Source | Study design | Subjects (n) | FU | Exposure | Outcome | Results (95% CI) |
---|---|---|---|---|---|---|
Clinical dementia | ||||||
Karp et al27; Swedish Kungsholmen Project/Sweden | Cohort | General population (931) | 2 FU at 3 year intervals | Complexity of data/people/things (Swedish census) | Dementia* | Cox proportional hazards model: Complexity/data: RR=0.85 (0.77 to 0.95), p=0.003; Complexity/people: RR=0.88 (0.80 to 0.97), p=0.011; Complexity/things: RR=1.04 (0.97 to 1.11). Subgroup analysis (education and complexity/data): low/low: 1.00; low/high: 0.52 (0.29 to 0.95); high/low: 0.55 (0.41 to 0.74); high/high: 0.48 (0.32 to 0.72).Subgroup analysis (education and complexity/people): low/low: 1.00; low/high: 0.69 (0.28 to 1.69); high/low: 0.57 (0.44 to 0.75); high/high: 0.48 (0.29 to 0.80) |
Kröger et al28; Canadian study of Health and Aging/Canada | Cohort | General population (3557) | 2 FU at 5 year intervals | Complexity with data/people/things (DOT) | Dementia*, MMSE | Cox proportional hazards model: Complexity/data: HR=1.14 (0.79 to 1.64), >23 years in job HR=1.77 (1.02 to 3.08); Complexity /people: HR=0.66 (0.44 to 0.98), >23 years in job HR=0.36 (0.20 to 0.66); Complexity/things: HR=0.72 (0.52 to 0.99), >23 years in job HR=0.45 (0.28 to 0.73) |
Andel et al30; Swedish twin registry/Sweden | Case–control | Twins (225 cases; 9854 controls) | Complexity with data/people/things (DOT) | Dementia*, clinical AD diagnosis | Logistic regression: Complexity/data: OR=1.00 (0.91 to 1.09), analysis of twin pairs-dementia OR=0.77 (0.43 to 1.38), analysis of twin pairs-AD OR=0.17 (0.15 to 0.57); Complexity/people: OR=0.86 (0.76 to 0.98), analysis of twin pairs-dementia OR=0.47 (0.25 to 0.88), analysis of twin pairs-AD OR=0.05 (0.01 to 0.35); Complexity/things: OR=1.07 (1.00 to 1.14), analysis of twin pairs-dementia OR=1.50 (0.83 to 2.73), analysis of twin pairs-AD OR=1.11 (0.45 to 2.72) | |
Potter et al29; Duke Twins Study of Memory in Aging/USA | Case–control | WWII veteran male twins (425 cases; 6075 controls) | Complexity with data/people/things (DOT) | Dementia* | Cox proportional hazards model: Complexity/data: HR=1.115 (1.015 to 1.226), p=0.019; twin pairs HR=1.158 (0.927 to 1.445), p=0.196; twin pairs discordant for dementia >6 years HR=1.41 (1.02 to 1.94), p=0.037; Complexity/people: HR=1.055 (0.946 to 1.177), p=0.343; twin pairs HR=1.034 (0.795 to 1.344), p=0.802; twin pairs discordant for dementia >6 years HR=1.12 (0.814 to 1.535), p=0.491; Complexity/things: HR=0.996 (0.903 to 1.099), p=0.941; twin pairs HR=1.058 (0.849 to 1.137), p=0.620; twin pairs discordant for dementia >6 years HR=1.10 (0.820 to 1.464), p=0.538 | |
Smyth et al32; USA | Case–control | General population (122 cases; 235 controls) | Mental and social demands (DOT) | Clinical AD diagnosis | ANOVA pairwise comparison, significance at p<0.01: mental demands p=0.007; social demands p=0.33 | |
Seidler et al34; Germany | Case–control | General population (195 cases; 122 controls) | Social climate; challenge; control possibilities; work load; perceived risks for error; social demands; supervisor support; working-time arrangement (FINJEM) | Dementia*†, (MMSE) | Logistic regression: social climate OR=1.3 (0.8 to 2.2), p for trend=0.42; challenge at work OR=0.5 (0.3 to 0.9), p for trend=0.01; control possibilities: OR=0.5 (0.3 to 1.1), p for trend=0.02; work load OR=2.0 (0.9 to 4.6), p for trend=0.24; perceived risks for error OR=1.1 (0.8 to 5.4), p for trend=0.001; social demands OR=0.2 (0.04 to 1.3), p for trend=0.05; supervisor support OR=0.7 (0.4 to 1.4), p for trend=0.23); working time arrangements OR 0.9 (0.3 to 2.6, p for trend=0.13 | |
Crowe et al33; Swedish twin registry/Sweden | Case–control | Twins (144 cases; 1905 controls) | Work-related stress | Dementia* | Logistic regression by pair: work-related stress via job demands OR=1.02 (0.96 to 1.08), p=0.527; adjusted for reactivity to stress OR=1.07 (0.96 to 1.19); analysis of twin pairs t=0.62, p=0.54 | |
Andel et al35; Swedish twin registry/Sweden | Case–control | Twins (257 cases; 9849 controls) | Job control; job demands, social support, job strain (ISCO) | Dementia* | Generalised estimating equation modelling: low job control OR=1.17 (1.04 to 1.31); high job demands OR=1.00 (0.90 to 1.11); low social support OR=1.15 (1.03 to 1.28); great job strain OR=1.10 (0.98 to 1.25) | |
Wang et al36; Swedish Kungsholmen Project/Sweden | Cohort | General population (913) | 2 FU at 3 year intervals | Job control; job demands, job strain (Nordic occupation classification) | Dementia*, clinical AD diagnosis | Cox proportional hazards models: low job control HR=1.9 (1.1 to 3.2), for AD HR=2.3 (1.2 to 4.3); low job demands HR=1.3 (0.8 to 1.9), for AD HR=1.3 (0.8 to 2.1); high job strain HR=1.7 (1.02 to 2.8); for AD HR=1.9 (1.0 to 3.4) |
Global cognitive functioning | ||||||
Andel et al8; SWEOLD/Sweden | Cohort | General population (911) | 1 FU after 24 or 34 years | Job control, job demands, job strain | MMSE score | Binary logistic regression: iob control (self-rated): OR=0.71, p<0.001; cut-off MMSE<7 OR=1.17, p=0.329; job control (DOT): OR=0.63, p=<0.001; cut-off MMSE<7 OR=1.86, p=0.002; job demands (self-rated): OR=1.03, p=0.737; cut-off MMSE<7 OR=1.07, p=0.643; job demands (DOT): OR=1.27, p=0.053; cut-off MMSE <7 OR=0.69, p=0.064; job strain (self-rated): OR=0.97, p=0.894; cut-off MMSE <7 OR=1.27, p=0.524; job strain (DOT): OR=0.71, p=0.063; cut-off MMSE <7 OR=1.51, p=0.129 |
Andel et al37; SWEOLD/Sweden | Cohort | General population (386) | 3 FU at intervals of 6, 7 and 10 years | Complexity with data/people/things (Swedish census) | MMSE score | Logistic regression: Complexity/data: b=0.15, SE=0.07; cut-off MMSE <7 OR=0.84 (0.66 to 1.08); Complexity/people: b=0.13, SE=0.08; cut-off MMSE <7 OR=0.71 (0.48 to 1.05); Complexity/things: b=0.00, SE=0.05; cut-off MMSE <7 OR=1.08 (0.89 to 1.30) |
Potter et al38; USA | Case–control | WWII veteran male twins (3880) | 2 FU at 3–4 year intervals | Intellectual demands; human interaction/communication; visual attention (DOT) | TIC-M | Least-squares regression model: intellectual demands: b=0.074, p=0.011; only monozygotic b=-0.053, p=0.194; human interaction/communication: b=0.004, p=0.836; only monozygotic b=0.008, p=0.769; visual attention: b=−0.036, p=0.023; only monozygotic b=0.013, p=0.551 |
Marquie et al39; VISAT cohort study/France | Cohort | Working population (3123) | 2 FU at 5 year intervals | Cognitive stimulation at work | Composite score of cognitive functioning (including word-list learning/recall; digit symbol task; selective attention; delayed retrieval) | Mixed model analysis: cognitive stimulation at work*follow-up: F=16.18, b=0.059/=0.099, T=3.89/5.3, p<0.01 |
Bosma et al40; Netherlands | Cohort | General population (630) | 1 FU at 3 year intervals | Mental work demands (coded by Dutch job experts) | Poorest 1/10th centile of the total score of cognitive functioning (Stroop Color-Word Test; Verbal Learning Test; Letter Digit Coding Test; Word Fluency Test) | Logistic regression: mental demands OR= 0.83 (0.68 to 1.02), p<0.1; strong concentration OR=0.80 (0.64 to 0.99), p<0.05; great precision OR=0.78 (0.64 to 0.95), p<0.05; time pressure OR=0.70 (0.54 to 0.9), p<0.01; task complexity OR=0.90 (0.76 to 1.06), p>0.1; composite score OR=0.79 (0.65 to 0.96), p<0.05 |
Specific cognitive abilities | ||||||
Virtanen et al41; Whitehall II study/UK | Cohort | London civil servants (2214) | 7 FU at 2–3 year intervals | Working hours | Verbal memory free recall; Alice Heim 4-I; Mill Hill vocabulary test; phonemic fluency S-words; semantic fluency | Multivariate analysis of covariance: memory p=0.118; reasoning p=0.010 (41–55 h) p=0.099; >55 h p=0.04; reasoning remains significant throughout stepwise confounder adjustments, test for linear trend p<0.04; vocabulary p=0.003 (41–55 h p=0.020; >55 h p=0.32; phonemic fluency p=0.088; semantic fluency p=0.430 |
Elovainio et al42; Whitehall II study/UK | Cohort | London civil servants (4531) | 7 FU at 2–3 year intervals | Organizational justice | verbal memory free recall; Alice Heim 4-I; Mill Hill vocabulary test; phonemic fluency S-words; semantic fluency | Multivariate regression analysis: memory p=0.008; reasoning p=0.002; vocabulary p<0.001; phonemic fluency p=0.011; semantic fluency p=0.031 |
Finkel et al43; Swedish Adoption/Twin Study of Aging/Sweden | Cohort | Twins (462) | 5 FU at 3 year intervals | Complexity with data/people/things (Swedish census) | SATSA cognitive test battery scores | Two-slope latent growth curve model centred at retirement age: Complexity/data: NS; Complexity/people: spatial performance p<0.05 for intercept, practice and slope; processing speed p<0.05 for intercept; memory NS; verbal NS; Complexity/things: NS |
*Dementia diagnosis based on DSM-III-R and NINCDS-ADRDA.
†Using MRI and CT scan.
AD, Alzheimer's disease; ANOVA, analysis of variance; Complexity/data, work complexity with data; Complexity/people, work complexity with people; Complexity/things, work complexity with things; DOT, 1970 US Census Dictionary of Occupational Titles; FINJEM, Finnish Job-Exposure Matrix; FU, follow-up; ISCO, International Standard Classification of Occupations; MMSE, Mini-Mental State Examination; NS, not significant; RR, risk ratio; SATSA, Swedish Adoption/Twin Study of Aging cognitive test battery; SE, standard estimate; SWEOLD, Swedish Panel Study of Living Conditions of the Oldest Old; TIC-M, telephone version of MMSE; VISAT, French Study on Aging, Health and Work; WWII, World War 2.