Diagnosis and Treatment for Mild Cognitive Impairment: A Systematic Review of Clinical Practice Guidelines and Consensus Statements

Background: Mild cognitive impairment (MCI) is an important stage between the normal cognitive decline of aging and dementia. The aim of this study was to compare and harmonize the recommendations for the diagnosis and treatment of MCI based on current clinical practice guidelines. Methods: We searched the PubMed, EMBASE, China National Knowledge Infrastructure, Wanfang Database, Chinese Science and Technology Periodical Database, and Chinese Biological Medicine Database from their inception date to April 24, 2021 to identify all published guidelines on MCI. The qualities of the eligible guidelines were appraised by two reviewers using the Appraisal of Guidelines for Research and Evaluation II instrument. Results: Thirteen guidance documents (four guidelines and nine consensus statements) with specific recommendations were included. Nine guidelines and consensus statements covered the screening and diagnosis of MCI. The evaluation of the documents showed that neuropsychological testing and biomarker assessments were the most common recommendations for the diagnosis of MCI. Nine of the 13 guidance documents covered the treatment and management of MCI. The recommendations for the treatment and management were classified into four categories, namely: intervention for risk reduction, pharmacologic interventions, non-pharmacologic interventions, and counseling. Regarding pharmacological interventions, three guidelines recommend no pharmacologic intervention. The use of cholinesterase inhibitors for MCI is contraindicated in three guidance documents, whereas one proposes that cholinesterase inhibitors and memantine should be deprescribed. EHb761®, Chinese herbal decoctions, and Chinese traditional patent medicine are recommended in two documents. A total of seven guidance documents recommend non-pharmacological interventions, including physical activity interventions, cognitive interventions, dietary and nutritional interventions, and acupuncture. Conclusion: An updated search for possible evidence on the diagnosis and treatment of MCI is needed. Potentially effective diagnoses and treatments, either conventional or complementary, and alternative therapies should be highly valued and addressed in correlation with the supporting evidence.

Background: Mild cognitive impairment (MCI) is an important stage between the normal cognitive decline of aging and dementia. The aim of this study was to compare and harmonize the recommendations for the diagnosis and treatment of MCI based on current clinical practice guidelines.

Methods:
We searched the PubMed, EMBASE, China National Knowledge Infrastructure, Wanfang Database, Chinese Science and Technology Periodical Database, and Chinese Biological Medicine Database from their inception date to April 24, 2021 to identify all published guidelines on MCI. The qualities of the eligible guidelines were appraised by two reviewers using the Appraisal of Guidelines for Research and Evaluation II instrument.
Results: Thirteen guidance documents (four guidelines and nine consensus statements) with specific recommendations were included. Nine guidelines and consensus statements covered the screening and diagnosis of MCI. The evaluation of the documents showed that neuropsychological testing and biomarker assessments were the most common recommendations for the diagnosis of MCI. Nine of the 13 guidance documents covered the treatment and management of MCI. The recommendations for the treatment and management were classified into four categories, namely: intervention for risk reduction, pharmacologic interventions, non-pharmacologic interventions, and counseling. Regarding pharmacological interventions, three guidelines recommend no pharmacologic intervention. The use of cholinesterase inhibitors for MCI is contraindicated in three guidance documents, whereas one proposes that cholinesterase inhibitors and memantine should be deprescribed. EHb761 ® , Chinese herbal decoctions, and Chinese traditional patent medicine are recommended in two documents. A total of seven guidance documents recommend non-pharmacological interventions, including physical activity interventions, cognitive interventions, dietary and nutritional interventions, and acupuncture.

INTRODUCTION
Mild cognitive impairment (MCI) is regarded as the transitional period between the normal cognitive decline of healthy aging and dementia (1). It is categorized into amnestic and non-amnestic subtypes (2). MCI is one of the most common diseases of the elderly, and it increases the risk of developing dementia (1). According to recent studies, the prevalence of MCI among elderly people in China is 15.5% (3), 13.11% in Greece (4), and 26.06% in South India (5). Moreover, it has been reported that 6-15% of patients with MCI develop Alzheimer's disease (AD) annually, and each patient with AD would spend $40,000 each year (6). The high conversion rate from MCI to AD places a large financial burden on public health. Thus, it is critical to pay attention to the diagnosis and treatment of MCI.
The first Mayo diagnostic criteria for MCI proposed by Petersen et al. (7) was mainly based on memory problems. The criteria comprised of memory complaint, normal activities of daily living, normal general cognitive function, abnormal memory for the age of the patient, and no signs of dementia (7). However, considering the non-memory problems that can also result in cognitive impairment, Petersen amended the Mayo criteria by adding non-memory criteria (8). Multiple tests, such as the mini-mental state examination (MMSE), the Montreal Cognitive Assessment (MoCA), the Free and Cued Selective Reminding Test (FCSRT), the California Verbal Learning Test, and the Boston Naming Test, have been proposed for screening for MCI. However, no specific accepted test and cutoff score have been recommended for the diagnosis of MCI. This lack of specific standardized tests of assessment cut-off points may influence the accuracy of the diagnosis (8)(9)(10). Furthermore, owing to the complexity of risk factors and the lack of clarity in the diagnosis of MCI, it is also difficult to provide specific treatment for patients. Currently, there are no accepted drugs for the treatment MCI (11). However, new management measures for MCI, such as cognitive interventions, exercise, and herbal medicine, are continuously introduced in clinical practice guidelines (12,13).
Guidance documents, such as clinical practice guidelines and consensus statements, are developed using a systematic method to provide guidance and recommendations for clinicians (14). Guidance documents may focus on different topics, such as screening, diagnosis, or treatment (14). There are several published guidelines and consensus statements for MCI. However, discrepancies exist between the guidelines because of variations in topics and the time-sensitive nature of some of the evidence. Another reason for the variations between the guidelines may be that the guidelines from different countries are based on different sources and qualities of evidence. Hence, it is necessary to summarize and compare MCI recommendations to enable the clinicians to make more thoughtful clinical decisions and to encourage MCI guideline developers to consider evidence comprehensively. The discovery and evaluation of controversial recommendations is an advantage for future research. Thus, the aim of this study was to systematically review, compare, and harmonize the recommendations for the diagnosis and treatment of MCI in the current clinical practice guidelines.

Eligibility Criteria
Clinical practice guidelines and expert consensus were included if they satisfied the following criteria: • with specific recommendations • focus on diagnosis and/or treatment of patients with MCI • not intended for MCI general practice, education, training, certification, and research • not protocols, abstracts, editorial comments, overview/review articles, and systematic reviews.

Basic Characteristics of the Included Documents and Extraction of Recommendations
The reviewers (NL and YC) extracted the basic characteristics and recommendation content of the included guidelines. The basic characteristics included guideline development information (development group, country, year, and number of guideline versions), guideline scope and content (target population, diagnosis, and treatment), evidence support (systematic search and number of references), grading systems of recommendations (level of evidence and strength of recommendations), and conflict of interest (type of funding). The included guidelines were reviewed, and the recommendations were extracted. Information on the level of evidence and strength of recommendations for diagnosis (or screening) and treatment (or management) were extracted separately. For guidelines with more than one version but which were developed by the same organization or group, the recommendations on the diagnosis and treatment of MCI were extracted from the most updated version.

AGREE II Assessment
The qualities of the eligible guidelines were independently appraised by two reviewers (YC and XL) using the Appraisal of Guidelines for Research and Evaluation (AGREE) II instrument. The results of the assessment were checked by the third author (NL). The AGREE II instrument comprises 23 items organized in six quality domains: scope and purpose, stakeholder involvement, rigor of development, clarity of presentation, applicability, and editorial independence. Each of the 23 items is rated on a seven-point scale. A score of 1 indicates "strongly disagree, " whereas a score of 7 represents "strongly agree." The score of each domain is calculated based on the scores of its specific items (15,16). Based on the methodology outlined in a previously published paper, we divided the domain scores into three groups: high quality (67-100%), sufficient quality (33-67%), and low quality (0-33%) (17,18).

Screening and Diagnosis
Nine guidelines and consensus statements covered the screening and diagnosis of MCI (10,12,13,19,24,26,28). Neuropsychological testing and biomarker assessments are the most recommended tests for the diagnosis of MCI. There was agreement between two guidance documents recommending that standard diagnostic criteria, such as the Vascular Behavioral and Cognitive Disorders Society criteria, Diagnostic and Statistical Manual of Mental Disorders 5, Vascular Impairment of Cognition Classification Consensus Study, or the American Heart Association consensus statement, should be used in the diagnostic process for VaMCI (19,26). Two guidelines recommend that self-report from patients should not be solely relied on for clinical history but should be supplemented by reports from people familiar with the patient (10,19). Three guidance documents indicate that clinicians should combine clinical history with neuropsychological testing in the diagnostic process (12,24,26). One guideline recommends making a diagnosis of MCI subtype (10). Two guidance documents recommend that clinicians identify the MCI risk factors that are potentially modifiable (12,28). Seven documents (10,12,13,19,24,26,28) recommend the use of neuropsychological testing for screening and diagnosis. Three (10,19,26) of these seven documents recommend cognitive testing, including assessment using the Modified Mini-Mental State (3MS) examination, the MMSE, the Rowland Universal Dementia Assessment Scale (RUDAS), MoCA, Toronto Cognitive Assessment (TorCA), National Institute for Neurological Disorders and Stroke and Canadian Stroke Network (NINDS-CSN), and FCSRT; three (10,13,19) recommend testing activities of daily living and functional assessment, including assessment using the Activity of Daily Living Scale (ADL), Instrumental Activity Daily Living (IADL) scale, and Functional Activities Questionnaire (FAQ); and one document (10) recommends behavioral and psychological assessment. Three documents (12,24,28) do not recommend a specific diagnostic tool, and one guidance document does not recommend cognitive testing for screening asymptomatic adults (26). Three guidance documents (10,19,26) propose that a physical examination needs to be conducted for diagnosis and for the prediction of the progression to dementia. Two (10,26) of these three documents recommend dual-task gait test, while one document recommends the olfactory function  test and hearing test (10). Regarding diagnostic radiologic examinations, two documents (12,19) recommend magnetic resonance imaging (MRI), and three documents (12,19,23) recommend positron emission tomography (PET). One guideline recommends the assessment of medial temporal lobe atrophy for the identification of hippocampal atrophy (19). Two guidelines recommend conducting blood tests to exclude other potential diseases (10,19). Five guidance documents (10,13,22,24,28) recommend biomarker assessments to help confirm the diagnosis of MCI. Three (10,22,28) of these documents recommend cerebrospinal fluid biomarker tests (CSF tau protein and CSF βamyloid 42), whereas another guideline (12) suggests that there is no accepted biomarker for a definite diagnosis. Three guidance documents recommend follow-up or monitoring the changes in the cognitive status of a patient with MCI (10, 12, 28). The details and levels of recommendations for screening and diagnosis are presented in Table 3 and Supplementary Table 3A.

Treatment and Management
Nine guidance documents covered recommendations for the treatment (10,12,13,20,21,(24)(25)(26)(27) Regarding counseling, a guideline proposes that clinicians should discuss prognosis and long-term planning with the patients and their families (12). The details and levels of treatment recommendations are presented in Table 4 and Supplementary Table 3B.

Key Findings
As MCI is a transitional stage between the cognitive decline of healthy aging and dementia, its management is critical and requires considerable attention. In this study, we systematically collated MCI guidelines and incorporated Cognitive testing for screening asymptomatic adults Activities of daily living testing and functional assessment  the recommendations for review after appraising the qualities of the guidelines. We included a total of 13 MCI guidance documents in this review. The assessment of the guidelines using the AGREE II tool showed that the AAN guideline has the highest methodological quality of all the included guidelines (12). We analyzed the consistency of the recommendations but found some variation among the MCI guidelines. Some differences were found among the nine guidance documents that include diagnostic recommendations. Neuropsychological assessment is recommended in seven guidelines; however, we noted some inconsistency in the guidelines (10,12,13,19,24,26,28). Three guidance documents do not recommend specific tests (12,24,28), whereas others recommend different tests, including MMSE, 3MS examination, MoCA, TorCA, RUDAS, NINDS-CSN neuropsychological protocols, FCSRT, IADL scale, and FAQ. Among the varied tests mentioned above, NINDS-CSN neuropsychological protocols, FCSRT, ADL, IADL scale, and FAQ are focused on testing deficits in a single cognitive domain (10,19), while the MMSE and MoCA tests, which are the most common cognitive screening instruments, are used to evaluate multiple cognitive domains (10,19). Three guidance documents recommend the use of more comprehensive neuropsychological tools for the assessment of multiple cognitive domains in patients with MCI (10,19,26). One document (19) states that cognitive assessment should include both a comprehensive test and at least four single-domain tests, while another document (10) states that cognitive assessment should include either a comprehensive test or several single-domain tests. An inconsistency was found regarding whether comprehensive tests and single-domain tests should be used alone or in combination. Moreover, both guidance documents do not clarify which specific test should be used. When choosing which cognitive test to use, clinicians should consider the sensitivity, specificity, and time efficiency of the test (29). The guidelines indicate that the sensitivity of the MMSE is low when it is used alone (10,26). A transversal study of 229 elderly participants in four medical centers demonstrated that the MMSE has lower sensitivity than the MoCA (30). Three systematic reviews also demonstrated that the sensitivity and specificity of the MoCA are superior to those of the MMSE (29,31,32). Hence, the MoCA may be better than the MMSE as a comprehensive neuropsychological screening tool. Moreover, we found that none of the guidance documents included in this review has details of a cutoff score for the diagnosis of MCI. However, a systematic review concluded that the best cutoff point for MoCA score is 24/25 (31). Another systematic review suggested that the influence of education should be considered when using cutoff values (33). In conclusion, there is still no commonly accepted cutoff score for the diagnosis of MCI. Future studies that involve a further evaluation of the diagnostic accuracy of MCI using different tests and clearly defined cutoff values are needed.
There is a controversial question regarding whether the measurement of biomarkers is effective for the detection of progression toward AD in patients with MCI. The biomarkers Cholinesterase inhibitors and memantine should be deprescribed can be divided into two categories: imaging-based biomarkers and CSF-based biomarkers (34). The imaging-based biomarkers include MRI-based biomarkers and PET-based biomarkers. MRIbased biomarkers contain the hippocampus and entorhinal cortex; PET-based biomarkers mainly contain amyloid PET, tau PET, and fluorine-18-fluorodeoxyglucose (FDG)-PET; CSFbased biomarkers mainly include CSF tau protein, CSF ttau protein, and CSF β-amyloid peptide (Aβ 1−42 ) (35)(36)(37). In the present study, six guidance documents covered biomarker assessments; however, they proposed different recommendations (10,12,13,22,24,28). Five documents (10,13,22,24,28) recommend that biomarker assessments may help confirm the diagnosis of MCI and identify prodromal AD in patients with MCI. One (10) of these documents indicates that the CSF tau protein and CSF β-amyloid 42 tests for amnestic MCI (aMCI) are advantageous for screening patients with AD. However, another guideline indicates that there is no accepted biomarker for predicting the progression of MCI (12). Although assessment of biomarker is not recommended in that guideline, the authors state that they have a positive view of the biomarker (12). Only one guidance document recommends using same the cutoff points for CSF biomarkers regardless of apolipoprotein E genotype (22). There are also some contraindications for the application of CSF biomarker test, such as increased intracranial pressure and coagulopathy (22). About imaging-based biomarkers, two documents give the recommendations about the use of FDG-PET to confirm the diagnosis of AD in MCI (19,23). FDG-PET is also recommended to ascertain the diagnosis of FTLD and DLB in MCI (23). The cost benefit of imaging-based and CSF-based biomarkers is another issue worth considering. The Manchester consensus of MCI includes three of 11 recommendations regarding the use of biomarkers for early detection. However, the consensus indicates that the cost utility of biomarkers remains controversial (28). Variations also exist in the nine documents that include recommendations for the treatment and management of MCI. Three documents do not recommend pharmacological interventions for the treatment of MCI (13,27). However, EHb761 R is recommended in one consensus statement for the improvement of MCI symptoms (13). EHb761 R is an extract from the Ginkgo biloba plant. A systematic review, which include 21 trials with 2,608 patients, concluded that G. biloba may help improve the cognitive function and activities of daily living of patients with MCI (38). Another consensus statement recommends Chinese herbal decoctions and Chinese traditional patent medicine for the treatment of VaMCI (27). A 2-year trial demonstrated that traditional Chinese medicine, such as Bushen capsules, may improve the cognitive performance of patients with aMCI (39). Herb medicine might be a potential treatment for the improvement of MCI symptoms; however, more high-quality evidence is needed to confirm their efficacy. Four guidance documents (12,13,20,26) recommend physical activity interventions for MCI. One (26) of these documents recommends a specific physical intervention, aerobic exercise. A systematic review of randomized control trials indicates that aerobic and resistance exercises could improve cognitive function (40). Some studies have shown that specific physical activity interventions, such as mind-body exercise, Tai Chi, functional task exercise, and yoga, may significantly improve the MCI symptoms (41)(42)(43)(44). The specific physical activity interventions suitable for patients with MCI, the duration of the exercises, and strength of the exercises need to be researched in future studies.
Although guidance on follow-up or monitoring of the changes of MCI patients have been provided by three guidance documents, no detailed information on how to manage it can be determined. As it is worthwhile to clarify how often the patients with MCI should be monitored and what specific assessments the patients should do, this problem needs to be addressed in future guidelines.
In general, most guidelines and consensus statements (seven of 11) included in the present study were of sufficient and low quality. The domain with the lowest score was the "applicability" domain (domain 5). "Applicability" is a key measure of the implementation of a guideline in clinical practice (16). Applicability limitations, lack of description of facilitators and barriers, provision of additional tools, and potential resources may potentially affect health improvement.

Strengths and Limitations
To the best of our knowledge, this is the first paper to summarize the published guideline recommendations for MCI. The main strength of this study is the comprehensive and systematic literature search that was conducted to identify guidance documents related to the diagnosis and treatment of MCI. The guidelines and consensus statements were independently appraised by two reviewers using AGREE II, and the finding on the methodological problems, mainly in the applicability domain, presented as lacking implementation strategies and related application resources, may help in the development of future MCI guidelines. The recommendations were summarized into key recommendations shown in tables, and the consistency of the recommendations and the difference between them were compared.
The limitation of this study is that, although a systematic search strategy was implemented, certain guidelines may have been missed because of language limitations. Moreover, there were no accepted cutoff points for the domain scores of AGREE II. In addition, the fact that we referred to a previous article for the grading of the domain scores might be a matter of dispute (17,18).

CONCLUSIONS
An updated search of possible evidence for the diagnosis and treatment of MCI is needed. Potentially effective diagnoses and treatments, either conventional or complementary, and alternative therapies should be highly valued and addressed in correlation with the supporting evidence. The AGREE II and Reporting Items for Practice Guidelines in Healthcare tools can be referred to when developing guidelines.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.