasma donepezil concentration

Steady-state plasma donepezil concentrations were measured using a validated high-performance liquid chromatography (HPLC) technique. Upon collection, blood samples were immediately centrifuged, and plasma was stored at −20°C until analysis.

For sample preparation, 1 ml of plasma was mixed with 10 μl of internal standard solution (200 ng/ml rosiglitazone in acetonitrile) for 1 min. Subsequently, 5 ml of methyl tert-butyl ether was added and mixed for another minute, followed by centrifugation at 4,500 rpm and 25°C for 10 min. The organic layer was transferred to a 15 ml conical glass tube, and the supernatant was evaporated to dryness. The residue was reconstituted with 150 μl of 50% methanol in deionized water, followed by another centrifugation at 4,500 rpm and 25°C for 15 min. Finally, 90 μl of the reconstituted solution was injected into the HPLC system (Shimadzu Corporation).

A reversed-phase chromatographic separation was performed using a Luna C18 column (250 mm × 4.6 mm, 5 μm, Phenomenex) under gradient elution conditions with a mobile phase composed of acetonitrile and 20 mM potassium dihydrogen phosphate (pH 5.85). Detection was achieved using a fluorescence detector set at an excitation wavelength of 330 nm and an emission wavelength of 380 nm. Calibration curves were linear over a range of 2.5–112 ng/ml (r^2 > 0.999). Analytical results were considered valid when more than 4 out of 6 quality control sample concentrations fell within 15% of the known values. Intra- and inter-day accuracies ranged from 96.3% to 104.3%.

Treatment response & adverse events

Cognitive function assessments using the Mini-Mental State Examination (MMSE) were conducted annually for up to three years following the initiation of treatment.

Donepezil can lead to several adverse effects which may include gastrointestinal disturbances such as nausea, vomiting, and diarrhea, as well as central nervous system effects like insomnia, dizziness, and headache. Additionally, cholinergic side effects such as bradycardia, urinary incontinence, and muscle cramps have been reported. Throughout the study period, adverse events were collected, with particular attention given during dosage escalation phases of donepezil.

Statistical analysis

Descriptive statistics were used to summarize the clinical characteristics of the study subjects, presented as mean (standard deviation, SD) or frequency. Mann–Whitney U-test was used to compare plasma donepezil concentration between genotypes. Linear mixed-effects models were used to test for associations between CYP 2D6 genotype and the MMSE score over time adjusting for age, sex, BMI, education period, APOE genotype and time examined. Spearman's rank correlation coefficient was used to examine the relationship between plasma donepezil concentration and MMSE changes from baseline. The chi-squared test was utilized to compare associations between CYP2D6 genotype and the occurrence of adverse events. Logistic regression was conducted to evaluate the impact of donepezil concentration on adverse events. Statistical analyses were performed using R (version 4.3.3), and significance was defined as p < 0.05 for a two-tailed distribution.

Demographic and clinical characteristics

In this study, a total of 76 patients with AD participated. Their MMSE scores prior to initiation of treatment ranged from 9 to 27 points, indicating generally mild to moderate disease severity. The mean age of the patients was 78.3 ± 6.7 years, with 51 females and 25 males among the participants. At the time of blood sampling, 71 patients were receiving a dosage of 10 mg, while 5 patients were receiving 5 mg of donepezil. No patients received AChEIs other than donepezil (such as galantamine or rivastigmine) or a NMDA receptor antagonist (memantine) which can affect cognitive function assessment during study period. Detailed demographic characteristics of the patients are presented in Table 1.

Table 1 . Baseline characteristics.

Clinical information	Value
Age (y)	78.3 ± 6.7
Sex (M:F)	25:51
BMI (kg/m2)	24.4 ± 3.3
AST (IU/L)	28.5 ± 10.8
ALT (IU/L)	21.8 ± 12.8
Creatinine (mg/dl)	0.9 ± 0.3
MMSE	20.6 ± 4.3
Dose (10 mg:5 mg)	71:5
Comorbidity
Hypertension	46 (60.5)
Diabetes mellitus	24 (31.6)
Hypercholesterolemia	19 (25.0)

Values are presented as mean ± standard deviation or number (%). BMI, body mass index; AST, aspartate aminotransferase; ALT, alanine aminotransaminase; MMSE, Mini-Mental State Examination.

CYP2D6 genotypes

The CYP2D6 *1/*1 genotype was observed in 14 out of 76 patients (18.4%). The most frequent genotype was CYP2D6 *1/*10, found in 28 patients (36.8%), followed by *10/*10 observed in 20 patients (26.3%). The CYP2D6 *1/*5, *2/*10, and *5/*10 genotypes were present in 4, 6, and 2 patients, respectively, while *2/*41 and *5/14 genotypes were each found in 1 patient. None of the subjects carried the CYP2D6 *65, *69, and *114 alleles in this cohort. According to CPIC guidelines, genotypes CYP2D6 *1/*1, *1/*10, *2/*10, and *2/*41 are classified as NMs, while *1/*5, *5/*10, *5/*14, and *10/*10 are classified as IMs in this cohort. There were no UMs or PMs identified. The frequencies per CYP2D6 genotypes were consistent with the previous study in Korean population [15]. The frequencies and AS for each genotype are presented in Table 2.

Table 2 . Frequencies and activity scores per CYP2D6 genotype.

CYP2D6 genotype	Activity score	Frequency (%)
*1/*1	2	14 (18.4)
*1/*5	1	4 (5.3)
*1/*10	1.25	28 (36.8)
*2/*10	1.25	6 (7.9)
*2/*41	1.5	1 (1.3)
*5/*10	0.25	2 (2.6)
*5/*14	0.5	1 (1.3)
*10/*10	0.5	20 (26.3)

Plasma donepezil concentration by CYP2D6 genotype

The mean (SD) and median (min-max) plasma donepezil concentrations among all 76 participants in this study were 59.7 ± 29.9 ng/ml and 56.3 (0.36–119.51) ng/ml, respectively.

Since there were no UMs or PMs, the donepezil plasma concentrations between NMs and IMs were compared, showing significant differences in donepezil plasma concentrations. The mean plasma concentration in NMs was 56.8 ± 27.1 ng/ml, whereas in IMs it was significantly higher at 69.6 ± 30.1 ng/ml (p = 0.042). After adjusted for administered dosage, NMs had a mean concentration of 5.84 ± 2.75 ng/ml, compared to 7.91 ± 4.49 ng/ml in IMs, still showing a significant difference (p = 0.032). Individual genotypes generally exhibited concentrations aligning with their respective AS (Table 3, Fig. 1).

Table 3 . Donepezil concentration per CYP2D6 genotype.

CYP2D6 genotype	N	Plasma donepezil concentration (ng/ml)	Plasma donepezil concentration/dose (ng/ml/mg)
NM	49	56.8 ± 27.1	5.84 ± 2.75
*1/*1	14	45.7 ± 27.1	4.76 ± 2.65
*1/*10	28	63.7 ± 26.0	6.55 ± 2.69
*2/*10	6	50.5 ± 26.5	5.05 ± 2.65
*2/*41	1	0.36	0.036
IM	27	69.6 ± 30.1	7.91 ± 4.49
*1/*5	4	48.9 ± 19.7	6.55 ± 4.74
*5/*10	2	53.2 ± 74.7	5.3 ± 7.47
*5/*14	1	67.6	6.76
*10/*10	20	72.0 ± 31.1	8.11 ± 4.70
p-value (NM vs. IM)		0.042	0.032

Values are presented as mean ± standard deviation or number only. NM, normal metabolizer; IM, intermediate metabolizer.

Figure 1. Donepezil plasma concentrations according to CYP2D6 metabolizer status. IM, intermediate metabolizer; NM, normal metabolizer.

Changes of cognitive function by CYP2D6 genotypes or donepezil plasma concentration

The MMSE scores before treatment and at 12, 24, and 36 months after treatment are presented in Table 4. There was no association between CYP 2D6 genotype and the MMSE score over time when analyzed using a linear mixed model (p = 0.782). Sex (p = 0.024) and education period (p = 0.003) showed significant associations with MMSE scores, consistent with existing knowledge [16,17].

Table 4 . The changes in MMSE scores after donepezil administration according to CYP2D6 metabolizer status.

CYP2D6 metabolizer status	N	MMSE
		Baseline	12 mo	24 mo	36 mo
NM	49	20.6 ± 4.60	18.7 ± 5.42	19.1 ± 5.21	16.8 ± 5.64
IM	27	20.7 ± 4.19	19.8 ± 4.42	19.2 ± 4.49	18.2 ± 4.11

Values are presented as mean ± standard deviation. NM, normal metabolizer; IM, intermediate metabolizer; MMSE, Mini-Mental State Examination.

There was also no significant correlation between MMSE changes from baseline and donepezil plasma concentration at 12, 24, and 36 months (p = 0.3517, 0.9257 and 0.2125).

Donepezil adverse events by CYP2D6 genotypes or donepezil plasma concentration

Among the total of 76 participants, 17 experienced adverse events, all of which were gastrointestinal-related side effects, specifically nausea and vomiting. Adverse effects including severe bradycardia, urinary incontinence, insomnia, or dizziness were not reported during the study period. Among those who experienced adverse events, 5 were unable to escalate from the starting dose of 5 mg to 10 mg of donepezil due to these effects, while 12 escalated to 10 mg despite experiencing adverse events. The remaining 59 participants started with 5 mg of donepezil without adverse effects and escalated to 10 mg.

Among the 49 participants classified as NMs of CYP2D6, 12 experienced adverse events, while among the 27 IMs, 5 experienced adverse events. Ultimately, there was no significant difference in adverse event occurrence based on CYP2D6 metabolizer status (p = 0.756).

Furthermore, logistic regression analysis, with adverse event occurrence as the dependent variable and donepezil plasma concentration, age, sex, among other variables as independent variables, also did not show a significant association between donepezil plasma concentration and adverse event occurrence (p = 0.829).

In this study, we investigated the impact of CYP2D6 genetic variants on donepezil concentration, treatment effect, and adverse events in Korean patients with AD dementia. Our findings revealed significant variability in donepezil plasma concentrations based on CYP2D6 genotype. However, no differences in treatment effect or adverse events were observed based on CYP2D6 genetic variants and donepezil concentration.

Regarding the impact of CYP2D6 genotype on donepezil concentration, multiple studies demonstrated that the CYP2D6 genotypes with low metabolizing activity exhibit high donepezil concentration.

Yaowaluk et al. [18] who investigated 85 patients with AD dementia observed elevated donepezil levels in homozygous CYP2D6*10/*10 individuals compared to wild-type or heterozygous CYP2D6*10 allele carriers. Chamnanphon et al. [6] reported significant differences in donepezil concentrations between NM and IM groups, though observed no significant differences in dose-normalized donepezil levels. Furthermore, Ortner et al. [19] demonstrated that gene dose of CYP2D6 alleles is significant predictor of donepezil serum concentration along with daily donepezil dose, duration of donepezil treatment and sex. The result of our study that NMs had a lower mean concentration compared to IMs aligns with prior reports. Additionally, in our study, donepezil concentrations generally corresponded well with the AS of the CYP2D6 genotypes. Specifically, patients with the CYP2D6 *1/*1 genotype, which has the highest AS, had the lowest donepezil concentrations. Conversely, patients with the CYP2D6 *10/*10 genotype, which has the second-lowest AS, had the highest donepezil concentrations. There were two patients with the CYP2D6 *5/*10 genotype, which has the lowest AS. Among these two patients, one showed a very high donepezil concentration, while the other had a concentration below 1, resulting in an average concentration that was not higher than that of the CYP2D6 *10/*10 genotype. The patients with concentrations below 1 included one with the CYP2D6 *5/*10 genotype and one with the CYP2D6 *2/*41 genotype. The reason for these abnormally low concentrations is unclear, but it is possible that poor adherence to the medication regimen prior to blood sampling could be a factor.

Our study found no significant differences in MMSE changes over 12, 24, and 36 months between NM and IM groups. The impact of CYP2D6 genotype on treatment response in patients with AD dementia treated with donepezil remains inconclusive based on research findings. Yaowaluk et al. [18] suggested a significant association between possessing the CYP2D6 *10 allele and cognitive function, as assessed by Thai Mental State Examination scores. The study claimed that CYP2D6 *10 transporters may improve cognitive function in AD through increased donepezil and greater inhibition of AChE in the prefrontal cortex. Similarly, Zhong et al. [20] identified a higher prevalence of the CYP2D6 *10 allele in responders compared to non-responders among Chinese AD patients treated with donepezil, suggesting a potential link between the CYP2D6 *10 allele and treatment response. However, a recent meta-analysis of 536 patients reported no difference in responder rates between decreased functional allele and normal allele groups [21], consistent with the findings of our study. These conflicting results may be due to differences in various potential factors that could influence the treatment effect of donepezil. In particular, the concurrent use of drugs that may affect the cholinergic pathway could be a major factor contributing to these differences. Furthermore, other genetic variation related to acetylcholine receptor binding may affect the treatment effect of donepezil. These conflicting results underscore the need for larger study which consider the various confounding factors to further elucidate the impact of CYP2D6 genetic variants on treatment outcomes in AD dementia.

An important aspect of our study was the examination of donepezil-associated adverse events in relation to CYP2D6 genotype and plasma concentration. Contrary to expectations, we found no significant correlation between donepezil concentration and the incidence of adverse effects. This observation suggests that adverse events associated with donepezil may be influenced by factors beyond plasma concentration alone, possibly including individual tolerance to acetylcholine-related effects, although this has not been extensively reported in previous studies. For instance, patient characteristics are pivotal in determining the likelihood of experiencing side effects. We investigated variables such as age, BMI, and underlying comorbidities, while recognizing that factors such as functional ability and frailty could also play a significant role. Furthermore, as with treatment effects, consideration should be given to potential interactions with other medications that affect the cholinergic pathway, as these interactions may alter the profile of side effects.

One limitation of this study is the lack of a thorough analysis of concomitant medications. Although no patients received AChEIs other than donepezil such as galantamine or rivastigmine or a NMDA receptor antagonist (memantine) during the study period, other various concomitant medications were not considered in the analysis. Donepezil is metabolized by CYP2D6 and CYP3A4 [10], and there are many inhibitors and inducers for these enzymes [22]. While the co-administration of CYP2D6 and CYP3A4 inhibitors or inducers is unlikely to be related to donepezil’s therapeutic outcome given that our study found no association between donepezil concentration and changes in MMSE or the occurrence of adverse events, it remains possible that these inhibitors and inducers could influence the donepezil concentration and thereby somewhat affect the assessment of CYP2D6 genotype's impact on donepezil levels.

Our study is significant as it is the first to elucidate the associations between CYP2D6 genotype and the concentration, clinical response or adverse events of donepezil in Korean patients with AD dementia. However, it remains difficult to draw definitive conclusions about the relationship between CYP2D6 and the treatment effect of donepezil. Larger studies are necessary to fully understand this association.

None.

The authors declare no conflicts of interest.

None to declare.