Loneliness, social isolation and risk of cardiovascular disease in the English Longitudinal Study of Ageing

In the paper ‘Loneliness, Social Isolation and Risk of Cardiovascular Disease in the English Longitudinal Study of Ageing’, Valtorta et al.1 report the correlation between the effects of loneliness (feeling lonely) and social isolation (being lonely) on cardiovascular (CV) events in a 6-year mean follow-up. This topic is very relevant to general health, since psychosocial status and social relations are recognized as important risk factors for mortality and are becoming an important issue in modern medicine, with relevant socio-sanitary implications.2 In this paper, the authors, in a secondary analysis of prospective follow-up data from the English Longitudinal Study of Ageing (ELSA) study, analysed a very large population of 5395 community-dwelling subjects aged >50 years with repeated and separate measurements of both loneliness and social isolation using well-validated scales, and found that, after adjustment for several confounders, loneliness, but not social isolation, was prospectively correlated with the risk of developing fatal or non-fatal CV events. The issue of whether social isolation (i.e. the objective quantity of relationships with friends, family members, others) or loneliness (i.e. the subjective perception of being lonely, independently of the number of social relations) is more correlated with long-term mortality or poor health is much debated in the literature, with some studies reporting that social isolation is more important that loneliness,3 others reporting the opposite,4 and others giving the same relevance to both in terms of predictive capacity for future negative health events.5 Differences among studies regarding type of population, age and number of subjects included, instruments adopted for assessing loneliness and social isolation, and selection bias due to exclusions may explain these differences, at least in part. The strengths of the present study are that a very large number of subjects were followed longitudinally with repeated simultaneous assessment of loneliness and social isolation, and it was possible to evaluate the trend in time and the complex interrelationship of these two domains. The cause–effect relationship between loneliness and social relations is indeed very complex and difficult to disentangle, each possibly being the cause or the consequence of the other. The method adopted by the authors, with longitudinal repeated measurement of both constructs in time and with the adoption of several corrections for confounders, gives some credit to the hypothesis that loneliness (i.e. the subjective perception of being isolated or alone) is more correlated with CV outcomes than the objective number of relations. This may be due to the fact that social relationships may be helpful and supportive, but are often disagreeable and nasty, and are not always the choice of the subject but, rather, an imposition due to blood links or convenience. On the other hand, a person may have a small number of social relations, but these may be really supportive and helpful, or may have no relations at all, but as a consequence of his/her choice, and may feel happy with that and not feel lonely at all. Although this study supports the concept that the subjective feeling of loneliness may be more important than the objective number of social relations in protecting against CV events, some limitations of the study, recognized by the same authors, must temper these conclusions. These include the subject selection, which removed a very large number of non-adherent or very sick or disabled patients, thus introducing a possible selection bias by excluding in advance the worst patients, who may be more socially isolated. The population finally included in the study may represent a relatively more healthy population, able to select satisfactory not necessarily a large number of relationships, therefore minimizing the prognostic role of social isolation in the study. In this population, therefore, loneliness emerges as more highly correlated with long-term CV outcome, likely due to the often-associated poor psychological status (mostly depression and anxiety) or poor health behaviours (alcohol, smoking, poor physical activity). The authors recognize the above limitations and concede that the results of this study cannot be extended to sicker or very elderly or poorly educated populations, and that further studies are needed to fully understand the complex interrelations between loneliness and social isolation. Another possible limitation of this study is that it paid attention only to CV disease, while the correlation of social isolation or loneliness with other health outcomes, such as cognition and cancer, is similarly important. In any case, this study paves the way for a better understanding of the prognostic relevance of loneliness, with its associated poor psychological status and poor health behaviours, to CV disease events, and indicates to cardiologists the necessity of taking into account and targeting these important additional risk factors in order to ensure favourable CV health for their patients in the future. It has been shown that social factors are powerful determinants of CV outcomes in the general population,6,7 and the neuro-biological basis of loneliness, including cardiovascular, inflammatory and neuroendocrine mechanisms, may well explain its relationship with CV disease.8 Since avoiding loneliness is so important for preserving CV health, what are the actions to be taken? At population level, primary prevention strategies should be designed to screen for loneliness and promote social networks through education and counselling in several settings (primary care, circles, schools, associations). In single subjects found with loneliness, primary prevention strategies with individual or group interventions must be designed to address loneliness, the quality and quantity of social interactions, psychological status and associated poor health behaviour, which may put patients at risk independently of the traditional risk factors. Despite the relevance of the problem, however, there is no evidence at the moment that improvement in loneliness has significant health benefits at the population level. In patients already affected by CV disease, interventions on loneliness may not be efficacious, since in these patients the associated psychological status and poor health behaviour have persisted for many years. However, secondary prevention intervention should in any case include psychological counselling and health education or cognitive behavioural therapy as a relevant component of the global approach to the patient, since psychosocial risk factors are important additional risk factors in patients with CV disease.9 In this regard, cardiac rehabilitation for patients recovering from an acute event represents a valid model of multidisciplinary intervention with multiple core components, including socio-psychological assessment and counselling;10 loneliness in patients following cardiac rehabilitation is associated with poor subjective health status,11 and several studies have shown that psychological interventions within cardiac rehabilitation programmes are associated with favourable results.12,13 The present study suggests that also in the environment of cardiac rehabilitation, patient evaluation should be widened to specifically address loneliness and social isolation, particularly in the elderly. In conclusion, this study, independently of the relative importance of loneliness or social isolation as a predictor of future CV events, draws the attention of cardiologists to the importance of adding the assessment of these two components when evaluating global CV risk in both primary and secondary prevention, and of designing opportune strategies for addressing them at population and patient level to reduce the burden of CV events. The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article.