Mental Vulnerability as a Predictor of Early Mortality
Eplov, Lene Falgaard*; Jørgensen, Torben*; Birket-Smith, Morten†; Segel, Stine*; Johansen, Christoffer‡; Mortensen, Erik Lykke§
From the *Copenhagen County Research Centre for Prevention and Health, Glostrup University Hospital, Copenhagen; †Liaison Psychiatric Unit, Bispebjerg University Hospital, Copenhagen; ‡The Danish Cancer Society, Institute of Cancer Epidemiology, Copenhagen; and §Department of Health Psychology, Institute of Public Health, University of Copenhagen, Copenhagen, Denmark.
Submitted 10 July 2003; final version accepted 23 November 2004.
Supported by Grant number 2000B087 from the Danish Health Insurance Foundation.
Supplemental material for this article is available with the online version of the journal at www.epidem.com
Correspondence: Lene Falgaard Eplov, Copenhagen County Research Center for Prevention and Health, Glostrup University Hospital, Nordre Ringvej 57, DK-2600 Glostrup, Denmark. E-mail: Falgaard@dadlnet.dk.
Background: Studies have demonstrated that mental vulnerability (ie, a tendency to experience psychosomatic symptoms or inadequate interpersonal interactions) is associated with various diseases. The objective of our study is to evaluate whether mental vulnerability is a risk factor for early mortality.
Methods: We conducted a prospective cohort study of 3 random samples of the population in Copenhagen County, Denmark selected in 1976, 1982–1984, and 1991 (n = 6435). Baseline data collection included measures of mental vulnerability, social factors, comorbidity, biologic risk markers (eg, blood pressure, lipid levels), and lifestyle factors. We determined vital status of the study sample through linkage to the Civil Registration System until 2001 and to the Cause of Death Registry until 1998. The mean follow-up time was 15.9 years for analysis of total mortality and 13.6 years for analysis of mortality as the result of natural causes. The association between mental vulnerability and survival was examined using Kaplan-Meir plots and Cox proportional-hazard models adjusting for possible confounding factors.
Results: With respect to mental vulnerability, 79% of the sample was classified as not vulnerable, 13% as moderately vulnerable, and 8% as highly vulnerable. Compared with the nonvulnerable group, highly vulnerable persons showed increased total mortality (hazard ratio = 1.6; 95% confidence interval = 1.3–1.9) and increased mortality from natural causes (1.6; 1.2–2.0). The inclusion of the mental vulnerability score as a continuous variable gave similar results.
Conclusions: Mental vulnerability may be an independent risk factor for premature mortality. The biologic mechanisms that may underlie this association need further exploration.
Click on the links below to access all the ArticlePlus for this article.
Please note that ArticlePlus files may launch a viewer application outside of your web browser.
Personality traits are stable patterns of thought, emotion, and behavior that characterize an individual through time and situations and generally are assumed to affect a person's vulnerability to various types of disease. A recently published review1 proposed that a combination of chronic distress and disagreeable social behavior (eg, selfishness and intolerance) may constitute a disease-prone personality. The authors of the review criticized the traditional approach of analyzing personality traits separately from essential elements of the surrounding social context. They propose instead that risk for development of diseases is conferred not through specific personality traits but through recurring transactions between individuals with certain personality traits and the social environment.
A mental vulnerability scale was developed in Denmark in the 1960s. According to this scale, the mentally vulnerable individual consistently has both psychosomatic reactions and negative reactions in social interactions, where psychosomatic reaction is defined as bodily reactions caused by mental processes.2,3 The theory behind the scale assumes the adaptation of an individual to be a result of an interaction between a person's characteristic and consistent behavior and the environment.2
A national Danish study demonstrated that mentally vulnerable persons, compared with nonvulnerable individuals, report more diseases and symptoms, use health services more often, and have a less healthy lifestyle.3 In other studies, mental vulnerability has been associated with upper dyspepsia,4 irritable bowel syndrome,5 and peptic ulcer6 but not with cancer.7 Furthermore, mental vulnerability predicted persistent pain after lumbar spine surgery8 and cholecystectomy9,10 independently of known risk factors.
Objective indicators of health status are not always available, and many epidemiologic studies are based on measures of self-reported health. These measures may be systematically biased because certain personality traits (eg, neuroticism) are associated with a tendency to experience and report physical symptoms and health problems.11 An alternative approach is to use mortality as the outcome measure. Although death is the extreme manifestation of disease, it is plausible that effects of psychologic factors on physical health could result in excess or premature mortality.1,12 The objective of the present study was to evaluate the association between mental vulnerability and premature mortality by using a prospective design and taking into account a number of well established risk factors.
The sample consisted of 3 study groups sampled randomly at different times (1976, 1982–1984, and 1991) from the southwestern part of Copenhagen County, Denmark (Table 1). The selected individuals who agreed to participate were examined at the Copenhagen County Research Center for Prevention and Health. The total sample included 8505 men and women of whom 6528 (77%) participated.13–15 Ninety-three persons did not complete the test for mental vulnerability, leaving 6435 individuals for this analysis.
Information on mental vulnerability, sex, age, social class, marital status, diseases, physical activity, tobacco use, and alcohol consumption was obtained from self-administered and interview-administered questionnaires. The general health examination included weight, height, blood pressure, and blood samples for measures of cholesterol and triglyceride levels.13–15 Body mass index (kg/m2) was calculated from the height and weight data.
Since 1968, all residents of Denmark have been assigned a 10-digit personal identification number, and this number was used to link the current sample to 2 relevant registries. Information on vital status and immigration was obtained from the Civil Registration System up to 18 September 2001, and information on causes of death was obtained from the Cause of Death Registry up to 31 December 1998. The Cause of Death Registry classifies according to the International Classification of Diseases (ICD). Cases of death from suicide or accidents were identified by codes E800-E999 in the ICD 8 and V00-Y99 in the ICD 10. (ICD-9 was never used in Denmark).
The Mental Vulnerability Scale
We define mental vulnerability as “a tendency to experience psychosomatic symptoms or inadequate interpersonal reactions.”3 The Mental Vulnerability Scale (formerly called the Psychic Vulnerability Scale) is a Danish scale based on a 27-item, yes/no questionnaire constructed by the Military Psychologic Services in Denmark in the 1960s.2 It is similar to international instruments developed in the same period, including 2 shortened versions of the Cornell Medical Index developed to find “emotionally disturbed persons,”16 and the “Twenty Two Item Screening Score” developed to measure impairment in life functioning due to common types of psychiatric symptoms.17 We used a shortened version 12-item version (available with the electronic version of this article). This 12-item scale has been validated as a reliable measure of a stable individual characteristic in studies comparing the measure to information collected from physicians on 1300 of the participants,3,18,19 and in a test-retest comparison study of 3000 participants with measures taken 1 year apart.18
We constructed 3 classes of mental vulnerability from the 12-item version of the Mental Vulnerability Scale, which was asymmetrically distributed in the sample (Fig. 1). Individuals with 0–2 affirmative answers were classified as not vulnerable (n = 5081, 79%), individuals scoring 3–4 as moderately vulnerable (n = 836, 13%), and those scoring 5 or more as highly vulnerable (n = 518, 8%). The cumulative total mortality by mental vulnerability classification was examined by Kaplan–Meier plots for men and women. We used the Cox proportional-hazards model, with age as the underlying time, to estimate the hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between mental vulnerability class and mortality risk. In subsequent models the association was adjusted for sex and morbidity (answer to the question: “Has a medical doctor ever told you, that you have a cardiac disease, a disease in the brain, bronchitis, diabetes or hypertension”), health status (systolic and diastolic blood pressure, triglyceride, cholesterol and body mass index), lifestyle (smoking habits, alcohol consumption, and physical activity), and social factors (social class and marital status).20 For each model backward selection was conducted, excluding nonsignificant variables at a 5% level. We performed the analysis in the total sample and also in the sample restricted to those who died of natural causes. Continuous variables fulfilled the criteria for linearity. The association between mental vulnerability and mortality was examined in the final model, including the mental vulnerability score as a continuous variable after testing for linearity, and for each sex separately. Finally, we examined the relationship among the 3 classes of mental vulnerability and survival for various causes of death using Cox proportional-hazard models, with age as the underlying time in a model adjusting for sex.
The mean follow-up time was 15.9 years (range, 0.1–25.4 years) for analysis of total mortality and 13.6 years (range, 0.1–22.7 years) for mortality due to natural causes. The mean age of the sample at inclusion was 46 years (range, 30–71 years). For the analysis of total mortality 1013 persons had died; and among these, 720 were not vulnerable, 160 were moderately vulnerable, and 133 were highly vulnerable. Summary statistics of covariates are shown in Table 2.
In both men and women, the Kaplan–Meier plots showed a clear association between mental vulnerability and total mortality, indicated by the risk estimates across categories of increasing mental vulnerability. The plots suggest a somewhat different association for men and women (Fig. 2), but there was no statistically significant interaction between sex and mental vulnerability (P = 0.46); therefore, men and women were analyzed together with sex included as a covariate.
Results from Cox proportional-hazards models showed highly vulnerable persons to have an increased mortality compared with individuals categorized as not vulnerable persons (Table 3). Adjusting for morbidity, lifestyle, and social class reduced hazard ratio from 2.1 to 1.6. Censoring individuals dying from unnatural causes (n = 27), had little effect on the associations (for moderately vulnerable: HR = 1.2, CI = 1.0–1.5; for highly vulnerable: HR = 1.6, CI = 1.2–2.0). Including the mental vulnerability score as a continuous variable in the final model gave similar results (data not shown). The final model was analyzed separately by sex, revealing the same hazard ratios for highly vulnerable men and women (for men 1.5, 1.1–2.1; for women 1.6, 1.2–2.1). Highly vulnerable persons had increased death rates compared with persons who were not vulnerable for each of the individual causes of death we examined (cardiovascular disease, cancer, suicides and accidents, and all others; Table 4).
In this population-based cohort study, we observed associations between mental vulnerability and premature mortality after adjusting for well-established risk factors for early death. Considerable evidence was found that psychosocial factors, including nonspecific somatic and psychologic symptoms, poor social network, low socioeconomic status, low social activity, and bereavement, can affect health and longevity.1,21–24 Hostility is the personality trait that has been most extensively studied in relation to mortality risk, and a recent meta-analysis of this literature concluded that hostility is an independent risk factor for all-cause mortality.25 Neuroticism has not been shown to be associated with mortality.26,27 Consequently, Costa and McCrae12 reasoned that although there is no evidence that neuroticism leads directly and causally to the development of life-threatening disease, this trait is intimately linked to health habits, somatic complaints, illness behavior, and even medical diagnosis. Other studies have shown associations of psychoneurosis, suppressed anger, suspiciousness, psychologic distress, and cynicism with mortality.23,27–30
Personality traits such as hostility and neuroticism are defined as stable patterns of thoughts, emotion, and behavior. Mental vulnerability describes a pattern of consistent psychosomatic reactions as well as negative reactions in interpersonal interactions.1–3 Mental vulnerability includes only 2 reaction patterns, whereas closely related personality traits such as neuroticism include anxiety, angry hostility, depression, self-consciousness, impulsiveness, and vulnerability.31 In a study of 49 patients after laparoscopic cholecystectomy, we observed a moderate correlation between the Eysenck Neuroticism Scale and the 12-item Mental Vulnerability Scale used in this analysis (Spearman's correlation 0.66).9 This level of correlation suggests that mental vulnerability is most likely a component of neuroticism, but neuroticism is a broader concept than mental vulnerability. In the present study, the Mental Vulnerability Scale was compared with 2 items on depressed mood in the SL-36, Mental Health Dimension (“Have you felt so down in the dumps that nothing could cheer you up?” “Have you felt downhearted and blue?”) in the 1936 cohort. The analysis revealed correlation coefficients of 0.41 and 0.44, respectively (Spearman's correlation). Thus, there is some overlap between depression and mental vulnerability, but the results give reason to believe that these are different constructs.
The Mental Vulnerability Scale (online appendix) includes several questions on bodily symptoms. One might argue that these physical symptoms indicate the early stages of medical diseases and that the presence of a serious medical disorder, although undiagnosed, might account for the observed association in the present study. However, the somatic symptoms in the scale are nonspecific, relatively common symptoms in the population, and are normally associated with stress and arousal, as well as mental conditions such as anxiety, depression and somatization.
A previous Danish study demonstrated that mental vulnerability is associated with an unhealthy lifestyle.3 Controlling for health-related behaviors did not substantially change the hazard ratios. Still, a possible explanation may be that mentally vulnerable persons may be less able to change their behavior in a more healthy direction. Further research in the area is needed. Mentally vulnerable persons may also have fewer interactions with the health care system, resulting in earlier death. The percentages of mentally vulnerable persons were higher in the lower social class, which could be related to health care utilization. However, in Denmark there is an equal and free access to the health care services, and in an earlier study a higher use of the health care system was seen in the mentally vulnerable compared with persons who were not vulnerable.3
An association between mental vulnerability and low social status also was observed in a previous study.3 The observed association between mental vulnerability and mortality might therefore reflect social status effects.21 However, controlling for social class and marital status did not change the hazard ratio substantially.
Our results imply that mental vulnerability is an independent risk factor for premature mortality. Mental vulnerability may be associated with higher levels of chronic physiological stress because vulnerable individuals not only react with bodily distress but also create stressful situations because of their negative reactions in interpersonal interactions. This high level of chronic psychologic stress may lead to diseases and early mortality.32 We used mortality as the outcome measure to avoid bias resulting from possible associations between mental vulnerability and tendencies to experience and report physical symptoms. However, total mortality also is problematic as an outcome measure because not all diseases lead to a premature mortality, and total mortality does not reveal associations between mental vulnerability and specific diseases. In this study, we found increased death rates for highly vulnerable persons compared with persons who were not vulnerable for some broad categories of diseases. Further research is needed to explore whether mental vulnerability is a risk factor for specific diseases and whether the mental vulnerability is associated with a higher level of chronic stress. The Mental Vulnerability Scale measures a person's reactions and does not include measures of the surrounding social environment. Therefore, the scale alone does not reflect the dynamic interaction of specific personality traits with recurring transactions that occur between the individual and the social environment and that may lead to disease.1 However, examining the possible link between mental vulnerability and mortality is a necessary first step before evaluating how changes in the social environment might influence the association between this personality trait and mortality.
The current study of mental vulnerability has certain advantages. The cohort was population-based with a relatively high response rate (77%), the study had a prospective design, and follow-up data on mortality were complete because it was based on valid central registers. Furthermore, the test applied in the study is a validated one-dimensional scale. This study also has certain limitations. The exposure assessment was limited since we only obtained information about mental vulnerability at one point in time. Finally, it is also a limitation that the Mental Vulnerability Scale has only been used among Danes, and, therefore, has not been validated in other populations.
In summary, this population-based prospective study suggests that mental vulnerability is an independent predictor of premature mortality. Further research is needed to examine if mental vulnerability is associated with a higher level of chronic stress, and to establish possible associations between mental vulnerability and specific diseases.
1. Smith TW, Gallo LC. Personality traits as risk factors for physical illness. In: Baum A, Revenson T, Singer J, eds. Handbook of Health Psychology
. Hillsdale, NJ: Lawrence Erlbaum; 2000;139–172.
2. Sabroe K, Kousgaard E, Knox-Seith B, Rieneck B, Bratfeldt EL. P-faktoren i FOUSHIP-klassificerings-systemet. Vol I-IV
. Copenhagen: MPT; 1968 (English summaries).
3. Kühl PH, Martini S. Psykisk Sårbare. Sociale livsbetingelser. (Psychically vulnerable persons. Their social and living conditions). Publication no. 102.
Copenhagen: The Danish National Institute of Social Research, 1981 (English abstract and summary).
4. Kay L, Jørgensen T. Upper dyspepsia in persons aged 50 to 60 years. Results from a 5-year follow-up population study in Glostrup. Ugeskr Laeger
5. Kay L, Jørgensen T, Jensen KH. Irritable colon among 30-60-year old subjects. A cohort study of 4581 Danes. Ugeskr Laeger
6. Rosenstock SJ. The Epidemiology of Peptic Ulcer Disease in a Danish County
. Copenhagen: FADL Publishers; 1996. (Thesis).
7. Schapiro IR, Nielsen LF, Jørgensen T, Boesen EH, Johansen C. Psychic Vulnerability and the associated risk for cancer. Cancer
8. Thorvaldsen P, Sørensen EB. Psychological vulnerability as a predictor for short-term outcome in lumbar spine surgery. Acta Neurochir
9. Jess P, Jess T, Beck H, Beck P. Neuroticism in relation to recovery and persisting pain after laparoscopic cholecystectomy. Scand J Gastroenterol
10. Jørgensen T, Teglbjerg JS, Wille-Jørgensen P, Bille T, Thorvaldsen P. Persisting pain after cholecystectomy. Scand J Gastroenterol
11. Watson D, Pennebaker JW. Health complaints, stress, and distress: exploring the central role of negative affectivity. Psychol Rev
12. Costa Pjr, McCrae R. Neuroticism, somatic complaints, and disease: is the bark worse than the bite? J Pers
13. Hollnagel H. The health structure of 40-year-old men and women in the Glostrup area, Denmark—an epidemiological survey. Dan Med Bull
14. Jørgensen T. Gallstones. An epidemiological investigation. Dan Med Bull
15. Thomsen KK, Larsen S, Schroll M. Cardiovascular risk factors and age: a cross-sectional survey of Danish Men and Women from the Glostrup Population Studies. Am J Geriatr Cardiol
16. Levav I, Arnin A, Portnoy A. Two shortened versions of the Cornell Medical Index—A new test of their validity. Int J Epidemiol
17. Langner TS. A. Twenty-Two Item screening score of Psychiatric symptoms indicating impairment. J Health Hum Behav
18. Andersen EB, Sørensen SL. Måling af psykisk sårbarhed. En analyse af visse egenskaber ved sårbarhedstesten. Research report no. 58
. Copenhagen: Universitetets Statistiske Institut; 1979. (In Danish).
19. Rasch G. An item analysis which takes individual differences into account. Br J Math Stat Psychol
20. Cox DR. Regression Models and life tables. J R Stat Soc B
21. Adler NE, Ostrove JM. Socioeconomic status of health: what we know and what we don't. In: Adler NE, Marmot M, McEwen BS, Stewart J, eds. Socioeconomic Status and Health in Industrial Nations. Social, Psychological and Biological Pathways. Annals of the New York Academy of Science, volume 896
. New York: The New York Academy of Science; 1999;3–15.
22. Daly MB, Tyroler HA. Cornell Medical Index response as a predictor of mortality. Br J Prev Soc Med
23. Somervell PD, Kaplan BH, Heiss G, Tyroler HA, Kleinbaum DG, Obrist PA. Psychologic distress as a predictor of mortality. Am J Epidemiol
24. Rosengren A, Orth-Gomér K, Wedel H, Wilhelmsen L. Stressful life events, social support, and mortality in men born in 1933. BMJ
25. Miller TQ, Smith TW, Turner C, Guijarro ML, Hallet AJ. A meta-analytic review of research on hostility and physical health. Psychol Bull
26. Hällström T, Lapidus L, Bengtsson C, Edström K. Psychosocial factors and risk of ischaemic heart disease and death in women: a twelve-year follow-up of participants in the population study of women in Gothenburg, Sweden. J Psychosom Res
27. Almanda SJ, Zonderman AB, Shekelle RB, et al. Neuroticism and cynicism and risk of death in middle-aged men: the western electric study. Psychosom Med
28. Keehn RJ, Goldberg ID, Beebe GW. Twenty-four year mortality follow-up of army veterans with disability separations for psychoneurosis in 1944. Psychosom Med
29. Julius M, Harburg E, Cottington EM, Johnson EH. Anger-coping types, blood pressure and all- cause mortality: a follow-up in Tecumseh, Michigan (1971–1983). Am J Epidemiol
30. Barefoot JC, Siegler IC, Nowlin JB, Peterson BL, Haney TL, Williams RB. Suspiciousness, health and mortality: a follow-up study of 500 older adults. Psychosom Med
31. Costa PT jr, Widinger TA. Appendix D. In: Costa PT Jr, Widinger TA, eds. Personality Disorders and the Five-Factor Model of Personality
. Washington: American Psychological Association; 1994;341–345.
32. McEwen BS, Stellar E. Stress and the individual. Mechanisms leading to disease. Arch Intern Med
Axel Fredrik Cronstedt (born 1722, died 1765) was a famous Swedish mineralogist and chemist, noted for his work on the chemistry of metallic elements. He was one of the first to recognize the importance of the chemical constituents of minerals and rocks, and to use the blowpipe in the study of minerals. He identified a previously undiscovered metallic element, published the finding in 1751, and three years later named it nickel.
“Kupfer-Nickel is the ore that carries the largest content of the previously described and published semi-metal, and therefore I have taken the opportunity to retain the same name for the metal, or, for more suppleness, to call it Nickel, until it can be proved that it is nothing but a composition of earlier known metals or semi-metals.
Translated from Swedish by Tom K Grimsrud - From: Cronstedt, AF. Fortsättning af rön och försök, gjorde med en malm-art från Los koboltgrufvor [Continuation of experience and experiments made with a type of ore from the Los cobalt mines]. Kongl Svenska vetenskapsacademiens handlingar [Acts of the Royal Swedish Academy of Sciences] 1754; 15:38–45.
Supplemental Digital Content
© 2005 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Data is temporarily unavailable. Please try again soon.