Article Text

Download PDFPDF

OP37 Are social engagement and loneliness differentially associated with neuro-immune markers in older age? A fixed effects analysis of time-varying associations in the english longitudinal study of ageing
  1. ES Walker1,
  2. G Ploubidis2,
  3. D Fancourt3
  1. 1Epidemiology and Public Health, University College London, London, UK
  2. 2Centre for Longitudinal Studies, University College London, London, UK
  3. 3Department of Behavioural Science and Health, University College London, London, UK


Background In recent years, associations between social isolation, loneliness and morbidity and mortality have become the focus of a great deal of cross-disciplinary research, public policy, media reporting and the wider public discourse. Some have suggested changes to our social structures have triggered a ‘loneliness epidemic’ to which older adults are particularly vulnerable. However, the mechanisms by which isolation may influence our biology and ultimately bring about disease states remains unclear. Furthermore, complex and longitudinal associations between social isolation, loneliness, socioeconomic position, health behaviours and symptoms of poor mental health further obscure our understanding of potential pathways. This study therefore aimed to explore time-varying associations between social isolation, living alone and loneliness and neuro-immune markers in older adults, whilst accounting for a comprehensive range of confounders.

Methods We analysed blood samples from 8780 adults aged 50 and above from the English Longitudinal Study of Ageing (ELSA), a nationally-representative longitudinal cohort study, across three waves of data collection: 2004/5, 2008/9 and 2012/2013. At baseline, the sample was 55% female, 66.4% married or in a partnership and 44.2% had no or basic qualifications. Multiple imputation was used to account for missing data. Fixed effects modelling was used to estimate independent relationships between loneliness, social isolation, living alone and levels of three inflammatory markers; fibrinogen, white blood cell (WBC) count and C-reactive protein (CRP), and the neuro-inflammatory regulator insulin like growth factor-1 (IGF-1).

Results ELSA participants who experienced an increase in social engagement were found to have decreased levels of the inflammatory markers fibrinogen (fixed effects coefficient: −0.007 [confidence interval: −0.015 – 0.001]), WBC (−0.012 [−0.021 – −0.003]) and CRP (−0.040 [−0.078 – −0.002]). Similarly, living status was inversely associated with fibrinogen (−0.057 [−0.097 – −0.018]), WBC (−0.098 [−0.147 – −0.048]) and CRP (−0.238 [−0.416 – −0.060]). By contrast, decreased loneliness was associated with increased IGF-1 (0.133 [0.026 – 0.240]). The findings were independent of time-invariant factors such as gender, medical history, unobserved aspects of social position, and genetics, and time-varying factors such as income, physical health, health behaviours, and depressive symptoms. The results were maintained in sensitivity analyses that accounted for BMI, gender interaction, survey weighting and exclusion of imputed values.

Conclusion Whilst causality cannot be assumed in observational studies, the results suggest that being alone and feeling alone are distinct biosocial stimuli. Assuming the methodology sufficiently accounted for confounding factors, this interpretation is especially relevant to the current social prescribing and healthy ageing movements.

  • Social Isolation
  • Biomarkers
  • Longitudinal

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.