Mild cognitive impairment (MCI), an intermediate stage between normal cognitive aging and dementia, is defined as a cognitive decline greater than expected for a person’s age and education level that does not significantly interfere with activities in daily life.1 The prevalence of MCI has gradually increased over time. The prevalence of MCI is 6.7% for individuals aged 60–64 years and 25.2% for individuals aged 80–84 years, and MCI patients present a significantly greater risk of developing Alzheimer’s diseasae (AD). The cumulative incidence of dementia is 14.9% in MCI patients aged greater than 65 years who were followed for 2 years.2 Additionally, MCI potentially poses a significant burden to individuals, families, and society.3 Therefore, early interventions and early prevention of the disease are particularly essential. However, the updated guidelines for MCI2 did not identify high-quality evidence supporting the administration of pharmacological treatments for MCI. In addition, pharmacological interventions may produce adverse side effects.4
Thus, more recent researches increasingly have focused on nonpharmacological interventions, such as physical exercise, cognitive interventions, dietary interventions, and among others. An increasing number of systematic reviews (SRs) has indicated that nonpharmacological interventions may be a viable alternative for MCI patient by potentially maintaining cognitive function and reducing dementia risk. However, due to the quality of both the SRs and their studies and diverse methodologies, the quality of evidence supporting these recommendations for improvements on cognitive functions is low and should be improved. Thus, this review is to obtain, summarize, and evaluate the high-level evidence, and hope that this summary is useful to support the development of future guidelines and clinical decisions. A summary of the evidence should follow the “6S” model of evidence, starting from the highest level of evidence. Therefore, this review focuses on summarizing and appraising the results and methodological quality of nonpharmacological interventions that have emerged from the relevant guidelines, SRs, and meta-analyses.
The inclusion criteria were guidelines including nonpharmacological interventions for MCI and SRs examining (1) patients who were diagnosed with MCI (no limitations on diagnostic criteria); (2) nonpharmacological interventions defined as nondrug, focused, and replicable interventions conducted with the patient that potentially improve some domains of cognitive impairment. In this evidence summary, we considered nonpharmacological interventions that were classified into the following four categories: physical exercise, cognitive intervention, nonpharmacological interventions using traditional Chinese medicine, and dietary interventions; and (3) studies reporting the following outcomes were considered for inclusion: the cognitive symptoms of individuals with MCI were rated using a variety of neuropsychological measures. Hence, we recorded measurement scales according to the following domains: (1) global cognition (e.g., Mini-Mental State Examination [MMSE], Alzheimer’s Disease Assessment Scale-cognition sub-scale [ADAS-cog], Montreal Cognitive Assessment [MOCA]); (2) executive function; (3) working memory; (4) attention; (5) immediate and delayed memory; and (6) orientation. We excluded studies that did not provide measures for any of the outcomes mentioned above, and duplicate publications were also excluded.
We searched PubMed, National Institute for Health and Clinical Excellence (NICE), Scottish Intercollegiate Guideline Network (SIGN), American Academy of Neurology (AAN), and Registered Nurses Association of Ontario (RNAO) for guidelines published from January 2014 to March 2019 to identify relevant publications on this topic.
We also searched China National Knowledge Infrastructure (CNKI), VIP, Wanfang Database, Cochrane Library, Web of Science, PubMed, and CINAHL for SRs and meta-analyses published from January 2014 to March 2019. According to the PICO framework, the following keywords were used: “cognitive dysfunction”, “mild cognitive impairment*”, “mild cognitive disorder”, “mild cognitive decline”, “MCI”, “MCD”, “meta-analysis*”, “systematic review*”, and “meta-analysis and systematic review”. We used the following Medical Subjects Headings and free-text words to search the aforementioned databases. In addition, reference sections of the retrieved reviews were also searched. The complete PubMed search strategy is presented in additional Figure 1.
All reviews identified in the search were independently assessed by two researchers (first and third authors). When eligibility was unclear, disagreements were resolved by a third reviewer, who approved the final list of included studies. For reviews that fulfilled the inclusion criteria, data were independently extracted from each included review in duplicate, including authors’ names, publication year, sample, patients, intervention, comparison, outcomes, and study characteristics.
We used the updated Appraisal of Guidelines for Research and Evaluation (AGREE II)5, 6 assessing the methodological quality of each guideline using which consists of 23 key items organized within 6 domains, to assess the quality of guidelines. Each of the AGREE II items and the two global rating items are rated on a 7-point scale (1—strongly disagree to 7—strongly agree).
We used the Assessment of Multiple Systematic Reviews (AMSTAR 2) tool, to assess the methodological quality of each SR, a 16-item checklist that is widely used to evaluate the methodological quality of quantitative SRs, 6 of which are key items. The evaluation results are divided into high, medium, low, and very low quality.7
In addition, we assessed the overall quality of the evidence8, 9 using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach, which included risk of bias (RoB), consistency, accuracy, indirectness, and risk of publication bias. The evidence level was divided into four grades: high, moderate, low, and very low.
Our search identified one clinical practice guideline (CPG) describing a nonpharmacological intervention for MCI,2 and the basic characteristics of this guideline are shown in Table 1.
In addition, the search strategy identified 1,600 studies in electronic databases. Five hundred and eleven duplicate records were excluded. After assessing the titles and abstracts, 1,044 SRs were omitted, as they did not meet the criteria of the reviews according to the PICO framework. After reviewing the full-text, another 14 SRs were excluded. Finally, 31 SRs were analyzed in this evidence summary. The PRISMA flow diagram for included reviews is presented in Figure 2. Fourteen SRs of physical exercise for MCI were included, six SRs describing cognitive interventions, four SRs analyzing acupuncture, and seven SRs describing dietary interventions (including four articles examining a Mediterranean diet, one article examining vitamin B supplementation, and two articles describing the effects of tea, coffee, and caffeine).
The 2017 updated AAN guideline2 for MCI included various nonpharmacological interventions, such as cognitive intervention, dietary interventions, and physical exercise, among others. The major recommendations are described as follows: (1) clinicians should recommend regular exercise twice a week (Level B) and encourage patients to engage in meaningful, interesting activities and regular exercise; 10, 11 (2) clinicians may recommend cognitive training (Level C); 12 (3) using vitamin E 2,000 IU/day might be ineffective for reducing AD risk (low confidence) and combine use of 300 mg of oral vitamin E and 400 mg of vitamin C daily for 12 months has an uncertain efficacy (very low confidence).13, 14, 15
Basic characteristics of the included guideline.
Study | Nation | Development organization | Published, updated times | Guideline type | References | Recommendation | Grade |
---|---|---|---|---|---|---|---|
AAN, 20172 | America | American Academy of Neurology | 2017 (update) | clinical practice guideline | 103 | In patients with MCI, treatment with exercise training for 6 months is likely to improve cognitive function10,11 | moderate confidence |
There is insufficient evidence to support or refute the use of any individual cognitive intervention 12 | very low confidence | ||||||
When various cognitive interventions are considered as a group, for patients with MCI, cognitive interventions may improve select measures of cognitive function 12 | low confidence | ||||||
In patients with MCI, there is insufficient evidence to support or refute the use of homocysteine lowering therapies in patients with MCI 13 | very low confidence | ||||||
In patients with MCI, use of vitamin E 2,000 IU daily is possibly ineffective for reducing progression to AD 14 | low confidence | ||||||
In patients with MCI, combined use of oral vitamin E 300 mg and vitamin C 400 mg daily over 12 months is of uncertain efficacy 15 | very low confidence |
We used the AGREE II to independently evaluate this guideline in duplicate (ICC>0.9). The total quality of this CPG was good. For this guideline, the appraisers assigned the highest scores to the clarity (72.2%), followed by the scope and purpose (66.70%), editorial independence (62.50%), and participation (53.30%).
Table 2 presents the characteristics of the 14 SRs16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 of exercise interventions and included data from 455 to 117,410 participants. All SRs were published in English. One SR was published in 2019, three in 2018, four in 2017, two in 2016, two in 2015 and two in 2014. Most SRs included randomized controlled trials (RCTs) and only two SRs23, 24 included prospective cohort studies. The studies examined various, diverse intervention types. In the case of Tai Chi, aerobic exercise, and walking were the most commonly used interventions. The average duration of the intervention was 30–90 min/session in 1–5 sessions/week for 3–12 months.
The results of the assessment of the methodological quality of the SRs of exercise interventions are presented in Table 3. Eleven of the remaining SRs16, 17, 18, 19, 20, 21, 24, 25, 27, 28, 29 were of moderate quality, two SRs22, 23 were considered low quality, and one SR26 was considered very low quality.
Commonly unreported items were the presence of an a priori design (
Summary of the characteristics of systematic reviews (SRs) assessing exercise interventions and detailed characteristics obtained from the full data abstraction.
Articles | Study (sample size) | Research design type | Intervention group/(exposure section) | Intervention frequency/time | Control group (unexposed group) | Outcome | Main conclusions |
---|---|---|---|---|---|---|---|
Wang et al. 201916 | 18 (1364) | RCT | aerobic exercise, resistance exercise, mind-body exercise | 30-90 min/time; 1-5 times/week; 6-12 weeks | health education/placebo | MMSE, MoCA, ADAS-Cog | Four exercise types all had significant benefits compared to the control, and resistance exercises outperformed mind-body exercises. |
Loprinzi et al. 2019" | 6 (355) | RCT | exercise intervention | 30-50 min/time; 2-4 times/week; 2-6 months | regular exercise | short term memory, long term memory | Exercise may help to improve MCI memory function |
Song et al. 201818 | 13 (956) | RCT | aerobic exercise, resistance exercise, multi-mode exercise | 30-60 min/time; 1-4 times/week; 3-12 months | regular exercise/placebo | global cognitive function, memory (short term memory, long term memory), executive function | Physical exercise can improve global cognitive function, but there was no significant effect on short-term memory, long-term memory, executive function |
Lam et al. 201819 | 43 (3988) | RCT | aerobic exercise (slow walking) | 60 min/time; 2-3 times/week | regular exercise | physical functions (BMI/strength/flexibility/stride/balance, walking endurance) | Aerobic exercise can improve physical function (strength/flexibility/stride/balance) |
Barreto et al. 201820 | 5 (2878) | RCT | aerobics, Tai Chi | 12/24 months | regular exercise | MMSE | Exercise intervention could not improve MCI cognitive function and reduce the risk of MCI |
Cammisuli et al. 201721 | 9 (768) | RCT | aerobic exercise (slow walking, handball) | 30-60 min; 2-4 times/week; 6-26 weeks | regular exercise/placebo | global cognitive function (MMSE, ADAS-Cog, MoCA) | Aerobic exercise can improve MCI cognitive function |
Zheng et al. 201722 | 3 (455) | RCT | Tai Chi | 24 style of Tai Chi (30 min, 3 times/week), Yang style of Tai Chi (60 min, 2 times/week), Dao style of Tai Chi (90 min, 2 times/week) | regular exercise | memory | Tai chi can improve MCI cognitive function |
Guure et al. 201723 | 45 (117410) | prospective study | physical exercise | more than once a week | — | — | Physical exercise has positive benefits for improving MCI and AD |
Quan et al. 201624 | 17 (24089) | prospective study | slow walking | — | — | — | Slow walking can improve cognitive function |
Zheng et al. 201625 | 11 (1497) | RCT | aerobic exercises (Tai Chi, walking, jogging) | 30-90 min/day; 1-5 times/week; 3-12 months | regular exercise | global cognitive function (MMSE, ADAS-Cog, MoCA), attention, executive function, memory (short- term, long- term memory) | Aerobic exercise can improve MMSE, MoCA, short-term memory and long-term memory, has no significant effect on ADAS-cog, attention and executive function |
Cai and Abrahamson 201526 | 13 (1171) | RCT | aerobic exercise (walking, Tai Chi) | 10 weeks-6/12 months | regular exercise | global cognitive function (MMSE, ADAS-Cog), attention, execution and memory function | Aerobic exercise has an ameliorative effect on MMSE, memory, endurance, attention and executive function |
Ströhle et al. 201527 | 5 (22689) | RCT | western medicine + exercise therapy | 6-12 months | western medicine + regular exercise/western medicine treatment | global cognitive function (ADAS-Cog, MMSE) | Exercise interventions can improve MCI cognitive function |
Wang et al. 201428 | 9 (795) | RCT | aerobic exercise (Tai Chi, walking, stretching) | 6 weeks-12 months | regular exercise | global cognitive function (ADAS-Cog, MMSE) attention, executive function, memory (short-term, long-term memory) | Exercise interventions can improve MCI cognitive function |
Öhman et al. 201429 | 22 (1699) | RCT | physical exercise, walking | 6 weeks-12 months | regular exercise/blank control | global cognitive function, executive function, long-term memory, attention | Exercise interventions can improve MCI global cognitive function, executive function, long-term memory and attention |
Assessment of the methodological quality of systematic reviews (SRs) of exercise interventions.
AMSTAR 2 | Wang et al. 201916 | Loprinzi et al. 2019" | Song et al. 201818 | Lam et al. 20199 | Barreto et al. 201820 | Cammisuli et al. 201721 | Zheng et al. 201722 | Guure et al. 201723 | Quan et al. 201624 | Zheng et al. 201625 | Cai and Abrahamson 2015 26 | Ströhle et al. 201527 | Wang et al. 201428 | Öhman et al. 201429 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Q1 | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ | √ |
Q2 | × | × | × | × | √ | × | × | × | × | √ | × | √ | × | × |
Q3 | × | × | × | × | × | × | × | × | × | × | × | × | × | × |
Q4 | √ | √ | √ | √ | √ | √ | √ | √ | ||||||
Q5 | √ | √ | √ | √ | √ | √ | √ | √ | × | √ | √ | √ | ||
Q6 | √ | √ | √ | × | √ | √ | √ | √ | √ | √ | × | √ | √ | √ |
Q7 | √ | √ | √ | √ | × | √ | × | × | √ | √ | √ | √ | √ | √ |
Q8 | √ | √ | √ | √ | √ | √ | √ | × | √ | √ | √ | √ | √ | √ |
Q9 | √ | √ | √ | √ | √ | √ | × | √ | √ | √ | × | √ | √ | √ |
Q10 | × | × | × | × | × | × | × | × | × | × | × | × | × | × |
Q11 | √ | √ | √ | √ | √ | √ | √ | √ | ||||||
Q12 | √ | √ | √ | √ | × | √ | √ | √ | ||||||
Q13 | √ | √ | √ | √ | √ | √ | √ | √ | ||||||
Q14 | √ | × | √ | √ | √ | √ | √ | √ | ||||||
Q15 | √ | √ | √ | √ | √ | √ | √ | √ | ||||||
Q16 | √ | √ | × | √ | × | √ | √ | × | √ | √ | ||||
Systematically the overall quality evaluate level | moderate | moderate | moderate | moderate | moderate | moderate | low | low | moderate | moderate | very low | moderate | moderate | moderate |
Five SRs16, 18, 22, 27, 28 analyzed the effect of physical exercise on cognitive function in MCI patients, as measured using the ADAS-Cog, MMSE, and MoCA. According to three SRs,18, 27, 28 aerobic exercise can improve global cognitive function significantly than controls (GRADE: moderate, moderate, and high). In another SR22 showed that aerobic exercise can significantly increase MoCA scores and MMSE scores, but it did not significantly influence ADAS-Cog scores (GRADE: low, low, and very low). One SR16 revealed significant beneficial effects of aerobic exercises (
Three SRs18, 22, 28 analyzed the effects of aerobic exercise on executive ability in MCI patients, and aerobic exercise exerted a significant effect on improving verbal fluency test (VFT) scores compared with controls (GRADE: very low, low, and moderate). One SR20 did not observe an effect of aerobic exercise on improving TMT scores (GRADE: moderate).
Three SRs18, 22, 28 analyzed the effects of aerobic exercise on memory in MCI patients. Three SRs all included a subgroup analysis, and only one SR22 showed that aerobic exercise group can significantly improve MCI patients immediate recall and delayed recall ability compared with controls (GRADE: low and low). However, the other two SRs18, 28 did not observe an effect of aerobic exercise on improving immediate recall and delayed recall abilities (GRADE: moderate, moderate, low, and low).
Nineteen evidences of eight SRs of physical exercise were evaluated according to the GRADE system for the quality of evidence. One of them was high quality, seven of them were moderate quality, eight of them were low quality, and three of them were very low quality, which was downgraded due to the RoB, inconsistencies, and imprecision.
Table 4 presents the characteristics of the six SRs30, 31, 32, 33, 34, 35 of cognitive intervention. Across the studies, the sample size varied considerably, ranging from 224 participants to 2,177 participants; the average sample size was 945 participants. One SR was published in 2019, one in 2018, two in 2017, and two in 2016. All SRs included RCTs, four SRs30, 31, 33, 35 used meta-analysis methods, whereas two SRs32, 34 describe their findings by narrative reviews.
The results of the assessment of the methodological quality of the SRs of cognitive intervention are presented in Table 5. Six SRs30, 31, 32, 33, 34, 35 evaluated the efficacy of cognitive interventions, and the SRs were rated as having moderate (
None of the SRs reported funding sources for the studies included in the SRs, and the results of industry-funded studies may occasionally favor sponsored products. Only two SRs employed an a priori design. Two SRs used narrative reviews to describe their findings and four SRs used meta-analysis. Commonly unreported items were an explanation of the methods used to select the study designs included in the review (
Four SRs30, 31, 32, 35 analyzed the effect of cognitive training on cognitive function in MCI patients. One SR32 showed little effect of memory training on improving global cognitive ability compared with controls, but the other three SRs30, 31, 35 reported a moderately significant effect of computerized cognitive training on improving cognitive function among MCI patients (GRADE: very low, very low, and low).
Four SRs31, 32, 33, 34, 35 analyzed the effect of cognitive interventions on memory in MCI patients. Two SRs32, 35 revealed moderate and statistically significant effects of cognitive interventions on working memory. Only one SR31 described a significant improvement in the delayed recall ability of participants who received mnemonic training compared with controls (GRADE: moderate). However, another SR33 did not observe an effect of computer cognitive training on improving memory (GRADE: very low).
A summary of the characteristics of SRs analyzing cognitive interventions and detailed characteristics obtained from the full data abstraction.
Articles | Study (sample size) | Intervention groups | Control groups | Time | Frequency | Outcome | Main conclusions |
---|---|---|---|---|---|---|---|
Wang 201930 | 21 (1470) | cognitive Intervention | blank controls/conventional therapy | — | — | MoCA | Cognitive intervention can effectively improve MCI cognitive function |
Zhao et al. 201831 | 11 (1069) | cognitive training | conventional health education | 4-48 weeks | — | MMSE, ADL, MoCA | Cognitive training can effectively improve MCI cognitive function |
Yang et al. 201732 | 27 (2177) | memory training/rehabilitation is conducted individually or in groups | blank controls/conventional therapy | 4-9 weeks | 30-120 min/time, 3-36 times (70%<10 times) | learning, memory function, immediate response, delayed response, overall cognitive function | Memory training has medium to high benefits for learning, memory, subjective memory, moderate benefits for delayed response and global cognitive function, low benefit for immediate reaction and no obvious effect for recognition |
Le et al. 201733 | 13(692) | conventional therapy + computerized cognitive function training | conventional therapy + conventional cognitive function training | 3-12weeks | 30-45 min/time, 5-7 times/week | immediate effects, different cognitive domains (memory, orientation, attention) | Short-term computerized cognitive training can improve patients' cognitive function; computerized cognitive function training was better than control group in terms of directional ability and attention improvement in different cognitive domains |
Chandler et al. 201634 | 6 (224) | computerized cognitive function training | conventional therapy /conventional cognitive function training | 2-36 weeks an average of 25.5 h (6-130h) | — | memory function | Computerized cognitive training could not improve MCI memory function. |
Hill et al. 201635 | 17 (686) | computerized cognitive training | blank controls/conventional therapy | 4h + | — | global cognition, verbal fluency, working memory, attention | Computer cognitive function training can improve MCI patients' global cognition, and it has significant effects on different cognitive domains (language learning, language memory, working memory, attention) |
Three SRs31,33,35 evaluated the efficacy of cognitive training on attention and showed that cognitive training exerted a statistically significant effect on the attention capacity compared with controls (GRADE: low, moderate, and moderate).
Two SRs evaluated the efficacy of cognitive training on orientation. One SR31 reported a statistically significant effect of cognitive training on orientation compared with controls (GRADE: moderate). However, another SR33 did not observe an effect of cognitive training (GRADE: very low).
Ten evidences of four SRs of cognitive intervention were evaluated according to the GRADE system for the quality of evidence. Four of them were moderate quality, two of them were low quality, and four of them were very low quality, which was downgraded due to the RoB of the included literatures, inconsistencies, and imprecision of the treatment effect.
Table 6 presents the characteristics of the four SRs36, 37, 38, 39 of nonpharmacological interventions using traditional Chinese medicine. The effects of acupuncture were only reported in, and the most recent search included studies published throughout 2016.36 The SRs included data from 5 to 18 primary prevention trials with 565 to 1,095 participants. All four SRs performed a meta-analysis and reported safety.
The results of the assessment of the methodological quality of the SRs of nonpharmacological interventions using traditional Chinese medicine are presented in Table 7. Only one SR36 was considered moderate quality and three SRs 37, 38, 39ere considered low quality.
None of the SRs reported funding sources for the studies included in the SRs (
Assessment of the methodological quality of systematic reviews (SRs) examining cognitive interventions.
AMSTAR 2 | Wang 201930 | Zhao et al. 201831 | Yang et al. 201732 | Le et al. 201733 | Chandler et al. 201634 | Hill et al. 201635 |
---|---|---|---|---|---|---|
Q1 | √ | √ | √ | √ | √ | √ |
Q2 | × | × | × | × | √ | √ |
Q3 | × | × | × | × | × | × |
Q4 | ≠ | ≠ | √ | √ | ≠ | √ |
Q5 | √ | √ | √ | √ | √ | √ |
Q6 | √ | √ | √ | √ | √ | √ |
Q7 | √ | √ | √ | √ | √ | √ |
Q8 | ≠ | √ | √ | √ | √ | √ |
Q9 | √ | √ | √ | √ | √ | √ |
Q10 | × | × | × | × | × | × |
Q11 | √ | × | —— | √ | —— | √ |
Q12 | × | × | —— | × | —— | √ |
Q13 | × | × | √ | × | √ | √ |
Q14 | × | × | √ | √ | √ | √ |
Q15 | × | × | —— | √ | —— | √ |
Q16 | × | × | × | × | √ | √ |
Systematically evaluate the overall quality level | very low | very low | low | low | moderate | moderate |
Three of four SRs36, 37, 39 compared the efficacy of Western medicine alone and acupuncture combined with Western medicine on MMSE scores, and showed a significant difference in MMSE scores. The combination of acupuncture combined with Western medicine produced a statistically significant decrease in cognitive decline (GRADE: very low, low, and low). In addition, one SR reported a significant effect on the outcome of MMSE scores between the two study groups that received acupuncture and acupuncture combined with aricept39 (GRADE: moderate). Two SRs36, 38 revealed a better MMSE score for patients who received acupuncture treatment for MCI than patients who received nimodipine alone (GRADE: moderate and low), as cognitive function was improved to some extent.
Only one SR38 showed an increase in the MoCA score for patients who received acupuncture at multiple sites on the head combined with cognitive training for MCI compared with patients subjected to cognitive training alone (GRADE: low). Importantly, cognitive function was improved to some extent.
All four SRs37, 38, 39 provided detailed descriptions of adverse events. The adverse events of acupuncture may include fainting during treatment, errhysis at the needle sites, and minor hematoma due to a shorter time of applying local pressure at the acupoint, but these adverse effects did not impact the treatment for MCI. Meanwhile, the adverse events of nimodipine therapy may include gastrointestinal reactions and mild headache.
Seven evidences of four SRs of nonpharmacological interventions using traditional Chinese medicine were evaluated according to the GRADE system for the quality of evidence. Two of them were moderate quality, four of them were low quality, and one of them was very low quality, which was downgraded due to the RoB of the included literatures, inconsistencies, and imprecision of the treatment effect.
A summary of the characteristics of systematic reviews (SRs) of acupuncture and detailed characteristics derived from the full data abstraction.
Articles | Study (sample size) | Intervention | Control | Treatment time | Frequency | Key findings | Safety |
---|---|---|---|---|---|---|---|
Deng and Wang 201736 | 5 (568) | Acupuncture | Nimodipine | 8 weeks | 30 min/time | MMSE (3RCT): MD=0.99, [0.71-1.28], P<0.01 | 3 RCT reported the safety of acupuncture, 2 RCT adverse reactions may occur in the area of acupuncture, 1 RCT may appear megrim; another 1 RCT mention of gastrointestinal reactions and minor headaches in the nimodipine group |
3 times/week | Picture cognition (2RCT): MD = 2.12, [1.48-2.75], P<0.01 | ||||||
Acupuncture + Nimodipine | Nimodipine | 8 weeks | 30 min/time 3 times/week | MMSE (2RCT): MD = 1.09, [0.29-1.89], P<0.01 | |||
Shuai et al.201637 | 18 (1095) | Acupuncture + medicine (Nimodipine, Duxil, Donepezil, Aniracetam) | Single medication (Nimodipine, Duxil, Donepezil, Aniracetam) | 4-24 weeks | 30-50 min/time | MMSE(12RCT): MD = 1.73, [1.28-2.18], P<0.00001 | 4 RCT reported security, 3 RCT mentioned bleeding reaction, 1RCT mention that intervention group and control group all vomiting |
4-6 times/week | ADL (6RCT): MD=5.63, [4.40-6.87],P<0.001 | ||||||
Mai and Zheng 201538 | 5 (565) | electric scalp acupuncture | Nimodipine | 8 weeks | 30 min/time 6 times/week | MMSE(3RCT): MD = 1.33, [0.85-1.82], P<0.0001 | 3 RCT mention causing headaches dizziness, bleeding, motion sickness, Subcutaneous ecchymosis, 1 RCT report no reaction, 1RCT unreported |
cluster needling of scalp acupuncture + cognitive training | cognitive training | 8 weeks | 30 min/time 6 times/week | MoCA(2RCT): MD = 2.12, [0.78-3.47], P = 0.0002 | |||
Hu et al. 201439 | 14 (1052) | Acupuncture + Nimodipine | Nimodipine | 8-9weeks | 30 min/time 3-4 times/week | MMSE(6RCT): MD = 1.19, [0.67-1.70], P<0.00001 | 5/14 RCT reported security, there were 240 cases in the acupuncture group, reported a total of 6 cases of fainting during acupunctures ecchymosis and 3 cases of fainting during acupuncture |
Acupuncture + Aricept | Aricept | 4-6weeks | 30 min/time 6 times/week | MMSE (2RCT): MD=0.70, [0.24-1.17], P =0.003 |
Table 8 presents the characteristics of the seven SRs40, 41, 42, 43, 44, 45, 46 of dietary interventions, and included data from 900 to 84,481 participants. Of these SRs, four examined Mediterranean diets; one analyzed vitamin B supplementation; and two assessed tea, coffee, and caffeine. One SR was published in 2017, two in 2016, two in 2015, and two in 2014. Only one SR explicitly included RCTs, three SRs40, 42, 46 explicitly included cohort studies, two SRs41, 44 included both cohort studies and longitudinal studies, and one SR43 included case–control studies, longitudinal and cross-sectional. Four SRs40, 42, 45, 46 used meta-analysis methods, whereas three SRs41, 43, 44 describe their findings by narrative reviews.
The results of the assessment of the methodological quality of the SRs of dietary interventions are presented in Table 9. Three SRs40, 45, 46 were considered to have moderate quality and majority of the included SRs41, 42, 43, 44 were of low and very low quality.
Commonly unreported items were an explanation of the methods used to select the study designs included in the review (
Only one SR45 analyzed the effects of multivitamin B supplements on people with MCI, and the results showed that had moderate beneficial on memory. However, no significant differences in general cognitive function, executive function, and attention were observed (GRADE: low, moderate, moderate, and low).
A Mediterranean diet is characterized by the consumption of large amounts of vegetables, fruit, cereals, legumes, and unsaturated fatty acids, small amounts saturated fatty acids and meat, small-to-moderate amounts of dairy products, moderate-to-large amounts of fish and regular but moderate consumption of alcohol.40, 41 At present, some articles have reported a positive effect of a Mediterranean diet on MCI.
Assessment of the methodological quality of systematic reviews (SRs) evaluating acupuncture.
AMSTAR 2 | Deng and Wang 201736 | Shuai et al. 201637 | Mai and Zheng 201538 | Hu et al. 201439 |
---|---|---|---|---|
Q1 | √ | √ | √ | √ |
Q2 | × | × | × | × |
Q3 | × | × | × | × |
Q4 | √ | ≠ | ≠ | √ |
Q5 | √ | × | × | × |
Q6 | √ | √ | √ | √ |
Q7 | √ | √ | √ | √ |
Q8 | √ | √ | √ | √ |
Q9 | √ | √ | √ | √ |
Q10 | × | × | × | × |
Q11 | √ | √ | √ | √ |
Q12 | × | × | × | × |
Q13 | √ | × | × | × |
Q14 | √ | √ | √ | √ |
Q15 | √ | √ | √ | √ |
Q16 | √ | √ | × | × |
Systematically evaluate the overall quality level | moderate | low | low | low |
A summary of the characteristics of systematic reviews (SRs) analyzing dietary interventions and detailed characteristics obtained from the full data abstraction.
Articles | Study (sample size) | Exposure group/ intervention group | Unexposed group/ control group | Design | Results |
---|---|---|---|---|---|
Wu and Sun 201740 | 9 (34168) | Mediterranean diet | unexposed group | cohort studies | High adherence to the Mediterranean diet can delay cognitive function decline, reduce the risk of AD |
Hardman et al. 201641 | 18 (59928) | Mediterranean diet | unexposed group | cohort, longitudinal studies, RCT | High Mediterranean adherence diet towards can delay cognitive decline and reduce the risk of AD |
Liu et al. 201642 | 11 (29155) | coffee | unexposed group | cohort studies | Moderate coffee intake can delay cognitive decline and reduce the risk of MCI/AD |
Panza et al. 201543 | 28 (56384) | coffee | unexposed group | cross-sectional, longitudinal studies, case control | Moderate coffee intake can delay cognitive decline and reduce the risk of MCI/AD |
Van de Rest et al. 201544 | 26 (84481) | Mediterranean diet | unexposed group | cross-longitudinal sectional studies, RCT | Higher Mediterranean adherence diet towards can delay cognitive decline and reduce the risk of AD |
Li et al. 201445 | 5 (900) | Vitamin B | placebo | RCT | Vitamin B intake had no significant effect on MCI global cognitive function, executive function and attention |
Singh et al. 201446 | 5 (3636) | Mediterranean diet | unexposed group | Cohort studies | Higher adherence towards Mediterranean diet can delay cognitive decline and reduce the risk of AD |
One SR40 included nine articles, and follow-up period ranged from 2.2 to 12 years. The Mediterranean diet score (category of high vs. low) was significantly associated with the incident risk of cognitive disorders (GRADE: moderate). However, the Mediterranean diet score (category of median vs. low) was not significantly associated with the risk of developing cognitive disorders (GRADE: low). In addition, there was no significant linear association between the Mediterranean diet and the incident risk of all types of cognitive disorders41. Singh B46 and colleagues reported that highest MeDi tertile had a 33% lower risk of cognitive impairment than the lowest MeDi score tertile (GRADE: moderate). In addition, another SR41 also revealed higher adherence to a Mediterranean diet can slower rate of cognitive decline and reduced conversion to Alzheimer’s disease. The specific cognitive domains that improved as the MeDi score increased were memory, executive function, and visual constructs. Furthermore, another SRs44 showed that better adherence to a Mediterranean diet can reduce cognitive decline.
Assessment of the methodological quality of systematic reviews (SRs) evaluating dietary interventions.
AMSTAR 2 | Wu and Sun 201740 | Hardman et al. 201641 | Liu et al. 201642 | Panzaetal.201543 | van de Rest et al. 201544 | Li et al. 201445 | Singh et al. 201446 |
---|---|---|---|---|---|---|---|
Q1 | √ | √ | √ | √ | √ | √ | √ |
Q2 | × | × | × | × | × | × | √ |
Q3 | × | × | × | × | × | × | × |
Q4 | ≠ | √ | ≠ | ≠ | ≠ | √ | |
Q5 | √ | × | √ | × | √ | √ | √ |
Q6 | √ | × | √ | × | √ | √ | √ |
Q7 | √ | √ | √ | √ | √ | √ | |
Q8 | √ | √ | √ | √ | √ | √ | √ |
Q9 | √ | × | × | × | × | √ | √ |
Q10 | × | × | × | × | × | × | × |
Q11 | √ | √ | √ | √ | |||
Q12 | √ | √ | √ | √ | |||
Q13 | √ | √ | √ | × | √ | √ | √ |
Q14 | √ | √ | √ | × | √ | √ | √ |
Q15 | √ | √ | √ | √ | |||
Q16 | √ | √ | × | √ | √ | √ | √ |
Systematically evaluate the overall quality level | moderate | low | low | very low | low | moderate | moderate |
Eleven prospective studies, including 29,155 participants, were included in the SRs.42 The dose– response analysis did not show an association between increased coffee intake and cognitive decline (GRADE: low) or cognitive impairment (GRADE: low). In addition, another SR43 included some cross-sectional, case–control, and longitudinal population-based studies that evaluated the long-term effects on brain function and provided some evidence supporting protective effects of tea, coffee, and caffeine used on late-life cognitive impairment/decline, but the association lacked a distinct dose–response association.
Nine evidences of four SRs of dietary intervention were evaluated according to the GRADE system for the quality of evidence. Four of them had moderate quality and five of them had low quality. GRADE evidence from observational studies was initially described as low-quality evidence, and observational studies generally had not been upgraded due to lack the upgrade factors.
This study evaluated and summarized the evidence from a guideline and many SRs of treatments for MCI patients using nonpharmacological interventions to provide the largest amount of data possible. However, some reviews exhibited a lack of inclusion of high-quality studies or SRs.
The updated guideline22 for mild cognitive impairment (MCI) indicates that clinicians should recommend cognitive training (Level C) and regular exercise (Level B); however, several dietary interventions were not recommended by the guideline. In addition, although the evidence was not robust, the use of Mediterranean diets may have some benefits, but the effects of tea, coffee, and caffeine remain unclear.
Many SRs of nonpharmacological interventions for changing (or at least maintaining) cognitive function have also been conducted; the impact of these changes on improving cognitive function and reducing the AD incidence has produced mixed results. An in-depth summary of the current SRs suggested that dietary interventions, physical exercise, and cognitive interventions are the most widely studied treatments. In general, nonpharmacological interventions for MCI are very important, some reviews showed that has a good effect and can be used as a supplementary treatment for MCI patients. Meanwhile, these findings were promising, but challenges remain.
High-quality SRs will likely provide less biased and relatively conclusive scientific evidence for clinical practice and health decision-making processes.47 Therefore, ARs require strict control over the quality of reviews and methodology.48
The methodological quality of included SRs was generally low. The most common reason was a lack of reporting of funding source for the studies included in all of the reviews. The quality of some SRs may be limited by the presence of an a priori study design, which may lead to publication bias. Moreover, some SRs lack a comprehensive literature search strategy. In addition, because of the variability and heterogeneity of the intervention methods, and assessment tools, data consolidation is difficult. Generally, the author often cannot subgroup analysis, thus it is difficult to draw strong conclusions about the efficacy of these interventions. Actually, different outcomes were reported by researchers focusing on the same health problems. However, due to the lack of a unified standard intervention and complete outcome description, the data were difficult to merge. Furthermore, one of the other new items was the new supplementary items of AMSTAR 2 assessing whether review authors explained their methods for selecting the study designs included in the review, a very important component RCTs were defined as high-quality evidence. However, observational studies may be the only studies available to answer some questions, for example, determine ethical reasons. Therefore, the assessment of the effects of the Mediterranean diet was better analyzed using observational studies than an RCT. Thus, authors should justify the inclusion of different study designs in SRs.7
In addition, based on the conclusions of SRs, we assessed the quality of evidence for outcomes of the meta-analysis, which were downgraded due to (1) RoB: the included studies have major defects in randomization method, allocation concealment and blinding methods. For example, the randomization method is not sufficiently reported, the allocation concealment is unreported or blinding method is not mentioned; (2)imprecision: information included in the study sample size did not meet the optimal sample size (line to a rough estimate, for the dichotomous variable data, if the total number of events of the quality of evidence is less than 300; for the continuous variable data, if some evidence total sample is less than 400, are more likely to consider the inconformity to the OIS); (3) inconsistencies: confidence interval of the overlap between different review is small, heterogeneity of
This review summarized and evaluated the evidence for treatment options utilizing nonpharmacological interventions in patients with MCI. To date, some physical exercise, acupuncture, Mediterranean diet, and cognitive training can significantly improve MCI cognitive function. However, the effects of coffee, tea, and caffeine remain unclear. In addition, due to the limited number and quality of the included publications, further research is necessary.