Psychological Distress as a Risk Factor for Dementia Death

Current estimates suggest that neuropsychiatric disorders account for 28% of the global burden of disease.¹ While depression and anxiety (commonly referred to as psychological distress) have been shown to be a consequence of dementia, the converse is less clear. The possibility that psychological distress might be a risk factor for dementia has major public health implications. However longitudinal studies—which are best placed to examine this relationship—have, with some exceptions,^2,3 been small in scale (affecting study precision), excluded individuals younger than 65 years (limiting insights into the pre–older age origins of dementia), or have used clinical samples (reducing generalizability). Accordingly, we examined the role of psychological distress as a risk factor for and dementia death by pooling 10 large community-based cohort studies.

Participants were recruited from the Health Survey for England,⁴ an annual general population-based cross-sectional study (with a longitudinal component) representative of household-dwelling individuals in England. Results from 1994 through 2004 were pooled. Participants gave informed consent; ethical approval was obtained from the London Research Ethics Council.

Psychological distress was measured during a household visit using the 12-item General Health Questionnaire (GHQ-12), a widely used measure of psychological distress in population studies comprising items rating anxiety, depression, social dysfunction, and loss of confidence. Higher scores indicate greater distress. We used a cut off score of 4 or greater to denote psychological distress as validated against standardized psychiatric interviews.⁵ Dementia was identified from death certification and coded according to the International Classification of Diseases, Ninth Revision (ICD-9) codes 290.0 through 290.4 and 294.9 and International Statistical Classification of Diseases, 10th Revision (ICD-10) codes F01, F03, F09, and G30. Follow-up was until date of death or February 15, 2008, whichever came first.

We used Cox proportional hazards models to compute hazard ratios with accompanying 95% confidence intervals for GHQ-12 score in relation to dementia-related deaths. Study members scoring zero (no apparent distress) denoted the reference group. Models were adjusted for age, sex, occupational social class (OSC),⁶ parental OSC, age at leaving full-time education, current smoking (yes/no), alcohol consumption (units per week), and existing cardiovascular disease (CVD) (yes/no), and diabetes (yes/no). Statistical analyses were conducted using PASW statistics, version 18.0 (SPSS, Chicago, Illinois), and R for Max OS X, version R-2.13.0.

The initial sample included 85 261 adults (in 1996 the GHQ-12 was not used). After removing individuals who declined linkage to mortality records (n = 9325) and those with missing GHQ-12 data (n = 2865), the analytic sample comprised 73 071 individuals (54.8% women) with a mean (SD) age of 55.9 (14.3) years (range, 35-102 years). Data were missing for 1 or more variables in 21% (n = 15 355) of the sample. Individuals with missing data were more likely to be older, be female, belong to a manual OSC, leave school later, be a nonsmoker, drink alcohol moderately, and have CVD and diabetes.

Of the 10 170 deaths during follow-up, 455 had dementia coding. A higher GHQ-12 score was associated with increased risk of dementia death in an age-adjusted model (GHQ-12 score of 1-3: HR, 1.44 [95% CI, 1.17-1.78]; GHQ-12 score of 4-12: HR, 1.74 [95% CI, 1.36-2.22]; P value for trend, <.001). Adding all remaining covariates (sex, OSC, parental OSC, age at leaving full-time education, current smoking, alcohol consumption, and existing CVD and diabetes) led to some attenuation of effect but statistical significance at conventional levels was retained (GHQ-12 score of 1-3: HR, 1.27 [95% CI, 1.00-1.61]; GHQ-12 score of 4-12: HR, 1.56 [95% CI, 1.17-2.07]; P value for trend, .005). In the Figure we relate 7 categories of GHQ score to dementia death to provide more detailed insight into the shape of the relationship. There was evidence of a dose-response effect (P value for trend, .001). Excluding individuals with any missing data (sample n = 57 716; 361 dementia deaths) or dementia deaths within 5 years (sample n = 72 926; 310 dementia deaths)—the latter to explore reverse causality—did not affect our results.

We found an association between elevated psychological distress and an increased risk of dementia death in a large general population sample of apparently dementia-free adults, which remained after adjustment for age, sex, OSC, education, alcohol use, smoking, and existing CVD and diabetes. Cardiovascular risk factors have been linked with dementia,⁷ but the association found in our study remained after controlling for them, thus implicating other explanations for the gradient seen. One possibility is a toxic effect of hypercortisolemia in depression on the hippocampus.⁸ Further research is required to investigate whether appropriate treatment of depression reduces dementia risk.

Correspondence: Dr Russ, Alzheimer Scotland Dementia Research Centre, University of Edinburgh, 7 George Sq, Edinburgh EH8 9JZ, Scotland (tom.russ@nhs.net).

Author Contributions:Study concept and design: Russ, Hamer, and Batty. Acquisition of data: Hamer, Stamatakis, and Batty. Analysis and interpretation of data: Russ, Starr, and Batty. Drafting of the manuscript: Russ and Batty. Critical revision of the manuscript for important intellectual content: Russ, Hamer, and Stamatakis, Starr, and Batty. Statistical analysis: Russ and Batty. Obtained funding: Stamatakis and Batty. Study supervision: Hamer, Starr, and Batty.

Funding/Support: Dr Russ is supported by Alzheimer Scotland and he is employed in the NHS by the Scottish Dementia Clinical Research Network, which is funded by the Chief Scientist Office (part of the Scottish Government Health Directorates). Dr Russ and Starr are members of the Alzheimer Scotland Dementia Research Centre funded by Alzheimer Scotland. Drs Russ, Starr, and Batty are members of the University of Edinburgh Centre for Cognitive Ageing and Cognitive Epidemiology, part of the cross council Lifelong Health and Wellbeing Initiative (G0700704/84698). Funding from the BBSRC, EPSRC, ESRC, and MRC is gratefully acknowledged. GDB is a Wellcome Trust Fellow.