Common psychosocial stressors in middle-aged women related to longstanding distress and increased risk of Alzheimer's disease: a 38-year longitudinal population study.

OBJECTIVE: To study the relation among psychosocial stressors, long-standing distress and incidence of dementia, in a sample of women followed from midlife to late life.
DESIGN: Prospective longitudinal population study.
SETTING: The analyses originate from the prospective population study of women in Gothenburg, Sweden, a representative sample of women examined in 1968 (participation rate 90%) and re-examined in 1974, 1980, 1992, 2000 and 2005. PARTICIPANTS: 800 women born in 1914, 1918, 1922 and 1930 who were systematically selected for a psychiatric examination at baseline, in 1968. PRIMARY AND SECONDARY OUTCOME MEASURES: 18 psychosocial stressors (eg, divorce, widowhood, work problems and illness in relative) were obtained at baseline. Symptoms of distress were measured according to a standardised question at each study wave. Dementia was diagnosed according to Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) criteria based on information from neuropsychiatric examinations, informant interviews, hospital records, and registry data, and measured through the whole study period.
RESULTS: During the 37 years of follow-up, 153 women developed dementia (104 of those had Alzheimer's disease (AD)). Number of psychosocial stressors in 1968 was associated (HR, 95% CI) with higher incidence of dementia (1.15, 1.04 to 1.27) and AD (1.20, 1.07 to 1.35) between 1968 and 2005, in multivariate Cox regressions. Number of psychosocial stressors in 1968 was also associated (OR, 95% CI) with distress in 1968 (1.48, 1.32 to 1.67), 1974 (1.31, 1.17 to 1.46), 1980 (1.27, 1.11 to 1.45), 2000 (1.39, 1.14 to 1.70) and 2005 (1.35, 1.02 to 1.79), in multivariate logistic regressions. Number of psychosocial stressors (HR 1.17, 95% CI 1.03 to 1.33) and long-standing distress (1968-1974-1980) (HR 1.58, 95% CI 1.03 to 2.45) were independently associated with AD.
CONCLUSIONS: Our study shows that common psychosocial stressors may have severe and long-standing physiological and psychological consequences. However, more studies are needed to confirm these results and investigate whether more interventions such as stress management and behavioural therapy should be initiated in individuals who have experienced psychosocial stressors.

To study the relation between psychosocial stressors, long-standing distress and incidence of dementia, in a sample of women followed over 38 years, from midlife to late life.

The study shows that the number of psychosocial stressors, measured in middle-aged women, was related to distress and incidence of AD almost four decades later.

The study also shows that the association between number of psychosocial stressors and AD was independent of long-standing perceived distress.

Midlife report of psychosocial stressors occurring long before dementia onset, the long follow-up period, the representative population and that multiple sources of information were used to detect and diagnose dementia.

The rating of stressors was related to the last year for some stressors and at any time before 1968 for other stressors. Some stressors were of a short duration, while others were chronic and lasting for many years. We only have information on a limited number of psychosocial stressors in our population. Individuals have different capacities to cope with stress and thus react differently when exposed to the same stressor. We did not have an individual weighting of the included stressors.

Introduction

Experiences of severe psychological stressors in adulthood (eg, combat,1 natural disasters2 and the Holocaust3) are known to influence mental and physical health decades later. Mild psychosocial stressors are common and could be considered as part of normal life. The long-term consequences of these more common stressors remain unclear. Epidemiological studies in the elderly with the follow-ups of less than 10 years have reported that history of early parental death,4–6 death of spouse7 and psychosocial risk factors in childhood6 increase the risk of dementia or Alzheimer's disease (AD). One explanation for the associations is that traumatic experiences may perhaps give rise to long-standing chronic distress many years after the trauma. This may lead to a cumulative burden to the brain with dysregulation in neuroendocrine systems.8–10 A study among Holocaust survivors found that higher levels of stress hormones remained decades after the traumatic experience.8

We have previously reported that long-standing distress in midlife leads to long-term consequences decades later, such as increased risk of dementia, AD11 and structural brain changes.12 To our knowledge, no population study has examined whether number of psychosocial stressors in midlife increase the risk of dementia in late life, and whether this is modified by long-standing distress.

The aim of this study was to examine whether common psychosocial stressors in midlife were related to distress, late-life dementia and AD, in women followed over 38 years. We further aimed to examine whether experiences of psychosocial stressors modify the previously reported association between long-standing midlife distress and AD.

Methods

Study population

This study is part of the Prospective Population Study of Women in Gothenburg, Sweden,13 14 which was initiated in 1968 with an examination of 1462 women (participation rate 90%) born in 1908, 1914, 1918, 1922 and 1930. The individuals were systematically sampled from the Swedish Population Registry based on specific birth dates in order to yield a representative sample at the ages studied. The follow-ups were performed in 1974, 1980, 1992, 2000 and 2005 with participation rates among survivors of 91%, 83%, 70%, 71% and 70%, respectively. The informed consent was obtained from all participants, in accordance with the provision of the Helsinki Declaration.

The current study included a subsample of 800 women who were systematically selected for a psychiatric examination in 1968. The women were aged 38 years (n=111), 46 years (n=309), 50 years (n=290) and 54 years (n=90). Among them, 713 participated in the follow-up examination in 1974, 639 in 1980, 472 in 1992, 368 in 2000 and 296 in 2005. Losses were mainly due to death.

Assessment of psychosocial stressors

At baseline 1968, 18 predefined psychosocial stressors were asked and rated by a psychiatrist during the psychiatric examination. These included divorce, widowhood, serious problem in children (eg, physical illness, death and abuse), extramarital childbirth, mental illness in spouse or first-degree relative, alcohol abuse in spouse or first-degree relative, physical illness or social problems related to husband, receiving help from Social Security, problem related to husband's or own work (eg, lost work) and limited social network. Some of the stressors (physical illness, mental illness and alcohol abuse in spouse; serious problem and mental illness in child; work-related problems and limited social network) were rated in the last year before examination in 1968. The others were rated as occurring at any time prior to the examination in 1968.

Assessment of distress

Symptoms of distress were rated according to a standardised question in 1968, 1974, 1980, 2000 and 2005. The question was worded identically at each examination; “Have you experienced any period of distress (1 month or longer) in relation to circumstances in everyday life, such as work, health or family situation? Distress refers to feelings of irritability, tension, nervousness, fear, anxiety or sleep disturbances.” Participants were asked to choose between; 0=have never experienced any period of distress; 1=have experienced period/s of distress more than 5 years ago; 2=have experienced one period of distress during the last 5 years; 3=have experienced several periods of distress during the last 5 years; 4=have experienced constant distress during the last year or 5=have experienced constant distress during the last 5 years. In the current study, distress is defined as a rating of 3–5.

Psychiatric examinations

The psychiatric examinations were conducted in 1968, 1974, 1980 and 1992 by psychiatrists and in 2000 and 2005 by experienced psychiatric research nurses. The examinations were semistructured and included a comprehensive neuropsychiatric examination and an extensive battery of neuropsychiatric tests.15 Close informant interviews were conducted in 1992, 2000 and 2005. These included questions about changes in behaviour and intellectual functions and, in cases of dementia, age of onset and disease course.15 Medical records were collected from all inpatient and outpatient departments and general practitioners’ offices in Gothenburg. The Swedish Hospital Discharge Registry provided diagnostic information for all individuals discharged from hospitals on a nationwide basis since 1978.

Diagnosis of dementia

The diagnosis of dementia was based on information from psychiatric examinations, close informant interviews, medical record examinations and the Swedish Hospital Discharge Registry. The diagnostic procedures have been described in detail previously.15 Dementia diagnosis at each examination was made according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) based on the combined information from the psychiatric examination and the close informant interview. Dementia diagnoses for individuals lost to the follow-up were based on information from medical records evaluated by geriatric psychiatrists in consensus conferences, and information from the Swedish Hospital Discharge Registry.16

AD was diagnosed according to the criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA).17 The criteria for vascular dementia (VaD) were similar to the criteria proposed by the National Institute of Neurological Disorders and Stroke and the Association Internationale pour la Recherce et l´Enseignement en Neurosciences (NINDS-AIREN).18 VaD was thus diagnosed when there was a temporal relationship (within 1 year) between a history of acute focal neurological symptoms and signs (haemiparesis or motor aphasia) and the first symptoms of dementia. Other dementias were diagnosed when other causes were likely to have caused the dementia. Person-years were calculated from the date of the baseline examination to (1) the time of dementia onset; (2) the date of death; (3) the date of the last follow-up examination for participants in 2005 or (4) 31 December 2006 for surviving drop-outs.

Potential confounders and mediators

Information on education, socioeconomic status, marital status and work status was obtained at the examination in 1968, and information on blood pressure, antihypertensive medication use, coronary heart disease (CHD), diabetes mellitus, stroke, waist and hip circumferences, cigarette smoking and wine consumption was obtained at the examinations in 1968, 1974 and 1980. Education was dichotomised as compulsory (6 years for those born during 1914–1922 and 7 years for those born in 1930) versus more than compulsory education. Socioeconomic status was based on husband's occupation for married women, and own occupation for unmarried women and was defined as higher middle, lower middle, skilled workers and unskilled workers.19 Marital status was classified as married and/or cohabiting versus single. Work status was measured as full-time work and/or part-time work versus no work outside home. Hypertension was defined as systolic blood pressure of 160 mm Hg or more, and/or diastolic blood pressure 95 mm Hg or more and/or taking antihypertensive medication. CHD was defined as angina pectoris according to the Rose criteria20 or documented history of myocardial infarction. Diabetes mellitus was defined as a diagnosis told by a doctor, death certificates, being on antidiabetes drugs or having two fasting blood glucose values of 7 mmol/L or more. Stroke was diagnosed based on information from the examinations and the Swedish Hospital Discharge Registry. High waist-to-hip ratio was defined as a ratio of waist and hip circumferences over 0.85. Cigarette smoking was defined as never, former or current smoker. Wine consumption was classified as none, less than once weekly and once weekly or more.

Statistical analyses

Logistic regressions were used to analyse the associations between number of psychosocial stressors in 1968 and report of distress in 1968, 1974, 1980, 2000 and 2005. The results are presented as ORs and 95% CIs in three separate models. The first model adjusts for age only. The second model adjusts for age, education, socioeconomic status, marital status, work status, hypertension, CHD, stroke, diabetes mellitus, waist-to-hip ratio, smoking and wine consumption. The third model adjusts for age and psychiatric family history, that is, mental illness in mother, father and/or sibling. (These three variables were then not counted as psychosocial stressors.)

Cox regressions were used to study the associations between number of psychosocial stressors and incidence of dementia and dementia subtypes. Associations are presented as HRs and 95% CIs, and model 1–3 adjust for the same covariates as listed above. The fourth model adjusts for age and long-standing midlife distress (ie, distress in all examinations 1968–1974–1980). Two interaction models were also added; (1) number of stressors × psychiatric family in relation to AD and (2) number of stressors × long-standing distress in relation to AD. Finally, we examined the associations between long-standing midlife distress and psychosocial stressors in relation to AD before and after age 75.

Results

Characteristics of the 800 participants are given in table 1. The proportion of women who reported specific life stressors in 1968 are shown in table 2. Twenty-five per cent of the women reported one psychosocial stressor, 23% reported two stressors, 20% three stressors and 16% four or more stressors. The most frequently reported psychosocial stressor was mental illness in first-degree relative (mother 27%, father 19% and sibling 32%).

Table 1

Characteristics of the study sample (N=800)

	N	Per cent
Birth year (age)
1914 (54 years)	89	11.1
1918 (50 years)	291	36.4
1922 (46 years)	309	38.6
1930 (38 years)	111	13.9
Education*
Compulsory	600	75.0
More than compulsory	200	25.0
Socioeconomic status*
Upper middle	161	20.2
Lower middle	267	33.4
Skilled workers	209	26.1
Unskilled workers	163	20.4
Marital status*
Married	638	79.8
Cohabited (not married)	94	11.7
Living alone (not married)	68	8.5
Work status*
Full-time work	270	33.8
Part-time work	258	32.3
No work outside home	272	34.0
Hypertension†	144	18.0
Coronary heart disease†	74	9.3
Diabetes mellitus†	24	3.0
Stroke†	5	0.5
Smoking†	341	42.6
Wine consumption†	246	30.8
High waist-to-hip ratio†	210	26.3

*Measured in 1968.

†Measured in 1968, 1974 and 1980.

Table 2

Prevalence of psychosocial stressors in women in 1968 (N=800)

	N	Per cent
Physical illness in spouse	62	7.8
Mental illness in spouse	98	12.3
Alcohol abuse in spouse	55	6.9
Social problem in spouse	81	10.1
Work related problems in spouse	32	4.0
Serious problem in children	70	8.8
Mental illness in child	139	17.4
Mental illness in father	151	18.9
Alcohol abuse in father	100	12.5
Mental illness in mother	212	26.5
Mental illness in sibling	255	31.9
Alcohol abuse in sibling	79	9.9
Divorced	65	8.1
Widowed	34	4.3
Limited social contacts	53	6.6
Work related problems	19	2.4
Received help from social security	10	1.3
Extramarital childbirth	84	10.5
Number of psychosocial stressors
0 psychosocial stressor	149	18.6
1 psychosocial stressor	197	24.6
2 psychosocial stressors	184	23.0
3 psychosocial stressors	143	19.9
4 psychosocial stressors	69	8.6
≥5 psychosocial stressors	58	7.2

Four hundred and twenty-five participants died during the follow-up (mean age 79 years). From 1968 to 2006, 153 (19.1%) women developed dementia during 25 131 person-years of follow-up, including 104 with AD, 35 with VaD and 14 with other dementias. The mean time from the baseline examination in 1968 to dementia onset was 29 years (26 had dementia onset before 1992, 73 between 1992 and 2000 and 54 after 2000). Mean age of dementia onset was 78 years (45 had dementia onset before age 75 years and 108 after age 75 years).

Number of psychosocial stressors in 1968 was associated with distress in 1968, 1974, 1980, 2000 and 2005, after adjustment for potential confounders (table 3). ORs were similar after further adjustment for psychiatric family history in model 3. Number of psychosocial stressors was associated with long-standing midlife distress (ie, distress in 1968–1974–1980) both in later born cohorts, born 1922 and 1930, (multiadjusted OR 1.32, 95% CI 1.14 to 1.52) and earlier born cohorts, born 1914 and 1918 (multiadjusted OR 1.58, 95% CI 1.30 to 1.94).

Table 3

Number of psychosocial stressors in 1968 in relations to report of distress in 1968, 1974, 1980, 2000 and 2005

	Cases, n (%)	Model 1	Model 2	Model 3
Distress in 1968	148 (18.5)	1.46 (1.30 to 1.63)	1.49 (1.31 to 1.70)	1.61 (1.22 to 2.13)
Distress in 1974	161 (20.1)	1.31 (1.18 to 1.46)	1.33 (1.17 to 1.50)	1.23 (1.05 to 1.44)
Distress in 1980	88 (11.0)	1.26 (1.10 to 1.43)	1.26 (1.08 to 1.47)	1.22 (1.00 to 1.50)
Distress in 2000	49 (6.1)	1.41 (1.17 to 1.72)	1.40 (1.13 to 1.74)	1.24 (0.95 to 1.64)
Distress in 2005	39 (2.6)	1.37 (1.05 to 1.80)	1.35 (1.00 to 1.85)	1.50 (1.05 to 2.20)

Logistic regression analyses presented as ORs with 95% CIs; model 1 adjust for age; model 2 adjust for age, education, socioeconomic status, marital status, work status at baseline (in 1968), and hypertension, CHD, stroke, diabetes mellitus, high waist-to-hip ratio, smoking and wine consumption (in 1968–80); and model 3 adjust for age and psychiatric family history (mental illness in mother, father and/or sibling is not included in number psychosocial stressors).

Number of psychosocial stressors in 1968 was associated with higher incidence of AD (HR 1.21, 95% CI 1.08 to 1.36) and all-type dementia (HR 1.15, 95% CI 1.05 to 1.27; table 4). The associations remained after adjusting for multiple confounders in model 2, psychiatric family history in model 3 and long-standing distress (ie, distress in 1968–1974–1980) in model 4. In the fourth model, long-standing distress (HR 1.58, 95% CI 1.01 to 2.46) and number of psychosocial stressors (HR 1.17, 95% CI 1.02 to 1.33) were independently associated with AD. There were no interactions between number of stressors and psychiatric family history in relation to AD (age-adjusted HR 1.05, 95% CI 0.75 to 1.45, p=0.79) or between number of stressors and long-standing distress in relation to AD (age-adjusted HR 1.04, 95% CI 0.77 to 1.40, p=0.82). The association between number of psychosocial stressors and incidence of AD were similar in those with early onset AD (aged <75 years; multiadjusted HR 1.25, 95% CI 1.02 to 1.54) and late onset AD (aged ≥75 years; multiadjusted HR 1.19, 95% CI 1.03 to 1.38). There were no visible associations between number of psychosocial stressors and VaD in any of the models.

Table 4

Number of psychosocial stressors in 1968 in relation to incidence of dementia over 38 years

	Casesn (%)	Model 1	Model 2	Model 3	Model 4
All-type dementia	153 (19.1)	1.15 (1.05 to 1.27)	1.16 (1.04 to 1.30)	1.10 (1.00 to 1.25)	1.13 (1.01 to 1.26)
Vascular dementia	35 (4.4)	0.94 (0.75 to 1.19)	0.97 (0.75 to 1.26)	0.79 (0.57 to 1.10)	0.93 (0.71 to 1.22)
Alzheimer's disease	104 (13.0)	1.21 (1.08 to 1.36)	1.21 (1.06 to 1.38)	1.16 (1.00 to 1.35)	1.17 (1.02 to 1.33)

Cox regression analyses presented as HRs with 95% CIs; model 1 adjust for age; model 2 adjust for age, education, socioeconomic status, marital status, work status at baseline (in 1968), and hypertension, CHD, stroke, diabetes mellitus, high waist-to-hip ratio, smoking, and wine consumption (in 1968–1980) and model 3 adjust for age and psychiatric family history (mental illness in mother, father and/or sibling is not included in number psychosocial stressors), and; model 4 adjust for age and long-standing distress (in 1968–1980).

Discussion

We found that number of common psychosocial stressors in midlife was associated with incidence of late-life dementia, especially AD, in a population-based sample of women followed for 38 years. The associations remained when controlling for long-standing distress. We also found that number of psychosocial stressors in 1968 was related to increased level of distress at every examination conducted between 1968 and 2005.

We have previously reported that long-standing distress in midlife increase risk of AD11 and structural brain changes.12 These findings are now extended by showing that number of psychosocial stressors and report of distress independently predicted AD, that is, increased distress could not completely explain the association between midlife stressors and dementia. One reason for this is that individuals respond differently to psychosocial stressors. Thus, biological responses may develop as a reaction to psychosocial stressors also in individuals who do not experience or report increased distress in association to the stressor.

There may be several biological explanations for the association between psychosocial stressors in midlife and dementia. One is related to the stress hypothesis. Stress may cause a number of physiological reactions in the central nervous, endocrine, immune and cardiovascular systems.10 21 Thus, psychological stress has been reported to increase the activity of the hypothalamic–pituitary–adrenal axis and the levels of glucocorticoid hormones,22 cause structural and functional damage to the hippocampus,22 influence learning and memory processes,23 increase the production of proinflammatory cytokines in the brain,10 increase the deposition of β-amyloid peptid and τ-protein in the brain24–26 and increase the frequency of cardiovascular disease27 28 and hypertension.29 All these factors have been linked to dementia.30

The associations between psychosocial stressors reported in midlife and perceived distress later in life was consistent through all follow-up years, as indicated by ORs of similar magnitude. Thus, even common psychosocial stressors (related to work and family) can cause distress over several decades. Our finding is supported by studies reporting that stress-hormones may remain elevated many years after traumatic events.8 Another explanation is that experiences of psychosocial traumas might make an individual more vulnerable to future stressors due to biological changes and dysfunctional stress coping mechanisms.31 32

Strengths and weaknesses of the study

The strengths of this study include midlife report of psychosocial stressors occurring long before the onset of dementia, the long follow-up period, the representative population and that multiple sources of information were used to detect and diagnose dementia. Some methodological issues need to be considered. First, the rating of stressors was related to the last year for some stressors and at any time before 1968 for other stressors. However, both these were related to the outcome in a similar way (data not shown). Second, some stressors were of a short duration, while others were chronic and lasted for many years. In addition, some stressors were severe and others more trivial. This might give an unbalanced weight among the factors studied. Third, we only have information on a limited number of psychosocial stressors in our population. Some events were not included, for example, physical abuse and own severe physical illness. The relationships might thus have been confounded by unmeasured factors. However, it is not likely that this had any major influence on our findings. Fourth, different individuals have varied capacities to cope with stress and thus react differently when exposed to the same stressor. We did not have an individual weighting of the included stressors. If anything, this might have decreased the strengths of associations. Fifth, some stressors are interrelated, for example, mental illness and alcohol abuse in spouse. However, these stressors independently increased stress reactions (data not shown). We, therefore, decided not to merge them. Sixth, distress in our study was based on self-report and we did not include an objective measure of stress reactions. However, most epidemiological studies use subjective report to assess stress or distress. Seventh, there are a number of risk factors occurring between baseline and development of dementia and these might potentially modify the association between common psychosocial stressors in midlife and dementia. However, these risk factors would most likely decrease the possibility of finding associations in a study with the long follow-up, as may exert competing risk, and controlling for future factors might lead to an over-adjustment. Eighth, psychiatric family history may have an impact on the predisposition to distress and dementia. However, after adjusting for psychiatric family history (ie, mental illness in mother, father and/or sibling) the associations between the number of stressors was still associated with both long-standing distress, AD and all-type dementia. Ninth, cumulative attrition is a problem in the long-term follow-up studies. While this problem was, to some extent, alleviated by using medical records and the hospital registry data to diagnose dementia in those lost to follow-up, these sources probably underestimate the number of dementia cases. It should be noted, however, that almost all people in Sweden received their hospital treatment within the public healthcare system during the time of the study and that the Swedish Hospital Discharge Register covers the entire country. Furthermore, the number of demented women detected in the different age groups is what could be expected from other incidence studies.33 Finally, it is difficult to diagnose dementia subtypes on clinical grounds alone. Individuals with AD often have cerebrovascular disease and individuals with VaD often have concomitant AD pathology. Furthermore, cerebrovascular disease may influence the presence and severity of clinical symptoms of AD, and vice versa.34 It is thus often difficult to make a clear distinction between AD and VaD in patients with a history of stroke or cerebrovascular disease, on clinical grounds and at autopsy, and mixed types are probably common.

Conclusion

To conclude, psychosocial stressors in midlife were associated with incidence of AD and long-standing distress, over several decades. This suggests that common psychosocial stressors may have severe and long-standing physiological and psychological consequences. However, more studies are needed to confirm these results and investigate whether more interventions such as stress management and behavioural therapy should be initiated in individuals who have experienced psychosocial stressors.