Loneliness and objectively measured physical capability in middle-aged adults

Loneliness and objectively measured physical capability in middle-aged adults Abstract Background Loneliness is associated with poor functional ability in older people. Little is known about this association in the middle-aged. The aim is to investigate if perceived loneliness is associated with lower physical capability among middle-aged men and women and if the associations of loneliness with physical capability interact with socioeconomic position and cohabitation status. Methods 5224 participants from Copenhagen Aging and Midlife Biobank (CAMB) aged 49–62 years (mean age 54) were included. Handgrip strength (measured by a dynamometer) and maximal number of chair rises in 30 s was recorded. Multivariate linear regression analyses were adjusted for age, occupational social class, cohabitation status, morbidity and personality traits. Results No association was found between loneliness and physical capability. For example estimates for handgrip strength in ‘often’ lonely men and women compared with the ‘not lonely’ were 1.2 kg (95% CI − 0.5;2.9)/1.0 kg (−0.7;2.6). Low occupational social class was associated with poorer physical capability, and living alone was associated with poorer handgrip strength in men [−2.4 kg (95% CI − 3.2;−1.5)] and poorer chair rise test in women [−0.8 rises (95% CI − 1.6;−0.1)]. There was no support for interactions. Conclusion In contrast to earlier studies among older people, no association between loneliness and physical capability was found in this cohort of middle-aged men and women. Loneliness may not yet have resulted in detectable differences in physical capability in this age group. Further research is needed to clarify if, and at what point in the life course loneliness begins to affect physical capability. Introduction Loneliness has been described as “the social equivalent of physical pain, hunger and thirst”1 and can be described as a feeling of being socially isolated, which is often accompanied by a perceived deficiency in quantity and quality of social relations.1 Loneliness has been associated with adverse health outcomes, including mortality, depression, cardiovascular disease, metabolic syndrome.2,3 Loneliness has also been associated with functional decline in older people,4,5 but little is known about this association in middle-aged adults. In the Northern European countries the prevalence of loneliness in the 60+ years old is around 6% which is lower compared with Southern (11%) and Eastern European countries (20%).6 The prevalence of loneliness furthermore varies with age,7 gender,7 socioeconomic position,7–9 living arrangements,7 health10 and personality.1 Suggested pathways through which loneliness accelerates physical ageing and contributes to poor health11 include: higher total peripheral resistance and blood pressure (1), higher concentration of epinephrine,11 dysregulation of the inflammatory and neuroendocrine responses,12 poorer self-reported sleep quality11 A clear social gradient in physical capability among middle-aged adults has been suggested,13 and furthermore that loneliness is associated with lower socioeconomic position.8,9 According to the concept of differential vulnerability the effect of a given risk factor (here loneliness) is intensified along the social strata.14 Similarly, living alone has been associated with poorer physical capability especially among older males15 and higher levels of loneliness are experienced by those living alone.16 The aim of this study is to investigate (1) if perceived loneliness is related to lower levels of physical capability evaluated using objective measures, including handgrip strength and number of chair rises, in a middle-aged population, and (2) if the associations of loneliness on physical capability are different in respondents with a high or low socioeconomic position and for those living alone versus those not living alone. Methods Study population The Copenhagen Aging and Midlife Biobank (CAMB) is based on three longitudinal studies: the Metropolit 1953 Danish Male Birth Cohort (MP), The Copenhagen Perinatal Cohort (CPC) born 1959–61 and The Danish Longitudinal Study on Work, Unemployment and Health (DALWUH) born 1949 or 1959. In total, 5575 persons answered the questionnaire and participated in the clinical examination (response rate 30%). Compared with non-respondents, CAMB participants were more educated and were less likely to be unemployed and experienced significantly higher all-cause mortality.17 Exclusion criteria Participants with missing information on loneliness, included covariates, job information for social class classification were excluded as were participants reporting to be a student or a homemaker (0.5% of the participants). There were no specific exclusion criteria for the handgrip test, but participants with systolic >160 and/or diastolic blood pressure >100 were not asked to complete the chair rise test.13 In total, 351 (6%) were excluded due to missing values in the analyses on hand grip strength (N = 5224) and 833 (15%) in the analysis on chair rise test (N = 4742). See figure 1 in Supplementary files for a flow chart of the study population. Loneliness, occupational social class and cohabitation status Loneliness (unwanted aloneness) was measured by a single item: “Are you ever alone, when you would much rather be with others?” Response categories were “Yes, often”,” Yes, sometimes”,” Yes, but seldom” and” No”, where “No” was used as the reference. For the analyses investigating the potential differential vulnerability, loneliness was dichotomized into either lonely (‘often’ or ‘sometimes lonely’) or not lonely (‘seldom’ or ‘not lonely’). Occupational social class was classified based on occupation in accordance with the Danish Occupational Social Class Measurement into occupational social class I–VI.18 Occupational social class I–V represent employed individuals ranging from professional occupation in social class I, to unskilled occupation in social class V. Social class VI represents people on transfer income, including unwaged persons and those receiving social security benefits or sickness benefits. Occupational social class was also dichotomized into high (social class I, II and III) and low (IV, V and VI). Loneliness and occupational social class/cohabitation status were combined into two joint variables with the following four levels; “lonely/low social class or living alone”, “lonely/high social class or not living alone”, “not lonely/low social class or living alone” and “not lonely/high social class or not living alone” (reference group). Physical capability Handgrip strength was measured with a dynamometer (model G100, Biometrics Ltd, Newport, UK) with the best maximal handgrip strength in kg recorded out of three to five attempts.13 During a 30-s chair rise test, the maximal number of chair rises was recorded (chair with a seat height of 45 cm and the seat wired to a computer providing information on mechanical contact).13 Covariates Potential confounders were identified based on prior knowledge and the method of directed acyclic graphs (see Supplementary file).19 Covariates included age, gender, occupational social class, cohabitation status, number of chronic physical and mental disorders and personality as they have been suggested as potential causes of both loneliness1,7–10 and physical capability/physical function.13,20 Loneliness is associated with high neuroticism and low conscientiousness and agreeableness,1 and both conscientiousness and neuroticism are associated with physical functioning and physical activity.21 Depressive symptoms can act both as a mediator and a potential confounder and was therefore included in a separate model.16 Body mass index (BMI), alcohol intake, smoking habits, physical activity were all identified as potential mediators of the relation between loneliness and physical functioning and they were added in a separate model. Gender and age were assessed from social security numbers given to all Danish citizens. The CAMB participants were asked to declare the following medical conditions: (1) allergies, (2) diabetes, (3) hypertension, (4) myocardial infarction, (5) stroke, (6) lung disease (incl. asthma), chronic bronchitis, emphysema, (7) autoimmune disease and arthritis (8) cancer incl. leukaemia and (9) other mental disorder than depression. These conditions were compiled to a morbidity index counting the number of medical conditions reported (0–9). Personality was assessed with the short Danish version of the NEO Five Factor Inventory.21,22 Depressive symptoms were assessed using the Major Depression Inventory with a score ranging from 0 to 50.23 For detailed description of measurement of BMI, alcohol intake, smoking habits, physical activity please see Supplementary file. Statistical methods The associations between loneliness and handgrip strength/number of chair rises were analysed with multiple linear regression, PROC GLM, using SAS 9.3. Data were stratified by gender because the chair rise test was non-normally distributed in the combined data. Plots of the residuals for the association between hand grip strength and chair rise, respectively, as well as loneliness were judged symmetric. Potential interaction (differential vulnerability) was assessed by analysing the association between joint variables of loneliness and cohabitation status/occupational social class, respectively, in regards to physical capability, as well as by the inclusion of product terms between loneliness and occupational social class and cohabitation status, respectively, in the regression analyses. Initially, crude multivariate regression models were run followed by a model adjusted for the identified confounders (Model 1). In Model 2, depressive symptoms at baseline were added, and in Model 3 health behaviour variables were added. In the joint effect analyses, the main effects of each of the joint effect variables (loneliness, occupational social class and cohabitation status) were not included. Weights were not used in this study as this method builds on the understanding that those who do in fact participate (given their measured variables) are representative of those who do not. This is difficult to validate and hence does not solve the problem of selection bias. Results Population demographics Sixteen per cent of the population reported ‘often’ or ‘sometimes’ feeling lonely, 3% often felt lonely. Approximately one-third of the population was men (n = 3596 men and n = 1628 women), with no gender differences in loneliness. Age ranged from 48 to 62 years, and mean age did not differ between the four categories of loneliness among women but differed slightly for men (in the ‘often lonely’ mean age was 1 year lower). The association between loneliness and each of the covariates showed a dose-dependent pattern except for alcohol intake. More men and women who often felt lonely, were in the lowest occupational social class group VI compared with the other groups. Fifteen per cent of both men and women were living alone. Among men in the often lonely group 61% were living alone, for often lonely women 34% was living alone. Respondents scoring high on loneliness were in general more likely to suffer from three or more chronic diseases, to be depressed, obese or being smokers, to be physical active less than 3–4 h/week (stronger for men) and to be either non-drinkers or heavy drinkers (men only). The often lonely scored higher on neuroticism and lower on conscientiousness (tables 1 and 2). Table 1 Men: distribution of demographic, personality, health and health behaviour variables by loneliness (N = 3596)   Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  110 (3%)  460 (13%)  1095 (31%)  1931 (54%)    Age, mean (years)  54 (SD 3.5)  55 (SD 3.2)  55 (SD 3.3)  55 (SD 3.2)  0.0083  Cohabitation status  <0.0001  Alone  67 (61%)  174 (38%)  193 (18%)  91 (5%)    Not alone  43 (39%)  286 (62%)  902 (82%)  1840 (95%)    Socioeconomic position  <0.0001  Social class I  12 (11%)  71 (15%)  204 (19%)  385 (20%)    Social class II  15 (14%)  89 (19%)  255 (23%)  549 (28%)    Social class III  23 (21%)  98 (21%)  244 (22%)  466 (24%)    Social class IV  15 (14%)  72 (16%)  178 (16%)  269 (14%)    Social class V  6 (6%)  51 (11%)  90 (8%)  166 (9%)    Social class VI  39 (36%)  79 (17%)  124 (11%)  96 (5%)    Number of chronic illnessesa  <0.0001  0  29 (26%)  252 (31%)  411 (38%)  763 (40%)    1–2  59 (54%)  252 (55%)  557 (51%)  998 (52%)    ≥3  22 (20%)  67 (15%)  127 (11%)  170 (9%)    Depressionb  <0.0001  Yes  27 (25%)  34 (7%)  36 (3%)  27 (1%)    No  83 (75%)  426 (93%)  1059 (97%)  1904 (99%)    Personality  Neuroticism  Median 25  Median 19  Median 17  Median 15  <0.0001  Q1c 17  Q1 15  Q1 13  Q1 11    Q3c 30  Q3 25  Q3 22  Q3 18    Agreeableness  Median 32  Median 32  Median 32  Median 33  <0.0001  Q1 28  Q1 29  Q1 29  Q1 30    Q3 35  Q3 35  Q3 36  Q3 37    Conscientiousness  Median 31  Median 32  Median 33  Median 35  <0.0001  Q1 25  Q1 28  Q1 30  Q1 31    Q3 35  Q3 35  Q3 37  Q3 38    BMI  0.0022  Underweight (<18.5)  2 (2%)  2 (1%)  6 (1%)  5 (1%)    Normal (18.5–25)  41 (37%)  183 (40%)  391 (36%)  727 (38%)    Overweight (25–30)  38 (35%)  204 (44%)  513 (47%)  929 (48%)    Obese (>30)  29 (26%)  71 (15%)  185 (17%)  270 (14%)    Smoking status  <0.0001  Smoker  42 (38%)  138 (30%)  256 (23%)  400 (21%)    Non-smoker  68 (62%)  322 (70%)  839 (77%)  1531 (79%)    Alcohol  <0.0001  Not drinking  19 (17%)  54 (12%)  90 (8%)  141 (7%)    0 units/week  50 (45%)  223 (48%)  562 (51%)  1109 (57%)    Low risk drinking  12 (11%)  75 (16%)  211 (19%)  313 (16%)    14 units/weekd  29 (26%)  108 (24%)  232 (21%)  368 (19%)    Elevated risk drinking  15–21 units/weekd            High risk drinking  >21 units/weekd            Physical activitye  0.0312  ≤3–4 h/week  48 (44%)  177 (38%)  366 (33%)  645 (33%)    ≥5 h/week  62 (56%)  283 (62%)  729 (67%)  1286 (67%)      Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  110 (3%)  460 (13%)  1095 (31%)  1931 (54%)    Age, mean (years)  54 (SD 3.5)  55 (SD 3.2)  55 (SD 3.3)  55 (SD 3.2)  0.0083  Cohabitation status  <0.0001  Alone  67 (61%)  174 (38%)  193 (18%)  91 (5%)    Not alone  43 (39%)  286 (62%)  902 (82%)  1840 (95%)    Socioeconomic position  <0.0001  Social class I  12 (11%)  71 (15%)  204 (19%)  385 (20%)    Social class II  15 (14%)  89 (19%)  255 (23%)  549 (28%)    Social class III  23 (21%)  98 (21%)  244 (22%)  466 (24%)    Social class IV  15 (14%)  72 (16%)  178 (16%)  269 (14%)    Social class V  6 (6%)  51 (11%)  90 (8%)  166 (9%)    Social class VI  39 (36%)  79 (17%)  124 (11%)  96 (5%)    Number of chronic illnessesa  <0.0001  0  29 (26%)  252 (31%)  411 (38%)  763 (40%)    1–2  59 (54%)  252 (55%)  557 (51%)  998 (52%)    ≥3  22 (20%)  67 (15%)  127 (11%)  170 (9%)    Depressionb  <0.0001  Yes  27 (25%)  34 (7%)  36 (3%)  27 (1%)    No  83 (75%)  426 (93%)  1059 (97%)  1904 (99%)    Personality  Neuroticism  Median 25  Median 19  Median 17  Median 15  <0.0001  Q1c 17  Q1 15  Q1 13  Q1 11    Q3c 30  Q3 25  Q3 22  Q3 18    Agreeableness  Median 32  Median 32  Median 32  Median 33  <0.0001  Q1 28  Q1 29  Q1 29  Q1 30    Q3 35  Q3 35  Q3 36  Q3 37    Conscientiousness  Median 31  Median 32  Median 33  Median 35  <0.0001  Q1 25  Q1 28  Q1 30  Q1 31    Q3 35  Q3 35  Q3 37  Q3 38    BMI  0.0022  Underweight (<18.5)  2 (2%)  2 (1%)  6 (1%)  5 (1%)    Normal (18.5–25)  41 (37%)  183 (40%)  391 (36%)  727 (38%)    Overweight (25–30)  38 (35%)  204 (44%)  513 (47%)  929 (48%)    Obese (>30)  29 (26%)  71 (15%)  185 (17%)  270 (14%)    Smoking status  <0.0001  Smoker  42 (38%)  138 (30%)  256 (23%)  400 (21%)    Non-smoker  68 (62%)  322 (70%)  839 (77%)  1531 (79%)    Alcohol  <0.0001  Not drinking  19 (17%)  54 (12%)  90 (8%)  141 (7%)    0 units/week  50 (45%)  223 (48%)  562 (51%)  1109 (57%)    Low risk drinking  12 (11%)  75 (16%)  211 (19%)  313 (16%)    14 units/weekd  29 (26%)  108 (24%)  232 (21%)  368 (19%)    Elevated risk drinking  15–21 units/weekd            High risk drinking  >21 units/weekd            Physical activitye  0.0312  ≤3–4 h/week  48 (44%)  177 (38%)  366 (33%)  645 (33%)    ≥5 h/week  62 (56%)  283 (62%)  729 (67%)  1286 (67%)    a Number of chronic illnesses includes: (1) Allergies, (2) Diabetes, (3) Hypertension, (4) Myocardial infarction, (5) Stroke, (6) Lung disease (incl. asthma, chronic bronchitis, emphysema) , (7) Autoimmune disease and arthritis, (8) Cancer incl. leukaemia and, (9) present other mental disorder. b Depression according to Major Depression Inventory-scale (41;42) ‘Yes’ includes moderate and severe depression and ‘no’ includes no and mild depression. c Q1: lower quartile, Q3: upper quartile. d Low-risk recommendations 1–14 units/week for men. Elevated risk drinking 15–21 units/week for men. High-risk drinking >21 for men. e Physical activity includes sports, physical training, house- and garden work, walks and bicycle rides incl. between home and work. Table 2 Women: distribution of demographic, personality, health and health behaviour variables by loneliness (N = 1628)   Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  44 (3%)  232 (14%)  509 (31%)  843 (52%)    Age, mean (years)  53 (SD 4.6)  52 (SD 4.1)  53 (SD 4.2)  53 (SD 4.4)  0.77  Cohabitation status  <0.0001  Alone  15 (34%)  86 (37%)  96 (19%)  63 (7%)    Not alone  29 (66%)  146 (63%)  413 (81%)  780 (93%)    Socio-economic position  <0.0001  Social class I  1 (2%)  13 (6%)  53 (10%)  95 (11%)    Social class II  8 (18%)  64 (28%)  152 (30%)  240 (28%)    Social class III  4 (9%)  49 (21%)  110 (22%)  234 (28%)    Social class IV  6 (14%)  45 (19%)  109 (21%)  170 (20%)    Social class V  4 (9%)  18 (8%)  41 (8%)  68 (8%)    Social class VI  21 (48%)  43 (19%)  44 (9%)  36 (4%)    Number of chronic illnessesa  <0.0001  0  7 (16%)  85 (37%)  192 (38%)  332 (39%)    1–2  14 (57%)  107 (46%)  270 (53%)  462 (55%)    ≥3  12 (27%)  40 (17%)  47 (9%)  49 (6%)    Depressionb  <0.0001  Yes  18 (41%)  30 (13%)  18 (4%)  17 (2%)    No  26 (59%)  202 (87%)  491 (96%)  826 (98%)    Personality  Neuroticism  Median 28  Median 23  Median 19  Median 17  <0.0001  Q1c 22  Q1 18  Q1 15  Q1 13    Q3c 31.5  Q3 27.5  Q3 24  Q3 21    Agreeableness  Median 35  Median 35  Median 35  Median 35  0.08  Q1 30  Q1 31  Q1 32  Q1 32    Q3 39  Q3 38  Q3 38  Q3 39    Conscientiousness  Median 30  Median 32  Median 34  Median 35  <0.0001  Q1 26  Q1 28  Q1 30  Q1 32    Q3 35  Q3 36  Q3 37  Q3 38    BMI  0.20  Underweight (<18.5)  1 (2%)  6 (3%)  8 (2%)  23 (3%)    Normal (18.5–25)  18 (41%)  119 (51%)  292 (57%)  440 (52%)    Overweight (25–30)  15 (34%)  71 (31%)  145 (28%)  278 (33%)    Obese (>30)  10 (23%)  36 (15%)  64 (13%)  102 (12%)    Smoking status  0.01  Smoker  17 (39%)  60 (26%)  116 (23%)  168 (20%)    Non-smoker  27 (61%)  172 (74%)  393 (77%)  675 (80%)    Alcohol  <0.0001  Not drinking  21 (48%)  47 (20%)  81 (16%)  137 (16%)    0 units/week  11 (25%)  104 (45%)  231 (45%)  374 (44%)    Low-risk drinking  7 (16%)  53 (23%)  142 (28%)  219 (26%)    7 units/weekd  5 (11%)  28 (12%)  55 (11%)  113 (13%)    Elevated risk drinking  8–14 units/weekd            High-risk drinking  >14 units/weekd            Physical activitye  0.1116  ≤3–4 h/week  13 (30%)  69 (30%)  113 (22%)  225 (27%)    ≥5 h/week  31 (70%)  163 (70%)  396 (78%)  618 (73%)      Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  44 (3%)  232 (14%)  509 (31%)  843 (52%)    Age, mean (years)  53 (SD 4.6)  52 (SD 4.1)  53 (SD 4.2)  53 (SD 4.4)  0.77  Cohabitation status  <0.0001  Alone  15 (34%)  86 (37%)  96 (19%)  63 (7%)    Not alone  29 (66%)  146 (63%)  413 (81%)  780 (93%)    Socio-economic position  <0.0001  Social class I  1 (2%)  13 (6%)  53 (10%)  95 (11%)    Social class II  8 (18%)  64 (28%)  152 (30%)  240 (28%)    Social class III  4 (9%)  49 (21%)  110 (22%)  234 (28%)    Social class IV  6 (14%)  45 (19%)  109 (21%)  170 (20%)    Social class V  4 (9%)  18 (8%)  41 (8%)  68 (8%)    Social class VI  21 (48%)  43 (19%)  44 (9%)  36 (4%)    Number of chronic illnessesa  <0.0001  0  7 (16%)  85 (37%)  192 (38%)  332 (39%)    1–2  14 (57%)  107 (46%)  270 (53%)  462 (55%)    ≥3  12 (27%)  40 (17%)  47 (9%)  49 (6%)    Depressionb  <0.0001  Yes  18 (41%)  30 (13%)  18 (4%)  17 (2%)    No  26 (59%)  202 (87%)  491 (96%)  826 (98%)    Personality  Neuroticism  Median 28  Median 23  Median 19  Median 17  <0.0001  Q1c 22  Q1 18  Q1 15  Q1 13    Q3c 31.5  Q3 27.5  Q3 24  Q3 21    Agreeableness  Median 35  Median 35  Median 35  Median 35  0.08  Q1 30  Q1 31  Q1 32  Q1 32    Q3 39  Q3 38  Q3 38  Q3 39    Conscientiousness  Median 30  Median 32  Median 34  Median 35  <0.0001  Q1 26  Q1 28  Q1 30  Q1 32    Q3 35  Q3 36  Q3 37  Q3 38    BMI  0.20  Underweight (<18.5)  1 (2%)  6 (3%)  8 (2%)  23 (3%)    Normal (18.5–25)  18 (41%)  119 (51%)  292 (57%)  440 (52%)    Overweight (25–30)  15 (34%)  71 (31%)  145 (28%)  278 (33%)    Obese (>30)  10 (23%)  36 (15%)  64 (13%)  102 (12%)    Smoking status  0.01  Smoker  17 (39%)  60 (26%)  116 (23%)  168 (20%)    Non-smoker  27 (61%)  172 (74%)  393 (77%)  675 (80%)    Alcohol  <0.0001  Not drinking  21 (48%)  47 (20%)  81 (16%)  137 (16%)    0 units/week  11 (25%)  104 (45%)  231 (45%)  374 (44%)    Low-risk drinking  7 (16%)  53 (23%)  142 (28%)  219 (26%)    7 units/weekd  5 (11%)  28 (12%)  55 (11%)  113 (13%)    Elevated risk drinking  8–14 units/weekd            High-risk drinking  >14 units/weekd            Physical activitye  0.1116  ≤3–4 h/week  13 (30%)  69 (30%)  113 (22%)  225 (27%)    ≥5 h/week  31 (70%)  163 (70%)  396 (78%)  618 (73%)    a Number of chronic illnesses includes: (1) Allergies, (2) Diabetes, (3) Hypertension, (4) Myocardial infarction, (5) Stroke, (6) Lung disease (incl. asthma, chronic bronchitis, emphysema), (7) Autoimmune disease and arthritis, (8) Cancer incl. leukaemia and (9) present other mental disorder. b Depression according to Major Depression Inventory-scale (41;42); ‘yes’ includes moderate and severe depression and ‘no’ includes no and mild depression. c Q1: lower quartile, Q3: upper quartile. d Low-risk recommendations 1–7 units/week or women. Elevated risk drinking 8–14 units/week for women. High-risk drinking >14 for women. e Physical activity includes sports, physical training, house- and garden work, walks and bicycle rides incl. between home and work. Loneliness and physical capability In men (crude models), loneliness was associated with lower handgrip strength [‘Often lonely’ −1.3 kg (95% CI −2.9;0.3) ‘Sometimes lonely’ −2.0 (−2.8;−1.1) reference group ‘Not lonely’] and for both genders with fewer chair raises [‘Often lonely’ women −3.3 rises/30 s (95% CI −5.5; −1.5, ‘Often lonely’ men −0.8(−2.0;0.4) ‘Sometimes lonely’ men −1.1 (−1.7;−0.5)] (table 3). After adjustment for confounders (Model 1), the observed associations attenuated. In women, there were no clear associations between loneliness and handgrip strength in any of the models (table 3). Adjusting for depressive symptoms in Model 2 and BMI, smoking status, alcohol consumption and physical activity in Model 3, did not change these conclusions (see Supplementary file). Table 3 Multivariate linear regression analysis of loneliness and handgrip strength/chair rise test, crude and adjusted models Handgrip strength  Chair rise test  Loneliness  N  Mean (kg)  Estimate (kg)  95% CI  Loneliness  N  Mean (rises/30 s)  Estimate (rises/30 s)  95% CI  Men  Crude model  Often lonely  110  48.9  −1.3  (−2.9;0.3)  Often lonely  89  21.1  −0.8  (−2.0;0.4)  Sometimes lonely  460  48.2  −2.0  (−2.8;−1.1)  Sometimes lonely  403  20.7  −1.1  (−1.7;−0.5)  Seldom lonely  1095  49.2  −0.9    Seldom lonely  986  21.5  −0.4  (−0.8;0.1)  Not lonely  1931  50.2  Ref.  (−1.5;−0.3)  Not lonely  1720  21.9  Ref.    P-value        <0.0001          0.004  Adjusted model  Often lonely  110  47.9  1.2  (−0.5;2.9)  Often lonely  89  20.0  0.2  (−1.0;1.4)  Sometimes lonely  460  46.2  −0.4  (−1.3;0.5)  Sometimes lonely  403  19.4  −0.3  (−0.9;0.3)  Seldom lonely  1095  46.4  −0.3  (−0.9;0.4)  Seldom lonely  986  19.7  −0.1  (−0.5;0.4)  Not lonely  1931  46.7  Ref.    Not lonely  1720  19.8  Ref.    P-value        0.23          0.77  Women  Crude model  Often lonely  44  30.8  −0.8  (−2.5;0.8)  Often lonely  43  18.3  −3.3  (−5.0; −1.5)  Sometimes lonely  232  31.2  −0.4  (−1.2;0.4)  Sometimes lonely  222  20.1  −1.5  (−2.3; −0.6)  Seldom lonely  509  31.1  −0.6  (−1.2;0.0)  Seldom lonely  489  21.3  −0.3  (−0.9;0.4)  Not lonely  843  31.6  Ref.    Not lonely  790  21.6  Ref.    P-value  0.25    <.0001  Model 1  Often lonely  44  31.3  1.0  (−0.7;2.6)  Often lonely  43  20.9  −0.3  (−2.0;1.4)  Sometimes lonely  232  30.6  0.3  (−0.6;1.1)  Sometimes lonely  222  20.9  −0.3  (−1.1;0.6)  Seldom lonely  509  30.0  −0.4  (−1.0;0.2)  Seldom lonely  489  21.2  0.1  (−0.5;0.7)  Not lonely  843  30.4  Ref.    Not lonely  790  21.2  Ref.    P-value  0.21    0.79  Handgrip strength  Chair rise test  Loneliness  N  Mean (kg)  Estimate (kg)  95% CI  Loneliness  N  Mean (rises/30 s)  Estimate (rises/30 s)  95% CI  Men  Crude model  Often lonely  110  48.9  −1.3  (−2.9;0.3)  Often lonely  89  21.1  −0.8  (−2.0;0.4)  Sometimes lonely  460  48.2  −2.0  (−2.8;−1.1)  Sometimes lonely  403  20.7  −1.1  (−1.7;−0.5)  Seldom lonely  1095  49.2  −0.9    Seldom lonely  986  21.5  −0.4  (−0.8;0.1)  Not lonely  1931  50.2  Ref.  (−1.5;−0.3)  Not lonely  1720  21.9  Ref.    P-value        <0.0001          0.004  Adjusted model  Often lonely  110  47.9  1.2  (−0.5;2.9)  Often lonely  89  20.0  0.2  (−1.0;1.4)  Sometimes lonely  460  46.2  −0.4  (−1.3;0.5)  Sometimes lonely  403  19.4  −0.3  (−0.9;0.3)  Seldom lonely  1095  46.4  −0.3  (−0.9;0.4)  Seldom lonely  986  19.7  −0.1  (−0.5;0.4)  Not lonely  1931  46.7  Ref.    Not lonely  1720  19.8  Ref.    P-value        0.23          0.77  Women  Crude model  Often lonely  44  30.8  −0.8  (−2.5;0.8)  Often lonely  43  18.3  −3.3  (−5.0; −1.5)  Sometimes lonely  232  31.2  −0.4  (−1.2;0.4)  Sometimes lonely  222  20.1  −1.5  (−2.3; −0.6)  Seldom lonely  509  31.1  −0.6  (−1.2;0.0)  Seldom lonely  489  21.3  −0.3  (−0.9;0.4)  Not lonely  843  31.6  Ref.    Not lonely  790  21.6  Ref.    P-value  0.25    <.0001  Model 1  Often lonely  44  31.3  1.0  (−0.7;2.6)  Often lonely  43  20.9  −0.3  (−2.0;1.4)  Sometimes lonely  232  30.6  0.3  (−0.6;1.1)  Sometimes lonely  222  20.9  −0.3  (−1.1;0.6)  Seldom lonely  509  30.0  −0.4  (−1.0;0.2)  Seldom lonely  489  21.2  0.1  (−0.5;0.7)  Not lonely  843  30.4  Ref.    Not lonely  790  21.2  Ref.    P-value  0.21    0.79  Crude model: outcome is solely dependent on loneliness. Model 1: Adjusted for age, occupational social class, cohabitation status, number of chronic illnesses and personality (neuroticism, agreeableness, conscientiousness). Joint effect of loneliness and occupational social class/cohabitation Occupational social class was associated with physical capability in both men and women. On average, men and women in the high occupational social class had 1 kg greater handgrip strength and could raise themselves two times more from the chair than those in the low occupational social class. Men who lived alone had, on average, 2 kg weaker grip strength, than men who did not live alone. Women who lived alone could on average raise themselves one fewer times from the chair, than women who did not live alone. Men who were jointly exposed to low occupational social class/living alone and loneliness generally showed lower hand-grip strength [Lonely/low social class −1.8 kg (95% CI −3.0;−0.6), Lonely/living alone −1.5 (−2.7;−0.4)] and fewer chair raises [Lonely/low social class −2.4 rises/30 s (95% CI −3.2;−1.6)] compared with the reference group of non-lonely persons with high occupational social class or who was not living alone. Of notice, lonely men living alone did not experience a poorer chair-rise test compared to the reference group [0.1 rises/30 s (95% CI −0.8;0.8)]. For women the joint variables were not associated with hand grip strength, but lonely women in low social class had fewer chair rises than the reference group (−2.3 rises/30 s (95% CI −3.3;−1.2)]. However, in conclusion the associations were no more than expected from the combination of the two exposures (P-values of the interaction term range 0.13–0.67) (table 4). Sensitivity analyses with those reporting loneliness ’often’, ’sometimes’ and ’seldom’ categorized together were run, but this did not change the overall conclusion. Table 4 Multivariate linear regression analysis Hand grip strength N = 5224  Men  Women  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  Loneliness  1. lonely (often/sometimes lonely)  570  46.2  −0.1  (−0.9;0.7)    276  30.3  0.4  (−0.3;1.1)    2. not lonely (seldom/not lonely)  3026  46.3  Ref.      1352  29.9  Ref.      Occupational social class  1. low social class (IV, V, VI)  1185  45.6  −1.3  (−1.9;−0.7)    605  29.6  −1.1  (−1.6;0.6)    2. high social class (I, II, III)  2411  46.9  Ref.      1023  30.7  Ref.      Cohabitation status  1. Living alone  525  45.0  -2.4  (-3.2;-1.5)    260  30.0  -0.2  (-0.9;0.5)    2. Not living alone  3071  47.4  Ref.      1368  30.2  Ref.      Loneliness/social class  1. Lonely/low social class  262  45.1  −1.8  (−3.0;−0.6)  0.44  137  30.4  −0.1  (−1.1;0.9)    2. Lonely/high social class  308  47.2  0.3  (−0.7;1.4)    139  30.2  −0.4  (−1.3;0.6)    3. Not lonely/low social class  923  45.8  −1.1  (−1.8;−0.4)    468  29.1  −1.4  (−2.0;−0.8)    4. Not lonely/high social class  2103  46.9  Ref.      884  30.5  Ref.    0.13  Loneliness/co-habitation status  1. Lonely/living alone  241  46.3  −1.5  (−2.7;−0.4)  0.14  101  30.9  0.6  (−0.5;1.7)    2. Lonely/not living alone  329  47.3  -0.6  (-1.5;0.4)    175  30.7  0.4  (-0.4;1.3)    3. Not lonely/living alone  284  45.1  -2.7      159  30.1  -0.2  (-1.1;0.7)    4. Not lonely/not living alone    47.8  Ref.  (-3.8;-1.7)    1193  30.3  Ref.    0.64  Chair rise test N = 4742  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Loneliness  1. Lonely (often/sometimes lonely)  492  19.2  −0.3  (−0.8;0.3)    265  20.2  −0.6  (−1.3;0.2)    2. Not lonely (seldom/not lonely)  2706  19.5  Ref.      1279  20.8  Ref.      Occupational social class  1. Low social class (IV, V, VI)  1033  18.4  −2.0  (−2.4;−1.5)    573  19.5  −2.0  (−2.6;−1.4)    2. High social class (I, II, III)  2165  20.3  Ref.      971  21.5  Ref.      Chair rise test N = 4742  Men  Women  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Cohabitation status  1. Living alone  462  19.2  −0.2  (−0.8;0.3)    249  20.1  −0.8  (−1.6;−0.1)    2. Not living alone  2736  19.5  Ref.      1295  20.9  Ref.      Loneliness/social class  1. Lonely/low social class  220  18.0  −2.4  (−3.2;−1.6)  0.35  129  19.5  −2.3  (−3.3;−1.2)    2. Lonely/high social class  272  20.4  0.0  (−0.8;0.7)    136  20.9  −0.9  (−1.9;0.1)    3. Not lonely/low social class  813  18.6  −1.9  (−2.3;−1.4)    444  19.6  −2.1  (−2.8;−1.5)    4. Not lonely/high social class  1893  20.4  Ref.      835  21.8  Ref.    0.26  Loneliness/co-habitation status  1. Lonely/living alone  204  19.8  0.1  (−0.8;0.8)  0.31  98  20.5  −1.1  (−2.2;0.1)    2. Lonely/not living alone  288  19.4  −0.4  (−1.1;0.3)    167  21.2  −0.4  (−1.2;0.5)    3. Not lonely/living alone  258  19.6  −0.2  (−0.9;0.5)    151  20.9  −0.7  (−1.6;0.2)    4. Not lonely/not living alone  2448  19.8  Ref.      1128  21.6  Ref.    0.67  Hand grip strength N = 5224  Men  Women  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  Loneliness  1. lonely (often/sometimes lonely)  570  46.2  −0.1  (−0.9;0.7)    276  30.3  0.4  (−0.3;1.1)    2. not lonely (seldom/not lonely)  3026  46.3  Ref.      1352  29.9  Ref.      Occupational social class  1. low social class (IV, V, VI)  1185  45.6  −1.3  (−1.9;−0.7)    605  29.6  −1.1  (−1.6;0.6)    2. high social class (I, II, III)  2411  46.9  Ref.      1023  30.7  Ref.      Cohabitation status  1. Living alone  525  45.0  -2.4  (-3.2;-1.5)    260  30.0  -0.2  (-0.9;0.5)    2. Not living alone  3071  47.4  Ref.      1368  30.2  Ref.      Loneliness/social class  1. Lonely/low social class  262  45.1  −1.8  (−3.0;−0.6)  0.44  137  30.4  −0.1  (−1.1;0.9)    2. Lonely/high social class  308  47.2  0.3  (−0.7;1.4)    139  30.2  −0.4  (−1.3;0.6)    3. Not lonely/low social class  923  45.8  −1.1  (−1.8;−0.4)    468  29.1  −1.4  (−2.0;−0.8)    4. Not lonely/high social class  2103  46.9  Ref.      884  30.5  Ref.    0.13  Loneliness/co-habitation status  1. Lonely/living alone  241  46.3  −1.5  (−2.7;−0.4)  0.14  101  30.9  0.6  (−0.5;1.7)    2. Lonely/not living alone  329  47.3  -0.6  (-1.5;0.4)    175  30.7  0.4  (-0.4;1.3)    3. Not lonely/living alone  284  45.1  -2.7      159  30.1  -0.2  (-1.1;0.7)    4. Not lonely/not living alone    47.8  Ref.  (-3.8;-1.7)    1193  30.3  Ref.    0.64  Chair rise test N = 4742  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Loneliness  1. Lonely (often/sometimes lonely)  492  19.2  −0.3  (−0.8;0.3)    265  20.2  −0.6  (−1.3;0.2)    2. Not lonely (seldom/not lonely)  2706  19.5  Ref.      1279  20.8  Ref.      Occupational social class  1. Low social class (IV, V, VI)  1033  18.4  −2.0  (−2.4;−1.5)    573  19.5  −2.0  (−2.6;−1.4)    2. High social class (I, II, III)  2165  20.3  Ref.      971  21.5  Ref.      Chair rise test N = 4742  Men  Women  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Cohabitation status  1. Living alone  462  19.2  −0.2  (−0.8;0.3)    249  20.1  −0.8  (−1.6;−0.1)    2. Not living alone  2736  19.5  Ref.      1295  20.9  Ref.      Loneliness/social class  1. Lonely/low social class  220  18.0  −2.4  (−3.2;−1.6)  0.35  129  19.5  −2.3  (−3.3;−1.2)    2. Lonely/high social class  272  20.4  0.0  (−0.8;0.7)    136  20.9  −0.9  (−1.9;0.1)    3. Not lonely/low social class  813  18.6  −1.9  (−2.3;−1.4)    444  19.6  −2.1  (−2.8;−1.5)    4. Not lonely/high social class  1893  20.4  Ref.      835  21.8  Ref.    0.26  Loneliness/co-habitation status  1. Lonely/living alone  204  19.8  0.1  (−0.8;0.8)  0.31  98  20.5  −1.1  (−2.2;0.1)    2. Lonely/not living alone  288  19.4  −0.4  (−1.1;0.3)    167  21.2  −0.4  (−1.2;0.5)    3. Not lonely/living alone  258  19.6  −0.2  (−0.9;0.5)    151  20.9  −0.7  (−1.6;0.2)    4. Not lonely/not living alone  2448  19.8  Ref.      1128  21.6  Ref.    0.67  Main and joint effects of loneliness, occupational social class and cohabitation status on physical capability Models have been adjusted for age, number of chronic illnesses and personality (neuroticism, agreeableness, conscientiousness). The loneliness/occupational social class model was also adjusted for cohabitation status, while the loneliness/cohabitation status model was adjusted for occupational social class. * Test for interaction by the inclusion of a product term: (1) loneliness + occupational social class + loneliness×occupational social class + covariates, (2) loneliness + cohabitation status + loneliness×cohabitation status + covariates, P values for the product term. Discussion In a large cross-sectional study of middle-aged adults we do not find evidence of poorer physical capability in lonely men and women. Men and women in low occupational social class or who were living alone had weaker hand grip strength and were limited in their chair raise ability, but there were no support for the differential vulnerability hypothesis. This study is among the first to investigate the association between loneliness and objectively measured physical capability among the middle-aged (50–60 years). In older populations the findings are mixed. A previous cross-sectional study among older people showed only a negligible association between loneliness and self-reported functional ability measured as independence from help in daily activities, which paradoxically showed that higher score on independence was associated with higher degree of loneliness.24 In contrast, a longitudinal study of middle-aged and older men and women (mean age 67 years), loneliness was found to increase the risk of self-reported functional limitations after taking into account the reciprocal association between functional limitations and loneliness as well as socio-demographic, health and health behaviour.5 Compared with this study, the cohort were older, and the findings were based on both self-reported measures of loneliness and physical limitation leaving a risk of dependent measurement error due to personality differences in reporting. Hence, studying objectively measured physical capability and adjusting for personality factors may produce results less prone to bias. In a longitudinal study, among older people (mean age 79 years), a strong association between loneliness and objectively measured motor decline was reported after adjustment for sociodemographic, baseline disability, mental and physical health and health behaviours.4 Despite its lack of adjustment for personality factors, this study lends support for a causal relation between loneliness and objectively measured physical function in older people. The younger age of the Copenhagen Aging and Midlife Biobank participants potentially in part explain our null-finding as the suggested physiological pathways11,12 between loneliness and health may not yet have resulted in detectable differences in physical capability. Furthermore, the prevalence of loneliness in middle age is rather low in the Northern European countries which we also find in this study (3% reports often to feel lonely), leaving only a smaller group exposed. A reason for this may be that most people by midlife have developed stable social networks, are well established on the labour market and most often have not suffered from social losses such as the death of a spouse or friend.8 The prevalence of loneliness is known to increase with age due to social losses, which may explain the negative association with physical capability in older people. Our findings suggest that even among the potential vulnerable groups in low occupational social class and those who live alone the effects of loneliness are not intensified. A wide variety of instruments have been used to measure loneliness, spanning from single item questions to comprehensive scales of many items e.g. the UCLA loneliness scale25 or de Jong Giervald scale.26 However, regardless of measurement differences there seems to be overall support for an association between loneliness and adverse health outcomes.1 The single item used in this study sheds light on the quantitative aspect of loneliness, where contact with others is sparse contrasting other measures evaluating feelings of loneliness.25 A Danish study among the elderly found that 43% who report unwanted aloneness are often lonely, while 59% of the often lonely also feel unwanted aloneness,7 indicating some difference in the unwanted aloneness-measure and the measure of loneliness. A meta-analysis found that older persons were more likely to associate loneliness with lack of activity, while middle-aged were more likely to recognize loneliness as absence of people.11 Therefore, the item used in this study may well encapsulate loneliness among the middle-aged. This one-item assessment has been used in prior research and the prevalence of often feeling lonely identified in this study (∼3%) is comparable to the prevalence in the middle-aged (4–5%) identified in National population-based surveys27 as well as that of a recent study in Denmark using the Three Item Loneliness UCLA Scale(4%).7 The characteristics of those who reported often to feel lonely in the CAMB participants were consistent with prior research,1,7,9,10,16,28 which may support the comparability of the single item measure of loneliness to the more comprehensive scores. Selection bias may have affected the results as CAMB non-respondents had lower education and were more likely to be unemployed, as well as they experienced higher all-cause mortality.17 These factors are associated with loneliness,4,5,8,9,28–30 and further it has been suggested that persons feeling lonely may be less inclined to participate in surveys.8 It is also possible that people, who are less physically capable, are less likely to participate in studies such as CAMB because of mobility limitations. Per design, we condition on participation which is a common cause of both loneliness and physical capability leaving a risk of selection bias.31 Considering the above-mentioned selection mechanisms, it is likely that lonely individuals that did in fact participate in CAMB will also be less likely to be physical incapable, and hence our findings are likely to underestimate the association between loneliness and physical capability. Strengths of this study include the large sample with comprehensive information on the participants, and that physical capability was measured by objective tests, which increases reliability and permits cross-cohort comparisons.20 A limitation of this study is the cross-sectional design which increases the risk of any associations being due to reverse causality. Due to demographic changes the number of 65+ years olds will increase dramatically in the years to come of which a considerable amount will be living alone, and thereby also in increased risk of loneliness and social isolation which potentially may harm the disadvantaged the most. Niedzwiedz et al. found loneliness to be most prevalent among the least wealthy groups across 14 European countries. This indicates the necessity for a continued focus on the potential health consequences of loneliness, social isolation in general among older people but especially among the least advantaged. In contrast to earlier studies among older people, we did not find an association between loneliness and physical capability in a large sample of middle-aged men and women, and there was no support for an interaction between loneliness and occupational social class and cohabitation status, respectively. This may be explained by the harmful effects of loneliness not yet to have had influence on physical capability in this age group. Selection bias may however have caused a true association to be underestimated. Further longitudinal studies are needed to clarify if and at what point in the life course loneliness begins to affect physical capability. Acknowledgements The authors thank the staff at Department of Public Health and National Research Centre for the Working Environment, who undertook the data collection. Further thanks to Kirsten Avlund, Helle Bruunsgaard, Nils-Erik Fiehn, Åse Marie Hansen, Poul Holm-Pedersen, Rikke Lund, Erik Lykke Mortensen and Merete Osler, who initiated and established the Copenhagen Aging and Midlife Biobank from 2009 to 2011. The authors acknowledge the crucial role of the initiators and steering committees of the Metropolit Cohort, The Copenhagen Perinatal Cohort and The Danish Longitudinal Study on Work, Unemployment and Health. The authors also want to thank professor Per Kragh Andersen, Section of Biostatistics for valuable advice regarding the statistical analyses. Disclaimers None. Supplementary data Supplementary data are available at EURPUB online. Funding This work and The Copenhagen Aging and Midlife Biobank were supported by grants from the VELUX FOUNDATION (VELUX26145, VELUX31539). Conflicts of interest: None declared. Key points In older people, loneliness has been associated with higher risk of decline in physical functioning. Little is known about this association in middle-aged individuals and if specific vulnerable groups can be identified. No association between loneliness and physical capability was found in this cohort of middle-aged men and women. This may be due to the harmful effects of loneliness not yet having resulted in detectable differences in physical capability in this age group. Lower occupational social class and living alone was associated with poorer physical capability, but there was no interaction between these factors and loneliness on physical capability. Further longitudinal studies are needed to clarify if and at what point in the life course loneliness begins to affect physical capability. References 1 Hawkley LC, Cacioppo JT. Loneliness. In: Reis HT, Sprecher S, editors. Encyclopedia of Human Relationships . Thousand Oaks: Sage, 2009: 985– 90. 2 Christiansen J, Larsen FB, Lasgaard M. Do stress, health behavior, and sleep mediate the association between loneliness and adverse health conditions among older people? Soc Sci Med  2016; 152: 80– 6. Google Scholar CrossRef Search ADS PubMed  3 Whisman MA. Loneliness and the metabolic syndrome in a population-based sample of middle-aged and older adults. Health Psychol  2010; 29: 550– 4. Google Scholar CrossRef Search ADS PubMed  4 Buchman AS, Boyle PA, Wilson RS, et al.   Loneliness and the rate of motor decline in old age: the Rush Memory and Aging Project, a community-based cohort study. BMC Geriatr  2010; 10: 77. Google Scholar CrossRef Search ADS PubMed  5 Luo Y, Hawkley LC, Waite LJ, Cacioppo JT. Loneliness, health, and mortality in old age: a national longitudinal study. Soc Sci Med  2012; 74: 907– 14. Google Scholar CrossRef Search ADS PubMed  6 Yang KM, Victor C. Age and loneliness in 25 European nations. Ageing Soc  2011; 31: 1368– 88. Google Scholar CrossRef Search ADS   7 Lasgaard M, Friis K. Loneliness in the population—prevalence and methodological considerations [in Danish Ensomhed i Befolkningen—forekomst og metodiske overvejelser.]Temaanalyse, vol. 3, Hvordan har du det? Aarhus, CFK Folkesundhed og kvalitetsudvikling, Denmark; 2015. 8 Pinquart M, Sorensen S. Influences on loneliness in older adults: a meta-analysis. Basic Appl Soc Psych  2001; 23: 245– 66. Google Scholar CrossRef Search ADS   9 Niedzwiedz CL, Richardson EA, Tunstall H, et al.   The relationship between wealth and loneliness among older people across Europe: is social participation protective? Prev Med  2016; 91: 24– 31. Google Scholar CrossRef Search ADS PubMed  10 Penninx BW, van Tilburg T, Kriegsman DM, et al.   Social network, social support, and loneliness in older persons with different chronic diseases. J Aging Health  1999; 11: 151– 68. Google Scholar CrossRef Search ADS PubMed  11 Hawkley LC, Cacioppo JT. Aging and loneliness: Downhill quickly? Curr Dir Psychol Sci  2007; 16: 187– 91. Google Scholar CrossRef Search ADS   12 Hackett RA, Hamer M, Endrighi R, et al.   Loneliness and stress-related inflammatory and neuroendocrine responses in older men and women. Psychoneuroendocrinology  2012; 37: 1801– 9. Google Scholar CrossRef Search ADS PubMed  13 Hansen AM, Andersen LL, Skotte J, et al.   Social class differences in physical functions in middle-aged men and women. J Aging Health  2014; 26: 88– 105. Google Scholar CrossRef Search ADS PubMed  14 Diderichsen F, Evans T, Whitehead M. The social basis of disparities in health. In: Evans T, Whitehead M, Diderichsen F, Bhuiya A, Wirth M, editors. Challenging Inequties in Health , 1st edn Oxford: OUP, 2001. 12– 23. Google Scholar CrossRef Search ADS   15 Lund R, Nilsson CJ, Avlund K. Can the higher risk of disability onset among older people who live alone be alleviated by strong social relations? A longitudinal study of non-disabled men and women. Age Ageing  2010; 39: 319– 26. Google Scholar CrossRef Search ADS PubMed  16 Cacioppo JT, Hughes ME, Waite LJ, et al.   Loneliness as a specific risk factor for depressive symptoms: cross-sectional and longitudinal analyses. Psychol Aging  2006; 21: 140– 51. Google Scholar CrossRef Search ADS PubMed  17 Lund R, Mortensen EL, Christensen U, et al.   Cohort Profile: The Copenhagen Aging and Midlife Biobank (CAMB). Int J Epidemiol. 2016; 45: 1044– 1053. 18 Christensen U, Krolner R, Nilsson CJ, et al.   Addressing social inequality in aging by the Danish occupational social class measurement. J Aging Health  2014; 26: 106– 27. Google Scholar CrossRef Search ADS PubMed  19 Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology  1999; 10: 37– 48. Google Scholar CrossRef Search ADS PubMed  20 Cooper R, Hardy R, Sayer A, Kuh DA. Life course approach to physical capability. In: Kuh D, Cooper R, Hardy R, Richards M, Ben-Shlomo Y, editors. A Life Course Approach to Healthy Ageing . London: Oxford University Press, 2014. 16– 31. 21 Rhodes RE, Smith NE. Personality correlates of physical activity: a review and meta-analysis. Br J Sports Med  2006; 40: 958– 65. Google Scholar CrossRef Search ADS PubMed  22 Mortensen EL, Flensborg-Madsen T, Molbo D, et al.   Personality in late midlife: associations with demographic factors and cognitive ability. J Aging Health  2014; 26: 21– 36. Google Scholar CrossRef Search ADS PubMed  23 Bech P, Rasmussen NA, Olsen LR, et al.   The sensitivity and specificity of the Major Depression Inventory, using the Present State Examination as the index of diagnostic validity. J Affect Disord  2001; 66: 159– 64. Google Scholar CrossRef Search ADS PubMed  24 Prieto-Flores ME, Forjaz MJ, Fernandez-Mayoralas G, et al.   Factors associated with loneliness of noninstitutionalized and institutionalized older adults. J Aging Health  2011; 23: 177– 94. Google Scholar CrossRef Search ADS PubMed  25 Russell DW. UCLA Loneliness Scale (Version 3): reliability, validity, and factor structure. J Pers Assess  1996; 66: 20– 40. Google Scholar CrossRef Search ADS PubMed  26 Gierveld JD, van TT. A 6-item scale for overall, emotional, and social loneliness—confirmatory tests on survey data. Res Aging  2006; 28: 582– 98. Google Scholar CrossRef Search ADS   27 The Danish Health Authority. Health of the Danes—The National Health Profile 2013 (In Danish: Danskernes sundhed - Den Nationale Sundhedsprofil 2013). 2014. http://www.danskernessundhed.dk/ (22nd May 2017) 28 Lauder W, Mummery K, Jones M, Caperchione C. A comparison of health behaviours in lonely and non-lonely populations. Psychol Health Med  2006; 11: 233– 45. Google Scholar CrossRef Search ADS PubMed  29 Newall NE, Chipperfield JG, Bailis DS, Stewart TL. Consequences of loneliness on physical activity and mortality in older adults and the power of positive emotions. Health Psychol  2013; 32: 921– 4. Google Scholar CrossRef Search ADS PubMed  30 Shiovitz-Ezra S, Ayalon L. Situational versus chronic loneliness as risk factors for all-cause mortality. Int Psychogeriatr  2010; 22: 455– 62. Google Scholar CrossRef Search ADS PubMed  31 Hernan MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology  2004; 15: 615– 25. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The European Journal of Public Health Oxford University Press

Loneliness and objectively measured physical capability in middle-aged adults

Loading next page...
1
 
/lp/ou_press/loneliness-and-objectively-measured-physical-capability-in-middle-aged-xwkDNgpVeE
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
ISSN
1101-1262
eISSN
1464-360X
D.O.I.
10.1093/eurpub/ckx069
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Loneliness is associated with poor functional ability in older people. Little is known about this association in the middle-aged. The aim is to investigate if perceived loneliness is associated with lower physical capability among middle-aged men and women and if the associations of loneliness with physical capability interact with socioeconomic position and cohabitation status. Methods 5224 participants from Copenhagen Aging and Midlife Biobank (CAMB) aged 49–62 years (mean age 54) were included. Handgrip strength (measured by a dynamometer) and maximal number of chair rises in 30 s was recorded. Multivariate linear regression analyses were adjusted for age, occupational social class, cohabitation status, morbidity and personality traits. Results No association was found between loneliness and physical capability. For example estimates for handgrip strength in ‘often’ lonely men and women compared with the ‘not lonely’ were 1.2 kg (95% CI − 0.5;2.9)/1.0 kg (−0.7;2.6). Low occupational social class was associated with poorer physical capability, and living alone was associated with poorer handgrip strength in men [−2.4 kg (95% CI − 3.2;−1.5)] and poorer chair rise test in women [−0.8 rises (95% CI − 1.6;−0.1)]. There was no support for interactions. Conclusion In contrast to earlier studies among older people, no association between loneliness and physical capability was found in this cohort of middle-aged men and women. Loneliness may not yet have resulted in detectable differences in physical capability in this age group. Further research is needed to clarify if, and at what point in the life course loneliness begins to affect physical capability. Introduction Loneliness has been described as “the social equivalent of physical pain, hunger and thirst”1 and can be described as a feeling of being socially isolated, which is often accompanied by a perceived deficiency in quantity and quality of social relations.1 Loneliness has been associated with adverse health outcomes, including mortality, depression, cardiovascular disease, metabolic syndrome.2,3 Loneliness has also been associated with functional decline in older people,4,5 but little is known about this association in middle-aged adults. In the Northern European countries the prevalence of loneliness in the 60+ years old is around 6% which is lower compared with Southern (11%) and Eastern European countries (20%).6 The prevalence of loneliness furthermore varies with age,7 gender,7 socioeconomic position,7–9 living arrangements,7 health10 and personality.1 Suggested pathways through which loneliness accelerates physical ageing and contributes to poor health11 include: higher total peripheral resistance and blood pressure (1), higher concentration of epinephrine,11 dysregulation of the inflammatory and neuroendocrine responses,12 poorer self-reported sleep quality11 A clear social gradient in physical capability among middle-aged adults has been suggested,13 and furthermore that loneliness is associated with lower socioeconomic position.8,9 According to the concept of differential vulnerability the effect of a given risk factor (here loneliness) is intensified along the social strata.14 Similarly, living alone has been associated with poorer physical capability especially among older males15 and higher levels of loneliness are experienced by those living alone.16 The aim of this study is to investigate (1) if perceived loneliness is related to lower levels of physical capability evaluated using objective measures, including handgrip strength and number of chair rises, in a middle-aged population, and (2) if the associations of loneliness on physical capability are different in respondents with a high or low socioeconomic position and for those living alone versus those not living alone. Methods Study population The Copenhagen Aging and Midlife Biobank (CAMB) is based on three longitudinal studies: the Metropolit 1953 Danish Male Birth Cohort (MP), The Copenhagen Perinatal Cohort (CPC) born 1959–61 and The Danish Longitudinal Study on Work, Unemployment and Health (DALWUH) born 1949 or 1959. In total, 5575 persons answered the questionnaire and participated in the clinical examination (response rate 30%). Compared with non-respondents, CAMB participants were more educated and were less likely to be unemployed and experienced significantly higher all-cause mortality.17 Exclusion criteria Participants with missing information on loneliness, included covariates, job information for social class classification were excluded as were participants reporting to be a student or a homemaker (0.5% of the participants). There were no specific exclusion criteria for the handgrip test, but participants with systolic >160 and/or diastolic blood pressure >100 were not asked to complete the chair rise test.13 In total, 351 (6%) were excluded due to missing values in the analyses on hand grip strength (N = 5224) and 833 (15%) in the analysis on chair rise test (N = 4742). See figure 1 in Supplementary files for a flow chart of the study population. Loneliness, occupational social class and cohabitation status Loneliness (unwanted aloneness) was measured by a single item: “Are you ever alone, when you would much rather be with others?” Response categories were “Yes, often”,” Yes, sometimes”,” Yes, but seldom” and” No”, where “No” was used as the reference. For the analyses investigating the potential differential vulnerability, loneliness was dichotomized into either lonely (‘often’ or ‘sometimes lonely’) or not lonely (‘seldom’ or ‘not lonely’). Occupational social class was classified based on occupation in accordance with the Danish Occupational Social Class Measurement into occupational social class I–VI.18 Occupational social class I–V represent employed individuals ranging from professional occupation in social class I, to unskilled occupation in social class V. Social class VI represents people on transfer income, including unwaged persons and those receiving social security benefits or sickness benefits. Occupational social class was also dichotomized into high (social class I, II and III) and low (IV, V and VI). Loneliness and occupational social class/cohabitation status were combined into two joint variables with the following four levels; “lonely/low social class or living alone”, “lonely/high social class or not living alone”, “not lonely/low social class or living alone” and “not lonely/high social class or not living alone” (reference group). Physical capability Handgrip strength was measured with a dynamometer (model G100, Biometrics Ltd, Newport, UK) with the best maximal handgrip strength in kg recorded out of three to five attempts.13 During a 30-s chair rise test, the maximal number of chair rises was recorded (chair with a seat height of 45 cm and the seat wired to a computer providing information on mechanical contact).13 Covariates Potential confounders were identified based on prior knowledge and the method of directed acyclic graphs (see Supplementary file).19 Covariates included age, gender, occupational social class, cohabitation status, number of chronic physical and mental disorders and personality as they have been suggested as potential causes of both loneliness1,7–10 and physical capability/physical function.13,20 Loneliness is associated with high neuroticism and low conscientiousness and agreeableness,1 and both conscientiousness and neuroticism are associated with physical functioning and physical activity.21 Depressive symptoms can act both as a mediator and a potential confounder and was therefore included in a separate model.16 Body mass index (BMI), alcohol intake, smoking habits, physical activity were all identified as potential mediators of the relation between loneliness and physical functioning and they were added in a separate model. Gender and age were assessed from social security numbers given to all Danish citizens. The CAMB participants were asked to declare the following medical conditions: (1) allergies, (2) diabetes, (3) hypertension, (4) myocardial infarction, (5) stroke, (6) lung disease (incl. asthma), chronic bronchitis, emphysema, (7) autoimmune disease and arthritis (8) cancer incl. leukaemia and (9) other mental disorder than depression. These conditions were compiled to a morbidity index counting the number of medical conditions reported (0–9). Personality was assessed with the short Danish version of the NEO Five Factor Inventory.21,22 Depressive symptoms were assessed using the Major Depression Inventory with a score ranging from 0 to 50.23 For detailed description of measurement of BMI, alcohol intake, smoking habits, physical activity please see Supplementary file. Statistical methods The associations between loneliness and handgrip strength/number of chair rises were analysed with multiple linear regression, PROC GLM, using SAS 9.3. Data were stratified by gender because the chair rise test was non-normally distributed in the combined data. Plots of the residuals for the association between hand grip strength and chair rise, respectively, as well as loneliness were judged symmetric. Potential interaction (differential vulnerability) was assessed by analysing the association between joint variables of loneliness and cohabitation status/occupational social class, respectively, in regards to physical capability, as well as by the inclusion of product terms between loneliness and occupational social class and cohabitation status, respectively, in the regression analyses. Initially, crude multivariate regression models were run followed by a model adjusted for the identified confounders (Model 1). In Model 2, depressive symptoms at baseline were added, and in Model 3 health behaviour variables were added. In the joint effect analyses, the main effects of each of the joint effect variables (loneliness, occupational social class and cohabitation status) were not included. Weights were not used in this study as this method builds on the understanding that those who do in fact participate (given their measured variables) are representative of those who do not. This is difficult to validate and hence does not solve the problem of selection bias. Results Population demographics Sixteen per cent of the population reported ‘often’ or ‘sometimes’ feeling lonely, 3% often felt lonely. Approximately one-third of the population was men (n = 3596 men and n = 1628 women), with no gender differences in loneliness. Age ranged from 48 to 62 years, and mean age did not differ between the four categories of loneliness among women but differed slightly for men (in the ‘often lonely’ mean age was 1 year lower). The association between loneliness and each of the covariates showed a dose-dependent pattern except for alcohol intake. More men and women who often felt lonely, were in the lowest occupational social class group VI compared with the other groups. Fifteen per cent of both men and women were living alone. Among men in the often lonely group 61% were living alone, for often lonely women 34% was living alone. Respondents scoring high on loneliness were in general more likely to suffer from three or more chronic diseases, to be depressed, obese or being smokers, to be physical active less than 3–4 h/week (stronger for men) and to be either non-drinkers or heavy drinkers (men only). The often lonely scored higher on neuroticism and lower on conscientiousness (tables 1 and 2). Table 1 Men: distribution of demographic, personality, health and health behaviour variables by loneliness (N = 3596)   Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  110 (3%)  460 (13%)  1095 (31%)  1931 (54%)    Age, mean (years)  54 (SD 3.5)  55 (SD 3.2)  55 (SD 3.3)  55 (SD 3.2)  0.0083  Cohabitation status  <0.0001  Alone  67 (61%)  174 (38%)  193 (18%)  91 (5%)    Not alone  43 (39%)  286 (62%)  902 (82%)  1840 (95%)    Socioeconomic position  <0.0001  Social class I  12 (11%)  71 (15%)  204 (19%)  385 (20%)    Social class II  15 (14%)  89 (19%)  255 (23%)  549 (28%)    Social class III  23 (21%)  98 (21%)  244 (22%)  466 (24%)    Social class IV  15 (14%)  72 (16%)  178 (16%)  269 (14%)    Social class V  6 (6%)  51 (11%)  90 (8%)  166 (9%)    Social class VI  39 (36%)  79 (17%)  124 (11%)  96 (5%)    Number of chronic illnessesa  <0.0001  0  29 (26%)  252 (31%)  411 (38%)  763 (40%)    1–2  59 (54%)  252 (55%)  557 (51%)  998 (52%)    ≥3  22 (20%)  67 (15%)  127 (11%)  170 (9%)    Depressionb  <0.0001  Yes  27 (25%)  34 (7%)  36 (3%)  27 (1%)    No  83 (75%)  426 (93%)  1059 (97%)  1904 (99%)    Personality  Neuroticism  Median 25  Median 19  Median 17  Median 15  <0.0001  Q1c 17  Q1 15  Q1 13  Q1 11    Q3c 30  Q3 25  Q3 22  Q3 18    Agreeableness  Median 32  Median 32  Median 32  Median 33  <0.0001  Q1 28  Q1 29  Q1 29  Q1 30    Q3 35  Q3 35  Q3 36  Q3 37    Conscientiousness  Median 31  Median 32  Median 33  Median 35  <0.0001  Q1 25  Q1 28  Q1 30  Q1 31    Q3 35  Q3 35  Q3 37  Q3 38    BMI  0.0022  Underweight (<18.5)  2 (2%)  2 (1%)  6 (1%)  5 (1%)    Normal (18.5–25)  41 (37%)  183 (40%)  391 (36%)  727 (38%)    Overweight (25–30)  38 (35%)  204 (44%)  513 (47%)  929 (48%)    Obese (>30)  29 (26%)  71 (15%)  185 (17%)  270 (14%)    Smoking status  <0.0001  Smoker  42 (38%)  138 (30%)  256 (23%)  400 (21%)    Non-smoker  68 (62%)  322 (70%)  839 (77%)  1531 (79%)    Alcohol  <0.0001  Not drinking  19 (17%)  54 (12%)  90 (8%)  141 (7%)    0 units/week  50 (45%)  223 (48%)  562 (51%)  1109 (57%)    Low risk drinking  12 (11%)  75 (16%)  211 (19%)  313 (16%)    14 units/weekd  29 (26%)  108 (24%)  232 (21%)  368 (19%)    Elevated risk drinking  15–21 units/weekd            High risk drinking  >21 units/weekd            Physical activitye  0.0312  ≤3–4 h/week  48 (44%)  177 (38%)  366 (33%)  645 (33%)    ≥5 h/week  62 (56%)  283 (62%)  729 (67%)  1286 (67%)      Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  110 (3%)  460 (13%)  1095 (31%)  1931 (54%)    Age, mean (years)  54 (SD 3.5)  55 (SD 3.2)  55 (SD 3.3)  55 (SD 3.2)  0.0083  Cohabitation status  <0.0001  Alone  67 (61%)  174 (38%)  193 (18%)  91 (5%)    Not alone  43 (39%)  286 (62%)  902 (82%)  1840 (95%)    Socioeconomic position  <0.0001  Social class I  12 (11%)  71 (15%)  204 (19%)  385 (20%)    Social class II  15 (14%)  89 (19%)  255 (23%)  549 (28%)    Social class III  23 (21%)  98 (21%)  244 (22%)  466 (24%)    Social class IV  15 (14%)  72 (16%)  178 (16%)  269 (14%)    Social class V  6 (6%)  51 (11%)  90 (8%)  166 (9%)    Social class VI  39 (36%)  79 (17%)  124 (11%)  96 (5%)    Number of chronic illnessesa  <0.0001  0  29 (26%)  252 (31%)  411 (38%)  763 (40%)    1–2  59 (54%)  252 (55%)  557 (51%)  998 (52%)    ≥3  22 (20%)  67 (15%)  127 (11%)  170 (9%)    Depressionb  <0.0001  Yes  27 (25%)  34 (7%)  36 (3%)  27 (1%)    No  83 (75%)  426 (93%)  1059 (97%)  1904 (99%)    Personality  Neuroticism  Median 25  Median 19  Median 17  Median 15  <0.0001  Q1c 17  Q1 15  Q1 13  Q1 11    Q3c 30  Q3 25  Q3 22  Q3 18    Agreeableness  Median 32  Median 32  Median 32  Median 33  <0.0001  Q1 28  Q1 29  Q1 29  Q1 30    Q3 35  Q3 35  Q3 36  Q3 37    Conscientiousness  Median 31  Median 32  Median 33  Median 35  <0.0001  Q1 25  Q1 28  Q1 30  Q1 31    Q3 35  Q3 35  Q3 37  Q3 38    BMI  0.0022  Underweight (<18.5)  2 (2%)  2 (1%)  6 (1%)  5 (1%)    Normal (18.5–25)  41 (37%)  183 (40%)  391 (36%)  727 (38%)    Overweight (25–30)  38 (35%)  204 (44%)  513 (47%)  929 (48%)    Obese (>30)  29 (26%)  71 (15%)  185 (17%)  270 (14%)    Smoking status  <0.0001  Smoker  42 (38%)  138 (30%)  256 (23%)  400 (21%)    Non-smoker  68 (62%)  322 (70%)  839 (77%)  1531 (79%)    Alcohol  <0.0001  Not drinking  19 (17%)  54 (12%)  90 (8%)  141 (7%)    0 units/week  50 (45%)  223 (48%)  562 (51%)  1109 (57%)    Low risk drinking  12 (11%)  75 (16%)  211 (19%)  313 (16%)    14 units/weekd  29 (26%)  108 (24%)  232 (21%)  368 (19%)    Elevated risk drinking  15–21 units/weekd            High risk drinking  >21 units/weekd            Physical activitye  0.0312  ≤3–4 h/week  48 (44%)  177 (38%)  366 (33%)  645 (33%)    ≥5 h/week  62 (56%)  283 (62%)  729 (67%)  1286 (67%)    a Number of chronic illnesses includes: (1) Allergies, (2) Diabetes, (3) Hypertension, (4) Myocardial infarction, (5) Stroke, (6) Lung disease (incl. asthma, chronic bronchitis, emphysema) , (7) Autoimmune disease and arthritis, (8) Cancer incl. leukaemia and, (9) present other mental disorder. b Depression according to Major Depression Inventory-scale (41;42) ‘Yes’ includes moderate and severe depression and ‘no’ includes no and mild depression. c Q1: lower quartile, Q3: upper quartile. d Low-risk recommendations 1–14 units/week for men. Elevated risk drinking 15–21 units/week for men. High-risk drinking >21 for men. e Physical activity includes sports, physical training, house- and garden work, walks and bicycle rides incl. between home and work. Table 2 Women: distribution of demographic, personality, health and health behaviour variables by loneliness (N = 1628)   Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  44 (3%)  232 (14%)  509 (31%)  843 (52%)    Age, mean (years)  53 (SD 4.6)  52 (SD 4.1)  53 (SD 4.2)  53 (SD 4.4)  0.77  Cohabitation status  <0.0001  Alone  15 (34%)  86 (37%)  96 (19%)  63 (7%)    Not alone  29 (66%)  146 (63%)  413 (81%)  780 (93%)    Socio-economic position  <0.0001  Social class I  1 (2%)  13 (6%)  53 (10%)  95 (11%)    Social class II  8 (18%)  64 (28%)  152 (30%)  240 (28%)    Social class III  4 (9%)  49 (21%)  110 (22%)  234 (28%)    Social class IV  6 (14%)  45 (19%)  109 (21%)  170 (20%)    Social class V  4 (9%)  18 (8%)  41 (8%)  68 (8%)    Social class VI  21 (48%)  43 (19%)  44 (9%)  36 (4%)    Number of chronic illnessesa  <0.0001  0  7 (16%)  85 (37%)  192 (38%)  332 (39%)    1–2  14 (57%)  107 (46%)  270 (53%)  462 (55%)    ≥3  12 (27%)  40 (17%)  47 (9%)  49 (6%)    Depressionb  <0.0001  Yes  18 (41%)  30 (13%)  18 (4%)  17 (2%)    No  26 (59%)  202 (87%)  491 (96%)  826 (98%)    Personality  Neuroticism  Median 28  Median 23  Median 19  Median 17  <0.0001  Q1c 22  Q1 18  Q1 15  Q1 13    Q3c 31.5  Q3 27.5  Q3 24  Q3 21    Agreeableness  Median 35  Median 35  Median 35  Median 35  0.08  Q1 30  Q1 31  Q1 32  Q1 32    Q3 39  Q3 38  Q3 38  Q3 39    Conscientiousness  Median 30  Median 32  Median 34  Median 35  <0.0001  Q1 26  Q1 28  Q1 30  Q1 32    Q3 35  Q3 36  Q3 37  Q3 38    BMI  0.20  Underweight (<18.5)  1 (2%)  6 (3%)  8 (2%)  23 (3%)    Normal (18.5–25)  18 (41%)  119 (51%)  292 (57%)  440 (52%)    Overweight (25–30)  15 (34%)  71 (31%)  145 (28%)  278 (33%)    Obese (>30)  10 (23%)  36 (15%)  64 (13%)  102 (12%)    Smoking status  0.01  Smoker  17 (39%)  60 (26%)  116 (23%)  168 (20%)    Non-smoker  27 (61%)  172 (74%)  393 (77%)  675 (80%)    Alcohol  <0.0001  Not drinking  21 (48%)  47 (20%)  81 (16%)  137 (16%)    0 units/week  11 (25%)  104 (45%)  231 (45%)  374 (44%)    Low-risk drinking  7 (16%)  53 (23%)  142 (28%)  219 (26%)    7 units/weekd  5 (11%)  28 (12%)  55 (11%)  113 (13%)    Elevated risk drinking  8–14 units/weekd            High-risk drinking  >14 units/weekd            Physical activitye  0.1116  ≤3–4 h/week  13 (30%)  69 (30%)  113 (22%)  225 (27%)    ≥5 h/week  31 (70%)  163 (70%)  396 (78%)  618 (73%)      Loneliness  Yes, often  Yes, sometimes  Yes, seldom  No  P-value  N  44 (3%)  232 (14%)  509 (31%)  843 (52%)    Age, mean (years)  53 (SD 4.6)  52 (SD 4.1)  53 (SD 4.2)  53 (SD 4.4)  0.77  Cohabitation status  <0.0001  Alone  15 (34%)  86 (37%)  96 (19%)  63 (7%)    Not alone  29 (66%)  146 (63%)  413 (81%)  780 (93%)    Socio-economic position  <0.0001  Social class I  1 (2%)  13 (6%)  53 (10%)  95 (11%)    Social class II  8 (18%)  64 (28%)  152 (30%)  240 (28%)    Social class III  4 (9%)  49 (21%)  110 (22%)  234 (28%)    Social class IV  6 (14%)  45 (19%)  109 (21%)  170 (20%)    Social class V  4 (9%)  18 (8%)  41 (8%)  68 (8%)    Social class VI  21 (48%)  43 (19%)  44 (9%)  36 (4%)    Number of chronic illnessesa  <0.0001  0  7 (16%)  85 (37%)  192 (38%)  332 (39%)    1–2  14 (57%)  107 (46%)  270 (53%)  462 (55%)    ≥3  12 (27%)  40 (17%)  47 (9%)  49 (6%)    Depressionb  <0.0001  Yes  18 (41%)  30 (13%)  18 (4%)  17 (2%)    No  26 (59%)  202 (87%)  491 (96%)  826 (98%)    Personality  Neuroticism  Median 28  Median 23  Median 19  Median 17  <0.0001  Q1c 22  Q1 18  Q1 15  Q1 13    Q3c 31.5  Q3 27.5  Q3 24  Q3 21    Agreeableness  Median 35  Median 35  Median 35  Median 35  0.08  Q1 30  Q1 31  Q1 32  Q1 32    Q3 39  Q3 38  Q3 38  Q3 39    Conscientiousness  Median 30  Median 32  Median 34  Median 35  <0.0001  Q1 26  Q1 28  Q1 30  Q1 32    Q3 35  Q3 36  Q3 37  Q3 38    BMI  0.20  Underweight (<18.5)  1 (2%)  6 (3%)  8 (2%)  23 (3%)    Normal (18.5–25)  18 (41%)  119 (51%)  292 (57%)  440 (52%)    Overweight (25–30)  15 (34%)  71 (31%)  145 (28%)  278 (33%)    Obese (>30)  10 (23%)  36 (15%)  64 (13%)  102 (12%)    Smoking status  0.01  Smoker  17 (39%)  60 (26%)  116 (23%)  168 (20%)    Non-smoker  27 (61%)  172 (74%)  393 (77%)  675 (80%)    Alcohol  <0.0001  Not drinking  21 (48%)  47 (20%)  81 (16%)  137 (16%)    0 units/week  11 (25%)  104 (45%)  231 (45%)  374 (44%)    Low-risk drinking  7 (16%)  53 (23%)  142 (28%)  219 (26%)    7 units/weekd  5 (11%)  28 (12%)  55 (11%)  113 (13%)    Elevated risk drinking  8–14 units/weekd            High-risk drinking  >14 units/weekd            Physical activitye  0.1116  ≤3–4 h/week  13 (30%)  69 (30%)  113 (22%)  225 (27%)    ≥5 h/week  31 (70%)  163 (70%)  396 (78%)  618 (73%)    a Number of chronic illnesses includes: (1) Allergies, (2) Diabetes, (3) Hypertension, (4) Myocardial infarction, (5) Stroke, (6) Lung disease (incl. asthma, chronic bronchitis, emphysema), (7) Autoimmune disease and arthritis, (8) Cancer incl. leukaemia and (9) present other mental disorder. b Depression according to Major Depression Inventory-scale (41;42); ‘yes’ includes moderate and severe depression and ‘no’ includes no and mild depression. c Q1: lower quartile, Q3: upper quartile. d Low-risk recommendations 1–7 units/week or women. Elevated risk drinking 8–14 units/week for women. High-risk drinking >14 for women. e Physical activity includes sports, physical training, house- and garden work, walks and bicycle rides incl. between home and work. Loneliness and physical capability In men (crude models), loneliness was associated with lower handgrip strength [‘Often lonely’ −1.3 kg (95% CI −2.9;0.3) ‘Sometimes lonely’ −2.0 (−2.8;−1.1) reference group ‘Not lonely’] and for both genders with fewer chair raises [‘Often lonely’ women −3.3 rises/30 s (95% CI −5.5; −1.5, ‘Often lonely’ men −0.8(−2.0;0.4) ‘Sometimes lonely’ men −1.1 (−1.7;−0.5)] (table 3). After adjustment for confounders (Model 1), the observed associations attenuated. In women, there were no clear associations between loneliness and handgrip strength in any of the models (table 3). Adjusting for depressive symptoms in Model 2 and BMI, smoking status, alcohol consumption and physical activity in Model 3, did not change these conclusions (see Supplementary file). Table 3 Multivariate linear regression analysis of loneliness and handgrip strength/chair rise test, crude and adjusted models Handgrip strength  Chair rise test  Loneliness  N  Mean (kg)  Estimate (kg)  95% CI  Loneliness  N  Mean (rises/30 s)  Estimate (rises/30 s)  95% CI  Men  Crude model  Often lonely  110  48.9  −1.3  (−2.9;0.3)  Often lonely  89  21.1  −0.8  (−2.0;0.4)  Sometimes lonely  460  48.2  −2.0  (−2.8;−1.1)  Sometimes lonely  403  20.7  −1.1  (−1.7;−0.5)  Seldom lonely  1095  49.2  −0.9    Seldom lonely  986  21.5  −0.4  (−0.8;0.1)  Not lonely  1931  50.2  Ref.  (−1.5;−0.3)  Not lonely  1720  21.9  Ref.    P-value        <0.0001          0.004  Adjusted model  Often lonely  110  47.9  1.2  (−0.5;2.9)  Often lonely  89  20.0  0.2  (−1.0;1.4)  Sometimes lonely  460  46.2  −0.4  (−1.3;0.5)  Sometimes lonely  403  19.4  −0.3  (−0.9;0.3)  Seldom lonely  1095  46.4  −0.3  (−0.9;0.4)  Seldom lonely  986  19.7  −0.1  (−0.5;0.4)  Not lonely  1931  46.7  Ref.    Not lonely  1720  19.8  Ref.    P-value        0.23          0.77  Women  Crude model  Often lonely  44  30.8  −0.8  (−2.5;0.8)  Often lonely  43  18.3  −3.3  (−5.0; −1.5)  Sometimes lonely  232  31.2  −0.4  (−1.2;0.4)  Sometimes lonely  222  20.1  −1.5  (−2.3; −0.6)  Seldom lonely  509  31.1  −0.6  (−1.2;0.0)  Seldom lonely  489  21.3  −0.3  (−0.9;0.4)  Not lonely  843  31.6  Ref.    Not lonely  790  21.6  Ref.    P-value  0.25    <.0001  Model 1  Often lonely  44  31.3  1.0  (−0.7;2.6)  Often lonely  43  20.9  −0.3  (−2.0;1.4)  Sometimes lonely  232  30.6  0.3  (−0.6;1.1)  Sometimes lonely  222  20.9  −0.3  (−1.1;0.6)  Seldom lonely  509  30.0  −0.4  (−1.0;0.2)  Seldom lonely  489  21.2  0.1  (−0.5;0.7)  Not lonely  843  30.4  Ref.    Not lonely  790  21.2  Ref.    P-value  0.21    0.79  Handgrip strength  Chair rise test  Loneliness  N  Mean (kg)  Estimate (kg)  95% CI  Loneliness  N  Mean (rises/30 s)  Estimate (rises/30 s)  95% CI  Men  Crude model  Often lonely  110  48.9  −1.3  (−2.9;0.3)  Often lonely  89  21.1  −0.8  (−2.0;0.4)  Sometimes lonely  460  48.2  −2.0  (−2.8;−1.1)  Sometimes lonely  403  20.7  −1.1  (−1.7;−0.5)  Seldom lonely  1095  49.2  −0.9    Seldom lonely  986  21.5  −0.4  (−0.8;0.1)  Not lonely  1931  50.2  Ref.  (−1.5;−0.3)  Not lonely  1720  21.9  Ref.    P-value        <0.0001          0.004  Adjusted model  Often lonely  110  47.9  1.2  (−0.5;2.9)  Often lonely  89  20.0  0.2  (−1.0;1.4)  Sometimes lonely  460  46.2  −0.4  (−1.3;0.5)  Sometimes lonely  403  19.4  −0.3  (−0.9;0.3)  Seldom lonely  1095  46.4  −0.3  (−0.9;0.4)  Seldom lonely  986  19.7  −0.1  (−0.5;0.4)  Not lonely  1931  46.7  Ref.    Not lonely  1720  19.8  Ref.    P-value        0.23          0.77  Women  Crude model  Often lonely  44  30.8  −0.8  (−2.5;0.8)  Often lonely  43  18.3  −3.3  (−5.0; −1.5)  Sometimes lonely  232  31.2  −0.4  (−1.2;0.4)  Sometimes lonely  222  20.1  −1.5  (−2.3; −0.6)  Seldom lonely  509  31.1  −0.6  (−1.2;0.0)  Seldom lonely  489  21.3  −0.3  (−0.9;0.4)  Not lonely  843  31.6  Ref.    Not lonely  790  21.6  Ref.    P-value  0.25    <.0001  Model 1  Often lonely  44  31.3  1.0  (−0.7;2.6)  Often lonely  43  20.9  −0.3  (−2.0;1.4)  Sometimes lonely  232  30.6  0.3  (−0.6;1.1)  Sometimes lonely  222  20.9  −0.3  (−1.1;0.6)  Seldom lonely  509  30.0  −0.4  (−1.0;0.2)  Seldom lonely  489  21.2  0.1  (−0.5;0.7)  Not lonely  843  30.4  Ref.    Not lonely  790  21.2  Ref.    P-value  0.21    0.79  Crude model: outcome is solely dependent on loneliness. Model 1: Adjusted for age, occupational social class, cohabitation status, number of chronic illnesses and personality (neuroticism, agreeableness, conscientiousness). Joint effect of loneliness and occupational social class/cohabitation Occupational social class was associated with physical capability in both men and women. On average, men and women in the high occupational social class had 1 kg greater handgrip strength and could raise themselves two times more from the chair than those in the low occupational social class. Men who lived alone had, on average, 2 kg weaker grip strength, than men who did not live alone. Women who lived alone could on average raise themselves one fewer times from the chair, than women who did not live alone. Men who were jointly exposed to low occupational social class/living alone and loneliness generally showed lower hand-grip strength [Lonely/low social class −1.8 kg (95% CI −3.0;−0.6), Lonely/living alone −1.5 (−2.7;−0.4)] and fewer chair raises [Lonely/low social class −2.4 rises/30 s (95% CI −3.2;−1.6)] compared with the reference group of non-lonely persons with high occupational social class or who was not living alone. Of notice, lonely men living alone did not experience a poorer chair-rise test compared to the reference group [0.1 rises/30 s (95% CI −0.8;0.8)]. For women the joint variables were not associated with hand grip strength, but lonely women in low social class had fewer chair rises than the reference group (−2.3 rises/30 s (95% CI −3.3;−1.2)]. However, in conclusion the associations were no more than expected from the combination of the two exposures (P-values of the interaction term range 0.13–0.67) (table 4). Sensitivity analyses with those reporting loneliness ’often’, ’sometimes’ and ’seldom’ categorized together were run, but this did not change the overall conclusion. Table 4 Multivariate linear regression analysis Hand grip strength N = 5224  Men  Women  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  Loneliness  1. lonely (often/sometimes lonely)  570  46.2  −0.1  (−0.9;0.7)    276  30.3  0.4  (−0.3;1.1)    2. not lonely (seldom/not lonely)  3026  46.3  Ref.      1352  29.9  Ref.      Occupational social class  1. low social class (IV, V, VI)  1185  45.6  −1.3  (−1.9;−0.7)    605  29.6  −1.1  (−1.6;0.6)    2. high social class (I, II, III)  2411  46.9  Ref.      1023  30.7  Ref.      Cohabitation status  1. Living alone  525  45.0  -2.4  (-3.2;-1.5)    260  30.0  -0.2  (-0.9;0.5)    2. Not living alone  3071  47.4  Ref.      1368  30.2  Ref.      Loneliness/social class  1. Lonely/low social class  262  45.1  −1.8  (−3.0;−0.6)  0.44  137  30.4  −0.1  (−1.1;0.9)    2. Lonely/high social class  308  47.2  0.3  (−0.7;1.4)    139  30.2  −0.4  (−1.3;0.6)    3. Not lonely/low social class  923  45.8  −1.1  (−1.8;−0.4)    468  29.1  −1.4  (−2.0;−0.8)    4. Not lonely/high social class  2103  46.9  Ref.      884  30.5  Ref.    0.13  Loneliness/co-habitation status  1. Lonely/living alone  241  46.3  −1.5  (−2.7;−0.4)  0.14  101  30.9  0.6  (−0.5;1.7)    2. Lonely/not living alone  329  47.3  -0.6  (-1.5;0.4)    175  30.7  0.4  (-0.4;1.3)    3. Not lonely/living alone  284  45.1  -2.7      159  30.1  -0.2  (-1.1;0.7)    4. Not lonely/not living alone    47.8  Ref.  (-3.8;-1.7)    1193  30.3  Ref.    0.64  Chair rise test N = 4742  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Loneliness  1. Lonely (often/sometimes lonely)  492  19.2  −0.3  (−0.8;0.3)    265  20.2  −0.6  (−1.3;0.2)    2. Not lonely (seldom/not lonely)  2706  19.5  Ref.      1279  20.8  Ref.      Occupational social class  1. Low social class (IV, V, VI)  1033  18.4  −2.0  (−2.4;−1.5)    573  19.5  −2.0  (−2.6;−1.4)    2. High social class (I, II, III)  2165  20.3  Ref.      971  21.5  Ref.      Chair rise test N = 4742  Men  Women  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Cohabitation status  1. Living alone  462  19.2  −0.2  (−0.8;0.3)    249  20.1  −0.8  (−1.6;−0.1)    2. Not living alone  2736  19.5  Ref.      1295  20.9  Ref.      Loneliness/social class  1. Lonely/low social class  220  18.0  −2.4  (−3.2;−1.6)  0.35  129  19.5  −2.3  (−3.3;−1.2)    2. Lonely/high social class  272  20.4  0.0  (−0.8;0.7)    136  20.9  −0.9  (−1.9;0.1)    3. Not lonely/low social class  813  18.6  −1.9  (−2.3;−1.4)    444  19.6  −2.1  (−2.8;−1.5)    4. Not lonely/high social class  1893  20.4  Ref.      835  21.8  Ref.    0.26  Loneliness/co-habitation status  1. Lonely/living alone  204  19.8  0.1  (−0.8;0.8)  0.31  98  20.5  −1.1  (−2.2;0.1)    2. Lonely/not living alone  288  19.4  −0.4  (−1.1;0.3)    167  21.2  −0.4  (−1.2;0.5)    3. Not lonely/living alone  258  19.6  −0.2  (−0.9;0.5)    151  20.9  −0.7  (−1.6;0.2)    4. Not lonely/not living alone  2448  19.8  Ref.      1128  21.6  Ref.    0.67  Hand grip strength N = 5224  Men  Women  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  N  Mean (kg)  Est (kg)  95% CI  Test for interaction*  Loneliness  1. lonely (often/sometimes lonely)  570  46.2  −0.1  (−0.9;0.7)    276  30.3  0.4  (−0.3;1.1)    2. not lonely (seldom/not lonely)  3026  46.3  Ref.      1352  29.9  Ref.      Occupational social class  1. low social class (IV, V, VI)  1185  45.6  −1.3  (−1.9;−0.7)    605  29.6  −1.1  (−1.6;0.6)    2. high social class (I, II, III)  2411  46.9  Ref.      1023  30.7  Ref.      Cohabitation status  1. Living alone  525  45.0  -2.4  (-3.2;-1.5)    260  30.0  -0.2  (-0.9;0.5)    2. Not living alone  3071  47.4  Ref.      1368  30.2  Ref.      Loneliness/social class  1. Lonely/low social class  262  45.1  −1.8  (−3.0;−0.6)  0.44  137  30.4  −0.1  (−1.1;0.9)    2. Lonely/high social class  308  47.2  0.3  (−0.7;1.4)    139  30.2  −0.4  (−1.3;0.6)    3. Not lonely/low social class  923  45.8  −1.1  (−1.8;−0.4)    468  29.1  −1.4  (−2.0;−0.8)    4. Not lonely/high social class  2103  46.9  Ref.      884  30.5  Ref.    0.13  Loneliness/co-habitation status  1. Lonely/living alone  241  46.3  −1.5  (−2.7;−0.4)  0.14  101  30.9  0.6  (−0.5;1.7)    2. Lonely/not living alone  329  47.3  -0.6  (-1.5;0.4)    175  30.7  0.4  (-0.4;1.3)    3. Not lonely/living alone  284  45.1  -2.7      159  30.1  -0.2  (-1.1;0.7)    4. Not lonely/not living alone    47.8  Ref.  (-3.8;-1.7)    1193  30.3  Ref.    0.64  Chair rise test N = 4742  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Loneliness  1. Lonely (often/sometimes lonely)  492  19.2  −0.3  (−0.8;0.3)    265  20.2  −0.6  (−1.3;0.2)    2. Not lonely (seldom/not lonely)  2706  19.5  Ref.      1279  20.8  Ref.      Occupational social class  1. Low social class (IV, V, VI)  1033  18.4  −2.0  (−2.4;−1.5)    573  19.5  −2.0  (−2.6;−1.4)    2. High social class (I, II, III)  2165  20.3  Ref.      971  21.5  Ref.      Chair rise test N = 4742  Men  Women  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  N  Mean rises pr. 30 s  Est  95% CI  Test for interaction*  Cohabitation status  1. Living alone  462  19.2  −0.2  (−0.8;0.3)    249  20.1  −0.8  (−1.6;−0.1)    2. Not living alone  2736  19.5  Ref.      1295  20.9  Ref.      Loneliness/social class  1. Lonely/low social class  220  18.0  −2.4  (−3.2;−1.6)  0.35  129  19.5  −2.3  (−3.3;−1.2)    2. Lonely/high social class  272  20.4  0.0  (−0.8;0.7)    136  20.9  −0.9  (−1.9;0.1)    3. Not lonely/low social class  813  18.6  −1.9  (−2.3;−1.4)    444  19.6  −2.1  (−2.8;−1.5)    4. Not lonely/high social class  1893  20.4  Ref.      835  21.8  Ref.    0.26  Loneliness/co-habitation status  1. Lonely/living alone  204  19.8  0.1  (−0.8;0.8)  0.31  98  20.5  −1.1  (−2.2;0.1)    2. Lonely/not living alone  288  19.4  −0.4  (−1.1;0.3)    167  21.2  −0.4  (−1.2;0.5)    3. Not lonely/living alone  258  19.6  −0.2  (−0.9;0.5)    151  20.9  −0.7  (−1.6;0.2)    4. Not lonely/not living alone  2448  19.8  Ref.      1128  21.6  Ref.    0.67  Main and joint effects of loneliness, occupational social class and cohabitation status on physical capability Models have been adjusted for age, number of chronic illnesses and personality (neuroticism, agreeableness, conscientiousness). The loneliness/occupational social class model was also adjusted for cohabitation status, while the loneliness/cohabitation status model was adjusted for occupational social class. * Test for interaction by the inclusion of a product term: (1) loneliness + occupational social class + loneliness×occupational social class + covariates, (2) loneliness + cohabitation status + loneliness×cohabitation status + covariates, P values for the product term. Discussion In a large cross-sectional study of middle-aged adults we do not find evidence of poorer physical capability in lonely men and women. Men and women in low occupational social class or who were living alone had weaker hand grip strength and were limited in their chair raise ability, but there were no support for the differential vulnerability hypothesis. This study is among the first to investigate the association between loneliness and objectively measured physical capability among the middle-aged (50–60 years). In older populations the findings are mixed. A previous cross-sectional study among older people showed only a negligible association between loneliness and self-reported functional ability measured as independence from help in daily activities, which paradoxically showed that higher score on independence was associated with higher degree of loneliness.24 In contrast, a longitudinal study of middle-aged and older men and women (mean age 67 years), loneliness was found to increase the risk of self-reported functional limitations after taking into account the reciprocal association between functional limitations and loneliness as well as socio-demographic, health and health behaviour.5 Compared with this study, the cohort were older, and the findings were based on both self-reported measures of loneliness and physical limitation leaving a risk of dependent measurement error due to personality differences in reporting. Hence, studying objectively measured physical capability and adjusting for personality factors may produce results less prone to bias. In a longitudinal study, among older people (mean age 79 years), a strong association between loneliness and objectively measured motor decline was reported after adjustment for sociodemographic, baseline disability, mental and physical health and health behaviours.4 Despite its lack of adjustment for personality factors, this study lends support for a causal relation between loneliness and objectively measured physical function in older people. The younger age of the Copenhagen Aging and Midlife Biobank participants potentially in part explain our null-finding as the suggested physiological pathways11,12 between loneliness and health may not yet have resulted in detectable differences in physical capability. Furthermore, the prevalence of loneliness in middle age is rather low in the Northern European countries which we also find in this study (3% reports often to feel lonely), leaving only a smaller group exposed. A reason for this may be that most people by midlife have developed stable social networks, are well established on the labour market and most often have not suffered from social losses such as the death of a spouse or friend.8 The prevalence of loneliness is known to increase with age due to social losses, which may explain the negative association with physical capability in older people. Our findings suggest that even among the potential vulnerable groups in low occupational social class and those who live alone the effects of loneliness are not intensified. A wide variety of instruments have been used to measure loneliness, spanning from single item questions to comprehensive scales of many items e.g. the UCLA loneliness scale25 or de Jong Giervald scale.26 However, regardless of measurement differences there seems to be overall support for an association between loneliness and adverse health outcomes.1 The single item used in this study sheds light on the quantitative aspect of loneliness, where contact with others is sparse contrasting other measures evaluating feelings of loneliness.25 A Danish study among the elderly found that 43% who report unwanted aloneness are often lonely, while 59% of the often lonely also feel unwanted aloneness,7 indicating some difference in the unwanted aloneness-measure and the measure of loneliness. A meta-analysis found that older persons were more likely to associate loneliness with lack of activity, while middle-aged were more likely to recognize loneliness as absence of people.11 Therefore, the item used in this study may well encapsulate loneliness among the middle-aged. This one-item assessment has been used in prior research and the prevalence of often feeling lonely identified in this study (∼3%) is comparable to the prevalence in the middle-aged (4–5%) identified in National population-based surveys27 as well as that of a recent study in Denmark using the Three Item Loneliness UCLA Scale(4%).7 The characteristics of those who reported often to feel lonely in the CAMB participants were consistent with prior research,1,7,9,10,16,28 which may support the comparability of the single item measure of loneliness to the more comprehensive scores. Selection bias may have affected the results as CAMB non-respondents had lower education and were more likely to be unemployed, as well as they experienced higher all-cause mortality.17 These factors are associated with loneliness,4,5,8,9,28–30 and further it has been suggested that persons feeling lonely may be less inclined to participate in surveys.8 It is also possible that people, who are less physically capable, are less likely to participate in studies such as CAMB because of mobility limitations. Per design, we condition on participation which is a common cause of both loneliness and physical capability leaving a risk of selection bias.31 Considering the above-mentioned selection mechanisms, it is likely that lonely individuals that did in fact participate in CAMB will also be less likely to be physical incapable, and hence our findings are likely to underestimate the association between loneliness and physical capability. Strengths of this study include the large sample with comprehensive information on the participants, and that physical capability was measured by objective tests, which increases reliability and permits cross-cohort comparisons.20 A limitation of this study is the cross-sectional design which increases the risk of any associations being due to reverse causality. Due to demographic changes the number of 65+ years olds will increase dramatically in the years to come of which a considerable amount will be living alone, and thereby also in increased risk of loneliness and social isolation which potentially may harm the disadvantaged the most. Niedzwiedz et al. found loneliness to be most prevalent among the least wealthy groups across 14 European countries. This indicates the necessity for a continued focus on the potential health consequences of loneliness, social isolation in general among older people but especially among the least advantaged. In contrast to earlier studies among older people, we did not find an association between loneliness and physical capability in a large sample of middle-aged men and women, and there was no support for an interaction between loneliness and occupational social class and cohabitation status, respectively. This may be explained by the harmful effects of loneliness not yet to have had influence on physical capability in this age group. Selection bias may however have caused a true association to be underestimated. Further longitudinal studies are needed to clarify if and at what point in the life course loneliness begins to affect physical capability. Acknowledgements The authors thank the staff at Department of Public Health and National Research Centre for the Working Environment, who undertook the data collection. Further thanks to Kirsten Avlund, Helle Bruunsgaard, Nils-Erik Fiehn, Åse Marie Hansen, Poul Holm-Pedersen, Rikke Lund, Erik Lykke Mortensen and Merete Osler, who initiated and established the Copenhagen Aging and Midlife Biobank from 2009 to 2011. The authors acknowledge the crucial role of the initiators and steering committees of the Metropolit Cohort, The Copenhagen Perinatal Cohort and The Danish Longitudinal Study on Work, Unemployment and Health. The authors also want to thank professor Per Kragh Andersen, Section of Biostatistics for valuable advice regarding the statistical analyses. Disclaimers None. Supplementary data Supplementary data are available at EURPUB online. Funding This work and The Copenhagen Aging and Midlife Biobank were supported by grants from the VELUX FOUNDATION (VELUX26145, VELUX31539). Conflicts of interest: None declared. Key points In older people, loneliness has been associated with higher risk of decline in physical functioning. Little is known about this association in middle-aged individuals and if specific vulnerable groups can be identified. No association between loneliness and physical capability was found in this cohort of middle-aged men and women. This may be due to the harmful effects of loneliness not yet having resulted in detectable differences in physical capability in this age group. Lower occupational social class and living alone was associated with poorer physical capability, but there was no interaction between these factors and loneliness on physical capability. Further longitudinal studies are needed to clarify if and at what point in the life course loneliness begins to affect physical capability. References 1 Hawkley LC, Cacioppo JT. Loneliness. In: Reis HT, Sprecher S, editors. Encyclopedia of Human Relationships . Thousand Oaks: Sage, 2009: 985– 90. 2 Christiansen J, Larsen FB, Lasgaard M. Do stress, health behavior, and sleep mediate the association between loneliness and adverse health conditions among older people? Soc Sci Med  2016; 152: 80– 6. Google Scholar CrossRef Search ADS PubMed  3 Whisman MA. Loneliness and the metabolic syndrome in a population-based sample of middle-aged and older adults. Health Psychol  2010; 29: 550– 4. Google Scholar CrossRef Search ADS PubMed  4 Buchman AS, Boyle PA, Wilson RS, et al.   Loneliness and the rate of motor decline in old age: the Rush Memory and Aging Project, a community-based cohort study. BMC Geriatr  2010; 10: 77. Google Scholar CrossRef Search ADS PubMed  5 Luo Y, Hawkley LC, Waite LJ, Cacioppo JT. Loneliness, health, and mortality in old age: a national longitudinal study. Soc Sci Med  2012; 74: 907– 14. Google Scholar CrossRef Search ADS PubMed  6 Yang KM, Victor C. Age and loneliness in 25 European nations. Ageing Soc  2011; 31: 1368– 88. Google Scholar CrossRef Search ADS   7 Lasgaard M, Friis K. Loneliness in the population—prevalence and methodological considerations [in Danish Ensomhed i Befolkningen—forekomst og metodiske overvejelser.]Temaanalyse, vol. 3, Hvordan har du det? Aarhus, CFK Folkesundhed og kvalitetsudvikling, Denmark; 2015. 8 Pinquart M, Sorensen S. Influences on loneliness in older adults: a meta-analysis. Basic Appl Soc Psych  2001; 23: 245– 66. Google Scholar CrossRef Search ADS   9 Niedzwiedz CL, Richardson EA, Tunstall H, et al.   The relationship between wealth and loneliness among older people across Europe: is social participation protective? Prev Med  2016; 91: 24– 31. Google Scholar CrossRef Search ADS PubMed  10 Penninx BW, van Tilburg T, Kriegsman DM, et al.   Social network, social support, and loneliness in older persons with different chronic diseases. J Aging Health  1999; 11: 151– 68. Google Scholar CrossRef Search ADS PubMed  11 Hawkley LC, Cacioppo JT. Aging and loneliness: Downhill quickly? Curr Dir Psychol Sci  2007; 16: 187– 91. Google Scholar CrossRef Search ADS   12 Hackett RA, Hamer M, Endrighi R, et al.   Loneliness and stress-related inflammatory and neuroendocrine responses in older men and women. Psychoneuroendocrinology  2012; 37: 1801– 9. Google Scholar CrossRef Search ADS PubMed  13 Hansen AM, Andersen LL, Skotte J, et al.   Social class differences in physical functions in middle-aged men and women. J Aging Health  2014; 26: 88– 105. Google Scholar CrossRef Search ADS PubMed  14 Diderichsen F, Evans T, Whitehead M. The social basis of disparities in health. In: Evans T, Whitehead M, Diderichsen F, Bhuiya A, Wirth M, editors. Challenging Inequties in Health , 1st edn Oxford: OUP, 2001. 12– 23. Google Scholar CrossRef Search ADS   15 Lund R, Nilsson CJ, Avlund K. Can the higher risk of disability onset among older people who live alone be alleviated by strong social relations? A longitudinal study of non-disabled men and women. Age Ageing  2010; 39: 319– 26. Google Scholar CrossRef Search ADS PubMed  16 Cacioppo JT, Hughes ME, Waite LJ, et al.   Loneliness as a specific risk factor for depressive symptoms: cross-sectional and longitudinal analyses. Psychol Aging  2006; 21: 140– 51. Google Scholar CrossRef Search ADS PubMed  17 Lund R, Mortensen EL, Christensen U, et al.   Cohort Profile: The Copenhagen Aging and Midlife Biobank (CAMB). Int J Epidemiol. 2016; 45: 1044– 1053. 18 Christensen U, Krolner R, Nilsson CJ, et al.   Addressing social inequality in aging by the Danish occupational social class measurement. J Aging Health  2014; 26: 106– 27. Google Scholar CrossRef Search ADS PubMed  19 Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology  1999; 10: 37– 48. Google Scholar CrossRef Search ADS PubMed  20 Cooper R, Hardy R, Sayer A, Kuh DA. Life course approach to physical capability. In: Kuh D, Cooper R, Hardy R, Richards M, Ben-Shlomo Y, editors. A Life Course Approach to Healthy Ageing . London: Oxford University Press, 2014. 16– 31. 21 Rhodes RE, Smith NE. Personality correlates of physical activity: a review and meta-analysis. Br J Sports Med  2006; 40: 958– 65. Google Scholar CrossRef Search ADS PubMed  22 Mortensen EL, Flensborg-Madsen T, Molbo D, et al.   Personality in late midlife: associations with demographic factors and cognitive ability. J Aging Health  2014; 26: 21– 36. Google Scholar CrossRef Search ADS PubMed  23 Bech P, Rasmussen NA, Olsen LR, et al.   The sensitivity and specificity of the Major Depression Inventory, using the Present State Examination as the index of diagnostic validity. J Affect Disord  2001; 66: 159– 64. Google Scholar CrossRef Search ADS PubMed  24 Prieto-Flores ME, Forjaz MJ, Fernandez-Mayoralas G, et al.   Factors associated with loneliness of noninstitutionalized and institutionalized older adults. J Aging Health  2011; 23: 177– 94. Google Scholar CrossRef Search ADS PubMed  25 Russell DW. UCLA Loneliness Scale (Version 3): reliability, validity, and factor structure. J Pers Assess  1996; 66: 20– 40. Google Scholar CrossRef Search ADS PubMed  26 Gierveld JD, van TT. A 6-item scale for overall, emotional, and social loneliness—confirmatory tests on survey data. Res Aging  2006; 28: 582– 98. Google Scholar CrossRef Search ADS   27 The Danish Health Authority. Health of the Danes—The National Health Profile 2013 (In Danish: Danskernes sundhed - Den Nationale Sundhedsprofil 2013). 2014. http://www.danskernessundhed.dk/ (22nd May 2017) 28 Lauder W, Mummery K, Jones M, Caperchione C. A comparison of health behaviours in lonely and non-lonely populations. Psychol Health Med  2006; 11: 233– 45. Google Scholar CrossRef Search ADS PubMed  29 Newall NE, Chipperfield JG, Bailis DS, Stewart TL. Consequences of loneliness on physical activity and mortality in older adults and the power of positive emotions. Health Psychol  2013; 32: 921– 4. Google Scholar CrossRef Search ADS PubMed  30 Shiovitz-Ezra S, Ayalon L. Situational versus chronic loneliness as risk factors for all-cause mortality. Int Psychogeriatr  2010; 22: 455– 62. Google Scholar CrossRef Search ADS PubMed  31 Hernan MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology  2004; 15: 615– 25. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.

Journal

The European Journal of Public HealthOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off