Table 2.
Overview Over Epidemiological Studies on the Link Between Alcohol Consumption and Dementia with Cohorts of Over 250 Participants Since 1994 Until the 1st September 2019, Listed on PubMed. Studies Which Do Not Diagnose Dementia with Standardized Protocol Including Neuropsychological Assessment, MRI (Magnetic Resonance Imaging) and Examination by a Psychiatrist and/or Neurologist are Not Mentioned. No Gender-Specific Studies or Beverage-Specific Studies are Mentioned. Only Studies Published in English are Mentioned. Multiple Analyses of Same Cohorts Were Excluded. The Effect Summary in Bold Letters Refers to the Effect of Low to Medium Dose Alcohol Consumption as Defined in the Study, as Definitions Vary. AD (Alzheimer’s Disease), HR (Hazard Ratio), MCI (Mild Cognitive Impairment), MMSE (Mini Mental State Examination), OR (Odds Ratio)
| Cohort Studies | ||||
|---|---|---|---|---|
| Author, Year | Participants (Dementia Cases) | Cohort Description, | Follow-up | Finding |
| Yoshitake et al, 199555 | 828 (103) | Aged over 65, non-demented at inception, Kyushu, Japan | 7 years, continuous assessment, 99.8% follow-up |
No effect: alcohol consumption was not significantly associated with AD Limitations: vague measurement of alcohol intake (current consumption: yes/no) |
| Broe et al, 199860 | 327 (47) | Aged over 75, Sydney, Australia | 3 years, 2 assessments |
No effect: alcohol consumption was not significantly associated with AD Limitations: very aged cohort (mean age 83.4 years), low power to detect effects |
| Huang et al, 200261 | 402 (66) | Aged over 75, one district of Stockholm, Sweden | 7 years, 3 assessments, 91% follow-up |
Protective effect: Lower risk of AD in patients consuming between 1–14 drinks per week (women) or 1–21 drinks per week (men), OR 0.5. Limitations: Participants had already low MMSE scores at study inception (<25). Sample was predominantly female (81.8%). When patients with MMSE>24 were included, the result was not significant. |
| Ruitenberg et al, 200210 | 7983 (197) | Aged 65 or older, residents of a Rotterdam suburb, Netherlands | 7 years average,4 assessments, 99.7% follow-up |
Protective effect: Lower risk of all-cause dementia and vascular dementia in individuals consuming 1–3 drinks/day Limitations: standard drink was not defined in terms of grams |
| Lindsay et al, 200262 | 4088 (194) | Aged over 65, Nationwide Canadian cohort | 5 years, 2 assessments |
Protective effect: Regular alcohol consumption was associated with lower risk of AD (OR 0.68) Limitations: alcohol consumption was categorized as “regular consumption” or “no regular consumption” |
| Truelsen et al, 200263 | 1709 (83) | Aged over 65 at inception, residents of Copenhagen, Denmark, | Interval of 15 years in between alcohol assessment and dementia assessment | No effect: the risk of all-cause-dementia was elevated in the group of heavy drinkers (15–21 drinks per week, OR 2.26), effect disappears when controlled for co-variates (age, sex, years of education, history of stroke, income, cohabitation status, smoking, systolic blood pressure) |
| Luchsinger et. al, 200464 | 980 (260) | Aged over 65 at inception, residents of New York City, USA | 4.1 years (mean), annual assessment |
No effect: neither light to moderate (1–3 drinks/day) nor heavy alcohol consumption (>3 drinks/day) was associated with risk of AD or all-cause-dementia Limitations: alcohol intake was measured on one occasion |
| Anttila et al, 200465 | 1018 (48) | Aged 42–56 at inception, inhabitants of two regions in Finland (Kuopio and Joensuu) | 23 years (mean), 2 assessments |
Protective effect: frequent drinkers (more than once a month) and abstainers (more than once a month) had higher OR for all-cause-dementia than infrequent drinkers (less than once a month) Limitations: very broad category of frequent drinkers ApoE4 enhanced the effect of frequent drinking (combined OR 7.07) |
| Ogunniyi et al, 200666 | 2480 (187) | Aged 65 or older; two cohorts from Indianapolis, USA (only African Americans) and Yoruba, Nigeria |
5 years, 3 assessments |
Protective effect: Alcohol consumption was associated with lower incidence of AD in both cohorts, but with lower OR (0.49) and significance only in the cohort from Indianapolis Limitations: low percentage of alcohol consuming participants, even lower in the Yoruba cohort |
| Deng et al 200667 | 2632 (121) | Aged 60 or older cohorts from six departments in China, |
2 years, 2 assessments |
Protective effect: Lower risk of AD in individuals consuming 1–21 units of alcohol (OR: 0.63) Limitations: only 84 AD cases |
| Langballe et al, 201556 | 40,435 (1084) | Aged 38–81 years at inception, cohort of an entire region of Northern Norway, | 27 years, 2 assessments, 98.2% |
Risk factor: no effect of low-dose alcohol consumption (less than five times during the last 14 days) on all-cause dementia, HR 14 for consumption of five or more times during the last 14 days for all-cause dementia. Limitations: no standardization of dementia diagnosis, no estimation of total alcohol dose |
| Heffernan et al 201657 | 821 (48) | Aged between 70–90 at inception, cohort of community dwelling participants from Sydney, Australia | 4 years, 2 assessments |
No effect on any form of dementia Limitations: short follow up of four years might underestimate long-term effects |
| Paganini-Hill et al, 201658 | 547 (268) | Participants of the 90+ study, aged over 90 at inception from California, USA | 26 years, 3 assessments, 95% follow-up |
No effect on any form of dementia Limitations: data partially obtained via telephone and relatives |
| Xue et al, 201759 | 437 (MCI), 106 (AD) | Aged over 65 at inception in 2010, participants with MCI, community-dwelling from Taiyuan, China | 5 years, up to 10 assessments | Study on the risk of transition from MCI to AD Risk factor: light to moderate drinkers (<2 drinks per day) showed higher HR for transition from MCI to all-cause dementia Limitations: assessment of only MCI patients, small sample, broad definition of alcohol dose |
| Sabia et al, 201870 | 9087 (397) | aged 35–55 at inception; cohort from London civil service employees |
23.2 years (mean), 5 assessments | Protective effect: risk for developing all-cause-dementia lowest in the individuals consuming between 1–14 drinks per week compared to abstainers and heavier drinkers |
| Schwarzinger et al 20189 | 31 624 156 (1 109 3434) |
Cohort of all discharged patients in metropolitan France aged over 20 years at inception | 5 years, continuous assessment |
Risk factor: Risk of dementia elevated in patients diagnosed with alcohol use disorder, hazard ratio 3.3 for all type dementia Limitations: use of discharge diagnoses, no standardized assessment of neither alcohol use nor dementia diagnosis |
| Case Control Studies | ||||
| Author, Year | Participants (dementia cases) | Cohort description | Finding | |
| Bachmann et al 200368 | 2779 (844) | Mean age 70 years, American patients with dementia and non-demented siblings |
Protective effect: white participants consuming moderate (0.25–2 drinks/day) to high (>2 drinks per day) levels of alcohol showed a lower OR (0.88) of all-cause dementia compared to abstainers. Limitations: Number of African American participants was low (343) |
|
| Mukamal et al 200369 | 746 (373) | Aged over 65 years at inception, cohort subsample of the cardiovascular health study, USA |
Protective effect: the subgroups drinking less than 1drink (OR 0.65), 1–6 drinks (OR 0.46), 7–13 drinks (0,69) had a lower risk of all-cause dementia than abstainers Thorough assessment of covariates |
|
| Twin Studies | ||||
| Author, Year | Participants (dementia cases) | Cohort description | Follow-Up | Finding |
| Handing et al, 201571 | 12326 (1958) | Swedish twin cohort born between 1907 and 1925, aged between under 65 at inception | Up to 43 years |
No effect: of low (1–5g/day) and moderate (6–12g/day) alcohol consumption in comparison to abstainers, elevated risk for heavy drinkers (>12g/day) Limitations: Diagnosis of dementia assessed only through death registry +Very long follow-up |
| Järvenpää et al, 200590 | 826 (103) | Finnish twin cohort, aged over 65 at assessment for dementia in 1999–2001 | Up to 25 years |
No effect: low-dose alcohol consumption was not associated with higher or lower risk Binge drinking or passing out had a very high OR for developing all-cause-dementia 25 years later (4.2/11.8) Limitations: Dementia was assessed by interview only, no neuropsychological assessment |
| Cross-sectional studies | ||||
| Author, Year | Participants (dementia cases) | Cohort description | Finding | |
| Harwood et al 200972 | 685 (685) | Cohort of patients of a memory clinics in Miami, USA |
No effect: only heavy drinking (>2 drinks per day) was associated with an earlier onset of AD compared to abstainers Limitations: age of onset was evaluated by interviews with caregivers |
|
| Toure et al, 201291 | 507 (47) | Cohort of patients of Social and medical center in Dakar, Senegal, aged over 65 |
No effect: alcohol had no effect on dementia risk Limitations: low percentage of alcohol consumption and dementia cases |
|
| Pilleron et al, 201592 | 1772 (135) | Two cohorts from the Central African Republic and Republic of Congo, aged over 65 at inception |
Protective effect: participants consuming any amount of alcohol had a lower risk of all-cause-dementia (OR 0.34) Limitations: generally low levels of alcohol consumption and dementia in the whole cohort |
|
| Radford et al, 201893 | 381 (45) | Cohort of Aboriginal Australians, aged over 65 years at inception |
Protective effect: low-risk alcohol consumption (audit-c) had lower OR for all-cause dementia than abstention Limitations: Aboriginal Australians differ in terms of rates of illiteracy and education level from the general Australian population |
|
| Meta-Analyses | ||||
| Author, Year | Participants (dementia cases) | Cohort description | Finding | |
| Anstey et al, 200994 | 14 studies included in the meta-analysis AD: 14646 Vascular D.: 10225 ( All-cause-dementia: 11875 |
Different cohorts: Australia, Canada, China, France, Germany, Japan, Netherlands, Nigeria, Sweden, UK, USA |
Protective effect: light to moderate drinkers OR 0.72 for AD and 0.74 for vascular dementia Limitations: dosage definitions vary in different studies included |
|
| Peters et al, 200876 | 23 studies included in the meta-analysis; Total number of included participants not given, in all studies showing significant results: 29,946 participant | Different cohorts: Australia, Canada, France, Finland, Germany, Japan, Netherlands, Sweden, UK, USA, |
Protective effect: participants consuming small amounts of alcohol show lower risk of AD (RR 0.57) and all-cause dementia (0.63) but not for vascular dementia (0.82) Limitations: dosage definitions vary in different studies included |
|
| Tyas et al, 201095 | 3 only Canadian case-control-studies included | 3cohorts from Canada |
No effect Limitations: high inconsistency, use of secondary data |
|
| Wei-Xu et al, 201779 | 11 studies with 73,330 participants and 4586 cases of all-cause dementia | 11 cohorts from China, Denmark, Finland, Netherlands, Norway, Sweden, USA | Dose-response analysis with the most beneficial dose being 6g alcohol/day or 2x/week Protective effect: consumption of less or equal of 2x/week, 7.5 drinks/week or 12.5g/day were associated with lower risk of all-cause-dementia Limitations: dosage definitions vary in different studies included |
|