Preventie cognitieve achteruitgang en dementie

Beoordeeld: 15-10-2024

Project nr: 627002001
Einddatum 10-2016

Uitgangsvraag

Welke adviezen en/of interventies zijn gewenst ter preventie van cognitieve achteruitgang?

Aanbeveling

Zet geen multicomponent interventie in met als enige doel cognitieve achteruitgang of dementie te voorkomen.

Indien een patiënt verhoogd cardiovasculair risico heeft, raadpleeg de richtlijn ‘Cardiovasculair risicomanagement’ (NHG / NIV / NVVC, 2019). Hou hierbij rekening met patiëntvoorkeuren en het risicoprofiel van de patiënt.

Overwegingen

Voor- en nadelen van de interventie en de kwaliteit van het bewijs

Negen studies hebben het effect geëvalueerd van een multicomponent interventie op cognitief en/of functioneel functioneren bij ouderen. In deze studies is er sprake van een risico op bias en de multicomponent interventie is op verschillende manieren geoperationaliseerd. De bewijskracht voor de uitkomstmaten (cognitieve achteruitgang, dagelijkse activiteiten, dementie, cognitieve stoornis en kwaliteit van leven) en de overall bewijskracht komen daarmee op zeer laag. Op basis van de huidige literatuur kan er geen conclusie worden getrokken dat een multicomponent interventie effectief is als preventie voor cognitieve achteruitgang en dementie. Dit betekent dat nieuwe studies met langdurige follow up (en eenduidige multicomponent interventie) nodig zijn voor meer bewijskracht.

De gevonden studies laten een zeer geringe verbetering cq minder achteruitgang zien op cognitieve testen bij multicomponent interventie, maar dit verschil is niet klinisch relevant. Op meer klinische uitkomstmaten zoals het krijgen van lichte cognitieve stoornissen (MCI) of dementie wordt geen effect gezien. Een recent gepubliceerde observationele follow-up voor een studiecohort opgenomen in de huidige literatuursamenvatting rapporteert ook geen risicoreductie op dementie na een follow-up van 12 jaar (dementie incidentie 11,6 % (oorspronkelijke interventiegroep) vs. 11,9% (oorspronkelijke controlegroep), hazard ratio 0.99; 95% CI, 0.89- 1.10; P = .92) (Hoevenaar-Blom, 2021).

Op grond van epidemiologische studies kunnen we patiënten informeren over welke risicofactoren voor dementie er zijn (Livingston, 2020), maar het is nog onduidelijk of al deze risicofactoren echt een causaal verband hebben met het krijgen van een dementiesyndroom en of het terugbrengen van die risicofactoren het risico op dementie verkleint. Hoewel er in deze module geen specifiek literatuuronderzoek is verricht naar unicomponent interventies, acht het cluster het benoemen van de behandeling van hoge bloeddruk als noemenswaardig. Een recente meta-analyse toonde dat langdurige intensieve behandeling van hypertensie (>2 jaar) bij personen van 55 tot 75 jaar het risico op dementie iets verlaagd (Hughes, 2020; Peters, 2022). Het netto-effect is gering maar consistent, 7.0% versus 7.5% bij een gemiddelde trial follow-up van 4.3 jaar. Hughes (2020) rapporteert odds ratio van 0.93 (95%CI 0.88 tot 0.98); absolute risicoreductie (ARR) 0.39% (95%CI 0.09 tot 0.68%) voor intensieve versus standaard bloeddrukverlaging. Peters (2022) rapporteert odds ratio van 0.87 (95%CI 0.75 tot 0.99) voor vermindering van ontwikkeling van dementie met een bloeddrukverlaging van 10/4 mm Hg ten op zichtte van placebo. Sproviero (2021) en Peters (2021) geven aan dat ondanks heterogeniteit in samples en studieopzet bloeddrukbehandeling wel degelijk veelbelovend is voor vermindering van het risico op dementie en dat aanvullende studies nodig zijn. Het individuele risicoprofiel kan bij verschillen per patient. Er zijn 9 risicofactoren geïdentificeerd, waaronder lage opleiding, overgewicht, hypertensie, slechthorendheid, roken, depressie, fysieke inactiviteit, diabetes en weinig sociale contacten. Een lagere opleiding en sociale klasse komen vaak gecombineerd voor met andere bovengenoemde risicofactoren, zoals diabetes mellitus, overgewicht, hoge bloeddruk, roken, overmatig alcoholgebruik (Livingston, 2020) waarop interventie wel effect zou kunnen geven. Toekomstige studies naar het effect van een multicomponente interventie in subgroepen (o.a. lagere opleiding, lagere sociale klasse), met een langere follow-up duur en vroegtijdige start van interventie (bijvoorbeeld vanaf middelbare leeftijd 40-60 jaar) kunnen beter bewijs opleveren.

De multicomponent interventie is op verschillende manieren geoperationaliseerd door de studies in Den Brok (2022) (zie tabel 1 in de literatuursamenvatting). Alle studies geven leefstijl en zelfmanagement advies in de interventie groep, maar in slechts vier studies is lichaamsbeweging onderdeel van de interventie en in drie daarvan ook nog cognitieve training. Hoewel een hoge kwaliteit van zorg aanwezig was in alle studies, waren de interventies in sommige studies minder intensief. Daarnaast varieerde de duur van de interventies van 12 maanden tot 10 jaar. Daarom kan geen uitspraak worden gedaan over het inzetten op interventies gericht op bepaalde risicofactoren of een bepaalde combinatie van risicofactoren, omdat onduidelijk is of specifieke interventies een groter effect hebben. Ook de vraag hoelang interventies zouden moeten duren en of meer intensieve interventies misschien meer effectief zijn is niet te beantwoorden op basis van deze studies. Denk daarbij aan interventies die een sterkere daling in bloeddruk of cholesterol bereiken, een sterkere gewicht reductie of toename in lichaamsbeweging.

Er is geen eenduidig bewijs dat een multicomponent interventie effect heeft op het cognitief functioneren. Dit hangt mede af hoe het cognitief functioneren in deze studies is gemeten. Grotere overeenstemming over hoe cognitieve achteruitgang beoordeeld moet worden, kan helpen om interventies beter te beoordelen. Hafdi (2021) rapporteert cognitieve achteruitgang aan de hand van neuropsychologische tests (NTB), MoCA en MMSE. Op basis van NTB is er bewijs voor een klein gunstig effect van multicomponent interventies op cognitief functioneren. Echter, dit effect is alleen waargenomen in de studies waarin cognitieve training onderdeel was van de multicomponent interventie en dus wellicht een leereffect voor cognitieve testen in de interventiegroep aanwezig is. Op basis van de MOCA is er zeer lage kwaliteit bewijs voor een klein gunstig effect van multicomponent interventies op cognitief functioneren. Er is geen bewijs voor een effect op de MMSE, de meest gebruikte uitkomstmaat.

Waarden en voorkeuren van patiënten (en evt. hun verzorgers)

Voor patiënten is het belangrijk om zelf iets te kunnen doen om cognitieve achteruitgang te remmen of zelfs te voorkomen. De patiënt hecht hierbij veel belang aan uitleg over de risicofactoren die geassocieerd zijn met ontwikkeling van dementie. Hierbij is het goed om uit te leggen dat wanneer men aan de slag wil met leefstijl een multicomponent interventie het meest zinvol is en dat de meeste ervaring en hoogste verwachting bestaat ten aanzien van gecombineerde interventies gericht op cardiovasculaire risicofactoren. De verwachting is dat de meeste patiënten positief staan tegenover de inzet van multicomponent interventies gericht op voeding, beweging, behandelen van cardiovasculaire risicofactoren, sociale activiteiten en cognitieve stimulatie (leren van nieuwe vaardigheden, bijvoorbeeld muziek, taal leren).

Kosten (middelenbeslag)

De kosten van de multicomponent interventie moeten in een groot deel van de gevallen door de patiënt zelf worden betaald. De meerwaarde voor preventie van dementie en cognitieve achteruitgang is niet aangetoond. Of deze kosten opwegen tegen de (beperkte) gezondheidswinst moet per patiënt worden afgewogen. Bij patiënten met een hoog cardiovasculair risicoprofiel is behandelen volgens CVRM-richtlijn zinvol gezien de wel bewezen risicoreductie op hart- en vaatziekten in het algemeen en het effect wat hierdoor secundair mogelijk op de hersenen verwacht kan worden.

Aanvaardbaarheid, haalbaarheid en implementatie

Multicomponent interventies op voeding, beweging, cardiovasculaire risicofactoren en cognitieve training zijn relatief eenvoudig en relatief weinig tijdrovend voor de zorgverlener. Echter gaat het grotendeels om leefstijlinterventies, die veel vragen van zowel de patiënt als diens omgeving. We verwachten meer belemmerende factoren voor personen met geringe gezondheidsvaardigheden, lager inkomen en lage sociale klasse. We weten dat de verandering in leefstijl vaak lastiger is voor deze groepen terwijl de omgeving er minder op is ingericht. Daarnaast worden de kosten vaak niet vergoed en dit is een grotere barrière voor de groepen met een laag inkomen.

Onderbouwing

Achtergrond

De komende jaren wordt een toename verwacht in het aantal mensen met dementie (Livingston, 2020). Een veel gestelde vraag in de spreekkamer van een patiënt die zich zorgen maakt om dementie te krijgen is: ‘dokter wat kan ik er zelf aan doen om cognitieve achteruitgang te remmen of zelfs te voorkomen?’ Leeftijd is de belangrijkste risicofactor. Belangrijke andere risicofactoren voor het ontwikkelen van dementie zijn hypertensie, diabetes, roken, overmatig alcoholgebruik en onvoldoende lichaamsbeweging (Livingston, 2020). Ondanks dat niet zeker is of elk van deze risicofactoren causaal verbonden is met het ontstaan van dementie, ligt een multicomponent interventie waarbij meerdere risicofactoren tegelijkertijd worden aangepakt op basis van de huidige inzichten het meest voor de hand. Ook omdat deze factoren los van dementie relevant zijn voor de preventie van andere aandoeningen, zoals hart- en vaatziekten. Daarom is er een literatuur search gedaan gericht op multicomponent interventies d.w.z. voeding, beweging, cardiovasculaire risicofactoren en cognitieve training in de preventie van cognitieve achteruitgang.

Conclusies

1. Cognitive decline (critical)

Very low GRADE

The evidence is very uncertain about the effect of a multicomponent intervention on cognitive decline when compared with placebo in patients at risk for dementia.

Sources: Hafdi, 2021 (including Ngandu 2015; Andrieu 2017; Richard 2019; Komulainen 2010; Lee 2014; Moll 2016; Li 2018; Chen 2020); Den Brok, 2022

2. Daily activities (critical)

Very low GRADE

The evidence is very uncertain about the effect of a multicomponent intervention on daily activities when compared with placebo in patients at risk for dementia.

Sources: Hafdi, 2021 (including Ngandu, 2015; Moll, 2016; Andrieu, 2017; Richard, 2019; Chen, 2020)

3. Incidence of dementia (important)

Very low GRADE

The evidence is very uncertain about the effect of a multicomponent intervention on the incidence of dementia when compared with placebo in patients at risk for dementia.

Sources: Hafdi, 2021 (including Moll, 2016; Espeland, 2017)

4. Incidence of mild cognitive impairment (important)

Very low GRADE

The evidence is very uncertain about the effect of a multicomponent intervention on the incidence of mild cognitive impairment when compared with placebo in patients at risk for dementia.

Source: Hafdi, 2021 (including Espeland, 2017)

5. Quality of life (important)

Very low GRADE

The evidence is very uncertain about the effect of a multicomponent intervention on the quality of life when compared with placebo in patients at risk for dementia.

Source: Hafdi, 2021 (including Strandberg, 2017)

Samenvatting literatuur

Description of studies

Hafdi (2021) performed a systematic review and meta-analysis and evaluated the effects of multicomponent interventions on the prevention of cognitive decline and dementia in older adults. A systematic literature search was performed up until 28 April 2021. Studies were included if they 1) were randomised controlled trials (RCTs) with a double-blinded, single-blinded, open and 'prospective, randomised, open, blinded end-point' (PROBE) design and an outcome of cognitive function or incident dementia, 2) had a follow-up of ≥12 months, 3) had ≥400 participants, 4) participants were ≥50 years of age, 5) included a multicomponent intervention with more than one component (e.g. pharmacological treatment, physical activity, dietary manipulations), but not including two or more drugs targeting the same therapeutic target (e.g. two antihypertensive medications), 6) had a control condition of no intervention, usual care, placebo or any other type of sham intervention, but not single-component interventions intended to reduce dementia risk, 7) assessed cognitive function or incidence of dementia according to any of the validated criteria current at the time of the study. In total nine randomized control trials (RCTs) comprising 18.452 patients were included (Komulainen, 2010; Ngandu, 2015; Richard, 2019; Lee, 2014; Li, 2018; Espeland, 2017; Andrieu, 2017; Moll, 2016; Chen, 2020). Characteristics of the individual studies are summarized in the evidence tables and table 1.

Den Brok (2022) performed a pooled analysis of individual participant data from two parallel RCTs comprising 4162 patients (Andrieu, 2017 n=1680; Moll, 2016 n=3526). Both RCTs were included in Hafdi (2021). Both RCTs examined the effects of a multicomponent intervention on the prevention of cognitive decline in a community-dwelling population free from dementia at baseline and over 70 years of age. The RCTs compared a multicomponent intervention to usual care. Moll (2016) compared two groups (one intervention group, one control group). Andrieu (2017) compared four groups (two intervention groups, two control groups) also containing a comparison of omega-3 capsules to placebo capsules. For the pooled analysis the omega-3 and placebo groups were merged, thus half of the patients in both groups were on (placebo) and the other half were on omega-3 capsules. Outcome measure was cognitive decline, which was measured with the Mini-Mental State Examination (MMSE). Characteristics of the individual studies are summarized in the evidence tables and table 1.

Table 1. Summary of study characteristics

Study	Population	Intervention	Control	Outcomes	Follow-up
Hafdi, 2021
Komulainen, 2010	1410 patients (group 1: 234, group 2: 234, control: 236), 65.9 yrs	Group 1: aerobic exercise + diet. Group 2: resistance training + diet.	Oral information with general advice on diet and physical activity.	Cognitive decline: MMSE	2 years
Ngandu, 2015 and Strandberg, 2017 (for the outcome quality of life)	1260 patients (631 intervention, 629 control), 69.3 yrs	Nutritional advice, physical exercise, and cognitive training with 6 monthly assessments by study nurse.	Regular health advice.	Cognitive decline: MMSE and composite score neuropsychological test battery (NTB) Quality of life: Medical Outcomes Study 36-Item Short-Form Health Survey (RAND-36) Daily activities: multiple Activities of Daily Living related questionnaires.	2 years
Richard, 2019	2724 patients (1389 intervention, 1335 control), 69 yrs High risk population with ≥2 cardiovascular risk factors or a history of cardiovascular disease and/or diabetes.	Continuous coaching to facilitate self-management of cardiovascular risk factors via internet-based platform with remote support from a coach, goal setting at start, boost-call at 12 months.	Static platform, similar in appearance, with limited general health information only, without interactive components or a remote coach.	Cognitive decline: MMSE and composite score neuropsychological test battery (NTB) Daily activities: late life function and disability instrument.	18 months
Lee, 2014	1115 patients (group 1: 215, group 2: 241, group 3: 211, group 4: 236, control: 212), 77.1 yrs	Group 1: every two months telephonic lifestyle review. Group 2: monthly telephonic lifestyle review plus educational items. Group 3: every two months health worker visit to review lifestyle. Group 4: every two months health worker visit to review lifestyle plus behavioural reinforcement through rewards.	Lifestyle advice.	Cognitive decline: MMSE	18 months
Li, 2018	510 patients (group 1: 116, group 2: 209 control group 1: 38, control group 2: 147), 45.6 yrs High risk population of mine workers with hypertension (SPB > 140, DBP > 90 or taking antihypertensive drugs 2 weeks before the study)	Group 1: lifestyle advice plus medication. Group 2: lifestyle advice only.	Group 1: medication without lifestyle advice. Group 2: usual care (no medication and no lifestyle advice).	Cognitive decline: MoCA	2 years
Espeland, 2017	5145 patients (2570 intervention, 2575 control), 58.8 yrs High risk population with BMI ≥ 25 kg/m2 (≥ 27 kg/m2 if currently taking insulin), Type 2 diabetes mellitus	Lifestyle: 7% weight loss goal with calorie goals and physical activities support.	Three group educational/social support sessions each year for 4 years after randomizations of the last volunteer.	Cognitive decline: modified version of the MMSE (3MSE, 100-point scale), Incidence of dementia Mild Cognitive Impairment	10-13 years
Andrieu, 2017*	1680 patients (group 1: 417, group 2: 420 control: 420), 75.3 yrs High risk population with at least one of the following: spontaneous memory complaint expressed to their physician, limitation in one instrumental activity of daily living, or slow gait speed (≤ 0.8 meter/second, or more than 5 seconds to walk 4 metre).	Group 1: Omega-3 capsules and multidomain intervention with cognitive stimulation, physical activity, and nutrition. Group 2: Placebo capsules and multidomain intervention with cognitive stimulation, physical activity, and nutrition.	Group 1: Daily placebo capsules (vs omega-3). Group 2: Usual care (vs multidomain).	Cognitive decline: MMSE and composite score neuropsychological test battery (NTB) Daily activities: Alzheimer’s Disease Cooperative Study Activities of Daily Living Prevention Instrument (ADCSADLPI)	36 months
Moll, 2016*	3526 patients (1890 intervention, 1636 control), 74.5 yrs	Nurse-led multidomain cardiovascular care (lifestyle advice and medication).	Usual care.	Cognitive decline: MMSE Incidence of dementia Daily activities: Amsterdam Linear Disability Scale	6-8 years
Chen, 2020	1082 patients (549 intervention, 533 control), 75.1 yrs High risk population with subjective memory impairment and/or loss of ≥ 1 instrumental activities of daily living (IADL), and/or timed 6-mertre walk speed ≤ 1 metre/second	Physical training and advice on managing/preventing chronic disease by physician.	Usual care, including offering conventional health education via three monthly telephone calls.	Cognitive decline: MoCA Daily activities: Instrumental Activities of Daily Living	12 months
Den Brok, 2022
Den Brok, 2022	4162 patients (median age 74 yrs)	Andrieu, 2017 Group 1: Omega-3 capsules and multidomain intervention. Group 2: Placebo capsules and multidomain intervention. Moll, 2016 Nurse-led multidomain cardiovascular care (lifestyle advice and medication).	Andrieu 2017 Usual care (vs multidomain), with half participants taking placebo capsules and other half omega-3 capsules. Moll, 2016 Usual care.	Cognitive decline: MMSE	3-4 years
*Included in pooled analysis of Den Brok (2022).

Results

1. Cognitive decline (critical)

Both studies (Hafdi, 2021; Den Brok, 2022) reported on cognitive decline, using 9 RCTs (Komulainen, 2010; Ngandu, 2015; Richard, 2019; Lee, 2014; Li, 2018; Espeland, 2017; Andrieu, 2017; Moll, 2016; Chen, 2020). The outcome cognitive decline was assessed with three different measures:

Composite score from a neuropsychological test battery (NTB)
Mini-Mental State Examination (MMSE)
Montreal Cognitive Assessment (MoCA)

Composite score from a neuropsychological test battery (NTB)

For the comparison between multicomponent intervention and control, three studies (Ngandu, 2015; Andrieu, 2017; Richard, 2019) with 4617 participants were combined by Hafdi (2021) in a meta-analysis. Analyses showed that the mean difference in the NTB score between the multicomponent-group (n=2482) and the control-group (n=2135) was 0.03 (95%CI 0.01 to 0.06) favouring multicomponent intervention, see Figure 1. This difference was not clinically relevant.

Figure 1 Results of the meta-analysis for the effect of a multicomponent intervention on cognitive decline assessed by a neuropsychological test battery (NTB).

Two studies (Ngandu, 2015; Andrieu, 2017) performed subgroup analysis for baseline cognitive status based on MMSE score (Andrieu, 2017: <30 points, Ngandu, 2015: <26 points) and were combined in a meta-analysis for 2331 patients by Hafdi (2021). Analyses showed that the mean difference in the NTB score favoured multicomponent intervention in the low baseline cognition subset 0.06 (95%CI 0.01 to 0.11; 1414 participants), but not in participants with a high baseline cognition (mean difference: 0.01, 95%CI -0.01 to 0.04; 917 participants). This difference was not clinically relevant.

Mini-Mental State Examination (MMSE)

For the comparison between multicomponent intervention and control, four studies (Ngandu, 2015; Andrieu 2017; Richard, 2019) from Hafdi (2021) were combined with Den Brok (2022) in a meta-analysis of 8891 participants. Analyses showed a mean difference between multicomponent-group (n=4774) and the control-group (n=4117) was 0.01 (95%CI -0.06 to 0.08), see Figure 2. This difference was not clinically relevant.

Figure 2 Results of the meta-analysis for the effect of a multicomponent intervention on cognitive decline assessed by MMSE.

Espeland (2017) from Hafdi (2021) was not included in the MMSE meta-analysis as it used a modified version of the MMSE (3MSE, 100-point scale). This study also did not observe a clinically relevant difference in cognitive decline measured with the 3MSE (mean difference: 0.06 point, 95%CI -0.31 to 0.44).

Den Brok (2022) performed a subgroup analysis for baseline cognition based on MMSE < 26 at baseline. In the multicomponent intervention group, the MMSE declined less in those with MMSE <26 at baseline (n=250; mean difference: 0.84 point, 95%CI 0.15 to 1.54) compared to those with MMSE ≥26 at baseline (n=3454; mean difference: -0.03, 95%CI -0.12 to 0.06). This difference was not clinically relevant.

Montreal Cognitive Assessment (MoCA)

For the comparison between multicomponent intervention and control, two studies (Li, 2018; Chen, 2020) with 1554 participants were combined by Hafdi (2021) in a meta-analysis. Both studies have severe study limitations with issues around randomization, trial registration and statistical analyses (described in more detail in the evidence table). Analyses showed that the mean difference between the multicomponent-group (n=874) and the control-group (n=680) was 0.76 (95%CI 0.05 to 1.46) favouring multicomponent intervention, see Figure 3. This difference was not clinically relevant.

Figure 3 Results of the meta-analysis for the effect of a multicomponent intervention on cognitive decline assessed by MoCA.

2. Daily activities (critical)

For the comparison between multicomponent intervention and control, five studies (Andrieu, 2017; Chen, 2020; Moll, 2016; Ngandu, 2015; Richard, 2019) with 7795 participants were combined by Hafdi (2021) in a meta-analysis. Analyses showed that the standardized mean difference between the multicomponent-group (n=4200) and the control-group (n=3595) was -0.00 (95%CI -0.05 to 0.04), see Figure 4. This difference was not clinically relevant.

Figure 4 Results of the meta-analysis for the effect of a multicomponent intervention on daily activities.

3. Incidence of dementia (important)

For the comparison between multicomponent intervention and control, two studies (Espeland 2017, Moll 2016) with 7256 participants were combined by Hafdi (2021) in a meta-analysis. Analyses showed that dementia was present in 155 out of 3771 (4.1%) patients in the multicomponent-group and in 146 out of 3485 (4.2%) patients in the control-group, resulting in a risk ratio of 0.94 (95%CI 0.76 to 1.18). This difference was not clinically relevant.

4. Incidence of mild cognitive impairment (important)

One study (Espeland, 2017) from Hafdi (2021) evaluated the effect of a multicomponent intervention on MCI incidence in 3802 patients. In the intervention-group in 122 out of 1918 patients (6.4%) and in the control-group 124 out of 1884 patients (6.6%) were diagnosed with MCI. This resulted in a risk ratio of 0.97 (95%CI 0.76 to 1.23). This difference was not clinically relevant.

5. Quality of life (important)

One study (Strandberg, 2017) from Hafdi (2021) evaluated the effect of a multicomponent intervention on quality of life in 1260 patients. The standardized mean difference in RAND-36 (SF-36) score between the multicomponent-group (n=631) and the control-group (n=629) was 0.18 for general health, 0.013 for physical function, 0.08 for role physical, 0.05 for role mental, 0.03 for vitality, 0.02 for mental health, 0.02 for social function and 0.01 for bodily pain, all favouring the multicomponent-group. For none of the RAND-36 components the difference between the groups was clinically relevant.

Level of evidence of the literature

1. The level of evidence regarding the outcome cognitive decline started at high but was downgraded by three levels to very low due to study limitations including lack of blinding to allocation of participants and personnel and attrition bias (-1; risk of bias), imprecision (-1; confidence interval crosses boundaries of clinical important difference), and heterogeneity between studies (-1, inconsistency).

2. The level of evidence regarding the outcome daily activities started at high but was downgraded by three levels to very low due to study limitations including lack of blinding to allocation of participants and personnel and attrition bias (-1; risk of bias), heterogeneity between studies and conflicting results between studies (-2, inconsistency).

3. The level of evidence regarding the outcome incidence of dementia started at high but was downgraded by three levels to very low due to study limitations including lack of blinding to allocation of participants and personnel and attrition bias (-1; risk of bias), imprecision (-1; confidence interval crosses boundaries of clinical important difference), and heterogeneity between studies (-1, inconsistency).

4. The level of evidence regarding the outcome mild cognitive impairment started at high but was downgraded by three levels to very low due to study limitations lack of blinding to allocation of participants and personnel (-1, risk of bias), imprecision (-1, confidence interval crosses boundaries of clinical important difference and the inclusion of patients from only one study) and all study participants had a history of diabetes type 2 which means there is uncertain external validity to other populations (indirectness, -1).

5. The level of evidence regarding the outcome quality of life started at high but was downgraded by three levels to very low due to study limitations including lack of blinding to allocation of participants and personnel and attrition bias (-2; risk of bias) and imprecision (-1; confidence interval crosses boundaries of clinical important difference and the inclusion of patients from only one study).

Zoeken en selecteren

A systematic review of the literature was performed to answer the following question: What are the effects of multicomponent interventions intended to improve cognitive/functional course in patients with cognitive symptoms, without pre-existing cognitive disorders, aged 50 years and above?

P: patients with cognitive symptoms, without pre-existing cognitive disorders, aged 50 years and above

I: multicomponent intervention with more than one component (e.g. pharmacological treatment, physical activity, dietary manipulations) with duration at least 12 months

C: placebo, no intervention/usual care, waiting list

O: cognitive decline, daily activities, incidence of dementia, incidence of mild cognitive impairment (MCI), well-being/quality of life

Relevant outcome measures

The guideline development group considered cognitive decline and a reduction in activities of daily living (henceforward: Daily Activities) as critical outcome measures for decision making; and incidence of dementia, incidence of mild cognitive impairment (MCI) as well as well-being/quality of life as important outcome measures for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

Per outcome, the working group defined the following differences as minimal clinically (patient) important differences:

Cognitive decline: RR ≤0.8 or ≥1.25 (dichotomous); 0.5 SD (continuous)
Activities of daily living: RR ≤0.8 or ≥1.25 (dichotomous); 0.5 SD (continuous)
Incidence of dementia: RR ≤0.8 or ≥1.25 (dichotomous)
Incidence of mild cognitive impairment: RR ≤0.8 or ≥1.25 (dichotomous)
Well-being/quality of life: RR ≤0.8 or ≥1.25 (dichotomous); 0.5 SD (continuous)

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until 17 January 2022. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 383 hits. Studies for this module were selected based on the following criteria:

systematic reviews (searched in at least two databases, and detailed search strategy, risk of bias assessment and results of individual studies available), randomized controlled trials, or observational comparative studies,
full-text English language publication,
adults aged ≥ 50 years,
studies including ≥ 20 (ten in each study arm) patients, and
studies according to the PICO.

A total of 207 studies were initially selected based on title and abstract screening. After reading the full text, 205 studies were excluded (see the table with reasons for exclusion under the tab Methods) and 2 studies were included.

Results

Two studies were included in the analysis of the literature, one is a systematic review including nine studies and the other is a pooled analysis for two of the nine studies. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.

Referenties

den Brok MGHE, Hoevenaar-Blom MP, Coley N, Andrieu S, van Dalen J, Meiller Y, Guillemont J, Brayne C, van Gool WA, Moll van Charante EP, Richard E. The Effect of Multidomain Interventions on Global Cognition, Symptoms of Depression and Apathy - A Pooled Analysis of Two Randomized Controlled Trials. J Prev Alzheimers Dis. 2022;9(1):96-103. doi: 10.14283/jpad.2021.53. PMID: 35098979.
Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, Brayne C, Burns A, Cohen-Mansfield J, Cooper C, Costafreda SG, Dias A, Fox N, Gitlin LN, Howard R, Kales HC, Kivimäki M, Larson EB, Ogunniyi A, Orgeta V, Ritchie K, Rockwood K, Sampson EL, Samus Q, Schneider LS, Selbæk G, Teri L, Mukadam N. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020 Aug 8;396(10248):413-446. doi: 10.1016/S0140-6736(20)30367-6. Epub 2020 Jul 30. PMID: 32738937; PMCID: PMC7392084.
Hafdi M, Hoevenaar-Blom MP, Richard E. Multi-domain interventions for the prevention of dementia and cognitive decline. Cochrane Database Syst Rev. 2021 Nov 8;11(11):CD013572. doi: 10.1002/14651858.CD013572.pub2. PMID: 34748207; PMCID: PMC8574768.
Hoevenaar-Blom MP, Richard E, Moll van Charante EP, van Wanrooij LL, Busschers WB, van Dalen JW, van Gool WA. Targeting Vascular Risk Factors to Reduce Dementia Incidence in Old Age: Extended Follow-up of the Prevention of Dementia by Intensive Vascular Care (preDIVA) Randomized Clinical Trial. JAMA Neurol. 2021 Dec 1;78(12):1527-1528. doi: 10.1001/jamaneurol.2021.3542. Erratum in: JAMA Neurol. 2022 Jan 1;79(1):93. Erratum in: JAMA Neurol. 2022 Jan 10;: PMID: 34633434; PMCID: PMC8506299.
Hughes D, Judge C, Murphy R, Loughlin E, Costello M, Whiteley W, Bosch J, O'Donnell MJ, Canavan M. Association of Blood Pressure Lowering With Incident Dementia or Cognitive Impairment: A Systematic Review and Meta-analysis. JAMA. 2020 May 19;323(19):1934-1944. doi: 10.1001/jama.2020.4249. PMID: 32427305; PMCID: PMC7237983.
Sproviero W, Winchester L, Newby D, Fernandes M, Shi L, Goodday SM, Prats-Uribe A, Alhambra DP, Buckley NJ, Nevado-Holgado AJ. High Blood Pressure and Risk of Dementia: A Two-Sample Mendelian Randomization Study in the UK Biobank. Biol Psychiatry. 2021 Apr 15;89(8):817-824. doi: 10.1016/j.biopsych.2020.12.015. Epub 2020 Dec 27. PMID: 33766239.
Peters R, Breitner J, James S, Jicha GA, Meyer PF, Richards M, Smith AD, Yassine HN, Abner E, Hainsworth AH, Kehoe PG, Beckett N, Weber C, Anderson C, Anstey KJ, Dodge HH. Dementia risk reduction: why haven't the pharmacological risk reduction trials worked? An in-depth exploration of seven established risk factors. Alzheimers Dement (N Y). 2021 Dec 8;7(1):e12202. doi: 10.1002/trc2.12202. PMID: 34934803; PMCID: PMC8655351.
Peters R, Xu Y, Fitzgerald O, Aung HL, Beckett N, Bulpitt C, Chalmers J, Forette F, Gong J, Harris K, Humburg P, Matthews FE, Staessen JA, Thijs L, Tzourio C, Warwick J, Woodward M, Anderson CS; Dementia rIsk REduCTion (DIRECT) collaboration. Blood pressure lowering and prevention of dementia: an individual patient data meta-analysis. Eur Heart J. 2022 Oct 25:ehac584. doi: 10.1093/eurheartj/ehac584. Epub ahead of print. PMID: 36282295.

Evidence tabellen

Table of quality assessment for systematic reviews of RCTs and observational studies

Study	Appropriate and clearly focused question?¹	Comprehensive and systematic literature search?²	Description of included and excluded studies?³	Description of relevant characteristics of included studies?⁴	Appropriate adjustment for potential confounders in observational studies?⁵	Assessment of scientific quality of included studies?⁶	Enough similarities between studies to make combining them reasonable?⁷	Potential risk of publication bias taken into account?⁸	Potential conflicts of interest reported?⁹
Hafdi, 2021	Yes	Yes	Yes	Yes	Not applicable	Yes	Yes, although the type of multicomponent intervention per study varies. Some studies also include medication, while others only include lifestyle interventions.	Unclear, cannot be formally ruled out as the authors could not assess publication bias in funnel plot because of the limited number of identified studies	Yes

Study

Appropriate and clearly focused question?¹

Comprehensive and systematic literature search?²

Description of included and excluded studies?³

Description of relevant characteristics of included studies?⁴

Appropriate adjustment for potential confounders in observational studies?⁵

Assessment of scientific quality of included studies?⁶

Enough similarities between studies to make combining them reasonable?⁷

Potential risk of publication bias taken into account?⁸

Potential conflicts of interest reported?⁹

Hafdi, 2021

Yes

Not applicable

Yes

Yes, although the type of multicomponent intervention per study varies. Some studies also include medication, while others only include lifestyle interventions.

Unclear, cannot be formally ruled out as the authors could not assess publication bias in funnel plot because of the limited number of identified studies

Yes

Research question (PICO) and inclusion criteria should be appropriate and predefined
Search period and strategy should be described; at least Medline searched; for pharmacological questions at least Medline + EMBASE searched
Potentially relevant studies that are excluded at final selection (after reading the full text) should be referenced with reasons
Characteristics of individual studies relevant to research question (PICO), including potential confounders, should be reported
Results should be adequately controlled for potential confounders by multivariate analysis (not applicable for RCTs)
Quality of individual studies should be assessed using a quality scoring tool or checklist (Jadad score, Newcastle-Ottawa scale, risk of bias table etc.)
Clinical and statistical heterogeneity should be assessed; clinical: enough similarities in patient characteristics, intervention and definition of outcome measure to allow pooling? For pooled data: assessment of statistical heterogeneity using appropriate statistical tests (e.g. Chi-square, I²)?
An assessment of publication bias should include a combination of graphical aids (e.g., funnel plot, other available tests) and/or statistical tests (e.g., Egger regression test, Hedges-Olken). Note: If no test values or funnel plot included, score “no”. Score “yes” if mentions that publication bias could not be assessed because there were fewer than 10 included studies.
Sources of support (including commercial co-authorship) should be reported in both the systematic review and the included studies. Note: To get a “yes,” source of funding or support must be indicated for the systematic review AND for each of the included studies.

Evidence table for systematic review of RCTs and observational studies (intervention studies)

Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Hafdi 2021

[individual study characteristics deduced from Hafdi 2021]

SR and meta-analysis of 9 parallel RCTs.

Literature search up to 28 April 2021

A: Komulainen, 2010 (DRs extra)

B: Ngandu, 2015 (Finger)

C: Richard, 2019 (Hatice 2019)

D: Lee, 2014

E: Li, 2018

F: Espeland, 2017 (Look AHEAD 2017)

G: Andrieu, 2017 France placebo-randomized superiority trial with four parallel groups including 30 memory centers (MAPT 2017)

H: Moll, 2016 Netherlands open-label cluster RCT including 116 general practices (preDIVA 2016)

I: Chen, 2020 (Taiwan Health 2020)

Three studies were carried out in a community-based setting (A,D,H) and six studies were carried out via a healthcare facility.

One study (Espeland, 2017) is at unclear risk of bias as the study was partially funded by commercial funding sources but did not explicitly state the exact role of their funding sources.

Inclusion criteria SR: double-blinded, single-blinded, open and 'prospective, randomised, open, blinded end-point' (PROBE) design studies with a minimum follow-up of 12 months and 400 or more participants with participants over 50 years of age; 9 studies included.

Participants were over 50 years of age, including

unselected and high-risk populations (with known risk factors for dementia and/or subjective cognitive symptoms or cognitive impairment not fulfilling criteria for a dementia diagnosis).

N, mean age

A: 1410 patients (group 1: 234, group 2: 234, control: 236), 65.9 yrs

B: 1260 patients (631 intervention, 629 control), 69.3 yrs

C: Total: 2724 (1389 intervention, 1335 control), 69 yrs

D: Total: 1115 (intervention 1: 215, intervention 2: 241, intervention 3: 211, intervention 4: 236, Control: 212), 77.1 yrs

E: Total 510 (Intervention: group 1: 116, group 2: 209 Control: group 1: 38, group 2: 147), 45.6 yrs

F: Total: 5145 (intervention: 2570, control: 2575), 58.8 yrs

G: Total: 1680 (Intervention: Group 1: 417, group 2: 420 Control: 420), 75.3 yrs

H: Total: 3526 (Intervention: 1890, Control: 1636), 74.5 yrs

I: Total: 1082 (Intervention: 549 Control: 533), 75.1 yrs

C, E,F,G&I includes a high risk population

C: >=2 cardiovascular risk factors or a history of cardiovascular disease and/or diabetes.

E: mine workers with hypertension (SPB > 140, DBP > 90 or taking antihypertensive drugs 2 weeks before the study)

F: BMI ≥ 25 kg/m2 (≥ 27 kg/m2 if currently taking insulin), Type 2 diabetes mellitus

G: At least one of the following: spontaneous memory complaint expressed to their physician, limitation in one instrumental activity of daily living, or slow gait speed (≤ 0.8 meter/second, or more than 5 seconds to walk 4 metre).

I: Subjective memory impairment and/or loss of ≥ 1 instrumental activities of daily living (IADL), and/or timed 6-mertre walk speed ≤ 1 metre/second

Multi-domain interventions with more than one component, not including two or more drugs targeting the same therapeutic target (e.g. two antihypertensive medications) delivered by trained professionals. Major components of modification include, but are not limited to, pharmacological treatment, physical activity, and dietary manipulations.

All nine studies incorporated lifestyle and self-management advice, no study used a predefined pharmacological treatment, four incorporated exercise (A,B,G,I) and three (B,G,I) also delivered cognitive training.

A: Group 1: aerobic exercise + diet. 5 x 60 minutes/week aerobic exercise. 5 individual exercise or nutrition session in 1st year and every 6 months thereafter
Group 2: resistance training + diet. 2 resistance training sessions/week. 5 individual exercise or nutrition session in 1st year and every 6 months thereafter

B: 6 monthly assessments with study nurse. Nutrition: 3 individual and 7-9 group sessions. Physical exer- cise: 1-3/week muscle strength, 2-5/week aerobic. Cognitive training: 10 group, 72 individual sessions.

C: Continuous coaching to facilitate self-management of cardiovascular risk factors via intern-based platform with remote support from a coach, goal setting at start, boost-call at 12 months.

D: Group 1: every two months telephonic care management (10-15 minutes) reviewing participants' lifestyle behaviour

Group 2: monthly telephonic care management and educational items

Group 3: every two months health worker-initiated visit care management (15-20 minutes) reviewing participants' lifestyle behaviour

Group 4: every two months health worker-initiated visit care management (15-20 minutes) reviewing participants' lifestyle behaviour plus behavioural reinforcement through rewards such as golden medals

E: Group 1: (using medication) Health education and behaviour counselling monthly. Regular health seminars and free education material on prevention of hypertension every 3 months. Monthly SMS reminders to improve lifestyle. Monthly phone calls concerning blood pressure and lifestyle

Group 2: (not using medication) Health education and behaviour counselling monthly. Regular health seminars and free education material on prevention of hypertension every 3 months. Monthly SMS reminders to improve lifestyle. Monthly phone calls concerning blood pressure and lifestyle

F: 7% weight loss goal with targeted daily calorie goals of 1,200– 1,800 according to initial weight and ≥ 175 minutes/week of physical activities such as brisk walking. Intervention sessions occurred weekly at months 1–6 and then tapered to 3 per month for the remainder of the first year, 6 months, and monthly thereafter, with additional support with monthly phone or e-mail contacts.

G: Group 1: Two capsules of PUFA daily and 2-hour group sessions focusing on three domains (cognitive stimulation, physical activity, and nutrition) and a preventive consultation (at baseline, 12 months, and 24 months) with a physician to optimise management of cardiovascular risk factors and detect functional impairments. Twelve small group sessions of the multi-domain intervention were done in the first 2 months of the trial. Each session included 60 minutes of cognitive training, 45 minutes of advice about and demonstrations of physical activity and 15 minutes of nutritional advice.

Group 2: Two capsules of placebo daily and 2-hour group sessions focusing on three domains (cognitive stimulation, physical activity, and nutrition) and a preventive consultation (at baseline, 12 months, and 24 months) with a physician to optimise management of cardiovascular risk factors and detect functional impairments. Twelve small group sessions of the multi-domain intervention were done in the first 2 months of the trial. Each session included 60 minutes of cognitive training, 45 minutes of ad- vice about and demonstrations of physical activity and 15 minutes of nutritional advice.

H: Nurse-led multidomain cardiovascular care (lifestyle advice supported by motivational interviewing techniques and drug treatment for hypertension, dyslipidaemia, and type 2 diabetes mellitus or antithrombotic drugs)

I: Four structured 2-hour training sessions in the first month, two during the second, and one in each of the next 10 months. In addition, every 3 or 4 months, some activities were curtailed, and a visiting doctor instead gave a 30 to 60 minutes class on preventing/man- aging chronic disease

Two studies delivered care as usual via the general practitioner (E,H), one used daily placebo pills and no lifestyle intervention (G), and six gave concise health and lifestyle advice (A-D,F,I).

A: Oral information about general public health advice on diet and physical activity.

B: Regular health advice

C: Static platform, similar in appearance, with limited general health information only, without interactive components or a remote coach.

D: Lifestyle advice

E: Group 1 (compared to intervention group 1): medication + no lifestyle intervention
Group 2 (compared to intervention group 2): no medication + no lifestyle intervention

F: Three group educational/social support sessions each year for 4 years after randomizations of the last volunteer

G: Daily placebo pills (vs omega-3) and usual care (vs multidomain).

H: Usual care

I: Conventional health education in the Efficacy Study control group entailed periodic telephone calls (~3 monthly) by local research site staff to offer participants health education and advice (the intervention group did not receive such calls)

Duration of follow-up:

Outcomes were assessed after follow-up times ranging from 12 months to 13 years.

A: 2 years (for cognition)

B: 2 years

C: 18 months

D: 18 months

E: 2 years

F: 10-13 years

G: 36 months

H: 6-8 years

I: 12 months

Complete outcome data

(intervention/control)

A: 96% complete for MMSE

B: 95% complete outcome data, no indicators of selective drop out

C: 91% of participants in control, 86% in intervention

D: More than half of the participants dropped out during the intervention and were not retrieved for final assessment. It seems like people were first randomised and afterwards they were asked if they were willing to participate

E: Authors report no dropout from the study (at all). No attrition or excluded participants

F: 74% of participants had data on the primary outcome.

G: 90% complete outcome data, number of excluded did not differ significantly between groups. Excluded participants were older and with lower education.

H: Follow-up data for the primary outcome of dementia were obtained for 3454 (98∙0%) participants.

I: 74,7% in intervention and 70.4% in control group available for analysis.

Cognitive decline, incidence of dementia, incidence of cognitive impairment, activities of daily living (ADL) or quality of life. Could be cognitive screening instruments, multi-domain cognitive assessment scales, neuropsychological test batteries, quality of life instruments, ADL instruments or dementia diagnosis according to any of the validated criteria current at the time of the study.

Cognitive decline

Defined as the decline in cognitive functioning measured at two separate occasions with any validated measure, including Mini-Mental State Examination (MMSE), composite score from a neuropsychological test battery (NTB) and Montreal Cognitive Assessment (MoCA).

Composite score, Mean difference, IV [95%CI]:

B: 0.04 [-0.02, 0.10]

C: 0.02 [-0.01, 0.05]

G: 0.08 [-0.03, 0.19] group 1 (multidomain plus PUFA)

0.09 [-0.00, 0.19] group 2 (multidomain plus placebo)

Pooled effect random effects model: 0.03 [0.01, 0.06] favouring multi-domain.

Heterogeneity (I²): 0%

MMSE Mean difference, IV [95%CI]:

A: 0.04 [-0.33, 0.41] group 1 (aerobic exercise plus diet)

-0.05 [-0.43, 0.33] group 2 (resistance training plus diet)

B: -0.03 [-0.35, 0.29]

C: -0.05 [-0.19, 0.09]

D: 0.32 [-0.76, 1.40] group 1 (bimonthly telephonic care)

0.14 [-0.86, 1.14] group 2 (monthly telephonic care)

0.38 [-0.66, 1.42] group 3 (bimonthly health worker-initiated visit care)

1.04 [0.00, 2.08] group 4 (bimonthly health worker-initiated visit care plus rewards)

G: 0.12 [-0.15, 0.38] group 1 (multidomain plus PUFA)

0.12 [-0.15, 0.38] group 2 (multidomain plus placebo)

H: 0.01 [-0.11, 0.13]

Pooled effect random effects model: 0.02 [-0.06, 0.09] favouring multi-domain.

Heterogeneity (I²): 0%

F was not included in the MMSE meta-analysis as it used a modified version of the MMSE (3MSE, 100-point scale). F also did not observe a difference in cognitive decline measured with the 3MSE (MD 0.06 point [95%CI -0.31 to 0.44]).

MoCA Mean difference, IV [95%CI]:

E: 1.10 [0.99, 1.21] group 1 (lifestyle plus medication)

0.30 [0.26, 0.34] group 2 (lifestyle only)

I: 1.03 [-0.19, 2.25]

Pooled effect random effects model: 0.76 [0.05, 1.46] favouring multi-domain intervention.

Heterogeneity (I²): 99%

Incidence of dementia or any subtype of dementia

Incidence of dementia, as assessed by an independent outcome committee.

RR [95%CI]:

F: 0.98 [0.61, 1.57]

H: 0.93 [0.73, 1.20]

Pooled effect random effects model:

0.94 [95%CI 0.76 to 1.18] favouring multi-domain

Heterogeneity (I²): 0%

Incidence of mild cognitive impairment (MCI)

Incidence of mild cognitive impairment, as assessed by an independent outcome committee

RR (M-H, Random, 95%CI)

F: 0.97 [0.76, 1.23] favouring multi-domain

Quality of life

Defined by mean change on the RAND-36 (SF-36) score in intervention/control.

B: general health 1.5/-1.6, P < 0.001

physical activity -2.3/-4.0,

role physical -17/-4.7,

role mental -0.1/-1.7,

vitality -0.3/-0.9, -0.6/-0.9,

social function -0.8/-1.2,

bodily pain -2.4/-2.1,

Activities of daily living (ADL)

Defined as measured by any validated questionnaire, including multiple ADL related questionnaires (B), the late life function and disability instrument (C), Alzheimer’s Disease Cooperative Study Activities of Daily Living Prevention Instrument (ADCSADLPI) (G), the Amsterdam Linear Disability Scale (H) and Instrumental Activities of Daily living (I). For most scales a lower score indicates more disability, but for study B the point estimate of the trial was multiplied with -1 to align the direction of the scales.

Standardized mean difference, IV [95%CI]:

Effect measure: RR, RD, mean difference [95%CI]:

B: 0.00 [-0.11, 0.11]

C: -0.05 [-0.15, 0.05]

G: 0.08 [-0.10, 0.26] group 1 (multidomain plus PUFA)

0.07 [-0.11, 0.25] group 2 (multidomain plus placebo)

H: -0.01 [-0.09, 0.07]

I: 0.00 [-0.12, 0.12]

Pooled effect random effects model: -0.00 [-0.05, 0.04]

Heterogeneity (I²): 0%

The authors found no convincing evidence that multi-domain interventions can prevent incident dementia, incident MCI, quality of life or activities of daily living.

There was a small effect of multi-domain interventions on cognitive functioning (assessed by NTB and MoCA). Although no evidence for an effect was found when cognitive decline was assessed with the MMSE, the most widely used outcome measure. The effect assessed by NTB was only observed in studies with cognitive training, thus a potential learning effect could be present in the cognitive testing in the intervention group.

Overall, the quality of included studies was low. All studies had no blinding to allocation of participants and personnel to the (lifestyle) interventions. Moreover, Li 2018 was at high risk of selection bias due to randomization per working district. Four studies were at high risk for incomplete outcome data due to various reasons: >25% loss to follow-up (Chen 2020 and Espeland 2017), unclear randomization that may have affected drop-out (Lee 2014) and differential drop-out for MMSE in the intervention (86%) versus control (91%) (Richard 2019). High risk for selective reporting in Komulainen 2010, because the authors did not report on individual tests nor total scores for a score that they obtained. Selective reporting could not be assessed for studies that either did not publish a protocol (Lee 2014, Li 2018, Chen 2020), register in a trial register (Lee 2014, Li 2018) or only identified primary outcomes in their trial registration (Chen 2020). Finally two studies were at high risk of bias due to their statistical analysis. Li 2018 reported one-way ANOVA for two group comparison. Lee 2014 used cluster as unit of analysis instead of taking a cluster effect into account in the analysis, used single instead of multiple imputation and was unclear about the follow-up MMSE for the control group which hampered effect size interpretation.

Level of evidence: GRADE

VERY LOW