- Email: [email protected]
Loneliness and All-Cause, Cardiovascular, and Noncardiovascular Mortality in Older Men: The Zutphen Elderly Study
Accepted Manuscript Title: Loneliness and All-Cause, Cardiovascular and Non-Cardiovascular Mortality in Older Men: The Zutphen Elderly Study Author: Jolien E. Julsing, Daan Kromhout, Johanna M. Geleijnse, Erik J. Giltay PII: DOI: Reference:
S1064-7481(16)00150-0 http://dx.doi.org/doi: 10.1016/j.jagp.2016.01.136 AMGP 568
To appear in:
The American Journal of Geriatric Psychiatry
Received date: Revised date: Accepted date:
25-3-2015 22-1-2016 22-1-2016
Please cite this article as: Jolien E. Julsing, Daan Kromhout, Johanna M. Geleijnse, Erik J. Giltay, Loneliness and All-Cause, Cardiovascular and Non-Cardiovascular Mortality in Older Men: The Zutphen Elderly Study, The American Journal of Geriatric Psychiatry (2016), http://dx.doi.org/doi: 10.1016/j.jagp.2016.01.136. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Word count text: 3663
Loneliness and all-cause, cardiovascular and non-cardiovascular mortality in older men: The Zutphen Elderly Study 1
2
2
Jolien E. Julsing , M.D., Daan Kromhout , Ph.D., M.P.H., Johanna M. Geleijnse , Ph.D., and Erik 1
J. Giltay , Ph.D., M.D. 1
Leiden University Medical Centre, Department of Psychiatry, Leiden, the Netherlands
2
Division of Human Nutrition, Wageningen University, Wageningen, The Netherlands
Correspondence to: J.E. Julsing, M.D. Leiden University Medical Centre, Department of Psychiatry B1-P Postbus 9600 2300 RC Leiden, the Netherlands Telephone: 00316 44444383 Fax: 003171 5248156 Email: [email protected]
The Zutphen Elderly Study was supported by the Netherlands Prevention Foundation (Preventiefonds)
Key words: Loneliness, mortality, aged, cardiovascular death, cohort studies
Page 1 of 29
Abstract Objectives - Loneliness, defined as the discrepancy between one’s desired and actual relationships, is prevalent in the elderly and can be distinguished in emotional and social loneliness. We aimed to determine whether loneliness is independently related to higher allcause, cardiovascular and non-cardiovascular mortality in elderly men. Design - A population-based cohort study with 25 years of follow-up from 1985 onward. Setting - The Zutphen Study started as the Dutch contribution to the Seven Countries Study. Participants - Seven hundred nineteen (76.2%) of 939 men (age range 64-84 years), who had complete data on loneliness at baseline and at least 2 years of survival. Measurements - Loneliness was assessed using a validated 11-item questionnaire in 1985, 1990, 1995 and 2000. Time-dependent Cox proportional hazards models were adjusted for socio-demographic characteristics and cardiovascular risk factors. Results - At baseline, point prevalence of moderate and severe loneliness was respectively 38.8% (n=279) and 3.2% (n=23). Loneliness –especially emotional loneliness– did significantly increase over 15 years with an overall reliability coefficient of 0.50. All-cause, cardiovascular, and non-cardiovascular mortality were not higher among moderately lonely participants (Hazard ratio [HR] 1.00; 95% confidence interval (CI) 0.84-1.17, HR 0.99; 95% CI 0.78-1.25, HR 0.99; 95% CI 0.79-1.24, respectively) and severely lonely participants (HR 1.40; 95% CI 0.85-2.31, HR 1.18; 95%CI 0.58-2.39, HR 1.63; 95% CI 0.80-3.31 respectively). Conclusions - Loneliness is common and increases during aging, due to the increase in its component emotional loneliness over time. No independent associations with risks of allcause, cardiovascular and non-cardiovascular death were found.
Page 2 of 29
Key words - Loneliness, mortality, aged, cardiovascular death, cohort studies
Page 3 of 29
Objective Loneliness is a complex concept involving psychological and social aspects, and is an important indicator of human well-being (1). It is the subjective feeling of isolation, explained as the discrepancy between one’s desired and actual relationships (2), and differs from several measures of social isolation such as living alone, marital status and size of social network. Thus, a person may suffer from loneliness even though he or she is surrounded by many people, and, conversely, persons who live alone may not feel lonely. Weiss (1) distinguished two types of loneliness: emotional and social loneliness. Emotional loneliness is caused by the absence of one or more intimate figures, whereas social loneliness refers to the absence of a social network with whom common interests and activities can be shared. When studying older subjects, emotional loneliness may be particularly relevant since they are more likely to suffer from bereavement. In general, these losses are more salient for men as they rely more on their spouses while forming less close and enduring relationships outside marriage (3). Although loneliness seems to increase with age (4), loneliness may level off at advanced age and does not depend on age any more (5). Nevertheless, loneliness emerges as a prevalent phenomenon with 20 to 40% of older adults reporting feeling lonely at any given time(4, 5). Of them, five to seven percent report feeling intense or persistent loneliness(6). Feelings of loneliness are associated with a large number of health-related problems. These include depressive symptoms (7), poor functional health (8), reductions in physical activity (9), and impaired cognition (10). At the biological level, loneliness has been linked with increased vascular resistance (11), increased systolic blood pressure (12), and markers of low-grade inflammation (13, 14). This could ultimately make lonely people susceptible to a variety of illness, and especially cardiovascular disease (15, 16).
Page 4 of 29
An increasing body of research focused on the hypothesis that loneliness is a risk factor for overall mortality. Where social isolation and low social support have consistently been associated with increased mortality (17-19), the concept of loneliness is only starting to be recognized as a separate entity from social isolation and therefore fewer studies have examined it as an independent risk factor for mortality. Previous prospective studies investigating the effects of loneliness on mortality in the elderly included men as well as women aged above 50. Follow-up time differed between 6 and 20 years. Except for one study that used the 11-item the Jong Gierveld Loneliness Scale (20), being lonely was assessed rather crudely, with a 3-item or even a single item scale (“do you feel lonely?”). Moreover, apart from three studies (20-22), loneliness was obtained only at baseline and not sequentially during follow-up. Furthermore, these studies focused on all-cause mortality, differed in controlling for potentially confounding factors, and have yielded mixed results. Where some studies found an association between loneliness and all-cause mortality (23-25), another one did not (21) or the association disappeared after multivariable adjustment (8, 20, 26). Three of these studies reported significant sex differences in these findings, with a higher mortality risk for lonely men (8, 22, 25). Therefore, the independent association with cardiovascular mortality has not yet been determined. Because of strong links that have been found between loneliness on the one hand and physical activity, optimism and depressive symptoms, having a partner, and somatic factors like blood pressure and chronic disease on the other hand, it is important to study whether a putative association with mortality is confounded by one of these factors. The purpose of this study was to investigate whether loneliness is an independent predictor for higher all-cause, cardiovascular and non-cardiovascular mortality in older men.
Page 5 of 29
Methods Participants The data for this study were collected as a part of The Zutphen Elderly Study, which is a prospective population-based study among older male inhabitants of Zutphen, a small industrial town with 47,000 inhabitants in the Netherlands. The study characteristics of the cohort have previously been described in detail (27). Briefly, in 1985 a random sample of elderly men in Zutphen born in 1900-1920 was asked to participate. There were a priori no selection criteria for other demographic or health characteristics. Participants were visited for face-to-face interviews between March 1 and June 30 in 1985 (response rate 70.1%), 1990 (response rate 78.0%), 1995 (response rate 74.1%), and 2000 (response rate 72.8%) and were followed for causes of mortality for up to 25 years until 2010. Informed consent was obtained from all study participants. The study was approved by the Medical Ethics Committee of the Leiden University Medical Centre, and in 1995 and 2000, by the Medical Ethics Committee of The Netherlands Organization for Applied Scientific Research (TNO). Of the 939 subjects who participated in the study in 1985, 153 subjects (16.3%) were excluded due to missing information on the loneliness scores. To reduce bias from reverse causation, we included only the men with at least 2 years of survival. Thus, analyses were based on 719 subjects. Compared at baseline, the 719 participating subjects were significantly younger, had a higher Body Mass Index (BMI), had a higher level of education, were more physical active and had less chronic or other cardiovascular diseases than the 153 excluded subjects (data not shown).
Disease ascertainment Information on the vital status of the participants until July 1, 2010, was obtained from municipal population registries. The causes of death were obtained from Statistics
Page 6 of 29
Netherlands and verified with information obtained from the hospital and the general practitioners in Zutphen. The final coding was done by an experienced clinical epidemiologist. Because it is often difficult to determine the underlying cause of death in elderly people, the primary and secondary causes of death were included in the analyses. Cardiovascular disease was defined as codes 390 to 459 according to the International Classification of Diseases, Ninth Revision (ICD-9). The presence of a prevalent cardiovascular disease at baseline, defined as having either myocardial infarction, stroke or heart failure, and other chronic diseases, defined as having either diabetes mellitus, cancer or chronic obstructive pulmonary disease (COPD), was determined by a survey questionnaire and confirmed with hospital discharge data and written information from the general practitioner. Only three men were lost to follow-up and were included in the analyses, but censored after 4.2, 5.8 and 5.1 years of follow-up.
Loneliness Feelings of loneliness were assessed in 1985, 1990, 1995 and 2000 using the Loneliness Scale of de Jong Gierveld, an 11-item questionnaire consisting of two subscales: the emotional loneliness scale and the social loneliness scale (2). The loneliness scale comprises five positive stated items (e.g. “There are enough people that I feel close to”) and six negative stated items (e.g. “I miss having a really close friend”), and ranges from 0-11 points, with higher scores being indicative of higher loneliness levels. The total loneliness score was calculated by assigning a score of one point to the responses indicating a (certain) feeling of loneliness. The emotional loneliness score was the result of the sum of the positively or neutral stated answers on six items, whereas the social loneliness score was the sum of the negatively and neutral stated answers on its five corresponding items. If a respondent had one missing value, the missing item was imputed with the median of the
Page 7 of 29
remaining items of the relevant subscale for that particular subject. The 11-item De Jong Gierveld Loneliness Scale has proved to be a valid and reliable instrument for the assessment of overall, emotional, and social loneliness (28).
Correlates of loneliness In 1985, 1990, 1995 and 2000, dispositional optimism was assessed using a questionnaire consisting of four items on a 3-point scale (29) from a survey of Statistics Netherlands (Centraal Bureau voor de Statistiek – CBS-). The additional answer category “do not know” was coded as the midpoint (score 1). If a respondent had one missing value, the missing item was imputed with the mean of the remaining three items for that particular subject. The optimism questionnaire score ranges from 0 to 8 points, with higher scores being indicative of higher optimism levels. The presence of depressive symptoms was assessed in 1990, 1995 and 2000 using the Dutch translation of the Zung self-rating depression scale, consisting of 20 items on a 4-point scale (30). Finally, the presence of disability was assessed in 1990, 1995 and 2000 with the 13-item Activities of Daily Living questionnaire (31). The following severity levels of disability status were distinguished: no disability (0); mild disability, disability in instrumental activities only (1); moderate disability, disability in instrumental activities and mobility (2); and severe disability, disability in instrumental activities, mobility, and basic activities of daily living (3).
Covariates Information on living arrangement, education, family history of stroke or myocardial infarction, and physical activity was obtained with a questionnaire that participants were asked to complete at home. A validated questionnaire on physical activity designed for retired men was used to calculate the total minutes spent in physical activity per week (of an
Page 8 of 29
intensity of >2kcal/kg per hour) (32) and coded along recommended level or not. Cigarette smoking (never, former, or current smoker) and alcohol use (0, 1-29, or ≥ 30 g/d of alcohol) were coded along three levels. Information on alcohol consumption was collected with the cross-check dietary history method. In 1985, 1990, 1995 and 2000, research assessments were performed by five trained physicians and included body mass index (BMI), blood pressure measurements (in duplicate), and medical history. The use of antihypertensive medication was also assessed. Total and high density lipoprotein (HDL) cholesterol levels were analysed enzymatically in a standardized lipid laboratory.
Statistical analyses The baseline characteristics of the participants were compared between not lonely, moderately lonely and severely lonely participants (respectively defined with a cut-off score between 3-8 and ≥9 on the De Jong Gierveld Loneliness Scale) (33) using the chi-square test (linear by linear term) and one-way analysis of variance (for linear trend), when appropriate. Results are presented as numbers and percentages for categorical variables, and means and standard deviations for continuous variables. Linear mixed models were used to assess the progression of loneliness and its two subscales over 15 years of follow-up. Reliability coefficients (i.e., an intraclass correlation coefficient with a two-way random effects model with single-measure reliability) were used to examine temporal stability of loneliness over the four time points in subjects with paired loneliness scores (meaning that a subject has a non-missing loneliness score in both mentioned time-points). Age of death according to loneliness was assessed by conducting analysis of covariance (ANCOVA) controlling for socio-demographic characteristics and cardiovascular
Page 9 of 29
risk factors. When a participant was alive at the end of the follow-up period, age of death was imputed with the mean remaining life expectancy for men of the same age group, for that particular subject according to Statistics Netherlands (34). The Kaplan-Meier method was used to present the rate of cardiovascular and noncardiovascular mortality in the not lonely, moderately lonely and severely lonely categories. After testing the proportional hazard assumption, hazard ratios (HR) with 95% confidence intervals (CI) of the three types of mortality were estimated by Cox proportional hazards models. Four multivariable models were tested in 1985. Model one was crude, model two adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model three additionally adjusted for dispositional optimism, serum total cholesterol, systolic and diastolic blood pressure, history of having a cardiovascular disease, having another chronic disease, family history of stroke or myocardial infarction and use of anti-hypertensive medication. Model four used the same adjustment factors, but included all covariates if possible as time-dependent variables that were updated using the mean of all previous measurements. If a subsequent score was missing, the missing score was imputed with the mean score of all previous measured scores for that particular subject. We repeated the four Cox proportional hazards models with emotional (ranging from 0 to 6 points) and social (ranging from 0 to 5 points) loneliness on a continuous scale as independent variables. To test whether an association between loneliness and death could be confounded by depressive symptoms or disability, we carried out a sensitivity analysis using 1990 as the starting date and additionally adjusted for Zung self-rating depression score and disability. All p-values are two tailed and considered statistically significant at p < 0.05. Data analyses were performed using SPSS for windows, version 20.0 (Chicago, IL, USA).
Page 10 of 29
Results Baseline characteristics Table 1 shows the baseline characteristics of 719 men according to the presence of subjective feelings of loneliness. Loneliness was present in 302 (42.0%) participants, of whom 279 (92.4%) were moderately lonely and 23 (7.6%) were severely lonely. Loneliness was positively associated with living alone, not being married, lower education and lower baseline dispositional optimism.
Loneliness in time Figure 1 shows a statistically significant increase of loneliness over time (F(1, 1134) = 7.737, p = 0.006). As for the loneliness subscales, we observed a significant increase in the emotional loneliness scores (F(1, 1143)=71.937, p < 0.001) and a statistically significant decrease (F(1, 1179)=15.492, p < 0.001) in social loneliness during follow-up. Among participants who provided paired loneliness scores, the reliability coefficient for loneliness was 0.55 between 1985 and 1990 (single measure; F(1, 441) = 3,398; n=442 pairs), 0.51 between 1990 and 1995 (single measure; F(1, 273) = 3,060; n=274 pairs), and 0.59 between 1995 and 2000 (single measure; F(1, 146) = 3,859; n=147 pairs) (p < 0.001 for all). The overall reliability coefficient was 0.50 (single measure over 4 time points; F(1, 137) = 4.984, p < 0.001; n=138 with complete data).
Loneliness and mortality
Page 11 of 29
During the 25-year follow-up, 683 (95.0%) of 719 men died. Mean age at death was 83.3 years (standard deviation [SD] 6.3). Of these 719 men, 321 (44.6%) died of cardiovascular causes. After controlling for socio-demographic characteristics and cardiovascular risk factors, age of death was similar between not lonely, moderately lonely, and severely lonely men (analysis of covariance, F = 1.083; df = 2, 620; p = 0.34). Kaplan-Meier analysis of survival according to the three categories of loneliness showed no evidence that moderately lonely and severely lonely men had higher cardiovascular or non-cardiovascular mortality rates (log rank test: = 1.967; df = 2; p = 0.37; and log rank test: = 0.508; df = 2; p = 0.78, respectively) compared to not lonely men in 1985 (Figure 2). Table 2 presents the results of the Cox proportional hazard analyses detailing crude and adjusted associations of loneliness with all-cause mortality, cardiovascular mortality and non-cardiovascular mortality. Compared with not lonely men, those who were moderately lonely and severely lonely showed in all four models no significant different hazard ratio’s for all three types of mortality. Table 3 summarizes the mortality risks according to emotional and social loneliness. The HR adjusted for age, demographic factors and lifestyle factors for emotional loneliness and cardiovascular mortality was significant (HR 1.08, 95% CI 1.00 – 1.17, Wald test: 2(1)=4.227; p = 0.04), but was attenuated and no longer significant when cardiovascular risk
factors and dispositional optimism were included in the model and remained so after timedependent emotional loneliness was included in the model. Similar associations were found for all-cause and non-cardiovascular mortality. All results depicting the mortality risks according to social loneliness were non-statistically significant.
Page 12 of 29
We carried out two sensitivity analyses (data not shown). First, a sensitivity analysis using 1990 as the starting date and additionally adjusted for Zung self-rating depression score and disability. The second sensitivity analysis used loneliness as a continuous variable, ranging from 0 through 11. In both sensitivity analyses and both the adjusted and timedependent models, no significant associations were found between loneliness on the one hand and overall, cardiovascular and non-cardiovascular mortality on the other hand..
Conclusions This prospective cohort study examined whether loneliness is related to a higher all-cause, cardiovascular and non-cardiovascular mortality. Although our results demonstrate loneliness as a common entity in older men that increases during aging, no independent associations between loneliness and risks of all-cause, cardiovascular and non-cardiovascular death were found. The percentages of moderately lonely and severely lonely men in this sample were comparable with those reported in the research program of Living arrangements and Social Networks of older adults (LSN) using the same measure (35). Our results that loneliness increases with age are mostly in line with other studies (4, 5, 36). Furthermore, this study demonstrates that social loneliness declines with older age, where emotional loneliness further extends. This is a remarkable finding and we are not aware of other studies regarding this subject. Aging men may be more prone to emotional loneliness because of the increasing risk of bereavement of the death of a partner or close life-long friends. At the same time, loss of these key figures may stimulate older men to come in contact with other people, hence our observed decline in social loneliness. Our observation that loneliness is not independently associated with all-cause mortality stands in contrast to other studies, which suggest that loneliness in older people is
Page 13 of 29
not only a source of great suffering but also contributes to declining longevity (22-25). However, in a number of studies the effect was no longer observed after adjustment for demographic factors (e.g. marital status and education), health indicators (e.g. chronic disease, depression, physical activity and smoking) (8, 20, 26, 37). The most recent study included 2911 elder subjects, with six follow-up measurements covering a time span of twenty years, during which loneliness was also measured with the 11-item De Jong Gierveld Loneliness Scale. Similarly, neither emotional, nor social loneliness were independently associated with mortality once mental health was added to their model. This and our findings suggest a potential pathway from loneliness to mortality through the effects of mental disorders. Moreover, our findings are in line with recent research from the Jerusalem Longitudinal Cohort Study (21). In this study, 407, 661 and 1113 participants born 1920-1921 were asked how often they felt lonely with one question. Loneliness was not associated with mortality among the participants aged 70-78 (HR 1.07; 95% CI 0.54 – 2.1), 78-85 (HR 1.10; 95% CI 0.69 – 1.77) and 85-90 (HR 0.84; 95% CI 0.56 – 1.27). To our knowledge, only Patterson et al. (37) investigated the relationship between mortality and cardiovascular death. In this prospective cohort of 6789 men and women aged 21 and over, one question to measure loneliness was used. After full adjustment for confounders, frequent loneliness was not significantly associated with mortality from ischemic heart disease and other ailments of the circulatory system. The question remains whether loneliness is a (subclinical) manifestation of another condition (e.g. genetic factor, vulnerability / frailty factor or unknown condition). The main strength of our study lies in its prospective design with a long follow-up period (25 years), with multiple measurements of loneliness over time. The reliability coefficient – recorded in paired samples over 5-year intervals- of about 0.5 for loneliness is of a moderate strength. This study used a 11-item loneliness questionnaire which has been
Page 14 of 29
shown to have good validity and reliability (28). This measure is an important improvement over previous studies on the loneliness-mortality relationship that measured loneliness with three questions or even a single item. We had an almost complete mortality follow-up and included sufficient men to estimate cardiovascular risk reliably. In addition, we adjusted for many important confounders. This wide range of medical, individual and social risk factors are, in this combination, rarely available in other studies. Furthermore, we conducted a sensitivity analysis including depressive symptoms, since it is well-known that lonely individuals are more susceptible to develop depressive symptoms compared with non-lonely individuals (38, 39). There are some limitations that ought to be discussed. One limitation is the degree to which our findings may be applied to different populations, since the Zutphen Elderly Study cohort is composed of relatively healthy, white, Dutch men. This also turns out from our missing cases analysis. Hence, the lack of association between loneliness and mortality needs to be confirmed in other, more diverse populations. Second, many men included in this study have one or more missing data on loneliness or on covariates. However, we tried to impute these missing values as precisely as possible and used mixed models analyses that included all the available time points in the analyses. Although the prospective study design, the possibility of reverse causation and selective survival deserves attention. This possibility cannot be ruled out completely, although we restricted the outcome to subjects who died only after a lag period of two years. An additional limitation is the lack of data regarding depressive symptoms and disability in 1985, although our sensitivity analysis using the 1990 data as a starting point did not show any significant associations. Finally, only a small group of men were severely lonely, which resulted in wider confidence intervals. However, our sensitivity analysis using loneliness as a continuous variable did display a similar null-finding
Page 15 of 29
and the time-dependent mixed models enabled us to use multiple time-points of severe loneliness which also led to similar conclusions. In spite of the relatively new finding that loneliness is not independently associated with all-cause, cardiovascular and non-cardiovascular mortality, the point prevalence of loneliness is high and regarding earlier demonstrated relationships between loneliness, lower quality of life and adverse health outcomes, loneliness warrants attention and intervention. Possible interventions should target emotional loneliness in particular, including promotion of expertise on loneliness in health-care professionals and psychotherapy, e.g. dealing with the loss of a partner and addressing maladaptive social cognitions (40). Moreover, focus on strategies to prevent loneliness may be a promising avenue to pursue (41). Future prospective studies in the elderly are warranted to further explore the relationship between aging and increasing emotional loneliness both in men and women, to confirm this relationship in men and to demonstrate this relationship in women.
Acknowledgements The Zutphen Elderly Study was supported by the Netherlands Prevention Foundation (Preventiefonds). The funder had no role in the design, analysis, conduct or reporting for the study. The opinions, results, and conclusions reported in this paper are those of the authors and are independent from the funding sources. No other conflicts of interest declared.
Page 16 of 29
References
1. Weiss RS: Loneliness: The experience of emotional and social isolation, Cambridge, MA, MIT Press, 1973 2. de Jong-Gierveld J: Developing and testing a model of loneliness. Journal of personality and social psychology 1987; 53:119-128 3. Dykstra PA,de Jong Gierveld J: Gender and marital-history differences in emotional and social loneliness among Dutch older adults. Canadian Journal on Aging/La revue canadienne du vieillissement 2004; 23:141-155 4. Savikko N, Routasalo P, Tilvis RS, et al: Predictors and subjective causes of loneliness in an aged population. Archives of gerontology and geriatrics 2005; 41:223-233 5. Theeke LA: Predictors of loneliness in US adults over age sixtyfive. Archives of psychiatric nursing 2009; 23:387-396 6. Victor C, Scambler S, Bond J, et al: Being alone in later life: loneliness, social isolation and living alone. Reviews in Clinical Gerontology 2000; 10:407-417 7. Cacioppo JT, Hawkley LC,Thisted RA: Perceived social isolation makes me sad: 5-year cross-lagged analyses of loneliness and depressive symptomatology in the Chicago Health, Aging, and Social Relations Study. Psychology and aging 2010; 25:453 8. Luo Y, Hawkley LC, Waite LJ, et al: Loneliness, health, and mortality in old age: A national longitudinal study. Social science & medicine 2012; 74:907-914 9. Hawkley LC, Thisted RA,Cacioppo JT: Loneliness predicts reduced physical activity: cross-sectional & longitudinal analyses. Health Psychology 2009; 28:354 10. Wilson RS, Krueger KR, Arnold SE, et al: Loneliness and risk of Alzheimer disease. Archives of general psychiatry 2007; 64:234-240 11. Hawkley LC, Burleson MH, Berntson GG, et al: Loneliness in everyday life: cardiovascular activity, psychosocial context, and health behaviors. Journal of personality and social psychology 2003; 85:105
Page 17 of 29
12. Hawkley LC, Thisted RA, Masi CM, et al: Loneliness predicts increased blood pressure: 5-year cross-lagged analyses in middleaged and older adults. Psychology and aging 2010; 25:132 13. Steptoe A, Owen N, Kunz-Ebrecht SR, et al: Loneliness and neuroendocrine, cardiovascular, and inflammatory stress responses in middle-aged men and women. Psychoneuroendocrinology 2004; 29:593-611 14. Hackett RA, Hamer M, Endrighi R, et al: Loneliness and stressrelated inflammatory and neuroendocrine responses in older men and women. Psychoneuroendocrinology 2012; 37:1801-1809 15. Brown MJ,Haydock S: Pathoaetiology, epidemiology and diagnosis of hypertension. Drugs 2000; 59:1-12 16. Danesh J, Collins R, Appleby P, et al: Association of fibrinogen, Creactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. Jama 1998; 279:14771482 17. House JS, Landis KR,Umberson D: Social relationships and health. Science 1988; 241:540-545 18. Kaplan GA, Salonen JT, Cohen RD, et al: Social connections and mortality from all causes and from cardiovascular disease: prospective evidence from eastern Finland. American journal of epidemiology 1988; 128:370-380 19. Eng PM, Rimm EB, Fitzmaurice G, et al: Social ties and change in social ties in relation to subsequent total and cause-specific mortality and coronary heart disease incidence in men. American journal of epidemiology 2002; 155:700-709 20. Ellwardt L, van Tilburg T, Aartsen M, et al: Personal networks and mortality risk in older adults: a twenty-year longitudinal study. PloS one 2015; 10:e0116731 21. Stessman J, Rottenberg Y, Shimshilashvili I, et al: Loneliness, health, and longevity. The journals of gerontology. Series A, Biological sciences and medical sciences 2014; 69:744-750 22. Luo Y,Waite LJ: Loneliness and mortality among older adults in China. The journals of gerontology. Series B, Psychological sciences and social sciences 2014; 69:633-645
Page 18 of 29
23. Perissinotto CM, Cenzer IS,Covinsky KE: Loneliness in older persons: a predictor of functional decline and death. Archives of internal medicine 2012; 172:1078-1084 24. Shiovitz-Ezra S,Ayalon L: Situational versus chronic loneliness as risk factors for all-cause mortality. International Psychogeriatrics 2010; 22:455-462 25. Holwerda T, Beekman A, Deeg D, et al: Increased risk of mortality associated with social isolation in older men: only when feeling lonely? results from the Amsterdam Study of the Elderly (AMSTEL). Psychological medicine 2012; 42:843-853 26. Steptoe A, Shankar A, Demakakos P, et al: Social isolation, loneliness, and all-cause mortality in older men and women. Proceedings of the National Academy of Sciences of the United States of America 2013; 110:5797-5801 27. Feskens EJ, Bloemberg BP, Pijls LT, et al: A longitudinal study on elderly men: the Zutphen Study, in Aging, Health and Competence. Edited by Schroots JJ. Amsterdam, Elsevier Science, 1993, pp 327-333 28. Pinquart M,Sörensen S: Gender Differences in Self-Concept and Psychological Well-Being in Old Age A Meta-Analysis. The Journals of Gerontology Series B: Psychological sciences and social sciences 2001; 56:P195-P213 29. Giltay EJ, Kamphuis MH, Kalmijn S, et al: Dispositional optimism and the risk of cardiovascular death: the Zutphen Elderly Study. Archives of internal medicine 2006; 166:431-436 30. Zung WW: A self-rating depression scale. Archives of general psychiatry 1965; 12:63-70 31. Hoeymans N, Feskens EJ, van den Bos GA, et al: Measuring functional status: cross-sectional and longitudinal associations between performance and self-report (Zutphen Elderly Study 1990– 1993). Journal of clinical epidemiology 1996; 49:1103-1110 32. Caspersen CJ, Bloemberg BP, Saris WH, et al: The prevalence of selected physical activities and their relation with coronary heart disease risk factors in elderly men: the Zutphen Study, 1985. American journal of epidemiology 1991; 133:1078-1092
Page 19 of 29
33. Scharf T,de Jong Gierveld J: Loneliness in urban neighbourhoods: an Anglo-Dutch comparison. European Journal of Ageing 2008; 5:103-115 34. Longevity; gender and age, from 1861, 2011 35. De Jong Gierveld J,Van Tilburg T: Manual of the Loneliness Scale 1999. Department of Social Research Methodology, Vrije Universiteit Amsterdam, Amsterdam (updated version 18.01. 02) 1999; 36. Holmen K, Ericsson K,Winblad B: Loneliness and living conditions of the oldest old. Scandinavian journal of social medicine 1994; 22:1519 37. Patterson AC,Veenstra G: Loneliness and risk of mortality: a longitudinal investigation in Alameda County, California. Social science & medicine (1982) 2010; 71:181-186 38. Green B, Copeland J, Dewey M, et al: Risk factors for depression in elderly people: a prospective study. Acta Psychiatrica Scandinavica 1992; 86:213-217 39. Prince MJ, Harwood RH, Blizard R, et al: Social support deficits, loneliness and life events as risk factors for depression in old age. The Gospel Oak Project VI. Psychological medicine 1997; 27:323-332 40. Cattan M, White M, Bond J, et al: Preventing social isolation and loneliness among older people: a systematic review of health promotion interventions. Ageing and society 2005; 25:41-67 41. Sha'ked A,Rokach A: Addressing Loneliness: Coping, Prevention and Clinical Interventions, Psychology Press, 2015
Page 20 of 29
Figure 1. Progression of loneliness and its two subscales over time
Notes: Notes: P-value were obtained by type III tests of fixed effects for time as a continuous variable (total loneliness F = 7,737; df = 1, 1134; p = 0.006; emotional loneliness F = 71,937; df = 1, 1143; p < 0.001; social loneliness F = 15,492; df = 1, 1179; p < 0.001). The error bars represent standard errors.
Figure 2. Kaplan-Meier analyses of survival due to cardiovascular death (A) and non-cardiovascular deaths (B), according to loneliness in 719 men, aged 64 to 85 years. Data are presented with a lag period, excluding the first 2 years of observation
Page 21 of 29
Table 1 - Baseline Characteristics in 1985 in the not lonely, moderately lonely and severely lonely groups for 719 elderly men
Not Lonely
Moderately Lonely
Severely Lonely
n
(n = 417)
(n = 279 )
(n = 23)
Test
for trend
Age (yr.) – mean (SD)
719
71.8 (5.0)
71.9 (5.3)
71.5 (5.2)
F(1, 716)=0.008
0.93
Living alone - n (%)
705
25 (6.1)
40 (14.6)
12 (57.1)
2(1)=41.246
<0.001
Marital status single / divorced / widowed – n (%)
719
49 (11.8)
59 (21.1)
15 (65.2)
2(1)=35.060
<0.001
Higher education – n (%)
718
100 (24.0)
49 (17.6)
4 (17.4)
2(1)=3.995
0.046
Smokers - n (%)
692
2(1)=0.478
0.49
2 (1)=0.640
0.42
Variable
Never
73 (18.2)
49 (18.1)
3 (15.0)
Former
209 (52.1)
137 (50.6)
9 (45.0)
Current
119 (29.7)
85 (31.4)
8 (40.0)
Alcohol intake – n (%)
701
0 g/day
112 (27.5)
94 (34.7)
8 (36.4)
1-29 g/day
236 (57.8)
134 (49.4)
8 (36.4)
P- value
Page 22 of 29
60 (14.7)
43 (15.9)
6 (27.3)
692
25.6 (3.0)
25.5 (3.0)
25.1 (3.4)
F(1, 689)=0.341
0.56
Diastolic blood pressure
692
85.9 (11.8)
85.2 (10.4)
81.2 (10.5)
F(1, 689)=2.302
0.13
Systolic blood pressure
692
151.8 (21.7)
149.3 (20.7)
151.9 (20.2)
F(1, 689)=1.421
0.23
Use of antihypertensive medication – n (%)
692
87 (21.7)
66 (24.4)
2 (10.0)
2(1)=0.003
0.96
Prevalent cardiovascular disease – n (%)
693
66 (16.5)
60 (22.1)
3 (14.3)
2(1)=1.758
0.19
Any other chronic disease – n (%)
719
76 (18.2)
48 (17.2)
3 (13.0)
2(1)=0.355
0.55
688
80 (20.0)
59 (22.0)
4 (20.0)
2 1)=0.257
0.61
693
5.7 (1.6)
5.8 (1.6)
5.5 (1.8)
F(1, 690)=0.019
0.89
Physical activity, recommended level – n (%)
719
112 (26.9)
76 (27.2)
6 (26.1)
2(1)=0.002
0.96
Dispositional optimism – mean (SD)
702
6.2 (1.6)
5.6 (1.9)
4.5 (2.3)
F(1, 699)=35.031
<0.001
≥ 30 g/day 2
Body mass index (kg/m ) – mean (SD) Blood pressure (mmHg) – mean (SD)
Family history of stroke or myocardial infarction – n (%) f
Ratio serum cholesterol level (mmol/L) – mean (SD)
Notes: Data are mean ± standard deviation (SD) or n (percentage). Being moderate lonely was defined as having a loneliness score between 3 and 8 points.
Page 23 of 29
Being severe lonely was defined as having a loneliness score ≥ 9 points. Higher education is defined as higher vocational education, secondary education, high level applied training or college or university. Body mass index was calculated as weight in kilograms divided by the square of the height in meters. Prevalent cardiovascular disease was defined as having myocardial infarction, stroke or heart failure. Chronic disease was defined as having cancer, diabetes mellitus or chronic obstructive pulmonary disease (COPD). Ratio serum cholesterol was defined as total serum cholesterol divided by serum high density lipoprotein (HDL) cholesterol. Recommended level physical activity was defined as >150 minutes active per week at moderate intensity.
Page 24 of 29
Table 2. Risk of mortality according to loneliness in 719 elderly men during 25 years of follow-up PNot
Moderately
Severely
Lonely
Lonely
Lonely
(n=417)
(n=279)
(n=23)
value Test
for Tren d
Total mortality: Cases, no (%)
396 (95.0)
265 (95.0)
22 (95.7)
Model 1
1.00
1.05 (0.90 – 1.22)
1.34 (0.87 – 2.06)
Wald (1)=1.210
0.27
Model 2
1.00
1.03 (0.88 – 1.22)
1.32 (0.82 – 2.10)
Wald (1)=0.761
0.38
Model 3
1.00
0.99 (0.83 – 1.17)
1.21 (0.72 – 2.01)
Wald (1)=0.027
0.87
Model 4
1.00
1.00 (0.84 – 1.17)
1.40 (0.85 – 2.31)
Wald (1)=0.213
0.65
186 (44.6)
123 (44.1)
12 (52.2)
Model 1
1.00
1.03 (0.82 – 1.29)
1.52 (0.84 – 2.72)
Wald (1)=0.743
0.39
Model 2
1.00
0.99 (0.78 – 1.25)
1.64 (0.91 – 2.98)
Wald (1)=0.435
0.51
2
2
2
2
CVD mortality: Cases, no. (%)
2
2
Page 25 of 29
Model 3
1.00
0.92 (0.72 – 1.18)
1.33 (0.69 – 2.55)
Wald (1)=0.020
0.89
Model 4
1.00
0.99 (0.78 – 1.25)
1.18 (0.58 – 2.39)
Wald (1)=0.010
0.92
210 (50.4)
142 (50.9)
10 (43.5)
Model 1
1.00
1.06 (0.86 – 1.32)
1.18 (0.63 – 2.23)
Wald (1)=0.488
0.49
Model 2
1.00
1.08 (0.86 – 1.35)
0.99 (0.46 – 2.12)
Wald (1)=0.287
0.59
Model 3
1.00
1.05 (0.83 – 1.33)
0.99 (0.43 – 2.29)
Wald (1)=0.120
0.73
Model 4
1.00
0.99 (0.79 – 1.24)
1.63 (0.80 – 3.31)
Wald (1)=0.152
0.70
2
2
Non-CVD mortality: Cases, no. (%)
2
2
2
2
Notes: Data are Hazard Ratio’s (with 95% confidence intervals). Being moderate lonely was defined as having a loneliness score between 3 and 8 points. Being severe lonely was defined as having a loneliness score ≥ 9 points. P-value was obtained by an analysis of variance for trend, linear term. Model 1: Crude Model 2: Adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model 3: Additionally adjusted for dispositional optimism, family history of stroke of myocardial infarction, serum cholesterol, blood pressure, history of cardiovascular disease, having another chronic disease and use of anti-hypertensive medication. Model 4: Same adjustment as model 3, but with time-dependent loneliness, dispositional optimism, smoking habits, alcohol use, physical activity, BMI, serum cholesterol, use of anti-hypertensive medication and blood pressure updated every 5 years.
Page 26 of 29
Table 3. Risk of mortality according to emotional and social loneliness in 719 elderly men during 25 years of follow-up Emotional
PP-
Loneliness
Social
Test value
Loneliness
Test
value
Total mortality: Model 1
1.11 (1.05 – 1.16)
Wald (1)=15.248
<0.001
1.00 (0.95 – 1.04)
Wald (1)=0.034
0.85
Model 2
1.08 (1.02 – 1.14)
Wald (1)=7.825
0.01
1.01 (0.96 – 1.05)
Wald (1)=0.037
0.85
Model 3
1.05 (0.99 – 1.11)
Wald (1)=2.334
0.13
1.00 (0.95 – 1.05)
Wald (1)=0.009
0.93
Model 4
1.04 (0.97 – 1.11)
Wald (1)=1.250
0.26
Wald (1)=0.843
0.36
2
2
2
2
1.03 (0.97 – 1.09)
2
2
2
2
CVD mortality: Model 1
1.10 (1.03 – 1.19)
Wald (1)=7.003
0.01
1.00 (0.94 – 1.06)
Wald (1)=0.003
0.96
Model 2
1.08 (1.00 – 1.17)
Wald (1)=4.227
0.04
1.00 (0.93 – 1.07)
Wald (1)=0.001
0.98
Model 3
1.03 (0.95 – 1.12)
Wald (1)=0.484
0.49
0.98 (0.92 – 1.06)
Wald (1)=0.222
0.64
Model 4
1.02 (0.93 – 1.12)
Wald (1)=0.246
0.62
1.02 (0.93 – 1.10)
Wald (1)=0.119
0.73
2
2
2
2
2
2
2
2
Non-CVD mortality:
Page 27 of 29
Model 1
1.11 (1.03 – 1.19)
Wald (1)=8.252
0.004
0.99 (0.94 – 1.06)
Wald (1)=0.041
0.84
Model 2
1.08 (1.00 – 1.16)
Wald (1)=3.534
0.06
1.01 (0.95 – 1.08)
Wald (1)=0.077
0.78
Model 3
1.07 (0.98 – 1.16)
Wald (1)=2.123
0.15
1.01 (0.94 – 1.08)
Wald (1)=0.077
0.78
Model 4
1.05 (0.95 – 1.15)
Wald (1)=0.880
0.35
1.04 (0.96 – 1.12)
Wald (1)=0.714
0.40
2
2
2
2
2
2
2
2
Notes: Data are Hazard Ratio’s (with 95% confidence intervals). Being moderate lonely was defined as having a loneliness score between 3 and 8 points. Being severe lonely was defined as having a loneliness score ≥ 9 points. P-value was obtained by an analysis of variance for trend, linear term. Model 1: Crude Model 2: Adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model 3: Additionally adjusted for dispositional optimism, family history of stroke of myocardial infarction, serum cholesterol, blood pressure, history of cardiovascular disease, having another chronic disease and use of anti-hypertensive medication. Model 4: Same adjustment as model 3, but with time-dependent loneliness, dispositional optimism, smoking habits, alcohol use, physical activity, BMI, serum cholesterol, use of anti-hypertensive medication and blood pressure updated every 5 years.
Page 28 of 29
Page 29 of 29
S1064-7481(16)00150-0 http://dx.doi.org/doi: 10.1016/j.jagp.2016.01.136 AMGP 568
To appear in:
The American Journal of Geriatric Psychiatry
Received date: Revised date: Accepted date:
25-3-2015 22-1-2016 22-1-2016
Please cite this article as: Jolien E. Julsing, Daan Kromhout, Johanna M. Geleijnse, Erik J. Giltay, Loneliness and All-Cause, Cardiovascular and Non-Cardiovascular Mortality in Older Men: The Zutphen Elderly Study, The American Journal of Geriatric Psychiatry (2016), http://dx.doi.org/doi: 10.1016/j.jagp.2016.01.136. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Word count text: 3663
Loneliness and all-cause, cardiovascular and non-cardiovascular mortality in older men: The Zutphen Elderly Study 1
2
2
Jolien E. Julsing , M.D., Daan Kromhout , Ph.D., M.P.H., Johanna M. Geleijnse , Ph.D., and Erik 1
J. Giltay , Ph.D., M.D. 1
Leiden University Medical Centre, Department of Psychiatry, Leiden, the Netherlands
2
Division of Human Nutrition, Wageningen University, Wageningen, The Netherlands
Correspondence to: J.E. Julsing, M.D. Leiden University Medical Centre, Department of Psychiatry B1-P Postbus 9600 2300 RC Leiden, the Netherlands Telephone: 00316 44444383 Fax: 003171 5248156 Email: [email protected]
The Zutphen Elderly Study was supported by the Netherlands Prevention Foundation (Preventiefonds)
Key words: Loneliness, mortality, aged, cardiovascular death, cohort studies
Page 1 of 29
Abstract Objectives - Loneliness, defined as the discrepancy between one’s desired and actual relationships, is prevalent in the elderly and can be distinguished in emotional and social loneliness. We aimed to determine whether loneliness is independently related to higher allcause, cardiovascular and non-cardiovascular mortality in elderly men. Design - A population-based cohort study with 25 years of follow-up from 1985 onward. Setting - The Zutphen Study started as the Dutch contribution to the Seven Countries Study. Participants - Seven hundred nineteen (76.2%) of 939 men (age range 64-84 years), who had complete data on loneliness at baseline and at least 2 years of survival. Measurements - Loneliness was assessed using a validated 11-item questionnaire in 1985, 1990, 1995 and 2000. Time-dependent Cox proportional hazards models were adjusted for socio-demographic characteristics and cardiovascular risk factors. Results - At baseline, point prevalence of moderate and severe loneliness was respectively 38.8% (n=279) and 3.2% (n=23). Loneliness –especially emotional loneliness– did significantly increase over 15 years with an overall reliability coefficient of 0.50. All-cause, cardiovascular, and non-cardiovascular mortality were not higher among moderately lonely participants (Hazard ratio [HR] 1.00; 95% confidence interval (CI) 0.84-1.17, HR 0.99; 95% CI 0.78-1.25, HR 0.99; 95% CI 0.79-1.24, respectively) and severely lonely participants (HR 1.40; 95% CI 0.85-2.31, HR 1.18; 95%CI 0.58-2.39, HR 1.63; 95% CI 0.80-3.31 respectively). Conclusions - Loneliness is common and increases during aging, due to the increase in its component emotional loneliness over time. No independent associations with risks of allcause, cardiovascular and non-cardiovascular death were found.
Page 2 of 29
Key words - Loneliness, mortality, aged, cardiovascular death, cohort studies
Page 3 of 29
Objective Loneliness is a complex concept involving psychological and social aspects, and is an important indicator of human well-being (1). It is the subjective feeling of isolation, explained as the discrepancy between one’s desired and actual relationships (2), and differs from several measures of social isolation such as living alone, marital status and size of social network. Thus, a person may suffer from loneliness even though he or she is surrounded by many people, and, conversely, persons who live alone may not feel lonely. Weiss (1) distinguished two types of loneliness: emotional and social loneliness. Emotional loneliness is caused by the absence of one or more intimate figures, whereas social loneliness refers to the absence of a social network with whom common interests and activities can be shared. When studying older subjects, emotional loneliness may be particularly relevant since they are more likely to suffer from bereavement. In general, these losses are more salient for men as they rely more on their spouses while forming less close and enduring relationships outside marriage (3). Although loneliness seems to increase with age (4), loneliness may level off at advanced age and does not depend on age any more (5). Nevertheless, loneliness emerges as a prevalent phenomenon with 20 to 40% of older adults reporting feeling lonely at any given time(4, 5). Of them, five to seven percent report feeling intense or persistent loneliness(6). Feelings of loneliness are associated with a large number of health-related problems. These include depressive symptoms (7), poor functional health (8), reductions in physical activity (9), and impaired cognition (10). At the biological level, loneliness has been linked with increased vascular resistance (11), increased systolic blood pressure (12), and markers of low-grade inflammation (13, 14). This could ultimately make lonely people susceptible to a variety of illness, and especially cardiovascular disease (15, 16).
Page 4 of 29
An increasing body of research focused on the hypothesis that loneliness is a risk factor for overall mortality. Where social isolation and low social support have consistently been associated with increased mortality (17-19), the concept of loneliness is only starting to be recognized as a separate entity from social isolation and therefore fewer studies have examined it as an independent risk factor for mortality. Previous prospective studies investigating the effects of loneliness on mortality in the elderly included men as well as women aged above 50. Follow-up time differed between 6 and 20 years. Except for one study that used the 11-item the Jong Gierveld Loneliness Scale (20), being lonely was assessed rather crudely, with a 3-item or even a single item scale (“do you feel lonely?”). Moreover, apart from three studies (20-22), loneliness was obtained only at baseline and not sequentially during follow-up. Furthermore, these studies focused on all-cause mortality, differed in controlling for potentially confounding factors, and have yielded mixed results. Where some studies found an association between loneliness and all-cause mortality (23-25), another one did not (21) or the association disappeared after multivariable adjustment (8, 20, 26). Three of these studies reported significant sex differences in these findings, with a higher mortality risk for lonely men (8, 22, 25). Therefore, the independent association with cardiovascular mortality has not yet been determined. Because of strong links that have been found between loneliness on the one hand and physical activity, optimism and depressive symptoms, having a partner, and somatic factors like blood pressure and chronic disease on the other hand, it is important to study whether a putative association with mortality is confounded by one of these factors. The purpose of this study was to investigate whether loneliness is an independent predictor for higher all-cause, cardiovascular and non-cardiovascular mortality in older men.
Page 5 of 29
Methods Participants The data for this study were collected as a part of The Zutphen Elderly Study, which is a prospective population-based study among older male inhabitants of Zutphen, a small industrial town with 47,000 inhabitants in the Netherlands. The study characteristics of the cohort have previously been described in detail (27). Briefly, in 1985 a random sample of elderly men in Zutphen born in 1900-1920 was asked to participate. There were a priori no selection criteria for other demographic or health characteristics. Participants were visited for face-to-face interviews between March 1 and June 30 in 1985 (response rate 70.1%), 1990 (response rate 78.0%), 1995 (response rate 74.1%), and 2000 (response rate 72.8%) and were followed for causes of mortality for up to 25 years until 2010. Informed consent was obtained from all study participants. The study was approved by the Medical Ethics Committee of the Leiden University Medical Centre, and in 1995 and 2000, by the Medical Ethics Committee of The Netherlands Organization for Applied Scientific Research (TNO). Of the 939 subjects who participated in the study in 1985, 153 subjects (16.3%) were excluded due to missing information on the loneliness scores. To reduce bias from reverse causation, we included only the men with at least 2 years of survival. Thus, analyses were based on 719 subjects. Compared at baseline, the 719 participating subjects were significantly younger, had a higher Body Mass Index (BMI), had a higher level of education, were more physical active and had less chronic or other cardiovascular diseases than the 153 excluded subjects (data not shown).
Disease ascertainment Information on the vital status of the participants until July 1, 2010, was obtained from municipal population registries. The causes of death were obtained from Statistics
Page 6 of 29
Netherlands and verified with information obtained from the hospital and the general practitioners in Zutphen. The final coding was done by an experienced clinical epidemiologist. Because it is often difficult to determine the underlying cause of death in elderly people, the primary and secondary causes of death were included in the analyses. Cardiovascular disease was defined as codes 390 to 459 according to the International Classification of Diseases, Ninth Revision (ICD-9). The presence of a prevalent cardiovascular disease at baseline, defined as having either myocardial infarction, stroke or heart failure, and other chronic diseases, defined as having either diabetes mellitus, cancer or chronic obstructive pulmonary disease (COPD), was determined by a survey questionnaire and confirmed with hospital discharge data and written information from the general practitioner. Only three men were lost to follow-up and were included in the analyses, but censored after 4.2, 5.8 and 5.1 years of follow-up.
Loneliness Feelings of loneliness were assessed in 1985, 1990, 1995 and 2000 using the Loneliness Scale of de Jong Gierveld, an 11-item questionnaire consisting of two subscales: the emotional loneliness scale and the social loneliness scale (2). The loneliness scale comprises five positive stated items (e.g. “There are enough people that I feel close to”) and six negative stated items (e.g. “I miss having a really close friend”), and ranges from 0-11 points, with higher scores being indicative of higher loneliness levels. The total loneliness score was calculated by assigning a score of one point to the responses indicating a (certain) feeling of loneliness. The emotional loneliness score was the result of the sum of the positively or neutral stated answers on six items, whereas the social loneliness score was the sum of the negatively and neutral stated answers on its five corresponding items. If a respondent had one missing value, the missing item was imputed with the median of the
Page 7 of 29
remaining items of the relevant subscale for that particular subject. The 11-item De Jong Gierveld Loneliness Scale has proved to be a valid and reliable instrument for the assessment of overall, emotional, and social loneliness (28).
Correlates of loneliness In 1985, 1990, 1995 and 2000, dispositional optimism was assessed using a questionnaire consisting of four items on a 3-point scale (29) from a survey of Statistics Netherlands (Centraal Bureau voor de Statistiek – CBS-). The additional answer category “do not know” was coded as the midpoint (score 1). If a respondent had one missing value, the missing item was imputed with the mean of the remaining three items for that particular subject. The optimism questionnaire score ranges from 0 to 8 points, with higher scores being indicative of higher optimism levels. The presence of depressive symptoms was assessed in 1990, 1995 and 2000 using the Dutch translation of the Zung self-rating depression scale, consisting of 20 items on a 4-point scale (30). Finally, the presence of disability was assessed in 1990, 1995 and 2000 with the 13-item Activities of Daily Living questionnaire (31). The following severity levels of disability status were distinguished: no disability (0); mild disability, disability in instrumental activities only (1); moderate disability, disability in instrumental activities and mobility (2); and severe disability, disability in instrumental activities, mobility, and basic activities of daily living (3).
Covariates Information on living arrangement, education, family history of stroke or myocardial infarction, and physical activity was obtained with a questionnaire that participants were asked to complete at home. A validated questionnaire on physical activity designed for retired men was used to calculate the total minutes spent in physical activity per week (of an
Page 8 of 29
intensity of >2kcal/kg per hour) (32) and coded along recommended level or not. Cigarette smoking (never, former, or current smoker) and alcohol use (0, 1-29, or ≥ 30 g/d of alcohol) were coded along three levels. Information on alcohol consumption was collected with the cross-check dietary history method. In 1985, 1990, 1995 and 2000, research assessments were performed by five trained physicians and included body mass index (BMI), blood pressure measurements (in duplicate), and medical history. The use of antihypertensive medication was also assessed. Total and high density lipoprotein (HDL) cholesterol levels were analysed enzymatically in a standardized lipid laboratory.
Statistical analyses The baseline characteristics of the participants were compared between not lonely, moderately lonely and severely lonely participants (respectively defined with a cut-off score between 3-8 and ≥9 on the De Jong Gierveld Loneliness Scale) (33) using the chi-square test (linear by linear term) and one-way analysis of variance (for linear trend), when appropriate. Results are presented as numbers and percentages for categorical variables, and means and standard deviations for continuous variables. Linear mixed models were used to assess the progression of loneliness and its two subscales over 15 years of follow-up. Reliability coefficients (i.e., an intraclass correlation coefficient with a two-way random effects model with single-measure reliability) were used to examine temporal stability of loneliness over the four time points in subjects with paired loneliness scores (meaning that a subject has a non-missing loneliness score in both mentioned time-points). Age of death according to loneliness was assessed by conducting analysis of covariance (ANCOVA) controlling for socio-demographic characteristics and cardiovascular
Page 9 of 29
risk factors. When a participant was alive at the end of the follow-up period, age of death was imputed with the mean remaining life expectancy for men of the same age group, for that particular subject according to Statistics Netherlands (34). The Kaplan-Meier method was used to present the rate of cardiovascular and noncardiovascular mortality in the not lonely, moderately lonely and severely lonely categories. After testing the proportional hazard assumption, hazard ratios (HR) with 95% confidence intervals (CI) of the three types of mortality were estimated by Cox proportional hazards models. Four multivariable models were tested in 1985. Model one was crude, model two adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model three additionally adjusted for dispositional optimism, serum total cholesterol, systolic and diastolic blood pressure, history of having a cardiovascular disease, having another chronic disease, family history of stroke or myocardial infarction and use of anti-hypertensive medication. Model four used the same adjustment factors, but included all covariates if possible as time-dependent variables that were updated using the mean of all previous measurements. If a subsequent score was missing, the missing score was imputed with the mean score of all previous measured scores for that particular subject. We repeated the four Cox proportional hazards models with emotional (ranging from 0 to 6 points) and social (ranging from 0 to 5 points) loneliness on a continuous scale as independent variables. To test whether an association between loneliness and death could be confounded by depressive symptoms or disability, we carried out a sensitivity analysis using 1990 as the starting date and additionally adjusted for Zung self-rating depression score and disability. All p-values are two tailed and considered statistically significant at p < 0.05. Data analyses were performed using SPSS for windows, version 20.0 (Chicago, IL, USA).
Page 10 of 29
Results Baseline characteristics Table 1 shows the baseline characteristics of 719 men according to the presence of subjective feelings of loneliness. Loneliness was present in 302 (42.0%) participants, of whom 279 (92.4%) were moderately lonely and 23 (7.6%) were severely lonely. Loneliness was positively associated with living alone, not being married, lower education and lower baseline dispositional optimism.
Loneliness in time Figure 1 shows a statistically significant increase of loneliness over time (F(1, 1134) = 7.737, p = 0.006). As for the loneliness subscales, we observed a significant increase in the emotional loneliness scores (F(1, 1143)=71.937, p < 0.001) and a statistically significant decrease (F(1, 1179)=15.492, p < 0.001) in social loneliness during follow-up. Among participants who provided paired loneliness scores, the reliability coefficient for loneliness was 0.55 between 1985 and 1990 (single measure; F(1, 441) = 3,398; n=442 pairs), 0.51 between 1990 and 1995 (single measure; F(1, 273) = 3,060; n=274 pairs), and 0.59 between 1995 and 2000 (single measure; F(1, 146) = 3,859; n=147 pairs) (p < 0.001 for all). The overall reliability coefficient was 0.50 (single measure over 4 time points; F(1, 137) = 4.984, p < 0.001; n=138 with complete data).
Loneliness and mortality
Page 11 of 29
During the 25-year follow-up, 683 (95.0%) of 719 men died. Mean age at death was 83.3 years (standard deviation [SD] 6.3). Of these 719 men, 321 (44.6%) died of cardiovascular causes. After controlling for socio-demographic characteristics and cardiovascular risk factors, age of death was similar between not lonely, moderately lonely, and severely lonely men (analysis of covariance, F = 1.083; df = 2, 620; p = 0.34). Kaplan-Meier analysis of survival according to the three categories of loneliness showed no evidence that moderately lonely and severely lonely men had higher cardiovascular or non-cardiovascular mortality rates (log rank test: = 1.967; df = 2; p = 0.37; and log rank test: = 0.508; df = 2; p = 0.78, respectively) compared to not lonely men in 1985 (Figure 2). Table 2 presents the results of the Cox proportional hazard analyses detailing crude and adjusted associations of loneliness with all-cause mortality, cardiovascular mortality and non-cardiovascular mortality. Compared with not lonely men, those who were moderately lonely and severely lonely showed in all four models no significant different hazard ratio’s for all three types of mortality. Table 3 summarizes the mortality risks according to emotional and social loneliness. The HR adjusted for age, demographic factors and lifestyle factors for emotional loneliness and cardiovascular mortality was significant (HR 1.08, 95% CI 1.00 – 1.17, Wald test: 2(1)=4.227; p = 0.04), but was attenuated and no longer significant when cardiovascular risk
factors and dispositional optimism were included in the model and remained so after timedependent emotional loneliness was included in the model. Similar associations were found for all-cause and non-cardiovascular mortality. All results depicting the mortality risks according to social loneliness were non-statistically significant.
Page 12 of 29
We carried out two sensitivity analyses (data not shown). First, a sensitivity analysis using 1990 as the starting date and additionally adjusted for Zung self-rating depression score and disability. The second sensitivity analysis used loneliness as a continuous variable, ranging from 0 through 11. In both sensitivity analyses and both the adjusted and timedependent models, no significant associations were found between loneliness on the one hand and overall, cardiovascular and non-cardiovascular mortality on the other hand..
Conclusions This prospective cohort study examined whether loneliness is related to a higher all-cause, cardiovascular and non-cardiovascular mortality. Although our results demonstrate loneliness as a common entity in older men that increases during aging, no independent associations between loneliness and risks of all-cause, cardiovascular and non-cardiovascular death were found. The percentages of moderately lonely and severely lonely men in this sample were comparable with those reported in the research program of Living arrangements and Social Networks of older adults (LSN) using the same measure (35). Our results that loneliness increases with age are mostly in line with other studies (4, 5, 36). Furthermore, this study demonstrates that social loneliness declines with older age, where emotional loneliness further extends. This is a remarkable finding and we are not aware of other studies regarding this subject. Aging men may be more prone to emotional loneliness because of the increasing risk of bereavement of the death of a partner or close life-long friends. At the same time, loss of these key figures may stimulate older men to come in contact with other people, hence our observed decline in social loneliness. Our observation that loneliness is not independently associated with all-cause mortality stands in contrast to other studies, which suggest that loneliness in older people is
Page 13 of 29
not only a source of great suffering but also contributes to declining longevity (22-25). However, in a number of studies the effect was no longer observed after adjustment for demographic factors (e.g. marital status and education), health indicators (e.g. chronic disease, depression, physical activity and smoking) (8, 20, 26, 37). The most recent study included 2911 elder subjects, with six follow-up measurements covering a time span of twenty years, during which loneliness was also measured with the 11-item De Jong Gierveld Loneliness Scale. Similarly, neither emotional, nor social loneliness were independently associated with mortality once mental health was added to their model. This and our findings suggest a potential pathway from loneliness to mortality through the effects of mental disorders. Moreover, our findings are in line with recent research from the Jerusalem Longitudinal Cohort Study (21). In this study, 407, 661 and 1113 participants born 1920-1921 were asked how often they felt lonely with one question. Loneliness was not associated with mortality among the participants aged 70-78 (HR 1.07; 95% CI 0.54 – 2.1), 78-85 (HR 1.10; 95% CI 0.69 – 1.77) and 85-90 (HR 0.84; 95% CI 0.56 – 1.27). To our knowledge, only Patterson et al. (37) investigated the relationship between mortality and cardiovascular death. In this prospective cohort of 6789 men and women aged 21 and over, one question to measure loneliness was used. After full adjustment for confounders, frequent loneliness was not significantly associated with mortality from ischemic heart disease and other ailments of the circulatory system. The question remains whether loneliness is a (subclinical) manifestation of another condition (e.g. genetic factor, vulnerability / frailty factor or unknown condition). The main strength of our study lies in its prospective design with a long follow-up period (25 years), with multiple measurements of loneliness over time. The reliability coefficient – recorded in paired samples over 5-year intervals- of about 0.5 for loneliness is of a moderate strength. This study used a 11-item loneliness questionnaire which has been
Page 14 of 29
shown to have good validity and reliability (28). This measure is an important improvement over previous studies on the loneliness-mortality relationship that measured loneliness with three questions or even a single item. We had an almost complete mortality follow-up and included sufficient men to estimate cardiovascular risk reliably. In addition, we adjusted for many important confounders. This wide range of medical, individual and social risk factors are, in this combination, rarely available in other studies. Furthermore, we conducted a sensitivity analysis including depressive symptoms, since it is well-known that lonely individuals are more susceptible to develop depressive symptoms compared with non-lonely individuals (38, 39). There are some limitations that ought to be discussed. One limitation is the degree to which our findings may be applied to different populations, since the Zutphen Elderly Study cohort is composed of relatively healthy, white, Dutch men. This also turns out from our missing cases analysis. Hence, the lack of association between loneliness and mortality needs to be confirmed in other, more diverse populations. Second, many men included in this study have one or more missing data on loneliness or on covariates. However, we tried to impute these missing values as precisely as possible and used mixed models analyses that included all the available time points in the analyses. Although the prospective study design, the possibility of reverse causation and selective survival deserves attention. This possibility cannot be ruled out completely, although we restricted the outcome to subjects who died only after a lag period of two years. An additional limitation is the lack of data regarding depressive symptoms and disability in 1985, although our sensitivity analysis using the 1990 data as a starting point did not show any significant associations. Finally, only a small group of men were severely lonely, which resulted in wider confidence intervals. However, our sensitivity analysis using loneliness as a continuous variable did display a similar null-finding
Page 15 of 29
and the time-dependent mixed models enabled us to use multiple time-points of severe loneliness which also led to similar conclusions. In spite of the relatively new finding that loneliness is not independently associated with all-cause, cardiovascular and non-cardiovascular mortality, the point prevalence of loneliness is high and regarding earlier demonstrated relationships between loneliness, lower quality of life and adverse health outcomes, loneliness warrants attention and intervention. Possible interventions should target emotional loneliness in particular, including promotion of expertise on loneliness in health-care professionals and psychotherapy, e.g. dealing with the loss of a partner and addressing maladaptive social cognitions (40). Moreover, focus on strategies to prevent loneliness may be a promising avenue to pursue (41). Future prospective studies in the elderly are warranted to further explore the relationship between aging and increasing emotional loneliness both in men and women, to confirm this relationship in men and to demonstrate this relationship in women.
Acknowledgements The Zutphen Elderly Study was supported by the Netherlands Prevention Foundation (Preventiefonds). The funder had no role in the design, analysis, conduct or reporting for the study. The opinions, results, and conclusions reported in this paper are those of the authors and are independent from the funding sources. No other conflicts of interest declared.
Page 16 of 29
References
1. Weiss RS: Loneliness: The experience of emotional and social isolation, Cambridge, MA, MIT Press, 1973 2. de Jong-Gierveld J: Developing and testing a model of loneliness. Journal of personality and social psychology 1987; 53:119-128 3. Dykstra PA,de Jong Gierveld J: Gender and marital-history differences in emotional and social loneliness among Dutch older adults. Canadian Journal on Aging/La revue canadienne du vieillissement 2004; 23:141-155 4. Savikko N, Routasalo P, Tilvis RS, et al: Predictors and subjective causes of loneliness in an aged population. Archives of gerontology and geriatrics 2005; 41:223-233 5. Theeke LA: Predictors of loneliness in US adults over age sixtyfive. Archives of psychiatric nursing 2009; 23:387-396 6. Victor C, Scambler S, Bond J, et al: Being alone in later life: loneliness, social isolation and living alone. Reviews in Clinical Gerontology 2000; 10:407-417 7. Cacioppo JT, Hawkley LC,Thisted RA: Perceived social isolation makes me sad: 5-year cross-lagged analyses of loneliness and depressive symptomatology in the Chicago Health, Aging, and Social Relations Study. Psychology and aging 2010; 25:453 8. Luo Y, Hawkley LC, Waite LJ, et al: Loneliness, health, and mortality in old age: A national longitudinal study. Social science & medicine 2012; 74:907-914 9. Hawkley LC, Thisted RA,Cacioppo JT: Loneliness predicts reduced physical activity: cross-sectional & longitudinal analyses. Health Psychology 2009; 28:354 10. Wilson RS, Krueger KR, Arnold SE, et al: Loneliness and risk of Alzheimer disease. Archives of general psychiatry 2007; 64:234-240 11. Hawkley LC, Burleson MH, Berntson GG, et al: Loneliness in everyday life: cardiovascular activity, psychosocial context, and health behaviors. Journal of personality and social psychology 2003; 85:105
Page 17 of 29
12. Hawkley LC, Thisted RA, Masi CM, et al: Loneliness predicts increased blood pressure: 5-year cross-lagged analyses in middleaged and older adults. Psychology and aging 2010; 25:132 13. Steptoe A, Owen N, Kunz-Ebrecht SR, et al: Loneliness and neuroendocrine, cardiovascular, and inflammatory stress responses in middle-aged men and women. Psychoneuroendocrinology 2004; 29:593-611 14. Hackett RA, Hamer M, Endrighi R, et al: Loneliness and stressrelated inflammatory and neuroendocrine responses in older men and women. Psychoneuroendocrinology 2012; 37:1801-1809 15. Brown MJ,Haydock S: Pathoaetiology, epidemiology and diagnosis of hypertension. Drugs 2000; 59:1-12 16. Danesh J, Collins R, Appleby P, et al: Association of fibrinogen, Creactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. Jama 1998; 279:14771482 17. House JS, Landis KR,Umberson D: Social relationships and health. Science 1988; 241:540-545 18. Kaplan GA, Salonen JT, Cohen RD, et al: Social connections and mortality from all causes and from cardiovascular disease: prospective evidence from eastern Finland. American journal of epidemiology 1988; 128:370-380 19. Eng PM, Rimm EB, Fitzmaurice G, et al: Social ties and change in social ties in relation to subsequent total and cause-specific mortality and coronary heart disease incidence in men. American journal of epidemiology 2002; 155:700-709 20. Ellwardt L, van Tilburg T, Aartsen M, et al: Personal networks and mortality risk in older adults: a twenty-year longitudinal study. PloS one 2015; 10:e0116731 21. Stessman J, Rottenberg Y, Shimshilashvili I, et al: Loneliness, health, and longevity. The journals of gerontology. Series A, Biological sciences and medical sciences 2014; 69:744-750 22. Luo Y,Waite LJ: Loneliness and mortality among older adults in China. The journals of gerontology. Series B, Psychological sciences and social sciences 2014; 69:633-645
Page 18 of 29
23. Perissinotto CM, Cenzer IS,Covinsky KE: Loneliness in older persons: a predictor of functional decline and death. Archives of internal medicine 2012; 172:1078-1084 24. Shiovitz-Ezra S,Ayalon L: Situational versus chronic loneliness as risk factors for all-cause mortality. International Psychogeriatrics 2010; 22:455-462 25. Holwerda T, Beekman A, Deeg D, et al: Increased risk of mortality associated with social isolation in older men: only when feeling lonely? results from the Amsterdam Study of the Elderly (AMSTEL). Psychological medicine 2012; 42:843-853 26. Steptoe A, Shankar A, Demakakos P, et al: Social isolation, loneliness, and all-cause mortality in older men and women. Proceedings of the National Academy of Sciences of the United States of America 2013; 110:5797-5801 27. Feskens EJ, Bloemberg BP, Pijls LT, et al: A longitudinal study on elderly men: the Zutphen Study, in Aging, Health and Competence. Edited by Schroots JJ. Amsterdam, Elsevier Science, 1993, pp 327-333 28. Pinquart M,Sörensen S: Gender Differences in Self-Concept and Psychological Well-Being in Old Age A Meta-Analysis. The Journals of Gerontology Series B: Psychological sciences and social sciences 2001; 56:P195-P213 29. Giltay EJ, Kamphuis MH, Kalmijn S, et al: Dispositional optimism and the risk of cardiovascular death: the Zutphen Elderly Study. Archives of internal medicine 2006; 166:431-436 30. Zung WW: A self-rating depression scale. Archives of general psychiatry 1965; 12:63-70 31. Hoeymans N, Feskens EJ, van den Bos GA, et al: Measuring functional status: cross-sectional and longitudinal associations between performance and self-report (Zutphen Elderly Study 1990– 1993). Journal of clinical epidemiology 1996; 49:1103-1110 32. Caspersen CJ, Bloemberg BP, Saris WH, et al: The prevalence of selected physical activities and their relation with coronary heart disease risk factors in elderly men: the Zutphen Study, 1985. American journal of epidemiology 1991; 133:1078-1092
Page 19 of 29
33. Scharf T,de Jong Gierveld J: Loneliness in urban neighbourhoods: an Anglo-Dutch comparison. European Journal of Ageing 2008; 5:103-115 34. Longevity; gender and age, from 1861, 2011 35. De Jong Gierveld J,Van Tilburg T: Manual of the Loneliness Scale 1999. Department of Social Research Methodology, Vrije Universiteit Amsterdam, Amsterdam (updated version 18.01. 02) 1999; 36. Holmen K, Ericsson K,Winblad B: Loneliness and living conditions of the oldest old. Scandinavian journal of social medicine 1994; 22:1519 37. Patterson AC,Veenstra G: Loneliness and risk of mortality: a longitudinal investigation in Alameda County, California. Social science & medicine (1982) 2010; 71:181-186 38. Green B, Copeland J, Dewey M, et al: Risk factors for depression in elderly people: a prospective study. Acta Psychiatrica Scandinavica 1992; 86:213-217 39. Prince MJ, Harwood RH, Blizard R, et al: Social support deficits, loneliness and life events as risk factors for depression in old age. The Gospel Oak Project VI. Psychological medicine 1997; 27:323-332 40. Cattan M, White M, Bond J, et al: Preventing social isolation and loneliness among older people: a systematic review of health promotion interventions. Ageing and society 2005; 25:41-67 41. Sha'ked A,Rokach A: Addressing Loneliness: Coping, Prevention and Clinical Interventions, Psychology Press, 2015
Page 20 of 29
Figure 1. Progression of loneliness and its two subscales over time
Notes: Notes: P-value were obtained by type III tests of fixed effects for time as a continuous variable (total loneliness F = 7,737; df = 1, 1134; p = 0.006; emotional loneliness F = 71,937; df = 1, 1143; p < 0.001; social loneliness F = 15,492; df = 1, 1179; p < 0.001). The error bars represent standard errors.
Figure 2. Kaplan-Meier analyses of survival due to cardiovascular death (A) and non-cardiovascular deaths (B), according to loneliness in 719 men, aged 64 to 85 years. Data are presented with a lag period, excluding the first 2 years of observation
Page 21 of 29
Table 1 - Baseline Characteristics in 1985 in the not lonely, moderately lonely and severely lonely groups for 719 elderly men
Not Lonely
Moderately Lonely
Severely Lonely
n
(n = 417)
(n = 279 )
(n = 23)
Test
for trend
Age (yr.) – mean (SD)
719
71.8 (5.0)
71.9 (5.3)
71.5 (5.2)
F(1, 716)=0.008
0.93
Living alone - n (%)
705
25 (6.1)
40 (14.6)
12 (57.1)
2(1)=41.246
<0.001
Marital status single / divorced / widowed – n (%)
719
49 (11.8)
59 (21.1)
15 (65.2)
2(1)=35.060
<0.001
Higher education – n (%)
718
100 (24.0)
49 (17.6)
4 (17.4)
2(1)=3.995
0.046
Smokers - n (%)
692
2(1)=0.478
0.49
2 (1)=0.640
0.42
Variable
Never
73 (18.2)
49 (18.1)
3 (15.0)
Former
209 (52.1)
137 (50.6)
9 (45.0)
Current
119 (29.7)
85 (31.4)
8 (40.0)
Alcohol intake – n (%)
701
0 g/day
112 (27.5)
94 (34.7)
8 (36.4)
1-29 g/day
236 (57.8)
134 (49.4)
8 (36.4)
P- value
Page 22 of 29
60 (14.7)
43 (15.9)
6 (27.3)
692
25.6 (3.0)
25.5 (3.0)
25.1 (3.4)
F(1, 689)=0.341
0.56
Diastolic blood pressure
692
85.9 (11.8)
85.2 (10.4)
81.2 (10.5)
F(1, 689)=2.302
0.13
Systolic blood pressure
692
151.8 (21.7)
149.3 (20.7)
151.9 (20.2)
F(1, 689)=1.421
0.23
Use of antihypertensive medication – n (%)
692
87 (21.7)
66 (24.4)
2 (10.0)
2(1)=0.003
0.96
Prevalent cardiovascular disease – n (%)
693
66 (16.5)
60 (22.1)
3 (14.3)
2(1)=1.758
0.19
Any other chronic disease – n (%)
719
76 (18.2)
48 (17.2)
3 (13.0)
2(1)=0.355
0.55
688
80 (20.0)
59 (22.0)
4 (20.0)
2 1)=0.257
0.61
693
5.7 (1.6)
5.8 (1.6)
5.5 (1.8)
F(1, 690)=0.019
0.89
Physical activity, recommended level – n (%)
719
112 (26.9)
76 (27.2)
6 (26.1)
2(1)=0.002
0.96
Dispositional optimism – mean (SD)
702
6.2 (1.6)
5.6 (1.9)
4.5 (2.3)
F(1, 699)=35.031
<0.001
≥ 30 g/day 2
Body mass index (kg/m ) – mean (SD) Blood pressure (mmHg) – mean (SD)
Family history of stroke or myocardial infarction – n (%) f
Ratio serum cholesterol level (mmol/L) – mean (SD)
Notes: Data are mean ± standard deviation (SD) or n (percentage). Being moderate lonely was defined as having a loneliness score between 3 and 8 points.
Page 23 of 29
Being severe lonely was defined as having a loneliness score ≥ 9 points. Higher education is defined as higher vocational education, secondary education, high level applied training or college or university. Body mass index was calculated as weight in kilograms divided by the square of the height in meters. Prevalent cardiovascular disease was defined as having myocardial infarction, stroke or heart failure. Chronic disease was defined as having cancer, diabetes mellitus or chronic obstructive pulmonary disease (COPD). Ratio serum cholesterol was defined as total serum cholesterol divided by serum high density lipoprotein (HDL) cholesterol. Recommended level physical activity was defined as >150 minutes active per week at moderate intensity.
Page 24 of 29
Table 2. Risk of mortality according to loneliness in 719 elderly men during 25 years of follow-up PNot
Moderately
Severely
Lonely
Lonely
Lonely
(n=417)
(n=279)
(n=23)
value Test
for Tren d
Total mortality: Cases, no (%)
396 (95.0)
265 (95.0)
22 (95.7)
Model 1
1.00
1.05 (0.90 – 1.22)
1.34 (0.87 – 2.06)
Wald (1)=1.210
0.27
Model 2
1.00
1.03 (0.88 – 1.22)
1.32 (0.82 – 2.10)
Wald (1)=0.761
0.38
Model 3
1.00
0.99 (0.83 – 1.17)
1.21 (0.72 – 2.01)
Wald (1)=0.027
0.87
Model 4
1.00
1.00 (0.84 – 1.17)
1.40 (0.85 – 2.31)
Wald (1)=0.213
0.65
186 (44.6)
123 (44.1)
12 (52.2)
Model 1
1.00
1.03 (0.82 – 1.29)
1.52 (0.84 – 2.72)
Wald (1)=0.743
0.39
Model 2
1.00
0.99 (0.78 – 1.25)
1.64 (0.91 – 2.98)
Wald (1)=0.435
0.51
2
2
2
2
CVD mortality: Cases, no. (%)
2
2
Page 25 of 29
Model 3
1.00
0.92 (0.72 – 1.18)
1.33 (0.69 – 2.55)
Wald (1)=0.020
0.89
Model 4
1.00
0.99 (0.78 – 1.25)
1.18 (0.58 – 2.39)
Wald (1)=0.010
0.92
210 (50.4)
142 (50.9)
10 (43.5)
Model 1
1.00
1.06 (0.86 – 1.32)
1.18 (0.63 – 2.23)
Wald (1)=0.488
0.49
Model 2
1.00
1.08 (0.86 – 1.35)
0.99 (0.46 – 2.12)
Wald (1)=0.287
0.59
Model 3
1.00
1.05 (0.83 – 1.33)
0.99 (0.43 – 2.29)
Wald (1)=0.120
0.73
Model 4
1.00
0.99 (0.79 – 1.24)
1.63 (0.80 – 3.31)
Wald (1)=0.152
0.70
2
2
Non-CVD mortality: Cases, no. (%)
2
2
2
2
Notes: Data are Hazard Ratio’s (with 95% confidence intervals). Being moderate lonely was defined as having a loneliness score between 3 and 8 points. Being severe lonely was defined as having a loneliness score ≥ 9 points. P-value was obtained by an analysis of variance for trend, linear term. Model 1: Crude Model 2: Adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model 3: Additionally adjusted for dispositional optimism, family history of stroke of myocardial infarction, serum cholesterol, blood pressure, history of cardiovascular disease, having another chronic disease and use of anti-hypertensive medication. Model 4: Same adjustment as model 3, but with time-dependent loneliness, dispositional optimism, smoking habits, alcohol use, physical activity, BMI, serum cholesterol, use of anti-hypertensive medication and blood pressure updated every 5 years.
Page 26 of 29
Table 3. Risk of mortality according to emotional and social loneliness in 719 elderly men during 25 years of follow-up Emotional
PP-
Loneliness
Social
Test value
Loneliness
Test
value
Total mortality: Model 1
1.11 (1.05 – 1.16)
Wald (1)=15.248
<0.001
1.00 (0.95 – 1.04)
Wald (1)=0.034
0.85
Model 2
1.08 (1.02 – 1.14)
Wald (1)=7.825
0.01
1.01 (0.96 – 1.05)
Wald (1)=0.037
0.85
Model 3
1.05 (0.99 – 1.11)
Wald (1)=2.334
0.13
1.00 (0.95 – 1.05)
Wald (1)=0.009
0.93
Model 4
1.04 (0.97 – 1.11)
Wald (1)=1.250
0.26
Wald (1)=0.843
0.36
2
2
2
2
1.03 (0.97 – 1.09)
2
2
2
2
CVD mortality: Model 1
1.10 (1.03 – 1.19)
Wald (1)=7.003
0.01
1.00 (0.94 – 1.06)
Wald (1)=0.003
0.96
Model 2
1.08 (1.00 – 1.17)
Wald (1)=4.227
0.04
1.00 (0.93 – 1.07)
Wald (1)=0.001
0.98
Model 3
1.03 (0.95 – 1.12)
Wald (1)=0.484
0.49
0.98 (0.92 – 1.06)
Wald (1)=0.222
0.64
Model 4
1.02 (0.93 – 1.12)
Wald (1)=0.246
0.62
1.02 (0.93 – 1.10)
Wald (1)=0.119
0.73
2
2
2
2
2
2
2
2
Non-CVD mortality:
Page 27 of 29
Model 1
1.11 (1.03 – 1.19)
Wald (1)=8.252
0.004
0.99 (0.94 – 1.06)
Wald (1)=0.041
0.84
Model 2
1.08 (1.00 – 1.16)
Wald (1)=3.534
0.06
1.01 (0.95 – 1.08)
Wald (1)=0.077
0.78
Model 3
1.07 (0.98 – 1.16)
Wald (1)=2.123
0.15
1.01 (0.94 – 1.08)
Wald (1)=0.077
0.78
Model 4
1.05 (0.95 – 1.15)
Wald (1)=0.880
0.35
1.04 (0.96 – 1.12)
Wald (1)=0.714
0.40
2
2
2
2
2
2
2
2
Notes: Data are Hazard Ratio’s (with 95% confidence intervals). Being moderate lonely was defined as having a loneliness score between 3 and 8 points. Being severe lonely was defined as having a loneliness score ≥ 9 points. P-value was obtained by an analysis of variance for trend, linear term. Model 1: Crude Model 2: Adjusted for age, smoking habits, alcohol use, physical activity, educational level and BMI. Model 3: Additionally adjusted for dispositional optimism, family history of stroke of myocardial infarction, serum cholesterol, blood pressure, history of cardiovascular disease, having another chronic disease and use of anti-hypertensive medication. Model 4: Same adjustment as model 3, but with time-dependent loneliness, dispositional optimism, smoking habits, alcohol use, physical activity, BMI, serum cholesterol, use of anti-hypertensive medication and blood pressure updated every 5 years.
Page 28 of 29
Page 29 of 29