Table 1 Systematic reviews on the associations between alcohol use and the incidence of cognitive impairment or dementia, including dose-response studies

Hersi et al. [23]

Onset and progression of AD

AD had to be determined using neuropathologic examination or by diagnosis with standardized instruments.

Light to moderate alcohol use was associated with a decreased risk of AD onset. Heavy and daily use was associated with an increased risk.

No evidence was found of the association between alcohol use and progression of AD.

Qualitative review based on seven moderate-quality systematic reviews on the association between alcohol use and risk of AD and two primary studies (see Additional file 1 for details). No age restriction as part of the inclusion/exclusion criteria.

Xu et al. [24]

All-cause D; AD; VD were analyzed separately.

No measurement specification for outcome variables

Non-linear association between alcohol use and all-cause dementia risk; the alcohol dose associated with a lower risk of dementia was confined to at most 12.5 g/day, and the risk hit bottom (RR 0.9) at roughly 6 g/day. Risk was elevated (about 10%) when the dose surpassed 23 drinks/week or 38 g/day.

MA based on 10 prospective (longitudinal) studies for all-cause dementia (Additional file 1). Relatively detailed dose-response relationships based on different dimensions (average level and frequency). No age restriction as part of the inclusion/exclusion criteria.

Cao et al. [25]

All-cause D; AD; mild cognitive impairment

AD had to be determined using clearly stated diagnostic criteria or identified through diagnostic codes with additional confirmation

Alcohol use (dichotomous) was not significantly related to the incidence of dementia as defined (see endpoint): RR 0.74, 95% CI 0.55–1.01.

MA based on eight prospective (longitudinal) studies (Additional file 1). Number of studies seems low given their inclusion/exclusion criteria. No age restriction as part of the inclusion/exclusion criteria.

LaFortune et al. [26]

All-cause D; cognitive impairment

No measurement specification for outcome variables

Consistent evidence demonstrating an association between alcohol abstinence and/or heavy drinking and cognitive impairment; compared with moderate alcohol intake, alcohol abstinence was associated with a higher risk of poor executive functioning and poor memory; one study reported no association with impairment of cognition or dementia.

Qualitative assessment of five longitudinal studies (see Additional file 1 for studies included). Minimum follow-up of 5 years. The populations covered included (1) mid-life adults (aged 40–64 years) and (2) adults aged 39 and younger in populations at a higher risk of health inequalities (people from disadvantaged and minority groups).

Cooper et al. [27]

All-cause D; AD

Mild cognitive impairment was identified from neuropsychological tests, in the absence of dementia or significant functional impairment.

Grade 2 evidence that heavy alcohol use was associated with conversion from any-type mild cognitive impairment to dementia and inconsistent evidence of whether light to moderate alcohol use predicts the risk of dementia.

Qualitative assessment of the role of alcohol in the course from mild cognitive impairment to D. Samples were based on either general populations (four studies) or clinical samples (three studies; all studies in the Additional file 1). No age restriction as part of the inclusion/exclusion criteria.

Ilomäki et al. [28]

All-cause D, AD, indicators of cognitive decline

No measurement specification for outcome variables

Light to moderate drinking was associated with the risk of AD (RR 0.72; 95% CI 0.61–0.86) and dementia (RR 0.74; 95% CI 0.61–0.91), whereas heavy to excessive drinking was not associated with either AD or D (RR 0.92; 95% CI 0.59–1.45; RR 1.04; 95% CI 0.69–1.56, respectively).

Systematic review of systematic reviews. Included only three systematic reviews [20, 37, 44] based on longitudinal studies and quality criteria (Additional file 1). No age restriction as part of the inclusion/exclusion criteria.

Xu et al. [30]

AD

No measurement specification for outcome variables except that it had to be AD and not unspecified dementia

Alcohol use, especially use of 1–3 drinks per day (RR 0.61 95% CI 0.54–0.68), but not heavier drinking, and alcohol use disorders showed a protective association (grade 1).

Included (longitudinal) cohort and retrospective case-control studies and conducted MA for different dimensions (ever versus never, 1–3 drinks per day versus never, and heavier use versus light or no use). Meta-analyses for 1–3 drinks per day based on five studies (Additional file 1). No age restriction as part of the inclusion/exclusion criteria. Excluded non-significant relationships.

Alzheimer’s Disease International [31]

Incident all-cause D, AD, and VD

No measurement specification for outcome variables except that it had to be all-cause D, AD, and VD had to be directly measured

Moderate drinkers (1–14 units for women and 1–21 for men) were at lower risk of AD (RR = 0.62, 95% CI 0.54–0.69) or any dementia (RR = 0.54, 95% CI 0.42–0.67) compared with abstainers. No significant difference between heavy drinkers and abstainers for either AD or D.

Based on Anstey et al. [20], Neafsey and Collins [40], and Peters et al. [44], plus an independent search for longitudinal studies which had measured and excluded cases with dementia at baseline (16 studies; Additional file 1). No age restriction as part of the inclusion/exclusion criteria.

Beydoun et al. [32]

Incident and prevalent D, AD, cognitive function, and cognitive decline

Well-defined criteria for all outcomes

Moderate alcohol use was associated with better cognitive function, a lower rate of cognitive decline, and a lower incidence/prevalence of dementia in the majority of cohort and cross-sectional studies (17 out of 30) included. Dose response was linear and curvilinear (J/U-shaped).

Based on 18 (longitudinal) cohort studies and 12 cross-sectional studies with qualitative assessment of results (Additional file 1) mainly with decline in cognitive function as the outcome variable. No age restriction as part of the inclusion/exclusion criteria.

Di Marco et al. [33]

Dementia or any subtype thereof

No measurement specification for outcome variables

Most studies found an association between mild to moderate alcohol use and a lower incidence of dementia.

Based on 13 cohort studies, population restricted to 35+ years old, which were free of dementia at baseline. Qualitative assessment of outcomes (Additional file 1).

Pei et al. [34]

Prevalence or incidence of any type of D

Diagnosis of dementia and types of dementia had to be based on internationally recognized criteria

Based on one study each, light to moderate alcohol use (<20 g pure alcohol per day for men and <16 g for women) was associated with a lower risk of dementia compared with those not drinking alcohol (OR 0.5; 95% CI 0.3–0.8). Daily alcohol use was associated with an increased risk of AD (OR 1.7; 95% CI 1.1–2.8).

Review was restricted to Chinese general populations, and only two studies were included with alcohol use as a risk factor (Additional file 1). Age restriction was for populations 60 years and older.

Piazza-Gardner et al. [37]

AD

No measurement specification for outcome variables except that it had to be AD and not unspecified dementia (exact definition of each underlying study was given)

No clear outcome: seven studies found an association between alcohol use and a decreased risk of AD, three studies found an association with an increased risk of AD (especially for heavier drinking), and nine studies reported no association between alcohol use and AD.

Review based on 19 studies of various designs (Additional file 1). Studies on single beverages or on symptoms only were excluded. No age restriction as part of the inclusion/exclusion criteria.

Neafsey and Collins [40]

Impaired cognition; all-cause D

No measurement specification for outcome variables

Between 1977 and 1997, mainly the associations of light to moderate alcohol use with cognitive function and impairment in young to middle-aged (18–50 years old) subjects were examined. Initial studies indicated an association of alcohol use with impaired cognition, but most later studies failed to confirm this, instead finding no difference in cognition between drinkers and non-drinkers. After 1998, mainly 55+-year-old subjects were assessed, overwhelmingly finding an association of moderate drinking with a reduced risk of dementia or cognitive impairment (RR 0.77; 95% CI 0.73–0.80) compared with non-drinkers. Heavy use (> 3–4 drinks/day) was associated with a higher risk of cognitive impairment/dementia.

Qualitative analyses of studies without a quantitative indicator of risk (OR; RR; hazard ratio) based on 69 articles. MA for quantitative studies (mainly after 1998) based on 74 studies (see Additional file 1 for listing). No age restriction as part of the inclusion/exclusion criteria; however, the results were interpreted based on age (see left column).

Lee et al. [41]

All-cause D, cognitive function, decline, or impairment

Measurement of outcomes was part of a quality index

Moderate alcohol use was associated with a lower risk of cognitive decline and dementia (compared with non-drinkers), but frequent and heavier use was associated with higher risks of dementia and cognitive impairment.

Prospective (longitudinal) studies in people older than 65 years of age were targeted, but others were included as well; qualitative analyses based on eight studies (Additional file 1).

Anstey et al. [20]

All-cause D, AD, VD, cognitive decline

No measurement specification for outcome variables

Pooled RRs of AD, VD, and all-cause D for light to moderate alcohol use compared with no use were 0.72 (95% CI 0.61–0.86), 0.75 (95% CI 0.57–0.98), and 0.74 (95% CI 0.61–0.91), respectively. Heavy use was not associated with an increased risk of any of the dementia categories.

MAs based on 15 prospective (longitudinal) studies (Additional file 1) with inclusion criteria that included ascertainment where at baselines there were no dementias, or some indication of cognitive status. No age restriction as part of the inclusion/exclusion criteria.

Purnell et al. [42]

Incident AD

Formal clinical assessment procedure and use of standardized definitions for AD

Four out of five studies found no association between alcohol use and the incidence of AD. One study reported an association with a decreased risk of AD, in part in interaction with apolipoprotein E ε4.

Qualitative assessment based on five prospective (longitudinal) studies (six articles; Additional file 1) limited to certain clinical assessment instruments of AD. No age restriction as part of the inclusion/exclusion criteria.

Peters et al. [44]

Incident all-cause D; AD; VD; cognitive decline

No measurement specification for outcome variables

MAs suggest that small amounts of alcohol were associated with a lower risk of dementia (RR 0.63; 95% CI 0.53–0.75) and AD (RR 0.57; 95% CI 0.44–0.74), but not of vascular dementia (RR 0.82; 95% CI 0.50–1.35) or cognitive decline (RR 0.89; 95% CI 0.67–1.17).

MAs based on 23 longitudinal studies (20 epidemiological cohort, three retrospective matched case-control studies nested in a cohort, reported in 26 publications; Additional file 1). Several MAs by type of dementia, sex, and various drinking measures. Only studies with subjects aged 65 years and older were included.

Patterson et al. [45]

Incident all-cause D; AD; VD

“Standardized criteria”

In the two studies identified, moderate wine consumption was associated with a reduced risk of all-cause D and AD.

Qualitative statement based on two (longitudinal) cohort studies in populations broadly similar to the Canadian population. No age restriction as part of the inclusion/exclusion criteria.

Weih et al. [46]

AD

No measurement specification for outcome variables

Most studies showed that heavy alcohol use increased dementia risk and moderate alcohol intake could reduce dementia (RR 0.55; low evidence level).

Based on seven prospective (longitudinal) studies (Additional file 1). No age restriction as part of the inclusion/exclusion criteria.

Reid et al. [47]

All-cause D, cognitive impairment, cognition disorders

No measurement specification for outcome variables

Ten studies found an increased risk of cognitive impairment associated with either a history of alcohol abuse, heavy use, or an average weekly consumption of more than 10 drinks when compared with individuals without a history of alcohol abuse or heavy drinking or compared with non-drinkers; 21 studies found no relationship between cognitive impairment and various alcohol measures. One study reported that consuming 2–5 drinks per day was associated with improved cognitive function in older women but not in men (compared with abstention).

Qualitative assessment based on 32 studies (Additional file 1) in populations older than 60 years of age. Special attention to alcohol measures, but no attempt to quantitatively summarize the findings as the measures were too different to quantitatively pool.