Abstract
Background:
In Japan, only oral formulation of donepezil hydrochloride is approved for the treatment of Alzheimer’s disease.
Objective:
To evaluate safety and efficacy of a donepezil patch 27.5 mg application for 52 weeks in patients with mild-to-moderate Alzheimer’s disease; and to evaluate safety on switching from donepezil hydrochloride tablets.
Methods:
This 28-week, open-label study (jRCT2080224517) is an extension of a 24-week double-blind (donepezil patch 27.5 mg versus donepezil hydrochloride tablet 5 mg) noninferiority study. The patch group (continuation group) continued administration of the patch and the tablet group (switch group) switched to the patch in this study.
Results:
A total of 301 patients participated (156 patients continued using patches; 145 patients switched). Both groups showed similar course on the Alzheimer’s Disease Assessment Scale-cognitive component-Japanese version (ADAS-Jcog) and ABC dementia scales. At weeks 36 and 52, changes in ADAS-Jcog from week 24 [mean (standard deviation)] were 1.4 (4.8) and 2.1 (4.9) in the continuation group, and 1.0 (4.2), and 1.6 (5.4) in the switch group. The incidence of adverse events at application site in the continuation group over 52 weeks was 56.6% (98/173). Erythema, pruritus, and contact dermatitis at application site were observed in more than 10 patients each. There was no additional adverse event of clinical concern, and no increase in their incidence from the double-blind study. During the four weeks following switching, no patient discontinued or suspended administration due to adverse events.
Conclusion:
Application of the patch for 52 weeks was well tolerated and feasible, including switching from tablets.
INTRODUCTION
Donepezil hydrochloride has been marketed in oral formulations for the treatment of Alzheimer’s disease [1]. It is an inhibitor of acetylcholine esterase with a piperidine backbone and increases acetylcholine in the brain.
A donepezil patch has been developed by Teikoku Seiyaku Co., Ltd. as a transdermal patch. This has the advantage that blood levels of the active drug increase slowly compared to the oral formulation, providing more stable blood levels. Furthermore, transdermal patches can be used in the patients who have trouble with swallowing [2], improving drug adherence.
A ‘Phase III Efficacy and safety study of a transdermal donepezil patch for patients with mild-to-moderate Alzheimer’s disease’ was conducted on donepezil patches 27.5 mg, and noninferiority to donepezil hydrochloride tablets 5 mg was shown in the first part, ‘a 24-week, randomized, multicenter, double-blind, parallel group, noninferiority study’ [3]. The current 28-week open-label study was conducted as a continuation of the 24-week double-blind study for the assessment of safety and efficacy of long-term administration of a donepezil patch 27.5 mg for up to 52 weeks. The enrolled patients had been treated with either a donepezil patch 27.5 mg or a donepezil hydrochloride tablet 5 mg in the double-blind period of the previous study, and safety on switching from donepezil hydrochloride tablet 5 mg to donepezil patch 27.5 mg was thus assessed as well.
METHODS
Study design
The phase III study consisted of an observation period, a double-blind period (donepezil patch 27.5 mg vs donepezil hydrochloride tablet 5 mg), and an open-label period (donepezil patch 27.5 mg). In the current study, which covers the open-label period, the patients from the double-blind period were enrolled and administered a donepezil patch 27.5 mg in an open-label manner, once a day for 28 weeks. The patch was applied to healthy skin of the upper arm, chest, or back, and when re-applied to the same region, at least seven days must have elapsed.
Evaluation of efficacy and safety was carried out in the ‘continuation’ group, i.e., patients who received a donepezil patch 27.5 mg from the start of the double-blind period until the end of the open-label period, and in the ‘switch’ group, i.e., patients who received donepezil hydrochloride tablet 5 mg during the double-blind period but then switched to a donepezil patch 27.5 mg during the open-label period.
The present study was carried out according to the principles that have their origin in the Declaration of Helsinki and all current legal regulations in Japan. Data collection and management were carried out by a third-party contract research organization to avoid any possible bias.
Inclusion/exclusion criteria
Major inclusion criteria included patients with diagnostic evidence of probable dementia of the Alzheimer-type consistent with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) [4], at screening. Major exclusion criteria included patients with central nervous system disease or systemic disease other than Alzheimer’s disease, affecting cognitive function or its evaluation, inadequate control of high blood pressure, high risks with cardiovascular symptoms, high risks of QT prolongation, an active skin lesion/disorder that may affect adhesion and skin symptom assessment at the application sit, risks with allergy of investigational drugs. Details of main inclusion/exclusion criteria are described in Supplementary Table 1.
Endpoints
Efficacy was evaluated by assessing the change in the Alzheimer’s Disease Assessment Scale-cognitive component-Japanese version (ADAS-Jcog) [5, 6], the ABC dementia scales [7], and the Three-Dimensional Distance (TDD) based on the ABC dementia scale [8], scored from baseline and week 24. In this report, ‘baseline’ referred to the starting point of the double-blind period in the previous study when treatment was started, and the number of weeks was counted from baseline. Safety assessments were applied to adverse events (AEs) including AEs at the application site, laboratory values, vital signs, body weight, and abnormalities on 12-lead electrocardiogram.
Statistical analysis
The full analysis set (FAS) included a group of patients who were administered the investigational drugs and from whom at least one evaluable efficacy datum was obtained after administration of the investigational drugs during the double-blind period.
For the analyses of efficacy, the FAS was analyzed. For the analyses of long-term safety from baseline, the safety analysis set acquired during the double-blind period was analyzed. For the analysis from week 24, the safety analysis set acquired during the open-label period was analyzed. AEs during the four weeks prior to (data not shown) and after switching from donepezil hydrochloride tablet to donepezil patch were analyzed using safety analysis set from the open-label period. Classification was in accordance with Medical Dictionary for Regulatory Activities (MedDRA) version 23.1. All analyses were performed using SAS software, version 9.4 for Windows.
Sample size
The target number of patients was 328 (164 patients/group).
The rationale for the number of patients has been described in the previous study [3].
RESULTS
Study population
Among 303 patients (156 donepezil patch 27.5 mg, 147 donepezil hydrochloride tablet 5 mg) who completed the double-blind period, a total of 301 patients entered the open-label period, with two patients discontinuing participation before administration started in the open-label period (Fig. 1). The reasons for discontinuation were withdrawal of consent and a caregivers’ decision (one patient each). Among the patients who entered the open-label period, 156 patients continued to receive a donepezil patch 27.5 mg and 145 patients switched from donepezil hydrochloride tablets 5 mg to a donepezil patch 27.5 mg. Among these patients, 17 (10.9% , 17/156) and 7 (4.8% , 7/145) discontinued, respectively; as a result, a total of 277 patients completed the open-label period. The reasons for discontinuation in the continuation group were AEs (six patients), withdrawal of consent (four patients), worsening of symptoms (three patients), poor compliance (one patient), and other reasons (three patients); in the switch group the reasons were AEs (two patients), withdrawal of consent (three patients), caregivers’ decision (one patient), and other reasons (one patient)(Fig. 1).

Patient disposition. †patients who continued their donepezil patch 27.5 mg from the double-blind period; ‡patients who switched from donepezil hydrochloride tablets 5 mg to a donepezil patch 27.5 mg.
The safety analysis set in the open-label period included 301 patients, and the FAS comprised 338 patients.
There was no major difference in demographics (sex, age, body weight, duration of illness) or clinical characteristics [Mini-Mental State Examination (MMSE), ADAS-Jcog, ABC dementia scale, use of prior cholinesterase inhibitor] between the continuation group and the switch group in safety analysis set (Supplementary Tables 2). In both groups, there were more women than men (66.5% and 61.4% , for each group). The mean ages were 79.7 (range: 60 to 93) and 79.3 (range: 61 to 95) for the respective groups. The mean body weights were respectively 51.2 kg (range from 30.8 to 77.3) and 53.6 kg (range from 33.0 to 82.2). The means ADAS-Jcog scores when the double-blind period started were similar (23.7 and 22.9). Patients in the FAS had similar demographics to those in the safety analysisset.
Efficacy
At weeks 28, 36, and 52, the mean changes in the ADAS-Jcog from baseline [mean (standard deviation [SD])] were 0.5 (4.3), 0.8 (5.1), and 1.2 (5.2) in the continuation group. The mean changes in the ADAS-Jcog from week 24 [mean (SD)] were 1.1 (4.6), 1.4 (4.8), and 2.1 (4.9) in the continuation group, and 1.1 (4.0), 1.0 (4.2), and 1.6 (5.4) in the switch group (Fig. 2, Supplementary Table 3). The mean changes in the ABC dementia scale (total scores) from baseline [mean (SD)] were – 2.4 (7.8), – 2.1 (8.2), and – 3.2 (9.4) in the continuation group. The mean changes in the ABC dementia scale (total scores) from week 24 [mean (SD)] were – 1.0 (4.3), – 1.1 (5.1), and – 2.6 (6.8) in the continuation group, and – 0.1 (4.4), and – 1.0 (4.9), and – 2.8 (6.6) in the switch group (Supplementary Figure 1, Supplementary Table 4).

Changes in the ADAS-Jcog from baseline (full analysis set). †screening; ‡baseline; §patients who continued their donepezil patch 27.5 mg from the double-blind period; ¶patients who switched from donepezil hydrochloride tablets 5 mg to a donepezil patch 27.5 mg; ‖ADAS-Jcog, Alzheimer’s Disease Assessment Scale-cognitive component-Japanese version.
Safety
No patient dropped out in double-blind period due to poor compliance (non-compliance for more than three weeks) in both groups. The incidence of all AEs over 52 weeks in the continuation group was 88.4% (153/173) (Table 1) and among these, the incidence of adverse reactions was 59.0% (102/173). No AE occurred that led to death. The incidence of serious AEs was 17.9% (31/173), of which decreased appetite and weight decreased (one patient each) were judged to be related to investigational drugs. AEs that led to discontinuation over the 52 weeks occurred in 7.5% (13/173) (Supplementary Table 5) and among these the incidence of adverse reactions was 2.9% (5/173).
AEs commonly observed were AEs for General disorders and administration site conditions 46.2% (80/173), Infections and infestations 26.6% (46/173), and Skin and subcutaneous tissue disorders 28.9% (50/173). The incidence of AEs for Gastrointestinal disorders was 17.9% (31/173) (Table 1). These included diarrhea (seven patients), constipation, nausea, and vomiting (five patients, each), and also dental caries and periodontal disease (two patients each) (data not shown). The incidence of AEs for Cardiac disorders was 11.6% (20/173) (Table 1), and electrocardiogram QT prolonged was also observed in seven patients and counted as an AE with respect to Investigations (data not shown).
AEs over 52 weeks; continuation group (safety analysis set)
MedDRA Ver. 23.1. †system organ class.
Grade 4 or 5 AEs did not occur. Grade 3 AEs occurred in 17.3% , Grade 2 AEs in 59.0% , and Grade 1 AEs in 12.1% ; most AEs were Grade 2 or lower (Supplementary Table 6). The incidence of AEs by time of occurrence was 63.0% (109/173) from week 0 to 12, 40.6% (67/165) from week 13 to 24, 48.1% (75/156) from week 25 to 36, 31.3% (46/147) from week 37 to 48, and 13.4% (19/142) from week 49 or later (Supplementary Table 7).
The incidence of AEs at the application site over 52 weeks in the continuation group was 56.6% (98/173). Those that occurred in more than 10 patients were application site erythema 29.5% (51/173), application site pruritus 26.6% (46/173) and dermatitis contact 12.7% (22/173) (Table 2). Among all AEs that led to discontinuation, those related to cutaneous symptoms at the application site were application site pruritus 1.7% (3/173) and dermatitis exfoliative generalized 0.6% (1/173) (Supplementary Table 5), while application site pruritus that occurred in three patients was considered by the investigators to be related to the investigational drug. However, all AEs at the application site over 52 weeks in the continuation group were Grade 2 or Grade 1 (Supplementary Table 8). The incidence of AEs at the application site by time of occurrence was 33.5% (58/173) from week 0 to 12, 15.8% (26/165) from week 13 to 24, 13.5% (21/156) from week 25 to 36, 5.4% (8/147) from week 37 to 48, and 3.5% (5/142) from week 49 or later (Supplementary Table 9).
AEs at application site over 52 weeks; continuation group (safety analysis set)
MedDRA Ver. 23.1. †system organ class; ‡preferred term.
The incidence of AEs during the four weeks after switching from donepezil hydrochloride tablets 5 mg to a donepezil patch 27.5 mg was 46.9% (68/145). AEs commonly observed were AEs for General disorders and administration site conditions 25.5% (37/145) and Skin and subcutaneous tissue disorders 8.3% (12/145) (Table 3). The incidence of these AEs in the continuation group during the early part of the open-label period (week 25 to 36) was 10.9% (17/156) and 2.6% (4/156), respectively (Supplementary Table 9). The incidence of AEs for Gastrointestinal disorders and AEs for Cardiac disorders was 0.7% (1/145) and 2.1% (3/145), respectively (Table 3). No patient discontinued or suspended administration during the four weeks after switching.
AEs during the four weeks after switching from donepezil hydrochloride tablets 5 mg to a donepezil patch\\ 27.5 mg; switch group (safety analysis set)
MedDRA Ver. 23.1. †system organ class.
DISCUSSION
Long-term safety and efficacy of a 27.5 mg donepezil patch was evaluated in patients who continued treatment from the 24-week double-blind period of the previous study through the 28-week open-label period of the current study. Safety was evaluated in the current study during the open-label period after switching from a donepezil hydrochloride tablet 5 mg to a donepezil patch 27.5 mg.
Both continuous group and switched group showed similar trend of cognitive decline up to 52 weeks. During the 28-week open-label period, the ADAS-Jcog and ABC dementia scale showed a similar time course in the continuation group and the switch group. As the donepezil patch 27.5 mg was shown to be noninferior to donepezil hydrochloride tablets 5 mg in the previous 24-week study [3], cognitive function in the continuation group tended to be maintained for up to 52 weeks and the difference between the continuation and switch groups tended to decrease at weeks 36 and 52.
By week 28, cognitive function had declined in both groups. An evaluation bias may have accounted for this decline, observed during the first evaluation after all of the patients allocated to tested formulation (donepezil patch). In double-blinded period, the evaluator unknowingly has positive bias, because half of the patients is allocated to widely used formulation, donepezil hydrochloride tablet, which efficacy is proven as a commercial product. However, in open-label period, all of the patients allocated to tested formulation, donepezil patch, which efficacy is not proven yet and the evaluator wonder effectiveness of the treatment.
The incidence of AEs at the application site during the 52 weeks was 56.6% (98/173) and there was concern about the effect of local AEs on compliance. However, AEs that led to discontinuation due to skin symptoms were infrequent [application site pruritus 1.7% (3/173) and dermatitis exfoliative generalized 0.6% (1/173)]. Application site pruritus occurring in three patients was judged to be related to the investigational drug. However, most AEs were mild (Grade 2 or Grade 1) and disappeared quickly. These results were consistent with a study in which donepezil patches were administered repeatedly to healthy, older Japanese men [9]. In this study, skin symptoms occurred in many subjects in the study, but were mild and disappeared completely within a few days. In relation to the critical adverse reactions described with donepezil hydrochloride oral formulations, the following events were observed in the current study: electrocardiogram QT prolonged in 4.0% (7/173); and cerebral hemorrhage, syncope, and liver disorder each in 0.6% (1/173). Although gastrointestinal symptoms and cardiovascular symptoms are described as the AEs that need attention with oral donepezil hydrochloride formulations [10], the incidence of AEs for Gastrointestinal disorders and AEs for Cardiac disorders was low during the 52 weeks of application of donepezil patches. No delayed AEs of clinical concern were observed during the late stage and no AE became more frequent with increasing duration from the beginning of administration, including AEs at the application site.
The incidence of AEs during the four weeks after switching from donepezil hydrochloride tablets 5 mg to a donepezil patch 27.5 mg was 46.9% (68/145). The type and incidence of AEs other than at the application site were similar to those in patients in the continuation group. Because of the switch, the incidence of AEs at the application site was higher compared to those in the continuation group. However, these AEs were largely mild and occurred relatively early after switching, i.e., AEs for General disorders and administration site conditions (45 events in 37 patients) accounted for about 70 – 80% of those that occurred in the open-label period (66 events in 48 patients; Supplementary Table 10); therefore, many of the AEs after switching occurred during the first four weeks. Furthermore, no patient discontinued or suspended the investigational drug during the four weeks after switching.
As the safety and efficacy profiles observed during the double-blind period continued through the open-label period, it was concluded that the application of a donepezil patch at a dose of 27.5 mg once a day for 52 weeks was well tolerated and feasible for the treatment of patients with mild-to-moderate Alzheimer’s disease, including in cases switched from donepezil hydrochloride tablets 5 mg once a day.
Limitations of the study include the fact that it was conducted under open-label conditions, after completion of the double-blind period; and that efficacy and safety in the switch group might have been affected by the donepezil hydrochloride tablet 5 mg, as no wash-out period was instituted between the double-blind and open-label period. Additional measures are under consideration.
Footnotes
ACKNOWLEDGMENTS
The authors appreciate the contributions of the following study investigators (listed by surname, in alphabetical order): Tetsuo Abe, Abe Clinic; Jun Akanuma, Mabashi Clinic; Masaharu Amagasa, Yamagata Tokushukai Hospital; Tsuguyoshi Asano, Asano Kanamachi Clinic; Mutsuo Enomoto, Enomoto Internal Medicine Clinic; Kengo Fujita, Fujita Neurology Clinic; Hiroyuki Hatsuta, Hatsuta Neurology Clinic; Hiroaki Hino, Yokohama Hoyu Hospital; Shinichiro Ikebe, Ikebe Clinic; Kunihiko Ikeguchi, Jichi-idai Station Brain Clinic; Yukimichi Imai, Wako Hospital; Yukari Imon, Imon Yukari Memory Clinic; Mitsunori Ishikawa, Ishikawa Clinic; Akira Izaki, Meiwakai Izaki Clinic; Makoto Izuta, Kishiwada Tokushukai Hospital; Taizo Kanetou, Tottori Medical Center; Sadao Katayama, Katayama Medical Clinic; Teruaki Kawasaki, Kyoto Clinical and Translational Research Center for Neurocognitive Disorders; Masayuki Kikukawa, Kikukawa Clinic; Osami Kinoshita, Akino Hospital; Tooru Kinoshita, Nozomi Memory Clinic; Masaki Kishiro, Kishiro Mental Clinic; Chiaki Kudoh, Medical Corporation Kudoh Chiaki Neurosurgery Clinic; Akito Kume, Kume Clinic; Tomoko Murakami, Funairi Brain Clinic; Tsutomu Nagamitsu, Nagamitsu Clinic; Kiyoshi Negoro, Negoro Neurology Clinic; Yoshihiko Nishida, Medical corporation Ichiekai Itsuki Hospital; Chika Nishimura, Kurumi Clinic; Mitsuru Nunomura, Teine Neurosurgical Clinic; Jun Ochiai, Nagoya Ekisaikai Hospital; Masayuki Osako, Osako Cocorono Clinic; Shinya Shimoji, Shimoji Neurology Clinic; Tomohiro Shirasaka, Teine Keijinkai Hospital; Takahiro Sugiyama, Kawasaki Saiwai Clinic; Madoka Tokuyama, Nagomi Clinic; Kimiaki Utsugisawa, General Hanamaki Hospital; Hideo Yagi, Koseikai Takeda Hospital; Yasuhiro Yoshii, Koukan Clinic.
The authors thank Professor Tatsuyuki Kakuma (Biostatistics center, Kurume University, Fukuoka, Japan) for his statistical advice. The authors acknowledge the technical assistance with monitoring of this clinical trial, data collection and management, and statistical analysis provided by Intellim Corp., Tokyo, Japan, and medical writing support provided by Kyoko Okamoto (SunFlare Co., Ltd., Tokyo, Japan), which were funded by Teikoku Seiyaku Co., Ltd.
FUNDING
This clinical trial was sponsored by Teikoku Seiyaku Co., Ltd., Kagawa, Japan.
Clinical Trial Registration: jRCT2080224517.
CONFLICT OF INTEREST
YN has received consultant fees from Teikoku Seiyaku (the sponsor of this study), Eisai, Ono Pharmaceutical and Otsuka Pharmaceutical; lecture fees from Daiichi Sankyo, Eisai, Eli Lilly Japan, Meiji Seika Pharma, Mochida, Mylan EPD, Ono Pharmaceutical, Otsuka Pharmaceutical, Sumitomo Pharma, Takeda Pharmaceutical, Towa Pharmaceutical and Viatris; contract research funding or clinical trials from Biogen Japan, MSD and Nippon Boehringer Ingelheim; grants from Daiichi Sankyo, Eisai, Novartis Pharma and Ono Pharmaceutical. TO, RK, KN, and HA are employees of Teikoku Seiyaku. TK is an Editorial Board Member of this journal but was not involved in the peer-review process nor had access to any information regarding its peer-review. II declare no conflict of interest.
DATA AVAILABILITY
Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data is not available.
