Traditional East Asian Herbal Medicine Treatment for Alzheimer’s Disease: A Systematic Review and Meta-Analysis

Alzheimer’s disease (AD) is a leading progressive neurodegenerative disease worldwide, and its treatment is a challenging clinical problem. This review was conducted to evaluate the efficacy and safety of herbal medicine for AD treatment. The PubMed, CENTRAL, EMBASE, CNKI, OASIS, KTKP, and CiNii databases were searched until June 2020 for randomized controlled trials (RCTs) on herbal medicine for AD, and a meta-analysis of 57 RCTs was conducted. For cognitive function, herbal medicine significantly improved the Mini-Mental State Examination (MMSE) and AD Assessment Scale-Cognitive Subscale (ADAS-cog) scores compared with conventional medicine. The MMSE scores showed no significant difference between the groups treated with herbal medicine and donepezil; however, herbal medicine significantly lowered the ADAS-cog score. Acori Graminei Rhizoma-containing and Cnidii Rhizoma-containing herbal medicine significantly improved the MMSE and ADAS-cog scores compared with conventional medicine. Ginseng Radix-containing herbal medicine showed a positive, but not statistically significant, tendency toward improving the MMSE score compared with conventional medicine. Herbal medicine with conventional medicine significantly improved the MMSE, ADAS-cog, and Montreal Cognitive Assessment (MoCA) scores compared with conventional medicine, and herbal medicine with donepezil also significantly improved these scores compared with donepezil. Acori Graminei Rhizoma or Cnidii Rhizoma-containing herbal medicine with conventional medicine significantly improved the MMSE and ADAS-cog scores compared with conventional medicine. Ginseng Radix-containing herbal medicine + conventional medicine significantly improved the MMSE score, but not the ADAS-cog score, compared with conventional medicine. For behavioral and psychological symptoms of dementia, the Neuropsychiatry Inventory (NPI) score was not significantly different between herbal and conventional medicines. Herbal medicine with conventional medicine significantly improved the NPI and Behavioral Pathology in Alzheimer’s Disease Rating Scale scores compared with conventional medicine. The NPI score showed no significant difference between the groups treated with herbal medicine and placebo. Furthermore, herbal medicine with conventional medicine significantly lowered plasma amyloid beta levels compared with conventional medicine alone. Herbal medicine, whether used alone or as an adjuvant, may have beneficial effects on AD treatment. However, owing to the methodological limitations and high heterogeneity of the included studies, concrete conclusions cannot be made.


Introduction
Dementia is a clinical syndrome that affects memory, thinking, behavior, and activities of daily life, and although 2-10% of cases start before the age of 65, this disease mainly

Eligibility Criteria
The initially screened studies were subsequently reviewed and selected according to the following eligibility criteria: only RCTs were included; non-RCTs, case reports, case series, uncontrolled trials, and laboratory studies were excluded, as were trials that failed to provide detailed results. Patients clinically diagnosed with AD by criteria such as the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) or the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) were included, regardless of their age, gender, nationality, and educational background. Patients with other disorders that might affect cognitive function, such as stroke, Parkinson's disease, and traumatic brain injury, were excluded. Patients with any other rare forms of dementia other than AD, such as Lewy Body, frontotemporal, or vascular dementia, were also excluded. Studies using herbal medicine as experimental interventions were included with no limitations on dosage, frequency, and duration of treatment. Various formulations of herbal medicine were allowed (e.g., decoctions, tablets, capsules, powders); however, only studies in which herbal medicine was administered orally were included. Herbal medicine alone and concurrent treatment with conventional therapy were both considered acceptable if herbal medicine was applied only to the treatment group. The control interventions included placebo or conventional medicine.

Outcome Assessment
For the primary outcome, the MMSE score was assessed to evaluate cognitive function. For secondary outcomes, the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-cog) and Montreal Cognitive Assessment (MoCA) Test for Dementia scores were assessed for further evaluation of cognitive function. Secondary outcomes also included the Neuropsychiatric Inventory (NPI) and Behavioral Pathology in Alzheimer's Disease Rating Scale (BEHAVE-AD) scores for behavioral and psychological symptoms of dementia (BPSD) and plasma levels of amyloid-beta. The number and severity of adverse events were also reviewed.

Data Extraction
Two independent reviewers (JEL and SK) extracted following data from the included studies: general information (e.g., authors, title, publication year), study methods (e.g., interventions and comparisons, intervention duration, study design, sample size, randomization details, blinding and any other bias information), participants (characteristics) and outcomes (primary and secondary outcomes, side effects). If there were disagreements between two reviewers, another independent reviewer (KHC) made decisions.

Quality Assessment of Individual Studies
The risk of bias in each included study was assessed using the Cochrane Risk of Bias Tool [19]. Two independent reviewers (JEL and SK) performed the assessment. If there were disagreements between two reviewers, another independent reviewer (KHC) made final decisions. The tool consists of seven sections, including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. Each domain was assessed as "high risk of bias", "unclear risk of bias", or "low risk of bias".

Statistical Analysis
The synthesis was carried out using RevMan V.5.3.5 provided by the Cochrane Collaboration. For continuous data, the pooled results were presented as the mean difference (MD) or standardized MD with 95% confidence intervals (CIs). For dichotomous data, the pooled results were presented as risk ratios with 95% CIs. In terms of heterogeneity, the Higgins I 2 statistics and Chi-square test were used; significant heterogeneity was thought to exist when the I 2 value was greater than 50 or the p value was less than 0.05.

Characteristics of Included Studies
A total of 2114 studies were identified by electronic searches. After eliminating duplicates, 1827 studies were left and screened by the abstracts.

Treatment
Period (Weeks)
Another subgroup analysis was conducted based on the treatment period. Among the studies that evaluated the MMSE score between herbal medicine with conventional medicine and conventional medicine, four studies had a treatment period of fewer than 12 weeks [40,41,48,70], and pooled results showed that herbal medicine with conventional medicine significantly improves the MMSE score compared with conventional medicine when administered for a period of fewer than 12 weeks [MD 2.86, 95% CI (1.35, 4.38), p = 0.0002, I 2 = 84%]. In studies with a treatment period of 12 or more weeks, herbal medicine with conventional medicine also significantly improved the MMSE score compared with conventional medicine [MD 2.76, 95% CI (1.94, 3.57), p < 0.00001, I 2 = 85%] ( Figure A1j).

Publication Bias
Funnel plot analyses were conducted when 10 or more studies were included in the analysis to assess publication bias ( Figure A2) [18]. All funnel plots were asymmetrically distributed, suggesting potential publication bias.

Discussion
AD is a growing global health concern and treatments proven to alter the disease pathology are not available [3]. Symptoms usually start with memory loss as the presenting symptom [77] and progress gradually. Behavioral and psychological symptoms are also common and cause significant distress to caregivers and increase mortality in patients [77]. However, there are no licensed treatments for these symptoms [3]. Dementia is a complex and multifaceted disease, and additional therapeutic methods are needed to help patients adapt to the disease's progress [6].
Amyloid beta accumulation in the brain is thought to initiate the AD process, and during the past 20 years, efforts have been made to develop drugs that affect the formation or clearance of Aβ; however, these treatments have failed to show positive results regarding cognitive function in all past clinical trials [7]. Not only did cognitive outcomes not improve, but for some patients, the symptoms worsened, despite the fact that the treatments reduced brain Aβ levels [78]. Most recently, aducanumab, a human monoclonal antibody that targets Aβ fibrils and oligomers, underwent two phase III trials (EMERGE and ENGAGE). Based on data from these trials, the FDA granted approval to aducanumab on 7 June 2021, the first Alzheimer's drug to be approved in almost 20 years [79]. However, one review had reported that it could not be concluded that aducanumab has clinical benefits based on the data from these trials [80], and many scientists still say that there is not enough evidence to say aducanumab is an effective treatment for Alzheimer's [81]. Therefore, alternative potential neuroprotective and disease-modifying agents are needed.
For cognitive function, herbal medicine showed statistically significant effects compared with conventional medicine in the MMSE and ADAS-cog scores with high statistical heterogeneity. Therefore, subgroup analyses were conducted in an attempt to lower heterogeneity, and herbal medicines used in two or more studies were evaluated. While Yizhi Jiannao Granule and Bushen Phlegm Stasis Compound did not have significant effects, Bushenyisui Formula significantly improved the MMSE score compared with conventional medicine with low heterogeneity. Herbal medicine with conventional medicine significantly improved the MMSE, ADAS-cog, and MoCA scores compared with conventional medicine alone. Through a subgroup analysis, Bushenyizhi Granule with conventional medicine showed a significant effect with low statistical heterogeneity compared with conventional medicine in the MMSE score. When herbal medicine was compared with placebo, there was no significant difference between the two groups regarding MMSE. Additional analyses were performed to compare herbal medicine with conventional medicine with the first-line therapy for AD, donepezil. When herbal medicine combined with donepezil was compared with donepezil alone, herbal medicine with donepezil significantly improved the MMSE, ADAS-cog, and MoCA scores. For BPSD, when herbal medicine was compared with conventional medicine, the NPI score was not significantly different between the two groups. Herbal medicine with conventional medicine was significantly more effective than conventional medicine in the NPI and BEHAVE-AD scores. One study compared the NPI score between herbal medicine and placebo, and the score was not significantly different between the two groups.
A large variety of herbs were used in each herbal medicine of the selected studies. Among them, Acori Graminei Rhizoma appeared in 30 studies and Cnidii Rhizoma in 25 studies, and both were included in the two formulas that showed significant results in the subgroup analyses-Bushenyizhi Granule and Bushenyisui Formula. In addition, Ginseng Radix, which appeared in 17 studies, is one of the most popular herbs, with continuously increasing use worldwide [82]. Therefore, additional analyses were conducted to evaluate the efficacy of herbal medicines containing Acori Graminei Rhizoma, Cnidii Rhizoma, or Ginseng Radix. In this review, results showed that herbal medicine that included Acori Graminei Rhizoma or Cnidii Rhizoma significantly improved MMSE and ADAS-cog scores compared with conventional medicine. Similar results were obtained when herbal medicine including Acori Graminei Rhizoma or Cnidii Rhizoma was combined with conventional medicine.
Acori Graminei Rhizoma was first mentioned in Shinnongbonchokyung (神農本草 經) as an herb that "treats wind-cold damp impediment, cough and counterflow qi ascent, opens the heart portals, supplements the five viscera, frees the nine orifices, brightens the eyes and sharpens the hearing, and helps the articulation of the voice (主治風寒濕痹, 咳 逆上氣, 開心孔, 補五臟, 通九竅, 明耳目, 出音聲)" [83]. Acori Graminei Rhizoma is one of the most frequently cited herbs in terms of memory improvement, forgetfulness, and aging in classical Chinese medical literature [84], and has been used for the treatment of neurodegenerative diseases in Chinese traditional medicine for thousands of years [85]. One of its main components is β-asarone, which is known to have neuroprotective effects via various mechanisms. An experimental study reported that in Aβ42-induced rats, βasarone ameliorated cognitive impairment and reversed the increase of apoptosis in the hippocampus by inhibiting JNK activation, upregulating Bcl-w and Bcl-2, and inhibiting caspase-3 activity [86]. Another experimental study reported that in Aβ42-induced rats, β-asarone improved cognitive impairment and alleviated Aβ deposition by protecting astrocytes, which was possibly achieved by reducing the levels of TNF-α and IL-1β and then downregulating AQP4 expression [87]. Other authors reported that β-asarone mitigated learning and memory impairments and improved synaptic plasticity by suppressing the excess release of IL-6, IL-1β, i-NOS, and COX-2 in dizocilpine-treated mice [88]. In one study, β-asarone reduced the APP, PS1, Aβ, BACE1, and p62 levels while increasing the SYN1, BECN1, and LC3 levels in PC12 cell AD models, suggesting that β-asarone protects the PC12 cell model against Aβ42 by promoting autophagy [89]. In another study, β-asarone increased the activities of SOD and GPX in Alzheimer rats [90]. Overall, β-asarone has anti-inflammatory, antioxidant, and anti-apoptotic effects and protects neurons from damage by amyloid-beta. Cnidii Rhizoma also first appeared in Shinnongbonchokyung (神農本草經) as an herb that "treats wind entering the brain, headache, cold impediment, hypertonicity of the sinews which are sometimes slack and sometimes tense, incised wounds, and blood block in females (主中風入腦, 頭痛, 寒痺, 筋攣緩急, 金創, 婦人 血閉無子)" [83], and is commonly used to treat cardiovascular and neurovascular diseases in traditional Chinese medicine [91]. One of the main components of this herb is tetramethylpyrazine (TMP)-also known as ligustrazine-which has shown therapeutic effects in various neurodegenerative conditions such as ischemic stroke, Parkinson's disease, and AD [92]. One animal study reported that TMP restores the cAMP levels and, thus, has a protective effect on the cAMP/PKA/CREB signaling pathway, which leads to the reversal of memory deficits induced by scopolamine [93]. Another study reported that by inhibiting caspase-3 activation and increasing the Bcl-2/Bax ratio, TMP protects PC12 cells against H 2 O 2 -induced apoptosis and plays a role in neuroprotection [94]. The complexity and multicausality of dementia have become increasingly recognized in recent studies [95]. The above mechanisms imply that Acori Graminei Rhizoma and Cnidii Rhizoma have multitarget neuroprotective effects and can be potential agents for AD.
Ginseng Radix is one of the most popular, high-value herbs worldwide [82], and is traditionally used to treat weakness and fatigue and to promote longevity [96]. Ginseng Radix was first mentioned in Shinnongbonchokyung (神農本草經) as an herb that "tonifies the five viscera, tranquilizes the essence-spirit, relieves ethereal and corporeal souls, suppresses fright palpitations, eliminates pathogens, improves vision, opens the heart, and enhances intellect. When taken for a long period, it makes the body light and prolongs life (主補五臟，安精神，定魂魄，止驚悸，除邪氣，明目、開心、益智。久服，輕身延 年)" [83]. Recently, studies on the use of Ginseng Radix for improving cognitive performance have increased, and Ginseng Radix is known to have antioxidant, anti-inflammatory, and anti-apoptotic effects [82]. Studies have shown that Ginseng Radix has two active components-ginsenosides and ginotonin-which inhibit βand γ-secretase activity, enhance hippocampal neurotrophic factor expression, attenuate neuroinflammation and mitochondrial oxidative stress, and stimulate non-amyloidogenic α-secretase activity [97]. Furthermore, one experimental study reported that fermented ginseng treatment significantly improves memory function in transgenic mouse models, and significantly reduces the soluble A-β42 levels in the cerebral cortex [98]. In this review, Ginseng Radix-containing herbal medicine with conventional medicine significantly improved the MMSE score compared with conventional medicine. However, Ginseng Radix-containing herbal medicine alone did not show statistically significant effects on MMSE and ADAS-cog scores compared with conventional medicine. These results suggest that Ginseng Radix-containing herbal medicine can be used as adjuvant therapy. Combining Ginseng Radix with an AD drug might ameliorate the symptoms of AD by inducing synergistic effects, enhancing efficacy, and having additional neuroprotective effects [97].
Traditional East Asian medicine emphasizes individuality, with its unique characteristic of "pattern identification," a traditional approach to classify disease symptoms into patterns [99,100], which helps practitioners select relevant treatments. Based on this theory, different herbal formulas can be used to treat the same disease [100]. It is considered that herbal medicine administered based on pattern identification is more effective; a previous study showed that herbal medicine taken in accordance with the diagnosed pattern is more effective in acute stroke patients [101]. Therefore, in clinical settings, herbal medicine is usually prescribed as multi-herb formulas based on pattern identification. In the included studies of this review, except for two studies that administered Ginseng Radix as a single herb, Acori Graminei Rhizoma, Cnidii Rhizoma, and Ginseng Radix were administered as components of a formula rather than a single herb. It might be suggested that the positive results of this review regarding Acori Graminei Rhizoma, Cnidii Rhizoma, and Ginseng Radix were achieved by synergistic interactions between the pharmacological actions of the herbs and use of pattern identification. Therefore, Acori Graminei Rhizoma, Cnidii Rhizoma, and Ginseng Radix may be beneficial in clinical settings when added to herbal medicine in accordance with pattern identification.
In this study, two studies reported that herbal medicine with conventional medicine intervention significantly lowers plasma amyloid beta levels compared with conventional medicine [46,68]. However, these results must be interpreted with caution. First, these two studies did not mention the evaluated isoforms of amyloid-beta, which is important because Aβ40 and Aβ42 might have different influences on AD pathology [102]. Moreover, previous studies have shown conflicting results regarding plasma amyloid-beta levels in AD patients. One study reported that plasma Aβ42 levels are significantly higher in AD than in healthy subjects [103], while another study reported that AD patients have reduced plasma Aβ42 and Aβ40 levels [104]. Other authors reported that the levels of plasma Aβ42/40 are significantly low in AD patients [105], and one meta-analysis revealed a statistically insignificant variation in plasma Aβ42 levels between AD patients and controls [106]. Therefore, it is difficult to conclude that AD patients have lower Aβ levels than normal subjects. Furthermore, biomarkers might remain dynamic even after the onset of AD. A previous study showed that patients with an advanced stage of dementia (CDR = 1-2) had increased Aβ40 levels compared to those with an early stage of dementia (CDR = 0.5) [107]. In the future, studies are needed not only to compare patients with controls but also to validate the change in plasma amyloid-beta levels based on disease progression [108]. Referring to these previous studies, concrete conclusions cannot be made on whether the results of this review regarding plasma amyloid beta are beneficial. Blood biomarkers are cost-effective and non-invasive compared with positron emission tomography (PET) imaging or cerebrospinal fluid (CSF) biomarkers and can be collected and analyzed routinely to detect or track the disease [109]. Therefore the trend of biomarkers of AD is moving from CSF to blood [110]. In the future, more studies should be conducted to elucidate the changes in the levels of plasma amyloid beta and other plasma biomarkers in the progression of AD.
A total of three studies compared herbal medicine with placebo [21,25,28], and when assessing the MMSE and NPI scores, no significant differences were found between the two groups. Among the three studies, one study stated that for the placebo group, the main cause for withdrawal was "lack of efficacy," which might have altered the results [25]. Another study that administered Yokukansan to the herbal medicine group was a multicenter trial and reported that the centers that reported strong placebo effects carried out non-pharmacotherapies, and the influence of these therapies cannot be ignored [21]. Furthermore, in this study, Yokukansan significantly improved the "agitation/aggression" and "hallucinations" subcategories of the NPI-Q scale in subgroups with an MMSE score of less than 20 or with an age younger than 74 compared with placebo. Owing to the small number of studies and reasons mentioned above, it is difficult to conclude that herbal medicine is not significantly beneficial compared with placebo. Herbal medicine might still be useful in clinical settings, and further trials with a large sample size and standardization of other therapies are needed.
Three studies compared the NPI score between herbal and conventional medicines [20,30,31], and there was no significant difference between the two groups. However, two of the three studies reported that the herbal medicine group had a significantly lower prevalence of adverse events than the conventional medicine group. Drug safety is an important aspect of long-term disease treatment. A previous pragmatic randomized, open-label trial reported that 81.2% of caregivers reported adverse events of the study medications, and adverse events were the main cause for discontinuing the study drug [111]. Moreover, a meta-analysis reported that all-cause discontinuation and discontinuation due to adverse events were higher with cholinesterase inhibitors than with placebo [112]. Although the differences between herbal and conventional medicines were not statistically significant, herbal medicine might be a safe therapeutic option to consider for BPSD in clinical settings.
This study has some strengths compared with previous meta-analyses. Various interventions and outcomes were investigated. Scales for cognitive function as well as BPSD were investigated, giving a broader perspective on the efficacy of herbal medicine in AD. Databases were searched regardless of the language and publication types and included 57 studies (4257 patients), a number larger than the previous studies, which might increase the validity of the study results. Subgroup analyses were conducted based on herbal medicines used in two or more studies as well as herbal medicines that included Acori Graminei Rhizoma, Cnidii Rhizoma, or Ginseng Radix. Although statistical heterogeneity among the studies could not be resolved, the results of this study might suggest specific herbal medicines that are promising in clinical settings and future research fields. In addition, changes in plasma amyloid-beta levels were analyzed, which has not been done in previous studies.
This systematic review and meta-analysis has several limitations. First, many of the studies showed bias in terms of qualitative research methodology. Many studies did not describe random sequence generation or allocation concealment, and thus selection bias might exist. In addition, the blinding of participants was mostly absent (performance bias), and the blinding of outcome assessors was unclear in most studies (detection bias); these results might have influenced the results of this study. Second, because of the differences in the degrees of dementia, duration of treatment, prescriptions, and doses of herbal medicine, high heterogeneity was observed among the studies. In particular, some studies that were analyzed did not show the severity of AD. This may have possibly affected the evaluation of treatment outcomes and the sake of fair comparison. Third, many studies only reported outcome results measured at the end of treatment, making it difficult to evaluate the longterm effects of herbal medicine in AD patients. Fourth, most of the studies included in this review were conducted in China. Ethnic and sociocultural factors might influence trial outcomes [113]; therefore, it is difficult to extend the results to patients of other ethnic backgrounds. Furthermore, studies conducted in China have a tendency to have high rates of success [114]. Finally, not all studies reported adverse effects and many studies had small sample sizes, making it difficult to draw concrete conclusions.
Although it is difficult to draw concrete conclusions because of methodological limitations of the included studies and high heterogeneity among them, the results of this study indicate that herbal medicine-especially when administered with conventional medicinemay improve the symptoms of AD patients and, therefore, be used as adjuvant therapies in clinical settings to improve the effectiveness and reduce the limitations of conventional medicine. In particular, according to the results of the present study, Bushenyisui Formula, Bushenyizhi Granule (for cognitive decline), and Yokukansan (for BPSD) are noteworthy. However, among the selected studies of the present study, only two studies evaluated plasma amyloid beta. Although the results of the meta-analysis confirmed the possibility that herbal medicine has an amyloid-beta alleviation effect, the amount of research is still insufficient and there is a scientific gap with the current research trend. Therefore, studies on a larger scale and with high quality are needed in future trials, as well as safety evaluations, assessment of amyloid-beta, and analyses of long-term follow-up.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
Acknowledgments: This paper is based on JiEun Lee's doctoral dissertation.

Conflicts of Interest:
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.   Figure A2. Funnel Plots.