Abstract
Objective
To determine whether there is an association between osteoarthritis (OA) and incident social isolation using data from the European Project on OSteoArthritis (EPOSA) study.
Design
Prospective, observational study with 12-18 months of follow-up.
Setting
Community-dwelling.
Participants
Older people living in six European countries.
Measurements
Social isolation was assessed using Lubben’s Social Network Scale and the Maastricht Social Participation Profile. Clinical OA of the hip, knee and hand was assessed according to ACR criteria. Demographic characteristics including, age, sex, multi-joint pain and medical comorbidities were assessed.
Results
Out of the 1,967 individuals with complete baseline and follow-up data, 382 (19%) were socially isolated and 1,585 were non-socially isolated at baseline; of these, 222 (13.9%) experienced social isolation during follow-up. Using logistic regression analyses, after adjustment for age, sex and country, four factors were significantly associated with incident social isolation: clinical OA, cognitive impairment, depression, and worse walking time. Compared to those without OA at any site or with only hand OA, clinical OA of the hip and/or knee, combined or not with hand OA, led to a 1.47 times increased risk of social isolation (95% confidence intervals: 1.03-2.09).
Conclusion
Clinical OA, present in one or two sites of the hip and knee, or in two or three sites of the hip, knee and hand increased the risk of social isolation, adjusting for cognitive impairment and depression, and worse walking times. Clinicians should be aware that individuals with OA may be at greater risk of social isolation.
Keywords: osteoarthritis, social isolation, EPOSA, epidemiology
Introduction
Social isolation has been defined as the consequence of a small social network with few social contacts. 1 It has been reported to be associated with poor physical and mental health. (2-4) Previous literature has suggested there is an association between musculoskeletal pain and social isolation. 1 This has been attributed to comorbid disease 2–6 and physical impairment. 7
The signs and symptoms commonly associated with osteoarthritis (OA), most notably joint pain and reduced function, may increase the risk of social isolation. 8, 9 People with OA often present with health risk factors that may increase their probability of social isolation. These include anxiety and depression, kinesiophobia, physical inactivity, reduced self-efficacy, which, depending on their severity, may reduce functional independence. 1, 10, 11 However, there has been limited research on the relationship between OA and social isolation. 1 Given that the high prevalence of OA 12, 13 in older people, affecting about 30% of persons over 65 and especially affecting lower limbs, is associated with poor quality of life and disability, a better understanding of social isolation in this specific population has become urgent. If studies show that age is associated with OA, preventative health and social interventions may be able to reduce the impact of social isolation and to improve quality of life. 8, 9
Given these considerations and basing its analysis on data from the European Project on OSteoArthritis (EPOSA) study, a large European cohort study with 12 to 18 months follow-up, the current study aimed to determine whether there is an association between OA and incident social isolation and to identify OA's unique contribution in the presence of other predictors for social isolation.
Methods
Population and Data Collection
Participants were identified from the EPOSA cohort. This is a population-based study of 2,942 adults between the ages of 65 to 85 years old, who are resident in six European countries (Germany, Italy, the Netherlands, Spain, Sweden, and the UK). More details about the EPOSA cohort are described elsewhere. 12
After obtaining written informed consent, all participants underwent a baseline assessment including a clinical examination and interview on health status performed at home or in a health care centre between November 2010, and November 2011. A follow-up interview was performed 12 to 18 months later.
The local research ethics committees approved the study (Germany: Universitat Ulm Ethikkommission [312/08]. Italy: Comitato Etico Provinciale Treviso [XLIV-RSA/AULSS7]. The Netherlands: Medisch Ethische Toetsingscommissie Vrije Universiteit Amsterdam [2002/141]. Spain: Comité Ético de Investigación Clínica del Hospital Universitario La Paz Madrid [PI-1080]. Sweden: Till forskningsetikkommittén vid Karolinska Instituted Stockholm [00-132]. UK: Hertfordshire Research Ethics Committee [10/H0311/59]).
Outcome
The current study’s primary outcome was the social isolation of the participants at baseline and 12 to 18 months later. Social isolation was assessed using two instruments (Supplementary Table S1): Lubben’s Social Network Scale (LSNS-6) 14 and the Maastricht Social Participation Profile (MSPP). 15
The LSNS-6 tool measures the number and frequency of social contacts with friends (three items) and family members (three items); each question is scored from zero (‘not at all’) to five (‘nine times or more a month’); the total score ranges from zero (indicating high isolation/few social resources) to 30 (indicating low isolation/many social resources); as proposed by Lubben a cut-off point of less than 12 indicates social isolation. 14
The MSPP measures the participant’s actual social participation over the preceding four weeks. 15 It is composed of three indexes: consumptive participation (CP) which refers to organised activities (six items), formal social participation (FSP) which refers, for example, to volunteer activities (three items) and informal social participation (ISP) which refers to contacts with family members, friends and acquaintances. The responses are classified using a Likert-type scale from 0 (‘not at all’) to 3 (‘more than twice a week’). Two types of scores are foreseen for each index: diversity (the number of items on which a respondent scored at least one) and frequency (mean score of the items). There is also a total diversity score that refers to the number of indices with a score of at least one. 15 Higher scores indicate more diverse or more frequent social participation.
Since the EPOSA study needed to harmonize data from 6 countries, it used 2 of the 3 MSPP subscales, the CP and FSP. As the third subscale (ISP) of the MSPP is similar to the Lubben scale, we used the latter, together with its cut-off value (12). The total diversity score of our analysis 15 was calculated considering the CP and the FSP of the MSPP and the LSNS-6. A participant’s diversified social participation index was calculated considering the median value of each of the following: total diversity, CP diversity, CP frequency, FSP diversity and FSP frequency. Since the MSPP does not define cut-off points, we used medians in our analysis as they are considered the most appropriate statistical method for evaluating continuous variable scores.
Social Isolation was defined as LSNS-6 < 12 15 or ≤ the median values of all five scores.
Clinical diagnosis of OA
The study’s primary aim was to estimate the effect of the clinical diagnosis of OA on the outcome variable (Supplementary Table S1).
In accordance with the clinical criteria of the American College of Rheumatology (ACR) 16 and European League Against Rheumatism, 17 the clinical diagnosis of OA was determined at baseline on the basis of the participant’s medical history and a physical examination, Clinical hand OA was diagnosed using specific sections of the AUSCAN. 18Clinical hip/knee OA, defined as the presence of OA in at least one or both of these joints, was diagnosed using specific WOMAC sections examining pain and stiffness. 8, 9 Pain in the hip/knee on at least one side was also evaluated during the physical examination. 18
As far as clinical OA was concerned, the participants were classified as: (1) no OA, (2) only hand OA, (3) hip and/or knee OA, (4) hip and/or knee OA combined with hand OA.
Baseline characteristics
The baseline characteristics (Supplementary Table S1) considered included: age, sex, country of residence, education level, marital status, income, comorbidity, medications being taken, joint replacements, clinical examination, health, and lifestyle characteristics.
Education level was categorized as up to elementary education versus higher levels of education. Marital status was categorized as being single or never married, divorced, widowed, living apart versus married or cohabitating, or a registered partnership. A monthly income capable of making ends meet was classified as ‘only with great difficulty’, ‘with some difficulty’, ‘fairly easily’ and ‘easily’. Comorbidity in our analysis referred to: obesity 19, cognitive impairment 20, anxiety and depression 21, self-reported presence of chronic conditions such as non-specific lung disease (i.e. asthma, chronic bronchitis or pulmonary emphysema, etc.), cardiovascular disease (i.e. cardiac valve disease, coronary heart diseases, arrhythmia, pacemaker, cardiac arrest, etc.), peripheral artery disease, diabetes mellitus, stroke, cancer and osteoporosis, lasting at least three months or which caused the individual to seek a physician’s attention (each dichotomized as present versus absent).
Medication used over the past two weeks referred to analgesic and/or anti-inflammatory drugs, the variable was dichotomized as medication use versus non-use. The presence of previous joint replacements was assessed by asking participants if they had ever had joint replacement surgery. If the response was affirmative, the participant was questioned about the location and time of and the reason for the joint replacement. Self-rated health assessment 22 classified as ‘fair’, ‘bad’, ‘very bad’, ‘good’ and ‘very good’. Health-related quality of life was assessed using the EuroQoL (EQ-5D, EQ VAS). 23
The clinical examination assessed grip strength and walking-test time. The mean of two right and left hand measurements by a dynamometer of the maximum grip strength was calculated. 24 The walking-test time was determined during a timed three-meter walk test. The participants' times were classified according to country-specific quartiles in order to take account the specific methodology used in each country.
Physical activity was measured using the validated LASA Physical Activity Questionnaire (LAPAQ) 24 which assesses the frequency and duration of activities such as: walking, cycling, gardening, light and heavy household work, and participation in sports over the past two weeks. The total time dedicated to physical activity was calculated in minutes/day and the total amount of energy was expressed as kcal/day.
Health characteristics considered physical function, pain, and stiffness in hand, hip and/or knee. These were assessed using subscales of the AUSCAN 18 and of WOMAC. 8, 9 Hip/knee pain and stiffness were defined as the maximum value reported across two joints.
All the AUSCAN and WOMAC subscales (responses ranging from 0=none to 4=extreme) were normalized to a 0 to 100 range: higher scores indicated worse health status. 8, 9
Statistical Analysis
Only participants with complete data on all the variables were included in the analyses. As the age and sex distribution varied in the cohorts from the different countries participating in the EPOSA study, they calculated a weighting variable for each individual within each country. The weights, which were based on sex and five-year age categories according to the 2010 Standard European Population, were applied only in the descriptive and not in the analytic statistics. 12 Categorical variables were reported as proportions, and continuous variables as means and standard deviations, or medians with interquartile ranges (IQRs). Significant differences between the groups of participants were evaluated using Wilcoxon rank-sum test, or χ2 testing.
The predictors of social isolation were assessed using logistic regression models adjusted for sex, age, and country. Each independent variable was tested using a significance level P≤0.20 as the screening criterion. 25 The appropriate categories for the categorical variables and the linearity in the logit for continuous variables were then examined, and the scale for the continuous variables in the logit was checked.
A multivariable model containing all the variables identified for inclusion was fitted using a stepwise selection procedure (P to enter =0.15 and P to remain =0.10) to select them. Those excluded were controlled for confounding effects. The collinearity of the predictor variables was assessed with the variance inflation factor, using a cut-off of 2 to exclude a variable. All the interactions between the variables in the final model were checked; interaction terms with P ≤0.10 were retained in the final model. Odds ratios (ORs) were presented with their 95% confidence intervals (CIs).
Statistical analyses were performed with SAS software (SAS Institute Inc, Cary, NC), version 9.4. All the tests were two-sided, and P <0.05 was considered statistically significant.
Results
Out of the 2,942 individuals originally enrolled in the EPOSA, 1,967 (67%) presented complete baseline and follow-up data on all the variables used in the analyses.
With respect to the participants with complete follow-up data (n=1,967), those whose data were uncomplete (n=488) were significantly older, more likely to be female, single/divorced/widowed or living apart, and predominantly Dutch (Supplementary Table S2).
The median age of the 1,967 participants was 73 years (IQR: 70 to 77 years); 50% were women, and almost 30% had a diagnosis of OA (Table 1). At baseline, 382 (19%) of the participants were categorized as socially isolated and 1,585 as non-socially isolated. The non-socially isolated individuals differed from the socially isolated participants in many important ways (data not shown), including being younger, being residents in all countries except Spain, being more educated and with higher income. The non-socially isolated people also presented a significantly lower prevalence of cognitive impairment, anxiety, depression, chronic lung disease, and stroke. They were less likely to use analgesic/anti-inflammatory medications. They reported a lower rate of clinical OA, but only when all the sites (hand, hip and/or knee) were considered; they had a better health status, and were more likely to partake in physical activity. They were quicker on the walking test, were stronger, and had less physical function impairment and a lower perception of pain.
Table 1. Baseline characteristics for social isolation at 12-18 months Follow-up.
| Baseline Characteristics | Total(n=1,967) | 12-18 months Follow-up(n=1,585) |
P-value | ||
|---|---|---|---|---|---|
| Isolated (n=222) | Not isolated(n=1,363) | ||||
| Age, mean±SD, median (IQR), years | 73.7±5.3 73 (70-77) | 74.0±5.0 73.5 (70-78) | 73.3±4.8 73 (70-76) | 0.058 | |
| Female sex, % | 49.6 | 49.7 | 54.1 | 0.234 | |
| Country, % | Germany | 13.3 | 9.8 | 15.2 | <.001 |
| Italy | 15.5 | 20.3 | 15.3 | ||
| The Netherlands | 17.5 | 20.5 | 17.6 | ||
| Spain | 19.7 | 25.9 | 14.8 | ||
| Sweden | 20.4 | 14.9 | 22.2 | ||
| UK | 13.6 | 8.7 | 14.9 | ||
| Up to elementary education, % | 41.7 | 47.0 | 36.8 | 0.005 | |
| Marital status (single/divorced/widowed/living apart), % | 32.4 | 32.6 | 31.7 | 0.804 | |
| Income, % | With great difficulty | 2.7 | 2.4 | 2.5 | 0.001 |
| With some difficulty | 13.9 | 18.5 | 10.9 | ||
| Fairly easily | 50.4 | 53.7 | 50.3 | ||
| Easily | 33.1 | 25.4 | 36.4 | ||
| Obesity (BMI ≥30 kg/m2), % | 24.7 | 26.1 | 23.8 | 0.479 | |
| Cognitive impairment (MMSE ≤23), % | 6.1 | 9.7 | 4.2 | 0.001 | |
| Anxiety (HADS ≥8), % | 17.8 | 19.1 | 16.1 | 0.292 | |
| Depression (HADS ≥8), % | 9.6 | 12.0 | 6.0 | 0.001 | |
| Chronic lung disease, % | 12.5 | 11.4 | 11.5 | 0.97 | |
| Cardiovascular disease, % | 23.6 | 22.1 | 23.1 | 0.744 | |
| Peripheral artery disease, % | 9.8 | 10.0 | 9.3 | 0.728 | |
| Diabetes mellitus, % | 11.6 | 12.7 | 10.9 | 0.453 | |
| Stroke, % | 4.5 | 5.8 | 3.7 | 0.143 | |
| Cancer, % | 13.9 | 10.6 | 14.9 | 0.101 | |
| Osteoporosis, % | 14.7 | 16.4 | 14.4 | 0.448 | |
| Analgesic/Anti-inflammatory medication, % | 24.8 | 27.6 | 22.8 | 0.128 | |
| Clinical osteoarthritis, % | No | 70.5 | 63.9 | 72.2 | 0.007 |
| Hand | 8.4 | 6.9 | 8.8 | ||
| Hip and/or knee | 13.6 | 20.3 | 12.5 | ||
| Hand and (hip and/or knee) | 7.6 | 9.0 | 6.5 | ||
| Joint replacements, % | 10.9 | 12.8 | 11.0 | 0.452 | |
| Self-rated health (fair/bed/very bed), % | 33.5 | 41.0 | 28.9 | <.001 | |
| EQ-5D# (time trade-off), mean±SD, | 0.8±0.2 | 0.82±0.20 | 0.84±0.18 | 0.12 | |
| median (IQR) | 0.8(0.7-1.0) | 0.8(0.7-1.0) | 0.85(0.73-1.0) | ||
| EQ VAS§ (health state today), mean±SD | 75.9±17.7 | 73.6±18.5 | 77.1±17.3 | 0.006 | |
| median (IQR) | 80(70-90) | 75 (65-90) | 80(70-90) | ||
| Grip strength†, mean±SD, | 28.0±10.1 | 26.5±9.4 | 28.5±10.3 | 0.054 | |
| median (IQR), kg | 26.5(20-35) | 25.5(20.0-32.5) | 27.0(20.5-36.5) | ||
| Walking time‡,, % | ≤ Q1 | 27.4 | 15.5 | 31.3 | <.001 |
| Q1-Q2 | 26.2 | 30.5 | 25.6 | ||
| Q2-Q3 | 23.6 | 27.3 | 22.8 | ||
| > Q3 | 22.8 | 26.8 | 20.3 | ||
| Total physical activity time (LAPAQ), mean±SD | 201.8±130.8 | 201.1±137.9 | 207.3±126.8 | 0.11 | |
| median (IQR), min/day | 180.0(110.7-262.5) | 169.3(105.1-258.9) | 184.3(120.0-267.9) | ||
| Total physical activity amount (LAPAQ), mean±SD | 870.8±644.4 | 824.0±635.2 | 907.3±635.4 | 0.01 | |
| median (IQR), kcal/day | 717.2(451.8-1101.1) | 6501.4(440.0-1026.1) | 754.5(489.6-1148.9) | ||
| WOMAC* hip/knee physical function score, mean±SD | 7.4±12.4 | 9.3±13.6 | 6.6±11.9 | 0.001 | |
| median (IQR) | 1(0-10) | 3 (0-13) | 0(0-8) | ||
| WOMAC* hip/knee pain score, mean±SD | 9.6±14.1 | 10.8±13.4 | 8.8±13.7 | 0.006 | |
| median (IQR) | 0(0-15) | 5 (0-20) | 0(0-10) | ||
| WOMAC* hip/knee stiffness score, mean±SD | 11.9±18.1 | 14.3±19.0 | 11.9±17.5 | 0.04 | |
| median (IQR) | 0(0-25) | 0(0-25) | 0(0-25) | ||
| AUSCAN* hand physical function score, mean±SD | 7.7±13.9 | 8.8±14.5 | 7.2±13.5 | 0.025 | |
| median (IQR) | 0(0-8) | 0(0-11) | 0(0-8) | ||
| AUSCAN* hand pain score, mean±SD, | 7.2±14.9 | 6.8±14.2 | 6.9±14.6 | 0.76 | |
| median (IQR) | 0(0-5) | 0(0-5) | 0(0-5) | ||
| AUSCAN* hand stiffness score, mean±SD, | 9.4±17.9 | 10.8±18.3 | 9.4±18.0 | 0.453 | |
| median (IQR) | 0(0-0) | 0(0-25) | 0(0-25) | ||
Weighted data except numbers of participants, age, and sex. BMI, body mass index; MMSE, Mini-Mental State Examination score; HADS, Hospital Anxiety and Depression Scales; Q1, Q2, Q3, quartiles; EQ-5D, Health status using five dimensions; EQ VAS, Health status using the visual analogue scale; LAPAQ, LASA Physical Activity Questionnaire; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index; AUSCAN, AUStralian CANadian Osteoarthritis Hand Index; SD, standard deviation; IQR, interquartile range.
Possible scores range from -0.594 to 1, lower values indicating worse health status.
Possible scores range from 0 to 100, with 0 indicating worse health status.
Lower values indicate worse performance.
By country quartiles, class ≤ Q1 indicates best performance, class > Q3 indicates worst performance.
Possible scores range from 0 to 100, with 100 indicating worse status.
According to logistic regression analyses, when clinical OA was adjusted for age, sex and country, it was associated with social isolation only when it was present in all three sites. But this association was not confirmed in the multivariable model in which low education, low income, depression, joint replacement (protective) and a pattern of fair/bad/very bad self-rated health and anxiety were found to be associated with social isolation (data not shown).
Baseline demographic and clinical characteristics of incident cases of social isolation
Out of the 1,585 non-socially isolated individuals at baseline, 222 (13%) incident cases of social isolation were found 12 to 18 months after baseline (Table 1).The participants who had become socially isolated were less educated and predominantly Spanish, Dutch and Italian. They reported that their income easily or fairly easily covered their needs. They presented higher percentages of cognitive impairment, depression, and clinical OA of the hip and/or knee and of the hand and hip and/or knee. They presented worse self-rated health status (self-rated health and EQ VAS), and lower levels of physical activity. They had slower walking times, higher levels of physical functioning impairments, worst stiffness in the hip/knee and hand, and a higher perception of pain in the hip/knee.
Predictors of incident social isolation
According to logistic regression analyses, adjusted for age, sex and country, clinical OA was associated with incident social isolation. As only the hip and/or knee level was significantly associated to social isolation (data not shown), at the next step “no OA” or “only hand OA” was compared to “hip and/or knee OA” and “hand OA and hip and/or knee OA”. The other 11 univariable predictors of social isolation that were identified were: income, cognitive impairment, depression, cancer, self-rated health, EQ-5D, EQ VAS, walking time, physical function, pain (dichotomized in correspondence with the third quartile as <15 vs ≥15) and stiffness of the WOMAC hip/knee (Table 2).
Table 2. Univariable and Multivariable models for social isolation 12-18 months after baseline.
| Univariable adjusted model | Multivariable adjusted model | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| β | SE | P-value | OR | 95% CI | β | SE | P-value | OR | 95% CI | |
| Clinical osteoarthritis | 0.004 | 0.032 | ||||||||
| no/only hand | 1.00 | 1.00 | ||||||||
| (hip and/or knee); hand and (hip and/or knee) | 0.487 | 0.169 | 1.63 | 1.17-2.27 | 0.384 | 0.179 | 0.032 | 1.47 | 1.03-2.09 | |
| Income | 0.160 | |||||||||
| With great difficulty | 1.00 | |||||||||
| With some difficulty | 0.6541 | 0.5138 | 0.203 | 1.92 | 0.70-5.27 | |||||
| Fairly easily | 0.3325 | 0.4938 | 0.501 | 1.39 | 0.53-3.67 | |||||
| Easily | 0.1102 | 0.5106 | 0.829 | 1.12 | 0.41-3.04 | |||||
| Cognitive impairment (MMSE ≤23) | 0.673 | 0.275 | 0.015 | 1.96 | 1.14-3.36 | 0.640 | 0.282 | 0.022 | 1.90 | 1.09-3.29 |
| Depression § (HADS ≥8) | 0.592 | 0.242 | 0.014 | 1.81 | 1.13-2.90 | -0.434 | 0.435 | |||
| Male sex | 0.332 | 0.66 | 0.28-1.54 | |||||||
| Female sex | 0.001 | 2.78 | 1.50-5.15 | |||||||
| Cancer | -0.390 | 0.235 | 0.097 | 0.68 | 0.43-1.07 | |||||
| Self-rated health (fair/bad/very bad) | 0.298 | 0.164 | 0.068 | 1.35 | 0.98-1.86 | |||||
| EQ-5D# (time trade-off) | -0.639 | 0.373 | 0.087 | 0.53 | 0.25-1.10 | |||||
| EQ VAS§ (health state today) | -0.007 | 0.004 | 0.076 | 0.99 | 0.99-1.00 | |||||
| Walking time‡ | 0.0008 | 0.003 | ||||||||
| ≤ Q1 | 1.00 | 1.00 | ||||||||
| Q1-Q2 | 0.742 | 0.224 | 0.001 | 2.10 | 1.36-3.26 | 0.748 | 0.225 | <0.001 | 2.11 | 1.36-3.28 |
| Q2-Q3 | 0.807 | 0.229 | <0.001 | 2.24 | 1.43-3.51 | 0.750 | 0.231 | 0.001 | 2.12 | 1.35-3.33 |
| > Q3 | 0.870 | 0.238 | <0.001 | 2.39 | 1.50-3.81 | 0.724 | 0.244 | 0.003 | 2.06 | 1.28-3.33 |
| WOMAC * hip/knee physical function score | 0.010 | 0.005 | 0.069 | 1.01 | 1.00-1.02 | |||||
| WOMAC* hip/knee pain score (≥15) | 0.436 | 0.160 | 0.006 | 1.55 | 1.13-2.12 | |||||
| WOMAC* hip/knee stiffness score | 0.006 | 0.004 | 0.139 | 1.01 | 1.00-1.01 | |||||
Models adjusted for age, sex and country.
BMI, body mass index; MMSE, Mini-Mental State Examination score; HADS: Hospital Anxiety and Depression Scales; OA: osteoarthritis; Q1, Q2, Q3; quartiles; EQ-5D: Health status using five dimensions; EQ VAS: Health status using the visual analogue scale; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index; β, regression coefficient; SE, standard error; OR, odds ratio; CI, confidence intervals.
Possible scores range from -0.594 to 1, lower values indicating worse health status.
Possible scores range from 0 to 100, with 0 indicating worse health status.
By country quartiles, reference class ≤ Q1 indicates best performance, class > Q3 indicates worst performance.
Possible scores range from 0 to 100, with 100 indicating worse health status.
Four variables proved significant in the multivariable analysis (Table 2): clinical OA, cognitive impairment, depression, and walking time. The distribution of these variables at baseline among those who will develop social isolation versus those who will remain socially active is presented in Figure 1. When we controlled for confounding factors, no mediators were found. The resulting model uncovered only one significant interaction: depression and sex.
Figure 1.
Proportion of isolated and not isolated participants at follow-up by baseline factors.
Weighted data.
Walking time is classified by country quartiles, reference class ≤ Q1 indicates best performance.
The estimate of the odds ratio for clinical OA of the hip and/or knee combined or not with hand OA was 1.47 (95% CI: 1.03 to 2.09) times greater than the odds for someone with similar characteristics (with respect to the other covariates in the model) without OA or with only hand clinical OA. The estimate of the odds ratio for cognitive impairment was 1.90 (95% CI: 1.09-3.29). Walking time was associated to social isolation with odds at each level greater than 1 (Q1-Q2: OR 2.11, 95% CI: 1.36-3.28; Q2-Q3: OR 2.12, 95% CI: 1.35-3.33; >Q3: OR 2.06, 95% CI: 1.28-3.33). The odds of incident social isolation for a person with worse walking times was 2 times greater than the odds for a person whose walking time was better (≤Q1).
There was an interaction between depression and sex: each increased the odds of social isolation in the presence of the other. Females who were depressed were found to be almost three times more likely to become socially isolated 12 to 18 months after baseline with respect to their female counterparts without depression (OR 2.78, 95% CI: 1.50-5.15).
Discussion
This study shows that OA increases the risk of incident social isolation onset. People with hip and/or knee OA combined or not with hand OA at baseline are at increased risk of social isolation in a community cohort. The presence of cognitive impairment and worse walking times in both genders and depression in the females also increased the risk of becoming socially isolated during the follow-up period.
While it is absolutely known that a complex of deficits, including mobility limitations predict isolation, we focused on OA, because in our opinion it is interesting that OA remains an independent predictor in the multivariate analyses, even after adjusting for functional limitations and pain. Moreover, as we have previously reported 26 OA has an independent effect also on self‐reported physical function impairment, even after adjusting for pain, which can probably be explained by the “expected pain” that OA may cause during physical activity. Probably the fear of pain is more important than pain itself as far as OA patients are concerned. This would explain why OA independently predicts isolation.
A large meta-analysis examining 148 studies assessing the association of social isolation and mortality reported that individuals who had more supportive social relationships had a lower mortality risk. 27 Similarly, socially isolated older adults tend to have an increased risk of experiencing a decline in mobility. 28 Finally, social isolation is associated with an increased risk of cardiovascular disease 29 and dementia. 30 Since social isolation is a potentially reversible condition, increasing research efforts are attempting to identify as early as possible socially isolated older people.
This study was the first analysis to assess an association between OA and social isolation based on prospective data. Several explanations for the association could be proposed. First, people with OA are more disabled and show poorer physical performance, which are both independent risk factors for social isolation. 31 Moreover, worse walking times in the patients studied were found to be a significant predictor of social isolation. Second, OA has also been associated with depression. 31 In the current study, depression was in fact another significant predictor of social isolation.32
The findings suggest that people with OA are at increased risk of social isolation. Given the important negative health outcomes associated with social isolation, interventions should be developed and tested to address this unmet healthcare need. These should include forms of physical activity, social engagement and community participation as well as some type of psychological assistance. 33 According to a systematic review on interventions to reduce social isolation, educational and social activities targeting specific groups can lower social isolation in older people. 34 Referring older adults with OA to social activity/senior centres in their area offering these types of activities may be useful especially when these interventions are specifically designed for older people with OA presenting physical impairments limiting social participation.
The study presents some limitations. First, the presence of comorbidity was evaluated on the basis of self-reported information and was not ascertained clinically. Self-reported information regarding comorbidities has nevertheless a good accuracy compared to gold standard methods of diagnosis. 35 Second, although a 12 to 18 month follow-up time may be considered insufficient to determine incident cases of social isolation, a large number of participants did indeed become socially isolated during that time period. Third, variables linked to life events such as the death of a family member or friend or being admitted to hospital, which may be important predictors of isolation, were not considered by the designers of the EPOSA study. Finally, the high number of participants whose data was incomplete might have caused a selection bias. Nevertheless, the high number of participants living in six different European nations who were studied can be considered the study’s strength. Moreover, standardized international guidelines were used for the clinical diagnosis of OA in all participants. 12
Conclusion
In conclusion, data from the EPOSA study suggest that OA is associated with incident social isolation, adjusting for cognitive impairment, depression, and worse walking times. Future research is warranted.
Supplementary Material
Impact Statement.
We certify that this work is novel.
The potential impact of this research on clinical care or health policy includes the following:
-
−
The study is the first to assess the association between OA and social isolation using prospective data.
-
−
OA is associated with incident social isolation, adjusting for cognitive impairment, depression, and worse walking time.
-
−
The findings suggest that interventions addressing this topic could/should be developed.
Acknowledgments
The corresponding author Siviero affirms that she has listed everyone who contributed significantly to the work and has obtained written consent from all contributors who are not authors and are named in this section.
Appreciation is expressed to Linda Inverso Moretti for assistance in editing the manuscript.
Funding
The study was supported by a non-commercial private funder.
The Indicators for Monitoring COPD and Asthma - Activity and Function in the Elderly in Ulm study (IMCA - ActiFE) was supported by the European Union (No.: 2005121) and the Ministry of Science, Baden-Württemberg. The Italian cohort study is part of the National Research Council Project on Aging (PNR). The Longitudinal Aging Study Amsterdam (LASA) is financially supported by the Netherlands Ministry of Health Welfare and Sports, Directorate of Long-Term Care. The Peñagrande study was partially supported by the National Fund for Health Research (Fondo de Investigaciones en Salud) of Spain (project numbers FIS PI 05/1898; FIS RETICEF RD06/0013/1013 and FIS PS09/02143). The Swedish Twin Registry is supported in part by the Swedish Ministry of Higher Education. The Hertfordshire Cohort Study is funded by the Medical Research Council of Great Britain, Arthritis Research UK, the British Heart Foundation and the International Osteoporosis Foundation.
Smith and Cooper are supported by the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre (BRC). The views expressed are those of these author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.
Footnotes
Conflicts of Interest: The authors have no conflicts.
Author Contributions: Siviero has full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design. Smith, Stubbs, Zambon, Dennison, Edwards, Cooper, van Schoor, van der Pas, Schaap, Denkinger, Peter, Otero, Castell, Pedersen, Deeg, Maggi.
Acquisition of data. Siviero, Limongi, Zambon, Cooper, Dennison, Edwards, Timmermans, van der Pas, Schaap, van Schoor, Denkinger, Peter, Herbolsheimer, Otero, Castell, Pedersen, Deeg, Maggi.
Analysis and interpretation of data. Siviero, Veronese, Smith, Stubbs, Limongi, Dennison, Edwards, Cooper, Timmermans, van Schoor, van der Pas, Schaap, Denkinger, Peter, Herbolsheimer, Castell, Pedersen, Deeg, Maggi.
Preparation of manuscript. All the authors contributed to the drafting and critical revision of the manuscript, and all the authors have approved the final manuscript.
Sponsor’s Role: The Sponsor had no role in the study design, the collection, analysis and interpretation of the data, in the writing of the report and in the decision to submit the article for publication.
References
- [1].Ethgen O, Vanparijs P, Delhalle S, et al. Social support and health-related quality of life in hip and knee osteoarthritis. Quality of Life Research. 2004;13:321–330. doi: 10.1023/B:QURE.0000018492.40262.d1. [DOI] [PubMed] [Google Scholar]
- [2].Veronese N, Cereda E, Maggi S, et al. Osteoarthritis and mortality: A prospective cohort study and systematic review with meta-analysis. Seminars in arthritis and rheumatism. 2016;46:160–167. doi: 10.1016/j.semarthrit.2016.04.002. [DOI] [PubMed] [Google Scholar]
- [3].Veronese N, Punzi L, Sieber C, et al. Sarcopenic osteoarthritis: a new entity in geriatric medicine? European Geriatric Medicine. 2018;9:141–148. doi: 10.1007/s41999-018-0034-6. [DOI] [PubMed] [Google Scholar]
- [4].Veronese N, Smith T, Reginster JY, et al. Osteoarthritis increases the risk of cardiovascular disease: Data from the osteoarthritis initiative. Osteoporosis International. 2017;28:S58–S59. doi: 10.1007/s12603-017-0941-0. [DOI] [PubMed] [Google Scholar]
- [5].Veronese N, Stubbs B, Solmi M, et al. Association between lower limb osteoarthritis and incidence of depressive symptoms: data from the osteoarthritis initiative. Age and ageing. 2016 doi: 10.1093/ageing/afw216. [DOI] [PubMed] [Google Scholar]
- [6].Veronese N, Stubbs B, Solmi M, et al. Knee Osteoarthritis and Risk of Hypertension: A Longitudinal Cohort Study. Rejuvenation research. 2017 doi: 10.1089/rej.2017.1917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].McAlindon T, Cooper C, Kirwan J, et al. Determinants of disability in osteoarthritis of the knee. Annals of the rheumatic diseases. 1993;52:258. doi: 10.1136/ard.52.4.258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Bellamy N, Buchanan WW, Goldsmith CH, et al. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. The Journal of rheumatology. 1988;15:1833–1840. [PubMed] [Google Scholar]
- [9].Roorda L, Jones C, Waltz M, et al. Satisfactory cross cultural equivalence of the Dutch WOMAC in patients with hip osteoarthritis waiting for arthroplasty. Annals of the rheumatic diseases. 2004;63:36–42. doi: 10.1136/ard.2002.001784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Herbolsheimer F, Schaap LA, Edwards MH, et al. Physical activity patterns among older adults with and without knee osteoarthritis in six European countries. Arthritis care & research. 2016;68:228–236. doi: 10.1002/acr.22669. [DOI] [PubMed] [Google Scholar]
- [11].Timmermans EJ, de Koning EJ, van Schoor NM, et al. Within-person pain variability and physical activity in older adults with osteoarthritis from six European countries. BMC musculoskeletal disorders. 2019;20:12. doi: 10.1186/s12891-018-2392-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Van Der Pas S, Castell MV, Cooper C, et al. European project on osteoarthritis: design of a six-cohort study on the personal and societal burden of osteoarthritis in an older European population. BMC musculoskeletal disorders. 2013;14:138. doi: 10.1186/1471-2474-14-138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Hunter DJ, Bierma-Zeinstra S. Osteoarthritis. Lancet (London, England) 2019;393:1745–1759. doi: 10.1016/S0140-6736(19)30417-9. [DOI] [PubMed] [Google Scholar]
- [14].Lubben J, Blozik E, Gillmann G, et al. Performance of an abbreviated version of the Lubben Social Network Scale among three European community-dwelling older adult populations. The Gerontologist. 2006;46:503–513. doi: 10.1093/geront/46.4.503. [DOI] [PubMed] [Google Scholar]
- [15].Mars GM, Kempen GI, Post MW, et al. The Maastricht social participation profile: development and clinimetric properties in older adults with a chronic physical illness. Quality of Life Research. 2009;18:1207. doi: 10.1007/s11136-009-9537-4. [DOI] [PubMed] [Google Scholar]
- [16].Altman RD. Seminars in arthritis and rheumatism. Vol. 20. Elsevier; 1991. Classification of disease: osteoarthritis; pp. 40–47. [DOI] [PubMed] [Google Scholar]
- [17].Zhang W, Doherty M, Leeb B, et al. EULAR evidence-based recommendations for the diagnosis of hand osteoarthritis: report of a task force of ESCISIT. Annals of the rheumatic diseases. 2009;68:8–17. doi: 10.1136/ard.2007.084772. [DOI] [PubMed] [Google Scholar]
- [18].Bellamy N, Campbell J, Haraoui B, et al. Clinimetric properties of the AUSCAN Osteoarthritis Hand Index: an evaluation of reliability, validity and responsiveness. Osteoarthritis and cartilage. 2002;10:863–869. doi: 10.1053/joca.2002.0838. [DOI] [PubMed] [Google Scholar]
- [19].Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organization technical report series. 2000;894:1–253. i-xii. [PubMed] [Google Scholar]
- [20].Tombaugh TN, McIntyre NJ. The mini-mental state examination: a comprehensive review. Journal of the American Geriatrics Society. 1992;40:922–935. doi: 10.1111/j.1532-5415.1992.tb01992.x. [DOI] [PubMed] [Google Scholar]
- [21].Zigmond A, Snaith R. The hospital anxiety and depression scale. ActaPsychiatrScand. 1983;67:361–370. doi: 10.1111/j.1600-0447.1983.tb09716.x. [DOI] [PubMed] [Google Scholar]
- [22].van Sonsbeek J. The self-rating of health: Methodological effects of the rating of health in health interview surveys. Maandbericht Gezondheid. 1991;10:15–23. [Google Scholar]
- [23].Brooks R, Rabin R, De Charro F. The measurement and valuation of health status using EQ-5D: a European perspective: evidence from the EuroQol BIOMED Research Programme. Springer Science & Business Media: 2013. [Google Scholar]
- [24].Roberts HC, Denison HJ, Martin HJ, et al. A review of the measurement of grip strength in clinical and epidemiological studies: towards a standardised approach. Age and ageing. 2011;40:423–429. doi: 10.1093/ageing/afr051. [DOI] [PubMed] [Google Scholar]
- [25].Hosmer DW, Jr, Lemeshow S, Sturdivant RX. Applied logistic regression. John Wiley & Sons; 2013. [Google Scholar]
- [26].Zambon S, Siviero P, Denkinger M, et al. Role of Osteoarthritis, Comorbidity, and Pain in Determining Functional Limitations in Older Populations: European Project on Osteoarthritis. Arthritis Care Res. 2016;68(6):801–810. doi: 10.1002/acr.22755. [DOI] [PubMed] [Google Scholar]
- [27].Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS medicine. 2010;7:e1000316. doi: 10.1371/journal.pmed.1000316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Luo Y, Hawkley LC, Waite LJ, et al. Loneliness, health, and mortality in old age: a national longitudinal study. Social science & medicine. 2012;74:907–914. doi: 10.1016/j.socscimed.2011.11.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Valtorta NK, Kanaan M, Gilbody S, et al. Loneliness and social isolation as risk factors for coronary heart disease and stroke: systematic review and meta-analysis of longitudinal observational studies. Heart (British Cardiac Society) 2016;102:1009–1016. doi: 10.1136/heartjnl-2015-308790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Rafnsson SB, Orrell M, d’Orsi E, et al. Loneliness, social integration, and incident dementia over 6 years: prospective findings from the English Longitudinal Study of Ageing. The Journals of Gerontology: Series B. 2017 doi: 10.1093/geronb/gbx087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Iredell H, Grenade L, Nedwetzky A, et al. Reducing social isolation amongst older people-implications for health professionals. Geriaction. 2004;22:13. [Google Scholar]
- [32].Santini ZI, Fiori KL, Feeney J, et al. Social relationships, loneliness, and mental health among older men and women in Ireland: A prospective community-based study. Journal of affective disorders. 2016;204:59–69. doi: 10.1016/j.jad.2016.06.032. [DOI] [PubMed] [Google Scholar]
- [33].Routasalo PE, Tilvis RS, Kautiainen H, et al. Effects of psychosocial group rehabilitation on social functioning, loneliness and well-being of lonely, older people: randomized controlled trial. Journal of advanced nursing. 2009;65:297–305. doi: 10.1111/j.1365-2648.2008.04837.x. [DOI] [PubMed] [Google Scholar]
- [34].Cattan M, White M, Bond J, et al. Preventing social isolation and loneliness among older people: a systematic review of health promotion interventions. Ageing & Society. 2005;25:41–67. doi: 10.7748/nop.17.1.40.s11. [DOI] [PubMed] [Google Scholar]
- [35].Galenkamp H, Huisman M, Braam AW, et al. Disease prevalence based on older people's self-reports increased, but patient–general practitioner agreement remained stable, 1992–2009. Journal of clinical epidemiology. 2014;67:773–780. doi: 10.1016/j.jclinepi.2014.02.002. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.