Search for a command to run...
Dementia prevalence changes over time, suggesting that modifiable factors affect dementia risk. Being unmarried is associated with a range of adverse health outcomes and may be a risk factor for dementia, due to life-course factors. We aimed to combine published evidence examining the effect of being married, widowed, divorced and single on risk of developing dementia and its subtypes, and the extent of modification of this effect by cultural context and study methodology. We searched EMBASE, PsycINFO and Medline from inception to 5th December 2016, hand-searched references and contacted experts in the field looking for relevant papers. Two researchers identified studies and rated their quality. We included observational research reporting age- and sex-adjusted association between marital status and dementia, in general older populations. Where studies did not report adjusted results, we contacted the authors to request these. We conducted random-effects meta-analyses of the effect of being widowed, divorced or single, compared to being married and secondary analyses for other hypotheses. We included 14 studies (8 cohort, 2 case-control, 4 cross-sectional) including 806,955 people. Compared to people who are married, single (pooled RR 1.42, 95% CI [1.07, 1.90]) and widowed (RR 1.25, 95% CI [1.05, 1.50]) people are at a higher risk of developing dementia. There was no increased risk for divorced people. Much of this association was explained by the effect of age, gender, education and physical health. In stratified secondary analyses, we did not find this effect modified by gender, cultural setting (Asian studies v others) or the participants’ mean year of birth. Studies which derived dementia diagnoses from routine clinical records found a smaller association than those which used a clinical examination. In the studies which examined risk of Alzheimer's (n=7) or vascular dementia (n=3), there was no significant association with marital status. There was no evidence of publication bias in the published studies. Being widowed and single elevate risk of dementia. These findings are affected by the method of dementia diagnosis ascertainment, but not by cultural context, time-period or gender. These associations may be explained by cognitive reserve and physical health.